UNIVERSIDADE ESTADUAL DE CAMPINAS

FACULDADE DE ENGENHARIA AGRÍCOLA

ANÁLISE DOS PARÂMETROS RELACIONADOS AO

RESFRIAMENTO A AR FORÇADO EM EMBALAGENS PARA PRODUTOS HORTÍCOLAS

LARISSA RODRIGUES DE CASTRO

CAMPINAS

NOVEMBRO DE 2004.

UNIVERSIDADE ESTADUAL DE CAMPINAS

FACULDADE DE ENGENHARIA AGRÍCOLA

ANÁLISE DOS PARÂMETROS RELACIONADOS AO

RESFRIAMENTO A AR FORÇADO EM EMBALAGENS PARA PRODUTOS HORTÍCOLAS

Tese de Doutorado submetida à banca

examinadora para obtenção do título de

Doutor em Engenharia Agrícola, na área de

concentração de Tecnologia Pós-Colheita.

LARISSA RODRIGUES DE CASTRO Orientador: Prof. Dr. Luís Augusto Barbosa Cortez Co-orientador: Prof. Dr. Clément Vigneault

CAMPINAS

NOVEMBRO DE 2004.

FICHA CATALOGRÁFICA ELABORADA PELA BIBLIOTECA DA ÁREA DE ENGENHARIA - BAE - UNICAMP

C279a

Castro, Larissa Rodrigues de Análise dos parâmetros relacionados ao resfriamento a ar forçado em embalagens para produtos hortícolas / Larissa Rodrigues de Castro.--Campinas, SP: [s.n.], 2004. Orientadores: Luís Augusto Barbosa Cortez, Clément Vigneault. Tese (Doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Agrícola. 1. Hortigranjeiros – Resfriamento - Embalagens. 2. Energia. 3. Fluidodinâmica. I. Cortez, Luís Augusto Barbosa. II. Vigneault, Clément. III. Universidade Estadual de Campinas. Faculdade de Engenharia Agrícola. IV. Título.

ii

EPÍGRAFE

A embalagem ideal

A laranja, como embalagem, apresenta “absoluta coerência entre forma, função e consumo”.

Cada receptáculo (gomo) é embrulhado individualmente através de uma película suficiente para conter o suco e muito facilmente manipulável. Dispostos em torno de um eixo central, eles são unidos por um

“adesivo delicadíssimo”, que facilita a tarefa de “decompor o objeto em suas partes”. Para proteger a fruta, primeiro há uma camada de tecido branco que serve de acolchoamento de proteção macio. Sobre ela, a casca brilhante e dura, numa cor que não poderia ser mais adequada. A casca é dura o suficiente

para proteger a fruta mas não a ponto de dificultar o acesso a ela. Citação do designer e teórico italiano Bruno Munari nos anos 70 e publicada na Gazeta Mercantil

12.08.01.

Fonte: http://www.horticiencia.com.br/news/news2.asp?id=229

iii

AGRADECIMENTOS

Agradeço a todos que contribuíram, de alguma forma, à realização deste trabalho.

Dentre eles:

• toda minha família;

• professores Dr. Luís Augusto Barbosa Cortez e Dr. Clément Vigneault;

• Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP), em especial ao

diretor científico Dr. José Fernando Perez;

• Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq);

• professores da banca do exame de qualificação e defesa de tese: Dr. Ângelo Pedro

Jacomino, Dr. Carlos Alberto Rodrigues Anjos, Dr. Roberto Funes Abrahão, Dr. Theo

Guenter Kieckbusch, Dr. Vivaldo Silveira Júnior;

• Rosângela Parreira, da Coordenadoria de Relações Institucionais e Internacionais

(CORI);

• professores Andrés Rodríguez, Mickey Waxman e Dr. Hosahalli S. Ramaswamy;

• professores e funcionários da Faculdade de Engenharia Agrícola, UNICAMP, em

especial Pedro Luís Magna Fonte, Edson Roberto Caires, Odorico Roza De Andrade,

Ana Paula Montagner, Rosângela Gomes e Marta Ap. Rigonatto Vechi e à

coordenadora do curso de pós-graduação Dra. Raquel Gonçalves;

• pesquisadores, técnicos e secretárias do Centre de Recherche et Développement en

Horticulture, Agriculture et Agroalimentaire Canada: Bernard Goyette, Naro

Markarian Roxane, Dominique Roussel, Isabel Lemay, Ginette Gelderbloom, Jocelyne

Martineau, Sylvain Côté, Vallée Martin, Mathieu Robitaille, Thérèse Otis, Sandra

Hudson, Roger Chagnon, Dra. Marie Thérèse Charles, Dr. Gaétan Bourgeois, Dr.

Bernard Panneton e aos alunos estagiários Guillaume Gautron, Emmanuelle

Tanglinefoot e Anna Belle Blouin.

iv

SUMÁRIO

LISTA DE FIGURAS ................................................................................................................vi LISTA DE TABELAS.................................................................................................................x LISTA DE SÍMBOLOS .......................................................................................................... xiii RESUMO.................................................................................................................................xvii ABSTRACT .............................................................................................................................xix 1. INTRODUÇÃO E JUSTIFICATIVA .....................................................................................1 2. OBJETIVOS............................................................................................................................8 3. REVISÃO BIBLIOGRÁFICA................................................................................................9

3.1 Embalagens para frutas e hortaliças ...............................................................................9 3.1.1 Materiais de embalagem e acessórios......................................................................9 3.1.2 Paletização .............................................................................................................10

3.2 Proteção contra ferimentos...........................................................................................11 3.2.1 Impacto, compressão e vibração............................................................................12

3.3 Controle de temperatura ...............................................................................................12 3.3.1 Área das aberturas nas embalagens .......................................................................13 3.3.2 Geometria e disposição das aberturas, e localização do produto na embalagem ..14 3.3.3 Altura das embalagens...........................................................................................14 3.3.4 Arranjo do produto ................................................................................................15

3.4 Refrigeração .................................................................................................................15 3.4.1 Métodos de resfriamento rápido ............................................................................16 3.4.2 Tempo de meio resfriamento.................................................................................18 3.4.3 Propriedades térmicas do produto .........................................................................20

4. SEÇÕES: ARTIGOS CIENTÍFICOS SUBMETIDOS E PUBLICADOS ...........................24 Artigo 1. Container opening design for horticultural produce cooling efficiency .................32 Artigo 2. Indirect airflow distribution measurement for horticultural crop package. Part I: Produce-simulator property evaluation ..................................................................................39 Artigo 3. Indirect airflow distribution measurement for horticultural crop package. Part II: Verification of the research tool applicability........................................................................47 Artigo 4. Effect of container opening on air distribution during precooling of horticultural produce ...................................................................................................................................61 Artigo 5. Cooling performance of horticultural produce in containers with peripheral openings .................................................................................................................................68 Artigo 6. Effect of gravity on forced-air precooling ..............................................................75 Artigo 7. Effet des poignées ouvertes sur les contenants lors du prérefroidissement de produits horticoles ..................................................................................................................80 Artigo 8. Effect of container openings and airflow rate on energy required for forced-air cooling of horticultural produce.............................................................................................87

5. DISCUSSÃO DOS RESULTADOS.....................................................................................98 6. CONCLUSÕES GERAIS....................................................................................................106 7. SUGESTÕES PARA TRABALHOS FUTUROS...............................................................108 REFERÊNCIAS BIBLIOGRÁFICAS ....................................................................................109 APÊNDICES ...........................................................................................................................119

APÊNDICE A - Resumo dos resultados obtidos nos experimentos descritos nos artigos de 4 a 7 ......................................................................................................................................120

v

APÊNDICE B - Artigo 9. Laminar to turbulent indirect airflow measurement method for horticultural crop package....................................................................................................124 APÊNDICE C - Artigo 10. A new approach to measure air distribution through horticultural crop packages .......................................................................................................................126

ANEXO ...................................................................................................................................134 Method to evaluate the average temperature at the surface of a horticultural crop .............135

vi

LISTA DE FIGURAS

Figura 1. Esfera plástica preenchida com ágar-ágar usada como produto-modelo nos experimentos iniciais............................................................................................................... 24Figura 2. Vista da matriz de esferas de ágar-ágar com placas plásticas no interior do túnel de ar forçado usado como aparato experimental.................................................................... 25Figura 3. Vista geral do aparato experimental dos testes com esferas de ágar-ágar na câmara fria, mostrando túnel de ar forçado, ventilador e dispositivos de medição de pressão do ar............................................................................................................................ 25Figura 4. Esferas sólidas (bolas de bilhar ou “snooker”) instrumentadas com termopar tipo T utilizadas como produtos-modelo nos experimentos seguintes........................................... 26Figura 5. Aparato experimental mostrando em detalhe tubo circular usado para determinar as primeiras correlações entre velocidade do ar e coeficiente de resfriamento da esfera (artigo 2).................................................................................................................................. 27Figura 6. Vista do grupo de 512 esferas empilhadas de maneira colunar para simular o produto acondicionado em embalagem................................................................................... 28Figura 7. Aparato experimental com grupo de 512 esferas usado nos últimos experimentos para testar a aplicabilidade da metodologia de pesquisa desenvolvida................................... 29Artigo 1 Figura 1. Vista geral da matriz das esferas. As esferas numeradas são instrumentadas. O número de três dígitos na esfera representa as posições X, Y, Z, respectivamente. Figure 1. Overview of the matrix of balls. The numbered balls are instrumented. The three digit number on a ball represents the X, Y, and Z positions, respectively………………..… 33Figura 2. Vista superior do aparato experimental. Figure 2. Top view of the experimental set-up…………………......………………………. 33Figura 3. Tempo de meio resfriamento (HCT) em função do fluxo de ar (AFR) e da área total de abertura (TOA). Figure 3. Half-cooling time (HCT) as a function of airflow rate (AFR) and total opening area (TOA). ………………………………………………………………….……………... 35Figura 4. Tempo de meio resfriamento (HCT) em função da área de total abertura total (TOA) e do fluxo de ar (AFR). Figure 4. Half-cooling time (HCT) as a function of total opening area (TOA) and airflow rate (AFR)………………… ……………………………………………………………….. 35Artigo 2 Figura 1. Esfera instrumentada usada como simulador do produto. Figure 1. Instrumented ball used as a produce simulator..……………………………..…… 45Figura 2. Aparato experimental usado para os ensaios de calibração. Figure 2. Experimental set-up used for calibration trials. ………………………………..… 45Figura 3. Distribuição do índice de resfriamento das esferas para intervalo de ±0,0025 min-1. Figura 3. Distribution of the cooling rate index of the balls per interval of ±0.0025 min-1… 46Figura 4. Resultados do coeficiente de resfriamento obtidos nos experimentos e através de predição com a Equação 4. Figure 4. Cooling rate results from trials and from prediction with Equation 6…….……… 46

vii

Artigo 3 Figura 1. Matriz cúbica de esferas. Figure 1. Cubic matrix of balls. ……………….…………………………............................ 56Figura 2. Aparato experimental mostrando túnel de ar forçado, matriz de esferas, ventilador e dispositivos de medição das pressões dinâmica e estática. Figure 2. Experimental set up showing forced air tunnel, balls matrix, fan, and dynamic and static pressures measuring devices..……………………………………………………. 57Figura 3. Placa plástica com nove orifícios de 0,67% de área (38,6mm de diâmetro) distribuídos uniformemente. Figure 3. Plastic plate with nine holes of 0.67% as area (38.6mm of diameter) uniformly distributed. …………………………………………………………....…………………….. 57Figura 4. Resultados do HCT para as esferas 1, 21, 39 e 56, das camadas z=1, 3, 5 e 7, respectivamente. Figure 4. HCT responses for balls 1, 21, 39, and 56, from layers z=1, 3, 5, and 7, respectively............................................................................................................................. 58Figura 5. Efeito do fluxo de ar no desvio médio padrão do HCT incluindo os quatro “outliers”. Figure 5. Effect of airflow rate on the mean standard deviation of the HCT responses including the four outliers..…………………………………………………………………. 58Figura 6: Efeito do número de repetições do fluxo de ar nas respostas de HCT após serem rejeitados os quatro “outliers”. Figure 6. Mean-replication effect of airflow rate on the HCT responses after rejecting the four outliers..………………………………………………………………………………... 59Figura 7. Porcentagem da massa total de ar determinada pelos dois métodos de medição indireta da velocidade do ar (CAV e SCSV) para cada camada na direção z nos níveis de fluxo de ar mínimo e máximo testados (L.s-1.kg-1). Figure 7. Percentage of the total mass of air measured by the two indirect air velocity measuring methods (CAV and SCSV) for each z-direction layer for the minimal and maximal airflow tested (L.s-1.kg-1)..………………………………………………………… 59Figura 8. Porcentagem da massa total de ar determinada pelo método de medição indireta da velocidade do ar SCSV para cada uma das camadas na direção z em função do fluxos de ar. Figure 8. Percentage of the total mass of air measured by the SCSV indirect air velocity measuring method for each z-direction layers as a function of the different airflow rates.… 60Figura 9. Efeito do fluxo de ar nas respostas de HCT baseadas nas médias dos resultados. Figure 9. Effect of airflow rate on the HCT responses based on the result averages…….… 60Artigo 4 Figura 1. Aparato experimental mostrando túnel de ar forçado, matriz de esferas, ventilador e dispositivos de medição das pressões dinâmica e estática. Figure 1. Experimental set up showing forced air tunnel, balls matrix, fan, and dynamic and static pressures measuring devices...…………………………………………………… 67Artigo 5 Figura 1. Aparato experimental mostrando túnel de ar forçado, matriz de esferas, ventilador e dispositivos de medição das pressões dinâmica e estática. Figure 1. Experimental set up showing forced air tunnel, balls matrix, fan, and dynamic and static pressures measuring devices...........……………………………………………… 74

viii

Figura 2. Dimensões das placas #A, #B e #C que formam as oito configurações de abertura usadas para investigar o efeito da posição e da área total da abertura na uniformidade da distribuição do ar. Figure 2. Dimension of the #A, #B and #C plates used to produce the eight opening configurations to investigate the effect of the opening position and total area on the air distribution uniformity.……………………………………………………………………...

74

Artigo 6 Fig. 1. Aparato experimental mostrando túnel de ar forçado, matriz de esferas, ventilador e dispositivos de medição. Fig. 1. Experimental set up including a forced air tunnel, ball matrix, fan and measuring devices...…………………………………………………………………………………….. 78Fig. 2. Dimensões das placas das configurações diagonais de aberturas #A, #B e #C usadas para testar o efeito da gravidade. Fig. 2. Dimension of the diagonal-opening-configuration plates #A, #B and #C used to test the effect of the gravity..………………………………………………………………... 79Artigo 7 Figura 1. Aparato experimental incluindo túnel de ar forçado, matriz de esferas, ventilador e dispositivos de medição. Figure 1. Experimental set up including a forced air tunnel, ball matrix, fan and measuring devices.…………………………………………………………………………………..….. 84Figura 2. Matriz cúbica de esferas utilizada para avaliar o efeito das aberturas na circulação de ar. Figure 2. Cubic ball matrix used to evaluate the effect of opening on air circulation…….... 85Figura 3. Exemplos de placas usadas para se avaliar o efeito da alça aberto e fechado na circulação do ar em embalagens de hortícolas durante o processo de resfriamento a ar forçado; A) abertura total de 2%, B) abertura total de 16% desconsiderando a área de abertura da alça. Figure 3. Examples of plates used to evaluate the effect of open and closed handle on air circulation in horticultural crop containers during forced air precooling process; A) 2% total opening, B) 16% total opening excluding the handle opening area…………………… 85Artigo 8 Figura 1. Aparato experimental mostrando túnel de ar forçado, matriz, ventilador e dispositivos de medição das pressões dinâmica e estática. Figure 1. Experimental set up with forced air tunnel, matrix, fan, and dynamic and static pressures measuring devices....………………………………………………….………….. 95Figura 2. Configurações de abertura estudadas: quatro orifícios de 0,5% de área de abertura distribuídos nos cantos (A) e embalagem com alças e áreas de abertura de 2% (B) e de 16% (C). Figure 2. Opening configurations studied: 4-0.5%-holes distributed in corners (A), and container with handles and 2% (B) and 16% (C) opening areas....…………………………. 95Figura 3. Evolução da taxa de energia adicionada (EAR, %) de acordo com a área de abertura em cada fluxo de ar para produtos de atividade respiratória A) baixa e B) muito elevada, respectivamente. Figure 3. Energy added ratio evolution according to opening area at each airflow rate for A) low and B) very high respiration produces, respectively...……………………………… 96

ix

Figura 4. Evolução do coeficiente de eficiência energética de acordo com o fluxo de ar para produtos de atividade respiratória A) baixa e B) muito elevada, respectivamente. Figure 4. Energy efficiency coefficient evolution according to airflow for A) low and B) very high respiration rate produce.…………………………………………………………. 96Figura 5. Evolução do fluxo de ar ótimo de acordo com a atividade respiratória do produto para porcentagens de abertura de 2, 4, 8, 16 e 100%. Figure 5. Optimal airflow rate evolution according to respiration produce for opening percentages of 2, 4, 8, 16 and 100%.…...…………………………………………………... 97APÊNDICES Apêndice B - Artigo 9 Figura 1. Efeito do fluxo de ar nas respostas de HCT baseadas nas médias dos resultados. Figure 1. Effect of airflow rate on the HCT responses based on the result averages..……... 125Apêndice C - Artigo 10 Figura 1. Aparato experimental mostrando túnel de ar forçado, matriz de esferas, ventilador e dispositivos de medição de pressão. Figure 1. Experimental set up showing forced air tunnel, balls matrix, fan and pressures measuring devices..........……………………………………………………………………. 132Figura 2. Placa plástica com nove orifícios de 19,3 mm de raio distribuídos uniformemente. Figure 2. Plastic plate with nine 19.3 mm radius holes uniformly distributed…………...… 133Figura 3. Efeito do número de repetições do fluxo de ar nos valores de HCT. Figure 3. Mean-replication effect of airflow rate on the HCT responses...............………… 133Figura 4: Porcentagem da massa de ar medida através do método indireto para camada na direção z a 0,125 e 3,9 L.s-1.kg-1. Figure 4. Percentage of the mass of air measured by the indirect method for each z-direction at 0.125 and 3.9 L.s-1.kg-1..……………………………………………………….. 133Figura 5. Porcentagem da massa total de ar medida através do método indireto em função do fluxo de ar (L.s-1.kg-1). Figure 5. Percentage of the total mass of air measured by the indirect measuring method as a function of the airflow rates (L.s-1.kg-1)……………………………….………………..

133

x

LISTA DE TABELAS Artigo 1 Tabela 1. Posição relativa das esferas instrumentadas na matriz. O fluxo de ar seguiu o sentido crescente da direção do eixo Z. Table 1. Relative position of the instrumented balls in the matrix. The air was flowing in the increasing number direction of the Z axis.....……....…………………………………. 34Tabela 2. Combinações entre o número de aberturas nas direções X e Y e a área de abertura individual (IOA) para produzir da área total de abertura (TOA). Table 2. Combinations of the number of openings in X and Y directions and individual opening area (IOA) to produce total opening area (TOA).....…………………………….. 36Tabela 3. Resultados da análise estatística mostrando o nível de significância das correlações entre as variáveis independentes e: a) a queda da pressão do ar e uniformidade de resfriamento, e b) tempo de meio-resfriamento. Table 3. Results of the statistical analysis showing the level of significance of the correlations between the independent variables and: a) air pressure drop and cooling uniformity, and b) half-cooling time.....…………………………………………………... 37Artigo 2 Tabela 1. Resultados do teste de Duncan para o índice de resfriamento, calor específico, difusividade térmica e condutividade térmica dos quatro grupos de produtos-modelo. Table 1. Duncan's Multiple Range Test results for cooling rate index, specific heat, thermal diffusivity, and thermal conductivity of the four groups of produce simulators…... 45Artigo 3 Tabela 1. Posição das esferas instrumentadas na matriz de 512 esferas. Table 1. Instrumented ball positioning through the matrix of 512 balls.....………………. 54Tabela 2. Diferença mínima entre dois resultados de HCT a ser considerada como significativamente diferente para um intervalo de confiança de 95% (alfa = 0,05) para diferentes níveis do fluxo de ar. Table 2. Minimum difference between two HCT results to be considered as significantly different at a level of confidence of 95% (Alpha = 0.05) for the different airflow rate…... 55Tabela 3. Teste de Tukey mostrando o efeito do fluxo de ar nos resultados da velocidade do ar obtidos através da aplicação de cada método de medição nos experimentos com as áreas de abertura de 0,67, de 2, e de 6%. Table 3. Tukey-test showing the effect of the measured airflow rate on the air velocity results obtained from each method applied on the results of 0.67, 2, and 6% as opening areas.............………………………………………………………………………………. 55Tabela 4. Exemplos do coeficiente de determinação e dos parâmetros empíricos que relacionam o tempo de meio resfriamento (HCT) à velocidade de aproximação do ar calculada de acordo com o método SCSV de oito das esferas usadas para formar a matriz. Table 4. Examples of goodness of fit coefficient and empirical parameters relating the half-cooling time (HCT) to the air approach velocity calculated according to the SCSV method of eight of the balls used to form the matrix………………………………………. 55

xi

Tabela 5. Nível de significância da ANOVA univariada para a diferença entre o balanço de massa calculado usando os métodos CAV e SCSV para cada camada na direção z e fluxo de ar. Table 5. One-way ANOVA level of significance for the difference between the mass-balance calculated using the CAV and SCSV method for each z-direction layer and airflow rate.....……………………………..………………………………………………... 56Artigo 4 Tabela 1. Resultados do teste Tukey para a velocidade do ar e do tempo de meio resfriamento para a área de abertura, o fluxo de ar e as posições das esferas (X, Y e Z) obtidos no SPSS. Table 1. Tukey results of the air velocity and half cooling time for the opening area, airflow rate and balls positions (X, Y and Z) obtained on SPSS..………… ……………… 66Tabela 2. Resultados do teste Tukey para a heterogeneidade da distribuição do ar e queda da pressão para a área de abertura total e o fluxo de ar. Table 2. Tukey results of air distribution heterogeneity and pressure drop for total opening area and airflow rate..……………………………………………………………………….. 66Tabela 3. Efeito da área de abertura na heterogeneidade da distribuição do ar para diferentes fluxos de ar. Table 3. Effect of opening area on air distribution heterogeneity for different airflow rates. 66Artigo 5 Tabela 1. Combinações entre configurações de abertura da embalagem para os fluxos de ar de 0,125, 0,25, 0,5, 1, 2, 3,9 L.s-1.kg-1. Table 1. Combinations between container opening configurations for airflow rates of 0.125, 0.25 ,0.5, 1, 2, 3.9 L.s-1.kg-1....………………………………………………………. 73Tabela 2. Efeito da área aberta total e configurações de abertura na velocidade do ar, HCT, Vi e APD. Table 2. Effect of total opening area and opening configurations on air velocity, HCT, Vi and APD………………………….…………………………………………………….…… 73Tabela 3. Efeito do fluxo de ar na velocidade do ar, HCT, Vi e APD. Table 3. Effect of airflow rates on air velocity, HCT, Vi and APD..……………………..… 73Tabela 4. Efeito da posição da esfera na velocidade do ar e HCT. Table 4. Effect of ball positions on air velocity and HCT...........................................…....... 73Artigo 6 Tabela 1. Efeito das configurações de abertura (OC) e do fluxo de ar (L.s-1.kg-1) na heterogeneidade de distribuição do ar (Vi) e APD (mm de água). Table 1. Effect of opening configurations (OC) and airflow rates (L.s-1.kg-1) on air distribution heterogeneity (Vi) and APD (mm of water)......……………………………… 79Tabela 2. Efeito das configurações de abertura (OC) e do fluxo de ar (L.s-1.kg-1) no HCT médio (min.). Table 2. Effect of opening configurations (OC) and airflow rates (L.s-1.kg-1) on average HCT (min)..................……………………………………………………………………... 79Tabela 3. Efeito das configurações de abertura (OC) e do fluxo de ar (L.s-1.kg-1) no HCTm (min). Table 3. Effect of opening configurations (OC) and airflow rates (L.s-1.kg-1) on HCTm (min)..………………………..……………………………………………………………… 79

xii

Tabela 4. Efeito das configurações de abertura diagonal nas partes superior (A) e inferior (BO) na velocidade do ar (m•s-1) para as direções Y e Z. Table 4. Effect of the top (TO) and bottom (BO) diagonal opening configuration on air velocity (m•s-1) for the Y and Z directions...…………....…………………………………..

79

Artigo 7 Tabela 1. Efeito da porcentagem de área aberta (%), do fluxo de ar (L.s-1.kg-1) e tipo de alça na velocidade média do ar em torno das esferas (m•s-1), na heterogeneidade da distribuição do ar através do produto (Vi), no tempo máximo de meio resfriamento (HCTm, min.) e na queda da pressão do ar (PP, mm de água). Table 1. Effect of vented area percentage (%), airflow rate (L.s-1.kg-1) and handles type on mean air velocity around the balls (m•s-1), air distribution heterogeneity through the produce (Vi), average maximum half-cooling time HCTm (min), and pressure drop (PP, mm of water). ……………………………………………………………………………... 86Artigo 8 Tabela 1. Taxa de energia adicionada (EAR, %) para cada combinação entre a área de abertura (OA, %), o fluxo de ar (Dair, L.s-1.kg-1) e a atividade da respiratória (L = baixa; M = moderada; H = alta; VH = muito elevada). Table 1. Energy added ratio (EAR, %) for each combination of opening area (OA, %), airflow rate (Dair, L.s-1.kg-1) and respiration activity (L = low; M = moderate; H = high; VH = very high).……………………………………………………………………………. 94Tabela 2. Taxa de energia adicionada (EAR, %) para 2% de área de abertura distribuída uniformemente e nos cantos da embalagem para cada nível de fluxo de ar testado (Dair, L.s-1.kg-1) e atividade respiratória (L = baixa; M = moderada; H = alta; VH = muito elevada). Table 2. Energy added ratio (EAR, %) of containers with 2% opening area as a function of airflow rate (Dair, L.s-1.kg-1) and respiration activity (L = low; M = moderate; H = high; VH = very high)…………………………………………………………………………….. 94APÊNDICES Apêndice A Tabela 1. Resultados obtidos para velocidade do ar (Vel, em m.s-1), uniformidade da distribuição do ar (Vi), queda da pressão do ar (APD, em Pa), média do tempo de meio resfriamento (HCTméd, min) e máximo tempo de meio resfriamento (HCTmáx, min) para cada combinação entre configuração de abertura de embalagem e fluxo de ar....................

120

Apêndice C - Artigo 10 Tabela 1. Exemplos do coeficiente de determinação e dos parâmetros empíricos que relacionam o tempo de meio resfriamento (HCT) à velocidade de aproximação do ar calculada de acordo com o método de medição indireta para oito das esferas usadas para formar a matriz. Table 1. Examples of goodness of fit coefficient and empirical parameters relating the half-cooling time (HCT) to the air approach velocity calculated according to the indirect measuring method of eight of the balls used to form the matrix............................................ 132Tabela 2. Diferença mínima entre dois resultados de HCT a ser considerada como significativamente diferente para um intervalo de confiança de 95% (alfa = 0,05) em função do número de amostras e do fluxo de ar. Table 2. Minimum difference between two HCT results to be considered as significantly different at a level of confidence of 95% (Alpha = 0.05) as a function of the number of samples and different airflow rates………………………………………………………… 132

xiii

LISTA DE SÍMBOLOS

1. Latinos

A= área da superfície efetiva (m2)

a e b = parâmetros empíricos obtidos para as equações que correlacionam tempo de meio

resfriamento (min) e velocidade de aproximação do ar (m.min-1) para cada esfera da matriz

AAV = velocidade de aproximação do ar ao redor de cada esfera (m.s-1). Também se refere ao

fluxo de ar (L.min-1) na equação de Lambrinos e Assimaki (1997) apresentada no artigo 2

AFR = fluxo de ar (L.s-1.kg-1)

ANOVA = análise de variância (estatística)

APD = queda da pressão do ar (Pa), correspondente a PP no artigo 7

Bi = número de Biot (adimensional)

BO = aberturas na base ou parte inferior da lateral da embalagem

c ou cp = calor específico a pressão constante (J.kg-1.K-1)

cv = calor específico a volume constante (J.kg-1.K-1)

CAV = velocidade média de aproximação do ar medida indiretamente através do método

baseado na seção circular do túnel de resfriamento a ar forçado (m.s-1)

CR = coeficiente ou taxa de resfriamento do produto (s-1)

CRIb = índice de resfriamento de cada esfera (s-1 ou min-1)

CU = uniformidade do processo de resfriamento, definido no artigo 1 como o inverso do

desvio padrão do tempo de meio resfriamento do produto. CU nos demais artigos corresponde

ao coeficiente de uniformidade da distribuição da velocidade do ar através da matriz de

produtos-modelo (adimensional)

Dair = fluxo de ar (m3.s-1.kg-1)

d = metade do intervalo de confiança (min para tempos de meio resfriamento)

Ep = calor de campo do produto (J)

Er = calor de respiração do produto (J)

Et = energia térmica total removida (J)

Ev = calor liberado pelo ventilador (J)

EC = coeficiente de energia (adimensional)

EE = energia elétrica (J)

xiv

EAR = taxa da energia adicionada durante o processo de resfriamento rápido devido à energia

de respiração do produto e àquela liberada pelo ventilador em relação ao calor de campo do

produto (adimensional)

F = valor obtido na tabela de distribuição de Fisher em função do intervalo de confiança

desejado e grau de liberdade da amostra (adimensional)

Fα = ângulo que um plano forma em relação a outro (graus)

FЄ = emissividade térmica (adimensional)

fh = -(índice de resfriamento de cada esfera)-1 (s ou min)

fm,loss = fração do calor liberado pelo motor e transferido ao ar (adimensional)

FO =configuração de 100% de área aberta, isto é, sem embalagem

H = entalpia específica (J.kg-1)

h=coeficiente de transferência de calor convectivo (W.m-2.K-1)

HCT = tempo de meio resfriamento do produto (min)

HCTm = tempo de sete oitavos de resfriamento do produto que se resfria mais lentamente

(min). Corresponde ao símbolo “t” usado no artigo 8

ICR (=CRIb) = índice de resfriamento de cada esfera (s-1 ou min-1)

IOA = porcentagem de área de cada orifício (%) em relação à área das laterais da embalagem

k = condutibilidade térmica do produto (W.m-1. K -1)

L = dimensão característica do material (m)

m = massa do produto (kg)

mb = massa da esfera usada como produto-modelo (kg)

mw = massa de água (kg)

Mdx = diferença mínima entre dois resultados de tempo de meio resfriamento para serem

considerados como significativamente diferentes para um intervalo de confiança de 95% (min)

n = número repetições dos experimentos (adimensional)

NOX = número de orifícios na direção horizontal ou largura da lateral da embalagem

NOY = número de orifícios na direção vertical ou altura da lateral da embalagem

OA = porcentagem de área de cada orifício (%) em relação à área das laterais da embalagem

(corresponde a IOA no artigo 1)

PO = porcentagem de área total aberta (%) em relação à área das laterais da embalagem

(corresponde a TOA)

xv

PP = queda da pressão do ar (Pa), correspondente a APD

PXD = posição do simulador em relação ao eixo horizontal perpendicular à direção do fluxo

de ar, equivalente à largura da embalagem

PYD = posição do simulador em relação ao eixo vertical perpendicular à direção do fluxo de

ar, equivalente à altura da embalagem

PZD = posição do simulador em relação à direção principal do fluxo de ar através da matriz de

esferas, equivalente à “profundidade” da embalagem a ser percorrida pelo ar de resfriamento

Q = energia fornecida ou removida (J)

QA = fluxo de ar (L.s-1.kg-1), corresponde a AFR (artigo 1) e Dair (artigo 8)

Qair = energia total transferida (J)

Qr,produce = taxa respiratória do produto (W.kg-1)

q = taxa de transferência de calor (W)

r = raio do produto (m)

R2i = coeficiente de determinação

Re = número de Reylnolds

s2 = variância das amostras (min2 para tempos de meio resfriamento)

So = dimensão característica do material (m), equivalente ao símbolo L

sb = calor específico da esfera usada como produto-modelo (J.kg-1.K-1)

sw= calor específico da água (J.kg-1.K-1)

SCSV = velocidade média de aproximação do ar medida indiretamente através do método

baseado na seção transversal do túnel de resfriamento a ar forçado (m.s-1)

STDr = desvio padrão entre valores obtidos nas repetições dos experimentos (s ou min para

tempos de meio resfriamento)

T = temperatura do produto em qualquer ponto ao longo do resfriamento (oC)

t = tempo de operação do ventilador usado no resfriamento rápido a ar forçado ou tempo

necessário para reduzir a temperatura do produto de menor coeficiente de resfriamento até

atingir a temperatura correspondente ao sete oitavos do processo (s). Equivale a HCTm usado

no artigo 6

Ta = temperatura do meio de resfriamento (oC)

Ti = temperatura inicial do produto (oC)

Tf = temperatura final do produto (oC)

xvi

T1 e T2 = temperatura absoluta de dois materiais (oC)

tbi = temperatura inicial da esfera usada como produto-modelo (oC)

tf = temperatura de equilíbrio entre as esferas e a água segundo procedimento para

determinação do calor específico da esfera (oC)

twi = temperatura inicial da água (oC)

TO = aberturas na parte superior da lateral da embalagem

TOA = porcentagem de área total aberta (%) em relação à área das laterais da embalagem

(correspondente a PO no artigo 7)

V = volume do material (m3)

v = velocidade do fluido (m.s-1)

Vi = coeficiente de heterogeneidade da distribuição da velocidade do ar através de um meio

poroso (adimensional), equivalente a CU

X = eixo horizontal perpendicular à direção do fluxo de ar, referente à largura da embalagem

Y = eixo vertical perpendicular à direção do fluxo de ar, referente à altura da embalagem

Z = direção principal do fluxo de ar através da matriz de esferas, referente à profundidade da

embalagem a ser percorrida pelo ar de resfriamento

2. Gregos

∆T = variação de temperatura em um material ao receber ou liberar energia Q (K)

α = difusividade térmica (m2.s-1)

ηm = eficiência do ventilador (adimensional)

θ = tempo (s ou min)

ρ = massa específica ou densidade (kg.m-3)

σ = constante de Stephan-Boltzmann (σ = 5.67051.10-8 W.m-2.K-4)

υ = viscosidade cinemática do fluido (m2.s-1)

xvii

RESUMO

Esta pesquisa teve como objetivo desenvolver uma metodologia para o projeto dos

orifícios de embalagens para frutas e hortaliças submetidas ao processo de resfriamento rápido

a ar forçado. Tendo em vista as altas perdas registradas anualmente sobretudo devido à falta de

embalagens adequadas e armazenagem frigorificada, tal metodologia poderá auxiliar de forma

prática e precisa a escolha da configuração dos orifícios da embalagem para maximização da

eficiência do resfriamento rápido de produtos hortícolas. Esta ferramenta deverá ser utilizada

em combinação com demais tecnologias existentes para o projeto de uma embalagem

suficientemente atrativa ao consumidor e que atenda não apenas às necessidades do produto

submetido a tratamentos pós-colheita, como também à resistência estrutural da caixa,

possibilitando inclusive sua higienização, reutilização e desmontagem, visando reduzir custos

de confecção e transporte.

Para atingir a meta da pesquisa, os produtos hortícolas acondicionados em

embalagens foram representados por produtos-modelo nos experimentos laboratoriais. Vários

materiais e arranjos experimentais foram testados até se chegar à simulação de uma

embalagem com um grupo de esferas plásticas sólidas instrumentadas. Estas simularam com

maior precisão a distribuição do ar através de um leito de produtos hortícolas durante o

resfriamento rápido a ar forçado. A partir dos resultados obtidos para o coeficiente de

resfriamento das esferas, foram desenvolvidas correlações para determinação indireta da

velocidade de aproximação do ar ao redor de cada uma delas, considerando sua posição na

embalagem relativa ao ar de entrada. Estas correlações foram posteriormente aprimoradas

através do refinamento da pesquisa na faixa de regime de fluxo transiente. A precisão das

correlações estabelecidas foi verificada através de análises de balanço de massa nas camadas

de produto ao longo da direção do fluxo de ar.

A metodologia foi aplicada para a investigação do efeito de diferentes configurações

de abertura de embalagem na distribuição do ar através do produto submetido ao resfriamento.

Tais configurações incluíram orifícios centrais, periféricos, diagonais e uniformemente

distribuídos, além das aberturas tipo “alça” para manuseio. Também foi realizada uma análise

energética envolvendo o calor adicionado ao sistema devido à taxa respiratória do produto e ao

funcionamento do ventilador usado no resfriamento rápido. Através da ferramenta

xviii

desenvolvida, foram definidos certos valores de operação do sistema para maximizar a

eficiência do processo de resfriamento, em termos de velocidade e uniformidade de

resfriamento e energia requerida, que afetarão a qualidade e preço finais do produto. Assim,

recomenda-se o projeto de orifícios uniformemente distribuídos na superfície da embalagem

com área total aberta entre 8 e 16% . A porcentagem a ser escolhida nesta faixa dependerá dos

limites de resistência estrutural do material, fluxo de ar fornecido pelo ventilador e taxa

respiratória do produto. Por exemplo, hortícolas com atividade metabólica muito elevada,

como brócolis, acondicionados em caixas mais abertas exigirão um maior fluxo de ar para

otimização do processo de resfriamento rápido.

Palavras-chave: resfriamento rápido a ar forçado; embalagem, coeficiente de resfriamento,

uniformidade da distribuição do ar, energia.

xix

ABSTRACT

The aim of this research was to develop a methodology for designing container

openings for fruits and vegetables submitted to a forced-air precooling process. Due to

significant annual losses of fruits and vegetables, especially because of inappropriate

packaging and storage, this tool could allow practical and accurate selection of the best

package opening configuration to maximize the precooling efficiency of horticultural produce.

The tool should be combined with other technologies currently in market use to design a

container sufficiently attractive to consumers. This container must not only meet the produce

requirements when submitted to postharvest operations, but also the material structural

constraints. Furthermore, reusable and foldable containers could be desirable for manufacture

and transport cost reductions.

To this end, packed horticultural produce were represented by produce simulators in

the trials. Several materials and experimental set-ups were tested before selecting an

arrangement of instrumented solid plastic spheres in a container. These spheres simulated the

air distribution through a horticultural produce stack during forced-air precooling with more

accuracy. Correlations were established by measuring the cooling rate of the instrumented

simulators for indirect determination of the surrounding air velocity, as a function of the

simulator locations in reference to the inlet air. These correlations were further improved by

refining the airflow range studied in laminar, transient, and turbulent phases. Their

applicability was verified by performing a mass balance through the produce layers

perpendicular to the main airflow direction.

The methodology was applied to investigate the container opening design on air

distribution through horticultural produce submitted to precooling. The configurations tested

included central, peripheral, diagonal, and uniformly distributed openings, besides the

container handle openings. Furthermore, an energy analysis was performed involving heat

added to the system due to produce respiratory rate and fan functioning during precooling. The

research tool developed here allowed defining some optimum values for system operation to

maximize the cooling efficiency regarding the process rate and uniformity and required

energy, which affect produce final quality and cost.

xx

Therefore, it is recommended that design openings be uniformly distributed on

package surface with total vented area between 8 and 16%. The exact percentage to be

selected in this range will depend on the material structural resistance, airflow and produce

respiratory rates. For instance, horticultural produce with very high respiratory activity, such

as broccoli, require higher airflow rate when submitted to larger venting package to optimize

precooling process.

Keywords: forced-air precooling; package, cooling rate, air distribution uniformity, energy.

1

1. INTRODUÇÃO E JUSTIFICATIVA

No início da década de 90 o volume de frutas e hortaliças comercializado no Brasil

cresceu consideravelmente. Apesar do aumento no consumo desses produtos, as perdas anuais

na pós-colheita têm atingido de 30 a 40% da produção devido à falta de tecnologias adequadas

na colheita, embalagem e armazenagem, assim como ineficiência dos sistemas de transporte e

comercialização (FAO, 1991).

Quanto às embalagens em específico, a partir do momento em que são colhidos até

atingirem o mercado consumidor, os produtos hortícolas ficam sujeitos a uma série de efeitos

mecânicos, como vibração, impacto e compressão, que poderão provocar ferimentos

irreversíveis, facilitando a propagação de patógenos ou levando à deterioração do vegetal.

Estima-se que tais efeitos mecânicos sejam responsáveis pela perda de cerca de 25% da

produção. Também pode ser citado que o transporte, em geral, é feito nas horas mais quentes

do dia e devido à amarração das caixas há dificuldade de ventilação entre elas, o que provoca a

ocorrência de perdas por alta temperatura (BORDIN, 1998).

RODRIGUES (2000) estima que das 45 milhões de toneladas de frutas produzidas no

Brasil, 30% deixam de ser comercializadas ou simplesmente são descartadas porque não são

acondicionadas em embalagens aptas a suportar fatores como mudanças climáticas, estradas

em péssima conservação ou veículos inapropriados ao transporte de cargas perecíveis. O autor

ainda cita que os frutos precisam ser armazenados em embalagens com tratamento para barrar

a umidade (no caso das caixas de papelão), como é o caso do modelo Plaform.

Existe uma imensa variedade de materiais, tamanhos e formatos de embalagens

utilizadas para colheita, manuseio, processamento, transporte e comercialização de frutas e

hortaliças. No Brasil, atualmente vive-se um período de transição marcado pelo abandono das

caixas de madeira tradicionais (como a “K”) para vários tipos diferentes de embalagens, de

acordo com as necessidades e o uso final. Embora se observe uma grande preocupação com a

questão de embalagens para produtos hortícolas, pode-se encontrar apenas alguns super e

hipermercados com novos tipos de caixas, mas ainda em fase de experiência. De acordo com

SILVEIRA (1999), as embalagens de madeira, que o mercado brasileiro ainda insiste em

utilizar, são inadequadas às frutas e hortaliças. A popular caixa K (hortaliças e algumas frutas),

a caixa M (citros, abacaxi e mamão), a caixa Torito (banana) e o engradado (hortaliças) não

2

cumprem nenhum dos requisitos necessários à preservação da qualidade. Uma pesquisa

realizada pela CEAGESP (Companhia de Entrepostos e Armazéns Gerais de São Paulo)

revelou que a caixa K é responsável pela maior parte dos danos às frutas.

Nos Estados Unidos existem mais de 500 tipos diferentes de embalagens. Alguns

esforços de padronização vêm sendo realizados, mas com pouco sucesso, sendo que a maioria

das mudanças que ocorrem é em resposta a considerações econômicas como: uso de materiais

mais baratos, necessidade de adaptação ao acondicionamento e processamento menos

dispendiosos ou capacidade de aumentar a densidade de carga transportada (KADER, 2002).

Nos Estados Unidos e no mercado nacional existe hoje uma forte tendência na substituição do

manuseio de embalagens isoladas para o uso de paletes e também das embalagens de madeira

pelas de papelão ondulado (com uso limitado ainda das caixas plásticas, 5%).

Para se ter uma idéia da proporção entre os dois tipos de embalagens, em dezembro

de 1996, passaram pelo CEASA-Campinas 1,6 milhões de caixas e deste total 85% eram de

madeira e apenas 15% de papelão (SILVEIRA, 1997). O segundo maior fabricante de caixas

de papelão do Brasil, a empresa RIGESA Celulose S.A., tem fornecido a produtores e

empresas equipamentos de montagem de embalagens de papelão para uso nas propriedades

rurais. Também é importante citar as empresas Klabin de Papel e Celulose S.A., Igaras Papéis

e Embalagens e Grupo Orsa pelo elevado crescimento no último ano na fabricação de

embalagens de papelão ondulado (TEIXEIRA, 1999). As mudanças que vêm ocorrendo têm

forçado uma revisão geral em termos de requerimentos de embalagens para uso com produtos

hortícolas.

As embalagens de madeira utilizadas no Brasil, de acordo com BORDIN (1998), se

tratam de caixas feitas com tábuas de madeira nativa ou de reflorestamento, montadas com

prego ou grampo, sem qualquer cuidado especial com a presença de nós, pregos e aspereza da

madeira que provocam ferimentos nos frutos mais sensíveis, além de transferir odor quando

reutilizadas. Além disso, conforme SILVEIRA (1999), tais caixas são altas e pesadas

provocando a compressão e abrasão dos produtos nelas inseridos, além das quinas e rebarbas

da madeira promoverem a disseminação de diversos patógenos que aceleram o processo de

apodrecimento do produto. As caixas K, por exemplo, foram embalagens utilizadas

inicialmente durante a década de 40 para latas de querosene e posteriormente empregadas para

vários produtos hortícolas, sem qualquer modificação do projeto ou configuração da caixa para

3

melhor atender tais produtos. Já as embalagens de madeira utilizadas em países europeus

como Espanha e Itália, possuem tábuas faqueadas, sem superfícies pontiagudas e com

procedimentos de montagem para evitar ao máximo o desenvolvimento de danos mecânicos

nas frutas e hortaliças comercializadas.

Quanto às caixas plásticas, apesar de possuírem molde específico de custo elevado,

várias empresas e mesmo institutos de pesquisa e universidades norte-americanas apontam-nas

como a tendência mundial, tendo em vista as inúmeras vantagens que oferecem, desde a alta

resistência e durabilidade, até a possibilidade de higienização, desmontagem e reutilização.

Embora vários comerciantes julguem as embalagens plásticas como vantajosas apenas para

transportes entre mercados geograficamente próximos, que permitiria o seu retorno, vale

lembrar que a idéia da reutilização também se aplica entre mercados que tenham relações de

compra e venda estáveis e duradouras (mas não necessariamente próximos). Pode-se citar

como exemplos brasileiros as empresas PLASTGRUP, PISANI e MILPLAST que têm

produzido contentores plásticos recicláveis e resistentes, adequados às aplicações específicas

de cada segmento, seja de frigoríficos, bebidas, frutas ou hortaliças. Já no Canadá e Estados

Unidos pode-se citar como exemplos as empresas IPL e IFCO, que vêm desenvolvendo caixas

plásticas desmontáveis, de maior resistência mecânica e padronizadas para grupos de produtos

hortícolas. No caso da IPL, os engenheiros dos institutos de pesquisa canadenses encarregados

do desenvolvimento, têm se preocupado não apenas com os requerimentos essenciais das

embalagens (VIGNEAULT et al., 2004a; VIGNEAULT et al., 2004b; VIGNEAULT e

GOYETTE, 2002a; VIGNEAULT e GOYETTE, 2002b; VIGNEAULT e GOYETTE, 2001),

como também vêm desenvolvendo caixas com itens especiais, como dispositivos para facilitar

a limpeza (VIGNEAULT e ÉMOND, 1998).

No Brasil, começam a surgir pequenos focos de mobilização para pesquisa e

produção de embalagens mais adequadas às frutas e hortaliças, além de resistentes, recicláveis,

retornáveis e desmontáveis, em resposta às exigências do mercado, cada vez mais seletivo.

Dentre as instituições, podem se destacar as universidades, o IPT, o CETEA/ITAL e a

EMBRAPA, que vêm pesquisando novas configurações de embalagens para a comercialização

de produtos hortícolas. A EMBRAPA, por exemplo, lançou no ano de 2000 uma embalagem

plástica de dimensões 300x500x230mm desenvolvida para tomate (13 kg) e pimentão (6,5 kg)

que já está sendo comercializada pela empresa GECAL Plásticos, em Itupeva, SP. Dentre as

4

vantagens desta nova embalagem, pode-se citar sua superfície interna lisa, cantos

arredondados, altura de 23 cm (inferior à da caixa K), e com dispositivos de encaixe para

maior estabilidade no empilhamento, além de ser lavável e auto-expositora. O

acondicionamento e transporte do produto na caixa EMBRAPA representa uma redução de

17% nas perdas de tomate e 10,14% no custo de produção desta hortaliça, em função da

redução dos ferimentos mecânicos e do menor custo da embalagem em relação à caixa K,

respectivamente (VILELA e LUENGO, 2002; LUENGO et al., 2003b).

É importante também citar o trabalho que vem sendo desenvolvido pela Secretaria de

Agricultura e Abastecimento de São Paulo representada pela parceria da CODEAGRO

(Coordenadoria de Defesa da Agropecuária) e da CEAGESP: “Programa Paulista para a

Melhoria dos Padrões Comerciais e Embalagens de Hortigranjeiros” (Programa Horti&Fruti).

Segundo SILVEIRA (1999), tal programa desenvolve embalagens no sentido de substituir as

caixas K por embalagens mais adequadas ao transporte e manutenção da qualidade de frutas e

hortaliças. Os fornecedores interessados procuram a câmara setorial correspondente da

secretaria e oferecem uma nova caixa, de qualquer material, para determinado alimento. Além

de ter que possuir as dimensões adequadas a cada produto, essas caixas deverão ser aprovadas

em testes de “simuladores de viagem” em equipamentos modernos de institutos de pesquisa

como CETEA/ITAL e IPT. A partir daí, poderão ser adotadas pelos produtores e

supermercados participantes do programa.

Além de uma embalagem apropriada, é fundamental a aplicação da tecnologia do

resfriamento como um meio de prolongar a vida útil dos produtos hortícolas, que combinado à

embalagem permite conservar as características desejáveis para comercialização por um

período maior de tempo, e reduzir as perdas. Segundo ASHRAE (2002), os produtos

hortícolas são em geral colhidos a uma temperatura superior à recomendada para

armazenagem. Deste modo, é necessário que passem por um processo de resfriamento rápido,

onde é retirado o “calor de campo” e de estocagem, dificultando, portanto, o surgimento,

desenvolvimento e propagação de microrganismos, reduzindo, conseqüentemente, a atividade

metabólica e prolongando o tempo de conservação do produto.

Segundo HARDENBURG et al. (1986), a temperatura influencia a respiração e

transpiração dos vegetais, sendo que níveis mais baixos reduzem a velocidade do processo de

maturação, minimizando também a ação de agentes deteriorantes e a perda de peso do produto.

