<%BANNER%>

Air Transport of Horticultural Products

Permanent Link: http://ufdc.ufl.edu/UFE0042079/00001

Material Information

Title: Air Transport of Horticultural Products A Thermal Analysis
Physical Description: 1 online resource (248 p.)
Language: english
Creator: Pelletier, William
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: aircraft, fruit, horticulture, temperature, transport, vegetables
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: AIR TRANSPORT OF HORTICULTURAL PRODUCTS: A THERMAL ANALYSIS By William Pelletier August 2010 Chair: Khe V. Chau Cochair: Ray A. Bucklin Major: Agricultural and Biological Engineering Temperature is the most important factor in the postharvest life of horticultural products and temperature management is essential to reach the optimum postharvest quality. Preserving the quality of fresh fruits and vegetables requires a proper cold chain from the field to the consumers' home. Too often, transport operations are responsible for breaks in the cold chain, particularly in the case of air shipments. Air transport plays a major role in the global food trade and, even during the recent worldwide economic slump the sector of air freight occupied by food products only experienced a small decline. Maintaining a proper cold chain during air transport operations is challenging since the cargo may spend extended periods of time on the tarmac, where they are exposed, with limited protection, to harsh environmental conditions. Studies on in-flight environmental conditions and their impacts on the temperature distribution in loads of fresh fruits and vegetables transported by air have been limited in number. However, such studies are important to develop new handling methods as well as to predict and preserve quality. A thermal analysis was conducted on loads of horticultural products in single boxes and in an aircraft container exposed to detrimental temperatures in laboratory conditions. Different fruit sizes and packing arrangements were used for the tests on individual boxes. These tests showed small temperature differences between the pulp and air temperatures within the boxes. Relatively fast rates of change of the temperatures were observed even in the core region of the boxes. For the aircraft container, a similar thermal behavior was observed in the fruit near the outside surface of the load, particularly for the boxes located on the top row. Even after 8 h of exposure to detrimental conditions, the temperatures within the core of the aircraft container remained almost constant. With the exception of the bottom layer, a vertical stratification of the temperatures was observed in the boxes as well as in the container. In addition to laboratory tests, temperatures were also monitored during several international flights using an instrumented aircraft container loaded with simulated horticultural products. Thermal behavior similar to that of the laboratory tests was observed during the air transport operations. The tests showed that the ramp transfers before and after flights were critical to maintaining a proper cold chain, mainly because of the effect of solar radiation. For some shipments, the ramp transfer exceeded 8 h and temperatures above 60degreeC were also measured on the inside surfaces of the aircraft container walls. Even though temperatures were not always optimum within the aircraft cargo compartments, their effects were not as detrimental as those associated with ramp transfers. Based on the experimental data collected through laboratory and air shipment tests, the validity of a heat transfer model based on an effective thermal conductivity was investigated. For individual boxes, simulations showed that an effective thermal conductivity approach provided acceptable results in the lower and lateral regions of the load but significantly underpredicted the temperatures in the core regions. Aircraft container simulations revealed that the temperatures were also underpredicted in the top of the boxes located on the bottom layer and at the bottom of the boxes located on the top layer. The model did not provide a good representation of the temperature distribution throughout the load of products, but did provide good results in the peripheral region. It appeared that the effect of natural convection must be included in the effective thermal conductivity via a variable dynamic component to improve the overall results of such a modeling approach. However, the results indicated that the modeling approach used could still be implemented as a useful tool for air shipments of horticultural products, since temperature abuses are generally observed in the peripheral region of the load, where the model provided useful results.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by William Pelletier.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Chau, Khe V.
Local: Co-adviser: Bucklin, Ray A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042079:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042079/00001