5

Para uma certa faixa de temperatura, quanto mais baixa for esta, menor será a atividade

respiratória do vegetal. VANEGAS (1987) cita que cada 10°C de aumento na temperatura

pode provocar uma atividade respiratória até sete vezes maior para uma faixa de 0-10oC, mas

se este aumento ocorrer quando a temperatura está entre 10 e 20oC, a atividade será apenas

dobrada ou triplicada.

No mercado brasileiro se observa pouca aplicação da refrigeração de produtos

hortícolas, principalmente do resfriamento rápido, exceto para maçã, mamão, pêssego, goiaba,

melão e caju. A incorporação de novas tecnologias na cadeia produção-comercialização, como

a instalação de câmaras frigoríficas, inclusive com atmosfera controlada, possibilitou

dinamismo no setor de maçãs, resultando em um salto das exportações de US$ 1,7 milhão em

1991 para US$ 20,9 milhões em 1992 (NEVES FILHO et al., 1997).

De acordo com ASHRAE (2002), as principais formas de resfriamento rápido são: à

água gelada; com uma mistura de água e gelo (gelo líquido); a ar forçado e a vácuo. A escolha

do método de resfriamento a ser utilizado é determinada pelas características do produto, como

forma, área da superfície, massa, condições de resistência da casca, dentre outros, além dos

fatores econômicos, conveniência, relação do equipamento de resfriamento com a operação de

embalagem total, e preferência pessoal.

Dentre os métodos citados destaca-se o ar forçado realizado em túneis de

resfriamento e fortemente recomendado para a maioria dos produtos do Programa Horti&Fruti

previamente mencionado. Neste caso, a embalagem utilizada tem papel fundamental na

eficiência da remoção do calor do produto. No resfriamento rápido a ar forçado, ASHRAE

(2002) cita que o tempo de resfriamento vai depender das dimensões das embalagens e área de

suas aberturas, formas com que são distribuídas e da característica do produto, temperatura,

umidade e velocidade do ar através do mesmo. Este método de resfriamento rápido envolve

padrões de empilhamento de caixas definidos onde o ar é forçado através, mais que ao redor,

de caixas individuais. Um uso bem sucedido de tal método requer embalagens com orifícios

localizados na direção em que o ar se moverá e com um mínimo de materiais que poderiam

impedir o fluxo livre do ar.

Conforme DINCER e GENCELI (1994), para se projetar um sistema de resfriamento

eficiente e efetivo para produtos hortícolas, é necessário fazer uma análise precisa da

transferência de calor transiente durante o processo. Segundo os mesmos, é necessário também

6

ter conhecimento das variáveis que afetam a taxa de resfriamento ou a velocidade de redução

da temperatura do vegetal, tais como as dimensões do produto, forma e propriedades térmicas,

condições do resfriamento e configuração do produto durante o processo. As interações entre

algumas dessas variáveis devem ser consideradas e a transferência de calor deve ser analisada

para determinar as condições ótimas do processo que podem resultar em um custo mínimo de

resfriamento para situações específicas, através do uso otimizado de energia. A determinação

de parâmetros do processo de resfriamento, tais como coeficiente de resfriamento, tempo de

meio e sete oitavos de resfriamento e parâmetros de transferência de calor, como

condutibilidade térmica, difusividade térmica, calor específico e coeficiente de transferência

de calor são importantes para o projeto adequado e operação de tais sistemas.

Em geral, tanto técnicas analíticas como numéricas têm sido utilizadas para analisar a

distribuição do ar e transferência de calor durante o processo de resfriamento rápido. No

entanto, o que dificulta seu cálculo é o fato do processo ser influenciado pelas características

do produto resfriado, sua configuração dentro da embalagem, a área, formato e disposição das

aberturas das caixas, distribuição das embalagens no túnel de resfriamento, dentre outros

fatores. Desta maneira, não existem na literatura, valores tabelados para as variáveis

envolvidas no processo de resfriamento rápido, como por exemplo taxa e uniformidade do

processo e consumo energético, para cada situação específica de resfriamento e embalagem.

No caso da embalagem por exemplo, a redução da área aberta acarreta o aumento da

queda da pressão do ar ao resfriar produtos hortícolas nela acondicionados, e portanto da

energia consumida pelo ventilador. A redução desta área pode também prejudicar a qualidade

final do produto em função da menor uniformidade e velocidade de distribuição do ar de

resfriamento. Por outro lado, o aumento da área aberta pode comprometer o suporte estrutural

da embalagem gerando também perdas por ferimentos mecânicos no produto. Nesse caso a

economia de material obtida com o acréscimo da área aberta, pode ser perdida com o aumento

da espessura necessário para reforço estrutural das laterais da embalagem.

Assim, a principal proposta desta pesquisa foi desenvolver uma metodologia prática e

precisa para determinar o efeito da configuração das aberturas da embalagem nos parâmetros

envolvidos no processo de resfriamento rápido a ar forçado de produtos hortícolas. Desta

forma, esta metodologia pôde ser aplicada para a otimização do projeto da área e disposição de

7

aberturas da embalagem e seleção da faixa de operação do sistema de ar forçado visando a

maximização da eficiência do processo de resfriamento rápido.

8

2. OBJETIVOS

O objetivo principal desta pesquisa foi analisar parâmetros relacionados à eficiência

de sistemas de resfriamento a ar forçado para produtos hortícolas, como fluxo de ar e orifícios

de embalagens. Para alcançar tal objetivo, propôs-se realizar os seguintes objetivos

específicos:

• determinar o melhor material e arranjo experimental para simular a distribuição do ar

através do leito de produtos hortícolas esféricos;

• desenvolver e avaliar uma forma indireta para determinação da velocidade de

aproximação do ar ao redor dos produtos no interior de embalagens, estabelecendo sua

correlação com parâmetros mais fácil e precisamente mensuráveis;

• aplicar a metodologia para determinar a distribuição do ar em diferentes configurações

de aberturas de embalagens, variando tanto área e número de orifícios como também

sua disposição nas laterais da caixa para diferentes níveis de fluxo de ar;

• determinar os efeitos positivos e negativos das configurações de embalagens testadas,

identificando o projeto mais favorável ao resfriamento rápido e uniforme para produtos

de diferentes atividades respiratórias, e verificar seu impacto no consumo energético

do sistema.

Este estudo também visou estabelecer, através da aplicação da metodologia

desenvolvida, algumas recomendações quanto à faixa de operação do sistema de ar forçado e

porcentagem máxima de abertura necessária da embalagem, para não comprometer a

resistência estrutural da embalagem e ainda promover maior uniformidade, rapidez e menor

restrição ao fluxo de ar durante o resfriamento do produto.

9

3. REVISÃO BIBLIOGRÁFICA

3.1 Embalagens para frutas e hortaliças

De nada adianta a utilização de moderna tecnologia agrícola para o aumento da

produção de alimentos se não existirem meios de conservar a qualidade de tais alimentos,

garantindo que sejam aproveitados pelo homem. Neste sentido, conforme KADER (2002), as

embalagens são unidades importantes para a comercialização e distribuição de produtos

hortícolas, possuindo muitos requerimentos especiais. Elas devem proteger o produto contra

ferimentos durante a distribuição e devem manter sua forma e tensão, freqüentemente por

longos períodos a uma umidade relativa próxima da saturação. Muitas devem ser projetadas

para facilitar o resfriamento rápido dos frutos das temperaturas altas de campo até baixas

temperaturas de armazenagem ou transporte, permitindo também a remoção contínua do calor

produzido pelo produto. Além de todos estes requisitos, é importante que as embalagens sejam

atrativas ao consumidor.

SINGH et al. (1992) mostram que a operação de embalagem é um dos fatores que

mais influenciam a qualidade da fruta “in natura” durante o transporte e distribuição, até

atingir o consumidor. Segundo BORDIN (1998), a importância das embalagens reside no fato

de que são as responsáveis por conter e proteger o produto hortícola contra as adversidades do

meio de distribuição, de modo a tornar mais conveniente e eficaz o seu manuseio e

comercialização. Outro fator importante a ser considerado é o custo que tais materiais

representam nas despesas gerais do comerciante.

3.1.1 Materiais de embalagem e acessórios

Para BORDIN (1998), as frutas e hortaliças em geral são acondicionadas em caixas

de diversos materiais e formas, podendo também ser ensacadas de acordo com sua resistência.

Para a escolha do tipo e material a ser utilizado, deve-se levar em conta fatores como:

resistência, disponibilidade, custo, adequação a tratamentos pós-colheita (como resfriamento)

e principalmente atender às necessidades do produto. A embalagem está relacionada à

classificação, método de resfriamento e armazenagem do produto, assim como com o sistema

de manipulação da mesma, como por exemplo, o sistema paletizado. Deste modo, são

10

utilizadas principalmente embalagens plásticas e de materiais celulósicos, como madeira,

papelão ondulado e cartão.

SOARES et al. (1993), estudaram a influência da utilização de embalagens de

madeira (caixas K) na conservação da qualidade do tomate. Os autores observaram sérios

problemas de dano mecânico nos frutos, cuja intensidade variou em função da posição do fruto

dentro da caixa.

BORDIN (1998) cita que em função das necessidades de cada produto hortícola,

podem ser inclusos nas embalagens alguns acessórios como: papéis para proteção individual

de frutos; luvas de espuma plástica para evitar ferimentos na ocorrência de choques; bandejas

divisórias de polpa de celulose moldada ou chapas plásticas termo-formadas, contendo

alojamento individual para cada fruto; dentre outros.

3.1.2 Paletização

Segundo BORDIN (1998) todas as embalagens de produtos hortícolas devem ser

projetadas para uma paletização segura, com caixas em empilhamento colunar, cruzado ou

misto. A paletização é fundamental na agilização das operações de carregamento e

descarregamento, otimizando o transporte dos produtos hortícolas, e economizando espaços de

armazenagem. Ao serem paletizadas, as embalagens são empilhadas sobre o palete, geralmente

de madeira, podendo ser descartável ou reutilizável, formando uma unidade. O empilhamento

influencia a estabilidade do conjunto e também deve ser adequado ao movimento do ar,

removendo o calor de respiração e facilitando o resfriamento rápido da carga de produto. Se as

embalagens são dispostas em empilhamento cruzado, seus orifícios devem coincidir nas

arestas que serão unidas permitindo a circulação de ar através de todas as caixas no palete. Tal

condição requer uma relação geométrica entre as dimensões horizontais da caixa e as

localizações dos orifícios nas laterais e cantos. É importante também que as dimensões das

caixas sejam compatíveis com aquelas dos paletes (1,00 x 1,20m, 1,00 x 1,00m, dentre outras).

Alguns valores recomendados para as dimensões externas e capacidade da embalagem para

cada hortícola podem ser encontrados na nova padronização americana proposta pelo MUM

(Modularization, Unitization, and Mechanization).

11

3.2 Proteção contra ferimentos

O acondicionamento adequado de frutas e hortaliças, plantas e flores de corte, é

essencial para a manutenção da qualidade do produto durante seu transporte e

comercialização. Além de proteção, a operação de embalagem e a paletização servem para

conter o produto em unidades de forma que possa ser manuseado. Não tem sentido transportar

produtos perecíveis de alta qualidade e preço em embalagens de baixa qualidade que

ocasionarão ferimentos, deterioração e preços baixos ou completa rejeição dos produtos pelo

consumidor. As embalagens devem suportar: o manuseio durante o carregamento e

descarregamento, a compressão do peso de outras caixas acima, impacto e vibração durante o

transporte e a alta umidade durante o resfriamento rápido, transporte e armazenagem

(MCGREGOR, 1987).

Conforme ASHRAE (2002), os cuidados pós-colheita envolvem o empilhamento das

caixas com produto de maneira adequada à ventilação e refrigeração. As caixas não devem ser

muito profundas para evitar que o peso excessivo danifique o produto próximo à base.

LUENGO et al. (2003a), por exemplo, estimaram a altura máxima da embalagem para

diversos produtos hortícolas para que a carga de compressão e acelerações de impacto durante

transporte não produzam uma deformação superior a 5% no produto.

O efeito provocado pelo manejo envolvendo embalagens de alta rugosidade, e

portanto abrasivas, é cumulativo. Vários ferimentos pequenos em um tomate podem produzir

perda de sabor. Danos mecânicos provocam perda de umidade.

De acordo com SARGENT (1998), após a colheita, as hortaliças perdem firmeza

devido aos processos de desenvolvimento, senescência e perda de água. Como resultado,

tornam-se mais susceptíveis aos danos mecânicos.

SOARES et al. (1993) afirmam que um outro inconveniente é o fato dos impactos

estimularem um aumento na taxa de respiração e na produção de etileno, acelerando o

amadurecimento e conseqüentemente reduzindo a vida útil do produto. Também descrevem

que os ferimentos enfraquecem ou destroem as defesas naturais do vegetal, favorecendo o

desenvolvimento de patógenos oportunistas, que provocam infecção, podendo se proliferar e

contaminar frutos próximos e saudáveis.

Para SARGENT et al. (1992), os hortícolas transportados a longas distâncias estão,

freqüentemente, sob condições que podem promover o crescimento de organismos

12

deteriorantes nos locais de danos mecânicos como ferimentos por impacto e compressão,

cortes e abrasões. Sendo os danos mecânicas cumulativos, os vários passos do manuseio, do

campo ao consumidor, devem ser cuidadosamente coordenados e integrados para minimizar as

perdas na qualidade do produto. Embora ferimentos internos não sejam normalmente

detectáveis em “packing houses”, podem ser percebidos pelo consumidor e sua presença

poderá comprometer compras subseqüentes.

3.2.1 Impacto, compressão e vibração

BORDIN (1998) cita que os danos por impacto são provocados na queda do produto

em superfícies rígidas, individualmente ou dentro de embalagens; tal ferimento pode não ser

visível na superfície do vegetal. Acolchoamentos nos pontos de queda e projeto de materiais

de preenchimento que elevam as caixas vazias para reduzir as alturas de queda durante a

operação de embalagem podem diminuir a incidência e severidade dos danos por impacto.

Acessórios acolchoados no fundo das caixas podem prover proteção adicional.

Os ferimentos de compressão se originam de operações de embalagem impróprias ou

da utililização de embalagens inadequadas. As caixas devem ser capazes de suportar o peso

especificado de caixas adicionais e não se deve empilhar um número de embalagens além do

limite definido.

Os danos de vibração ou abrasão ocorrem quando os produtos se movem dentro da

embalagem durante algum tipo de movimento (BARTSCH et al., 1987; MCLAUGHLING,

1987; MCLAUGHLING e PITT, 1984). Para prevenir tais ferimentos deve-se imobilizar o

produto dentro da embalagem através de correto dimensionamento da mesma para as

dimensões e quantidade de produtos e de acordo com sua densidade. Alguns materiais

adicionais como bandejas, copos, “capas” e “enchimentos” podem ser úteis.

3.3 Controle de temperatura

Devem ser considerados igualmente a natureza do produto contido na embalagem e o

tratamento a que será submetido após o acondicionamento na caixa (como resfriamento,

tratamento com etileno, CO2, dentre outros). O controle da temperatura na armazenagem

depende de um bom contato entre o produto na embalagem e o ambiente externo (KADER,

2002). Em todas as frutas e hortaliças embrulhadas individualmente e inseridas em caixas de

13

madeira, papelão, cestas, sacos ou no caso de produtos que são colocados em embalagens que

não contêm orifícios, como vagens, ervilhas, frutas pequenas e hortaliças folhosas que

naturalmente são acondicionadas bem unidas/apertadas, o calor é removido apenas pela

condução da superfície da embalagem. Nesse caso, a espessura de cada face de tais

embalagens vai depender da atividade respiratória do vegetal. Se essa espessura for maior que

a crítica, os produtos do centro da embalagem vão se aquecer pois o calor de respiração não foi

dissipado rápido suficientemente. Na prática, esse calor se torna significativo para produtos

que respiram intensamente como ervilha, feijão, alface, brócolis em caixas grandes ou pilhas

de embalagens e envolvidas por acessórios. Tal problema pode ser evitado utilizando-se

embalagens menores ou através de ventilação em grandes caixas e pilhas (KADER, 2002).

3.3.1 Área das aberturas nas embalagens

A ventilação é geralmente necessária para que o fluxo de ar que atravessa a

embalagem remova o calor rapidamente (VIGNEAULT e ÉMOND, 1998). Dentro de

determinados limites, aumentando as dimensões das aberturas da embalagem ocorre uma

mudança na velocidade do fluxo de calor. Para caixas de papelão ondulado, 5% de orifícios

nas arestas permite um resfriamento rápido sem comprometer a resistência mecânica da

embalagem. Segundo KADER (2002) poucos orifícios grandes são mais eficazes que muitos

pequenos mas tal afirmação considera o aspecto estrutural das embalagens e não a questão de

eficiência energética. Aberturas verticais a pelo menos 5 cm dos cantos das embalagens têm

melhor desempenho. O efeito desses orifícios durante o transporte depende do padrão de

carregamento que determinará se o ar frio atingirá as aberturas (KADER, 2002). As

recomendações apresentadas por KADER (2002) não coincidem com a opinião de outros

autores. VIGNEAULT e ÉMOND (1998), por exemplo, recomendam uma superfície de

abertura total consideravelmente superior, 25%, assim como caixas menos espessas, de 9 mm.

Tais discrepâncias observadas nas recomendações dos diferentes autores requerem uma

pesquisa mais aprofundada para elucidação do assunto.

É importante que os orifícios não sejam obstruídos por acessórios internos, como

bandejas, plásticos, papéis, linhas, etc. Se tais materiais que restringem a passagem do ar são

essenciais para o produto, então o fluxo de ar deve ser adequadamente elevado para compensar

seus efeitos (KADER, 2002).

14

Quanto maior a área disponível para a entrada do ar de resfriamento na embalagem,

menor será a energia necessária para operar o ventilador e sistema frigorífico. Estudos

mostram que abaixo de 3% de abertura em caixas de papelão, o custo de resfriamento aumenta

significativamente. Uma caixa de papelão com 5% de abertura, submetida a uma velocidade

do ar de 0,25m/s requereu 3,75h, em média, para ser resfriada, com custo equivalente a US$

0,13/kg de produto, enquanto que uma caixa com 2% de área de abertura, à mesma velocidade

do ar, precisou de 4,65h para o resfriamento e o custo foi de US$ 0,18/kg de produto. No caso

extremo da caixa apresentar somente 1% de área de abertura, o tempo de resfriamento passou

para 6,6h e o custo foi elevado para US$ 0,43/kg (BAIRD et al., 1988).

3.3.2 Geometria e disposição das aberturas, e localização do produto na

embalagem

Além da área de abertura, a geometria e disposição dos orifícios na embalagem e a

localização do produto dentro da caixa com relação ao ar de entrada, também influenciam a

velocidade de resfriamento das frutas e hortaliças. No caso de caixas que contenham apenas

aberturas próximas à parte superior da embalagem, o ar forçado para resfriamento da carga

tende a percorrer um caminho preferencial formado entre a parte superior da caixa e o fundo

da embalagem imediatamente acima, no palete, não passando portanto pela maioria do

produto. Por isso, os orifícios devem ser projetados para promover a melhor distribuição do ar

dentro das caixas (ARIFIN e CHAU, 1987). Quanto à geometria das aberturas, aquelas com

forma elíptica conferem melhores características de resistência mecânica à embalagem que as

circulares, quadradas ou retangulares. No entanto, VIGNEAULT e GOYETTE (2002)

demonstraram que a queda da pressão do ar através do produto acondicionado em embalagem

é mais afetada pela porcentagem de abertura que pela forma dos orifícios nas laterais da caixa.

3.3.3 Altura das embalagens

A altura das embalagens (e conseqüentemente o número de camadas de produto no

seu interior) e a altura dos paletes influenciam a circulação do ar (forçada ou não). Além

desses fatores, as dimensões do produto e a forma como está disposto dentro da embalagem

também afetam o tempo necessário para a remoção do calor e resfriamento do mesmo. A

eficiência da circulação do ar é reduzida conforme a altura do leito de produto aumenta,

15

acarretando um incremento tanto no tempo de resfriamento como na perda de pressão (BAIRD

et al., 1988).

3.3.4 Arranjo do produto

CHAU et al. (1985) e VIGNEAULT et al. (2004a) mostram que a disposição do

produto dentro da embalagem tem maior influência na resistência ao fluxo de ar que as

dimensões da fruta, por exemplo no caso de laranjas. Quanto maior a porosidade do arranjo,

menor a resistência ao fluxo de ar, independente do diâmetro do produto. Os arranjos

estudados pelos autores constituíam: randômico, em que as frutas foram dispostas de forma

randômica; reto, onde as laranjas foram empilhadas diretamente umas sobre as outras,

resultando em colunas retas; quadrado, em que cada quatro frutas adjacentes formavam um

quadrado e cada fruta da segunda camada se apoiava nos quatro produtos da camada anterior;

e alternado, que consiste em dispor a fruta em cada camada de forma que cada três frutas

adjacentes formem um triângulo eqüilátero e cada produto da próxima camada se apóie em

três frutos da camada anterior. A classificação dos arranjos, em ordem decrescente de

resistência ao fluxo de ar foi a seguinte: arranjo alternado>arranjo quadrado>arranjo

randômico>arranjo reto.

3.4 Refrigeração

Segundo ASHRAE (2002), após a colheita, os vegetais mais perecíveis devem ser

removidos do campo tão rápido quanto possível e colocados sob refrigeração, ou devem ser

classificados e embalados para comercialização. Tendo-se em vista que os processos de

envelhecimento e amadurecimento dos vegetais continuam após a colheita, a vida útil depende

fortemente da temperatura e cuidado no manuseio físico. A manutenção da qualidade é

auxiliada pelos seguintes processos (KADER, 2002):

• colheita em maturidade ou qualidade ótima;

• manuseio cuidadoso, para evitar danos mecânicos, e rápido, para minimizar a

deterioração;

• utilização de embalagens adequadas de proteção e acondicionamento;

• uso de tratamentos químicos, calor ou atmosfera modificada para preservação;

• emprego de procedimentos sanitários;

16

• resfriamento rápido para remover o “calor de campo”;

• alta umidade relativa na estocagem frigorífica para minimizar as perdas de peso

do produto;

• refrigeração apropriada durante todo o processo de comercialização.

Segundo RAGHAVAN et al. (2003), KADER (2002) e GHAOUTH et al. (1992), um

método que complementa a refrigeração, mas não a substitui, é a atmosfera modificada para

conservação da produção, retardando a senescência, reduzindo a sensibilidade do fruto à ação

do etileno, aliviando algumas desordens fisiológicas resultantes da injúria do frio e inibindo a

incidência de organismos patógenos. As atmosferas modificadas correspondem à remoção ou

adição de gases resultando em uma composição atmosférica diferente daquela do ar

(CARVALHO FILHO et al., 2003). Em geral envolve a redução das concentrações de

oxigênio e/ou elevação do dióxido de carbono. A modificação atmosférica em torno do

produto pode ser obtida, por exemplo, através do uso de filmes plásticos e com cobertura de

cera (MARKARIAN et al., 2003).

3.4.1 Métodos de resfriamento rápido

Segundo ASHRAE (2002), o resfriamento rápido dos produtos após a colheita, antes

ou depois de serem embalados, e antes de serem armazenados ou transportados, previne a

deterioração dos vegetais mais perecíveis. Quanto mais rápido o “calor de campo” é removido

após a colheita, por mais tempo o produto pode ser mantido em boas condições de

comercialização. O processo de resfriamento reduz a velocidade da deterioração natural

(amadurecimento e senescência) e do crescimento dos organismos deteriorantes (e deste modo

o desenvolvimento da podridão/putrefação), reduzindo também o murchamento visto que a

perda de água ocorre muito mais lentamente em temperaturas baixas que em altas.

CORTEZ et al (2002b) cita que o resfriamento rápido pode ser realizado com água

gelada; com uma mistura de água e gelo (gelo líquido); a ar forçado e a vácuo. Apesar de

muito utilizado, CORTEZ et al. (2002a) mostram que o “room cooling” não é um método de

resfriamento rápido. Consiste apenas em colocar o produto na câmara e deixar que este se

resfrie naturalmente, sem a circulação forçada do ar, o que pode gerar um tempo de

resfriamento de até dias. Para KADER (2002), o uso mais comum do “room cooling” se aplica

17

a produtos com vida útil relativamente longa que são armazenados na mesma sala em que são

resfriados.

Em relação ao método do ar forçado, ASHRAE (2002) descreve que, devido às

características físicas de frutas e hortaliças, principalmente geometria, várias delas respondem

diferentemente a tratamentos semelhantes de fluxo e temperatura do ar. Um exemplo disto é o

caso de pêssegos que se resfriam mais rapidamente que batatas quando são resfriados em

embalagens sob condições similares de fluxo e temperatura do ar.

Mitchell et al. (1972) citados por ASHRAE (2002), observaram que a operação do

resfriamento a ar forçado geralmente resfria o produto em 0,25 a 0,1 do tempo necessário para

o convencional “room cooling” mas que ainda gasta 2 a 3 vezes mais tempo que o

resfriamento a água ou a vácuo. Por exemplo, estudos reportam um tempo de meio

resfriamento de uma hora para o resfriamento a ar forçado de pêssegos comparado a seis horas

para pêssegos embalados de modo semelhante em convencional “room cooling”.

Após o resfriamento, o produto deve ser mantido refrigerado às temperaturas e níveis

de umidade recomendados. A temperatura de armazenagem depende do produto, sua cultivar,

estado de maturidade, dentre outros. Quanto à umidade, esta afeta a perda de peso do produto

que, além de poder provocar uma deterioração da qualidade e reduzir sua possibilidade de

comercialização, tendo em vista a formação de enrugamento, murchamento e perda de

firmeza, também representa uma perda direta de peso comercializável (SHOWALTER, 1981).

O autor reporta que a susceptibilidade das hortaliças à perda de água varia significativamente,

mostrando resultados como a perda diária de 0,9% para tomates e de 2,5% para pepinos

mantendo-se cerca de 27oC e 60% de umidade relativa durante a armazenagem. Tendo em

vista que essas perdas são multiplicadas várias vezes em função da duração do período de

comercialização do produto, torna-se importante tentar reduzi-las através de resfriamento,

manutenção de alta umidade durante a armazenagem, redução do movimento do ar, uso de

embalagens protetoras e ceras, dentre outros.

Se ocorrer o re-aquecimento, quebrando a “cadeia do frio” muitos dos benefícios

obtidos no resfriamento rápido imediato podem ser perdidos (KADER, 2002; CORTEZ et al.,

2002b). Um dos problemas certamente encontrados é a condensação de água na superfície do

produto ao sair da cadeia do frio. Essa umidade na superfície associada à elevação da

18

temperatura pode acelerar a atividade de microrganismos e conseqüentemente a deterioração

do produto.

3.4.2 Tempo de meio resfriamento

Uma grande preocupação é determinar o tempo necessário para um “resfriamento

completo” que em geral corresponde ao tempo requerido para atingir uma temperatura

desejada antes de transferir o produto à estocagem ou transporte. De acordo com KADER

(2002), nas comparações entre os métodos, utiliza-se os conceitos “tempo de meio

resfriamento” ou “7/8 de resfriamento”. O tempo de meio resfriamento consiste no tempo

necessário para resfriar o produto até a temperatura média entre a temperatura inicial (Ti) e a

do meio de resfriamento (Ta). Em outras palavras, refere-se ao tempo em que a diferença da

temperatura entre o produto e o meio ao seu redor se torna igual à metade da diferença inicial

de temperatura, isto é: (T - Ta) = 1/2 (Ti - Ta) onde T é a temperatura do produto em qualquer

ponto ao longo do resfriamento.

Este é um valor constante para um determinado sistema e a velocidade de

resfriamento pode parecer diminuir enquanto se dá o processo. Por exemplo, se uma carga de

produto em uma câmara com ar a 0°C gasta 4 horas para resfriar da temperatura de 20°C até

10°C (tempo de meio resfriamento), então serão necessárias outras 4 horas adicionais para

resfriar a carga até 5°C e mais 4 horas para atingir cerca de 2,5°C, e assim por diante.

Freqüentemente se utiliza o termo 7/8 de resfriamento (três “tempos de meio resfriamento”,

neste exemplo valendo 12 horas) como um ponto de referência. Sete oitavos de resfriamento

poderia ser definido como o tempo necessário para resfriar o produto 7/8 da diferença entre a

temperatura inicial e a temperatura da média do resfriamento.

Assim, segundo ASHRAE (2002) após um “atraso” inicial, a temperatura do centro

do produto diminui exponencialmente e o tempo do processo de resfriamento dependerá do

quociente entre as resistências à transferência de calor externa e interna do produto. Este valor

é definido como número de Biot ou Bi=hL/k, onde h=coeficiente de transferência de calor

convectivo, L=dimensão característica do material e k=condutibilidade térmica do produto.

Tais propriedades serão descritas em maiores detalhes ao longo deste capítulo. Assim, por

exemplo, Bi>40 indica que a resistência à condução de calor no produto é muito superior à

gerada pela convecção (h superior à k/L) e portanto a temperatura da superfície do produto

19

poderia ser assumida igual à do meio (câmara fria). Os produtos hortícolas apresentam

0,2<Bi<10, e nesse caso o tempo de resfriamento dependerá, dentre outros fatores, da

condutividade térmica do produto, velocidade do ar, e demais parâmetros que influenciam o

movimento e contato do fluido no produto, como formato e arranjo em embalagens

(MOHSENIN, 1980).

O tempo de meio resfriamento pode ser determinado a partir do coeficiente ou taxa

de resfriamento (CR), que por sua vez é obtido pela equação CR=h.A/(c.ρ.V), a partir dos

dados ou simplesmente a partir dos dados da relação tempo-temperatura conforme

MOHSENIN (1980) e ASHRAE (2002):

a) através do cálculo da inclinação da curva de resfriamento se o quociente entre

temperaturas (Ti –Ta / T - Ta) definido anteriormente for traçado na escala logarítmica do eixo

vertical e o tempo θ na escala aritmética do eixo horizontal de um papel semi-log. Se a

mudança de temperatura seguir a forma exponencial, uma linha reta deve ser obtida através da

interceptação no eixo y na unidade. A inclinação da linha reta será então o coeficiente de

resfriamento, determinado pela equação CR = ln (Ti –Ta / T - Ta) / θ (vide GOYETTE et al.,

1996 no Anexo). Na prática (Ti –Ta / T - Ta) é traçada em função do tempo θ para fornecer

uma curva de resfriamento inclinada para baixo à direita, que pode ser vista reescrevendo a

última equação: -CR = ln (T - Ta / Ti –Ta ) / θ ou (T - Ta / Ti –Ta ) = e-CR(θ);

b) pelo logaritmo da diferença de temperatura média entre o produto e o ambiente:

)T(T)T(Tln)T(T)T(T

θT)(TCR

aai

aai

i

−−−−−

−= que se refere à redução da temperatura em um certo período do

resfriamento dividido pelo logaritmo da diferença de temperatura média do produto e do meio;

c) através da relação médiaa

i

)T(TθT)(TCR

−−

= onde (T – Ta)média equivale à diferença de

temperatura média para o período de tempo θ, que pode ser determinada através da área sob a

curva tempo-temperatura dividida pelo tempo de resfriamento correspondente. Como o

produto é resfriado durante o processo, a diferença da temperatura entre o produto e o

ambiente diminui, e seu valor depende portanto do momento ao longo do processo de

resfriamento.

20

Assim, sob as condições de resfriamento Newtoniano (quando o gradiente de

temperatura dentro do produto durante o processo é considerado desprezível) e substituindo o

tempo θ por z e fazendo (T - Ta) = 0,5 (Ti - Ta), tem-se que:

naia

ai

ai TTTTCRzTT

TTzCR 5,0)()(2ln

)(5,0)(

ln)( =−−⇒=⇒−

−= ,

onde n denota múltiplos de meio-tempos de resfriamento (MOHSENIN, 1980).

3.4.3 Propriedades térmicas do produto

As propriedades térmicas dos produtos hortícolas resumem-se à sua habilidade de

conduzir, armazenar, e liberar calor. Tais propriedades são essenciais na análise da

transferência de calor que ocorre em processos térmicos tais como refrigeração, congelamento,

aquecimento e extrusão. Seu conhecimento é necessário para modelagem e dimensionamento

do projeto de equipamentos de refrigeração, cálculo do consumo de energia, e

desenvolvimento de sistemas de esterilização e assepsia (FONTANA et al., 1999).

As propriedades termo-físicas das frutas e hortaliças afetam também sua qualidade

sensorial e variam consideravelmente com o teor de água, seu principal constituinte. Este teor,

por sua vez, depende da cultivar, das condições de cultivo, do estádio de amadurecimento do

vegetal quando colhido e umidade perdida após a colheita (CORTEZ, 2002b). Algumas das

propriedades térmicas mais importantes envolvidas no resfriamento rápido do produto são

descritas a seguir.

1. A massa específica ou densidade (ρ) pode ser definida pela relação entre a massa

(m) e o volume (V) ocupado por um material biológico. Existem três tipos de densidade: a)

“bulk”ou a granel: equivale à massa individual de cada produto acondicionado em embalagem

em um determinado volume total incluindo o espaço poroso dentro da embalagem; b)

aparente: corresponde à massa total do produto dividida pelo seu volume total considerando

também a porosidade; c) real é obtida pelo quociente entre a massa e o volume totais do

produto sem incluir o espaço poroso (MOHSENIN, 1980).

2. Calor específico corresponde à energia necessária para alterar a temperatura da

massa unitária do produto em um grau. Baseia-se estritamente na quantidade de energia

21

requerida e não na taxa em que ocorre essa mudança de temperatura (FONTANA et al., 1999).

Considerando-se Q como a energia fornecida para uma alteração ∆T da temperatura, obtém-se

o calor específico (c) pela equação c= Q/(ρV∆T). Dependendo de como o calor é armazenado

no produto, a transferência de calor pode ocorrer à pressão (cp) ou volume constante (cv)

segundo MOHSENIN (1980). Dentre os vários métodos para se determinar o calor específico

de produtos hortícolas pode-se citar por exemplo o método empírico das equações, das

misturas (mais usado devido a sua simplicidade e precisão) e calorímetro adiabático

(FONTANA et al., 1999).

O calor específico é essencial no cálculo da quantidade de calor imposta ao

equipamento de refrigeração durante o resfriamento ou congelamento de produtos hortícolas.

Através dessa propriedade pode-se obter a variação de entalpia específica H de um produto

utilizada na estimativa da energia a ser removida ou adicionada para se reduzir ou elevar a

temperatura T do produto: cp = (∂H/∂T)p, conforme ASHRAE (2002).

3. Condutividade térmica (k) consiste na quantidade de calor que flui entre duas

superfícies de diferentes temperaturas através de uma área unitária (A) em função do tempo

(θ). Essa propriedade relaciona a taxa de transferência de calor por condução (dQ)1 com o

gradiente de temperatura (dT): q= k A dT/dx ou dQ/Adθ = -k dT/dx (MOHSENIN, 1980). A

condutividade térmica consiste na capacidade do produto em conduzir o calor quando

submetido a processos de resfriamento, congelamento ou secagem, por exemplo. Tal

propriedade, portanto, é essencial na busca da manutenção da qualidade do produto e da

eficiência do sistema (FONTANA et al., 1999). A condutividade térmica varia com a

composição (como umidade), estrutura e temperatura do produto (MOHSENIN, 1980).

4. Difusividade térmica corresponde à taxa em que o calor se propaga ou é difundido

através dos produtos, sendo portanto essencial para a estimativa da duração de processos como

aquecimento, resfriamento ou congelamento (FONTANA et al., 1999). A difusividade

equivale à relação entre a condutividade térmica k, a densidade ρ e o calor específico c: α =

k/ρc (MOHSENIN, 1980). Assim como outras propriedades já citadas, seu valor é 1 Condução - quando não há nenhum movimento apreciável do líquido circundante e este não é transparente, a transferência de calor de uma porção do sólido opaco homogêneo à outra ocorre principalmente por condução. Neste caso, a transferência de calor é influenciada apenas por um gradiente de temperatura (MOHSENIN, 1980).

22

influenciado pela composição, teor de umidade, temperatura e porosidade do leito de produto

(FONTANA et al., 1999). Existe uma grande variação entre os valores de difusividade térmica

de um mesmo produto medidos por pesquisadores diferentes. Os métodos de determinação

podem ser longos ou exigir muitos cálculos (ANSARI e AFAQ, 1986).

5. A emissividade térmica (FЄ) é a relação entre o poder emissivo total de qualquer

material e aquele de um corpo negro à mesma temperatura. Pode ser correlacionada à taxa de

transferência de calor por radiação2 de um material “quente” a um “frio” pela equação: q=A FЄ

Fα σ (T14-T2

4), onde A equivale à área da superfície efetiva, T1 e T2 se referem à temperatura

absoluta dos dois materiais, Fα ao ângulo de um deles em relação ao outro e σ corresponde à

constante de Stephan-Boltzmann (MOHSENIN, 1980).

6. Coeficiente de transferência de calor superficial (h) é a taxa de transferência de

calor para cada grau de diferença de temperatura (∆T) através da interface sólido-fluido por

unidade de área da superfície (A) do material sólido. Embora não se trate de uma propriedade

térmica dos produtos hortícolas, é essencial no projeto de equipamentos de transferência de

calor envolvendo convecção3, daí a ser denominada também de fator de proporcionalidade na

equação da convecção: q = h A ∆T. Este coeficiente depende de várias propriedades do fluido,

como densidade, viscosidade, calor específico, condutividade térmica, velocidade e

temperatura média, assim como parâmetros do produto que podem afetar o movimento do

fluido ao seu redor, por exemplo a forma, tamanho, área e aspereza da superfície e arranjo na

embalagem (ASHRAE, 2002; MOHSENIN, 1980).

Dentre aquelas propriedades do fluido, a velocidade do ar exerce crucial influência

na taxa de transferência de calor do produto a ser resfriado. Como são muitas as variáveis que

podem alterar o perfil de velocidade do ar no leito de hortícolas acondicionado em

embalagem, geralmente os instrumentos e técnicas de laboratório para sua determinação direta 2 A radiação é particularmente significativa sob altas temperaturas e não requer nenhum líquido. Ocorre quando o sólido, ou fluido ao seu redor, é transparente. Como a condição do sólido transparente é rara no caso de materiais biológicos, os efeitos da radiação no produto são menores (MOHSENIN, 1980). 3 A convecção ocorre quando há um movimento relativo entre o sólido e o meio circundante (ao redor), em função do fluxo do fluido. O transporte do fluido aquecido pode ser natural, devido à simples mudança na densidade fluida causada por variação de temperatura (convecção livre ou natural), ou forçado, através de instrumentos mecânicos tais como ventiladores ou bombas (convecção forçada). A principal resistência à transferência de calor neste caso está na camada laminar do fluido na interface sólido-líquido (MOHSENIN, 1980).

23

não fornecem resultados precisos. A dificuldade e imprecisão da determinação direta da

velocidade do ar em torno do produto (ALVAREZ e FLICK, 1999) aumenta à medida que o

fluxo passa da faixa laminar à turbulenta. Esta faixa, por sua vez, depende da relação entre as

forças inerciais e viscosas, definida como número de Reylnolds (Re = ρ VL/υ), onde L é o

comprimento característico do material, ρ é a densidade, V, a velocidade e υ a viscosidade

cinemática do fluido (ASHRAE, 2001).

Quando o fluxo encontra-se na faixa laminar, isto é Re<2300, as forças viscosas

prevalecem e portanto o movimento das partículas do fluido é bastante ordenado. Nesse caso,

a velocidade em geral é superior na zona porosa e diminui ao atingir obstáculos como as

paredes da embalagem ou pontos de contato entre produtos. Se Re>4000 as forças inerciais

dominam provocando instabilidade no fluxo, caracterizado então por flutuações longas e

intensas de velocidade. Assim, a turbulência acelera a transferência de calor e massa através

do fluido durante o processo de resfriamento, mas a formação de vórtices pode tornar o fluxo

imprevisível. Na faixa transiente, isto é, 2300<Re<4000, o fluxo se torna uma mistura de

laminar e turbulento que se comportam de maneira diferenciada quanto à perda de energia por

fricção ao circular através do meio poroso, gerando ainda mas instabilidade e imprecisão nas

medidas (ASHRAE, 2001).

24

4. SEÇÕES: ARTIGOS CIENTÍFICOS SUBMETIDOS E PUBLICADOS

Esta seção descreve as principais etapas utilizadas na pesquisa para alcançar o

objetivo proposto, sob forma de artigos científicos, submetidos e/ou já publicados. Os

principais materiais e métodos utilizados para execução dos experimentos da pesquisa são

citados sucintamente a seguir e descritos em maiores detalhes em cada artigo.

Conforme previamente mencionado, este trabalho visou o desenvolvimento e

aplicação de uma metodologia para o projeto dos orifícios de embalagens para frutas e

hortaliças submetidas ao resfriamento rápido a ar forçado. Para tal propósito, os produtos

hortícolas foram inicialmente representados nos experimentos laboratoriais por esferas ocas de

PVC preenchidas com um gel de ágar-ágar e água (Figura 1 e descrição detalhada no artigo 1).

Vários inconvenientes foram observados nessa primeira tentativa quanto à heterogeneidade do

gel, seja durante o preenchimento das esferas ou devido a vazamentos e evaporação da água ao

longo do experimento, à variação da posição do termopar no interior das esferas (aparato

experimental mostrado nas Figuras 2 e 3), à variação entre os diâmetros das esferas e sua baixa

resistência à deformação.

Figura 1. Esfera plástica preenchida com ágar-ágar usada como produto-modelo nos

experimentos iniciais.

termopar

25

Figura 2. Vista da matriz de esferas de ágar-ágar com placas plásticas no interior do túnel de

ar forçado usado como aparato experimental.

Figura 3. Vista geral do aparato experimental dos testes com esferas de ágar-ágar na câmara

fria, mostrando túnel de ar forçado, ventilador e dispositivos de medição de pressão do ar.

termopares

placa plástica

placa plástica

túnel de ar-forçado ventilador

dispositivos de medição de pressão do ar

26

Para reduzir a heterogeneidade do produto-modelo usado no experimentos, as esferas

ocas foram substituídas por esferas sólidas de polímero desconhecido impermeáveis

previamente testadas (Figura 4). Assim, as esferas a serem utilizadas nos experimentos foram

primeiramente calibradas em resfriamento rápido em água gelada sob agitação, para garantir a

seleção de um grupo uniforme em termos de taxa de resfriamento e capacidade térmica. Cada

esfera foi então inserida em um tubo de seção transversal circular, considerada como área total

aberta, e submetida a um fluxo de ar pré-determinado e conhecido (Figura 5). A partir dos

resultados gerados, estabeleceu-se uma correlação exponencial entre o coeficiente de

resfriamento das esferas e o índice de resfriamento obtido anteriormente em água gelada

(artigo 2). Esta correlação foi então aplicada na determinação do perfil de velocidade do ar no

leito de produtos-modelo, apresentando limitação de cálculo para fluxos de ar mais baixos

(0,125 e 0,250 L·s-1·kg-1) e gerando erros para produtos localizados nas camadas mais

profundas da embalagem.

Figura 4. Esferas sólidas (bolas de bilhar ou “snooker”) instrumentadas com termopar tipo T

utilizadas como produtos-modelo nos experimentos seguintes.

termopares

27

Figura 5. Aparato experimental mostrando em detalhe tubo circular usado para determinar as

primeiras correlações entre velocidade do ar e coeficiente de resfriamento da esfera (artigo 2).

A ferramenta desenvolvida não representou com precisão a distribuição do ar em

produtos acondicionados em embalagem submetidos ao resfriamento rápido, produzindo

valores de velocidade do ar até 50% menores aos medidos instrumentalmente em avaliações

de balanço de massa através do eixo Z. Tal imprecisão é oriunda do fato de não considerar o

efeito de aquecimento do ar ao atravessar as camadas subseqüentes de produto e super estimar

a superfície da esfera exposta ao ar frio, que por ser resfriada individualmente permite a

circulação do ar livremente ao seu redor.

Um novo aparato foi então desenvolvido, onde um grupo de 512 esferas empilhadas

de maneira colunar (Figura 6) foi utilizado para simular o produto acondicionado em

embalagem e submetido ao resfriamento rápido por ar forçado (Figura 7). Neste caso,

desenvolveu-se uma relação que permite obter a velocidade de aproximação do ar

considerando a seção transversal da embalagem inserida no túnel de resfriamento como a área

total aberta. Uma correlação foi desenvolvida entre a velocidade do ar e o coeficiente de

28

resfriamento para cada esfera instrumentada, considerando sua posição específica no interior

da embalagem (artigo 3). Portanto, tais equações levaram em conta as pequenas diferenças

existentes entre as propriedades térmicas de cada esfera e a variabilidade do movimento do ar

ao redor de cada esfera de acordo com sua posição no leito de produto. O balanço de massa

mostrou que tais equações produziram resultados iguais a apenas ±1% do fluxo de ar esperado.

Estas correlações foram mais uma vez aperfeiçoadas através de pesquisa detalhada em fluxos

de ar correspondentes à faixa de transição entre regimes laminar e turbulento, para reduzir

imprecisões geradas na região transiente. Este trabalho seguiu mesma metodologia e

procedimentos utilizados e descritos no artigo 3, exceto que a faixa de fluxo de ar testada foi

ampliada. Como os resultados obtidos ainda não foram submetidos para publicação, um

resumo desta última pesquisa é mostrado no apêndice B (artigo 9).

Figura 6. Vista do grupo de 512 esferas empilhadas em arranjo colunar para simular o produto

acondicionado em embalagem.

29

Figura 7. Aparato experimental com grupo de 512 esferas usado nos últimos experimentos

para testar a aplicabilidade da metodologia de pesquisa desenvolvida.

A metodologia estabelecida possibilitou a investigação do efeito da configuração de

aberturas de embalagens na velocidade e uniformidade de distribuição do fluxo de ar em seu

interior. Dessa forma, as equações desenvolvidas foram aplicadas no estudo da circulação do

ar durante o resfriamento para diferentes configurações de áreas e posicionamento de aberturas

nas laterais da embalagem, como centro, periferia, diagonal, e orifícios para manuseio (artigos

4, 5, 6 e 7). As aberturas distribuídas na diagonal foram analisadas para verificação do efeito

que a gravidade exerce na distribuição do ar para fluxos laminares. Assim como os orifícios do

tipo “alça” para o manuseio foram investigados para verificar sua influência no processo de

resfriamento rápido do produto, em função da área relativa das demais aberturas da

sistema de aquisição de dados de pressão e temperatura

esferas plásticas

trocador de calor

instrumentos de medição do fluxo de ar

túnel de ar-forçado

ventilador

30

embalagem e fluxo de ar aplicado. Uma tabela contendo os principais resultados obtidos nas

pesquisas descritas nos quatro artigos é apresentada no apêndice A.

Também foi realizada uma análise energética, incluindo os efeitos da taxa respiratória

do produto e do ventilador usado no resfriamento rápido, visando auxiliar o projeto dos

orifícios das caixas de hortícolas considerando também os custos decorrentes do

funcionamento do sistema (artigo 8). Desta forma, foi possível verificar e estabelecer algumas

recomendações quanto à configuração mais adequada para embalagens e faixa de fluxo de ar

para operação do sistema que proporcionem um processo de resfriamento mais eficiente,

respeitando as peculiaridades do hortícola e restrições estruturais do material.

Lista de artigos submetidos e publicados:

Artigo 1. CASTRO, L.R.; VIGNEAULT, C.; CORTEZ, L.A.B. Container opening

design for horticultural produce cooling efficiency. International Journal of Food,

Agriculture and Environment, Helsinki, Finlândia, v. 2, n. 1, p. 135-140, 2004.

Artigo 2. VIGNEAULT, C.; CASTRO, L.R. Indirect airflow distribution

measurement for horticultural crop package, Part I: Produce-simulator property evaluation.

Transactions of the ASAE. 2004. (no prelo)

Artigo 3. VIGNEAULT, C.; CASTRO, L.R.; GOYETTE, B.; MARKARIAN, N.R.;

CHARLES, M.T.; BOURGEOIS, G.; CORTEZ, L.A.B. Indirect airflow distribution

measurement for horticultural crop package; Part II: Verification of the research tool

applicability. Transactions of the ASAE. 2004. (submetido em abril/2004)

Artigo 4. CASTRO, L.R.; VIGNEAULT, C.; CORTEZ, L.A.B. Effect of container

opening on air distribution during precooling of horticultural produce. Transactions of the

ASAE. 2004. (no prelo)

Artigo 5. CASTRO, L.R.; VIGNEAULT, C.; CORTEZ, L.A.B. Cooling performance

of horticultural produce in containers with peripheral openings. Postharvest Biology and

Technology. 2004 (submetido em julho/2004)

Artigo 6. VIGNEAULT, C.; CASTRO, L.R.; CORTEZ, L.A.B. Effect of gravity on

forced-air precooling. IASME Transactions on Mechanical Engineering. 2004. (no prelo)

Artigo 7. VIGNEAULT C.; CASTRO, L.R.; GAUTRON, G. Effet de la présence de

poignées ouvertes sur le prérefroidissement de produits horticoles (Effect of handle opening

31

on cooling efficiency of horticultural produce). La Revue Internationale du Froid.

(submetido em julho/2004)

Artigo 8. CASTRO, L.R.; VIGNEAULT, C.; CORTEZ, L.A.B. Effect of container

openings and airflow rate on energy required for forced-air cooling of horticultural produce.

Canadian Agricultural Engineering. 2004. (no prelo)

No apêndice A:

Artigo 9. VIGNEAULT C.; CASTRO, L.R.; PANNETON, B. Laminar to turbulent

indirect airflow measurement method for horticultural crop package. Canadian Agricultural

Engineering. 2004 (em preparação, artigo a ser submetido em dezembro/2004)

Além destes, esta tese ainda contém o artigo 10 (Apêndice C) que reúne informações

de pesquisas realizadas anteriormente, incluindo os trabalhos apresentados nos artigos 1, 2 e 3,

e apresenta o conceito geral da metodologia de projeto de embalagem, através da determinação

indireta da distribuição do ar através de produtos hortícolas submetidos ao resfriamento

rápido:

Artigo 10. VIGNEAULT C.; CASTRO, L.R.; CORTEZ, L.A.B. A new approach to

measure air distribution through horticultural crop packages. Acta Horticulturae, Leuven,

2004. (no prelo)

32

Food, Agriculture & Environment, Vol.2 (1), January 2004 135-140 Artigo 1. Container opening design for horticultural produce cooling efficiency

Larissa R. de Castro1, 2

, Clément Vigneault1, 2 *

and Luís A. B. Cortez2

1Horticultural Research and Development Centre, Agriculture and Agri-Food Canada, 430 Gouin, Saint-Jean-

sur-Richelieu, Québec, CANADA, J3B 3E6. 2College of Agricultural Engineering, State University of Campinas

(UNICAMP), Cidade Universitária Zeferino Vaz, s/n, Campinas, SP, Brazil, CP 6011, 13083-970. *e-mail: vigneaultc@agr.gc.ca

Received 12 September 2003, accepted 15 January 2004.

Abstract Rapid cooling is recommended to decrease postharvest losses of fruits and vegetables. Cooling rate and homogeneity are affected by the package design. Thus, the aim of this research was to evaluate the effect of different container opening configurations on cooling process. The variables analyzed were: thirteen opening designs and four airflow rates. Twenty-eight out of the 224 produce simulators inside the container were instrumented with thermocouple. The configurations of the openings on the package walls were compared in terms of three individual orifice areas (0.67, 1 and 2%) distributed in 4 and 3 positions in width and height (X and Y directions), respectively. The statistical analysis demonstrated that airflow rate was the most important variable in explaining the variations of half-cooling time. The opening position in Y direction had no significant effect on pressure drop, cooling rate and its uniformity. At the maximum airflow rate studied, individual area and position of the openings had no significant effect on half-cooling time. Increasing total opening area for more than 8% did not produce significant cooling rate increment. The 14% opening area generated cooling rates as uniform as the ones obtained using fully opened configuration as well as negligible differences in pressure drop.

Key words: Package openings, airflow, cooling rate, uniformity, pressure drop.

Introduction Postharvest losses are estimated to range from 30-40% of total harvested produce in the world1. Precooling is a critical step in slowing down the metabolic processes of the perishable produce, extending their shelf life, and decreasing the rate of deterioration by microbiological activity and water loss2. Forced-air cooling is the most common and efficient precooling technique used commercially, particularly for produce that are sensitive to water contact3-5.

The efficiency of the forced-air cooling process is mainly evaluated by the rapidity and the uniformity in produce temperature reduction compared to the energy input. Therefore, process improvement is a function of the cooling rate optimization. In terms of air distribution, uniformity is directly related to the number and distribution of openings6

on the container walls. Increasing the total opening area causes a decrease in pressure drop through the container6, 7

but a decline in its structural resistance. Thus, the design of a container for handling horticultural produce must consider both the cooling efficiency and the structural aspect.

A wide variety of containers for storage and transportation of fruits and vegetables have been adopted over time. Container design was generally based on production capability and structural rigidity while adequate air distribution was not considered8. Rather than shape, opening area plays a more

important role in airflow restriction2,9-11. A total opening area less than 25% of the container surface restricts airflow considerably6, 11. Reducing the opening area below 10% generates a lower cooling rate and sharply increases the cooling cost10.

Opening position has also an important effect on cooling rate uniformity12. Openings must be evenly distributed on package walls to provide homogeneous temperature reduction of the entire packed produce. When choosing opening positions, the package stack pattern must be considered to ensure that the openings of side by side packages are aligned to avoid obstruction by other container walls13-14.

Since air movement is driven by the pressure differential, it follows the path of least resistance, generally creating air dead zones inside of the container. Secondary packaging and produce reduce the air movement15-16. Even produce positioning has a significant effect on air circulation; for example: although straight stacking of oranges resulted in the same porosity value as randomly stacked ones, straight stack generated a better defined airflow path, and consequently, a lower pressure drop for the same airflow rate17. The number of layers and orientation of the produce inside a package also generates an important effect on airflow resistance9, 11, 17-19. Many authors also observed that increasing the distance of the produce from the air inlet caused lower cooling rates10, 12, 20-23. Increasing airflow rate enhances

33

average cooling rate of produce12, 14, 16, 22, 24 as well as

decreases the temperature gradient inside the container10, 21.

Horticultural crops exhibit differences in physical and chemical properties, which also modify as they ripen25. For this reason, it is very difficult to maintain similar produce positioning patterns and thermal properties when replicating experiments with packed produce. Therefore, they are occasionally replaced by more stable produce simulators which help to minimize test variability6, 12, 27, 27.

The objective of this research was to evaluate the effect of the area and distribution of openings on cooling rate and and uniformity using straight stacked produce simulators.

Material and Methods Produce simulator: Two hundred twenty-four hollow polychlorinated vinyl (PVC) balls, 68.0 mm outside diameter, 0.8 mm wall thickness, and 9.50 g mean weight, were used to simulate spherical horticultural produce. Twenty-eight balls were filled with 3% mass-base agar-agar and water solution, and instrumented with a 24-gage type T thermocouple positioned at their center. The instrumented balls were stacked along with non-instrumented balls forming a 224-ball orthogonal matrix of eight rows (X direction) by four layers (Y) and seven columns (Z) (Figure 1). The matrix was supported by a metallic grid to provide stability. The metal grid consisted of a 1.5 mm-diameter wire, 25.4 mm orthogonally spaced resulting in an 88.5% total opening area. The matrix dimensions were 560 by 280 by 490 mm in the X, Y and Z directions respectively, which resulted in 47.6% porosity. Table 1 presents the relative position of the 28 instrumented balls. Experimental set-up: Figure 2 shows a diagram of the experimental set-up which consisted of a tunnel containing the ball matrix, a fan, and an airflow measuring device. The setup was controlled using a data acquisition and control unit. The tunnel was made of wood with a 560 by 280 mm inside cross-section and 1700 mm in length and insulated with 12-mm-thick rubber foam. The matrix was positioned at a 140 mm distance from the end of the air-inlet tunnel. The air-outlet of the tunnel consisted of a 600 mm long plenum allowing the measurement of the air pressure drop (APD) across the matrix of balls with the use of either pressure transmitters 0-0.250±0.0025 or 0-6.25±0.0625 mm of water (Model 607-0 and 607-1 respectively, Dwyer Instruments Inc. Michigan City, IN, USA). The end of the air-outlet tunnel was air-tightly connected to the inlet of a backward-curved centrifugal fan driven by a 2.3 kW variable speed motor. The experimental setup was placed inside a refrigerated room maintained at 5ºC.

Figure 1. Overview of the matrix of balls. The numbered balls are instrumented. The three digit number on a ball represents the X, Y, and Z positions, respectively.

Figure 2. Top view of the experimental set-up. The airflow measuring device consisted of a 97 mm-

diameter and 1830 mm-long tube attached to the fan outlet. A Pitot tube was placed at the center of the 97 mm tube. The two pressure transmitters 0-0.250±0.0025 and 0-6.25±0.0625 mm of water were also used to measure the dynamic pressure of the air circulating during the tests. These transmitters were selected to cover the entire range of measurement and guaranty an acceptable precision according to standard recommendations25. The air-measuring device was calibrated before the experimentation using a 16-point measuring pattern, according to the same recommendations25. A calibration curve was developed to express the airflow rate as a function of the dynamic pressure measured at the center of the tube.

Twelve pairs of polypropylene plates, 560 wide by 280 high and by 2 mm thick, were perforated. One, two, four and seven columns (X direction), and one,

34

two and three layers (Y direction) of square openings were made through these plates (Table 2). The total opening areas (TOA) through the plates were 2, 4, 8, and 14% using holes having individual opening areas (IOA) of 0.67, 1, and 2%. These pairs of plates were spaced 490 mm apart to simulate a package. A “fully-open-configuration” was also tested which covered 88.5% of the cross-sectional area of the tunnel due to the metallic grid.

Four airflow rates equivalent to 0.5, 1, 2, and 4 L•s-1•kg-1

were tested, by adjusting the fan speed until the desired dynamic pressure was obtained.

A data acquisition system (DAS, model CAD1232/ MCS1000, Lynx Tecnologia Eletrônica, São Paulo, Brazil) was used to record on a 20 s-interval base simultaneously, the temperature inside the balls along with the air temperature before and after the ball matrix, the pressure drop and the dynamic pressure at the Pitot tube.

Experimental procedure: Thirteen opening areas and four airflow rates were tested in a complete block design, to determine the effect of the total opening percentage, the airflow rate and the ball positioning on the cooling rate of the balls. Each test was repeated three times. Prior to the start of each test, the tunnel holding the balls was placed at ambient temperature, approximately 25ºC. An axial fan forced the air through the matrix to reheat the balls for approximately 60 minutes. After this conditioning period, a pair of perforated plates was installed and the tunnel was placed in the cold room. The tunnel was connected to the inlet of the centrifugal fan, which was then turned on immediately. The results were recorded until the temperature of the warmest ball had reached 7ºC. The temperature-time data recorded were used to calculate the cooling coefficient of each ball for all treatments based on their half-cooling time (HCT) using a dedicated ExcelTM

macro developed by Goyette et al29. The uniformity (CU) of the cooling process was determined as the inverse of the standard deviation of the HCT of the balls for each test. A Stepwise Forward-Backward Regression30

was performed to determine the effect and the interactions between the airflow rate (AFR), the number of openings in the X (NOX) and Y (NOY) directions, the individual opening area (IOA), as well as the position of the balls in the X, Y, and Z directions (PXD, PYD, PZD) on the HCT, CU, and APD. The goodness of fit coefficient (R2

i) was calculated to determine the increase in variance explained by adding a dependent variable i to the other variables in the regression equation.

Table 1. Relative position of the instrumented balls in the matrix. The air was flowing in the increasing number direction of the Z axis.

# of ball X Y Z 1 1 1 1 2 3 1 1 3 5 1 1 4 2 3 1 5 4 3 1 6 6 3 1 7 8 3 1 8 4 2 3 9 6 2 3

10 8 2 3 11 1 4 3 12 3 4 3 13 5 4 3 14 7 4 3 15 2 1 5 16 4 1 5 17 6 1 5 18 8 1 5 19 1 3 5 20 5 3 5 21 7 3 5 22 1 2 7 23 3 2 7 24 5 2 7 25 7 2 7 26 4 4 7 27 6 4 7 28 8 4 7

Results and Discussion

The statistical analysis showed a rather large experimental error. The latter is most likely due to difficulties of uniformly filling the balls with agar-agar solution and the wide imprecision of airflow measurements, which allowed identifying only large differences.

As expected, the following parameters produced a significant effect (P<0.0001) on the HCT: AFR (F3, 4360 = 6633), position of the balls in X (F3,4360 = 387), Y (F1, 4360 = 134), and Z directions (F3, 4360 = 169), NOX (F 3, 4360 = 183), and IOA (F3, 4360

= 31), and NOY did not have a significant effect on HCT (F2, 4360 = 3.68, P = 0.055). The Stepwise Forward-Backward Regression allowed evaluating the relative importance of the independent variables in explaining linearly the variation of HCT. In a decreasing order, AFR explained 55.8% of HCT variation, followed by PXD (3.2%), PZD (1.6%), NOX (1.5%), PYD (1.1%), IOA (0.2%), and NOY (<0.1%) for a total of 63.3% in the linear regression. The variables AFR, IOA, and NOX provided all a significant effect on CU and APD, but not NOY, as observed for HCT. These results are shown in Table 3a.

Airflow rate: When increasing AFR from 0.5 to 1, 2, and 4 L•s-1•kg-1, the average HCT was reduced by 24, 50, and 61% respectively, while the coefficient of uniformity (CU) increased by 33, 156, and 290% respectively. Similar results in the diminution of HCT

35

were noticed12 when doubling an AFR of 2 L•s-1•kg-1.

A regression analysis determined a non-linear model to infer the HCT response as a function of AFR and TOA (Equation 1).

HCT = 26.1544AFR -58.5849 ln(AFR) + 2.6618 ln(TOA) R2=0.892 (1) Where: HCT = half-cooling time, min; AFR = airflow rate, L•s-1•kg-1; TOA = total opening area, %. The fairly high goodness of fit coefficient (R2)

obtained confirms the previous assertion, which stated that airflow rate explained most of the variation in produce cooling time response. The latter agrees with the results found by Arifin and Chau21

who also claimed a non-linear relationship between AFR and cooling time, and disagrees with other author

findings21. On the other hand, Boyette’s16 results point to a linear correlation between airflow rate and cooling time for the tested range, but the author suggests an exponential function over a larger range.

At the maximum AFR (4 L•s-1•kg-1), the only variables to be considerate in the HCT response are PXD and PYD (Table 3b). Thus, the opening area had an effect on HCT only for the lower values of AFR (Figure 3) which agrees with Arifin and Chau’s findings21. Therefore, it is possible to compensate the effect of the open area on both cooling rate and uniformity by increasing the airflow rate.

Ball position: In general, ball position in every direction had a significant effect on the HCT, which agrees with the results presented by many authors12,20,22.

0

10

20

30

40

50

60

70

0 1 2 3 4

AFR ( L· s- 1 · k g- 1 )

2

4

8

14

88.5

TOA ( %)

Figure 3. Half-cooling time (HCT) as a function of airflow rate (AFR) and total opening area (TOA).

0

10

20

30

40

50

60

70

0 10 20 30 40 50 60 70 80 90

TOA ( %)

0.5

1.0

2.0

4.0

AFR ( L. s - 1 . k g - 1 )

Figure 4. Half-cooling time (HCT) as a function of total opening area (TOA) and airflow rate (AFR).

36

For the statistical analysis of the effect of the ball positions in X and Y directions (PXD and PYD), the matrix center was considered as the origin reference, instead of the matrix left inferior edge as it had been used for ball positioning purposes. All openings were also distributed symmetrically relative to that point, permitting a symmetric approach. This allowed detecting a significant effect of the distance separating the balls and openings on HCT.

The position of the balls in Z direction (PZD) had significant effect (Table 3b) on HCT for the lower airflow rates (<1 L•s-1•kg-1), but not for the highest AFR value (4 L•s-1•kg-1). No linear decrease in cooling rate was observed as the depth increased (PZD), which does not agree with the results obtained by other authors22. This result could be explained by the fact that most balls were not filled with agar-agar; hence the air was not warmed up as it usually occurs.

However, because the balls were empty, it was possible to identify another phenomenon. Although, in general, the balls of the first layers cooled faster, the highest HCT values were observed in the penultimate instrumented layer (PZD = 5), not in the last layer (PZD=7). A possible reason for the values found in the last layer is the proximity of layer 7 to the air outlet openings which could have increased the air velocity through the layer, forming a turbulent zone, and compensated for the lower temperature differential between the air and the balls of this last layer. The same phenomenon stated by different authors12, 10, 20, 22 was observed at the beginning of cooling process, regarding the increase of air temperature as it passes through the hot produce, keeping warm the rear produce. The lower the AFR, the longer is the period of warm air production and the higher the difference of air temperatures between the inlet and the outlet sides of a package are.

Opening area: NOX had an effect on HCT for the lowest airflow rates (AFR<2 L•s-1•kg-1) (Table 3b). This effect disappeared at higher AFR. NOY had a significant effect on HCT only at the airflow rate of 1 L•s-1•kg-1

(Table 3b). The proximity of the openings

to the balls could explain this almost absent effect. In fact, the largest distance separating the balls and the openings in this direction was equal to double the diameter of a ball.

Lower NOX resulted in lower cooling uniformity (CU). The CU could then be predicted by a linear equation model considering AFR (L•s-1•kg-1) and NOX (Equation 2).

CU = 0.079 AFR + 0.014 NOX R2=0.911 (2) Pressure drop: Using different IOA to form specific TOA did not produce any significant effect on APD but, varying TOA produced a significant effect on APD (F2, 69 = 152.07, P<0.0001). The latter results demonstrate the TOA affects the cooling rate and not the individual value of the opening, which is in agreement with Vigneault and Goyette’s findings

6.

The APD (mm water) could then be modeled as a function of AFR (L•s-1•kg-1) and TOA (%) using Equation 3.

APD = 37.487 TOA-1.5 AFR2

R2 =0.993 (3) This relationship agrees with many authors6, 17

who observed that APD through a vented package follows approximately a quadratic relationship with the average velocity which is related to airflow rate. Other authors7

also stated that pressure drop scales to TOA-1.5.

The cooling rate was generally stable when the total opening area >8% (Figure 4). Similar results10

stated that an opening area < 10% increases the cooling time. TOA of 14% produced a pressure drop response 2.5 times greater than a fully opened end, 5 compared to 2 mm of water, respectively. This difference could be considered negligible in terms of its impact on the energy required compared to the increase in pressure drop of 71 mm of water generated by using a TOA of 2%. Furthermore, no significant improvement was observed on HCT or on CU when increasing TOA from 14% to fully opened end. This 14% TOA could be considered as the maximum TOA required for the container design in view of cooling rate and energy.

Table 2. Combinations of the number of openings in X and Y directions and individual opening area (IOA) to produce total opening area (TOA)

IOA (%)

2 1 0.67

TOA (%)

NOX NOY NOX NOY NOX NOY

2 1 1 1 2 1 3 4 2 1 2 2 2 3 8 4 1 4 2 4 3

14 7 1 7 2 7 3

37

Table 3. Results of the statistical analysis showing the level of significance of the correlations between the independent variables and: a) air pressure drop and cooling uniformity, and b) half-cooling time.

a) AFR NOX IOA NOY

APD <0.0001*** <0.0001*** <0.0001*** 0.516

CU <0.0001*** <0.0001*** 0.002*** 0.698

b)

HCT AFR NOX 0.5 1 2 4 1 2 4 7 Full

AFR <0.0001*** <0.0001*** <0.0001*** <0.0001*** <0.0001***

PXD <0.0001*** <0.0001*** <0.0001*** <0.0001*** <0.0001*** 0.016* <0.0001*** <0.0001*** 0.012*

PYD <0.0001*** <0.0001*** <0.0001*** 0.002*** <0.0001*** <0.0001*** <0.0001*** <0.0001*** 0.344

PZD <0.0001*** <0.0001*** 0.057 0.511 <0.0001*** <0.0001*** <0.0001*** <0.0001*** 0.001***

NOY 0.312 0.015* 0.326 0.264 0.741 0.638 0.889 0.760

IOA <0.0001*** 0.003*** 0.0001*** 0.363 0.881 0.944 0.713 0.810

NOX <0.0001*** <0.0001*** <0.0001*** 0.094

*** = has a significant effect, P<0.01 * = has a significant effect, P<0.05

Where: AFR = Airflow rate APD = Air pressure drop CU = Cooling uniformity HCT = Half-cooling time IOA = Individual opening area NOX = Number of openings in X direction NOY = Number of openings in Y direction PXD = Position of the ball on X-axis direction PYD = Position of the ball on Y-axis direction PZD = Position of the ball on Z-axis direction TOA = Total opening area

Conclusions The effect of position and surface for the openings

of a container used for horticultural crop handling on the cooling rate was partially determined. The accuracy and repeatability of the results can be improved by designing a more stable produce simulator. Among all independent variables studied, airflow rate was found to contribute the most for half-cooling time of horticultural produce. Increasing the airflow rate could compensate for the negative effect of less opening area but, it would also raise the pressure drop through the container, and consequently the energy required for the process. Distribution of openings in Y direction did not produce a significant effect on the pressure drop and the cooling rate and its uniformity. The effect of ball position in the Z direction (depth) on HCT was significant only for airflow rates <1 L•s-1•kg-1 but this finding should be carefully used because it could be due to the presence of a large proportion of empty balls. A total opening area of 14% could be recommended as a maximum for the container design in terms of cooling efficiency, i.e., considering cooling rate and its uniformity and energy

costs, as long as the package structural resistance restrictions are also taken into account.

Acknowledgement The authors would like to thank the Fundação de

Amparo à Pesquisa do Estado de São Paulo (FAPESP), for the financial support and scholarship provided.

References 1FAO 1991. Production Yearbook. Food and

Agriculture Organiza-tion, v. 44. FAO Statistics. Series n. 99, Roma.

2Kader, A. A. (Ed.) 1992. Postharvest technology of horticultural crops. University of California: Publication No 3311, 296 p.

3Cortez, L.A.B., de Castro, L.R. and Vigneault, C. 2002. Resfriamento Rápido a ar: métodos da câmara frigorífica e do ar forçado. In : L.A.B. Cortez, S.L. Honório C.L. Moretti, Eds. Resfriamento de frutas e hortaliças. Embrapa Informação Tecnológica, Brasília, DF, Brasil. pp. 231-272.

4Sargent, S. A., Talbot, M. T. and Brecht, J. K. 1991. Evaluating pre-cooling methods for vegetable

38

packinghouse operations. Univer-sity of Florida, Gainesville. 7 pp.

5Hardenburg, R. E., Watada, A. E. and Wang, C. Y. 1986. The com-mercial storage of fruits, vegetables and florist and nursery stocks. Department of Agriculture, Agricultural Research Service, Wash-ington, 136 pp.

6Vigneault, C. and Goyette, B. 2002. Design of plastic container open-ing to optimize forced-air precooling of fruits and vegetables. Trans-actions of the ASAE 38(1):73-76.

7Van der Sman 2002. Prediction of airflow through a vented box by the Darcy-Forchheiner equation. Journal of Food Engineering 55: 49-57.

8Vigneault, C., Bordin, M. R. and Abrahão, R. F. 2002. Embalagem para frutas e hortaliças. In: Cortez, L. A. B., Honório, S. L. and Moretti, C. L. (eds). Resfriamento de Frutas e Hortaliças, Embrapa Informação Tecnológica, Brasília. pp. 95-119.

9Edeogu, I., Feddes, J. and Leonard, J. 1997. Comparison between vertical and horizontal airflow for fruit and vegetable precooling. Canadian Agricultural Engineering 39(2): 107-112.

10Baird, C. D., Gaffney, J. J. and Talbot, M. T. 1988. Design criteria for efficient and cost effective forced-air cooling systems for fruits and vegetables. ASHRAE Transactions 94: 1434-1453.

11Haas, E., Felsenstein, G., Shitzer, A. and Manor, G. 1976. Factors affecting resistance to airflow through packed fresh fruit. ASHRAE Transactions 82(2): 548-554.

12Émond, J. P., Mercier, F., Sadfa, S. O., Bourré, M. and Gakwaya, A. 1996. Study of parameters affecting cooling rate and temperature distribution in forced-air precooling of strawberry. Transactions of the ASAE 39(6): 2185-2191.

13Parsons, R. A., Mitchell, F. G. and Mayer, G. 1970. Forced-air cool-ing of palletized fresh fruit. Transactions of the ASAE 15: 729-731.

14Henry, F. E, Bennett, A. H. and Segall, R. H. 1979. Hydraircooling vegetables in pallet loads. American Vegetable Grower 27(10): 8-9, 46-47.

15Faubion, D. F. and Kader, A. A. 1997. Influence of place packing or tray packing on the cooling rate of palletized ´Anjou´ pears. HortTechnology 7(4):378-382.

16Boyette, M. D. 1996. Forced-air cooling packaged blueberries. Ap-plied Engineering in Agriculture 12(2): 213-217.

17Chau, K. V., Gaffney, J. J., Baird, C. D. and Church, G. A. 1985. Resistance to airflow of oranges in bulk and in cartoons. Transac-tions of the ASAE 8(6): 2083-2088.

18Vigneault C., Markarian, N.R., da Silva, A. and Goyette, B. 2004. Pressure drop during forced-air circulation of various horticultural produce. Transaction of the ASAE (submitted).

19Irvine, D. A., Jayas, D. S. and Mazza, G. 1993. Resistance to airflow through clean and solied potatoes. Transactions of the ASAE 36(5): 1405-1410.

20Alvarez, G. and Trystram, G. 1995. Design of a new strategy for the control of the refrigeration process: fruit and vegetables conditioned in a pallet. Food Control 6(6):347-355.

21Arifin, B. B. and Chau, K. V. 1987. Forced-air cooling of strawber-ries. Summer Meeting of American Society of Agricultural Engineers (ASAE), 87-6004.

22Lindsay, R. T., Neale, M. A. and Messer, H. J. M. 1983. Ventilation rates for the positive ventilation of vegetables in bulk bins. Journal of Agricultural Engineering Res. 28: 33-44.

23Minh, T. V., Perry, J.S. and Bennett, A. H. 1969. Forced-air precool-ing of white potatoes in bulk. ASHRAE Transactions 75(2): 143- 152.

24Kopelman, I., Blaisdell, J. L. and Pflug, I. J. 1966. Influence of fruit size and coolant velocity on the cooling of Jonathan Apples in wa-ter and air. ASHRAE Transactions 72(1): 209-216.

25ASHRAE 1998. Vegetables. In: Refrigeration Systems and Applica-tions Handbook. American Society of Heating, Refrigerating and AirConditioning Engineers, Inc. Atlanta, Georgia 17: 1-14.

26Maul, F., Vigneault, C., Sargent, S.A., Chau, K.V. and Caron, J. 1997. Nondestructive sensor system for evaluation of cooling effi-ciency. Proceeding from the Sensors for Nondestructive Testing International Conference and Tour. February 18-21. pp. 351-360.

27Alvarez, G., and Flick, D. 1999. Analysis of heterogeneous cooling of agricultural products inside bins. Part I: Aerodynamic study. Journal of Food Engineering 39(2):227-237.

28ASHRAE 2001. Measurement and Instruments. In: Fundamentals Handbook. American Society of Heating, Refrigerating and AirConditioning Engineers, Inc. Atlanta, Georgia, 13:12-17.

29Goyette, B., Vigneault, C., Panneton, B. and Raghavan, G. S. V. 1996. Method to evaluate the average temperature at the surface of a horticultural crop. Canadian Agricultural Engineering 38(4): 291-295. (no Anexo)

30SAS Institute 1988. SAS/STAT User’s guide, Release 6.03 ed. SAS Institute, Cary, North Carolina.

39

Artigo 2. Indirect airflow distribution measurement for horticultural crop

package. Part I: Produce-simulator property evaluation

Clement VIGNEAULT1,2,3, Larissa R. De CASTRO1,2

1Horticultural Research and Development Centre Agriculture and Agri-Food Canada

430 Gouin, Saint-Jean-sur-Richelieu (Qc) CANADA J3B 3E6

3 Corresponding author Tel : 450-346-4494 ext. 170 Fax: 450-346-7740 Email : vigneaultc@agr.gc.ca

2College of Agricultural Engineering State University of Campinas

Cidade Universitaria Zeferino Vaz 13083-970, Campinas SP, BRAZIL

ABSTRACT The possibility of using an indirect measurement method of air velocity pattern inside a container during precooling to aid the container openings design was discussed. This indirect measurement of air velocity is based on using the correlation between the cooling rate of a produce simulator and air approach velocity. Plastic balls were chosen to simulate packed horticultural produce in experiments with forced-air cooling. Specific heat capacity, thermal diffusivity, thermal conductivity, and a cooling rate index of the plastic balls were measured. The cooling rate of the balls was measured using still air and five approach velocities. A regression analysis demonstrated that 98.6% of the variation of the cooling rate of the balls could be explained by the variation of the air approach velocity and their cooling rate index. These findings will allow the prediction of valuable results while using indirect measurement of air velocity to establish the airflow pattern within horticultural produce containers during the next step of this research.

KEY WORDS Produce simulator, package, forced-air, hydrocooling, specific heat.

INTRODUCTION One of the main concerns of package industry engineers and researchers is to design containers that allow consistent airflow distribution during cooling of fruits and vegetables. The container must have enough openings to promote high and uniform airflow through the produce while ensuring that the mechanical resistance is not compromised. Although there are references regarding some package parameters such as overall dimensions and opening area (Kader 2002; Vigneault and Émond 1997), these provide mainly standard recommendations. Furthermore, because these recommendations were obtained under specific and often undefined cooling conditions, they cannot be applied for all situations.

Produce simulator Most authors recommend experimentation with real horticultural produce to determine airflow distribution through a porous medium (Dincer 1994; Irvine et al. 1993; Hackert et al. 1987; Chau et al. 1985; Gorini and Borinelli 1974). Nevertheless, it becomes very difficult to maintain similar thermal properties and produce positioning pattern when replicating experiments with packed fruits or vegetables. Differences in size, shape, weight, external surface roughness, and heat and mass transfer properties may be determined for each variety of produce, but their physical and chemical properties

change with time as they ripen and go through other physiological changes (ASHRAE 2002, Leyte and Forney, 1999). Sometimes, experiments can not be repeated because of a lack of produce uniformity (Tanner et al. 2002). For this reason, other authors propose to complement experimental analysis with airflow simulation through mathematic-computational techniques (Smale et al 2003, Tanner et al. 2002, Van der Sman 2002, Verboven et al. 2001, Nakayama and Kuwahara 1999). In this case, however, the difficulty in reproducing the same real conditions of packed produce such as thermal properties, stacking arrangement, and bed porosity, can compromise the airflow analysis accuracy when modeling cooling process and designing equipments (Fontana et al. 1999, Chau et al 1985).

The standard ways of determining air distribution is by introducing measuring instruments in the air pathway (ASHRAE 2001). The drawback of this method is the difficulty in obtaining consistent results. For instance, it is quite impossible to place a sensor exactly perpendicular to the airflow direction in a porous medium such as packed horticultural produce. The air circulating is continuously changing direction and the measuring device insertion generally disturbs the air pathway, and consequently, modifies the velocity field, as observed by Alvarez and Flick (1999a).

40

Another method consists of injecting a gas, such as CO2 or N2 into the air stream and measuring the proportion of the flow going through each part of the package (Smale et al. 2003 and Tanner et al. 2002). However, the precision of this method has not been published yet and it requires expensive equipment to measure the gas concentration at multiple locations.

The use of a stable material to replace horticultural produce would minimize the variability in tests and the experiment error. Indirect measurement of ambient condition has been used with success by many authors. Amongst them, Vigneault et al. (1995) employed instrumented nylon cylinder to measure ice distribution uniformity through a liquid ice system; Maul et al. (1997) used rubber balls filled with water/agar-agar solution to measure water distribution through a hydrocooling system; Alvarez and Flick (1999b) utilized aluminum balls to measure air distribution through a forced-air precooling system; Vissoto (1999) used carrageen balls to simulate forced-air cooling of orange considering water loss, and Castro et al. (2003) applied plastic balls filled with water/agar-agar solution to simulate horticultural produce in a forced-air precooling system. The last authors concluded that their system offered great opportunities for airflow measurement and could be used for validation of models describing airflow distribution in porous medium. However, the plastic balls filled with water/agar-agar solution and instrumented with thermocouple permitted only determining major effect of container openings on precooling process, requiring large differences between airflow rates due to the relatively important experimental error they produced (Castro et al. 2003). The best precision was obtained by using stable material such as aluminum (Alvarez and Flick 1999b) or nylon (Vigneault et al. 1995).

Produce simulators seem to be the most appropriate method to measure the effect of the different parameters on air distribution through porous medium, such as the container opening. These simulators must be low-cost, stable within the time, and uniform in any characteristic affecting the heat transfer and air circulation such as specific heat, thermal conductivity, density, and shape. The size, heat capacity and thermal conductivity of the simulator should be representative of horticultural produce. It must be possible to establish a relationship between the velocities of the air circulating around the simulator and its cooling rate. By satisfying these conditions, it would then be possible to insert the simulator into a mass of produce or of other simulators and indirectly measure the mean air velocity at the insertion position as well as establish the airflow pattern.

The relationship between air velocity and produce cooling rate was previously established by Lindsay et

al. (1983) and Émond et al. (1996). Although the authors have found linear correlation between airflow rate and cooling time, they suggested an exponential relationship for large airflow ranges. Lambrinos and Assimaki (1997) also established an exponential correlation (Eq. 1).

CT = e5.029 - 0.3336AAV (1)Where CT = cooling time (min) and AAV = airflow

rate (L·min-1).

Heat-mass transfer process The heat-transfer process can be divided in three classes depending on Biot number, Bi (Mohsenin 1980). Bi is defined as the ratio of the heat transfer external resistance of the body to its internal resistance (Eq. 2).

kSh

=Bi o (2)

For a given characteristic length of a body (So, m), when Bi > 10, the convective heat-transfer coefficient (h, W·m-2·K-1) is high compared to its thermal conductivity (k, W·m-1·K-1), which becomes the limiting factor to heat transfer. This is the case of a body immersed in a highly turbulent heat sink. From this theory, one could develop a cooling rate index that allows the comparison of different warm bodies by immersing them into a highly turbulent heat sink such as high-speed-stirred cold water bath. Under these conditions, only the thermal conductivity (k) of the body would affect the cooling rate which could be used as the cooling rate index.

When Bi < 0.2, k is high compared to h·S0, the body temperature is considered uniform throughout its whole volume and h is considered as the limiting factor to heat transfer. This is the case of a thin body immersed in a still environment. By definition, such a condition does not occur with solid foods during the precooling process since k of solid foods is relatively small compared to h, obtained under a fast cooling condition (Ramaswamy et al 1982).

The third class is when Bi ranges between 0.2 and 10 which is the case where there are finite internal and external resistances to heat transfer (Mohsenin 1980). Precooling processes for horticultural crops are considered to be within this range. In this case, the cooling rate is related to the fluid approach velocity and the produce physical and thermal properties.

Objective The aim of the global project is to develop a practical research tool and methodology that allows the determination of airflow pattern inside horticultural crop containers. This tool should provide enough flexibility and precision so that it can be applied for most produce, packing, and cooling conditions found in postharvest of horticultural crops. The specific objective of the first part of this project was to

41

determine the physical and thermal characteristics of a thermocouple-instrumented solid-plastic ball eventually used as a horticultural produce simulator.

MATERIAL AND METHODS Produce simulator Uniform plastic balls (Figure 1) were used to simulate spherical horticultural produce and generate replicable results when measuring the effect of package openings on air distribution. This produce simulator consisted of solid plastic balls made with an unknown polymer, with approximately 52 mm as outer diameter and 125 g as weight. These balls are commonly used for ′′Snooker game′′. First, the diameter and mass of each ball was measured using a micrometer (±0.001 mm) and an electronic scale (±0.001g) respectively. Next, a 1.587-mm-diameter hole was precisely drilled into the center of each ball (±0.025 mm). A 30-gage, 5 m-long insulated copper constantan thermocouple (Type T) was then carefully fixed into this hole and sealed with ‘’Super Strong Water Resistant 5-minute Epoxy’’ glue (LePage, Brampton, On, Canada).

Cooling rate index Cooling rate index is defined as the change in the logarithm time-temperature ratio between the ball center and the water and was calculated by using a dedicated ExcelTM macro developed by Goyette et al. (1996). The cooling rate index was determined for sixteen groups of eight balls each to find a total of 64 balls having similar thermal characteristics. The evaluation of the cooling rate index of each ball was performed by total immersion of a group of eight balls at a time in a 10-L thermostatic bath, Model 1187 (VWR Scientific, Gaithersburg, MD, USA). A plastic support was built to ensure a uniform distance between the balls while holding them into place. The water temperature was maintained at 5oC and vigorously stirred continuously to provide uniform condition for all the balls. The immersion lasted 35 minutes. Each ball was tested three times at different positions on the plastic support.

The temperatures of the water and in the center of each ball were recorded simultaneously by using a data acquisition system DATAshuttle (Strawberry Tree, Sunnyvale, CA, USA) and the Quicklog software (Strawberry Tree, Sunnyvale, CA, USA) at a 60-s interval. An ANOVA analysis followed by Duncan test was performed on the data (SAS Institute 1988) to identify the significant differences between the balls.

Specific heat The specific heat of the balls was also measured following the standard mixture method (Weast and Astle 1979) using a homemade calorimeter. It consisted of a 290 x 290 x 260 mm Styrofoam cube having a 115 x 115 x 75 mm cavity in its center. A 1.5-L plastic bag was inserted in this cavity and filled

with 300±0.01 g of distilled water at 7oC and four balls at a time. The calorimeter was hermetically sealed. Before each test, the balls were preheated at 47oC by immersing them for one hour in the 10-L thermostatic bath. Two thermocouples were used to monitor the water temperature in the calorimeter during the experiment. The temperatures of the water and balls were recorded on a Data Acquisition System (Agilent Technology, HP, Palo Alto, California, USA) at a 30-s-interval. The time-temperature data was recorded until the equilibrium was reached between the balls and the water (approximately 40 minutes). The tests were conducted on six groups of four balls each and repeated three times.

The mean specific heat of each set of four balls (sb) was calculated from the specific heat (sw), mass (mw), and initial temperature (twi) of water; the mass (mb) and the initial temperature (tbi) of balls; and the equilibrium temperature (tf) of balls and water using Equation 3 (Weast and Astle 1979). The results were analyzed using ANOVA test (SAS Institute 1988).

)t(tm)t(tsm

=sfbib

fwiwwb

(3)

Thermal conductivity and thermal diffusivity According to Ramaswamy et al. (1982), the thermal conductivity (k, J·mm-1·s-1·K-1) is directly related to the thermal diffusivity (α, mm2·s-1), the density (ρ, g·mm-1) and the specific heat (cp, J·g-1·K-1) (eq. 4).

pcραk = (4) In turn, the thermal diffusivity (α, mm2·s-1) was

obtained from the cooling rate index (s-1) by considering that the heat transfer from the ball to the turbulent-water heat sink was only limited by the thermal property of the ball (eq. 5).

=

2h r98.9

f

303.2α (5)

Where r (mm) is the ball radius and fh = -(cooling rate index)-1.

Indirect air velocity A correlation was determined between the ball cooling rate and air approach velocity. Although thermal conductivity and diffusivity are more extensively used as material thermal properties, the cooling rate index has the advantage to take into account the variations among the ball physical characteristics, such as radius, density, and specific heat. Therefore this variable allowed simplifying models.

The experimental design to determine the correlation between ball cooling rate and air approach velocity consisted of testing six air velocities and four different balls in a complete block design with 3 repetitions. One ball from each of four 8-ball groups was randomly selected out of those showing a nominal

42

cooling rate index of 0.14 min-1 (Table 1) to be evaluated through the following setup.

The forced-air set-up shown in Figure 2 was made up of a wooden chamber 10 mm in thickness, separated in two compartments (upper and lower). The outer dimensions of the chamber were 900 x 600 x 600 mm. An 80-mm-diameter air inlet hole was drilled through the top cover of the upper chamber and a 77.9-mm-diameter PVC-tube, 310 mm in length was vertically attached to it. Three screws were used as a triangular support to maintain one ball at a time horizontally centered in the PVC-tube and the superior part of the ball vertically distanced at 100 mm from the top of the PVC-tube. A centrifugal fan driven by a 2.3-kW-variable speed motor was mounted between the two inner chambers. The fan created a negative pressure in the upper chamber and forced the outside air to through the 77.9-mm diameter tube into the lower chamber. The fan-motor speed was controlled by a speed controller (Model 174450, Leeson Electric Corporation, Grafton, WI, USA) which was in turn controlled by an Agilent data acquisition system (Model 34970 Data Acquisition/Switch Unit – Agilent Technology, HP, Palo Alto, California, USA). The air was finally released to the atmosphere through a 500 mm long and 34.7 mm-diameter tube. The velocity of the air circulating through the setup was measured using an in-house constructed Pitot-tube device mounted at the air outlet tube. The static and dynamic pressures at the Pitot tube were measured using a pressure transmitter ranging from 0 to 12.70±0.13 mm of water (Model 607-0, Dwyer Instruments Inc. Michigan City, IN, USA). The pressures were recorded simultaneously with air and ball temperatures at a 3-sec-interval using the Agilent data acquisition system. This air-measuring device was calibrated before the experiments using a 16-point measuring pattern, following the standard recommendations of ASHRAE (2001). All tests were performed in a refrigerated room maintained at 2.5°C.

Approach air velocities of 0, 0.12, 0.21, 0.38, 0.70, and 1.03 m·s-1 were selected for this study with respect to the capacity and precision of the forced-air set-up. To produce approach air velocities close to 0 m·s-1, the tested ball was positioned at the center of a top covered vertical tube, 77.9-mm in diameter and 300-mm long. This tube was kept along with the forced-air set-up in the cold chamber. Since there was some air circulation inside the covered tube, the correct air approach velocity value was later obtained from the regression equation established with the ball cooling rates.

Prior to each test, each ball was preheated to 22oC by immersion in the same 10-L thermostatic bath described previously for one hour. The ball was dried with absorbent paper, placed on the triangular support

of air inlet tube of the forced-air set-up and submitted to the cooling process. The trials were completed when the temperature in the center of the ball reached 3.5oC. The cooling rates of the balls (CR) were calculated and used to establish the effect of AAV based on a non-linear regression analysis using the statistical software R (Gentleman and Ihaka 2003).

RESULTS AND DISCUSSION Produce simulator physical properties The physical proprieties of the balls were measured and proved to be uniform spheres of 52.36±0.08 mm in outer diameter and 125.5±2.5 g in mass. Other than color, no physical differences were found amongst the balls.

Since the balls had regular geometry their density was calculated (1.67 g·cm-3) directly from their mass and volume and compared to literature data. Phenolic vinyl (1.95 g.cm-3, PlasticsUSA 2003) was found to be the polymer with the closest properties with respect to specific gravity.

Cooling rate index The frequency distribution of the CR index clearly illustrated two main groups of balls (Fig. 3). The Duncan test (Table 1) showed a difference at a level of significance of 0.05 in the CR index of the balls and divided them in four different groups. The 63.1% difference found in the CR index between the group 1 and 4 can not be explained only by the significant difference between their densities (2.2%). The experimental method, the procedure, and the data acquisition system were carefully verified but no explanation was found to justify this large difference. Communications were made to tract the manufacturer and get more information about the balls without success. Thus, only balls from the 0.14-nominal group were used to determine the effect of AAV on CR.

Specific heat determination The results from the measurement of the specific heat showed no significant difference between the two main groups of balls (Table 1). The average specific heat (Cp) of the balls was 1.112 ±0.0616 kJ·kg-1·oC-1. Phenolic vinyl was also found to be the polymer with the closest specific heat, which is 1.17 kJ·kg-1·oC-1 (Durez Corporation 2000). This Cp value is comparable to dried product such alfalfa (1.047 kJ·kg-1·oC-1), fruits (1.26 to 1.34 kJ·kg-1·oC-1), beans (1.26 kJ·kg-1·oC-1) and corn (1.17 kJ·kg-1·oC-1) (Mohsenin 1980). However, this value is much lower than the specific heat of fresh fruits and vegetables which is generally around 3.68 kJ·kg-1·oC-1 (ASHRAE, 2002). The specific heat of the produce simulator could be different from the real produce as observed by different authors; for example: Minh et al. (1969) used acrylic plastic spheres (Cp = 1.47 kJ·kg-1·oC-1) to predict heat transfer of potatoes during forced-air

43

precooling with success; and, Castro et al. (2003), Vissoto (1999), and Émond et al. (1996) obtained important findings from hollow spheres filled with a liquid solution (Cp ≈ 4.18 kJ·kg-1·oC-1).

Thermal conductivity and thermal diffusivity The results obtained for thermal conductivity and thermal diffusivity are shown on Table (1) and four different groups of balls were identified. The results obtained for thermal diffusivity are about 2.5 times higher than mean values observed for fruits and vegetables (0.15 mm2·s-1) (ASHRAE 2002). However, the ball thermal conductivity (683.9 W·mm-1·K-1) compared well to produce such as pear (595 W·mm-

1·K-1), beet (601 W·mm-1·K-1), and carrot (669 W·mm-

1·K-1) (ASHRAE 2002).

Indirect air velocity measurement The response of the balls to the forced air system using different air approach velocities are presented in Figure 4. As predicted, the cooling rate is higher when AAV is increased. As suggested by Lambrinos and Assimaki (1997) and Émond et al. (1996), a regression considering an exponential approach was performed (Eq. 6) between ball CR (min-1) and air approach velocity (AAV, m·s-1).

CR = -0.0402 eAAV (R2=0.958) (6) However, a better goodness of fit coefficient (R2)

was obtained when adding the effect of the CR index of each ball (CRIb, min-1) on CR (Eq. 7). This new approach allowed the explanation 98.6% (R2) of the variation of the cooling rate of the balls. CR = -0.02615 ln (AAV) + 0.6989 CRIb (R2=0.986) (7)

The inverse of this relation allows the determination of AAV around a ball, based on CRIb (min-1) and CR (Eq. 8). This method will allow the evaluation of the effect of the design of a container on the inside air distribution (Castro et al. 2004).

AAV = e26.73 CRIb

– 38.24 CR (8) For the lowest values of air velocity, the effects of

natural convection of the air and conduction have higher importance compared to the effect of forced convection around the ball. Furthermore, the balls submitted to still air obtained a cooling rate corresponding to the predicted extrapolate AAV value of 0.065 m·s-1. This value is equal to half of the minimum studied air velocity in forced-air measurements. The latter clearly shows that using CR as an indirect tool to measure the mean air velocity around a body has its limitation. Considering the balls used in this study, it might be concluded that any cooling rate lower than 0.025 min-1 couldn’t be used for airflow measurement and is likely due to natural convection.

CONCLUSION A material was tested to be used as produce simulator in forced-air cooling investigations. The results indicated satisfactory stability and accuracy of the material evaluated. Although further investigations may be necessary to explain the difference in the cooling rate index of each individual ball, the results obtained allow using this type of material for indirect airflow measurement after determining the cooling rate index of each ball.

A significant correlation was obtained between cooling rate of the produce simulator and air approach velocity when the individual cooling rate index effect was considered. Therefore, the use of the instrumented polymer balls will allow investigating air distribution for different package opening configurations (Castro et al. 2004).

ACKNOWLEDGEMENT This project was accomplished with the financial support from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), São Paulo, SP, Brazil, and the Horticultural Research and Development Centre of Agriculture and Agri-Food Canada, Saint-Jean-sur-Richelieu, Qc, Canada. The authors would like to thank also Bernard Goyette, Emmanuelle Tang-line-foot, and Naro R. Markarian for the technical support, and Dr H.S. Ramaswamy, McGill University, for his scientific support.

REFERENCES Alvarez, G. and D. Flick. 1999a. Analysis of

heterogeneous cooling of agricultural products inside bins. Part I: Aerodynamic study. Journal of Food Engineering 39: 227-237

Alvarez, G. and D. Flick. 1999b. Analysis of heterogeneous cooling of agricultural products inside bins. Part II: Thermal study. Journal of Food Engineering 39: 239-245

ASHRAE. 2002. Thermal properties of Foods. In: Refrigeration. American Society of Heating, Refrigerating and Air Conditioning Engineers, Inc. Atlanta, Georgia. Chapter 8.

ASHRAE 2001. Measurement and Instruments. In: Fundamentals Handbook. American Society of Heating, Refrigerating and Air Conditioning Engineers, Inc. Atlanta, Georgia. Chapter 13.

Castro, L.R., C. Vigneault and L.A. B. Cortez. 2004. Effect of container opening on efficiency of horticultural produce cooling. Transaction of the ASAE. (In press).

Castro, L. R., C. Vigneault and L. A. B. Cortez. 2003. Effect of container area and position on horticultural crop half-cooling time. Montréal: CSAE/SCGR 2003 Meeting. Paper no. 03-309.

Chau, K. V., J.J. Gaffney, C. D. Baird and G. A. Church. 1985. Resistance to airflow of oranges in

44

bulk and in cartoons. Transactions of ASAE. 8(6): 2083-2088

Dincer, I. 1994. Precooling of cylindrically shaped grapes: experimental and theorical heat transfer rates. Journal of Food Process Engineering. 17: 57-71

Durez Corporation. 2000. Dallas, Texas, USA. http://www.durez.com/datasheets/ 31735.pdf

Émond, J. P., F. Mercier, S.O. Sadfa, M. Bourré and A. Gakwaya. 1996. Study of parameters affecting cooling rate and temperature distribution in forced-air precooling of strawberry. Transactions of the ASAE, 39(6): 2185-2191

Fontana, A. J., J. Varith, J. Ikediala, J., Reyes, and B. Wacker. 1999. Thermal properties of selected foods using a dual needle heat-pulse sensor. Toronto: ASAE/CSAE-SCGR 1999 Meeting. Paper no. 996063, 10p.

Gentleman, R. and R. Ihaka. 2003. The R Project for Statistical Computing. Statistics Department, University of Auckland. Auckland, CA, USA. http://www.r-project.org

Gorini, F. and G. Borinelli. 1974. Some trials of salad precooling. Acta Horticulturae. 38: 465-490

Goyette, B., C. Vigneault, B. Panneton and G. S. V. Raghavan. 1996. Method to evaluate the average temperature at the surface of a horticultural crop. Canadian Agricultural Engineering. 38(4): 291-295

Hackert, J.M., R.V. Morey and D.R. Tompson. 1987. Precooling of fresh market broccoli. Transactions of the ASAE. 24(4): 1073-1076

Irvine, D. A., D.S. Jayas and G. Mazza. 1993. Resistance to airflow through clean and solid potatoes. Transactions of ASAE. 36(5): 1405-1410

Kader A.A. (ed) 2002. Postharvest technology of horticultural crops. 3rd edition. Coop. Ext. Uni. of Ca. Div. Agric and Nat. Res. Univ. of CA, Davis, CA. Publ. no. 3311. 535p.

Lambrinos, G. and H. Assimaki, 1997. Air precooling and hydrocooling of Hayward kiwifruit. Acta Horticulturae. 44 (2): 561-566

Leyte, J. C., and C. F. Forney. 1999. Optmizing flat design for forced-air cooling of blueberries packaged in plastic clamshells. HortTechnology. 9(2):202-205.

Lindsay, R. T., M. A. Neale, H. J. M. Messer. 1983. Ventilation rates for the positive ventilation of vegetables in bulk bins. Journal of Agricultural Engineering Ressources. 28: 33-44

Maul, F., C. Vigneault, S.A. Sargent, K.V. Chau and J. Caron. 1997. Nondestructive sensor system for evaluation of cooling efficiency. Proceeding from the Sensors for Nondestructive Testing

International Conference and Tour. February 18-21. p. 351-360

Minh, T. V., J.S. Perry and A.H. Bennett. 1969. Forced-air precooling of white potatoes in bulk. ASHRAE Transactions. 75(2): 143-152

Mohsenin, N.N. 1980. Thermal properties of foods and agricultural materials. Gordon and Breach. New York, NY. 405 p.