Material Information

Title: Air Transport of Horticultural Products A Thermal Analysis
Physical Description: 1 online resource (248 p.)
Language: english
Creator: Pelletier, William
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: aircraft, fruit, horticulture, temperature, transport, vegetables
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: AIR TRANSPORT OF HORTICULTURAL PRODUCTS: A THERMAL ANALYSIS By William Pelletier August 2010 Chair: Khe V. Chau Cochair: Ray A. Bucklin Major: Agricultural and Biological Engineering Temperature is the most important factor in the postharvest life of horticultural products and temperature management is essential to reach the optimum postharvest quality. Preserving the quality of fresh fruits and vegetables requires a proper cold chain from the field to the consumers' home. Too often, transport operations are responsible for breaks in the cold chain, particularly in the case of air shipments. Air transport plays a major role in the global food trade and, even during the recent worldwide economic slump the sector of air freight occupied by food products only experienced a small decline. Maintaining a proper cold chain during air transport operations is challenging since the cargo may spend extended periods of time on the tarmac, where they are exposed, with limited protection, to harsh environmental conditions. Studies on in-flight environmental conditions and their impacts on the temperature distribution in loads of fresh fruits and vegetables transported by air have been limited in number. However, such studies are important to develop new handling methods as well as to predict and preserve quality. A thermal analysis was conducted on loads of horticultural products in single boxes and in an aircraft container exposed to detrimental temperatures in laboratory conditions. Different fruit sizes and packing arrangements were used for the tests on individual boxes. These tests showed small temperature differences between the pulp and air temperatures within the boxes. Relatively fast rates of change of the temperatures were observed even in the core region of the boxes. For the aircraft container, a similar thermal behavior was observed in the fruit near the outside surface of the load, particularly for the boxes located on the top row. Even after 8 h of exposure to detrimental conditions, the temperatures within the core of the aircraft container remained almost constant. With the exception of the bottom layer, a vertical stratification of the temperatures was observed in the boxes as well as in the container. In addition to laboratory tests, temperatures were also monitored during several international flights using an instrumented aircraft container loaded with simulated horticultural products. Thermal behavior similar to that of the laboratory tests was observed during the air transport operations. The tests showed that the ramp transfers before and after flights were critical to maintaining a proper cold chain, mainly because of the effect of solar radiation. For some shipments, the ramp transfer exceeded 8 h and temperatures above 60degreeC were also measured on the inside surfaces of the aircraft container walls. Even though temperatures were not always optimum within the aircraft cargo compartments, their effects were not as detrimental as those associated with ramp transfers. Based on the experimental data collected through laboratory and air shipment tests, the validity of a heat transfer model based on an effective thermal conductivity was investigated. For individual boxes, simulations showed that an effective thermal conductivity approach provided acceptable results in the lower and lateral regions of the load but significantly underpredicted the temperatures in the core regions. Aircraft container simulations revealed that the temperatures were also underpredicted in the top of the boxes located on the bottom layer and at the bottom of the boxes located on the top layer. The model did not provide a good representation of the temperature distribution throughout the load of products, but did provide good results in the peripheral region. It appeared that the effect of natural convection must be included in the effective thermal conductivity via a variable dynamic component to improve the overall results of such a modeling approach. However, the results indicated that the modeling approach used could still be implemented as a useful tool for air shipments of horticultural products, since temperature abuses are generally observed in the peripheral region of the load, where the model provided useful results.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by William Pelletier.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Chau, Khe V.
Local: Co-adviser: Bucklin, Ray A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042079:00001


This item has the following downloads:


Full Text

PAGE 1

1 AIR TRANSPORT OF HORTICULTURAL PRODUCTS: A THERMAL ANALYSIS By WILLIAM PELLETIER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 William Pelletier

PAGE 3

3 To my parents Nicole and Guy and my wife, Sophie Marcoux

PAGE 4

4 ACKNOWLEDGMENTS I would like first to thank Khe Van Chau for his valuable teaching s, support and friendship throughout my studies. I considered myself very lucky to have worked with him and I will always think of him as an important role model in my personal and professional life. It is very important for me to thank my friend Dr. Jean P ierre mond who encouraged me to study at the University of Flor ida, supported me and provided invaluable professional opportunities throughout my years as a graduate student. I also want to thank Dr. Ray Bucklin for his constant help in the last stretch of this project as well as for all the motivation he provided to me. I would like to express my sincere gratitude to the Fonds Qubecois de Recherche sur la Nature et les Technologies as well as The International Air Cargo Association for their financial support. Finally, I would like to e specially thank my parents Nicole and Guy, my sister Sophie as well as my wife Sophie for their u nconditional love and encouragement.