Nakayama, A. and F. Kuwahara, 1999. A macroscopic turbulence model for flow in a porous medium. Journal of Fluid Engineering. 121: 427-433

PlasticsUSA. 2003. http://www.polymerweb.com/_misc/specgrav.html. Specific density of major polymers.

Ramaswamy, H.S., K.V. Lo and M.A. Tung. 1982. Simplified equations for transient temperature in conductive foods with convective heat transfer at the surface. Journal of Food science. 47: 2042-2047

SAS Institute. 1988. SAS/STAT User’s guide, Release 6.03 ed. SAS Institute, Cary, North Carolina.

Smale, N. J., D. J Tanner, N. D. Amos, A. C. Cleland. 2003. Airflow properties of packaged horticultural produce - a practical study. ISHS Acta Horticulturae. 599: 443-450

Tanner, D. J., A. C. Cleland, L. U. Opara, 2002. A generalized mathematical modelling methodology for the design of horticultural food packages exposed to refrigerated conditions. Part 2: Heat transfer modelling and testing. International Journal of Refrigeration. 25 (1): 43-53

Van der Sman, R. G. M. 2002. Prediction of airflow through a vented box by the Darcy-Forcheheimer equation. Journal of Food Engineering 55: 49-57

Verboven, P., M. L. Hoang, A. Schenk, B.M. Nicolai. 2001. Effect of pallet box stacking on cooling of horticultural products in cool stores. Acta Horticulturae. 566: 239-244

Vigneault, C., B. Goyette, and G.S.V. Raghavan. 1995. Continuous flow liquid-ice system tested on broccoli. Can. Agric. Eng. 37(3):225-230

Vigneault, C and J.P. Émond. 1997. Reusable container for the preservation of fresh fruits and vegetables. United States Patent Application Office Washington, WA Dc, USA. Patent Number: 5,727,711. 60p.

Vissoto, F. Z. 1999. Estudo dos parâmetros que influenciam no pre-resfriamento de um leito de produtos esfericos disposto no interior de uma embalagem de papelao. Master thesis. Unicamp, Campinas, SP, Brasil. 115p.

Weast, R. C. and M.J. Astle. 1979. Handbook of Chemistry and Physics. 59th ed. CRC Press Inc. 119 p.

1

45

Table 1: Duncan's Multiple Range Test results for cooling rate index, specific heat, thermal diffusivity, and thermal conductivity of the four groups of produce simulators.

Group nominal

name

Balls per

group

Diameter mm

Mass g

Density g cm-3

Cooling rate index min-1

Specific heat J·g-1·K-1

Thermal diffusivity

mm2·s-1

Thermal conductivityW·mm-1·K-1

0.14 85 52.37a 124.5a 1655a -0.141±0.007a 1.12±0.06a 0.371a 683.9a 0.15 1 52.43a 125.0b 1656a -0.155b Not tested 0.412b 759.7b 0.17 3 52.30a 126.1c 1683b -0.172±0.019c Not tested 0.453c 847.6c 0.23 39 52.36a 127.1d 1692b -0.230±0.005d 1.10±0.06a 0.605d 1138.6d

Numbers with the same letter within a column are not significantly different according the Duncan statistic test at a level of significance of 0.05.

52.36 mm Ø

26.18 mm

1.59 mm Ø

thermocouple

glue

Figure 1. Instrumented ball used as a produce simulator.

Figure 2. Experimental set-up used for calibration trials.

46

0

5

10

15

20

25

30

35

0.10 0.15 0.20 0.25

Cooling rate (min-1)

Num

ber o

f bal

lspe

r int

erva

l

Figure 3. Distribution of the cooling rate index of the balls per interval of ±0.0025 min-1.

-0.12

-0.10

-0.08

-0.06

-0.04

-0.02

0.0 0.2 0.4 0.6 0.8 1.0Air approach velocity (m s-1)

-0.14213

-0.14227

-0.14227

-0.14230

-0.13773

-0.13980

-0.14230

-0.14317

P di t d l (E 4)

Ball cooling rate

Coo

ling

rate

min

-1

Figure 4. Cooling rate results from trials and from prediction with Equation 6

.

47

Artigo 3. Indirect airflow distribution measurement for horticultural crop

package. Part II: Verification of the research tool applicability

Clément VIGNEAULT1,2,3, Larissa R. DE CASTRO1,2, Bernard GOYETTE1

Naro R. MARKARIAN1, Marie T. CHARLES1, Gaétan BOURGEOIS1, Luis A. B. CORTEZ2

1Horticultural Research and Development Centre Agriculture and Agri-Food Canada

430 Gouin, Saint-Jean-sur-Richelieu (Qc) CANADA J3B 3E6

3 Corresponding author Tel : 450-346-4494 ext. 170 Fax: 450-346-7740 Email : vigneaultc@agr.gc.ca

2College of Agricultural Engineering State University of Campinas

Cidade Universitaria Zeferino Vaz 13083-970, Campinas SP, BRAZIL

ABSTRACT A group of 64 plastic balls were instrumented and used as horticultural produce simulators and strategically distributed in an orthogonal matrix along with other 448 plastic spheres to simulate precooling of column stacked produce. This research tool was based on the correlation between the cooling rate of produce simulators, their cooling indexes, and the air approach velocity. The applicability of using these instrumented balls as an indirect measurement of air velocity was evaluated. A second method was also proposed to analyze the effect of the container design on produce cooling rate. Correlations were determined by measuring the half-cooling time of 64 instrumented simulators positioned at fixed locations inside a two-end-fully-open ball matrix and submitted to different airflows. The surrounding air velocity was inferred as a function of the simulator locations in reference with the air entrance. These two methods were evaluated comparing the data obtained for three package opening areas (0.67%, 2%, and 6%), and six airflow rates (ranging from 0.125 to 3.9 L·s-1·kg-1), and performing a mass balance. This comparison showed only slightly better results in predicting the variation of the cooling rate using the second method, but this method produced a much better performance in predicting the airflow rate. A statistical analysis was also performed to determine the number of replications required to discriminate the limit of significant effect of airflow rate and opening percentage on produce cooling rate and its uniformity. KEY WORDS Produce simulator, container, forced-air cooling, half-cooling time, air approach velocity, precooling. INTRODUCTION Because of the significant levels of losses registered for horticultural produce during postharvest operations and increase of energy cost, the reduction of the general process costs is intensely searched as well as their quality conservation. Among postharvest processes, optimal precooling process generally plays the most important role in the achievement of both aims. The efficiency of this process affects produce quality considering the rapidity and uniformity of temperature reduction. Heterogeneous cooling may cause moisture loss, freezing and severe loss of quality by microbiological infection (Alvarez and Flick, 1999b; Van der Sman et al., 1996). Precooling efficiency influences the grower benefits not only when considering the advantages of shelf-life extension, but also the expenses of acquisition and operation of the refrigeration system. In this case, process efficiency and air pressure drop through the produce and openings of the package affect the required energy to operate the precooling system.

The design of a container is decisive in promoting sufficient and uniformly distributed fluid through the produce during forced-air precooling method (Castro et al. 2004). The container must have enough opening area allowing uniform and fast air circulation through the produce without over passing its mechanical resistance (Vigneault and Goyette 2002a; 2002b). A well-designed container must result in adequate produce conservation, as well as avoiding excessive energy input (Vigneault et al., 2002).

A wide variety of containers have been developed over time for fruit and vegetable market. However, the selection of the area and position of their openings seems arbitrary. Detailed recommendations have not been documented yet. The different existing recommendations are based in so specific experimental conditions that they may be found even contradictory (Van der Sman, 2002; Vigneault et al., 2002; Kader, 2002). Considering the holes position on the package surface, despite enough evidence of its influence on air distribution, there are no clear conclusions whether

48

it may produce (Leyte and Forney 1999; Émond et al., 1996) or not produce (Castro et al., 2004; Smale and Tanner, 2003) a significant effect on cooling efficiency.

An original concept was developed for the indirect measurement of the mean air velocity around a produce simulator during forced-air precooling (Vigneault and Castro, 2004). The resulting method has potential to become an accurate and useful tool for container design since it should allow air velocity profile prediction within horticultural produce containers with precision. The objectives of the present part of the research were: 1) to verify the applicability of this indirect airflow measuring method in revealing the airflow pattern through packed produce; 2) to exploit new avenue to increase the accuracy of indirect airflow measurement; 3) to determine a procedure to validate the results obtained; and 4) to determine the effect of the number of replications on the experimental design capability to identify significant differences between treatments.

MATERIAL AND METHODS A research tool was proposed by Vigneault and Castro (2004) to investigate the air distribution in containers during forced-air precooling process. This method based on the circular-area mean approach air velocity (CAV) consisted of using a relationship between the cooling rate (CR) and an individual ball cooling index for indirect measurement of surrounding air velocity. The applicability of this method was evaluated using the following experimental device.

Produce simulator The horticultural produce simulator described in detail by Vigneault and Castro (2004) consists of solid polymer balls 52.36 mm in diameter and weighing 125.55 g. Sixty-four balls were selected for their relatively high uniformity in terms of cooling index (-0.1414 ± 0.0081 min-1) and heat capacity (1.1252 ± 0.0657 kJ·kg-1·oC-1). Each of the 64 balls was instrumented with a 30-gage 5 m-long insulated copper constantan thermocouple wire (Type T) placed in their center with a precision of ±0.025 mm.

Experimental set-up Sixty four instrumented balls were stacked uniformly distributed along with other 448 balls on a columnar pattern to form a cubic matrix of 8-ball-side dimension (Figure 1). Table 1 presents the orthogonal positioning reference system of the instrumented balls; the z axe corresponding to the airflow direction. The arrangement resulted in 47.64% of porosity. The balls forming the two end layers on the “z” direction were assembled together using 12-mm-long and 6-mm-diameter plastic pins inserted into 6mm depth hole perforated at each ball to ball or ball at wall contact point. Thirty groups of five balls each were

also assembled to form five-ball stars to insure sufficient stability through the matrix. These five-ball stars were assembled using one centered surrounded by the four other balls in contact within a single geometric plan, using the same 12-mm-long plastic pin assembling system, and distributed within the matrix.

Figure 2 shows the experimental set up used during the trials. Four transparent acrylic plates were assembled to simulate a forced-air cooling tunnel of 420 mm inside square cross-section, and 1250 mm long. The ball matrix was positioned at a distance of 220 mm from the tunnel-end air inlet. The portion of the tunnel containing the balls was insulated with a 25 mm-thick polystyrene foam to reduce heat conduction. The air-outlet of the tunnel consisted on a 610 mm long plenum enabling air pressure drop (APD) measurements across the ball matrix using a pressure transmitter in the range of 0-127±6 mm of water (Model 607-7, Dwyer Instruments Inc. Michigan City, IN, USA).

The end of the air-outlet tunnel was air-tightly attached to the aspiration chamber built from 10 mm thick plywood with a 900x600x600 mm of outer dimensions. The aspiration chamber was divided in two inner compartments, so-called upper and lower divisions. A 0.75 kW, direct drive radial blade fan (Model PW-11, Peerless Electric, hot springs, NC, USA) driven by a 0.75 kW variable speed motor (Model G344, Marathon Electric, Wausau, WI, USA) was fixed between the two inner divisions. A variable AC motor drive (Model M1215SB, AC Tech, Uxbridge, MA, USA) connected to the motor allowed the proportional control of the fan speed. The fan created a negative pressure in the upper chamber and forced the outside air to circulate through the cooling tunnel. The air was released to the atmosphere through a 500 mm long, 101.6 mm-diameter tube instrumented with a Pitot-tube device allowing airflow measurements. Two transmitters, 0-12.7+0.6 and 0-25.4+0.13 mm of water (Models 607-2 and 607-3, Dwyer Instruments Inc. Michigan City, IN, USA) were used to measure the static and total pressures at the Pitot tube. Two other 500 mm long tubes of 76.2 and 31.8 m inner diameters were air tightly assembled in the 101.6 mm diameter tube to measure lower airflow. These three tubes were used to provide acceptable measurement of the dynamic pressure. The whole experimental set-up was placed in a cold chamber maintained at 4oC to generate precooling process to the ball matrix.

A fully open configuration was initially tested to determine the correlation between the half-cooling time (HCT) of the 64 produce simulators and air approach velocity. Based on the ball matrix volume, the airflow rates studied were equivalent to 0.125, 0.25, 0.5, 1, 2, and 3.9 L·s-1·kg-1 of apple (Vigneault et

49

al., 2004). Three other opening configurations were investigated by placing a pair of plates next to the first and eighth layers of balls to enclose the matrix within a simulate two-perforated-side package. The pair of square plates was made from 3-mm-thick 420 mm polypropylene plates and perforated using circular metal saws. Nine holes 38.6 mm in diameter, or 0.67% of the total plate area, were uniformly distributed on the plate surface (Figure 3). Three total opening areas (TOA) were evaluated and corresponded to 0.67, 2, and 6% of the plate area using the central hole, the central line of holes, or the nine holes, respectively. These configurations were obtained by covering the other holes with airtight seal tapes as necessary.

The pressure drop obtained with the 1, 2 and 3.9 L·s-1·kg-1 airflow rates with the central hole configuration overpasses the limit capacity of the fan. Thus, the four airflow rates were readjusted for this configuration to 0.125, 0.25, 0.5, and 0.75 L·s-1·kg-1. For similar reasons, the experiments with the plates with 3 holes were conducted only up to a 2 L·s-1·kg-1 airflow rate. Furthermore, the transmitter upper-limit of 125 mm of water was settled as the limit for pressure drop in the experiments, since higher values would not be considered as practical for commercial storage rooms (Kader, 2002).

A 520 x 840 x 1100 mm heat-exchanger was built to minimize the temperature variation at the air inlet during the experiments resulting from the oscillation of the cold chamber temperature. This heat-exchanger consisted of 30 perforated aluminum plates, 520 x 1100 mm and 1.04 mm in thickness, spaced at every 25 mm. These plates were uniformly perforated with 4.76 mm diameter holes resulting in a 44 % opening area. They were attached together and insulated longitudinally with a 50 mm-thick polystyrene foam. The laterals of the exchanger were wrapped with plastic to avoid air infiltration. One end-side of this heat exchanger was also insulated with a polystyrene frame where the forced-air tunnel could be attached during the experiments (Figure 2). Thus, the air suctioned from the cold chamber by the fan was forced to pass first through the 30-layer aluminum plate heat exchanger before reaching the balls.

Instrumentation and Control All instruments and motors in the experimental setup were monitored and controlled using a data acquisition system (34970 Data Acquisition/Switch Unit – Agilent Technology, HP, Palo Alto, California, USA) via a custom made desktop software using VEE language (Agilent Technology, HP, Palo Alto, California, USA). Preliminary tests were performed to tune the control system; the relationship between the voltage required by the motor to provide a certain air velocity, with respect to the various number of openings was determined and integrated into the software. The

airflow rates were calculated by computing the air velocity using the pressure velocity relationship (ASHRAE, 2001) of the Pitot tube and proportionally adjusting the fan motor through the AC motor drive. At the start of each experiment, the user was prompted to input the opening distribution and the set point air velocity. The experiment was subsequently entirely automated. The temperature inside the sixty-four balls along with air temperature before and after crossing the ball matrix, the temperature in the centre of the cold chamber, the pressure drop through the ball matrix and the plates, and the dynamic pressure of the air, were simultaneously recorded at a 20-s interval. The set point velocity was maintained due to a continuous error monitoring system which readjusted the fan voltage as necessary.

Experimental procedure The four opening configurations were tested with the different airflow rates and repeated three times. Prior to the start of each test, the forced-air tunnel containing the balls was placed in a warm chamber maintained at 27±1.0oC. An axial fan circulated ambient air through the ball matrix for about 120 minutes which resulted in a uniform temperature at the center of the 64 instrumented balls. After this conditioning period, the perforated plates were installed and the tunnel was placed in the cold room. The tunnel-end air inlet was connected to the aluminum heat-exchanger and the air outlet to the aspiration chamber and the centrifugal fan was turned on immediately. The data were recorded until the temperature of the warmest ball had reached 6.9oC, which correspond to a 7/8 cooling process (Goyette et al., 1996) and at which point the software terminated the control process and turned off all devices. The temperature-time data recorded was used to calculate the HCT and CR of each ball for all treatments by using a dedicated ExcelTM macro developed by Goyette et al. (1996).

Air-mass balance research tool evaluation Two different approaches were considered and resulted in two methods that were compared to determine their potential in indirectly measuring the mean approach air velocity around an instrumented ball. The first method consisted of using the general exponential correlation developed by Vigneault and Castro (2004) considering the circular-area mean air approach velocity (CAV, m·s-1) around the instrumented balls based on the circular area of an air-inlet tube holding the ball as the total opening area. This equation was applied to calculate the air approach velocity from the CR of the 64 balls in the matrix. At the two lowest airflows (0.125 and 0.250 L·s-1·kg-1), the natural convection was predominant and produced CR lower than 0.025 min-1. These values prevented the

50

use of the correlation to establish air velocity profile in this range of airflow (Vigneault and Castro, 2004).

For the second method, the fully open matrix was submitted to the six airflow rates and the HCT (min) was determined for each ball. A square-cross-section mean air approach velocity (SCSV, m·s-1) was calculated for each airflow rate based on the square cross-section of the forced-air cooling tunnel as the total area. A specific correlation was developed to explain the relation between this SCSV and the CR of each ball, considering that the ball was at a specific position in the matrix. Thus, the new equations included the specific thermal properties of each ball (Vigneault and Castro 2004) and the variability of the air movement around each ball according to its specific position in the matrix.

Both methodologies were used to infer the air approach velocity at the 64 different positions inside the container from the HCT. They were evaluated with the data obtained for the three container opening areas (0.67%, 2%, and 6%), and the airflow rates ranging from 0.125 to 3.9 L·s-1·kg-1. The mass of air circulation around each ball was calculated based on the assumption of a vertical and horizontal symmetry of air circulation. Therefore, the velocities of the air around the other positions occupied by the 56 non-instrumented balls in each layer were determined from the hypothesis of symmetry in x and y directions. Thus, the calculation based on each of the two methods resulted in empiric masses of air crossing each z-direction layer. The total masses of air calculated with these two indirect airflow measuring methods were compared to the mass of air measured by the airflow measuring device through a Tukey test using SPSS v. 11.5 (SPSS Inc. 2004. Chicago, Illinois. USA).

The air-mass balance analysis considered the air approach velocity obtained with the three cointainer-opening-area data in the surroundings of the balls contained in each z-direction layer. Therefore, a one-way ANOVA and Tukey test were performed on SPSS v. 11.5 to identify any significant difference between the masses of air calculated in each layer Z with the two indirect measuring methods.

Outlier rejection A result validation procedure was developed based on the identification and rejection, within a confidence level of 99.9%, of any HCT replication standard deviation larger than the outlier upper limit, which is equal to the mean standard deviation plus 3.09 standard deviations (Montgomery, 1996). Thus, the standard deviation of the HCT was calculated for each ball and plotted as a function of the airflow rate. A regression equation was calculated and the outlier upper limit was calculated for each airflow rate. The outliers were identified and rejected. New regression

equations and outlier upper limits were calculated from the remaining HCT results and so on, until any outlier was identified as recommended by Montgomery (1996).

Number of replications The determination of the effect of the number of replication (n) on the experimental design power for identifying significant differences between treatments was performed based on the equation (1).

nsFd

22 = (1)

Where F is a tabulated Fisher-Distribution value for the desired confidence level and the degree of freedom of the initial sample, s2 is the variance of the samples, and d is the half-width of the resulting confidence interval (Steel and Torrie, 1980).

RESULTS In general, the whole set up, including all the measurement and control systems and software, allowed practical procedures and precise performance when executing the experiments. The HCT and CR calculated from the temperature variation at the center of each of the 64 instrumented balls inside the matrix for each opening configuration and airflow rate were mostly accurate.

The HCT obtained with the fully-open configuration from four randomly chosen balls 1, 21, 39 and 56 from the odd z-layers, 1, 3, 5 and 7 respectively, are presented in Figure 4, as examples. The experimental system demonstrated great stability and a good relation between HCT with the airflow rate for all ball positions with the smaller regression goodness of fit coefficient (R2) equal to 0.9589 (result not presented). The fully-open configuration also showed a significant effect of the airflow rates on the variance of HCT (F5,

378 = 9.692, P<0.0005). The highest variances obtained corresponded to the minimum airflow rate (0.125 L·s-1·kg-1).

Effect of airflow rate on replication standard deviation The equation (2) shows the effect of the airflow rates (AFR) on the replication standard deviation (STDr). The airflow rate explains 93.41% of the replication standard deviations. STDr = 0.6326 AFR-0.4538 R2 = 0.9341 (2)

This equation was used to calculate the mean standard deviation of the results corresponding to each air flow and calculated the outlier limits.

Outlier rejection Among the 1152 HCT calculated (3 replicates, 6 airflow rates, 64 balls), only four results were rejected as being identified as outliers by the Montgomery (1996) method (Figure 5). Three outliers were from the lowest airflow rates tested, likely showing more

51

variability and instability in the experimental results while using very low flow rate. In fact, the variability increased with a decrease of the airflow rate (Table 2). The forth one outlier was from the 0.5L·s-1·kg-1 airflow rate, which will be discussed later. The remaining data obtained from the fully-opened configuration were used to establish the square-cross-section mean air approach velocity (SCSV, m·s-1) and the specific correlation between the SCSV and the HCT of each ball positioned at a permanent location in the matrix (Fig. 6).

This validation technique was also performed to identify and reject any other outlier while using the other opening configurations. When it was possible, the causes for producing outliers or wrong results during the following experiments were identified and corrected; one experiment containing manipulation error was also performed again.

Number of replications The effect of the number of replication (n) on the experimental design power for identifying significant differences between treatments is presented in Table 2. As expected, the minimum difference between two HCT results to be considered as significantly different at a level of confidence of 95% decreased with the increase of airflow rate and the number of replicate. These results could be used to determine the number of replications required to demonstrate the effect of any airflow rate or opening configuration as long as the expected mean and variance of two experimental conditions are known. In the present case, one replicate of each of the six airflow rates would have been sufficient to discriminate the effect of the airflow rate used. However, three replicates were necessary to discriminate the effect of different opening configurations.

Research tool evaluation based on air-mass balance CAV method

The equation 3 explained 98.65% (R2) of the variation of the cooling rate of the balls (CR, min-1) as a function of the circular-area mean air approach velocity (CAV, m·s-1) and the individual cooling index of the balls (CRIb, min-1) (Vigneault and Castro, 2004). The equation 4 used to produce the mass balance according to the CAV was developed from the equation (2) and the relation between half cooling time (HCT, min) and the CR (min-1) (Equation 5) (Goyette et al., 1996).

bCRI 0.6989 ln(CAV) -0.02615CR += R2=0.9865 (3) -1HCT 26.51 ICR 26.73e CAV += (4)

-1ln(0.5)CR HCT = (5) The mean air velocities based on the CAV produced

a mass balance explaining 92.9% of the variation of the airflow rates measured mechanically and showed a

significant correlation between the results of these two measuring methods (F6, 353 = 610.2, P<0.0005). However, there were no significant differences between the air velocities calculated with that method in the airflow rate range of 0.125 to 0.5 L.s-1.kg-1 and no significant difference were either detected between the air velocities calculated at 0.5 and 0.75 L.s-1.kg-1

and 0.75 and 1 L.s-1.kg-1 (Table 3). Finally, the CAV method produced always lower mass flow result then the mass flow measured mechanically. This lower result could be explained by observing that this methodology considered one ball being cooled separately at a time during the calibration of the method. Therefore, the air circulating around the ball did not reach any obstacle and flow fairly well all around the ball surface. However, when performing the test using the air flow through the ball matrix, the actual surface of the ball directly exposed to cold air was considerably reduced. In fact, in a columnar stacking pattern, the balls are touching six other objects (other balls, plastic slabs or tunnel walls), creating several obstructions to the air circulation. These obstructions generate a lower HCT that would have been resulted without any obstruction which results in lower airflow rate calculations.

SCSV method The mass balance based on SCSV (m.min-1) was based on the 64 equations developed from the whole matrix in fully open configuration submitted to the 6 airflow rates. A regression analysis was performed (Eq. 6a) and showed a good overall correlation (R2=0.8841) when considering all the results as a whole (Fig. 6). Individual ball performances were then used to determine a correlation equation between air velocity (SCSV, m.min-1) and the half-cooling time (HCT, min) for each ball (Eq. 6b). The 64 resulting equations explained 98.77 ± 0.89% (R2) of the variation of the CR of the balls for the operating conditions tested. The a and b empirical parameters and individual goodness of fit coefficients (R2) of eight equations are presented in Table 4, as examples. These equations were then used to produce the mass balance according to the SCSV method. HCT = 25.416 SCSV-0,6406 R2=0. 8841 (6a) HCT = a·SCSVb R2=0.9877±0.0089 (6b)

The results of this mass balance demonstrated that the balls area taken into account allowed determining the mean air velocity around each instrumented ball. The mean air velocities based on the SCSV produced a mass balance explaining 99.3% of the variation of the airflow rates measured mechanically and showed a significant correlation between the results of these two measuring methods (F6, 353 = 4405.2, P<0.0005). This method allowed to discriminate the significant difference between each air velocity calculated

52

corresponding to the different airflow rate tested (Table 3).

Finally, the highest level of correlation achieved with SCSV equations, compared to the CAV results, lead to consider this method as the indirect airflow measuring method to be used, revealing more accurately the air mass flowing through the balls matrix. Air-mass balance in z-direction Lower CAV mass of air were calculated at the deepest layers in z direction and showed a significant negative Pearson correlation with z-direction at 0.05 level (Pearson144 = -0.175, P2-tailed =0.036). These results were likely due to the use of solid balls producing a thermal mass effect.

The SCSV approach developed included the warming effect of the air when crossing the matrix and reaching the balls at the rear layers. Therefore, the SCSV approach validated the velocity symmetry hypothesis assumed for the z-depth direction and did not show any air-mass significant correlation with z-direction (Pearson144 = 0.010, P2-tailed =0.908).

An analysis of variance for both SCSV and CAV by z-direction at each airflow rate was performed (Table 5). For SCSV, the position of the ball in z-direction had a significant effect on the measurement of the airflow when ranging from 0.125 to 1 L.s-1.kg-1. This global difference was generally due to the significant difference showed between the first layer (z=1) and the other z layers. For 1 L.s-1.kg-1 the velocities obtained in that layer were significantly higher only than the values found on z=8. The CAV air velocities showed significant differences between the 8 layers for all airflow rates tested (Table 5).

The graphics of the percentage of the total mass of air calculated at each z-layer and the minimal and maximal airflow rate tested for both indirect air velocity measuring methods (Figure 7) showed a drastic increases of the airflow-rate-indirect-measurement accuracy and precisions when using the SCSV method (107.4 ± 5.94% and 105.8 ± 5.75%) compared to CAV method (109.3 ± 17.49% and 37.7 ± 7.16%), respectively. Furthermore, the SCSV method results were fairly close to the results obtained from the Pitot tube measuring method for each z-direction layer and obtained a 100.99 ± 9,18% overall average of the air mass but gave significantly different results when varying the airflow rate (Figure 8). The same figure shows an increase of mass balance performance as the airflow rate increased. In fact, fairly constantly lower results were obtained when experimenting with the 0.5 L.s-1.kg-1 giving a 85.77% mass balance difference. This lower result is explained by the regression equation underestimation of the HCT as a function of the airflow rate at this particular airflow rate (Figure 9). The calculation of the

Reynolds number resulted on 2081, which corresponds to the transient phase of airflow pattern from laminar to turbulent. Thus, a particular study should be performed to clarify the correlation between HCT and airflow rate for this particular case to increase the accuracy of the setup.

The performances of the SCSV method are considered as very good for an indirect measurement method of physical phenomena when compared to the mean precision of 77% obtained by Vigneault et al. (1992), which was already considered as fairly good result (Orsat et al., 1993). Furthermore, the SCSV indirect measurement method produced much more precise and more stable results and had less disturbing effect on the airflow pattern than the aluminum sphere method or direct measurement method presented by Alvarez and Flick (1999a, b).

Experimental setup performances The experimental method permitted to discriminate the airflow rate effect. In fact, using only one replication would have been sufficient to realize this discrimination. Some effects of different opening configurations were also discriminated. The analysis of variance showed that the opening area had a significant effect on the variance results of HCT (F3,

1340 = 115.047, P<0.0005). Furthermore, the results obtained allowed to demonstrate that: o The variance increases as the opening area is

reduced; o No significant difference exists between the data

for 6% and fully open; o At the minimum opening configuration (total

surface area = 0.67%), the airflow rate has a significant effect on the HCT variance (F3, 252 = 22.47, P<0.0005). The highest variation could be observed for the data produced with 0.5 L·s-1·kg-1;

o The 0.125 and 0.25 L·s-1·kg-1 airflow rate produces lower variance.

The performance obtained showed the great potential of the experimental setup and the SCSV method to discriminate the different effects of the airflow rate and opening configuration. Further research is necessary to establish these effects, but the experimental setup is considered as being able to discriminate between them.

CONCLUSION Two indirect airflow rate measuring approaches were investigated in terms of their applicability of investigating the airflow profile through a porous medium. An experimental setup allowed calibrating the two methods. Application of statistical methods permitted to eliminate some experimental error and determine the effect of the number of replications on the power of the method to discriminate significant effect of airflow rate and opening configuration.

53

The use of the indirect measure of the air based on circular-area mean approach air velocity (CAV) showed some limitation. Moreover, this method did not produce accurate results concerning the expected values likely due to the thermal-mass effect.

The method based on the square-cross-section mean air approach velocity (SCSV) improved the measurement of airflow pathway since it considered the variability of the air movement in different positions in the porous medium and resulted in a more accurate mass balance through the airflow direction layers. This method allowed the consistent and precise determination of the air approach velocities for the package opening areas and airflow rates investigated. This tool showed to be very practical, accurate, reliable, and therefore useful to determine the effect of different parameters potentially affecting the air distribution and cooling rate through a porous medium.

ACKNOWLEDGEMENT This project was accomplished with the financial support from the Horticultural Research and Development Centre of Agriculture and Agri-Food Canada and the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP). The authors wish to thank Emanuelle Tang Line Foot for her assistance during the construction of the experimental setup and the experimentation, Sandra Hindson for her help during the preparation of the article, and Mickey Waxman for the technical support in the statistical analysis.

REFERENCES Alvarez, G. and D. Flick. 1999a. Analysis of

heterogeneous cooling of agricultural products inside bins. Part I: aerodynamic study. J. Food Eng. 39: 227-237.

Alvarez, G. and D. Flick. 1999b. Analysis of heterogeneous cooling of agricultural products inside bins. Part II: thermal study. J. Food Eng. 39: 239-245.

Arifin, B.B. and K.V. Chau. 1987. Forced-air cooling of strawberries. ASAE Paper no. 87-6004. St. Joseph, Mich.: ASAE

ASHRAE. 2001. Measurement and Instruments. 1998 Fundamentals Handbook. Atlanta, Ga.: Am. Soc. Heating, Refrigerating and Air-Conditioning Eng: 13.12-13.17.

ASHRAE. 2002. Methods of precooling fruits, vegetables and cut flowers. 2002 Refrigeration Handbook. Atlanta, Ga.: Am. Soc. Heating, Refrigerating and Air-Conditioning Eng: 14.1-14.10.

Baird, C. D., J.J. Gaffney, M.T. Talbot. 1988. Design criteria for efficient and cost effective forced-air cooling systems for fruits and vegetables. ASHRAE Trans. 94: 1434-1453.

Boyette, M.D. 1996. Forced-air cooling packaged blueberries. Appl. Eng. Agr. 12(2): 213-217.

Castro, L.R., C. Vigneault, L.A.B. Cortez. 2003. Container opening design for horticultural produce cooling efficiency. Int. J. Food Agr. Environ. 2 (1): 135-140.

Émond, J. P., F. Mercier, S.O., Sadfa, M. Bourré, A. Gakwaya. 1996. Study of parameters affecting cooling rate and temperature distribution in forced-air precooling of strawberry. Trans. ASAE, 39(6): 2185-2191.

Flick, D., A. Leslous, G. Alvarez. 2003. Modélisation semi-empirique des écoulements et des transferts dans un milieu poreux en régime turbulent. Rev. Int. Froid. 26: 349-359.

Goyette, B., C.Vigneault, B. Panneton, and G.S.V. Raghavan. 1996. Method to evaluate the average temperature at the surface of a horticultural crop. Can. Agric. Eng. 38(4): 291-295.

Haas, E., G. Felsenstein, A. Shitzer, G. Manor. 1976. Factors affecting resistance to airflow through packed fresh fruit. ASHRAE Trans. 82(2): 548-554.

Jamet, Y. 1984. Transfert de chaleur et de masse au cours du refroidissement des denrées périssables à bord des navires - Durée minimale de refroidissement des cargaisons de fruits et des légumes dans les cales ou à l’intérieur des conteneurs. Rev. Int. Froid. 7(5) : 299-307.

Kader A.A. (ed) 2002. Postharvest technology of horticultural crops. 3rd edition. Coop. Ext. Uni. of Ca. Div. Agric and Nat. Res. Univ. of CA, Davis, CA. Publ. no. 3311.

Kopelman, I., J.L. Blaisdell, I.J. Pflug. 1966. Influence of fruit size and coolant velocity on the cooling of Jonathan Apples in water and air. ASHRAE Trans. 72(1): 209-216.

Leyte, J. C. and C.F. Forney. 1999. Optmizing flat design for forced-air cooling of blueberries packaged in plastic clamshells. HortTechnology. 9(2):202-205.

Lindsay, R. T., M.A. Neale, H.J.M. Messer. 1983. Ventilation rates for the positive ventilation of vegetables in bulk bins. J. Agric. Eng. Res. 28: 33-44.

Orsat, V., C. Vigneault, and G.S.V. Raghavan. 1993. Air diffusers characterization using a digitized image analysis system. Applied Engineering in agriculture. 9 (1): 115-121.

SAS Institute. 1988. SAS/STAT User’s guide, Release 6.03 ed. SAS Institute, Cary, NC, USA.

Smale, N. J., D.J. Tanner, N.D. Amos, A.C. Cleland. 2003. Airflow properties of packaged horticultural produce - a practical study. Acta Hortic. 599: 443-450

54

Steel R.G.D., J.H. Torrie. 1980. Principles and procedures of statistics: A biometrical approach. 2nd. New York, NY, USA. McGraw Hill Book Compagny.

Montgomery, D. 1996. Design and analysis of experiments. 2 ed. John Wiley & Sons, New York NY.

Van der Sman, R. G. M. 2002. Prediction of airflow through a vented box by the Darcy-Forcheheimer equation. J. Food Eng. 55: 49-57.

Van der Sman, R. G. M., R.G. Evelo, E.C. Wilkinson, W.G. Van Doorn, 1996. Quality loss in packed rose flowers due to Botrytis cinerea infection as related to temperature regimes and packaging design. Postharvest Biol. Tech. 7: 341-350.

Vigneault, C. and B. Goyette. 2002a. Design of plastic container openings to optimize forced-air precooling of fruits and vegetables. Appl. Eng. Agr. 18(1): 73-76.

Vigneault, C. and B. Goyette. 2002b. Largeur des ouvertures au fond de contenants de plastique utilisés pour la manutention des produits horticoles frais. Can. Biosyste. Eng. 44 (3): 7-10

Vigneault, C and L.R. de Castro. 2004. Indirect airflow distribution measurement for horticultural crop package, Part I: Development of the research tool. Trans. ASAE. (Submitted)

Vigneault, C., B. B. Panneton, and G.S.V. Raghavan. 1992. Image analysis of 3-D clouds of bubbles. Can. Agric. Eng. 34 (4): 347-352.

Vigneault, C., N.R. Markarian, A. da Silva, B. Goyette. 2004. Pressure drop during forced-air cooling of various horticultural produce. Trans. ASAE. 47(3):807-814.

Vigneault, C., M.R. Bordint and R.F. Abrahão. 2002. Embalagem para frutas e hortaliças. In: L.A.B. Cortez, S.L. Honório and C.L. Moretti (Eds.). Resfriamento de frutas e hortaliças. Embrapa Informaçã Tecnológica, Brasília, DF, Brésil. 95-140.

Xu, Y and D. Burfoot. 1999. Simulating the bulk storage of foodstuffs. J. Food Eng. 39: 23-29.

Table 1. Instrumented ball positioning through the matrix of 512 balls.

Ball Code X Y Z Ball

Code X Y Z Ball Code X Y Z Ball

Code X Y Z

1 1 1 1 17 1 1 3 33 1 1 5 49 1 1 7 2 2 2 1 18 2 2 34 2 2 5 50 2 2 7 3 3 3 1 19 3 3 3 35 3 3 5 51 3 3 7 4 4 4 1 20 4 4 3 36 4 4 5 52 4 4 7 5 7 5 1 21 7 5 3 37 7 5 5 53 7 5 7 6 8 6 1 22 8 6 3 38 8 6 5 54 8 6 7 7 5 7 1 23 5 7 3 39 5 7 5 55 5 7 7 8 6 8 1 24 6 8 3 40 6 8 5 56 6 8 7 9 2 1 2 25 2 1 4 41 2 1 6 57 2 1 8

10 1 2 2 26 1 2 4 42 1 2 6 58 1 2 8 11 4 3 2 27 4 3 4 43 4 3 6 59 4 3 8 12 3 4 2 28 3 4 4 44 3 4 6 60 3 4 8 13 8 5 2 29 8 5 4 45 8 5 6 61 8 5 8 14 7 6 2 30 7 6 4 46 7 6 6 62 7 6 8 15 6 7 2 31 6 7 4 47 6 7 6 63 6 7 8 16 5 8 2 32 5 8 4 48 5 8 6 64 5 8 8

55

Table 2. Minimum difference between two HCT results to be considered as significantly different at a level of confidence of 95% (Alpha = 0.05) for the different airflow rate.

Airflow rate (L.s-1.kg-1)

S2 Md1 (min)

Md2 (min)

Md3 (min)

Md4 (min)

0,125 3,527 3,72 2,33 1,78 1,48 0,25 1,859 2,71 1,70 1,30 1,08 0,50 0,979 1,98 1,24 0,95 0,79 1,00 0,516 1,45 0,91 0,69 0,58 2,00 0,272 1,06 0,66 0,51 0,42 3,90 0,147 0,78 0,49 0,37 0,31

Mdx = minimum difference when using x replicates (x = 1, 2, 3 or 4) Table 3. Tukey-test showing the effect of the measured airflow rate on the air velocity results obtained from

each method applied on the results of 0.67, 2, and 6% as opening areas. Measured velocity

(m.s -1) Airflow rate (L.s-1.kg-1)

Number of samplea

SCSV velocity (m.s -1)

CAV velocity (m.s -1)

0.029 0.125 72 0.022a 0.029 a 0.058 0.25 72 0.055b 0.036 a 0.116 0.50 72 0.114c 0.048 ab 0.174 0.75 24 0.171d 0.060 bc 0.232 1.00 48 0.244e 0.076 c 0.465 2.00 48 0.507f 0.151 d 0.906 3.90 24 1.019g 0.418 e

Means followed by the same letter in the same column are not significantly different at α = 0.05. Table 4. Examples of goodness of fit coefficient and empirical parameters relating the half-cooling time

(HCT) to the air approach velocity calculated according to the SCSV method of eight of the balls used to form the matrix.

Ball Code a b R2

0 8.950 0.3590 0.9672 8 9.161 0.5027 0.9808

16 10.121 0.5413 0.9774 24 10.083 0.6154 0.9866 32 11.903 0.5951 0.9792 40 11.287 0.6418 0.9786 48 13.374 0.5976 0.9734 56 11.439 0.6403 0.9736

56

Table 5. One-way ANOVA level of significance for the difference between the mass-balance calculated using the CAV and SCSV method for each z-direction layer and airflow rate.

Method Airflow rate (L.s-

1.kg-1) Degree of freedom

F P

0.125 7, 16 35.35.080 < 0.001 0.25 7, 16 27.341 < 0.001 0.5 7, 16 12.495 < 0.001 1 7, 16 3.689 0.015 2 7, 16 1.069 0.426

SCSV

3.9 7, 16 1.767 0.164 0.125 7, 16 9409.762 < 0.001 0.25 7, 16 1574.059 < 0.001 0.5 7, 16 1512.670 < 0.001 1 7, 16 548.864 < 0.001

CAV

2 7, 16 127.795 < 0.001 3.9 7, 16 17.156 < 0.001

Figure 1. Cubic matrix of balls.

57

pitot tube

outlet tube

fan

air flow

air flow

mobile plastic tunnel

mobile fan set up

mobile heat exchanger

aluminum plate

ball matrix

static pressuremeasuring device plywood box

air flow

plastic wrappolystyrene foam

Figure 2. Experimental set up showing forced air tunnel, balls matrix, fan, and dynamic and static

pressures measuring devices.

419.1

104.

8

104.8

38.7

Ø

104.8

419.

1

Figure 3. Plastic plate with nine holes of 0.67% as area (38.6mm of diameter) uniformly distributed.

58

Figure 4. HCT responses for balls 1, 21, 39, and 56, from layers z=1, 3, 5, and 7, respectively

Sdt = 0.5259 Flowrate-0.4715

R2 = 0.3935

0

2

4

6

8

10

12

14

16

0 1 2 3 4

Flow rate (L•s-1•kg-1)

Sta

ndar

d de

viat

Experimental resultsUpper limit for outliersPower (Experimental results)

Figure 5. Effect of airflow rate on the mean standard deviation of the HCT responses including the four

outliers.

HCT = 8.9504 v R2 = 0.9672

HCT = 9.1245 v -0.7098 R2 = 0.9921

HCT = 8.9431 v -0.6876 R2 = 0.9968

HCT = 10.285 v -0.7578 R2 = 0.9907

0

10

20

30

40

50

60

70

80

90

100

110

120

130

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Velocity (m.s-1)

HC

T (m

in)

1

21

39

56

Ball #

-0.359

59

HCT = 25.416 Flowrate-0.6405

R2 = 0.8841

0

20

40

60

80

100

120

140

160

180

200

0 1 2 3 4

Flow rate (L.s-1.kg-1)

Experimental results

Experimental result tendency

HC

T (m

in)

Figure 6. Mean-replication effect of airflow rate on the HCT responses after rejecting the four outliers.

0

20

40

60

80

100

120

140

0 2 4 6 8

z-direction

Air-

mas

s ra

tio (%

)

SCSV : 0.125 CAV : 0.125 SCSV : 3.9 CAV : 3.9

Figure 7. Percentage of the total mass of air measured by the two indirect air velocity measuring methods

(CAV and SCSV) for each z-direction layer for the minimal and maximal airflow tested (L.s-1.kg-1).

60

70

80

90

100

110

120

0 1 2 3 4

Flow rate (L•s -1•kg -1)

Air-

mas

s ra

tio (

%)

1 23 45 67 8

Figure 8. Percentage of the total mass of air measured by the SCSV indirect air velocity measuring method

for each z-direction layers as a function of the different airflow rates.

HTC = 25.416 Flow rate-0.6406

R2 = 0.9971

0

20

40

60

80

100

120

0 1 2 3 4Flow rate (L•s-1•kg-1)

Experimental results

Experimental result tendency

HC

T (m

in)

Figure 9. Effect of airflow rate on the HCT responses based on the result averages.

61

Artigo 4. Effect of container opening on air distribution during precooling of

horticultural produce

Larissa R. DE CASTRO1,2, Clément VIGNEAULT1,2,3, Luis A. B. CORTEZ2

1Horticultural Research and Development Centre

Agriculture and Agri-Food Canada 430 Gouin, Saint-Jean-sur-Richelieu (Qc)

CANADA J3B 3E6

2College of Agricultural Engineering State University of Campinas

Cidade Universitaria Zeferino Vaz 13083-970, Campinas SP, BRAZIL

3 Corresponding author : Tel : 450-346-4494 ext. 170 Fax: 450-346-7740, Email : vigneaultc@agr.gc.ca ABSTRACT An existing research tool was used to investigate air distribution in horticultural produce containers during forced-air precooling. This tool consisted of instrumented produce simulator allowing indirect measurement of surrounding air velocity at different positions inside a package. Using this new research tool for different forced-air cooling conditions, the surrounding air velocity was inferred as a function of the simulators location in reference to the air entrance. The air pathway during cooling process was investigated for three total package-opening areas (0.67%, 2%, 6%) at airflow rates ranging from 0.125 to 3.9 L·s-1·kg-1. The air approach velocity at each position inside a package rose as the opening area increased. More vented area also enhanced the cooling efficiency. However, increasing airflow rate resulted in more air pressure drop. Opening areas more than 6% of the package surface should be tested to achieve the maximum necessary vented configuration that meets the structure constraints and generates air distribution and cooling efficiency not significantly different from fully open. KEY WORDS Produce simulator, package, forced-air, cooling rate, cooling uniformity, air approach velocity. INTRODUCTION The main goal of the worldwide research consists of reducing the high losses currently observed during postharvest operations of fruits and vegetables. The maintenance of horticultural produce quality can be achieved by applying advanced technology. This technology involves from modern procedures for harvest to rapid cooling and refrigerated storage throughout the produce distribution to the market. In this scenario, palletization is essential to increase efficiency during handling, storing, and shipping processes (Neves Filho 2002). The pallet surface optimization can be attained through the standardization of the containers outside dimensions.

There are currently several international and national standards indicating the size and the vents alignment for reusable containers, as well as regulations to insure produce integrity and environment safety. Although many softwares are available to calculate the impact of the container vented area and positioning in its structural resistance, there are rare specifications concerning their influence in the produce cooling efficiency (Kader 2002; Vigneault et al. 2002).

The design of a reusable container should take into account its multiple uses. Therefore, its general dimensions must be suitable to the different postharvest procedures required for a variety of horticultural produce (Vigneault and Émond 1998). For instance, the open area must be large enough not to restrict the airflow during forced-air cooling (Arifin

and Chau 1987; Vigneault and Goyette 2002) but sufficiently narrow to minimize loss of ice particles in a liquid-ice process (Vigneault and Goyette 2001). Moreover, the openings must be well distributed on the package walls (Castro et al 2004) and bottom surface (Vigneault et al. 2004a) not to compromise the uniformity of the cooling operation, whatever the method chosen.

Combining a certain percentage of container opening area and positioning with airflow rate enhances the forced-air cooling efficiency up to a certain point (Castro et al 2004). However, the reduction obtained in cooling time could not justify an important increment in energy costs nor the risk of compromising package structural resistance. Furthermore, a considerable decrease in produce shelf-life deriving from slow and heterogeneous cooling process, may hazard the cooling postharvest treatment.

Fruits and vegetables as live organisms exhibit physiological changes after harvest. Produce from the same variety can present considerable differences in physical and chemical properties (ASHRAE 2002, Leyte and Forney, 1999). These factors added to the variability of produce positioning due to the packing procedure can affect cooling efficiency. Uniform produce and stacking condition are necessary to generate replicable results to allow comparison between opening configurations or airflow rates. These replicated data could be attained by representing horticultural produce with stable simulators. This tool

62

allows evaluating more accurately the airflow pattern through packages during forced-air cooling process, aiding the container and equipments design (Vigneault et al. 2004b).

The aim of this research was to evaluate the effect of different combinations between airflow rates and opening areas in the forced-air cooling of produce simulators. They were evaluated through indirect determination of air velocity profile at different locations inside a container. A new methodology (Vigneault and Castro 2004) was applied producing a reference source for the container design and the precooling operational settings to increase process efficiency.

MATERIAL AND METHODS Produce simulator Uniform plastic balls used to simulate spherical horticultural produce consisted of solid polymer balls made with 52.36 mm of mean outer diameter and 125.55 g of mean weight. Sixty-four balls were selected for their uniformity in terms of thermal properties: average cooling index of -0.1414±0.0081 min-1 and heat capacity of 1.1252±0.0657 kJ·kg-1·oC-1. Each of the 64 balls were instrumented with a 254µm-diameter, 5 m-long insulated copper constantan (Type T) thermocouple wire placed in their center with a precision of ±0.025 mm (Vigneault and Castro, 2004).

Experimental set-up The instrumented balls were stacked along with other 448 balls on a columnar pattern and uniformly distributed to form a cubic matrix of 8-ball-side dimension. Each instrumented ball had a defined relative orthogonal position (X=width, Y=high, and Z=depth of a container) in reference to the airflow direction (Z). The balls arrangement resulted in 48.06% of porosity and is described in detail by (Vigneault et al. 2004b). Figure 1 shows the experimental set up used during the trials. The calibration, the control system and the precision of the setup are also presented in detail by Vigneault et al. 2004b. The setup consists of a heat exchanger to stabilize the temperature of the air entering into the system, the tunnel containing the matrix of the balls, an aspiration chamber, and a control system. The whole experimental set-up was placed in a cold chamber maintained at 4oC to generate the precooling process of the ball matrix.

After positioning the experimental set-up in the cold chamber, the user was prompted by the control system to indicate the opening configuration and the set point air velocity in a user-friendly window. Then, the automated control system proceeded with the air velocity adjustments and all data acquisition until the experiment was completed. The temperature inside the sixty-four balls along with air temperature before and

after crossing the ball matrix, the temperature in the centre of the cold chamber, the pressure drop through the ball matrix and plates, and the air dynamic pressure through the airflow measuring device, were simultaneously recorded at a 20-s-interval. The set point velocity was maintained due to the continuous monitoring of the system error by the software and if necessary the readjustment of the fan voltage to obtain the desired air velocity.

Container Opening Configurations Three opening configurations for containers were randomly chosen considering the packages currently used for horticulture produce. These configurations were investigated by placing a pair of plates next to the first and eighth layers of balls to enclose the matrix and simulate the two sides of a package. One pair of square polypropylene plates of 420 mm and 3 mm in thickness was drilled with circular metal saws. Nine holes of 0.67% as area (38.6 mm of diameter) were uniformly distributed on the surface of the plates. The three opening areas studied in the research were 0.67, 2, and 6%, which were obtained by leaving 1, 3, and 9 holes, respectively, uncovered with airtight duct sealing tapes. A fully open configuration was also tested for comparison by not using any plastic slab.