PAGE 5

5 TABLE OF CONTENTS ACKNOWLEDGMENTS .................................................................................................. 4 page LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES ........................................................................................................ 10 LIST OF SYMBOLS ...................................................................................................... 17 ABSTRACT ................................................................................................................... 20 CHAPTER 1 INTRODUCTION .................................................................................................... 23 1.1 Temper ature and Fresh Horticultural Commodities ........................................ 23 1.2 Air Transport and its Cold Chain .................................................................... 24 1.3 Statement of the Problem ............................................................................... 29 2 LI TERATURE REVIEW .......................................................................................... 31 2.1 Air Transport .................................................................................................. 31 2.1.1 General Publications ........................................................................... 31 2.1.2 Temperature and Environmental Conditions ....................................... 32 2.2 Heat Transfer Modeling .................................................................................. 35 2.2.1 Models Applied to Packed Beds and Other Porous Media .................. 35 2.2.1.1 Introduction to effective thermal conductivity .......................... 36 2.2.1.2 Effective thermal conductivity models ..................................... 37 2.2.2 Models Applied to Boxed or Bulk Produce .......................................... 45 2.2.3 Models Applied to Air Transport .......................................................... 51 2.2.4 Computational Fluid Dynamics ............................................................ 54 3 MATERIAL AND METHODS .................................................................................. 56 3.1 Laboratory Tests ............................................................................................ 56 3.1.1 Monitoring Equipment .......................................................................... 56 3.1.2 Single Box Tests ................................................................................. 57 3.1.2.1 Fi rst series of tests .................................................................. 58 3.1.2.2 Second series of tests ............................................................ 59 3.1.2.3 Third series of tests ................................................................ 61 3.1.3 Aircraft Container Tests ....................................................................... 62 3. 1.3.1 Produce and packaging .......................................................... 62 3.1.3.2 Instrumentation and methodology ........................................... 63 3.2 Air Transport Tests ......................................................................................... 65 3.2.1 Product and Packaging System ........................................................... 66

PAGE 6

6 3.2 .2 Instrumentation .................................................................................... 67 3.2.3 Flights .................................................................................................. 70 4 RESULTS AND DISCUSSION OF LABORATORY TESTS .................................... 87 4.1 Single Box Tests ............................................................................................ 87 4.1.1 First Series of Tests ............................................................................. 88 4.1.1.1 Analysis of the temperature distribution within the box ........... 88 4.1.1.2 Comparisons of the results between test replications ............. 90 4.1.2 Second Series of Tests ....................................................................... 92 4.1.2.1 Analysis of the temperature distribution with the box .............. 92 4.1.2.2 Comparisons of the results between test replications ............. 94 4.1.3 Third Series of Tests ........................................................................... 95 4.1. 3.1 Analysis of the temperature distribution within the box ........... 95 4.1.3.2 Comparisons of the results between test replications ............. 96 4.1.4 Temperature Differences within the Fruit and with the Surrounding Air ........................................................................................................ 97 4.1.4.1 First series of tests .................................................................. 98 4.1.4.2 Second series of tests ............................................................ 99 4.1.4.3 Third series of tests .............................................................. 100 4.1.4.4 Theoretical calculations on the uniformity of the temperature within a fruit ...................................................... 101 4.2 Aircraft Container Test ................................................................................. 103 4.2.1 Temperatures of the Walls of the Aircraf t Container .......................... 104 4.2.2 Average Layer Temperatures ............................................................ 105 4.2.3 Temperatures of Fruit Located in the Outer Regions of the Load ...... 106 4.2.4 Air and Pulp Temperatures within the Load ....................................... 108 4.2.5 Comparisons of the Results between t he Test Replications .............. 110 4.2.6 Effect of Heat Generation .................................................................. 111 4.3 Conclusions from Laboratory Tests .............................................................. 112 5 RESULTS AND DISCUSSION OF AIR TRANSPORT TESTS ............................. 140 5.1 JFK GOT DXB ............................................................................................. 141 5.1.1 JFK Ramp Transfer ........................................................................... 141 5.1.2 Onboard the Aircraft .......................................................................... 143 5.1. 3 DXB Ramp Transfer .......................................................................... 145 5.1.4 Time and Delays ................................................................................ 146 5.2 First Shipment DXB NBO ............................................................................. 146 5.2.1 DXB Ramp Transfer .......................................................................... 146 5.2.2 On board the Aircraft .......................................................................... 148 5.2.3 NBO Ramp Transfer .......................................................................... 149 5.2.4 Time and Delays ................................................................................ 150 5.3 First Shipment NBO DXB ............................................................................. 150 5.3.1 NBO Ramp Transfer .......................................................................... 151 5.3.2 Onboard the Aircraft .......................................................................... 152 5.3.3 DXB Ramp Transfer .......................................................................... 152