Experimental procedure The opening configurations were tested with the six airflow rates 0.125, 0.25, 0.5, 1, 2, and 3.9 L·s-1·kg-1 in a complete block design and repeated three times. However, the pressure drop obtained with the greatest airflow rates and less holes configurations over passed the monitoring system and these tests could not be performed. Therefore, the highest airflow rates tested for the plates with one and three holes were 0.75 L·s-1·kg-1 and 2 L·s-1·kg-1, respectively.

Prior to the start of each test, the forced-air tunnel containing the balls was placed in a warm chamber maintained at approximately 27oC. An axial fan circulated the air through the matrix of the balls for about 120 minutes which resulted in a uniform temperature at the center of the 64 instrumented balls. After this conditioning period, the perforated plates were installed and the tunnel was placed in the cold room. The tunnel-end air inlet and outlet were connected to the complete setup (Fig. 1), and the fan was turned on immediately. The data were recorded until the temperature of the warmest ball had reached 6.9oC, at which point the software terminated the control process and turned off all devices. The temperature-time data recorded was used to calculate the half-cooling time (HCT) of each ball for all treatments by using a dedicated ExcelTM macro developed by Goyette et al. (1996).

63

Statistical analysis The air approach velocity (AAV) in the surroundings of 64 different positions inside the balls matrix was deduced from the half-cooling time (HCT) obtained for the produce simulators with the methodology developed by Vigneault et al. (2004b). The coefficient of uniformity of the velocity distribution in the matrix (CU) during the experiments was determined as the quotient between the standard deviation and the mean velocity for the balls. The cooling process efficiency was also evaluated through air pressure drop analysis. Multivariate Analysis of Variance was performed on the results obtained for the 64 balls to verify the effect of the opening configurations and the airflow rates in the cooling time and air approach velocity. The influence of the X, Y, and Z directions was also tested to verify the hypothesis of cooling symmetry in these directions.

Average values for uniformity of air distribution and pressure drop were compared through multivariate ANOVA. The outcome was deeper examined through follow up analysis of variance and Tukey test to determine the effect of the opening area on air uniformity at each level of airflow rate and vice-versa, at a 5% significance level. All the statistical analyses were executed on SPSS v. 11.5 (SPSS Inc. 2004).

RESULTS AND DISCUSSION Cooling rate Opening area and airflow rate Increasing the opening area lowered the half cooling time (HCT) and raised the mean air approach velocity in the balls’ surroundings. These results corresponded to the findings presented by Castro et al. (2004). However, the mean air velocity obtained when using 6% of package area was significantly higher than for 100%. This likely has occurred because of the effect of openings in converging airflow and generating more intense local velocities. The latter increases the average velocity value which was not necessarily uniform though the porous medium. These results were in agreement with Alvarez and Flick’s (1999a) assertions regarding the higher velocity and lower turbulence generated on the container vented area compared to the non perforated zone. Doubling the value of airflow rate approximately doubled the mean air velocity found. Arifin and Chau (1987) also noticed that cooling rate doubled when increasing airflows not more than 2 L·s-1·kg-1.

Positions X, Y and Z Table 1 shows that the farther the balls were from the air entrance the longer they took to cool down. Alvarez and Flick’s (1999b) also claimed that the balls from the first layers in Z or depth direction cooled first. On the

other hand, Castro et al. (2004), who used empty balls to overcome thermal mass effect, noticed that the simulators from the center cooled more slowly than those from the last layer. Analyzing the air approach velocity data instead of HCT on Table 1, allows to identify lower values not only in the deepest layers closer to the air outlet (Z=7, 8) but also in those closer to the air inlet (Z=1, 2) while finding the highest velocities in the middle layer of the balls matrix (Z=5). Therefore, the inference of air pathway from its velocity results obtained by applying Vigneault et al. 2004b’s method allowed disregarding the air heating inaccuracy often found through the produce layers. The proximity of the latter to the package walls may have increased turbulence while reducing the local velocity of air that was “blocked” by the non perforated zone aforementioned.

Regarding X and Y positions, the highest air velocity values were found in the middle matrix positions (X=Y=4 and X=Y=3). This can be explained by the fact that the end package configurations tested had most holes closer to the center of the package, creating a preferential pathway and therefore, increasing the air velocity at this location. The superiority of values found for the bottom of the matrix (Y=3, 4) in comparison to the top (Y=5, 6) in the openings region could be a result of the influence of natural convection, which increased at the lowest levels of airflow rate, as already mentioned by Vigneault and Castro (2004) and Vigneault et al. (2004b). Therefore, the cold air located in the bottom part induced a faster cooling process since its temperature, and consequently viscosity, was lower, creating a higher temperature gradient with the balls. On the other hand, Alvarez and Flick (1999b) and Castro et al. (2004) did not notice the effect of the ball height-direction (Y) in the heat transfer or produce cooling rate, probably because of insufficient number of layers tested in that direction which was two and four layers respectively.

The lowest velocities were observed on X=Y=1, which is the left corner on the bottom taking the air direction when crossing the balls matrix as the reference. At this position the air pathway had a mean velocity approximately 1.7 times lower than at venting passage. The reduction in the velocity results found when the air flowed to the package walls are comparable to the 3 to 4 time-difference between corner and perforated zones obtained by Alvarez and Flick 1999a on the width or X direction. If, instead of air velocity, the analysis of air distribution inside the packed produce had been based on HCT results, these would point out to the right corner at the top (X=8, Y=6, 8) as the slowest cooling process. This result may have been generated due to the natural convection effect previously explained, which lickely decreased

64

the cooling rate on the top portion of the matrix. Using the research tool developed by Vigneault et al. 2004b, this source of error is considerably reduced and the air pathway inside the porous medium can be predicted and therefore, assists the package design.

Air velocity heterogeneity and pressure drop Increasing the opening area enhanced the uniformity of air distribution inside the package and decreased the air pressure drop (Table 2). The results of pressure drop were in agreement with Vigneault and Goyette (2002) who claimed a significant influence of opening areas less than 25% on the air restriction. The minimal opening area produced a cooling process in average 2.5 more heterogeneous than 6% of area which were 0.796 and 0.324, respectively, and a pressure drop 4.2 times higher (543 Pa compared to 129 Pa). The most heterogeneous air distribution and highest pressure drop were identified for 0.75 L·s-1·kg-1, probably because this airflow rate level was tested only with the minimal opening area. The minimal airflow (0.125 L·s-1·kg-1) produced a process more uniform than for 0.5 L·s-1·kg-1, likely due to a natural convection effect. This caused an inversion of the tendency of heterogeneity results in the 0.125 - 0.75 L·s-1·kg-1 range, which was expected to decrease with the airflow rate increase as observed from 1 to 3.9 L·s-1·kg-1.

The same pattern was observed for air pressure drop, which reached the highest value at 0.75 L·s-1·kg-1. Again, the unexpected superior values of air resistance for 0.5 L·s-1·kg-1 in comparison to 1 L·s-1·kg-1 and for 2 L·s-1·kg-1 compared to 3.9 L·s-1·kg-1 were likely generated by the container openings configurations tested at these levels of airflow rate. Therefore, follow up analysis were performed to identify the influence of airflow rate on air distribution uniformity and rapidity, and pressure drop at each level of opening area.

Cooling efficiency Opening area Less container vented area generated more heterogeneous air pathway or steeper velocity gradients due likely to higher turbulence in the container (Table 2). However, the cooling uniformity results between the two least opened configurations (0.67 and 2%) were not statistically different at 0.125 and 0.5 L·s-1·kg-1 (P>0.183 and P>0.669, respectively). For airflow of 0.5 L·s-1·kg-1, there was no significant difference between the mean air velocities produced with the three package areas (0.67, 2 and 6%, P>0.610). From this airflow rate on, no statistical differences were found for any of the areas studied, including fully open (F2, 6 = 1.029, P>0.413). Therefore, although enlarging the total opening area enhances air distribution uniformity (Table 3), it seemed to increase the mean air velocity only at low

airflow ranges (0.125 and 0.25 L·s-1·kg-1). These results agreed to those found by Castro et al. (2004) and Arifin and Chau (1987) who also noticed that the opening area has a major effect in cooling rate at reduced airflow rates, although the authors have still found a significance difference at 1 and 2 L·s-1·kg-1.

Airflow rate For 0.67% of opening area, the highest values of air velocity and pressure drop were obtained with the airflow of 0.75 L·s-1·kg-1. However, for this container opening configuration an increment in airflow did not represent a great improvement in the uniformity of the cooling process (Table 3). The maximum airflow tested with the 2% total opening area generated the highest mean air velocities and the better uniformity, but also the highest pressure drop. The cooling efficiency obtained with 6% opening submitted to 0.125 L·s-1·kg-1 was not significantly different than at 0.25 L·s-1·kg-1 (P>0.064). Furthermore these levels of airflow produced a process as uniform as cooling at the maximum airflow rate (Table 3). For all opening areas, especially 2%, the air distribution heterogeneity at 0.5 L·s-1·kg-1 was higher than expected. This may have occurred due to the inferior accuracy of the research tool applied to determine the airflow distribution in its transient phase (Vigneault et al. 2004b). Since this airflow rate level generates a Reynolds Number of 2081, the undefined dominance of inertial over viscous forces when air passes from laminar to turbulent phase results in higher fluctuation or poorer precision of air velocity data. The sum up of the more opening area was the higher as the cooling uniformity increased (Table 2). On the other hand, the air pressure drop generated through the package increased inversely to the total opening area, but proportionally to the airflow rate.

When the limiting factor is material structure, resistance forces a minimization of package opening area, and therefore 0.75 L·s-1·kg-1 of airflow rate would be sufficient to produce a fast and uniform cooling process. Nevertheless, if there is a limitation of fan power, related to the pressure drop, which reduces the availability of airflow for the process, the opening area should enlarge to more than 6% to optimize cooling efficiency. By using this container opening area the mean air velocity may be compatible to the fully open configuration but the restriction for air distribution through packed produce is still critical. In this case, improvement of cooling performance could be achieved by increasing the number or area of holes and/or by changing their locations on the container surface. Therefore, future research is necessary to optimize package opening design for better air distribution, such as testing configurations that could

65

benefit of natural convection effect to enhance produce cooling rate.

CONCLUSION The application of the research tool to infer the surrounding air velocity in different locations inside the container allowed determining accurately the air distribution through packed produce during precooling. The lowest air velocities were observed on the left corner at the bottom of the container in reference to the air inlet and the highest values were nearer the center or the package holes. Although the different container opening configurations had created variable air pathway patterns, they did not have significant influence in the mean air velocity for airflows equal or higher than 0.5 L·s-1·kg-1.

Enlarging the opening area resulted in a decrease in the air distribution heterogeneity and pressure drop through the porous medium. Therefore, among the three opening areas studied here, 6% could be recommended to optimize cooling efficiency in case of a container structural restriction. Despite of the high air approach velocities obtained with 6%, larger opening areas should be tested to improve the results for air pathway restrain and heterogeneity through produce.

ACKNOWLEDGEMENT This project was accomplished with the financial support from Fundação de Amparo à Pesquisa do Estado de Sao Paulo (FAPESP), and the Horticultural Research and Development Centre of Agriculture and Agri-Food Canada.

REFERENCES Alvarez, G. and Flick, D. 1999a. Analysis of

heterogeneous cooling of agricultural products inside bins. Part I: aerodynamic study. J. Food Eng. 39: 227-237.

Alvarez, G. and Flick, D. 1999b. Analysis of heterogeneous cooling of agricultural products inside bins. Part II: thermal study. J. Food Eng. 39: 239-245.

Arifin, B. B. and Chau, K. V. 1987. Forced-air cooling of strawberries. ASAE Paper no. 87-6004. St. Joseph, Mich.: ASAE

ASHRAE. 2002. Methods of precooling fruits, vegetables and cut flowers. In: Refrigeration Handbook. Am. Soc. Heating, Refrigerating and Air-Conditioning Eng. 14.1-14.10.

Castro, L. R., Vigneault, C., Cortez, L. A. B. 2004. Container opening design for horticultural produce cooling efficiency. Int. J. Food Agr. Environ. 2 (1): 135-140.

Goyette, B., C.Vigneault, B. Panneton, and G. S. V. Raghavan. 1996. Method to evaluate the average temperature at the surface of a horticultural crop. Can. Agric. Eng. 38(4): 291-295.

Kader A.A. (ed) 2002. Postharvest technology of horticultural crops. 3rd edition. Coop. Ext. Uni. of Ca. Div. Agric and Nat. Res. Univ. of CA, Davis, CA. Publ. no. 3311. 535p.

Leyte, J.C. and Forney, C.F. 1999. Optimizing flat design for forced-air cooling of blueberries packaged in plastic clamshells. HortTechnology. 9(2):202-205.

Neves Filho, L. C. 2002. Armazenamento e Distribuiçao figorificados. In Cortez, L. A. B., Honorio, S. L. and Moretti, C. L. (eds). Resfriamento de Frutas e Hortaliças. Embrapa Informação Tecnológica, Brasilia. 165-189.

SPSS Inc. 2004. Chicago, Illinois. USA. http://www.spss.com

Vigneault, C and Castro, L.R. 2004. Indirect airflow distribution measurement for horticultural crop package, Part I: Development of the research tool. Trans. ASAE. (Submitted)

Vigneault, C. and Émond, J.P. 1998. Reusable container for the preservation of fresh fruits and vegetables. United States Patent No. 5,727,711.

Vigneault, C. and Goyette, B. 2001. Loss of ice through container openings during liquid-ice cooling of horticultural crops. Can. Agric. Eng. 43, 3.45-3.48.

Vigneault, C. and Goyette, B. 2002. Design of plastic container opening to optmize forced-air precooling of fruits and vegetables. Appl. Eng. Agr. 18(1): 73-76.

Vigneault, C., Goyette, B., Markarian, N.R., Hui, C.K. P., Côté, S., Charles, M.T. and Émond, J.P. 2004a. Plastic container opening area for optimum hydrocooling. Can. Biosys. Eng. (In press).

Vigneault, C., Bordin, M. R. and Abrahão, R. F. 2002. Embalagem para frutas e hortaliças. In Cortez, L. A. B., Honorio, S. L. and Moretti, C. L. (eds). Resfriamento de Frutas e Hortaliças. Embrapa Informação Tecnológica, Brasilia. 95-119.

Vigneault, C., Castro, L.R., Goyette, B., Markarian, N.R., Charles, M.T., Bourgeois, G. and Cortez, L.A.B. 2004b. Indirect airflow distribution measurement for horticultural crop package; Part II: Verification of the research tool applicability. Trans. ASAE. (Submitted).

66

Table 1. Tukey results of the air velocity and half cooling time for the opening area, airflow rate and balls positions (X, Y and Z) obtained on SPSS.

Factor** Levels N of samples

Air velocity (m.s-1)

HCT (min)

0.67 768 0.089a 76.19d

2 960 0.187b 58.31c

6 1152 0.330d 44.22b

Opening area (%)

100 1152 0.309c 41.50a

0.125 768 0.024a 118.85g

0.25 768 0.057b 68.54f

0.5 768 0.111c 44.87e

0.75 192 0.174d 36.19d

1 576 0.240e 25.97c

2 576 0.496f 15.86b

Airflow rate (L.s-1.kg-1)

3.9 384 0.987g 9.93a

1 504 0.191a 52.99cd 2 504 0.216b 51.04c

3 504 0.295d 41.43b

4 504 0.342e 37.49a

5 504 0.244c 54.26d

6 504 0.218b 60.78e

7 504 0.235c 60.62e

X position

8 504 0.212b 64.48f

1 504 0.198a 53.35d

2 504 0.209b 50.67c

3 504 0.305e 40.94b

4 504 0.331f 37.98a

5 504 0.230c 60.80e

6 504 0.217b 64.29f

7 504 0.248d 55.51d

Y position

8 504 0.214b 59.52e

1 504 0.238ab 33.07a

2 504 0.231a 40.60b 3 504 0.247cd 46.60c

4 504 0.252de 51.27d

5 504 0.256e 56.04e 6 504 0.253de 60.75f

7 504 0.242bc 66.09g

Z position

8 504 0.233ab 68.65h

** Significant at the 0.05 level (P<0.05)

Table 2. Tukey results of air distribution heterogeneity and pressure drop for total opening area and airflow rate.

Factor** Levels Air velocity heterogeneity

Pressure drop (Pa)

0.67 0.796d 543.3d 2 0.624c 354.0c

6 0.324b 129.4b

Opening area (%)

100 0.099a 4.9a 0.125 0.406cd 7.8a 0.25 0.509de 46.1b 0.5 0.528e 176.5d

0.75 0.824f 1358.2g 1 0.348bc 123.6c 2 0.269ab 497.2f

Airflow rate (L.s-1.kg-1)

3.9 0.197a 306.0e ** Significant at the 0.05 level (P<0.05)

Table 3. Effect of opening area on air distribution heterogeneity for different airflow rates. Airflow rate (L.s-1.kg-1)

Opening area(%) 0.125 0.25 0.5 0.75 1 2 3.9 0.67 0.543c*

A** 0.952d

B 0.866b

AB 0.824AB

2 0.632c

B 0.667c

BC 0.772b

C 0.604cB

0.445cA

6 0.301b

AB 0.317b

AB 0.354a

CD 0.360b

D 0.325b

BC 0.286b

A 100 0.147a

D 0.100a

BC 0.123a

CD 0.081a

B 0.038a

A 0.107a

BC * In the same column, numbers with the same lowercase superscript letter are not significantly different at an

alpha level of 5%** In the same row, numbers with the same uppercase subscript letter are not significantly different at an alpha level of 5%

67

pitot tube

outlet tube

fan

air flow

air flow

mobile plastic tunnel

mobile fan set up

mobile heat exchanger

aluminum plate

ball matrix

static pressuremeasuring device plywood box

air flow

plastic wrappolystyrene foam

Figure 1. Experimental set up showing forced air tunnel, balls matrix, fan, and dynamic and static pressures measuring devices.

68

Artigo 5. Cooling performance of horticultural produce in containers with

peripheral openings

Larissa R. DE CASTRO1,2, Clément VIGNEAULT1,2,*, Luis A. B. CORTEZ2

1 Horticultural Research and Development Centre

Agriculture and Agri-Food Canada Saint-Jean-sur-Richelieu, Québec, Canada, J3B

3E6

2College of Agricultural Engineering State University of Campinas

Cidade Universitaria Zeferino Vaz 13083-970, Campinas SP, BRAZIL

* Corresponding author, Tel : 450-346-4494 ext. 170, Fax: 450-346-7740, Email : vigneaultc@agr.gc.ca ABSTRACT Air pathways were investigated for peripheral and central opening configurations of package during horticultural produce forced-air cooling process using a research tool previously developed. Total opening areas of 0.67, 2, 4, and 8%, formed by combining three, four or eight holes of 0.67%, 1% or 2% distributed at the bottom and top, corners or center line of the package surface, were tested for airflow rates ranging from 0.125 to 3.9 L•s-1•kg-1 and compared to the fully open configuration. Air pressure drop, rates and uniformities of cooling were measured. Enlarging opening area increased the cooling efficiency. The higher the airflow the greater the rate and the uniformity of cooling process. However, gravity force effect influenced the enhancement of air distribution uniformity at the minimum airflow rate studied. When the container design options are limited to central or peripheral openings, the bottom and top opening configuration is preferred for greater cooling performance. KEY WORDS Package, forced-air cooling, air velocity, air distribution.

INTRODUCTION Forced-air cooling is the most common technology used to extend horticultural produce shelf life reducing deterioration and water loss rates, especially for those sensitive to water exposure (Kader, 2002). The efficiency of this process could be increased through optimization of the rate and uniformity of cooling and through energy effectiveness (Cortez et al., 2002; Kader, 2002; Castro et al., 2004a; Castro et al., 2004b).

During precooling of packed horticultural produce, the air pathway is affected by the pressure differentials formed at the container opening and produce bulk porosity, among other several parameters (Vigneault et al., 2004c). Since the air tends to circulate though zones that offer less resistance, a preferential pathway is created in the void space. Concurrently a kind of back-mixing effect often occurs in the corners, known as “dead zones”, decreasing the uniformity of the cooling process (Alvarez and Flick, 1999a).

Therefore, the diameter and positioning of the produce, the stacking arrangement of containers as well as their overall dimensions and total opening area (TOA) have an influence in the intensity and homogeneity of the air velocity profile through the packed produce (Alvarez and Flick, 1999a; Van der Sman, 2002).

Concerning the container, the less the TOA the higher the restriction to the air circulation through the packed produce bulk (Castro et al., 2004b). Opening position also plays a considerable role in the air

pathway, influencing the cooling homogeneity rather than the air resistance or pressure drop (Vigneault and Goyette, 2002; Castro et al., 2004a). Moreover, the opening position on the package surface must be designed so that the holes are aligned to avoid air obstruction when stacking containers side by side (Parsons et al., 1970; Henry et al., 1979).

As the air penetrates through the porous medium, its velocity becomes more uniform (Flick et al., 2003) and the temperature differential between the air and the produce is reduced (Arifin and Chau, 1987; Baird et al., 1988; Xu and Burfoot, 1999). Consequently, the cooling rate is expected to be lower in the last layers than in the first ones (Kopelman et al., 1966; Lindsay et al., 1983, Boyette, 1996) although some authors had noticed the opposite in some specific circumstances (Leyte and Forney, 1999).

A ‘Vi’ number was developed by Vigneault et al. (2004c) to compare the performance of different opening configurations. This dimensionless number is the coefficient of heterogeneity of air velocity distribution through a porous medium. It is defined as the ratio between the standard deviation and the mean of the velocity of the air circulating through a mass of horticultural produce packed in a container. Thus a Vi number equals to 0 means a perfect air distribution uniformity with no variation through the porous medium. Vigneault et al. (2004c) also demonstrated the possibility of using this number to discriminate the effect of different parameters on air distribution uniformity.

69

The present research aimed to investigate the effect of peripheral and central openings of containers on air distribution during forced-air precooling process. The research tool developed in a previous work (Vigneault and Castro, 2004; Vigneault et al., 2004b) was used to identify the significance of peripheral and central openings of container design on the overall cooling efficiency and uniformity of air distribution through packed produce.

MATERIAL AND METHODS Produce simulator The produce simulator described in detail by Vigneault and Castro (2004) consists of solid polymer balls 52.36 mm in diameter and weighing 125.55 g. Sixty-four balls were selected for their relatively high uniformity in terms of cooling rate index (-0.1414 ± 0.0081 min-1) and heat capacity (1.125 ± 0.066 kJ·kg-1·oC-1). Each ball was instrumented with a 30-gage insulated copper-constantan thermocouple wire placed in their center with a precision of ±0.025 mm.

Experimental set-up The 64 instrumented balls were stacked uniformly distributed along with other 448 balls on a columnar pattern to form a cubic matrix of 8-ball-side dimension positioned in a forced air cooling tunnel. The arrangement resulted in 47.64% of porosity. The balls forming the two end layers perpendicularly to the airflow direction were assembled together using 12-mm-long and 6-mm-diameter plastic pins inserted into 6mm depth hole perforated at each ball to ball or ball to wall contact point.

Figure 1 shows the experimental set up used during the trials. Four acrylic plates were assembled to simulate a forced-air cooling tunnel of 420 mm inside square cross-section, and 1250 mm long. The ball matrix was positioned at 220 mm from the tunnel-end air inlet. The portion of the tunnel containing the balls was insulated to reduce heat conduction. The air-outlet of the tunnel consisted of a 610 mm long plenum enabling air pressure drop (APD, mm of water) measurements across the ball matrix.

The end of the air-outlet tunnel was air-tightly attached to the aspiration chamber. A direct drive radial blade fan and a 0.75 kW variable speed motor assembly created a negative pressure in the aspiration chamber and forced the outside air to circulate through the cooling tunnel. The air was released to the atmosphere through a 500 mm long, 101.6 mm-diameter tube instrumented with a Pitot-tube device allowing airflow measurements.

The whole experimental set-up was placed in a cold chamber kept at 4oC to simulate the precooling process of the balls matrix. The end of the air-inlet tunnel was attached to a 520 x 840 x 1100 mm aluminum heat-

exchanger to reduce the temperature variation due to the oscillation of the cold chamber temperature (Vigneault et al., 2004b). A data acquisition and control system, described in detail by Vigneault et al. (2004b) was used to control the air flow at the desired level and record, simultaneously, the temperature inside the balls along with air temperature before and after crossing the matrix of balls, the temperature in the centre of the cold chamber, the pressure drop through the ball matrix and plates, and the airflow rate, at a 20-s-interval.

Container Opening Configurations Eight opening configurations for container were investigated by placing pairs of plates next to the first and eighth layers of balls to enclose the matrix and simulate the two sides of the package perpendicular to the airflow direction. Three pairs of 420 mm square polypropylene plates 3 mm thick were drilled with circular metal saws. Nine holes of 0.67% of the plate total area (38.6 mm Ø) were uniformly distributed on the surface of the plates #A (Fig. 2). The plates #B were drilled with four holes of 0.5% as area (33.4 mm Ø) on each of their upper and lower parts. The #B plate center line was drilled with 4 holes of 1% area each (47.3 mm Ø). The plates #C was prepared as the #B plates, but the hole areas were 1% for those on the upper and lower parts and 2% (66.9 mm Ø) for the central line holes. Eight TOA configurations were obtained by covering with sealing tape either the peripheral or the central holes of these pairs of plates (Table 1). A fully open configuration (100%-TOA) was also tested for comparison by not using any plate.

Experimental procedure The opening configurations were tested with five or six airflow rates among 0.125, 0.25, 0.5, 1, 2, and 3.9 L•s-1•kg-1 when possible (Table 1). The values of airflow used for each configuration were chosen basing on the transmitter upper-limit settled for air pressure drop through horticultural produce, 125 mm of water based on Vigneault et al. (2004d)’s results. The experiments were performed in a complete block design and repeated twice as recommended by Vigneault et al. (2004b).

Prior to the start of each test, the forced-air tunnel containing the balls was placed in a chamber containing a heat system set to operate at approximately 28oC. An axial fan forced the air through the matrix reheating the balls within two hours. After this period, the perforated plates were installed and the tunnel was placed in the cold chamber. The tunnel-end air inlet was connected to the heat-exchanger and the air outlet to the aspiration chamber, and the fan was turned on immediately. The control software recorded the results until the

70

temperature of the warmest ball had reached 7oC and then turned off the system.

Statistical analysis The temperature-time data recorded was used to calculate the half-cooling time (HCT) of each ball for all treatments by using a dedicated ExcelTM macro developed by Goyette et al. (1996). The HCT results were applied to infer the air approach velocity at the 64 different positions inside the ball matrix according to the method developed by Vigneault et al. (2004b) for each tested combination of opening configuration and airflow rate. In turn, the air approach velocities at the 64 balls were used to test the effect of opening positioning at y and z directions on cooling process. This analysis allowed verifying the cooling symmetry through the height and general airflow directions, respectively.

A Multivariate Analysis of Variance was also performed to verify the effect of TOA, opening position configurations, and airflow rates on cooling time, air velocity, and pressure drop. The effectiveness of the different parameters to homogenize the air distribution through the ball matrix was evaluated with the coefficient of heterogeneity, Vi. Since the airflow range was not the same for all opening configurations tested, the statistical analyses were performed only with the results obtained at 0.125 – 2 L•s-1•kg-1. The outcome was further examined through Tukey test to discriminate the effect of opening configurations on cooling efficiency. All the statistical analyses were executed on the SPSS v. 11.5 (SPSS Inc. 2004) at a 0.05 level of significance.

RESULTS AND DISCUSSION Effect of TOA and opening configuration In general, enlarging the TOA enhanced cooling rate and air velocity while decreased APD (Table 2). Although no significant difference had been found for HCT between 8%-TOA and 100%-TOA, the package with the 4 and 8% TOA generally produced mean air velocity higher than fully open configuration. In these cases, the airflow concentrated and directed through smaller openings increased considerably the velocity intensity in the perforated region as claimed by Alvarez and Flick (1999a) and Castro et al. (2004b).

Although the mean air velocity for 100%-TOA was not significantly different from the value obtained with 2%-TOA, the latter produced slower cooling process than fully open configuration. This result was explained by the 7 times higher Vi produced with 2%-TOA than with 100% TOA (Table 2). In fact, the most uniform process (smallest value of Vi) and the least air restriction (lowest APD) were obtained when testing the produce in bulk (fully open).

TOA=2% Four holes on corners with 0.5% individual area generated higher cooling time and air distribution heterogeneity (Vi) than three 0.67%-area central openings. This outcome could suggest that less individual hole area and corners positioning had more influence on the results rather than the number of openings (four against three). Since no significant differences were found for the results of air velocity and pressure drop, it is believed that corners positioning contributed the most for the poorer cooling efficiency for the airflow rate tested. Castro et al. (2004a) and Vigneault and Goyette (2002) also noticed that for a same TOA, the dimensions of the holes had no considerable influence on APD.

TOA=4% For 4%-TOA, increasing the number of holes and placing them on a bottom and top configuration decreased Vi and APD but, extended the cooling time by decreasing the mean air velocity. The perforated region showed a significant effect in intensifying air velocities through the produce, thus converging the open area by reducing number of openings, enhanced these values. Therefore, less holes increased air velocity and APD. The improvement of uniformity would be attributed either to the position (bottom and top) or/and number of openings rather than to the individual hole area, 0.5% compared to 1%.

Considering airflows ranging from 0.125 to 3.9L•s-1•kg-1 for the same 4%-TOA, four holes of 1%-area positioned at the corners showed air distribution heterogeneity (Vi=0.576) significantly higher than the center line hole position with Vi= 0.511. Hence, the influence of hole position must be considered on optimizing package design. Since each positioning (bottom and top, center line or corners) was only tested with the same number of holes (eight, four and four respectively), further investigation would be necessary to identify which one of these variables most influenced the outcome.

TOA=8% The same air distribution pattern outcome at 4%-TOA was observed for 8% but, no HCT significant difference was found between 8-1%-holes on bottom and top and 4-central-2% holes. This result suggested that the reduction in Vi was considerable enough to compensate the lower mean air velocity generated, these findings correspond to the results obtained by Castro et al. (2004a). Thus, the only design variation was in the number of holes in height-direction (Y) and their area, which did not affect significantly the cooling process time, in agreement with the same authors.

71

Airflow rate In most cases, increasing the airflow rate reduced the cooling time and Vi, and increased the mean air velocities and APD (Table 3) in agreement with previous findings (Arifin and Chau 1987; Castro et al. 2004a,b). However, at 0.125 L•s-1•kg-1 the air distribution through the balls matrix was more uniform than at 0.5 L•s-1•kg-1, possibly due to the influence of gravity force effect (Vigneault et al. 2004a). Therefore, among the lowest airflow rates studied, 0.125 L•s-1•kg-1 seemed to be preferred under refrigeration operational limitation since it produced also one of the lowest APD. On the other hand, the HCT resulting from 0.125 and 0.25 L•s-1•kg-1 were 2.6 and 1.5 times higher than the value obtained with 0.5 L•s-1•kg-1, respectively.

Produce positioning on Y and Z directions Considering the mean air velocity for all airflow rates tested (Table 4), the highest value was found at the bottom of the ball matrix (Y=1). This result may first appear contradictory to other findings (Alvarez and Flick 1999a; Castro et al. 2004b) that pointed the locations closest to the package walls as for the lowest air velocities. However, this height corresponded to one of the two opening positions of most package configurations studied here which include bottom and top and corners position. Furthermore, the greater mean air velocity at the bottom (Y=1 and 2) compared to the top layer air velocity (Y=5, 6, 7 and 8) corresponded to the Vigneault et al. (2004a)’s finding when measuring the importance of gravity force effect. Finally, the container top layer (Y=8) did not generate lower velocities than the middle layers (Y=3 to 7) because the influence of opening area prevailed over the expected blockage effect by the package walls proximity.

The HCT was inferior for the balls closer to the bottom openings (Y=1 to 4) than for the top holes (Y=5 to 8) likely due to the gravity force effect as shown by Vigneault et al. (2004a). Among the balls on the top layers, the longest HCT was observed for the middle (Y=5-6) likely because they were not advantaged by the wall/openings proximity.

Regarding Z positions, the farther the balls are from the air inlet, the higher their HCT (Table 4), corresponding to the results in the literature (Arifin and Chau, 1987; Baird et al., 1988; Xu and Burfoot, 1999) showing a thermal mass effect on the air temperature as it crosses a porous medium. Alvarez and Flick (1999b) also noticed that the balls from the first layers (Z=1, 2) had the lowest cooling times, but the highest values were found not on the last but on the middle layers of their matrix (Z=3, 4).

The first layer (Z=1) presented the same air velocity as the middle layers (Z=3 to 7). The inference of air approach velocity from the equations developed by

Vigneault et al. (2004b) minimized the thermal mass error and allowed identifying lower air velocities not only on the deepest level (Z= 8) but also on other layers of the matrix (Z=1, 2, 6, 7).

CONCLUSION Peripheral and central opening positions produced a significant effect on air distribution through the porous medium. The results suggested that the uniformity of air circulation increases as the holes are moved from corners to center and finally to the bottom and top portions of the package surface. For a same total open area, the position did not influence the air approach velocity or pressure drop, except when the number of holes was varied. In this case, the mean air velocity and pressure drop decreased as more holes were added to the package. This last assertion could be applied to understand the lowest values obtained for fully open, if this configuration was seen as formed with an “infinite number” of small holes.

When comparing packages with the same number and position of openings, those with larger holes produced higher air velocity, cooling rate and air velocity uniformity and less air pressure drop. This improvement in cooling performance was rather justified by the increase in total vented area and could be also obtained by increasing airflow rate. The combination among high airflow rate, large vented area and holes covering most of the container surface was revealed as the best for enhancement of cooling rate.

ACKNOWLEDGEMENTS This project was accomplished with the financial support from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, Brazil), and Agriculture and Agri-Food Canada (Canada); and is the result of a close collaboration project between the Horticultural Research and Development Center (Canada) and UNICAMP (Brazil). SYMBOLS AND ABREVIATIONS APD = air pressure drop, mm of water; HCT = half-cooling time, min; TOA = Total opening area, %; Vi = Vigneault number, coefficient of heterogeneity, dimensionless number; X = horizontal direction perpendicular to the airflow direction or width; Y = vertical direction perpendicular to the airflow direction or height; Z = main airflow direction through the matrix of ball. REFERENCES Alvarez, G., Flick, D. 1999a. Analysis of

heterogeneous cooling of agricultural products

72

inside bins. Part I: aerodynamic study. Journal of Food Engineering. 39: 227-237.

Alvarez, G., Flick, D. 1999b. Analysis of heterogeneous cooling of agricultural products inside bins. Part II: thermal study. Journal of Food Engineering. 39: 239-245.

Arifin, B. B., Chau, K. V. 1987. Forced-air cooling of strawberries. ASAE Paper no. 87-6004.

Baird, C. D., Gaffney, J. J., Talbot, M. T. 1988. Design criteria for efficient and cost effective forced-air cooling systems for fruits and vegetables. ASHRAE Transactions. 94: 1434-1453.

Boyette, M. D. 1996. Forced-air cooling packaged blueberries. Applied Engineering in Agriculture. 12(2): 213-217.

Castro, L. R., Vigneault, C., Cortez, L. A. B. 2004a. Container opening design for horticultural produce cooling efficiency. International Journal of Food, Agriculture and Environment. 2 (1): 135-140.

Castro, L. R., Vigneault, C., Cortez, L. A. B. 2004b. Effect of container opening on air distribution during precooling of horticultural produce. Transactions. of the ASAE. (Submitted for publication)

Cortez, L.A.B., Castro L.R., Vigneault, C. 2002. Resfriamento Rápido a ar : métodos da câmara frigorífica e do ar forçado (Refroidissement à l’air forcé). In : Cortez, L.A.B., Honório, S.L., Moretti, C.L. (Eds.) Resfriamento de frutas e hortaliças. Embrapa Informaçã Tecnológica, Brasília, DF, Brésil. pp: 231-272.

Flick, D., Leslous, A., Alvarez, G. 2003. Modélisation semi-empirique des écoulements et des transferts dans un milieu poreux en régime turbulent. Revue International du Froid. 26: 349-359.

Goyette, B., Vigneault, C., Panneton B., Raghavan, G. S. V. 1996. Method to evaluate the average temperature at the surface of a horticultural crop. Canadian Agricultural Engineering. 38(4): 291-295.

Henry, F. E., Bennett, A. H., Segall, R. H. 1979. Hydraircooling vegetables in pallet loads. American Vegetable Grower, 27(10): 8-9, 46-47.

Kader A.A. (ed) 2002. Postharvest technology of horticultural crops. 3rd edition. Cooperative Extension of University of California. Division of Agriculture and Natural Resources. University of California, Davis, CA. Publ. no. 3311.

Kopelman, I., Blaisdell, J. L., Pflug, I. J. 1966. Influence of fruit size and coolant velocity on the

cooling of Jonathan Apples in water and air. ASHRAE Transactions. 72(1): 209-216.

Leyte, J. C., Forney, C. F. 1999. Optmizing flat design for forced-air cooling of blueberries packaged in plastic clamshells. HortTechnology. 9(2):202-205.

Lindsay, R. T., Neale, M. A., Messer, H. J. M. 1983. Ventilation rates for the positive ventilation of vegetables in bulk bins. Journal of Agricultural Engineering Research. 28: 33-44.

Parsons, R. A., Mitchell, F. G., Mayer, G. 1970. Forced-air cooling of palletized fresh fruit. Transactions of the ASAE, 15: 729-731.

SPSS Inc. 2004. Chicago, Illinois. USA. http://www.spss.com

Van der Sman, R. G. M. 2002. Prediction of airflow through a vented box by the Darcy-Forcheheimer equation. Journal of Food Engineering. 55: 49-57.

Vigneault, C., Goyette, B. 2002. Design of plastic container openings to optimize forced-air precooling of fruits and vegetables. Applied Engineering in Agriculture. 18(1) :73-76

Vigneault, C., L.R.de Castro, L.A.B. Cortez . 2004a. Effect of gravity on forced-air precooling. IASME Transactions on Mechanical Engineering. (Submitted for publication)

Vigneault, C., Castro, L. R., Goyette, B., Markarian, N. R., Charles, M. T., Bourgeois, G., Cortez, L. A. B. 2004b. Indirect airflow distribution measurement for horticultural crop package : Part II: Verification of the research tool applicability. Transactions of the ASAE. (Submitted for publication)

Vigneault, C., Castro, L. R. 2004. Indirect airflow distribution measurement for horticultural crop package, Part I: Development of the research tool. Transactions of the ASAE. (Submitted for publication)

Vigneault C., Castro, L. R., Gautron, G. 2004c. Effet de la présence de poignées ouvertes sur le prérefroidissement de produits horticoles. La Revue Internationale du Froid. (Submitted for publication)

Vigneault, C., Markarian, N.R., da Silva, A., Goyette, B. 2004d. Pressure drop during forced-air circulation of various horticultural produce. Transaction of the ASAE. 47 (3): 807-814.

Xu, Y., Burfoot, D. 1999. Simulating the bulk storage of foodstuffs. Journal of Food Engineering. 39: 23-29.

73

Table 1. Combinations between container opening configurations for airflow rates of 0.125, 0.25 ,0.5, 1, 2, 3.9 L•s-1•kg-1.

Number of holes

Area (%)

Position TOA (%)

3 0.67 Center 2* 4 0.5 Corners 2* 4 1 Center 4 4 1 Corners 4 8 0.5 Bottom and top 4 4 2 Center 8 8 1 Bottom and top 8

*These configurations were not tested for 3.9 L•s-1•kg-1 airflow rate due to the limitation of the system. Table 2. Effect of total opening area and opening configurations on air velocity, HCT, Vi and APD.

TOA (%)

Openings configuration

Air velocity (m•s-1)

HCT (min)

Vi APD (mm water)

3-0.67% center 0.195b 58.04c 0.609f

37.1f 2

4-0.5% corners 0.188ab 65.36d 0.750g

35.4f

8-0.5% bottom and top 0.175a 60.19c 0.487c

8.7d

4-1% center 0.221c 49.11b 0.541d

11.4e 4

4-1% corners 0.217c 48.37b 0.567de

11.7e

8-1% bottom and top 0.219c 45.55a 0.377b

3.0b 8

4-2% center 0.241d 45.61a 0.582ef

6.1c

100 Fully open 0.187ab 47.90ab 0.094a 0.0a

Means in the same column and the same group of number followed by the same letter are not significantly different based on t-Test using α = 0.05.

Table 3. Effect of airflow rates on air velocity, HCT, Vi and APD. Airflow rate (L•s-1•kg-1)

Air velocity (m•s-1)

HCT (min)

Vi APD (mm water)

0.125 0.030a 111.58f 0.505c 0.0a

0.25 0.066b 65.80e 0.519cd 0.9a

0.5 0.128c 42.80d 0.534d 3.3b

1 0.268d 26.07c 0.495c 13.6c

2 0.535e 16.35b 0.452b 53.0d

3.9 1.043f 10.35a 0.409a 90.8e

Table 4. Effect of ball positions on air velocity and HCT.

Y Z Levels Air velocity

(m•s-1) HCT (min)

Air velocity (m•s-1)

HCT (min)

1 0.397d 40.57a 0.304abc 32.90a

2 0.322c 41.47a 0.282ab 38.57b

3 0.282ab 46.03b 0.307bc 42.51c

4 0.296b 46.41b 0.306bc 46.63d

5 0.265a 55.72de 0.308c 50.99e

6 0.266a 56.48e 0.304abc 55.94f

7 0.271a 53.09cd 0.292abc 60.61g 8 0.281ab 51.72c 0.278a 63.35h

Means in the same column and the same group of number followed by the same letter are not significantly different based on t-Test using α = 0.05.

74

LIST OF FIGURES Figure 1. Experimental set up showing forced air tunnel, balls matrix, fan, and dynamic and static pressures

measuring devices. Figure 2. Dimension of the #A, #B and #C plates used to produce the eight opening configurations to

investigate the effect of the opening position and total area on the air distribution uniformity.

pitot tube

outlet tube

fan

air flow

air flow

mobile plastic tunnel

mobile fan set up

mobile heat exchanger

aluminum plate

ball matrix

static pressuremeasuring device plywood box

air flow

plastic wrappolystyrene foam

Figure 1. Castro, Vigneault, Cortez #A #B #C Figure 2. Castro, Vigneault, Cortez

75

Artigo 6. Effect of gravity on forced-air precooling

Clément VIGNEAULT1,2,3, Larissa R. DE CASTRO1,2,4, Luis A. B. CORTEZ2,5

1Horticultural Research and Development Centre Agriculture and Agri-Food Canada

430 Gouin, Saint-Jean-sur-Richelieu (Qc) CANADA J3B 3E6

2College of Agricultural Engineering State University of Campinas

Cidade Universitaria Zeferino Vaz 13083-970, Campinas SP, BRAZIL

3vigneaultc@agr.gc.ca 4larissarcastro@yahoo.com 5cortez@reitoria.unicamp.br

Abstract: - Air pathways were investigated for both top and bottom diagonal opening configurations of package during horticultural produce forced-air precooling. The effect of gravity force on cooling process was evaluated with a research tool previously developed. Six opening configurations were tested for five airflow rates ranging from 0.125 to 2 L•s-1•kg-1 and compared to the fully open configuration. Air pressure drop and rates and uniformities of cooling were measured. Increasing the airflow rate enhanced the cooling efficiency and reduced the cooling time. Pressure drop and mean air velocity through package rose as airflow was increased, but the gravity force also produced a significant effect on the air pathway and cooling efficiency. Its magnitude was determined by the airflow rate applied. The effect of the gravity force should not be neglected when building simulation model of airflow through such porous media mainly at low airflow rates.

Key-Words: - package, cooling, air distribution, pressure drop, opening, coefficient of heterogeneity.

INTRODUCTION Forced-air precooling is the most common technology used to extend horticultural produce shelf life reducing deterioration and water loss rates, especially for those sensitive to water exposure [1]. The efficiency of this process could be increased through optimization of the rate and uniformity of cooling and through energy effectiveness [2][3][4]. Models exist to simulate the air circulation through porous medium but the effect of the gravity is generally neglected.

A research tool was developed in previous studies [5][6]0 to identify the effects of opening on containers on the overall cooling efficiency and uniformity of air distribution through the packed produce.

A ‘Vi’ number was developed [8] to compare the significance of the effect of different opening configurations on the air distribution uniformity. This dimensionless number is the coefficient of heterogeneity of the air velocity distribution through a porous medium and was defined as the ratio between the standard deviation and the mean of the air velocities circulating through a mass of packed produce. A Vi number equals to 0 would mean a perfect air distribution uniformity with no variation through the porous medium.

A cooling process is considered as completed when the warmest produce has reached the desired final temperature which occurs generally when 7/8 of the produce sensible heat is extracted [8]. Since the half cooling time (HCT) is directly related to the 7/8 cooling time, the maximum half cooling time (HCTm)

could be used to compare the efficiency of any cooling process as recommended by Cortez et al. [1].

Several researchers have been exploring different configurations to evaluate whether or not the container design meets the cooling requirements [9] for a variety of horticultural produce. Since the gravity force presented some effect on previous results [2][3][6], a diagonal air circulation system has been suggested to exploit this natural force to increase the uniformity of air circulation through the horticultural produce packed in standard container.

The aim of the present research was to investigate the effect of gravity force on air distribution uniformity during forced-air precooling process using diagonal openings through walls of packed-horticultural-produce container.

MATERIAL AND METHODS Produce simulator Uniform polymer balls with 52.36 mm of mean outer diameter and 125.55 g of mean weight were used to represent spherical horticultural produce. Sixty-four balls were selected for their uniformity in terms of thermal properties regarding cooling index (-0.1414 ±0.0081 min-1) and heat capacity (1.1252±0.0657 kJ•kg-1•oC-1), and instrumented with a 30-gage, 5 m-long Type T thermocouple [5].

Experimental set-up The instrumented balls were stacked along with other 448 balls on a columnar pattern and uniformly distributed to form a cubic matrix of 8-ball-side dimension. The calibration, the control system, and the

76

precision of the setup were also presented in detail by Vigneault et al. [6]. Fig. 1 shows the experimental set up used during the trials. The setup consisted of a heat exchanger to stabilize the temperature of the air entering into the system, the tunnel containing the matrix of the balls, an aspiration chamber, and a control system. The whole experimental set-up was placed in a cold chamber maintained at 4oC to generate the precooling process of the ball matrix.

At the start of each experiment, the user was prompted by the control system to input the opening distribution and the set point air velocity in a user-friendly window. The experiment was entirely automated. The temperature inside the sixty-four balls along with air temperature before and after crossing the ball matrix, the temperature in the centre of the cold chamber, the air pressure drop (APD, mm of water) through the ball matrix and plates, and the air dynamic pressure through the airflow measuring device, were simultaneously recorded at a 20-s-interval. The set point air velocity was maintained through a continuous monitoring system driven by dedicated software which readjusted the fan voltage as necessary.

Container Opening Configurations Six opening configurations for containers were

investigated by placing a pair of plates next to the first and eighth layers of balls to enclose the matrix and simulate the two sides of a package. Three pairs of 420 mm square polypropylene plates 3 mm thick were drilled with circular metal saws. Pairs of plate #A and #B were drilled with one row of four opening areas (OA) of 0.5% (33.4 mm Ø) and 1% (47.3 mm Ø) respectively (Fig. 2). Similar procedure was followed for the pair of plate #C to obtain the same total open area (TOA) as #A (TOA=2%) using one row of three openings of OA=0.67% (38.6 mm of diameter). The six opening configurations tested were obtained by using the three pairs of plate at two positions by switching the hole on either the top or the bottom part of the plate. The diagonal air circulation patterns were obtained by placing the plates with top openings at the inlet air position and the bottom ones at the outlet, or with the reverse position. The configurations formed in these opening were referred based on the opening position of air inlet plates which were top (TO) and bottom (BO) diagonal openings. A fully open (FO) configuration was also tested for comparison by not using any plastic plates.

Experimental procedure The opening configurations were tested with the five airflow rates 0.125, 0.25, 0.5, 1 and 2 L•s-1•kg-1in a complete block design and repeated twice as recommended by Vigneault et al. [6].

Prior to the start of each test, the forced-air tunnel containing the balls was placed in a warm chamber

maintained at approximately 28oC. An axial fan circulated the air through the matrix of the balls for about 120 minutes which resulted in a uniform temperature at the center of the 64 instrumented balls. After this conditioning period, the perforated plates were installed and the tunnel was placed in the cold room. The tunnel-end air inlet and outlet were connected to the complete setup (Fig. 1), and the fan was turned on immediately. The data were recorded until the temperature of the warmest ball had reached 6.9oC, at which point the software terminated the cooling process and turned off all devices.

Statistical analysis The temperature-time data recorded was used to calculate the HCT of each ball for all treatments by using a dedicated ExcelTM macro developed by Goyette et al.[10].

The air approach velocity in the surroundings of 64 different positions inside the ball matrix was deduced from the HCT obtained for the produce simulators with the methodology developed by Vigneault et al. 0. These data were associated with the results of APD, the Vi number [8] and the HCTm to analyze the cooling process efficiency and the air distribution uniformity.

A multivariate analysis of variance was performed on the results to evaluate the effect of the opening configurations on cooling time, uniformity of air velocity distribution, and APD. The outcome was deeper examined through follow up analysis of variance and Tukey test to verify if this effect varied depending on the airflow rate tested. All the statistical analyses were executed on SPSS v. 11.5 [11] at a 5% significance level.