PAGE 7

7 5.3.4 Ti me and Delays ................................................................................ 153 5.4 Second Shipment DXB NBO ........................................................................ 154 5.4.1 DXB Ramp Transfer .......................................................................... 154 5.4.2 Onboard the Aircraft .......................................................................... 155 5.4. 3 NBO Ramp Transfer .......................................................................... 156 5.4.4 Time and Delays ................................................................................ 1 56 5.4.5 Comparison with the Previous DXB NBO Flight ................................ 157 5.5 Second Shipment NBO DXB ........................................................................ 157 5.5. 1 NBO Ramp Transfer .......................................................................... 158 5.5.2 Onboard the Aircraft .......................................................................... 158 5.5.3 DXB Ramp Transfer .......................................................................... 159 5.5.4 Ti me and Delays ................................................................................ 160 5.6 Shipment DXB GOT JFK ............................................................................. 160 5.6.1 DXB Ramp Transfer .......................................................................... 161 5.6.2 Onboard the Aircraft .......................................................................... 161 5.6.3 JFK Ramp Transfer ........................................................................... 163 5.6.4 Ti me and Delays ................................................................................ 163 5.7 Air Transport Logistics and Detrimental Temperatures during Tests ........... 164 5.8 Impacts of the Results for Fresh Horticultural Products ............................... 166 5.8.1 Effect of Heat Generation .................................................................. 166 5.8.2 Effect of Gas Concentrations ............................................................. 167 5.8.3 Mechanical Damage .......................................................................... 169 5.9 Conclusions from Air Transport Tests .......................................................... 170 6 EVALUATION OF HEAT TRANSFER MODELS BASED ON AN EFFECTIVE THERMAL CONDUCTIVITY ................................................................................. 198 6.1 Model and General Assumptions ................................................................. 199 6.2 Sensitivity Analysis on the Effective Thermal Conductivity Model ................ 201 6.3 Simulation Parameters and Boundary Conditions ........................................ 203 6.4 Simulation Results ........................................................................................ 207 6.4.1 Single Box Tests ............................................................................... 207 6.4.1.1 First series of tests ................................................................ 207 6. 4.1.2 Second series of tests .......................................................... 208 6.4.1.3 Third series of tests .............................................................. 209 6.4.2 Aircraft Container Tests ..................................................................... 209 6.4.2.1 Laboratory tests .................................................................... 209 6.4.2. 2 Air transport tests .................................................................. 211 6.5 General Discussion on the Simulation Results ............................................. 212 6.6 Conclusions on the Effective Thermal Conductivity Modeling Approach ...... 213 7 CONCLUSIONS ................................................................................................... 239 LIST OF REFERENCES ............................................................................................. 241 BIOGRAPHICAL SKETCH .......................................................................................... 248

PAGE 8

8 LIST OF TABLES Table page 3 1 Schedule of the six flights as well as the type of aircraft and the position of the experimental aircraft container during the 2007 inflight tests. ...................... 86 4 1 Dimensions of the boxes, number or fruit and layers as well as characteristics of the oranges used for the three series of s ingle box tests. .... 137 4 2 P ulp temperature differences between the half rad ius distance and the center of fruit F1 and F63 for the first series of tests. ....................................... 137 4 3 D ifferences between the air and pulp temperatures at the half radius distance of fruit F1 and F63 for the first series of tests. .................................................. 137 4 4 Pulp temperature differences between the half radius distance and the center of fruit F1, F3 a nd F42 for the second series of tests. ........................... 138 4 5 D ifferences between the air and pu lp temperatures at the half radius distance of fruit F1, F23 and F42 for the second series of tests. .................................... 138 4 6 P ulp temperature differences between t he half radius distance and the ce nter of fruit F1 for the third series of tests. .................................................... 138 4 7 D ifferences between the air and pulp temperature s at the half radius distance of fruit F1, F43 and F72 for the third series of tests. ......................................... 139 4 8 Biot number calculations for a single spherical fruit with constant properties exposed to natural convection in air. ................................................................ 139 5 1 Initial and final temperatures of the inside walls of the instrumented container while outside on the tarmac at JFK airport (04 292007). ................................. 190 5 2 Initial and final temper atures of the inside walls of the instrumented container while in the plane during ramp transfer at GOT airport (0430 2007). .............. 190 5 3 Temperatures of the inside walls of the container when unloaded of the aircraft and after a 30 min on the tarmac at DXB airport (04302007). ............ 190 5 4 Initial and final temperatures as well as temperatures at sunrise of the inside walls of the container while on the tarmac at DXB airport (05032007). .......... 191 5 5 Initial and final average temperatures of the layer s of products in the container while on the tarmac at DXB airport (05032007). ............................. 191 5 6 T emperatures of the layers of products in the container between the time the products were loaded ( DXB ) and unloaded ( NBO ) (05 032007). .................... 192