RESULTS AND DISCUSSION The APD increased as airflow rose and the open area was decreased, corresponding to results of previous researches [2][3][12]. However, the APD was not influenced by the position of the openings in reference to the air inlet and outlet at a same TOA level (Table 1). Considering the effect of individual opening area on APD, the result reveals that for TOA=2%, four openings of 0.5% produced less APD than three openings of 0.67%. This outcome may be due to the number of holes rather than to their individual areas since previous research [4] had not identified a significant influence of opening area ≤ 2% on APD. On the other hand, four 0.5%-openings seemed to generate the least uniform air distribution through packed produce, increasing Vi.

Increasing the TOA decreased the HCT and Vi through the packed produce. Comparing these results with fully open, this configuration generally produced the most uniform air pattern but the highest HCT (Tables 1 and 2). However, the opening areas did not have a significant effect on HCT at the maximum

77

airflow rate (2 L•s-1•kg-1). A different effect was showed when comparing the HCTm (Table 3) which determines the duration of the cooling process. Due to the large heterogeneity of the results, produce packed in a BO configuration box would not performe significantly better in terms of cooling rate.

As Castro et al. [2] observed, the effect of opening position was significant on air distribution. On average, the top diagonal openings (TO) produced more heterogeneous air pattern (Vi=0.689) than BO (Vi=0.590). These results are comparable to the Vi~0.5 obtained by Castro et al. [2] for 8-0.5% holes distributed on top and bottom of a same plate (TOA=4%), showing clearly a gravity force effect on air circulation.

The gravity force effect can also be seen from the analysis of the air velocity. At the first layer in airflow direction (Z=1), the air velocity was higher at the position closer to the openings and lower at the blocked walls (results not shown). This tendency continued as air crossed the produce, keeping faster airflow or a preferential pathway on the same height position as air inlet. These findings were confirmed when analyzing the results for air velocity on Y axe (Table 4). At TO configuration, air velocity was more intense on the superior (Y=5-8) quadrant in reference to the air entrance. On the other hand, for BO configuration, the intensity increased on the inferior zone. These results are in agreement with Castro et al. [2][3] and Alvarez and Flick’s [13] who also noticed the effect of perforated areas in converging airflow and increasing local velocities.

In reference to the Z-direction, in general air velocity intensified on the last layers (Z=7, 8 on Table 4) in opposition to Castro et al.’s [2] findings for peripheral and central openings. Furthermore, the farther from the fan the lower the velocity observed for the produce layer. Table 4 also presents that the difference between maximum and minimum velocities at Z-direction for TO was higher than for BO, suggesting a more uniform process for BO.

Moreover, the inlet cold air was heated up as it crossed the warmer produce layers. This temperature rise drove less dense air to the upper zones of the package. When the outlet openings were located on the top, forced and natural convection worked in the same direction. They forced air to flow from bottom to top hastening its velocity distribution. However, when the cold air entered through the top openings, as its temperature raised it accelerated at the superior zones of the package. Since the fan did not supply air velocity high enough to force air through the BO, the produce inferior portion was exposed to lower air velocities. Table 1 evidences the more uniform air pattern created when BO benefited from natural convection at laminar airflow range (<0.5 L•s-1•kg-1).

Increasing airflow canceled this advantageous effect of BO over TO, and even inverted the situation when there was more turbulence. For instance, Table 1 shows this last remark for more uniform air pattern through packages with 4 holes (0.5 and 1%) distributed on the top and submitted to 2 L•s-1•kg-1.

CONCLUSION The position of the openings rather than their area has a considerable influence on air distribution through the porous medium. The gravity force has a significant effect on air distribution. The diagonal opening configuration allowed demonstrating some of the gravity force effect on the air distribution in a packed produced container.

The air entering by the top openings of the container has a tendency to flow down due to its higher density compared to the warm air reaching the still region of the package. Thus, it is more rapidly distributed through the whole height of the container compared to the air entering through the bottom opening. However, the opposite effect was also present since the cold air entering by the bottom warm up in contact with the warm produce, decreasing its density and forcing to rise to the top of the package.

The airflow rate had a significant effect on determining which of these two opposite phenomena was predominant. Increasing airflow canceled this advantageous effect of bottom diagonal openings over top, and even inverted the situation when there was more turbulence. Hence the gravity effect should be considered in models developed to simulate airflow through horticultural produced packed in a partially open container. This is especially important when simulating very low airflow rates.

ACKNOWLEDGEMENTS This project was accomplished with the financial support from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, Brazil), and Agriculture and Agri-Food Canada (Canada).

REFERENCES: [1] Cortez, L.A.B., Castro L.R., Vigneault C.

Resfriamento Rápido a ar : métodos da câmara frigorífica e do ar forçado. In : Resfriamento de frutas e hortaliças. Embrapa Informaçã Tecnológica, Brasília, DF, Brazil, 2002

[2] Castro, L. R., Vigneault, C., Cortez, L. A. B. Effect of peripheral opening on cooling efficiency of horticultural produce. Postharvest Biology and Technology. 2004 (Submitted for publication)

[3] Castro, L. R., Vigneault, C., Cortez, L. A. B. Effect of container opening on air distribution during precooling of horticultural produce. Transactions of the ASAE. 2004 (Submitted for publication)

78

[4] Castro, L. R., Vigneault, C., Cortez, L. A. B. Container opening design for horticultural produce cooling efficiency. International Journal of Food, Agriculture and Environment. Vol.2, No 1, 2004. pp. 135-140.

[5] Vigneault, C., Castro, L. R. Indirect airflow distribution measurement for horticultural crop package, Part I: Development of the research tool. Transactions of the ASAE. 2004 (Submitted for publication)

[6] Vigneault, C., Castro, L. R., Goyette, B., Markarian, N. R., Charles, M. T., Bourgeois, G., Cortez, L. A. B. Indirect airflow distribution measurement for horticultural crop package: Part II: Verification of the research tool applicability. Transactions of the ASAE. 2004 (Submitted for publication)

[7] Vigneault C., Castro, L. R., Panneton, B. Laminar to turbulent indirect airflow measurement method for horticultural crop package. Canadian Agricultural Engineering. 2004 (Submitted for publication)

[8] Vigneault C., Castro, L. R. Gautron, G. Effet de la présence de poignées ouvertes sur le

prérefroidissement de produits horticoles. International Journal of Refrigeration. 2004 (Submitted for publication)

[9] Vigneault, C., Émond, J.P. Reusable container for the preservation of fresh fruits and vegetables. United States Patent Office. 1998. Patent no 5,727,711.

[10] Goyette, B., Vigneault, C., Panneton B., Raghavan, G. S. V. Method to evaluate the average temperature at the surface of a horticultural crop. Canadian Agricultural Engineering. Vol.38, No 4, 1996. pp. 291-295.

[11] SPSS Inc. Chicago, Illinois. USA. 2004. http://www.spss.com

[12] Vigneault, C., Goyette, B. Design of plastic container openings to optimize forced-air precooling of fruits and vegetables. Applied Engineering in Agriculture. Vol.18, No 1, 2002. pp. 73-76.

[13] Alvarez, G. & Flick, D. Analysis of heterogeneous cooling of agricultural products inside bins. Part II: thermal study. Journal of Food Engineering. Vol.39, 1999. pp. 239-245.

pitot tube

outlet tube

fan

air flow

air flow

mobile plastic tunnel

mobile fan set up

mobile heat exchanger

aluminum plate

ball matrix

static pressuremeasuring device plywood box

air flow

plastic wrappolystyrene foam

Fig. 1. Experimental set up including a forced air tunnel, ball matrix, fan and measuring devices.

79

#A #B #C Fig. 2. Dimension of the diagonal-opening-configuration plates #A, #B and #C used to test the effect of the

gravity. Table 1. Effect of opening configurations (OC) and airflow rates (L•s-1•kg-1) on air distribution heterogeneity (Vi) and APD (mm of water).

Vi Airflow rate OC 0.125 0.25 0.5 1 2 APD

2TO_0.5 0.80a 0.79a 0.75a 0.71a 0.64b 1.43b

2BO_0.5 0.76a 0.64b 0.58bc 0.65b 0.72a 1.46b 2TO_0.67 0.58c 0.62b 0.59b 0.53cd 0.55c 1.53a

2BO_0.67 0.54c 0.61b 0.52d 0.55c 0.56c 1.55 a 4TO_1 0.69b 0.61b 0.57c 0.50d 0.42d 1.34c 4BO_1 0.47d 0.47c 0.46e 0.52cd 0.51c 1.31 c

FO 0.19e 0.08d 0.10f 0.06e 0.04e 0.02d

APD 0.04d 0.03d 0.22c 1.02b 3.67a Means in the same column or the same group of numbers followed by the same letter are not significantly different based on t-Test using α = 0.05. Table 2. Effect of opening configurations (OC) and airflow rates (L•s-1•kg-1) on average HCT (min).

HCT Airflow rate OC 0.125 0.25 0.5 1 2

2TO_0.5 94.2a 53.7b 35.4bc 24.0b 17.0ab 2BO_0.5 83.8ab 53.0b 33.4c 23.0cd 16.7ab 2TO_0.67 92.4ab 55.7ab 35.8b 23.4c 17.1a 2BO_0.67 81.9ab 54.7ab 35.8b 22.6d 16.5ab

4TO_1 79.5ab 47.7c 31.0d 19.6e 13.8c 4BO_1 76.0b 46.3c 30.7d 19.1e 13.7c

FO 96.1a 59.5a 43.7a 24.7a 15.5b Means in the same column followed by the same letter are not significantly different based on t-Test using α = 0.05.

Table 3. Effect of opening configurations (OC) and airflow rates (L•s-1•kg-1) on HCTm (min).

HCTm Airflow rate OC 0.125 0.25 0.5 1 2

2TO_0.5 148.7b 90.3ab 70.9a 53.5a 39.8a

2BO_0.5 151.8b 87.3ab 53.6cd 46.4b 35.3ab

2TO_0.67 144.0b 93.0a 60.1b 47.5b 38.8a

2BO_0.67 135.3bc 87.2ab 57.4bc 39.2c 28.9bc

4TO_1 113.6c 78.0cd 51.3de 34.5cd 23.3cd

4BO_1 125.9bc 74.5d 49.3e 30.7d 21.4d

FO 183.3a 81.6bcd 59.7b 35.0cd 21.1d

Means in the same column followed by the same letter are not significantly different based on t-Test using α = 0.05. Table 4. Effect of the top (TO) and bottom (BO) diagonal opening configuration on air velocity (m•s-1) for the Y and Z directions.

TO BO Y Z Y Z

1 0.12d 0.16e 0.47a 0.22cd 2 0.13cd 0.19de 0.45a 0.22d 3 0.19c 0.23cde 0.34b 0.26bcd

4 0.19c 0.24cd 0.30b 0.27bcd

5 0.40a 0.26bcd 0.17c 0.28abcd

6 0.40a 0.29bc 0.16c 0.30abc 7 0.33b 0.32ab 0.18c 0.32ab

8 0.32b 0.38a 0.16c 0.35a

Means in the same column followed by the same letter are not significantly different based on t-Test using α = 0.05

80

Artigo 7. Effet des poignées ouvertes sur les contenants lors du prérefroidissement

de produits horticoles

Effect of open handles on packages during precooling process of horticultural

produce

Clément Vigneault1,2, Larissa R. de Castro1,3, Guillaume Gautron4

1Horticulture Research and

Development Centre, Agriculture and Agri-Food

Canada, St-Jean-sur-Richelieu, Québec, Canada, J3B 3E6

3 Faculdade de Engenharia Agrícola, UNICAMP, Cidade

Universitária Zeferino Vaz 13087-970, Campinas, SP,

Brazil

4Institut Universitaire de Technologie d’Amiens,

Département Génie Biologique, option Agronomie, 80 000

Amiens, France

2Corresponding author : vigneaultc@agr.gc.ca ABSTRACT Open handles on horticultural crop container may affect the airflow distribution through these packed produce. Their effects were evaluated using an experimental set up previously developed for airflow rates between 0.25 and 2 L•s-1•kg-1, and for opening area of 2 to 16% of the container surface. The results demonstrated that the type of opening has effects differing based on the container operating conditions. In general, open handles increase the uniformity of the air distribution with large-opening containers. However, the containers with small openings and open handles contribute to air distribution heterogeneity which increases the produce refrigeration time. This increase in time and lack of uniformity in air distribution could be compensated by enhancing the airflow rate.

KEY WORDS Package, precooling, uniformity, forced-air, air distribution.

INTRODUCTION A wide variety of containers have been developed for horticultural produce handling. However, their design is mostly founded on only production capability and structural resistance while the air distribution problematic during precooling process is neglected [0, 2]. A container must have enough openings to allow uniform and fast fluid circulation through the packed produce to assure their conservation and avoid excessive energy consume [0, 2].

For forced-air precooling method, openings area and position on the package surface have considerable effect on the process efficiency [3]. Vigneault and Goyette [1] do not recommend containers with opening area less than 25% because of the significant restriction of airflow or fan power required. On the other hand, a design with maximum area of 14% is suggested as sufficient for fast and uniform produce cooling process [3]. The performance of cooling process may be attained also by enhancing airflow rate, but if the vented area is lessened to below 10% the cooling cost will sharply increase[4].

Regarding opening location, the presence of handle openings may result in heterogeneous air distribution during precooling, enhancing chances of hot or

freezing spot formations [5]. Since air tends to circulate through the container region that offers least resistance and the handles are usually located near their top, a preferential pathway may be formed in the empty space above the produce [6, 7]. Therefore, the air bypasses most of the produce creating a wake that reduces the cooling efficiency [8]. The amount of air effectively circulating through the produce depends on the bulk porosity [7], the relative vented area percentage [1, 5] and position between the handles and the other openings on the container [3, 5, 6]. Depending on this area percentage, the gravity effect due to the density difference between cold inlet air and warm interstitial air may be significant [9]. This would change the direction and impact of air distribution through the produce [10]. Furthermore, the holes must be uniformly distributed on container surface to improve air circulation inside the package avoiding the air deviation to gaps [1, 11, 12].

The objective was to verify the effect of the presence of handle openings for horticultural produce containers on air circulation during forced-air cooling. The methodology developed by Vigneault et al. [13, 14] was used for the indirect measurement of airflow distribution through a porous medium. This allowed

81

pointing out the influence of the container designs on the rate and uniformity of cooling process and air distribution [9, 14].

MATERIAL AND METHODS Horticulture produce simulator The horticultural produce simulator described in detail by Vigneault and Castro [13] consists of solid polymer balls 52.36 mm in diameter and weighing 125.55 g. Sixty-four balls were selected for their relatively high uniformity in terms of cooling index (-0.1414 ± 0.0081 min-1) and heat capacity (1.1252 ± 0.0657 kJ·kg-1·oC-1). Each of the 64 balls was instrumented with a 30-gage 5 m-long insulated copper constantan thermocouple wire (Type T) placed in their center with a precision of ±0.025 mm.

Experimental set-up Figure 1 shows the experimental set up used during the trials and described in detail by Vigneault et al. [14]. Four transparent acrylic plates were assembled to simulate a forced-air cooling tunnel of 420 mm inside square cross-section, and 1250 mm long.

Sixty four instrumented balls were stacked uniformly distributed along with other 448 balls on a columnar pattern to form a cubic matrix of 8-ball-side dimension (Fig. 2), the z axe corresponding to the airflow direction [14]. The arrangement resulted in 47.64% of porosity. The two end layers on the “z” direction were constructed assembling the balls together using 12-mm-long and 6-mm-diameter plastic pins inserted into 6mm depth hole perforated at each ball to ball or ball at wall contact point.

The ball matrix was positioned at a distance of 220 mm from the tunnel-end air inlet. The portion of the tunnel containing the balls was insulated with a 25 mm-thick polystyrene foam to reduce heat conduction. The air-outlet of the tunnel consisted on a 610 mm long plenum enabling air pressure drop (PP) measurements across the ball matrix using a pressure transmitter (0-127±6 mm of water, Model 607-7, Dwyer Instruments Inc. Michigan City, IN, USA).

During the trials, the end of the air-outlet tunnel was air-tightly attached to the aspiration chamber. A direct drive radial blade fan driven by a 0.75 kW variable speed motor created a negative pressure in the aspiration chamber and forced the outside air to circulate through the cooling tunnel. The air was released to the atmosphere through a 500 mm long, 101.6 mm-diameter tube instrumented with a Pitot-tube device allowing airflow measurements. Two transmitters, 0-12.7+0.6 and 0-2.54+0.13 mm of water (Models 607-2 and 607-3, Dwyer Instruments Inc. Michigan City, IN, USA) were used to measure the static and total pressures at the Pitot tube.

The whole experimental set-up was placed in a cold chamber kept at 4+1oC to simulate the precooling

process of the balls matrix. The end of the air-inlet tunnel was attached to a heat-exchanger formed by thirty 520 x 1100 x 1.5 mm aluminum plates placed at a 30 mm distance from each other. This exchanger allowed reducing the temperature variation due to the oscillation of the cold chamber temperature to less than 0.1oC [14].

A data acquisition system was used to record, simultaneously, the temperature inside the sixty-four balls along with air temperature before and after crossing the matrix of balls, the temperature in the centre of the cold chamber, the pressure drop through the matrix and plates, and the air dynamic pressure through the airflow measuring device, at a 20-s-interval.

Container opening configurations Four pairs of plates were used to simulate different vented area percentages and two standard openings configurations, open and closed handles. Several openings 9.52 mm wide and 11.1, 22.3, 44.5 and 77.8 mm long were drilled in square polypropylene plates of 420 mm and 3 mm in thickness to obtain the total venting areas (2, 4, 8, and 16%). The openings representing standard handles [12] were also perforated on the plates, so the closed configuration could be obtained by covering the handles with airtight sealing tape. Figure 3 presents the dimensions of the 2 and 16% plates with open handles. At each test, a pair of plates was placed next to the first and eighth layers of balls to enclose the matrix, perpendicular to the airflow direction. A fully open configuration was also tested for comparison by not using plastic plates.

Experimental procedure The nine opening configurations were tested with

four airflow rates corresponding to 0.25, 0.5, 1, and 2 L•s-1•kg-1 of produce according to the recommendations of Vigneault et al. [14]. The maximum air pressure drop through the porous medium was settled as 125 mm of water, which corresponded to the transmitter upper-limit [14]. The required airflow rates were achieved by a computerized speed control system that adjusted the fan motor rotation according to the dynamic pressure at the Pitot tube. Each experiment was performed in a complete block design and repeated two times following the recommendations of Vigneault et al. [14].

Prior to the start of each test, the forced-air tunnel containing the balls was placed in a chamber kept at approximately 28oC. An axial fan forced the air through the matrix to reheat the balls for two hours. After the heating period, the perforated plates were installed using the sealing tape and the tunnel was placed in the cold room. The tunnel-end air inlet was connected to the heat-exchanger and the air outlet to

82

the aspiration chamber. The fan was then turned on immediately. The data acquisition and control software recorded the results until the temperature of the warmest ball had reached 6.9oC and then turned off the system. The temperature-time data recorded were used to calculate the half-cooling time (HCT) of each ball for all treatments by using a dedicated ExcelTM macro developed by Goyette et al. [15].

Compared Parameters Uniformity coefficient

A ‘Vi’ number was developed to compare the performance of different opening configurations. This dimensionless number is the coefficient of heterogeneity of the air velocity distribution through a porous medium. It was defined as the ratio between the standard deviation and the mean of the air velocities in the surroundings of the 64 instrumented balls obtained with the method described by Vigneault et al. [14]. The minimum value for ‘Vi’ number is 0 which would mean perfect air distribution uniformity with no variation through the porous medium. The ‘Vi’ obtained with the 100% vented configuration (without any plate) was used as a reference to establish the effect of each of the other opening configurations.

Precooling duration The required period to cool down a mass of produce should be an indicator of the precooling performance. This time could be also used to compare different cooling systems efficiencies or any other variables influencing the produce cooling. Moreover, this time would allow evaluating the energy necessary for produce precooling according to the refrigeration, type of container, or any combination of parameters capable to affect the process performance. This performance index was denominated here as the precooling time (HCT m, min) and corresponded to the time necessary to reduce the centre temperature of the most slowly chilled ball by a 7/8 difference between its initial and the fluid temperatures. HCT m index was based on the 7/8 cooling time [3].

RESULTS AND DISCUSSION Table 1 summarizes the data obtained in the trials including the results of the statistical analysis for the effects of the vented percentage (PO, %), airflow rate (L•s-1•kg-1), and the type of handles (open or closed) on the mean air velocity around the balls (m•s-1), its heterogeneity (Vi), the average maximum half-cooling time (HCTm, min) and pressure drop (PP, mm of water).

Influence of the vented area percentage The vented area percentage had no significant effect

on the surrounding mean air velocity at low airflow rate (0.25 L•s-1•kg-1). However, raising the airflow progressively increased its effect until generate

significant differences among all openings percentages at the maximum airflow rate tested (2 L•s-1•kg-1).

The air distribution heterogeneity through the produce and the pressure drop were inversely proportional to the total vented area percentage on the container walls. Nevertheless, the same opening effect was not observed on cooling time. These results correspond to Vigneault et al.’s [9] findings for a similar study concerning the gravity effect on the air pathway in horticultural produce containers.

Airflow rate effect The mean air velocity around the balls was directly proportional to the airflow. However, the effect of airflow rate on the air distribution heterogeneity through the produce (Vi) was less evident. For less vented area (2%) and open handles, Vi number was inversely proportional to the airflow rates tested. The opposite effect was observed when enlarging the opening area. The influence of airflow on Vi number was not as considerable when using closed handles.

The average maximum half-cooling time (HCTm, min) was inversely proportional to the airflow whereas the pressure drop (PP, mm of water) increased as the flow rose, for all openings and handles types investigated. These results correspond to those presented by Castro et al. [3] when verifying the effect of openings positioning on air distribution through horticultural produce containers.

Influence of handles type The handles presence increased the mean air velocity at all the airflow rates and vented area percentages tested though this effect was not always significant.

For less total vented percentage (2%) and low airflow rate (0.25 L•s-1•kg-1), the open handles produced a negative effect on air distribution through the containers by generating higher Vi values. In this case, most air flowing through this opening bypassed a large amount of produce. When the vented area was increased, this effect was attenuated by enhancing airflow rate until be completely inversed at a high airflow rate and large opening percentage. The points of airflow rate effect inversion were 2, 1, 0.25 L•s-1•kg-

1 for 4, 8 and 16%, respectively. The values of air distribution heterogeneity through the produce (Vi) for a container with open or closed handles can be obtained using the equations 1 and 3, respectively. Thus, the effect of open handles on Vi varied with airflow rate and vented area percentage ( OP , %) whereas it depended only on the vented percentage for the closed handles configuration.

The cooling time has also varied in different ways according to the airflow rate depending on the handles configuration used. Hence, equations 2 and 4 allow determining the maximum half-cooling time (HCTm,

83

min) as a function of airflow rate ( AQ , L•s-1•kg-1) for open and closed handles, respectively.

Open handles 2004.0084.0078.0728.0 OOA PPQVi +−−= (R2=0.863) (1)

281.313.1119.118 AAm QQHCT +−= (R2=0.978) (2)

Closed handles 20017.00312.03453.0 OO PPVi +−= (R2=0,634) (3)

232.416.1363.131 AAm QQHCT +−= (R2=0.977) (4)

In most cases the presence of the handles contributed to reduce pressure drop (PP, mm of water). For less vented containers (2 and 4%) and airflow rates ≤ 1 L•s-1•kg-1, PP was higher when using closed handles than for open handles. This difference in results due to the handles configuration disappeared when enlarging the total vented area. Indeed, for larger opening areas (8 and 16%) at lower QA (0.25 and 0.5 L•s-1•kg-1) the handles did not produce any significant effect on PP. At the maximum area (16%) and minimum airflow rate (0.25 L•s-1•kg-1), the handles had no influence on any of the studied variables (mean air velocity, air distribution heterogeneity, average half-cooling time, and pressure drop).

Consequently, when the container chosen has open handles and also a structural restraint forcing to minimize vented area percentage, the airflow rate should be substantially raised to assure a fast and uniform precooling. On the other hand, if the container has no structural limitation, such as the plastic packages that can contain 25% of opening area or more, a reduced airflow rate, with even open handles, will contribute for air circulation uniformity. Furthermore, it could generate maximum precooling time (HCTm) values relatively low by decreasing Vi.

CONCLUSION The effect of open or closed handles on the air distribution through horticultural produce was evaluated using a previous experimental set up. This effect was evaluated under different conditions of airflow rate and vented area percentage. In general, the open handles increased the air distribution uniformity through containers with large vented surfaces. On the other hand, containers with open handles and less vented area percentages enhanced the heterogeneity of air circulation, which in turn increased produce precooling time. Increasing airflow rate could compensate for the lower performance of those containers.

REFERENCES 1. Vigneault, C. and B. Goyette. 2002. Design of

plastic container opening to optimize forced-air

precooling of fruits and vegetables. Transactions of the ASAE, 38(1), p.73-76.

2. Vigneault, C. and B. Goyette. 2002. Largeur des ouvertures au fond de contenants de plastique utilisés pour la manutention des produits horticoles frais. Can. Biosyste. Eng. 44(3): 7-10

3. Castro, L. R., C. Vigneault, and L. A. B. Cortez. 2003. Container opening design for horticultural produce cooling efficiency. Int. J. Food Agr. Environ. 2(1): 135-140.

4. Baird, C. D., J. J. Gaffney, M. T. Talbot. 1988. Design criteria for efficient and cost effective forced-air cooling systems for fruits and vegetables. ASHRAE Transactions, 94: 1434-1453.

5. Émond, J. P., F. Mercier, S.O. Sadfa, M. Bourré and A. Gakwaya. 1996. Study of parameters affecting cooling rate and temperature distribution in forced-air precooling of strawberry. Transactions of the ASAE, 39(6): 2185-2191.

6. Alvarez, G. and D., Flick. 1999. Analysis of heterogeneous cooling of agricultural products inside bins. Part I: aerodynamic study. J. Food Eng. 39: 227-237.

7. Vigneault C., N.R. Markarian, A. da Silva and B. Goyette. 2004. Pressure drop during forced-air circulation of various horticultural produce. Transaction of the ASAE. 47 (3): 807-814.

8. Leyte, J. C. and C. F. Forney. 1999. Optmizing flat design for forced-air cooling of blueberries packaged in plastic clamshells. HortTechnology. 9(2):202-205.

9. Vigneault, C., L.R.de Castro, L.A.B. Cortez . 2004. Effect of gravity on forced-air precooling. IASME Transactions On Mechanical Engineering. (In press)

10. North, M. F., S. J. Lovatt, and A. C. Cleland. 1998. Methods for evaluating the effect of large voids on food package cooling times. Acta Hort. 476: 95-103

11. Arifin, B. B. and K. V. Chau. 1987. Forced-air cooling of strawberries. ASAE Paper no. 87-6004. St. Joseph, Mich.: ASAE.

12. Vigneault, C, Émond, J.P. 1998. Reusable container for the preservation of fresh fruits and vegetables. United States Patent Application Office Washington, WA Dc, USA. Patent Number: 5,727,711. 60pp.

13. Vigneault, C. and L. R. de Castro. 2004. Indirect airflow distribution measurement for horticultural crop package: Part I: Development of the research tool. Transaction of the ASAE (submitted).

14. Vigneault, C., N. R. Markarian, B. Goyette and L. R. de Castro. 2004. Indirect airflow

84

distribution measurements in horticultural crop packages. ASAE Annual International Meeting Ottawa, Ontario. 1 - 4 August. ASAE paper No. 04-6163. 14pp.

15. Goyette, B., Vigneault, C., Panneton B., Raghavan, G. S. V. Method to evaluate the average temperature at the surface of a horticultural crop. Canadian Agricultural Engineering. Vol.38, No 4, 1996. pp. 291-295.

ACKNOWLEDGEMENT Funding for this project was provided by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) and Horticulture Research and Development Centre of Agriculture and Agri-Food Canada. This is a result of collaboration among a Canadian, Brazilian, and French engineer.

.

Figure 1. Experimental set up including a forced air tunnel, ball matrix, fan and measuring devices.

85

Figure 2. Cubic ball matrix used to evaluate the effect of opening on air circulation.

A) B)

Figure 3. Examples of plates used to evaluate the effect of open and closed handle on air circulation in

horticultural crop containers during forced air precooling process; A) 2% total opening, B) 16% total opening excluding the handle opening area.

86

Table 1. Effect of vented area percentage (%), airflow rate (L•s-1•kg-1) and handles type on mean air

velocity around the balls (m•s-1), air distribution heterogeneity through the produce (Vi), average

maximum half-cooling time HCTm (min), and pressure drop (PP, mm of water).

Opening (%)

Airflow (L•s-

1•kg-1) Handle Air velocity

(m•s-1) Vi HCTm

(min) PP

(mm of water) 2 0.25 open 0.080abcd 0.610n 100.04jkl 2.41bcd

0.25 closed 0.060a 0.269fgh 105.73l 3.66fg 0.5 open 0.151f 0.611n 73.75g 4.14gh

0.5 closed 0.127ef 0.291ghij 64.65ef 8.76k 1 open 0.324j 0.509m 44.94d 11.79m

1 closed 0.289hij 0.307hijk 37.33bc 30.84o 2 open 0.650p 0.335jk 24.60a 43.41p

2 closed 0.644p 0.289ghi 20.73a 116.9q 4 0.25 open 0.071abc 0.419l 96.74ijk 2.01abc

0.25 closed 0.062ab 0.234def 102.97kl 2.13abc 0.5 open 0.140ef 0.413l 67.67fg 2.62cde

0.5 closed 0.122ef 0.267efgh 67.33f 3.02def 1 open 0.294ij 0.343k 41.34cd 5.41i

1 closed 0.266ghi 0.265efgh 39.69bcd 8.10j 2 open 0.588no 0.235def 24.19a 17.17n

2 closed 0.585no 0.229def 22.07a 30.20o 8 0.25 open 0.065ab 0.252efg 94.50ij 1.80ab

0.25 closed 0.060a 0.167c 104.55l 1.80ab 0.5 open 0.126ef 0.263efgh 64.23ef 2.01abc

0.5 closed 0.109cde 0.199cd 67.95fg 2.08abc 1 open 0.268ghi 0.232def 40.39bcd 3.15ef

1 closed 0.249gh 0.222de 40.08bcd 3.63fg 2 open 0.558mn 0.166c 22.47a 7.65j

2 closed 0.513l 0.223def 24.82a 9.86l 16 0.25 open 0.067ab 0.203cd 90.53i 1.75a

0.25 closed 0.064ab 0.223de 101.10kl 1.75a 0.5 open 0.124ef 0.256efg 62.71ef 1.83ab

0.5 closed 0.116def 0.321ijk 66.15f 1.83ab 1 open 0.263ghi 0.233def 39.32bcd 2.29abc

1 closed 0.258ghi 0.283ghi 39.71bcd 2.36abc 2 open 0.610op 0.252efg 21.77a 4.09gh

2 closed 0.534lm 0.281ghi 23.80a 4.67h 100 0.25 0.064ab 0.075ab 81.57h 1.75a

0.5 0.103bcde 0.102b 59.62e 1.78a 1 0.243g 0.062ab 34.93b 1.93ab 2 0.495l 0.039a 21.13a 3.15ef

Means in the same column followed by the same letter are not significantly different at a 0.05 significance level.

87

Artigo 8. Effect of container openings and airflow rate on energy required for

forced-air cooling of horticultural produce

LARISSA R. DE CASTRO1,2, CLEMENT VIGNEAULT1,2,3, LUIS A. B. CORTEZ2

1Horticultural Research and Development Centre

Agriculture and Agri-Food Canada 430 Gouin, Saint-Jean-sur-Richelieu (Qc)

CANADA J3B 3E6

2College of Agricultural Engineering State University of Campinas

Cidade Universitaria Zeferino Vaz 13083-970, Campinas SP, BRAZIL

3 Corresponding author: Tel : 450-346-4494 ext. 170, Fax: 450-346-7740, Email : vigneaultc@agr.gc.ca

ABSTRACT A research tool previously developed to investigate air distribution in horticultural produce containers during forced-air precooling was used to determine the effect of airflow rate and opening configuration on air pressure drop and rate and uniformity of cooling process. Further analysis performed on previously tested opening configurations determined their influence on energy efficiency. A system efficiency coefficient, consisting of the overall Energy Added Ratio (EAR) was demonstrated as a functional tool during the container design, since it considers peculiarities of the forced-air cooling system and produce physiology. The results obtained for containers with handling openings and 2, 4, 8, and 16% opening area were used to evaluate the additional energy required to remove the heat generated by the forced-air fan and produce respiration. These results were also compared to produce in bulk and to produce packed in containers having 4-0.5%-holes in the corners to analyze the influence of hole positioning. The four large 0.5% opening configuration resulted in poor energy performance and cooling uniformity when compared to uniformly distributed smaller holes. Furthermore, the airflow rate could be optimized based on the respiration rate of the produce and container opening area.

KEY WORDS Efficiency, design, respiration rate, box, handling, packing.

INTRODUCTION The efficiency of a forced-air cooling process for fruits and vegetables is mainly indicated by the cooling rate and cooling uniformity it produces (Vigneault et al. 2004b) in contrast to the energy input required by precooling and refrigeration systems (Thompson and Chen 1988). However, a more efficient process is needed to better maintain produce quality and reduce energy consumption. Cooling process performance determines the amount of electrical energy consumed directly to operate compressor and fans (ASHRAE 2000). Energy consumption of the forced-air circulation system depends on the airflow rate through the fans, which is related to the fan’s physical characteristics, operating speed and pressure drop (Brooker et al. 1974).

The energy effectiveness of a particular cooling method is expressed in terms of an energy coefficient (EC), which is defined as the ratio of total thermal energy removed (Et, kJ) and the sum of electrical energy (EE, kJ) used through the cooling process (Eq. 1) (ASHRAE 2000). Et refers to the heat load to be removed such as produce heat (field and respiration) and heat transferred to the cold chamber through doors, walls, ceiling, floor, lights, people, and machines (Thompson and Chen 1988). Produce field and respiration heat account for the majority of Et, but

the portion of each factor also varies with the system. The second term, EE, includes the energy consumed by the refrigeration system and by the forced-air precooling equipment.

∑=

EEE

EC t (1)

In this case, EC is also known as COP, the Coefficient Of Performance (ASHRAE 2000), and varies between 2.5 and 3.5. However, this value does not include the energy used by the forced-air circulation system required for the precooling process, which depends on the characteristics of the cooling system. Thompson and Chen (1988) claimed a COP of 0.4 for forced-air cooling, which is in the 0.25 and 0.47 range reported by Kader (2002). Nevertheless, this last author also recommends 5% as the maximum container opening area, which affects the COP of the system (Vigneault and Goyette 2002).

The ventilation system used to force air through the horticultural produce consumes some energy to transfer the thermal energy to the refrigeration system. This energy is released to the air during the cooling process when it crosses the fan blades or circulates around the driving motor. The amount of energy is proportional to the electrical input to run the fan (ASHRAE 2001). The same author presented the

88

method to calculate the total heat transferred (Qair, J) by an air circulation system to the environment by using equation (2) where fm, h is the fraction of motor heat loss transferred to air stream, ηm is the motor efficiency and EE the electrical energy (J) used by the motor driving the fan.

)hm,fmη - (1 m(ηEE airQ )+= (2)

There are several factors affecting the efficiency of forced air cooling of packed horticultural crops. Some parameters may be set to decrease the process time and improve the cooling uniformity but also increase the energy consumption. For example, increasing air velocity through the produce enhances the cooling homogeneity (Castro et al. 2004a) but, requires more energy to drive the fan and more refrigeration energy to remove all the additional heat produced by air friction and fan inefficiency (Baird et al. 1988). The additional energy resulting from faster air circulation could be compensated, up to a certain limit, by more uniform air distribution and faster cooling rate.

On the other hand, lowering the air velocity through a mass of produce could reduce the energy required per unit of time but the operating time often increases the total energy required. Vigneault et al. (2004a) developed a dimensionless number (Vi) to compare the performance of different forced air cooling systems. The Vi number was defined as the coefficient of heterogeneity of air velocity distribution through a porous medium.

Optimal operating conditions should be determined for maximal efficiency. One way of enhancing the efficiency is by enlarging the container opening area which reduces the pressure drop through the whole system. However, since this method decreases the structural resistance of the container and the supporting surface supplied to the produce, it should be carefully studied. The container opening area and position, rather than the shape, play an important role in cooling efficiency (Vigneault and Goyette 2002; Kader 2002; Edeogu et al. 1997; Arifin and Chau 1988; Baird et al. 1988; Haas et al. 1976). For liquid-ice processes, the vented area must be sufficiently narrow to minimize the loss of ice particles (Vigneault and Goyette 2001). In the case of forced-air cooling, however, restricting the openings to less than 25% of the container surface significantly increases the air pressure drop and consequently, the energy required to run the fan (Vigneault and Goyette 2002; Haas et al. 1976). Baird et al. (1988) also reported that energy consumption increases as the opening area decreases, suggesting 10% as the minimum opening area needed to not compromise the cooling rate.

Performance calculations of the fan and air distribution systems require a detailed pressure balance of the entire network. Air is driven by the pressure differential, so any obstruction in its path restricts its

circulation through packed produce. Therefore, cooling efficiency can be jeopardized by misalignment of the openings of palletized containers, inappropriate produce stacking arrangement, or secondary packaging (Faubion and Kader, 1997; Chau et al. 1985). Since so many factors affect the cooling efficiency, a research tool was developed by Vigneault and Castro (2004) and Vigneault et al. (2004b) to perform tests under controlled and stable conditions.

The aim of this research was to evaluate the effect of opening configuration, total area and position, and airflow rate on the overall cooling system efficiency. Particular objectives were to develop a coefficient to verify the influence of produce respiration rate on energy requirement and to establish additional criteria for designing containers for fruit and vegetable handling.

MATERIAL AND METHODS Produce simulator Solid polymer spherical balls of 52.4 mm diameter and 125.6 g were used to represent horticultural produce of the same shape, as described in detail by Vigneault and Castro (2004). The balls were selected for their relatively high uniformity in terms of cooling index (-0.141±0.008 min-1) and heat capacity (1.12±0.07 kJ•kg-1•oC-1). Each ball was instrumented with thermocouple positioned at its center.

Experimental set-up Sixty four instrumented balls were stacked uniformly distributed throughout a stack of other 448 other balls to form a cubic matrix of 8x8x8 balls. Four acrylic plates were assembled to simulate a forced-air cooling tunnel containing the ball matrix (Figure 1). The air-outlet of the tunnel consisted of a 610 mm long plenum enabling air pressure drop (APD) measurements across the ball matrix. The end of the air-outlet tunnel was tightly sealed to an aspiration chamber. The air was released to the atmosphere through an airflow measurement device. The whole experimental set-up was placed in a cold chamber maintained at 4oC to generate a precooling process for the ball matrix. A heat-exchanger was built to minimize the temperature variation at the air inlet during the experiments. The center temperature of the sixty-four balls, the air temperature in the cooling tunnel before and after the ball matrix, the temperature of the cold chamber, the pressure drop through the ball matrix and plates, and the dynamic pressure of the air circulating through the airflow measurement device, were simultaneously recorded at 20 s intervals.

Opening configurations A fully open configuration was initially tested to determine the correlation between the half-cooling time (HCT) of the 64 produce simulators and air approach velocity. Based on the ball matrix volume,

89

the airflow rates studied were equivalent to 0.5, 1, 2 and 3.9 L•s-1•kg-1 of apple (Vigneault et al. 2004c). A group of opening configurations was investigated by placing a pair of plates on both opened sides of the matrix to simulate a two-side-perforated package. The openings configurations (Fig. 2) were selected among those studied in previous research (Castro et al. 2004b; Vigneault et al. 2004a), according to their responses to air distribution heterogeneity (Vi) and air pressure drop (APD) when compared to the fully open matrix. The first configuration with peripheral openings only, which is commonly used in the fruit and vegetable packing industry, consisted of 4-0.5%-holes distributed near the corners of the package and accounting for a total open area of 2%. The four other configurations consisted of a standard handle and uniformly distributed openings (Vigneault et al. 2004a). The openings were made for 3 mm-width opened slots of different lengths uniformly distributed on the plate resulting in total venting areas of 2, 4, 8 and 16%.

Experimental procedure Prior to the beginning of each test, the forced-air tunnel containing the balls was placed in a warm chamber maintained at 28±1.0oC. The balls were warmed using a forced air system. After this conditioning period, the perforated plates were installed and the tunnel was placed in the 4oC cold room. The forced air cooling system was immediately turned on. The data were recorded until the temperature of the warmest ball had reached 6.9oC. The temperature-time data recorded was used to calculate the HCT and cooling rate (CR) of each ball for all treatments by using a dedicated ExcelTM macro developed by Goyette et al. (1996).

Energy consumption The total thermal energy (Et, kJ) (Eq. 1) was estimated by calculating independently the different energy sources involved in the process, which included the produce field heat (Ep, kJ), respiration heat (Er, kJ) and the ventilation energy required by the forced air system (Ev, kJ).

The field heat (Eq. 3) consists of the energy removed from a mass of produce (m, kg) with a specific heat (cp, kJ•kg-1•oC-1) to reduce its initial temperature (Ti, oC) to a final temperature (Tf, oC).

)T-(T c m E ifpp = (3) The respiration heat (Er, kJ) generated during the

cooling process refers to the bio-chemical energy generated by physiological activity of living vegetable or fruit. The heat released in this process depends on the type and the mass of produce, the temperature (T, oC) along the cooling process and the duration of the precooling process (t, s) which corresponds to the time required for the slowest cooling ball to reach a 7/8

cooling process (Eq. 4). All the calculations were made assuming a 1 kg mass of produce.

tr ∆∑= produce,-6

r Qm10 E (4) The four horticultural produce selected for

comparison covered the full range of respiration rates suggested by Kader (2002): low (apple), moderate (lettuce), high (strawberry) and very high (broccoli). These respiration rates (Qr,produce, mW•kg-1) are calculated with exponential equations (Eq. 5a, b, c, d) obtained by regression analysis of data presented by ASHRAE (2002) for temperatures (T, oC) ranging from 0 to 25 oC.

0.0933Tapple r, 8.848e Q = (5a)

0.0807Tlettuce r, 36.54e Q = (5b)

0.1048Tstrawberry r, 48.67e Q = (5c)

0.1197Tbroccoli r, 87.01e Q = (5d)

Considering a precooling system where the fan and motor assembly is mounted inside the air stream, the fraction of motor heat loss transferred to the air stream (fm,h, Eq. 2) is equal to 1 (ASHRAE 2000). The ventilation energy (Ev, kJ) (Eq. 6) depends only on the airflow (Dair, m3•s-1), the total pressure drop (APD, kPa) across produce and container, the fan efficiency (ηm) generally considered as 0.6 (ASHRAE 2000), and the fan operating time (t, s).

tη

APD D E

m

airv

= (6)

The APD was calculated by adding the APD through the produce and through the container openings, which were experimentally obtained by Vigneault et al. (2004c), and Vigneault et al. (2004a) and Castro et al. (2004b) respectively.

The fan operation time was considered here as only package and airflow rate dependent. The produce cooling process was considered the same for the fruits and vegetables used as examples (apple, lettuce, strawberry, and broccoli) and was based on the results obtained using the produce simulator. This simplification does not compromise the analyses performed because the main objective was to determine the effect of the different package configurations and airflow rates on energy efficiency. Besides, the aim was to represent produce from four different categories rather than specific produce. Using the cooling rates of specific fruits and vegetables would not achieve either of these research goals. Nevertheless, respiration energy (Er) and the ventilation energy (Ev) are both dependent on the duration of the cooling process. Therefore, the absolute values of the different sources of energy would change if the actual cooling rates of those fruits and vegetables were used, but their relative values should be about the same.

90

Energy Added Ratio Since the mass of produce and the temperature differential, and thus the field heat, were the same for each produce and airflow comparison, an overall energy added ratio (EAR) was developed to measure the effect of container openings and airflow on the energy to be removed during the cooling process. EAR (Eq. 7) is the ratio of the energy added during the cooling process, which are the respiration energy (Er, J) and the ventilation energy (Ev, J), compared to the initial energy that the produce contains at the beginning of this process, namely the field heat energy (Ep, J).

p

vr

EEE +

= EAR (7)

The advantage of using EAR to compare the different systems instead of the standard COP is that EAR considers Ep, Er and Ev which do not depend on the mechanical characteristics of the cooling system. The other sources of energy to be removed as well as the total electrical energy that would be required for cooling process, including the cold chamber operation energy, are not considered. Therefore, a performance analysis using EAR could be applied as a simple and more practical method to compare the performances of any precooling system for different airflow rates and opening configurations. A value of 0 would represent a system that extracts instantaneously the field heat of the produce; thus, any higher value of EAR represents a decrease of energy performance.

Statistical analysis The effect of opening area and airflow rate on energy were investigated for produce of different respiration rates and porosities, resulting in various operating times, pressure drops, and energy requirements. EAR was calculated for each combination of opening configuration, airflow rate and respiration activity level. The EAR results were then analyzed through a Multivariate Analysis of Variance followed by a Tukey test at 0.05 significance level using SPSS v. 11.5 (SPSS Inc. 2004).

RESULTS The opening configuration and airflow rate both had a significant effect on the energy added ratio (EAR). Nevertheless, their effects were highly influenced by produce respiration rates (Fig. 3).

In general, EAR of the cooling process decreased as the opening area increased. Enlarging the opening area to 8% reduced the added energy but, only a slight decrease was observed when increasing from 8 to 100% (Table 1).

For all respiration rates, the difference between the EAR results with 8 and 16% opening area was not significant, except with an airflow rate of 2 L•s-1•kg-1.

No statistical difference was found between 8 and 16% for the full duration of the process, and for all the pressure drops, with an airflow rate up to 0.5 L•s-1•kg-1. Since both respiration and ventilation energies depend on the cooling time, the shorter cooling time resulting from a higher airflow rate (1 L•s-1•kg-1) compensated for the higher pressure drop caused by less opening area. However, at the maximum airflow rate and 8% opening area, the cooling process was not fast enough to maintain EAR at low levels. At this airflow rate the 16% opening showed an EAR as low as 0.426% demonstrating very little opportunity for increasing the performance of the system when the opening area is fairly large at low produce respiration rates, which explains the non-significant difference obtained between 16 and 100% openings at 2 L•s-1•kg-1.

Respiration Rates Low respiration rate Figure 4 shows a general increase of EAR while increasing the airflow rate for low respiration produce. According to the same figure, decreasing the opening area and increasing the airflow rate required more additional energy. However, the effect of airflow rate on energy progressively declined as the opening area was increased. In fact, for the largest opening areas (16 and 100%), no significant difference was found between 1 and 2 L•s-1•kg-1. Moreover, an inversion of effect occurred since at these two opening percentages the highest airflow rate required less energy than at 0.5 L•s-1•kg-1. This was due to the fact that when the airflow rate is enhanced at these opening areas the increase in added energy (20%) is less considerable than the energy reduction (40%) caused by the decrease of the cooling-process time resulting from the respiration and ventilation energies. Therefore, when the container is designed for low respiration rate produce, the pressure drop will be the limiting factor for lower opening areas but not for larger opening areas. For containers with less opening area, increasing the airflow rate to reduce cooling time and improve uniformity of air distribution does not result in energy savings because it is not enough to overcome the higher increment in pressure drop. Baird et al. (1988) also stated that this increment would result in a critical increase of cooling costs especially for areas less than 3%. On the other hand, for areas of 16% and higher the cooling time determined the energy efficiency since the container openings did not generate an important APD. In this case, increasing the airflow had greater effect on increasing cooling rate and therefore limiting respiration activity, than on increment of APD.

Besides reducing opening area, the stacking arrangement may also aggravate the air restriction. It is important to mention that this research considered the

91

data produced by a columnar stack pattern, but the actual pressure drop would be even higher for a random stacking as observed by Chau et al. (1985) with oranges in cartons having 4% opening area. However, the results would be approximately the same if low respiration produce of other shapes were considered, such as carrot or celery, since the arrangement porosity was previously showed to influence APD only on spherical produce (Vigneault et al 2004c).

Very high respiration rate For very high respiration rate produce, the EAR reached minimum values as a function of the airflow rate for each opening area (Fig. 4). Beyond these points, any increase in airflow rate to reduce cooling time and consequently respiration and ventilation energy, did not compensate for the energy increase due to the APD. Likewise, below these optimum airflow rates, lowering the airflow, and thus the pressure drop, would not compensate for the longer processing time and higher respiration activities.

The regression equations presented for each curve (Fig. 4) were used to calculate the optimum airflow rates, which increased as the area increased. Optimum airflow rates of 1.35, 1.56, 1.73 and 2.08 L•s-1•kg-1 were found for opening areas of 2, 4, 8, and 16%, respectively. This optimum airflow rate tendency was also noticed at a moderate respiration rate for opening areas equal or superior to 4%. However, this tendency for optimum airflow rates increased for produce generating more respiration heat (Fig. 5). The curve tendency could not be identified at low respiration rate likely because the inflexion point occurred at a lower value than the one studied (0.5 L•s-1•kg-1), especially for smaller opening areas.

Therefore, with higher produce respiratory activity, increased airflow is required to hasten cooling process and compensate for the respiration energy increase. When a smaller opening area is used, a lower airflow rate is necessary to produce this equilibrium due to a larger APD. On the other hand, with a larger opening area, the energy efficiency must be optimized by decreasing the magnitude of respiration energy through faster cooling obtained with a higher airflow rate. This outcome is confirmed with the cooling time responses found by Arifin and Chau (1988) for strawberries packed in carton with four openings and 18% total vented area. They reported a 50% reduction in cooling time (136 to 72 min) when increasing airflow from 1 to 2 L•s-1•kg-1. Similar results were found by the present authors for this high respiration product (129 to 65 min).

Opening Positions The 2% opening area formed by four 0.5%-holes

distributed on corners of package surface presented the

highest energy added during the cooling process. Increasing the airflow increased EAR, although no significant difference was found between the results produced with 0.5 and 1 L•s-1•kg-1. Table 2 shows the magnitude of the effect for the maximum airflow rate on energy efficiency compared to produce respiration activity. The four-hole container configuration submitted to 2 L•s-1•kg-1 generated the highest energy added regardless of the respiration activity. These high values were partially due to the poor uniformity of air distribution (Vi=0.83) and especially to the high APD, 1.27 kPa (Castro et al. 2004b), compared to the 2% uniformly distributed opening area package which produced a Vi of 0.34 and APD of 0.43 kPa at the same airflow rate (Vigneault et al. 2004a).

On the other hand, the lowest EAR was obtained with the holes distributed uniformly on the package surface at the lowest respiration and airflow rates (Table 2). At high and very high respiration activity, EAR tended to first decrease but then to increase as more air was circulated through the system. Lowering the produce respiration activity decreased the airflow rate value correspondent to the minimum EAR result.

The main result obtained for the comparison of the two types of 2% opening configurations was the sharp difference between their additional energy required at the maximum airflow supplied. This large difference was due to the contrary effect of airflow rate on cooling uniformity between the two opening positions. When airflow rose, heterogeneity of air distribution (Vi) was reduced in containers with evenly distributed openings, but enhanced in peripherally positioned openings. Furthermore, this increase of Vi limited the improvement of the cooling rate at higher air velocities. Thus, for the corner opening configuration, the reduction of cooling time at the maximum airflow level is not enough to offset the greater increase of Vi and APD.

For instance, at 0.5 and 2 L•s-1•kg-1 the maximum half-cooling times (HCTmax) obtained with the peripheral openings are 12 and 42% higher than for uniformly distributed holes, respectively. The difference between the Vi values for the two opening configurations was even larger than for HCTmax, going from 10 to 60%. Yet for the same airflow rate levels, the pressure drop difference showed a lower increase, from 42% at 0.5 L•s-1•kg-1 and 66% at 2 L•s-1•kg-1 (Castro et al. 2004b and Vigneault et al. 2004c), but in the last case, this is equivalent to a considerable APD value of 0.85 kPa. Therefore, container design involving many small openings uniformly distributed on the package surface would be preferable for a more efficient cooling process. This is contrary to Kader’s (2002) recommendation of a few larger holes.

The optimum airflow value was not verified for corner holes at very high respiration likely because it

92

did not occur within the range of the airflow rates studied. Table 2 shows that with the peripheral configuration no significant effect has been identified for 0.5 and 1 L•s-1•kg-1airflow rate on the EAR, however the difference between these values decreased as the respiration activity increased. Thus this outcome could suggest that EAR tended to decline beyond 0.5 L•s-1•kg-1, reaching a minimum point at some value around 1 L•s-1•kg-1 and then rising until reaching 2 L•s-

1•kg-1. As aforementioned, decreasing the opening area and

the produce respiration activity lowered the optimum airflow rate required to balance respiration and ventilation heat. Thus, the inflexion point likely occurs at a lower airflow rate than the minimum value studied (0.5 L•s-1•kg-1) for low respiration activity produce. The smallest opened package (2%) had pressure drop as the limiting factor in the cooling process and, although four-0.5% holes form the same total vented area, the effect of the holes positioning on cooling heterogeneity likely added a further restriction, reducing even more the optimum airflow rate. In this case, the selection of airflow to maximize the energy efficiency should be carefully analyzed since long cooling time can also compromise the produce quality by modifying the gas atmosphere. Even low respiration produce, such as pear, can be harmed if the package opening and airflow rate are not sufficiently high to dissipate the gases released during respiration (Faubion and Kader, 1997). These authors claimed that although accumulation of carbon dioxide can reduce the sensitivity to ethylene, concentrations more than 10% could cause critical internal carbon dioxide injuries.

CONCLUSION The system efficiency coefficient EAR was demonstrated as a functional tool during container design, since it considers the peculiarities of the forced air cooling system and produce physiology. The coefficient rapidly decreased as the opening area was gradually increased from 2% to 16% and continued decreasing, although only slightly until reaching a fully open condition. Therefore, the results pointed to an opening area between 8 and 16% for energy use optimization. Further investigation would be necessary to identify a specific value within this range, however container design, manufacturing and raw material costs, and many other parameters should be considered in determining the best opening configuration.

The optimum airflow rate for the cooling process, however, was closely dependent on the produce respiration rate and the container opening area investigated. A higher airflow rate is required for produce with high respiration activity in a container with a larger opening area to enhance the cooling time

and balance the heat produced by respiration and ventilation.

Since decreasing the opening area restricts air circulation, the pressure drop becomes the limiting factor and the optimum airflow rate is reduced. In addition, if this opening area is formed by non-uniformly distributed holes, any increase in airflow will generate an even higher pressure drop and air distribution heterogeneity, which also increases the cooling time and the total energy to be removed by the refrigeration system. Therefore, for the non-uniform configuration, only a very low airflow would be able to offset ventilation and respiration energies with a slight improvement of cooling time. In this case, optimization of energy consumption should be carefully analyzed not to generate produce quality deterioration.

By comparing the results of four 0.5%-holes distributed on corners to 8% uniformly distributed openings on the package surface, it could be concluded that the 0.5%-holes distributed on corners should generally be avoided. The reason is that the four 0.5% hole configuration generates from 1.75 up to 34% more energy to be removed by the cooling system when using low airflow rate (0.5 L•s-1•kg-1) with high respiration activity produce, and high airflow rate (2 L•s-1•kg-1) with low respiration activity produce, respectively.

ACKNOWLEDGEMENT This project was accomplished with the financial support from the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), and the Horticultural Research and Development Centre of Agriculture and Agri-Food Canada.

REFERENCES Arifin, B.B. and K.V Chau,. 1988. Cooling of

strawberries in cartons with new vent hole designs. ASHRAE Transactions, 94(1):1415-1426.

ASHRAE. 2002. Refrigeration. American Society of Heating, Refrigerating and Air Conditioning Engineers, Inc. Atlanta, Georgia.

ASHRAE. 2000. Systems and Equipments. American Society of Heating, Refrigerating and Air Conditioning Engineers, Inc. Atlanta, Georgia.

ASHRAE. 2001. Fundamentals Handbook. Atlanta, Ga. American Society of Heating, Refrigerating and Air-Conditioning Engineers,Inc. Atlanta, Georgia.

Baird, C. D., J.J. Gaffney and M.T Talbot. 1988. Design criteria for efficient and cost effective forced-air cooling systems for fruits and vegetables. ASHRAE Transactions, 94: 1434-1453.

93

Brooker, D.B., F.W. Bakker-Arkema and C.W. Hall. 1974. Drying cereal grains. The AVI Publishing Company, Inc. Westport, Co, USA. 265pp.

Castro, L.R., C.Vigneault and L.A.B. Cortez. 2004a. Container opening design for horticultural produce cooling efficiency. International Journal of Food, Agriculture and Environment. 2 (1): 135-140.

Castro, L.R., C. Vigneault and L.A.B. Cortez. 2004b. Effect of peripheral openings on cooling efficiency of horticultural produce. ASAE Annual International Meeting Ottawa, Ontario. 1 - 4 August. ASAE paper No. 04-6110. 11pp.

Chau, K.V., J.J. Gaffney, C.D. Baird and G.A. Church. 1985. Resistance to airflow of oranges in bulk and in cartons. Transactions of the ASAE 8(6): 2083-2088.

Edeogu, I., J. Feddes and J. Leonard. 1997. Comparison between vertical and horizontal airflow for fruit and vegetable precooling. Canadian Agricultural Engineering, 39(2): 107-112.

Faubion, D.F. and A.A. Kader. 1997. Influence of place packing or tray packing on the cooling rate of palletized ´Anjou´ pears. HortTechnology 7(4):378-382.

Goyette, B., C. Vigneault, B. Panneton and G.S.V. Raghavan. 1996. Method to evaluate the average temperature at the surface of a horticultural crop. Canadian Agricultural Engineering. 38(4): 291-295.

Haas, E., G. Felsenstein, A. Shitzer and G. Manor. 1976. Factors affecting resistance to airflow through packed fresh fruit. ASHRAE Transactions, 82(2): 548-554.

Kader A.A. (ed) 2002. Postharvest technology of horticultural crops. 3rd edition. Cooperative Extension of University of California. Division of Agriculture and Natural Resources. University of California, Davis, CA. Publ. no. 3311.

SPSS Inc. 2004. Chicago, Illinois. USA. http://www.spss.com

Thompson, J.F. and Y.L. Chen. 1988. Comparative energy use of vacuum, hydro, and forced air coolers for fruits and vegetables. ASHRAE Transactions, 94(1):1427-1432.

Vigneault, C. and L.R. de Castro. 2004. Indirect airflow distribution measurement for horticultural crop package, Part I: Development of the research tool. Transactions of the ASAE. (In press)

Vigneault, C., L.R. de Castro and G. Gautron. 2004a. Effect of the presence of openings as container handles on cooling efficiency of horticultural produce. ASAE Annual International Meeting

Ottawa, Ontario. 1 - 4 August. ASAE paper No. 04-6105. 8pp.

Vigneault C., L. R. de Castro, and L.A.B. Cortez. 2004b. A new approach to measure air distribution through horticultural crop packages. Acta Horticulturae. (In press)

Vigneault, C. and B. Goyette. 2002. Design of plastic container openings to optimize forced-air precooling of fruits and vegetables. Applied Engineering in Agriculture. 18(1) :73-76.

Vigneault, C. and B.Goyette, 2001. Loss of ice through container openings during liquid-ice cooling of horticultural crops. Canadian Agricultural Engineering. 43: 3.45-3.48.

Vigneault, C., N.R. Markarian, A. da Silva and B. Goyette. 2004c. Pressure drop during forced-air circulation of various horticultural produce. Transactions of the ASAE. 47 (3): 807-814.

LIST OF TABLES: Table 1. Energy added ratio (EAR, %) for each combination of opening area (OA, %), airflow rate (Dair, L•s-1•kg-1) and respiration activity (L = low; M = moderate; H = high; VH = very high). Table 2. Energy added ratio (EAR, %) of containers with 2% opening area as a function of airflow rate (Dair, L•s-1•kg-1) and respiration activity (L = low; M = moderate; H = high; VH = very high).

LIST OF FIGURES: Figure 1. Experimental set up with forced air tunnel, ball matrix, fan, and dynamic and static pressure measuring devices. Figure 2. Opening configurations studied: 4-0.5%-holes distributed in corners (A), and container with handles and 2% (B) and 16% (C) opening areas. Figure 3. Energy added ratio evolution according to opening area at each airflow rate for A) low and B) very high respiration produce, respectively. Figure 4. Energy efficiency coefficient evolution according to airflow for A) low and B) very high respiration rate produce. Figure 5. Optimal airflow rate evolution according to respiration produce for opening percentages of 2, 4, 8, 16 and 100%

94

Table 1. Energy added ratio (EAR, %) for each combination of opening area (OA, %), airflow rate (Dair, L•s-

1•kg-1) and respiration activity (L = low; M = moderate; H = high; VH = very high). Respiration activity OA

(%) Dair

(L•s-1•kg-1) L M H VH 0.5 1.158abcdef 2.968lmnop 5.805uvx 13.432ε 1 2.387ghijklmn

3.582opq 5.045stu 9.104βθ 2 2 8.045α 9.363βθ 9.656θ 11.208δ

0.5 0.794abcd 2.351ghijklm

n4.709rst 11.033δ

1 1.004abcde 2.064fghijkl

m3.358nopq 6.946z 4

2 3.008mnop 4.329qrs 4.665rst 6.331vxz 0.5 0.582abc 1.891efghijk 3.705pqr 8.544αβ 1 0.538abc 1.447bcdefgh 2.476hijklmn 5.329stuv 8 2 1.069abcdef 2.234ghijklm 2.475hijklmn 3.802pqr

0.5 0.578abc 1.921efghijkl 3.824pqr 8.906βθ 1 0.432ab 1.398bcdefg 2.530ijklmn 5.667tuvx

Con

tain

er w

ith h

andl

es

16 2 0.426ab 1.506cdefghi 1.695defghij 2.835klmnop

0.5 0.434ab 1.486cdefghi 2.860klmnop 6.512xz 1 0.279a 1.026abcdef 1.774defghij 3.858pqr Fully open 2 0.271a 1.343bcdefg 1.509cdefghi 2.591jklmno

Means followed by the same Arabic or Greek letter are not significantly different based on Tukey Test using α = 0.05. Table 2. Energy added ratio (EAR, %) of containers with 2% opening area as a function of airflow rate (Dair, L•s-1•kg-1) and respiration activity (L = low; M = moderate; H = high; VH = very high).

Respiration activity Position of opening Dair (L•s-1•kg-1) L M H VH

0.5 2.347ab 4.299abcd 7.203bcdef 14.967h

1 7.354bcdef 8.486cdefg 9.727efg 11.208gh On corners 2 36.940i 38.873i 39.324i 41.650i

0.5 1.158a 2.968ab 5.805abcde 13.432gh

1 2.387ab 3.582abc 5.045abcde 9.104defg Uniformly distributed

2 8.045 cdef 9.363efg 9.656efg 11.208fgh

Means followed by the same letter are not significantly different based on Tukey Test using α = 0.05.

95

pitot tube

outlet tube

fan

air flow

air flow

mobile plastic tunnel

mobile fan set up

mobile heat exchanger

aluminum plate

ball matrix

static pressuremeasuring device plywood box

air flow

plastic wrappolystyrene foam

Figure 1. Experimental set up with forced air tunnel, matrix, fan, and dynamic and static pressure measuring devices.

18,2

419,1

419,

1

11,1

111,1

25,4

28,6

108

4,8 18,2

419,1

39,7

419,

125,4

111,1

69,9

4,852,1

Ø33,4

419,

1

314,

9

28,6

R12,7

R12,7

a) b) c)

Figure 2. Opening configurations studied: 4-0.5%-holes distributed in corners (A), and container with handles and 2% (B) and 16% (C) opening areas.

A) B) C)

96

A) B)

y = 14.950x-0.2147 R2 = 0.8436

y = 16.751x-0.6686 R2 = 0.9813

y = 10.079x-0.2434 R2 = 0.8130

0

5

10

15

0 5 10 15 20Opening area (%)

0.512

EA

R (%

)

Figure 3. Energy added ratio evolution according to opening area at each airflow rate for A) low and B) very high respiration produces, respectively. A) B)

Figure 4. Energy efficiency coefficient evolution according to airflow for A) low and B) very high respiration rate produce.

y = 21.37x-1.4208 R2 = 0.9995

y = 3.6372x-0.8297 R2 = 0.9396

y = 1.358x-0.3462 R2 = 0.8853

0

5

10

15

0 5 10 15 20Opening area (%)

0.512

EA

R (%

)

y = 2.135x2 - 0.747x + 0,999 y = 1.056x2 - 1,164x + 1.111 y = 0.413x2 - 0.708x + 0.832

y = 0.190x2 - 0.576x + 0.818y = 0.202x2 - 0.613x + 0.690

0

5

10

15

0.5 1.0 1.5 2.0Airf low rate (L•s-1•kg-1)

24816100

EAR

(%)

y = 7.173x2 - 19.417x + 21.347y = 5.040x2 - 15.736x + 17.641y = 3.269x2 - 11.334x + 13.394y = 2.431x2 - 10.126x + 13.361y = 2.694x2 - 9.349x + 10.513

0

5

10

15

0.5 1.0 1.5 2.0

Airf low rate (L•s-1•kg-1)

24816100

EAR

(%)

97

y = -2E-6x2 + 0.003x + 0.0232 R2 = 0.998

y = -2E-6x2 + 0.0029x + 0.407 R2 = 0.998

y = -2E-6x2 + 0.0026x + 0.713 R2 = 0.999

y = -1.5E-6x2 + 0.0023x + 0.944 R2 = 0.999

y = -2.1E-6x2 + 0.0032x + 0.951 R2 = 0.993

0

1

2

0 200 400 600 800 1000Respiration rate (mW•kg-1)

Opt

imal

airf

low

rate

val

ues

(L•s-1

•kg-1

)

2

4

8

16

100

Figure 5. Optimal airflow rate evolution according to respiration produce for opening percentages of 2, 4, 8, 16 and 100%.

98

5. DISCUSSÃO DOS RESULTADOS

Apesar do erro experimental expressivo observado nos experimentos iniciais

realizados com esferas preenchidas com ágar-ágar pôde-se perceber a forte correlação não

linear existente entre o fluxo de ar e o tempo de meio resfriamento. Elevando-se o fluxo reduz-

se o tempo necessário para se resfriar o produto e a uniformidade do processo, sobretudo nas

menores faixas (< 2 L.s-1.kg-1). A área da abertura também mostrou exercer efeito na taxa de

resfriamento do produto nessa faixa de fluxo de ar, no entanto de menor significância. Assim,

o aumento do fluxo poderia compensar o efeito negativo gerado por embalagens com

pequenas porcentagens de área de abertura. Os resultados apontaram valores medianos de

fluxo de ar (1 a 2 L.s-1.kg-1) e área de abertura de 14% na embalagem como mais adequados

para que a pressão gerada pelo ventilador não encareça os custos energéticos da operação.

Desde que a resistência estrutural da embalagem não seja comprometida, tal porcentagem é

recomendada por promover um resfriamento praticamente tão rápido e uniforme quanto a

situação de produtos dispostos a granel (sem acondicionamento em embalagens), e gerar

valores de queda da pressão do ar aceitáveis.

Nesses experimentos iniciais não se observou efeito significativo da posição do

orifício da embalagem segundo o eixo Y (altura). A influência da localização do produto no

sentido da profundidade da embalagem foi detectada apenas para fluxos de ar mais baixos.

Tais resultados podem ser explicados sobretudo pelo arranjo experimental escolhido, que

adotou pequena distância na altura dos orifícios da embalagem às esferas assim como utilizou

simuladores ocos que evitaram o aquecimento do ar ao atravessar o produto. O resfriamento

ocorreu mais rapidamente nas últimas que antepenúltimas camadas de produto, em função da

proximidade daquelas com as aberturas da embalagem e maiores velocidades do ar de saída.

Apesar da área total de abertura da embalagem produzir efeito significativo na

eficiência do processo de resfriamento, a área individual de cada orifício não influenciou a

queda da pressão do ar ao atravessar o produto acondicionado.

Esta metodologia utilizada para projeto de embalagens foi então aperfeiçoada durante

o andamento da pesquisa, sendo modificados tanto produto-modelo como também arranjo

experimental. Assim, conseguiu-se reduzir o erro experimental obtido nos primeiros testes ao

se substituir as esferas plásticas ocas por esferas sólidas impermeáveis. Como estas novas

99

esferas foram previamente calibradas para seleção de um grupo com características térmicas

similares, pôde-se atribuir os resultados gerados, principalmente, à taxa de fluxo de ar e

configuração de embalagem testados e não mais à heterogeneidade dos simuladores. O novo

material usado para representar hortícolas esféricos possibilitou o estabelecimento de uma

correlação significativa entre seu coeficiente de resfriamento sob ar forçado, a velocidade da

aproximação do ar e a taxa de resfriamento de cada esfera submetida ao tratamento com água

gelada. No entanto, apesar do novo simulador ser mais uniforme e estável, o aparato

experimental utilizado para a obtenção da correlação mencionada apresentou bastante

limitação e imprecisão nos testes seguintes, principalmente para a faixa laminar de fluxo de ar.

Até então, a metodologia baseava-se na velocidade de aproximação do ar em uma área

circular, visto que cada esfera foi inserida em tubo e submetida individualmente a diversos

fluxos de ar. Desta forma, os valores obtidos para a taxa de resfriamento refletiram o valor

superestimado da superfície da esfera exposta ao ar frio. Na situação real essa área seria

limitada pelos pontos de contato entre produtos e destes com as paredes da caixa, reduzindo a

área exposta ao ar frio e portanto os coeficiente de resfriamento encontrados. Além disso, os

resultados também não incluem o efeito de aquecimento do ar que ocorre quando atravessa

camadas de produto “quente” acondicionados em embalagem, já que uma única esfera foi

submetida ao resfriamento rápido. Portanto, a metodologia foi aperfeiçoada mantendo-se os

simuladores, que geraram resultados satisfatórios, mas adotando-se um novo aparato e

procedimento experimental. Desta vez, o empilhamento colunar das esferas, para simulação de

um leito de produto acondicionado em embalagem, permitiu o desenvolvimento de uma

ferramenta mais precisa e confiável. Esse novo aparato foi construído visando corrigir os

inconvenientes encontrados no primeiro experimento (com esferas de ágar-ágar) através,

inicialmente, do aumento do número de camadas de produto na direção Y-altura da

embalagem. Isso permitiu investigar o efeito do posicionamento dos orifícios nessa direção

sobre o resfriamento das esferas. Também não foi necessário utilizar uma grade metálica para

retenção e simulação da pilha de produto, removendo portanto a resistência oferecida à

passagem do ar e garantindo área de abertura de 100% quando testado o produto a granel

(“fully open”).

Assim, a distribuição de 64 esferas instrumentadas em uma pilha de simuladores

submetidas a uma configuração de 100% de área aberta e no total quatorze fluxos de ar (oito

100

deles no experimento cujo resumo é mostrado nos apêndices – Artigo 9), permitiu o

estabelecimento de novas equações da velocidade da aproximação do ar para cada um desses

simuladores. Neste caso então, o método se baseou na área da seção transversal quadrada da

embalagem (SCSV) e considerou a variabilidade do fluxo de ar nas diferentes posições dentro

do meio poroso. Isto porque as esferas instrumentadas representaram melhor um leito de

produto acondicionado em embalagem, gerando resultados que incluíram a variação da

velocidade de aproximação do ar ao redor de cada uma delas (em função da posição, pontos de

contato e obstáculo à circulação do ar) e o aquecimento ao longo das camadas de produto

“quente” (ar à diferentes temperaturas ao atingir o produto). Além disso, toda a estrutura física

montada e equipamentos e softwares utilizados se mostraram bastante estáveis, possibilitando

que os experimentos fossem executados e repetidos de maneira muito prática, confiável e

precisa.

Através do estabelecimento desta “ferramenta de pesquisa”, que inclui o novo aparato

experimental e as correlações entre a velocidade do ar e a taxa de resfriamento dos

simuladores nas diferentes posições dentro do leito, foi possível investigar o efeito de diversas

configurações de aberturas de embalagens na eficiência do resfriamento rápido de produtos

hortícolas esféricos. Para cada configuração estudada, pôde-se extrair várias conclusões a

respeito da variação da taxa de resfriamento e distribuição do ar em função do fluxo aplicado

combinado à área aberta, posição e tamanho dos orifícios da embalagem. A aplicação da

metodologia desenvolvida possibilitou a determinação indireta do perfil de velocidade do ar, a

partir dos resultados de tempo ou taxa de resfriamento do produto. A medição direta dessas

variáveis com os instrumentos atualmente disponíveis no mercado seria extremamente difícil e

pouco precisa.

Assim, pôde-se verificar que, em geral, a velocidade do ar se revelou mais intensa na

direção dos orifícios da embalagem e menor nas camadas de produto próximas às paredes,

variando portanto em função da posição da abertura testada. A velocidade também foi superior

quando o número de orifícios foi reduzido, devido à convergência do fluxo de ar em um

número menor de áreas abertas disponível. Quanto maior a área total aberta da embalagem,

maior a uniformidade e velocidade de distribuição do ar e menor o tempo de resfriamento do

produto e a queda da pressão do ar ao atravessar o leito.

101

Observou-se também que, quando a superfície disponível aos orifícios da embalagem

for restrita, impedindo sua distribuição de maneira uniforme, é recomendável utilizar posição

central à periférica. Poucos furos distribuídos nos quatro cantos da parede da embalagem

reduziram a uniformidade da circulação de ar através da embalagem, embora não tenham

afetado a velocidade média e queda da pressão do ar. Um maior número de orifícios na

embalagem mostrou reduzir a velocidade e a pressão média do ar e também a heterogeneidade

de distribuição do ar quando combinado a orifícios melhor distribuídos na superfície da

embalagem (no caso, verificou-se tal comportamento para orifícios periféricos no alto e

embaixo). Dependendo da área total testada, o aumento da uniformidade obtido através da

distribuição uniforme de um número maior de orifícios permitiu compensar a menor

velocidade do ar, não afetando então o tempo de resfriamento do produto. Mas para menores

áreas totais de abertura (<=4%), o efeito da velocidade média reduzida prevaleceu provocando

o aumento do tempo de processo.

Um ponto importante a ser ressaltado é o efeito da correlação entre as variáveis

independentes (área total, área e posição do orifício, e fluxo de ar) nos resultados observados

para as variáveis dependentes (velocidade, uniformidade, pressão do ar, tempo de

resfriamento). Por exemplo, a área individual de cada orifício exerceu influência significativa

na eficiência do processo de resfriamento rápido, sobretudo por estar relacionada à área total

aberta da embalagem, e portanto, seguindo mesmo comportamento que aquele observado para

esta última. Sendo assim, mantendo-se um mesmo número de orifícios mas aumentando-se a

área individual de cada um deles poderá auxiliar na otimização do resfriamento do produto

visto que a área total aberta será também elevada. Do mesmo modo que, quando foram

analisadas configurações com a mesma área total aberta, aquela variável não influenciou os

valores de pressão do ar, conforme observado desde a pesquisa com esferas de ágar-ágar. Isto

indica, então, que os benefícios obtidos pela elevação da área individual de cada orifício não

deverão ser observados se a área total for reduzida, através por exemplo, da diminuição do

número de furos.

Além da correlação entre área dos orifícios e área total aberta da embalagem, notou-

se que os parâmetros estudados para embalagens também estiveram fortemente relacionados

com o fluxo de ar aplicado para a determinação da eficiênca do processo. Em geral, a elevação

do fluxo do ar acelerou a taxa de resfriamento do produto, aumentando a velocidade e

102

uniformidade da distribuição do ar, mas com um efeito negativo para a pressão exigida do

ventilador. Mas além do fluxo, o efeito da força gravitacional gerado pela diferença de

densidade entre o ar frio de resfriamento e o ar quente dentro da embalagem também foi

relevante na distribuição do ar através do produto acondicionado. Desta forma este efeito

também deverá ser previsto e considerado no projeto da embalagem.

Os experimentos para se avaliar o efeito da força gravitacional em fluxos de ar

menores (faixa laminar) confirmaram a maior importância da posição dos orifícios em relação

à sua área individual nos resultados da eficiência do processo de resfriamento. Quando a área

disponível às aberturas é restrita, o projetista poderia se beneficiar desse efeito na escolha do

posicionamento dos orifícios da embalagem. Orifícios dispostos de maneira a forçar o ar frio a

entrar na embalagem pela parte inferior e serem succionados pelo ventilador pela região

superior se beneficiaram da gravidade para acelerar e uniformizar a distribuição do ar a baixos

fluxos de ar. No entanto, para fluxos turbulentos, os orifícios fornecem melhor resultado

quando forçam o ar à atravessar a embalagem no sentido da região superior do leito à inferior.

Neste caso, o ar frio ao entrar na embalagem e atingir o ar “intersticial” (dentro dos poros,

entre os produtos) mais aquecido, tende a fluir rapidamente para a região mais baixa da

embalagem devido à diferença de densidades. Como o fluxo é alto, o ar então tende a se

distribuir mais uniformemente por toda a altura da camada de produto. Por outro lado quando

o fluxo é reduzido, o ar ao atingir o produto mais aquecido tem sua temperatura elevada e por

isso tende a “subir” e atuar em conjunto ou no sentido da sucção propiciada pelo ventilador.

Isto explica o fato de embalagens com orifícios que forçam o ar de baixo para cima serem

mais indicadas para aumentar a eficiência do processo de resfriamento sob condições de

menores fluxos de ar.

A área total da embalagem foi analisada ainda considerando-se seu efeito no processo

de resfriamento do produto no caso de projetos que incluem aberturas para facilitar o

manuseio. Nessa situação, se a estrutura da embalagem restringir a área total dos orifícios

(máximo 8%), o fluxo de ar deverá ser maior para vencer a resistência fornecida pelas

aberturas. Isso evitaria a formação de um caminho preferencial no sentido da abertura de

manuseio e portanto, a heterogeneidade da distribuição do ar pelo leito de produto

acondicionado. No entanto, essa situação desfavorável energeticamente quanto ao

requerimento de pressão maior do ventilador para fornecer fluxos mais altos, poderia ser

103

revertida caso a abertura da embalagem fosse elevada (16%). Nesse caso, sob fluxos

moderados (0.5 a 1 L.s-1.kg-1) a resistência imposta pelos seus orifícios equivale relativamente

àquela das alças, assim o ar tende a entrar e distribuir na embalagem de maneira mais

uniforme por todas as aberturas. Portanto, é recomendável que o técnico ou engenheiro inclua

aberturas de manuseio no projeto apenas no caso de não haver restrição estrutural da

embalagem para evitar o comprometimento do processo (seja prejudicando a qualidade do

produto, seja encarecendo consideravelmente os custos de operação do sistema).

Todos os resultados foram confirmados pela distribuição da velocidade do ar no leito

de produto, em função da posição dos orifícios analisados em relação à direção do fluxo de ar.

Para embalagens com aberturas tipo “alça” por exemplo, para menores valores de área aberta e

fluxo de ar, a velocidade tendeu a ser maior na camada superior (Y=5, 6) da embalagem, o que

comprovaria a formação do caminho preferencial anteriormente discutido. No entanto,

observou-se um inconveniente do aparato experimental montado. Como as esferas foram

distribuídas apenas do lado esquerdo para a parte inferior da pilha, e somente no lado direito

para a parte superior da caixa (em relação ao sentido do fluxo do ar), não foi possível se

verificar de maneira adequada a influência da abertura em relação à posição da esfera segundo

o eixo X (largura da embalagem).

Finalmente, realizou-se uma análise energética com cinco das configurações de

embalagens estudadas na pesquisa, incluindo-se o caso mais crítico observado para

heterogeneidade de distribuição do ar (quatro orifícios de 0,5% de área cada posicionados nos

cantos) e um dos mais favoráveis (embalagem de 16% de área de abertura com orifícios

uniformemente distribuídos e tipo “alça”). Esta análise permitiu estabelecer a influência da

área de abertura da embalagem na eficiência do resfriamento rápido em função da energia

adicionada pela respiração do hortícola e pelo ventilador em operação. Para todos os tipos de

produto investigados, embalagens com alças e 8% de área aberta em geral geraram um

processo tão eficiente quanto embalagens de mesma configuração mas área de 16%. No

entanto, quando o fluxo de ar é elevado para 2 L.s-1.kg-1 a energia consumida com a

embalagem mais aberta atinge valores inferiores devido à menor restrição e heterogeneidade

da circulação do ar. Neste caso, os valores maiores de velocidade observados nos orifícios da

embalagem de 16% contribuíram para reduzir o consumo energético a um nível comparável

àquele obtido com o produto resfriado a granel (100% de área).

104

Quanto menor a atividade respiratória do produto, menor será o fluxo necessário para

otimizar a eficiência do processo. Por exemplo, 1 L.s-1.kg-1 já seriam suficientes para agilizar o

resfriamento do produto de baixa taxa respiratória acondicionado em embalagem com abertura

igual ou maior que 16%. No entanto, para embalagens menos abertas, quando o fluxo de ar é

elevado, os benefícios obtidos na redução do tempo do resfriamento não compensam o

incremento da potência requerida do ventilador.

Assim, um ponto ótimo para operação do ventilador usado no resfriamento rápido

deverá ser buscado para minimização do requerimento energético do sistema. Este valor

deverá balancear a energia gerada da respiração do vegetal, influenciada pela taxa de

resfriamento, com aquela proveniente do funcionamento do ventilador, que também é afetada

pela pressão exigida. Quanto mais lentamente o produto respirar, e menor a área aberta da

embalagem, maior o efeito da pressão do ar na limitação do fluxo. Da mesma forma que

produtos que respiram mais rapidamente, como brócolis, acondicionados embalagem que

fornecem menor restrição à passagem do ar, tendem à exigir fluxos maiores de ar para agilizar

o resfriamento e controlar a energia adicionada ao sistema pela atividade respiratória.

A posição dos orifícios da embalagem também exerce influência significativa na

eficiência energética do sistema. Quando eles não são uniformemente distribuídos, o aumento

do fluxo de ar provoca maior heterogeneidade do processo de resfriamento. Nesse caso são

exigidas maiores velocidades do ar para melhorar a taxa de resfriamento, mas às custas de

maior potência requerida do ventilador. O fluxo de ar ótimo então para minimizar o consumo

energético no caso de orifícios distribuídos perifericamente, deveria ser limitado para

compensar a maior restrição e heterogeneidade da distribuição do ar, o que poderia

comprometer a qualidade final do produto resfriado.

Portanto, recomenda-se o projeto de um maior número de orifícios uniformemente

distribuídos nas paredes da embalagem, capazes de fornecer uma área total aberta entre 8 e

16%, combinados a um fluxo de ar ótimo em função da atividade respiratória do produto.

Hortícolas com alta taxa respiratória, como brócolis, se acondicionados em caixas com área

igual a 8 e 16% deveriam ser submetidos a fluxos na ordem de 1,73 e 2,08 L.s-1.kg-1,

respectivamente, para minimização dos custos energéticos. Já para vegetais de taxa

respiratória moderada, como alface, tais valores seriam reduzidos para 1,15 e 1,52 L.s-1.kg-1

105

para 8 e 16%, respectivamente, por exemplo, já que a energia adicionada pela respiração do

vegetal devido à taxa de resfriamento não tem impacto tão forte na eficiência do sistema.

106

6. CONCLUSÕES GERAIS

Tendo em vista os objetivos propostos nesta pesquisa, pôde-se extrair as principais

conclusões:

• Dentre todos materiais testados, esferas plásticas sólidas impermeáveis, calibradas e

empilhadas em arranjo colunar permitiram simular de maneira mais precisa a

distribuição do ar durante o resfriamento rápido a ar forçado de um leito de produtos

hortícolas acondicionados em embalagem;

• O perfil de velocidade do ar no meio poroso foi determinado através do

desenvolvimento de uma correlação entre o fluxo de ar e o coeficiente de resfriamento

das esferas utilizadas como produto-modelo;

• Essas esferas não substituem os produtos hortícolas com relação às propriedades

intrínsecas de cada vegetal, como a atividade respiratória e transpiração. No entanto a

aplicação da correlação desenvolvida usando o produto-modelo possibilitou a análise

precisa do efeito de diferentes configurações de aberturas de embalagem e fluxos de ar

na eficiência do resfriamento do produto, revelando, por sua vez, as seguintes

conclusões:

o A área, número e posição dos orifícios e sobretudo o fluxo de ar aplicado

tiveram efeito significativo na eficiência do processo de resfriamento rápido do

produto;

o Em geral, o aumento do fluxo de ar gera maior velocidade, homogeneidade e

restrição à passagem do ar. Sendo assim, maior fluxo combinado à maior área

aberta da embalagem, que fornece menor resistência, seriam vantajosos para

acelerar o resfriamento do produto, sobretudo aqueles de alta taxa respiratória.

Se o produto estiver acondicionado em caixas com orifícios distribuídos de

maneira não uniforme (partes superior e inferior das laterais), a elevação do

fluxo de ar acarreta maior heterogeneidade da taxa de resfriamento dentro da

embalagem. Produtos hortícolas em caixas com tais configurações de abertura e

submetidos a fluxo laminar se beneficiam mais da convecção natural, que

ameniza a heterogeneidade do processo;

107

o Quanto maior a área aberta total da embalagem, melhores serão todos os

parâmetros envolvidos na eficiência do resfriamento. No entanto, esta

contribuição da elevação da área é menos significativa acima de 8%. Por

exemplo, embalagens de orifícios uniformemente distribuídos com 8% de área

total aberta em geral tiveram desempenho equivalente às caixas com 16%. A

utilização de embalagens com esta última porcentagem também reduziu o

consumo energético a níveis comparáveis às totalmente abertas. Até mesmo

com 14% foi possível se obter alta taxa de resfriamento a um nível aceitável de

pressão do ar. Isto porque, o efeito dos orifícios de embalagens na convergência

e intensificação da velocidade do ar, é muitas vezes capaz de balancear a

heterogeneidade e restrição da passagem do ar, sobretudo através de um

número elevado de orifícios uniformemente posicionados;

o O efeito observado para a área individual no resfriamento do produto foi devido

à variação da área total aberta da embalagem. Dessa forma, para embalagens de

mesma área total, a variação da área de cada orifício não influenciou a queda da

pressão do ar no ventilador. Assim, recomenda-se aumentar a área total,

principalmente através da elevação do número de orifícios. Neste caso, a

velocidade média do ar é reduzida, em função da menor convergência do fluxo

de ar, que deverá se distribuir por várias aberturas da embalagem. No entanto,

devido à considerável suavização da heterogeneidade do processo e pressão do

ar, uma embalagem com vários orifícios pequenos se torna uma opção mais

vantajosa;

o Enfim, a seleção da área e disposição das aberturas para o projeto da

embalagem será limitada pelas restrições do material, e o sucesso da operação

de resfriamento dependerá do produto acondicionado e capacidade do sistema

frigorífico. No caso de embalagens de papelão, que restringem a área aberta, o

fluxo de ar poderá ser reduzido se fornecido a frutas e hortaliças de menor

atividade respiratória, para compensar a queda da pressão do ar. Já para

produtos de intensa atividade respiratória, a taxa de resfriamento será decisiva

na manutenção da qualidade do produto, exigindo portanto, um fluxo de ar um

pouco mais elevado para otimizar a eficiência energética do sistema.

108

7. SUGESTÕES PARA TRABALHOS FUTUROS

• Verificar a influência do modo de empilhamento ou arranjo do produto na embalagem

na eficiência do resfriamento rápido;

• Pesquisar o efeito de valores intermediários de área de aberturas, entre 8 e 16%, na

energia adicionada ao sistema de resfriamento, visando encontrar a área máxima de

orifício necessária para otimizar o processo de resfriamento, tendo em vista inclusive a

resistência estrutural da embalagem;

• Estudar a influência da posição e área dos oríficios da embalagem na eficiência dos

demais processos de resfriamento rápido de produtos hortícolas, como com água ou

gelo, visando o projeto de uma embalagem com configuração de aberturas padronizada

que atenda a uma maior gama de frutas e hortaliças.

109

REFERÊNCIAS BIBLIOGRÁFICAS

ALVAREZ, G.;TRYSTRAM, G. Design of a new strategy for the control of the refrigeration

process: fruit and vegetables conditioned in a pallet. Food Control, London, v. 6, n. 6,

p.347-355, 1995.

ALVAREZ, G.; FLICK, D. Analysis of heterogeneous cooling of agricultural products inside

bins. Part I: Aerodynamic study. Journal of Food Engineering, Dublin 4, Irlanda, v. 39,

p. 227-237, 1999.

ALVAREZ, G.; FLICK, D. Analysis of heterogeneous cooling of agricultural products inside

bins. Part II: Thermal study. Journal of Food Engineering, Dublin 4, Irlanda, v. 39, p.

239-245, 1999.

ANSARI, F.A.; AFAQ, A. New method of measuring thermal diffusivity of spherical produce.

International Journal of Refrigeration, New York, v. 9, p. 158-160, 1986.

ARIFIN, B.B; CHAU, K.V. Cooling of strawberries in cartons with new vent hole designs.

ASHRAE Transactions, Atlanta, v. 94, n. 1, p. 1415-1426, 1988.

ARIFIN, B.B.; CHAU, K.V. Forced air cooling of strawberries. In: SUMMER MEETING OF

AMERICAN SOCIETY OF AGRICULTURAL ENGINEERS, 1987, Baltimore.

Proceedings… St Joseph: ASAE, 1987. n. 87-6004.

ASHRAE. Refrigeration. Atlanta, Georgia: American Society of Heating, Refrigerating and

Air Conditioning Engineers, Inc., 2002.

ASHRAE. Fundamentals Handbook. Atlanta, Georgia: American Society of Heating,

Refrigerating and Air Conditioning Engineers, Inc., 2001.

ASHRAE. Systems and Equipment. Atlanta, Georgia: American Society of Heating,

Refrigerating and Air Conditioning Engineers, Inc., 2000

BAIRD, C.D.; GAFFNEY, J.J.; TALBOT, M.T. Design criteria for efficient and cost effective

forced-air cooling systems for fruits and vegetables. ASHRAE Transactions, Atlanta, v.

94, p. 1434-1453, 1988.

BARTSCH, J.A.; MCLAUGHLING, N.B.; PITT, R.E. A computerized control and data

acquisition system for a universal testing machine. Journal of Texture Studies,

Auckland, New Zealand, v. 17, p. 315-330, 1987.

110

BORDIN, M.R. Embalagem para frutas e hortaliças. In: CURSO DE ATUALIZAÇÃO EM

TECNOLOGIA DE RESFRIAMENTO DE FRUTAS E HORTALIÇAS, 2, 1998,

Campinas. Anais... Campinas: Faculdade de Engenharia Agrícola, UNICAMP, 1998. p.19-

27.

BOYETTE, M.D. Forced-air cooling packaged blueberries. Applied Engineering in

Agriculture, St Joseph, MI, USA. v. 12, n. 2, p. 213-217, 1996.

BROOKER, D.B.; BAKKER-ARKEMA, F.W.; HALL, C.W. Drying cereal grains.

Westport, CO, USA: The AVI Publishing Company, Inc. 1974. 265 p.

CASTRO, L.R.; VIGNEAULT, C.; CORTEZ, L.A.B. Effect of container area and position on

horticultural crop half-cooling time. In: CSAE/SCGR 2003 MEETING, 2003, Montreal,

QC, Canada. Proceedings… Winnipeg: CSAE, 2003. Paper n. 03-309.

______. Container opening design for horticultural produce cooling efficiency. International

Journal of Food, Agriculture and Environment, Helsinki, Finlândia, v. 2, n. 1, p. 135-

140, 2004.

______. Effect of container opening on air distribution during precooling of horticultural

produce. Transactions. of the ASAE. 2004. (no prelo)

______. Cooling performance of horticultural produce in containers with peripheral openings.

Postharvest Biology and Technology. 2004. (submetido)

______. Effect of container openings and airflow rate on energy required for forced-air

cooling of horticultural produce. Canadian Agricultural Engineering. 2004. (no prelo)

CARVALHO FILHO, C.D.; HONÓRIO, S.L.; VIGNEAULT, C.; RAGHAVAN, G.S.V.;

PRANGE, R.K. Técnicas de armazenagem com a atmosfera modificada e controlada. In:

CORTEZ, L.A.B.; HONÓRIO, S.L.; MORETTI, C.L. (Eds.) Resfriamento de frutas e

hortaliças. Brasília, DF, Brasil: Embrapa Informação Tecnológica, 2002, p. 385-409.

CHAU, K.V.; GAFFNEY, J.J.; BAIRD, C.D.; CHURCH, G.A. Resistance to air flow of

oranges in bulk and in cartoons. Transactions of the ASAE, St Joseph, MI, USA. v. 8, n.

6, p. 2083-2088, 1985.

CORTEZ, L.A.B.; CASTRO, L.R.; VIGNEAULT, C. Resfriamento rápido a ar : métodos da

câmara frigorífica e do ar forçado. In: CORTEZ, L.A.B.; HONÓRIO, S.L.; MORETTI,

C.L. (Eds.) Resfriamento de frutas e hortaliças. Brasília, DF, Brasil: Embrapa

Informação Tecnológica, 2002a, p. 231-272.

111

CORTEZ, L.A.B.; HONÓRIO, S.L.; MORETTI, C.L. (Eds.) Resfriamento de frutas e

hortaliças. Brasília, DF, Brasil: Embrapa Informação Tecnológica, 2002b. 428 p.

DEELL, J.R.; VIGNEAULT, C.; LEMERRE, S. Water temperature for hydrocooling field

cucumbers in relation to chilling injury during storage. Postharvest biology and

technology, Pullman, v. 18, n. 1, p. 27-32, 2000.

DINCER, I. Precooling of cylindrically shaped grapes: experimental and theorical heat

transfer rates. Journal of food process engineering, Trumbull, n. 17: p. 57-71. 1994.

DINCER, I.; GENCELI, O.F. Cooling process and heat transfer parameters of cylindrical

products cooled both in water and in air. International Journal of Heat and Mass

Transfer, New York, v. 37, n. 4, p. 625-633, 1994.

DUREZ - Thermoset molding composites. 24 Jun. 2003. Produced by Durez Corporation.

Disponível em: <http://www.durez.com/datasheets/31735.pdf>. Acesso em: ago. 2003.

EDEOGU, I.; FEDDES, J.; LEONARD, J. Comparison between vertical and horizontal

airflow for fruit and vegetable precooling. Canadian Agricultural Engineering,

Winnipeg, v. 39, n. 2, p. 107-112, 1997.

ÉMOND, J.P.; MERCIER, F.; SADFA, S.O.; BOURRÉ, M.; GAKWAYA, A. Study of

parameters affecting cooling rate and temperature distribution in forced-air precooling of

strawberry. Transactions of the ASAE, St Joseph, MI, USA. v. 39, n. 6, p. 2185-2191,

1996.

FAO. Production Yearbook. Roma: Food and Agriculture Organization, FAO Statistics

Series n.99, v. 44., 1991.

FAUBION, D.F.; KADER, A.A.; Influence of place packing or tray packing on the cooling

rate of palletized ´Anjou´ pears. HortTechnology, Alexandria, v. 7, n. 4, p. 378-382, 1997.

FLICK, D.; LESLOUS, A.; ALVAREZ, G. Modélisation semi-empirique des écoulements et

des transferts dans un milieu poreux en régime turbulent. Revue Internationale du Froid,

Villeurbanne, v. 26, p. 349-359, 2003.

FONTANA, A.J.; VARITH, J.; IKEDIALA, J.; REYES, J.; WACKER, B. Thermal properties

of selected foods using a dual needle heat-pulse sensor. In: ASAE/CSAE-SCGR 1999

MEETING, 1999, Toronto. Proceedings… St Joseph: ASAE, 1999, 10p. Paper no.

996063.

112

GHAOUTH, A.E.; PONNAMPALAM, R.; CASTAIGNE, F.; ARUL, J. Chitosan coating to

extend the storage life of tomatoes. Hortscience, Alexandria, v. 27, n. 9, p. 1016-1018,

1992.

GORINI, F.; BORINELLI, G. Some trials of salad precooling. Acta Horticulturae, Leuven,

Bélgica, v. 38, p. 465-490, 1974.

GOYETTE, B.; VIGNEAULT, C.; PANNETON, B.; RAGHAVAN, G.S.V. Method to

evaluate the average temperature at the surface of a horticultural crop. Canadian

Agricultural Engineering, Winnipeg, v. 38, n. 4, p. 291-295, 1996. (cópia no Anexo)

HAAS, E.; FELSENSTEIN, G.; SHITZER, A.; MANOR, G. Factors affecting resistance to

airflow through packed fresh fruit. ASHRAE Transactions, Atlanta, v. 82, n. 2, p. 548-

554, 1976.

HACKERT, J.M.; MOREY, R.V.; TOMPSON, D.R. Precooling of fresh market broccoli.

Transactions of the ASAE, St Joseph, MI, USA. v. 24, n. 4, p. 1073-1076, 1987.

HARDENBURG, R.E.; WATADA, A.E.; WANG, C.Y. The commercial storage of fruits,

vegetables and florist and nursery stocks. USDA Agricultural Handbook Number 66.

Washington: Department of Agriculture Research Service, 1986. 136p.

HENRY, F.E; BENNETT, A.H.; SEGALL, R.H. Hydraircooling vegetables in pallet loads.

American Vegetable Grower, Westville, v. 27, n. 10, p. 8-9, 46-47, 1979.

IRVINE, D.A.; JAYAS, D.S.; MAZZA, G. Resistance to airflow through clean and solid

potatoes. Transactions of the ASAE, St Joseph, MI, USA. v. 36, n. 5, p. 1405-1410,

1993.

JAMET, Y. Transfert de chaleur et de masse au cours du refroidissement des denrées

périssables à bord des navires - Durée minimale de refroidissement des cargaisons de fruits

et des légumes dans les cales ou à l’intérieur des conteneurs. Revue International du

Froid, Villeurbanne, v.7, n.5, p. 299-307, 1984.

KADER A.A. (Ed.) Postharvest technology of horticultural crops, 3. ed. Publication n.

3311. Oakland: Cooperative Extension of University of California. Division of Agriculture

and Natural Resources, 2002, 535p.

KOPELMAN, I.; BLAISDELL, J.L.; PFLUG, I.J. Influence of fruit size and coolant velocity

on the cooling of Jonathan Apples in water and air. ASHRAE Transactions, Atlanta, v.

72, n. 1, p. 209-216, 1966.

113

LAMBRINOS, G.; ASSIMAKI, H. Air precooling and hydrocooling of Hayward kiwifruit.

Acta Horticulturae, Leuven, Bélgica, v. 44, n. 2, p. 561-566, 1997.

LEYTE, J.C.; FORNEY, C.F. Optimizing flat design for forced-air cooling of blueberries

packaged in plastic clamshells. HortTechnology, Alexandria, v. 9, n. 2, p. 202-205, 1999.

LINDSAY, R.T.; NEALE, M.A.; MESSER; H.J.M. Ventilation rates for the positive

ventilation of vegetables in bulk bins. Journal of Agricultural Engineering Research,

Philadelphia, v. 28, n. 1, p. 33-44, 1983.

LUENGO, R.F.A.; CALBO, A.G.; JACOMINO, A.P.; PESSOA, J.D.C. Avaliação da

compressão em hortaliças e frutas e seu emprego na determinação do limite físico da altura

da embalagem de comercialização. Horticultura Brasileira, Brasília, v. 21, n. 4, p. 704-

707, outubro/dezembro 2003a.

LUENGO, R.F.A.; CAMARGO FILHO, W.; JACOMINO, A.P. Participação do custo da

embalagem na composição do custo de produção e do preço de atacado do tomate de mesa.

Horticultura Brasileira, Brasília, v. 21, n. 4, p. 719-721, outubro-dezembro 2003b.

MARKARIAN, N.R.; VIGNEAULT, C.; GARIÉPY Y.; RENNIE, T. Design of an advanced

control software for controlled atmosphere storage. Computers and Electronics in

Agriculture, Washington DC, USA, v. 39, p. 23-37, 2003.