PAGE 9

9 5 7 Initial and final temperatures of the layers of products in the container during overnight cold room storage at NBO airport (05032007 to 0503 2007). ....... 192 5 8 Initial an d final temperatures of the layers of products in the container while on the tarmac at DXB airport (0504 2007). ..................................................... 193 5 9 Initial and final temperatures as well as temperatures at sunrise of the inside walls of the container while on the tarmac at DXB airport (05062007). .......... 193 5 10 Initial and final average temperatures of t he layers of products in the container while on the tarmac at DXB airport (05062007). ............................. 194 5 11 T emperatures o f the layers of products in the container between the time the products were loaded ( DXB ) and unloaded (N BO ) (05 062007). .................... 194 5 12 Initial and final temperatures of the layers of products in the container during overnight cold room storage at NBO airport (05062007 to 0507 2007). ....... 195 5 13 Initial and final temperatures of the layers of products in the container while on the tarmac at DXB airport (0506 2007 and 0507 2007). ........................... 195 5 14 T emperatures of the layers of products while the container is onboard the aircraft for the DXB GOT JFK flight (05 122007 and 05132007). ................. 196 5 15 Scheduled and actual departure and arrival times as well as corresponding delays for all flights. .......................................................................................... 196 5 16 Total transit time of all flights as well as the corresponding time spent by the container in flight, on the tarmac and onboard the aircraft. ............................... 197 6 1 Sensitivity analysis of Zehner's model .............................................................. 237 6 2 Parameters used with Zehner's model ............................................................. 238 6 3 Parameters as well as thermal and physical properties used for the different simulations. ....................................................................................................... 238 6 4 Convective and radiative heat tra nsfer coefficients used for the aircraft container simulations. ....................................................................................... 238

PAGE 10

10 LIST OF FIGURES Figure page 2 1 Heat transfer mechanisms in packed beds. ........................................................ 55 3 1 Experimental set up used for all singlebox tests. .............................................. 72 3 2 Bottom part and lid of a fully telescopic box. ...................................................... 72 3 3 Staggered arrangement of the oranges for the first series of tests. .................... 72 3 4 Schematic of the first series of tests. .................................................................. 73 3 5 Staggered arrangement of the oranges for the second series of tests. .............. 74 3 6 Schematic of the second series of tests. ............................................................ 74 3 7 Staggered arrangement of the oranges for the third series of tests. ................... 75 3 8 Schematic of the third series of tests. ................................................................. 76 3 9 Internal dimensions of the aircraf t container. ...................................................... 77 3 10 Schematic of the aircraft container for the laboratory tests. ................................ 78 3 11 Experimental set up with the aircraft container for the laboratory tests. ............. 79 3 12 Schematic of the water bottles used for the inflight tests. .................................. 79 3 13 Collapsible reusable plastic container (RPC). ..................................................... 80 3 14 Staggered arrangement of the water bottles for the inflight tests ..................... 80 3 15 Schematic of the location of the temperature sensor inside a bottle of water used for the inflight tests. .................................................................................. 81 3 16 Schematic of a loaded RPC. ............................................................................... 82 3 17 Schematic of the aircraft container for the air transport tests. ............................ 83 3 18 Experimental set up with the aircraft container for the air transport tests. .......... 84 3 19 Unit load device (ULD) positions in the cargo compartments of the aircraft ....... 85 4 1 Identification numbers used for the walls of the boxes of produce. .................. 113 4 2 Temperatures of the inside surfaces of the walls of the box for the first series of tests. ............................................................................................................. 113