MAUL, F.; VIGNEAULT, C.; SARGENT, S.A.; CHAU; K.V. ; CARON, J. Nondestructive

sensor system for evaluation of cooling efficiency. In: SENSORS FOR

NONDESTRUCTIVE TESTING INTERNATIONAL CONFERENCE AND TOUR,

1997, Orlando, FL. Proceedings… Ithaca, NY, NRAES, 1997. p. 351-360.

MCGREGOR, B.M. Tropical products transport handbook. Agriculture Handbook, n.668.

Washington, DC: U.S. Department of Agriculture, 1987. 148p.

MCLAUGHLING, N.B. Statistical models for failure of apple tissue under constant-strain-rate

loading. Journal of Texture Studies, Auckland, New Zealand, v. 18, p. 173 186, 1987.

MCLAUGHLING, N.B. AND R.E. PITT. Failure characteristics of apple tissue under cycle

loading. Transactions of the ASAE, St Joseph, MI, USA, v. 27, n. 1, p. 311 320, 1984.

MINH, T.V.; PERRY, J.S.; BENNETT, A.H. Forced-air precooling of white potatoes in bulk.

ASHRAE Transactions, Atlanta, v. 75, n. 2, p. 143-152, 1969.

MOHSENIN, N.N. Thermal properties of foods and agricultural materials. New York:

Gordon and Breach. 1980. 405 p.

114

MONTGOMERY, D. Design and analysis of experiments. 2. ed., New York: John Wiley &

Sons, 1996.

NAKAYAMA, A.; KUWAHARA, F. A macroscopic turbulence model for flow in a porous

medium. Journal of Fluid Engineering, New York, v. 121, p. 427-433, 1999.

NEVES FILHO, L.C. Armazenamento e distribuição frigorificados. In: CORTEZ, L.A.B.;

HONÓRIO, S.L.; MORETTI, C.L. (Eds.) Resfriamento de frutas e hortaliças. Brasília,

DF, Brasil: Embrapa Informação Tecnológica, 2002, p. 165-189.

NEVES FILHO, L.C.; VIGNEAULT, C.; CORTEZ, L.A.B. A cadeia do frio no Brasil e sua

importância para frutas e hortaliças. In: CURSO DE RESFRIAMENTO DE FRUTAS E

HORTALIÇAS, 1997, Campinas. Anais... Campinas: Faculdades de Engenharia Agrícola

e de Engenharia de Alimentos, UNICAMP, 1997.

NORTH, M.F.; LOVATT, S.J.; CLELAND, A.C. Methods for evaluating the effect of large

voids on food package cooling times. Acta Horticulturae, Leuven, Bélgica, v. 476, p. 95-

103, 1998.

ORSAT, V.; VIGNEAULT, C. ; RAGHAVAN, G.S.V. Air diffusers characterization using a

digitized image analysis system. Applied Engineering in Agriculture, St Joseph, MI,

USA, v. 9, n. 1, p. 115-121, 1993.

PARSONS, R.A.; MITCHELL, F.G.; MAYER, G. Forced-air cooling of palletized fresh fruit.

Transactions of ASAE, St Joseph, MI, USA, v. 15, p. 729-731, 1970.

PLATICSUSA. Specific density of major polymers. Produced by Eastpoint Oltean Co. since

1997. Disponível em: <http://www.plasticsusa.com/specgrav.html> Acesso em: ago. 2003.

R. The R Project for Statistical Computing, version 1.7: Auckland: Statistics Department of

the University of Auckland, , 2004. Programa gratuito disponível eletronicamente em

<http://www.r-project.org>

RAGHAVAN, G.S.V.; VIGNEAULT, C.; GARIÉPY, Y.; MARKARIAN, N.R.; ALVO, P.

Refrigerated and controlled/modified atmosphere storage. In: BARRETT, D. (ed).

Biology, Principles and Application. New York, NY, USA: CRC Press, 2003, p. 23-53.

RODRIGUES, D. Vizinhos, vizinhos, negócios à parte – fruticultura. Revista Panorama

Rural, São Paulo, p. 31-34, fev. 2000.

115

SARGENT, S.A. Handling and cooling techniques for maintaining postharvest quality. In:

HOCHMUTH, G.J., MAYNARD, D.N. Vegetable production guide for Florida.

Gainesville: University of Florida, 1998. chap. 17.

SARGENT, S.A.; BRECHT, J.K.; ZOELLNER, J.J. Sensitivity of Tomatoes at Mature-green

and Breaker Ripeness Stages to Internal Bruising. Journal of the American Society for

Horticultural Science, Alexandria, v. 117, n. 1, p. 119-123, 1992.

SARGENT, S.A.; TALBOT, M.T.; BRECHT, J.K. Evaluating precooling methods for

vegetable packinghouse operations. Gainesville: University of Florida, 1991. 7p.

SAS/STAT version 8.2. Cary, North Carolina: SAS Institute, 1988. Software and User’s

guide.

SHOWALTER, R.K. Postharvest water loss control for vegetables. Vegetable Crops Fact

Sheet, VC-28. Gainesville: Vegetable Crops Department, University of Florida, 1981. 4p.

SILVEIRA, A. C. Rigesa vai embalar fruta no campo. Gazeta Mercantil, São Paulo, p. c-47,

3 a 5 de out. 1997.

SILVEIRA, C. De olho na qualidade dos alimentos. Revista OESP - Alimentação, São

Paulo, p.36-40, jul./ago. 1999.

SINGH, A.; SINGH, Y.; BATHEJA, S. Suitability of Packaging Boxes for Tomatoes. Journal

of Food Science and Technology, Whiteknights, v. 29, n. 6, p. 381-383, 1992.

SMALE, N.J.; TANNER, D. J.; AMOS, N. D.; CLELAND, A. C. Airflow properties of

packaged horticultural produce - a practical study. ISHS Acta Horticulturae, Leuven,

Bélgica, v. 599, p. 443-450, 2003.

SOARES, A.G.; CORRÊA, T.B.S.; SARGENT, S.A.; ROBBS, C.F. Perdas na Qualidade de

Tomate na Cadeia Produtiva, Rio de Janeiro: EMBRAPA, 1993.

SPSS version 11.5.0. Chicago: SPSS Inc., 2002. Software for statistical analysis. 1 CD-ROM.

STEEL R.G.D.; TORRIE, J.H. Principles and procedures of statistics: A biometrical

approach. 2. ed. New York: McGraw Hill Book Compagny, 1980.

TANNER, D.J.; CLELAND, A.C.; OPARA, L.U. A generalized mathematical modeling

methodology for the design of horticultural food packages exposed to refrigerated

conditions. Part 2: Heat transfer modeling and testing. International Journal of

Refrigeration, New York, v. 25, n. 1, p. 43-53, 2002.

116

TEIXEIRA, V.L. (jornalista responsável) O mercado de embalagens. Revista OESP -

Embalagens, São Paulo, p.46-47, set./out. 1999.

THOMPSON, J.F.; CHEN, Y.L. Comparative energy use of vacuum, hydro, and forced air

coolers for fruits and vegetables. ASHRAE Transactions, Atlanta, v. 94, n. 1, p. 1427-

1432, 1988.

VAN DER SMAN, R.G.M. Prediction of airflow through a vented box by the Darcy-

Forcheheimer equation. Journal of Food Engineering, Dublin 4, Irlanda, v. 55, p. 49-57,

2002.

VAN DER SMAN, R.G.M.; EVELO, R.G.; WILKINSON, E.C.; VAN DOORN, W.G.

Quality loss in packed rose flowers due to Botrytis cinerea infection as related to

temperature regimes and packaging design. Postharvest Biology and Technology,

Pullman, v.7: 341-350, 1996.

VANEGAS, J.A.G. Fisiologia pós-colheita de tomate (Lycopersicon esculentum Mill)

cultivar Ângela. 1987. 123 p. Dissertação de Mestrado - Faculdade de Engenharia

Agrícola, Universidade Estadual de Campinas, Campinas.

VERBOVEN, P.; HOANG, M. L.; SCHENK, A.; NICOLAI, B.M. Effect of pallet box

stacking on cooling of horticultural products in cool stores. Acta Horticulturae, Leuven,

Bélgica, v. 566, p. 239-244, 2001.

VIGNEAULT, C.; BORDIN, M.R.; ABRAHÃO, R.F. Embalagem para frutas e hortaliças. In:

CORTEZ, L.A.B.; HONÓRIO, S.L.; MORETTI, C.L. (Eds.) Resfriamento de frutas e

hortaliças. Brasília, DF, Brasil: Embrapa Informação Tecnológica, 2002, p. 95-121.

VIGNEAULT, C.; CASTRO, L.R. Indirect airflow distribution measurement for horticultural

crop package, Part I: Produce-simulator property evaluation. Transactions of the ASAE.

2004. (no prelo)

VIGNEAULT, C.; CASTRO, L.R.; CORTEZ, L.A.B. Effect of gravity on forced-air

precooling. IASME Transactions on Mechanical Engineering. 2004. (no prelo)

______. A new approach to measure air distribution through horticultural crop packages. Acta

Horticulturae, Leuven, 8p. (no prelo)

VIGNEAULT C.; CASTRO, L.R.; GAUTRON, G. Effet de la présence de poignées ouvertes

sur le prérefroidissement de produits horticoles. La Revue Internationale du Froid.

2004. (submetido)

117

VIGNEAULT, C.; CASTRO, L.R.; GOYETTE, B.; MARKARIAN, N.R.; CHARLES, M.T.;

BOURGEOIS, G.; CORTEZ, L.A.B. Indirect airflow distribution measurement for

horticultural crop package; Part II: Verification of the research tool applicability.

Transactions of the ASAE. 2004. (submetido)

VIGNEAULT C.; CASTRO, L.R.; PANNETON, B. Laminar to turbulent indirect airflow

measurement method for horticultural crop package. Canadian Agricultural

Engineering. 2004 (a ser submetido)

VIGNEAULT, C., ÉMOND, J.P. Reusable container for the preservation of fresh fruits

and vegetables. United States Patent Application Office, Washington, DC, USA. Patent

Number: 5,727,711. 60p. 1998.

VIGNEAULT, C.; GOYETTE, B. Design of plastic container opening to optimize forced-air

precooling of fruits and vegetables. Applied Engineering in Agriculture, St Joseph, MI,

USA, v. 38, n. 1, p.73-76, 2002a.

______. Largeur des ouvertures au fond de contenants de plastique utilisés pour la

manutention des produits horticoles frais. Canadian Biosystem Engineering, Winnipeg,

v. 44, n. 3, p. 7-10. 2002b.

______. Loss of ice through container openings during liquid-ice cooling of horticultural

crops. Canadian Agricultural Engineering, Winnipeg, v. 43, p. 3.45-3.48, 2001.

VIGNEAULT, C.; GOYETTE, B.; MARKARIAN, N.R.; HUI, C.K.P.; CÔTÉ, S.;

CHARLES, M.T.; ÉMOND, J.P. Plastic container opening area for optimum

hydrocooling. Canadian Biosystem Engineering, Winnipeg. 2004b. (no prelo).

VIGNEAULT, C.; GOYETTE, B.; RAGHAVAN, G.S.V. Continuous flow liquid-ice system

tested on broccoli. Canadian Agricultural Engineering, Winnipeg, v. 37, n. 3, p. 225-

230, 1995.

VIGNEAULT, C.; MARKARIAN, N.R.; DA SILVA, A.; GOYETTE, B. Pressure drop

during forced-air circulation of various horticultural produce. Transaction of the ASAE,

St Joseph, MI, USA. v. 47, n. 3, p. 807-814, 2004a.

VIGNEAULT, C.; MARKARIAN, N.R.; GOYETTE, B.; CASTRO, L.R. Indirect airflow

distribution measurements in horticultural crop packages. In: ASAE ANNUAL

INTERNATIONAL MEETING, 2004, Ottawa, ON, Canada. Proceedings… St Joseph:

ASAE, 2004. Paper n. 04-6163. 14p.

118

VIGNEAULT, C.; PANNETON, B. B.; RAGHAVAN, G.S.V. Image analysis of 3-D clouds

of bubbles. Canadian Agricultural Engineering, Winnipeg, v. 34, n. 4, p. 347-352, 1992.

VIGNEAULT, C.; PANNETON, B. B.; RAGHAVAN, G.S.V.; GRANGER, R.L. Evaluation

of a high pressure water scrubber for controlled atmosphere storage. Canadian

Agricultural Engineering, Winnipeg, v. 37, n. 4, p. 311-319, 1995.

VILELA, N. J.; LUENGO, R.F.A. Technical and economic viability of the Embrapa crate for

marketing of fresh tomatoes. Horticultura Brasileira, Brasília, v.20, n. 2, p. 222-227.

2002, ISSN 0102-0536.

VISSOTO, F. Z. Estudo dos parâmetros que influenciam no pré-resfriamento de um leito

de produtos esféricos disposto no interior de uma embalagem de papelão. 1999. 115 p.

Dissertação de Mestrado - Faculdade de Engenharia Química, Universidade Estadual de

Campinas, Campinas.

WEAST, R.C.; ASTLE, M.J. Handbook of Chemistry and Physics. 59. ed. CRC Press Inc.,

Boca Raton, FL, 1979. 119 p.

XU, Y; BURFOOT, D. Simulating the bulk storage of foodstuffs. Journal of Food

Engineering, Dublin 4, Irlanda, v. 39, p. 23-29, 1999.

119

APÊNDICES

120

APÊNDICE A - Resumo dos resultados obtidos nos experimentos descritos nos artigos de

4 a 7

Tabela 1 – Resultados obtidos para velocidade do ar (Vel, em m.s-1), uniformidade da distribuição do ar (Vi), queda da pressão do ar (APD, em Pa), média do tempo de meio resfriamento (HCTméd, min) e máximo tempo de meio resfriamento (HCTmáx, min) para cada combinação entre configuração de abertura de embalagem e fluxo de ar. TOA (%) Número

Área ind. (%)

Posição Fluxo de ar Vel Vi APD HCTméd

HCTmáx

0,125 0,018 a * 0,543 CC...NN 62,50 n...p 139,3 YY 232,3 ZZ

0,25 0,049 a...g 0,952 AAA 201,99 z 82,3 QQRR 148,0 TTUU

0,5 0,117 n...s 0,866 ZZ 650,10 NN 47,0 BB...DD 85,7 EE...HH0,67 1 0,67 Centro

0,75 0,174 u...w 0,824 YYZZ 1390,39 TTUU 36,2 t...w 62,3 u...y

0,125 0,049 a...g 0,761 V V...YY 24,64 a...i 83,8 RR 151,8 UUVV

0,25 0,095 i...o 0,636 NN...UU 45,81 e...n 53,0 EEFF 87,3 EE...II

0,5 0,191 v...w 0,576 FF...RR 116,55 s...v 33,4 s...t 53,6 r...u

1 0,395 II 0,648 QQ...UU 409,48 FFGG 23,0 l...o 46,4 n...r

Diagonal inferior

2 0,693 SS 0,718 UU...XX 1379,68 SSTT 16,7 f...j 35,3 h...m

0,125 0,042 a...e 0,802 XX...ZZ 23,89 a...h 94,2 SSTT 148,7 TTUU

0,25 0,103 k...p 0,787 WW...ZZ 46,06 f...n 53,7 EE...GG 90,3 FF...JJ

0,5 0,205 w...y 0,753 V V...YY 118,05 t...v 35,4 t...v 70,9 y...CC

1 0,373 HHII 0,712 TT...WW 390,80 EEFF 24,0 n...p 53,5 r...u

4 0,5

Diagonal superior

2 0,585 OO...QQ 0,644 PP...UU 1369,72 SSTT 17,0 f...j 39,8 k...p

0,125 0,045 a...f 0,539 CC...MM 26,88 a...l 81,8 QQRR 135,3 SS

0,25 0,086 h...n 0,612 JJ...SS 51,29 l...n 54,7 FF...HH 87,2 EE...II

0,5 0,164 t...v 0,523 CC...KK 124,52 uv 35,8 t...w 57,4 s...v

1 0,360 GG...HH 0,554 DD...QQ 453,57 JJKK 22,6 k...n 39,2 k...p

Diagonal inferior

2 0,600 QQ 0,555 DD...QQ 1436,73 WW 16,5 f...j 28,9 f...j

0,125 0,036 a...d 0,580 GG...RR 26,13 a...k 92,3 SS 144,0 SS...UU

0,25 0,083 g...n 0,617 KK...SS 51,29 l...n 55,7 FF...I I 93,0 HH...MM

0,5 0,163 t...v 0,592 HH...SS 125,27 uv 35,8 t...w 60,1 t...x

1 0,338 FFGG 0,532 CC...LL 431,90 HHII 23,3 m...o 47,5 o...r

3 0,67

Diagonal superior

2 0,556 NNOO 0,547 CC...OO 1437,22 WW 17,1 f...j 38,8 k...o

Unif distribuído com alça

0,25 0,080 f...m 0,610 JJ...SS 23,64 a...h 57,2 GG...JJ 100,0 KK...OO

Unif distribuído sem alça

0,25 0,060 b...h 0,269 e...m 35,85 a...m 64,7 MM 105,7 OOPP

Unif distribuído com alça

0,5 0,151 s..u 0,611 JJ...SS 40,58 a...n 38,4 v...y 73,7 z...DD

Unif distribuído sem alça

0,5 0,127 o...s 0,292 f...o 85,92 qr 39,8 w...AA 64,6 v...z

2

77 0,026

Unif distribuído com alça

1 0,324 EEFF 0,509 y...HH 115,56 s...v 22,7 k...o 44,9 m...r

121

Unif distribuído sem alça

1 0,289 CCDD 0,307 g...p 302,37 BB 23,4 m...o 37,3 i...o

Unif distribuído com alça

2 0,650 RR 0,335 j...s 425,67 GGHH 13,7 a...f 24,6 c...f

Unif distribuído sem alça

2 0,644 RR 0,289 e...o 1146,54 PP 13,7 a...f 20,7 a...f

0,125 0,024 a 0,602 HH...SS 22,40 a...f 130,0 XX 262,2 AAA

0,25 0,060 b...i 0,663 RR...UU 44,57 d...n 70,6 NN 125,5 QQ

0,5 0,122 o...s 0,780 WW...ZZ 108,33 s...u 46,3 BBCC 79,6 CC...EE

1 0,256 z...CC 0,580 GG...RR 373,36 DDEE 27,1 o...q 46,0 n...r 3 0,67 Centro

2 0,515 LLMM 0,423 s...z 1368,47 SS 16,2 f...j 27,7 f...i

0,125 0,022 a 0,675 SS...V V 23,15 a...g 140,5 YY 209,1 YY

0,25 0,050 a...g 0,786 WW...ZZ 42,82 b...n 84,8 RR 128,2 QQRR

0,5 0,106 l...q 0,679 SS...V V 101,11 r...t 50,3 DDEE 83,1 DD...GG

1 0,227 x...z 0,775 WW...YY 356,68 CCDD 32,3 st 51,8 q...t

4 0,5 Cantos

2 0,534 MMNN 0,834 YYZZ 1309,44 QQ 18,9 h...k 35,1 h...m

0,125 0,051 a...h 0,471 v...DD 19,41 a...c 76,0 OOPP 125,9 QQ

0,25 0,107 m...q 0,469 u...DD 27,63 a...l 46,3 BBCC 74,5 AA...DD

0,5 0,201 wx 0,456 u...CC 50,30 j...n 30,7 q...s 49,3 p...s

1 0,429 JJ 0,517 AA...JJ 151,68 wx 19,0 i...l 30,7 f...k

2 0,713 SS 0,514 z...II 471,26 KK 13,7 a...f 21,4 a...f

Diagonal inferior

3,9 1,163 XX 0,525 CC...KK 1426,26 V V WW 10,1 ab 16,7 a...e

0,125 0,048 a...f 0,686 SS...V V 19,41 ab 79,5 PPQQ 113,6 PP

0,25 0,103 k...p 0,610 JJ...SS 27,38 a...l 47,7 CC DD 78,0 BB...EE

0,5 0,197 wx 0,568 EE...QQ 51,04 k...n 31,0 rs 51,3 q...t

1 0,389 HHII 0,495 x...GG 153,67 wx 19,6 j...m 34,5 g...l

2 0,646 RR 0,424 s...z 502,14 LL 13,8 a...f 23,3 b...f

Diagonal superior

3,9 1,055 V V 0,303 g...p 1436,73 WW 9,8 a 15,7 a...e

0,125 0,033 a...c 0,613 JJ...SS 19,16 ab 105,8 V V 225,2 ZZ

0,25 0,072 e...l 0,544 CC...NN 28,13 a...l 61,0 JJ...MM 101,8 MM...OO

0,5 0,149 r...u 0,578 GG...RR 48,05 g...n 38,8 v...y 67,2 v...AA

1 0,293 DDEE 0,520 BB...JJ 129,51 uv 24,4 n...p 41,4 l...p

2 0,560 NNOO 0,454 u...CC 430,41 HHII 15,7 e...j 27,6 f...i

Centro

3,9 1,037 UUV V 0,359 m...t 1413,81 V V 10,2 a...c 15,2 a...d

0,125 0,034 a...c 0,553 DD...PP 19,41 ab 99,9 UU 157,9 VV

0,25 0,072 e...l 0,561 DD...QQ 27,38 a...l 61,3 JJ MM 92,1 GG...LL

0,5 0,145 r...u 0,573 FF...RR 48,55 h...n 39,0 v...y 61,7 u...y

1 0,283 CCDD 0,608 II...SS 125,52 uv 25,5 n...p 42,2 l...q

2 0,552 NN 0,543 CC...NN 450,83 II...KK 16,090 f...j 25,265 d...g

4 1

Cantos

3,9 1,062 V V 0,617 KK...SS 1409,82 UU V V 10,615 a...d 14,785 a...c

4

77 0,052 Unif

distribuído com alça

0,25 0,071 d...k 0,419 s...y 19,66 a...d 58,6 HH...KK 96,7 II...NN

122

Unif distribuído sem alça

0,25 0,062 c...i 0,234 d...i 20,9 a...e 62,5 KK...MM 103,0 NN...OO

Unif distribuído com alça

0,5 0,140 q...t 0,413 r...x 25,64 a...j 37,8 v...y 67,7 w...AA

Unif distribuído sem alça

0,5 0,122 o...s 0,267 e...m 29,62 a...m 40,5 x...AA 67,3 v...AA

Unif distribuído com alça

1 0,295 DDEE 0,343 j...t 53,04 m...o 23,0 l...o 41,3 l...p

Unif distribuído sem alça

1 0,266 AA...DD 0,265 e...l 79,44 p...r 24,3 n...p 39,7 k...p

Unif distribuído com alça

2 0,588 OO...QQ 0,236 d...i 168,36 xy 14,3 c...g 24,2 c...f

Unif distribuído sem alça

2 0,585 OO...QQ 0,230 d...h 296,15 BB 14,4 d...g 22,1 a...f

0,125 0,021 a 0,426 s...AA 18,66 ab 131,6 XX 199,8 X

0,25 0,052 a...h 0,475 w...EE 25,64 a...j 74,2 OO 113,5 PP

0,5 0,100 j...o 0,538 CC...MM 36,60 a...m 49,0 CC DD 78,5 BB..EE

1 0,234 y...AA 0,483 x...FF 94,88 q...s 27,8 p...r 44,0 l...r

2 0,471 KK 0,514 z...II 347,21 CC 18,4 g...j 38,9 k...o

8 0,5

Parte superior e inferior da

lateral

3,9 0,989 TT 0,629 MM...UU 1330,86 RR 11,7 a...e 23,2 b...f

0,125 0,026 ab 0,300 g...p 17,17 a 109,8 WW 210,1 YY

0,25 0,061 c...i 0,316 g...q 19,91 a...d 62,6 KK...MM 98,8 JJ...OO

0,5 0,115 n...r 0,354 l...t 44,57 c...n 42,0 y...AA 69,3 x...BB

1 0,247 z...BB 0,360 m...t 63,25 n...p 25,4 n...p 44,1 l...r

2 0,516 LLMM 0,325 i...r 181,07 y 15,6 e...j 25,6 e...h

6 9 0,67 Unif distribuído

3,9 1,014 TTUU 0,286 e...n 616,72 MM 9,9 ab 13,8 ab

0,125 0,039 a...e 0,641 OO...UU 18,16 ab 96,5 TTUU 209,9 YY

0,25 0,084 g...n 0,666 RR...UU 22,65 a...f 57,2 GG JJ 99,5 JJ...OO

0,5 0,165 t...v 0,623 LL...TT 33,36 a...m 36,4 u...x 60,2 t...x

1 0,322 EEFF 0,551 DD...PP 76,95 pq 23,1 l...o 38,5 j...o

2 0,594 PPQQ 0,431 t...BB 246,33 AA 14,9 e...i 23,9 b...f

4 2 Centro

3,9 1,108 WW 0,348 k...t 803,79 OO 9,7 a 14,9 a...c

0,125 0,036 a...d 0,338 j...t 17,17 a 92,1 SS 140,1 SSTT

0,25 0,073 e...l 0,386 p...w 20,16 a...d 57,8 HH JJ 85,6 EE...HH

0,5 0,136 p...t 0,404 q...x 24,64 a...i 38,8 v...y 58,2 s...w

1 0,288 CCDD 0,378 n...u 49,30 i...n 23,7 m...p 36,9 i...n

2 0,564 NN...PP 0,381 o...v 134,74 vw 15,3 e...i 24,1 c...f

8 1

Parte superior e inferior da

lateral

3,9 1,051 V V 0,388 p...w 442,61 HH...JJ 10,1 a...c 15,4 a...d

8

77 0,104 Unif

distribuído com alça

0,25 0,065 c...i 0,252 d...j 17,67 ab 61,2 JJ...MM 94, 5 HH...NN

123

Unif distribuído sem alça

0,25 0,060 b...i 0,167 b...d 17,67 ab 63,8 LLMM 104,6 OO

Unif distribuído com alça

0,5 0,127 o...s 0,263 e...l 19,66 a...d 39,5 v...z 64,2 v...z

Unif distribuído sem alça

0,5 0,109 m...q 0,199 d...f 20,41 a...d 43,3 z...BB 68,0 w...AA

Unif distribuído com alça

1 0,268 BB...DD 0,232 d...i 30,87 a...m 23,9 n...p 40,4 k...p

Unif distribuído sem alça

1 0,249 z...BB 0,222 d...g 35,60 a...m 25, 2 n...p 40,8 l...p

Unif distribuído com alça

2 0,559 NNOO 0,166 b...d 74,96 o...q 14,6 d...h 22,5 b...f

Unif distribuído sem alça

2 0,513 LLMM 0,224 d...g 96,63 q...t 15,7 e...j 24,8 c...f

Unif distribuído com alça

0,25 0,067 c...j 0,203 d...f 17,17 a 60,3 JJ...LL 90,5 FF...KK

Unif distribuído sem alça

0,25 0,064 c...i 0,223 d...g 17,17 a 62,3 KK...MM 101,1 LL...OO

Unif distribuído com alça

0,5 0,124 o...s 0,257 d...k 17,91 a 40,5 x...AA 62,7 u...y

Unif distribuído sem alça

0,5 0,116 n...r 0,322 h...r 17,91 a 43,2 z...BB 66,1 v...AA

Unif distribuído com alça

1 0,264 AA...DD 0,234 d...i 22,40 a...f 24,4 n...p 39,3 k...p

Unif distribuído sem alça

1 0,259 z...CC 0,283 e...m 23,15 a...g 25,0 n...p 39,7 k...p

Unif distribuído com alça

2 0,610 QQ 0,252 d...j 40,08 a...n 14,1 b...g 21,8 a...f

16 77 0,208

Unif distribuído sem alça

2 0,534 MMNN 0,282 e...m 45,81 e...n 15,4 e...j 23,8 b...f

0,125 0,033 a...c 0,195 c...e 16,92 a 96,1 TT 183,3 WW

0,25 0,064 c...i 0,075 a 17,17 a 59,5 I I KK 81,6 DD...FF

0,5 0,103 k...p 0,102 ab 17,42 a 43,7 AABB 59,6 t...x

1 0,243 z...BB 0,063 a 15,92 a 24,7 n...p 34,9 h...m

2 0,495 KKLL 0,040 a 30,87 a...m 15,5 e...j 21,1 a...f

100

Sem embalagem

3,9 1,009 TTUU 0,114 a...c 61,75 n...p 9,8 a 12,30 a

*Médias representadas por letras iguais na mesma coluna não diferem significativamente a 5% de significância (segundo teste Duncan).

124

APÊNDICE B - Artigo 9. Laminar to turbulent indirect airflow measurement method

for horticultural crop package

Abstract

A group of 64 plastic balls were instrumented and used as horticultural produce

simulators and strategically distributed in an orthogonal matrix along with other 448 plastic

spheres to simulate precooling of column stacked produce. The half-cooling times of the

instrumented simulators inside the two-end-fully-open ball matrix were determined at different

airflows. Thus, a research tool to determine air distribution through porous medium was

developed basing on the correlation between the cooling rate of produce simulators and the air

approach velocity. Particular cares were applied to the evaluation of transient zone of airflow

rate since it was previously recognized as particularly difficult for air velocity indirect

measurement. This particular attention included the refinement of airflow range studied in

each phase: laminar (Re<2000; 0.125, 0.1875, 0.25 and 0.3125 L·s-1·kg-1), transient

(2000<Re<3000; 0.375, 0.5, and 0.625L·s-1·kg-1), and turbulent airflows (Re>3000; 0.75,

0.875, 1, 1.5, 2, 3, and 3.9 L·s-1·kg-1).

Although from 0.375 to 0.625 L·s-1·kg-1 an underestimation of the HCT has still been

observed, the accuracy of the indirect airflow measuring method previously developed was

significantly improved by examining thoroughly the air pattern behavior at transient zone.

This refinement generated more precise results for the air mass balance on the airflow

direction layers, which were fairly close to the overall results obtained from the Pitot tube

(100.03 ± 9.23%). This new research tool also allowed establishing different models to deduce

airflow rate from produce average half-cooling time at each turbulent, transient and laminar

segments with high correlation coefficient (99.13 ± 0.46% and Figure 1). Therefore, this tool

has vast potential for container design since it can provide accurate air distribution pattern

through packed horticultural produce.

125

0

1

1

2

2

3

3

4

4

5

0 25 50 75 100

HCT (min-1)

Laminar-flow prediction ModelTransient-flow prediction ModelTurbulent-flow prediction ModelExperimental Data

Flow

rate

(L

•kg-1

•s-1

)

Figure 1: Effect of airflow rate on the HCT responses based on the result averages.

126

APÊNDICE C - Artigo 10. A new approach to measure air distribution through

horticultural crop packages C. Vigneault Horticultural Research and Development Centre, Agriculture and Agro-Food Canada, Saint-Jean-sur-Richelieu, (Qc) Canada

L.R. de Castro and L.A.B. Cortez FEAGRI, State University of Campinas, SP, Brazil

Key words: Simulator, forced-air, performance, container, handling, precooling

Abstract The applicability of using instrumented balls as an indirect measurement of air velocity was evaluated. A group of 64 instrumented plastic balls were used as horticultural produce simulators and strategically distributed in an orthogonal matrix along with other 448 plastic spheres to simulate precooling of column stacked produce. The matrix containing the simulators was submitted to cooling process under controlled conditions. Correlations were determined by measuring the half-cooling time of 64 instrumented simulators positioned at fixed locations inside a two-end-fully-open ball matrix. The surrounding air velocity was inferred as a function of the simulator locations in reference to the air entrance. This method was then evaluated by comparing the data obtained for three package opening areas (0.67%, 2%, and 6%), and six airflow rates (ranging from 0.125 to 3.9 L•s-1•kg-1), and by performing a mass balance. These analyses attested the capacity of this method in predicting the variation of the cooling rate as a function of the airflow rate and its potential in predicting the mean air velocity through the ball matrix.

INTRODUCTION One of the main concerns of package industry engineers and researchers is to design containers with adequate fluid-dynamic attributes for fruit and vegetable handling. The container must have enough openings to promote satisfying airflow through the produce while ensuring that the mechanical resistance is not compromised. Although there are references regarding some package parameters such as overall dimensions and opening area (Kader 2002; Vigneault and Émond 1998), these provide mainly standard recommendations.

While some authors propose to predict airflow through mathematic-computational simulation (Smale et al 2003, Tanner et al. 2002), most existing methodologies recommend experimentation with real horticultural produce (Dincer 1994; Chau et al. 1985). However, it becomes very difficult to maintain similar thermal properties and produce positioning pattern when replicating experiments with packed produce. The thermal properties, stacking arrangement, and bed porosity are crucial to establish airflow pattern while modeling cooling process and designing equipments (Fontana et al. 1999). Differences in size, shape, weight, external surface roughness, and heat and mass transfer properties may be determined for each variety of produce, but their physical and chemical properties change with time as they ripen and go through other physiological changes (ASHRAE 2002). Sometimes, experiments can not be repeated because of a lack of produce uniformity (Tanner et al. 2002).

The standard ways of determining air distribution is by introducing measuring instruments in the air pathway (ASHRAE 2001). The drawback of this method is the difficulty in obtaining consistent results. For instance, it is quite impossible to place a sensor exactly

127

perpendicular to the airflow direction in a porous medium such as packed horticultural produce. The air circulating is continuously changing direction and the measuring device insertion generally disturbs the air pathway(Alvarez and Flick, 1999).

Ideally, the use of a stable material to replace real horticultural produce would minimize the variability in tests. Indirect measurement of ambient condition has been used with success by many authors. Among them, Vigneault et al. (1995) employed instrumented nylon cylinder to measure ice distribution uniformity; Maul et al. (1997) used rubber balls filled with water/agar-agar solution to measure water distribution; Alvarez and Flick (1999) utilized aluminum spheres to measure air distribution, and Castro et al. (2003) applied plastic balls filled with water/agar-agar solution to simulate horticultural produce in a forced-air precooling system. The last authors concluded that their system offered great opportunities for the airflow measurement and could be used for validation of a model describing airflow distribution in porous medium. However, the agar-agar plastic balls were not precise enough (Castro et al. 2003). The best results presented in the literature were obtained by using stable water proof material such as aluminum (Alvarez and Flick 1999) or nylon (Vigneault et al. 1995).

The aim of this project is to develop a practical research tool that allows the determination of airflow pattern inside horticultural crop containers.

MATERIAL AND METHODS

Produce simulator The horticultural produce simulator described in detail by Vigneault and Castro (2004)

consists of solid polymer balls 52.36 mm in diameter and weighing 125.55 g. Sixty-four balls were selected for their relatively high uniformity in terms of cooling index (-0.1414 ± 0.0081 min-1) and heat capacity (1.1252 ± 0.0657 kJ·kg-1·oC-1). Each of the 64 balls was instrumented with a 30-gage 5 m-long insulated copper constantan thermocouple wire (Type T) placed in their center with a precision of ±0.025 mm.

Experimental set-up The sixty four instrumented balls were stacked uniformly distributed along with other 448 balls on a columnar pattern to form a cubic matrix of 8-ball-side dimension. The orthogonal positioning reference system of the instrumented balls is presented by Vigneault and Castro (2004). The arrangement resulted in 47.64% of porosity.

Figure 1 shows the experimental set up used during the trials. Four transparent acrylic plates were assembled to simulate a forced-air cooling tunnel of 420 mm inside square cross-section, and 1250 mm long. The ball matrix was positioned at a distance of 220 mm from the tunnel air-inlet. The portion of the tunnel containing the balls was insulated with a 25 mm-thick polystyrene foam to reduce heat conduction. The air-outlet of the tunnel consisted on a 610 mm long plenum enabling air pressure drop measurements across the ball matrix using a pressure transmitter in the range of 0-127±6 mm of water. The air-outlet was air-tightly attached to the aspiration chamber built from 10 mm thick plywood with a 900x600x600 mm of outer dimensions and containing a direct drive radial blade fan driven by a 0.75 kW variable speed motor. The fan created a negative pressure inside the aspiration chamber and forced the outside air to circulate through the cooling tunnel. The air was released to the atmosphere through a 500 mm long, 101.6 mm-diameter tube instrumented with a Pitot-tube device

128

allowing airflow measurements. The whole experimental set-up was placed in a cold chamber maintained at 4oC to precool the ball matrix.

A fully open configuration was initially used to generate the correlations between the half-cooling time (HCT) of the 64 produce simulators and air approach velocity. Based on the ball matrix volume, the airflow rates studied were equivalent to 0.125, 0.25, 0.5, 1, 2, and 3.9 L·s-1·kg-1 of apple (Vigneault et al., 1995). Three other opening configurations were then investigated to verify the equations developed. For this purpose, a pair of plates was placed next to the first and eighth layers of balls to enclose the matrix within a simulate two-perforated-side package. The pair of square plates was made from 3-mm-thick 420 mm polypropylene plates and perforated using circular metal saws. Nine holes 38.6 mm in diameter, or 0.67% of the total plate area, were uniformly distributed on the plate surface (Figure 2). Three total opening areas (TOA) were evaluated and corresponded to 0.67, 2, and 6% of the plate area using the central hole, the central line of holes, or the nine holes, respectively.

The pressure drop obtained with the greatest airflow rates and the less holes configurations over passed the transmitter upper-limit (125 mm of water), set as the maximum feasible value for the experiments. Therefore, the highest airflow rate tested for the plates with one and three holes were 0.75 L•s-1•kg-1 and 2 L•s-1•kg-1, respectively.

Experimental procedure Four opening configurations were tested with the different airflow rates and repeated

three times. Prior to the start of each test, the forced-air tunnel containing the balls was placed in a warm chamber maintained at 27±1.0oC for a uniform warming up. After this process, the perforated plates were installed and the tunnel was placed in the cold room. The tunnel-end air inlet was connected to an aluminum thermal mass. The air outlet was attached to the aspiration chamber, and the centrifugal fan was turned on immediately. The data were recorded until the temperature of the warmest ball had reached 6.9oC, which corresponds to a 7/8 cooling process (Goyette et al. 1996). At this point the software ended the control process and turned off all devices. The temperature-time data recorded were used to calculate the HCT and CR (cooling rate) of each ball for all treatments by using a dedicated ExcelTM macro developed by Goyette et al. (1996).

Air-mass balance research tool evaluation The mean air approach velocity (MAAV, m·s-1) was calculated for the airflow rates

based on the square cross-section of the forced-air cooling tunnel as the total area. A different correlation was developed between the MAAV and the CR for each ball. These equations included the specific thermal properties of the balls (Vigneault and Castro 2004) and the variability of the air movement around every simulator according to its specific position in the matrix. This method was used to infer the air approach velocity at the 64 different locations inside the container from the CR data obtained for the three container opening areas (0.67%, 2%, and 6%), and the airflow rates ranging from 0.125 to 3.9 L·s-1·kg-1. The mass of air circulation around each ball was calculated assuming a vertical and horizontal symmetry of air distribution. The total mass of air was determined for each air-direction layer and compared to the mass of air measured by the airflow measuring device through a one-way ANOVA and Tukey statistical test.

129

Number of replications The effect of the number of replication (n) on the experimental design power for

identifying significant differences between treatments was determined with equation (1).

nsFd

22 = (1)

Where F is a tabulated Fisher-Distribution value for the desired confidence level and the degree of freedom of the initial sample, s2 is the variance of the samples, and d is the half-width of the resulting confidence interval (Steel and Torrie, 1980).

Results

Research tool evaluation based on air-mass balance The mass balance based on MAAV (m•min-1) was performed on the 64 equations

developed from the whole matrix in fully open configuration submitted to the 6 airflow rates. A regression analysis (Eq. 2) showed good overall correlation (R2=0.8841) when considering all the results as a whole (Fig. 3). Individual ball performances were then used to determine the correlations between air velocity MAAV (m•min-1) and the half-cooling time (HCT, min) (Eq. 3). The 64 resulting equations explained 98.77 ± 0.89% (R2) of the variation of the CR of the balls for the operating conditions tested. The a and b empirical parameters and individual goodness of fit coefficients (R2) of eight equations are presented on Table 1, as examples. HCT = 25.416 MAAV -0,6406 R2=0. 8841 (2) HCT = a· MAAV b R2=0.9877±0.0089 (3)

The results of this mass balance demonstrated that the balls area taken into account allowed determining the mean air velocity around each instrumented ball. The mean air velocities based on the SCSV produced a mass balance explaining 99.3% of the variation of the airflow rates measured mechanically and showed a significant correlation between the results of these two measuring methods (F6, 353 = 4405.2, P<0.0005).

Number of replications The effect of the number of replication (n) on the experimental design power for

identifying significant differences between treatments is presented on Table 2. As expected, the minimum difference between two HCT results to be considered as significantly different at a level of confidence of 95% decreased with the increase of airflow rate and the number of replicate. These results allow determining the number of replications required to reveal the effect of any airflow rate or opening configuration as long as the expected mean and variance of two experimental conditions are known. In the present case, one replicate at each airflow rate level would suffice to discriminate the effect of this variable. However, three replicates were necessary to identify the effect of different opening configurations.

Air-mass balance in z-direction The MAAV approach developed included the warming effect of the air when crossing

the matrix and reaching the balls at the rear layers. Therefore, this approach validated the velocity symmetry hypothesis assumed for the z-depth direction and did not show any air-mass significant correlation with z-direction (Pearson144 = 0.010, P2-tailed =0.908).

An analysis of variance based on the air-direction layer at each airflow rate was performed (Table 2). The position of the ball in air pathway-direction had a significant effect on the measurement of the airflow when ranging from 0.125 to 1 L•s-1•kg-1. This global

130

difference was generally due to the significant difference found among the first and the other layers. For 1 L.s-1.kg-1 the velocities obtained in that layer were significantly higher only than the values found on the last layer.

The graphics of the percentage of the total mass of air calculated at each airflow-layer and the minimal and maximal airflow rates tested (Figure 4) showed a very good accuracy and precisions (107.4 ± 5.94% and 105.8 ± 5.75%), respectively. Furthermore, this method results were fairly close to those obtained from the Pitot tube measuring technique for each z-direction layer. Although the overall average of the air mass had represented 100.99 ± 9,18% of the instrumentally measured data, they were significantly different when varying the airflow rate (Figure 5). The same figure shows an increase of mass balance performance as the airflow rate raised. In fact, fairly constantly lower results were obtained when experimenting at 0.5 L.s-1.kg-1 giving a 85.77% mass balance difference. This finding is due to the regression equation underestimation of the HCT as a function of the airflow rate at this particular airflow rate level (Figure 3). The calculation of the Reynolds number resulted on 2081, which corresponds to the transient phase of airflow pattern from laminar to turbulent. Thus, a particular study should be performed to clarify the correlation between HCT and airflow rate for this particular case to increase the accuracy of the setup.

The performances of this method are considered as very satisfactory for an indirect measurement method of physical phenomena when compared to the mean precision of 77% obtained by Vigneault et al. (1992), which was already judged as fairly good result (Orsat et al., 1993). Furthermore, this indirect measurement method produced even more precise and stable results, and had less disturbing effect on the airflow pattern than the aluminum sphere or direct measurement methods (Alvarez and Flick, 1999).

Experimental setup performances The experimental method permitted to discriminate the airflow rate effect. In fact,

using only one replication would be enough for this accomplishment. Some effects of different opening configurations were also discriminated. The analysis of variance showed that the opening area had a significant effect on the variance results of HCT (F3, 1340 = 115.047, P<0.0005). The results obtained also allowed to demonstrate that:

The variance increases as the opening area is reduced; No significant difference exists between the data for 6% and fully open; At the minimum opening configuration (total surface area = 0.67%), the airflow rate

has a significant effect on the HCT variance (F3, 252 = 22.47, P<0.0005). The highest variation is observed for the data produced with 0.5 L·s-1·kg-1;

The 0.125 and 0.25 L·s-1·kg-1 airflow rates produce lower variance. The performance obtained showed the great potential of the experimental setup and the

MAAV method to discriminate the different effects of the airflow rate and opening configuration. Further research is necessary to establish these effects, but the experimental setup is considered as capable of distinguishing them.

ACKNOWLEDGEMENT This project was accomplished with the financial support from Fundação de Amparo à

Pesquisa do Estado de São Paulo (FAPESP), São Paulo, SP, Brazil, and the HRDC of Agriculture and Agri-Food Canada. The authors would like to thank also Bernard Goyette and Naro R. Markarian for the technical support.

131

Literature Cited Alvarez, G. and D. Flick. 1999. Analysis of heterogeneous cooling of agricultural products

inside bins. Part I: Aerodynamic study. J. Food Eng. 39: 227-237. ASHRAE. 2001. Measurement and Instruments. 1998 Fundamentals Handbook. ASHRAE,

Atlanta. 13.12-13.17. ASHRAE. 2002. Thermal properties of Foods. Refrigeration Handbook. ASHRAE, Atlanta.

8.1-8.12. Castro, L. R., Vigneault, C., Cortez, L. A. B. 2003. Container opening design for horticultural

produce cooling efficiency. Int. J. Food Agr. Env. 2 (1): 135-140. Chau, K. V., J.J. Gaffney, C. D. Baird and G. A. Church. 1985. Resistance to airflow of

oranges in bulk and in cartoons. Trans. ASAE. 8(6): 2083-2088. Dincer, I. 1994. Precooling of cylindrically shaped grapes: experimental and theorical heat

transfer rates. J. Food Process Eng. 17: 57-71. Fontana, A. J., J. Varith, J. Ikediala, J., Reyes, and B. Wacker. 1999. Thermal properties of

selected foods using a dual needle heat-pulse sensor. ASAE Paper no. 99-6063, Goyette, B., C. Vigneault, B. Panneton and G. S. V. Raghavan. 1996. Method to evaluate the

average temperature at the surface of a horticultural crop. Can. Agric. Eng. 38(4): 291-295.

Kader A.A. (ed) 2002. Postharvest technology of horticultural crops. 3rd edition. Coop. Ext. Uni. of Ca. Div. Agric and Nat. Res. Univ. of CA, Davis, CA. Publ. no. 3311.

Lambrinos, G. and H. Assimaki, 1997. Air precooling and hydrocooling of Hayward kiwifruit. Acta Hort. 44 (2): 561-566

Maul, F., C. Vigneault, S.A. Sargent, K.V. Chau and J. Caron. 1997. Nondestructive sensor system for evaluation of cooling efficiency. Proc. of the Sensors for Nondestructive Testing Int. Conf. and Tour. February 18-21. p. 351-360

Smale, N. J., D. J Tanner, N. D. Amos, A. C. Cleland. 2003. Airflow properties of packaged horticultural produce - a practical study. ISHS Acta Hort. 599: 443-450

Steel R.G.D., J.H. Torrie. 1980. Principles and procedures of statistics: A biometrical approach. 2nd.. McGraw Hill Book Compagny, New York.

Tanner, D. J., A. C. Cleland, L. U. Opara, 2002. A generalized mathematical modelling methodology for the design of horticultural food packages exposed to refrigerated conditions. Part 2: Heat transfer modelling and testing. Int. J. of Refri. 25 (1): 43-53

Vigneault, C and J.P. Émond. 1998. Reusable container for the preservation of fresh fruits and vegetables. United States Patent Application Office, Washington. Patent Number: 5,727,711. 60pp.

Vigneault, C and L.R. de Castro. 2004. Indirect airflow distribution measurement for horticultural crop package, Part I: Development of the research tool. Trans. ASAE. (Submitted)

Vigneault, C., B. Goyette, and G.S.V. Raghavan. 1995. Continuous flow liquid-ice system tested on broccoli. Can. Agric. Eng. 37(3):225-230

132

Table 1. Examples of goodness of fit coefficient and empirical parameters relating the half-cooling time (HCT) to the air approach velocity calculated according to the indirect measuring method of eight of the balls used to form the matrix.

Ball Code a b R2 Ball Code a b R2 0 8.95 0.359 0.967 32 11.90 0.595 0.979 8 9.16 0.503 0.981 40 11.29 0.642 0.979 16 10.12 0.541 0.977 48 13.37 0.598 0.974 24 10.08 0.615 0.987 56 11.44 0.640 0.974

Table 2. Minimum difference between two HCT results to be considered as significantly different at a level of confidence of 95% (Alpha = 0.05) as a function of the number of samples and different airflow rates.

Airflow rate (L.s-1.kg-1)

S2 Md1 (min)

Md2 (min)

Md3 (min)

Md4 (min)

Number of sample

Velocity* (m.s-1)

0,125 3,527 3,72 2,33 1,78 1,48 72 0.022a

0,25 1,859 2,71 1,70 1,30 1,08 72 0.055b 0,5 0,979 1,98 1,24 0,95 0,79 72 0.114c 1,0 0,516 1,45 0,91 0,69 0,58 48 0.244 d 2,0 0,272 1,06 0,66 0,51 0,42 48 0.507 e 3,9 0,147 0,78 0,49 0,37 0,31 24 1.019f

Mdx = minimum difference when using x replicates (x = 1, 2, 3 or 4) *Means followed by the same letter in the same column are not significantly different at α = 0.05.

pitot tube

outlet tube

fan

air flow

air flow

mobile plastic tunnel

mobile fan set up

mobile heat exchanger

aluminum plate

ball matrix

static pressuremeasuring device plywood box

air flow

plastic wrappolystyrene foam

Figure 1. Experimental set up showing forced air tunnel, balls matrix, fan and pressures measuring devices.

133

419,141

9,1

104.

8104.8 104.8

38.7

Ø

Figure 2. Plastic plate with nine 19.3 mm radius holes uniformly distributed. Figure 3: Mean-replication effect of airflow rate on the HCT responses.

Figure 4: Percentage of the mass of air measured by the indirect method for each z-direction at 0.125 and 3.9 L.s-1.kg-1. Figure 5: Percentage of the total mass of air measured by the indirect measuring method as a function of the airflow rates (L.s-1.kg-1)

80

100

120

0 2 4 6 8z-direction

Air-

mas

s ra

tio (%

)

0.125

3.9

y = 25,416x-0,6406

R2 = 0,9971

0

20

40

60

80

100

120

0 1 2 3 4Flow rate (L•s 1•kg 1)

HTC

Experimental results

Experimental result tendency

80

90

100

110

120

0 1 2 3 4Flow rate (L•s -1•kg -1)

Air-

mas

s ra

tio (%

)

134

ANEXO

135

136

137

138

139