PAGE 11

11 4 3 Average temperatures of the layers of fruit in the box of the first series of tests. ................................................................................................................. 114 4 4 Average temperature as well as temperatures of fruit located in corners and near the center of the first layer for the first series of tests. .............................. 114 4 5 Average temperature as well as temperatures of fruit located in corners and near the center of the second layer for the first series of tests. ........................ 115 4 6 Average temperature as well as temperatures of fruit located in corners and near the center of the third layer for the first series of tests. ............................. 115 4 7 Average temperature as well as temperatures of fruit located in corners and near the center of the fourth layer for the first series of tests. ........................... 116 4 8 Average temperature as well as temperatures of fruit located in corners and near the center of the fifth layer for the first series of tests. .............................. 116 4 9 Overview of the temperature distribution throughout the box for the first series of tests. .................................................................................................. 117 4 10 Comparison of the temperatures of the corner and center fruit located on the first layer of the box for the three tests of the first series. ................................. 117 4 11 Comparison of the temperatures of the corner and center fruit located on the second layer of the box for the three tests of the first series. ........................... 118 4 12 Comparison of the temperatures of the corner and center fruit located on the third layer of the box for the three tests of the first series. ................................ 118 4 13 Compariso n of the temperatures of the corner and center fruit l ocated on the fourth layer of the box for the three tests of the fir st series. .............................. 119 4 14 Compariso n of the temperatures of the corn er and center fruit located on the fifth layer of the box for the three tests of the first series. ................................. 119 4 15 Temperatures of the inside surface of the walls of the box for the second series of tests. .................................................................................................. 120 4 16 Average temperatures of the layers of fruit in the box of the second series of tests. ................................................................................................................. 120 4 17 Average temperature as well as temperatures of fruit located in corners and near the center of the first layer for the second series of tests. ........................ 121 4 18 Average temperature as well as temperatures of fruit located in corners and near the center of the second layer for the second series of tests. ................... 121

PAGE 12

12 4 19 Average temperature as well as temperatures of fruit located in corners and near the center of the third layer for the second series of tests. ....................... 122 4 20 Overview of the temperature distribution throughout the box for the second series of tests. .................................................................................................. 122 4 21 Comparison of the t emperatures of the corner and center fruit located on the first layer of the box for the three tests of the second series. ........................... 123 4 22 Comparison of the temperatures of the corner and center fruit located on the second layer of the box for the three tests of the second series. ...................... 123 4 23 Comparison of the temperatures of the corner and center fruit located on the third layer of the box for the three tests of the second series. .......................... 124 4 24 Temperatures of the inside surface of the walls of the box for the third series of tests. ............................................................................................................. 124 4 25 Average temperatures of the layers of fruit in the box of the third series of tests. ................................................................................................................. 125 4 26 Average temperature as well as temperatures of fruit located in corners and near the center of the first layer for the third series of tests. ............................. 125 4 27 Average temperature as well as temperatures of fruit located in corners and near the center of the second layer for the third series of tests. ....................... 126 4 28 Average temperature as well as temperatures of fruit located in corners and near the center of the third layer for the third series of tests. ............................ 126 4 29 Average temperature as well as temperatures of fruit located in corners and near the center of the fourth layer for the thi rd series of tests. ......................... 127 4 30 Overview of the temperature distribution throughout the box for the third series of tests. .................................................................................................. 127 4 31 Comparison of the temperatures of the corner and center fruit located on the first layer of the box for the three tests of the third series. ................................ 128 4 32 Comparison of the temperatures of the corner and center fruit located on the second layer of the box for the three tests of the third series. .......................... 128 4 33 Comparison of the t emperatures of the corner and center fruit located on the third layer of the box for the three tests of the third series. ............................... 129 4 34 Comp arison of the temperatures of the c orner and center fruit located on the fourth layer of the box for the three tests of the third series. ............................. 129 4 35 Identification numbers used for the walls of the aircraft container. ................... 130

PAGE 13

13 4 36 Temperatures of the inside surfaces of the walls of the aircraft container. ....... 130 4 37 Average temperatures of the layers of boxes in the aircraft container. ............. 131 4 38 Average temperature of the bo ttom layer of boxes and pulp temperatures of four fruit located near the outer surface of the load. ......................................... 131 4 39 Average temperature of the third layer of boxes and pulp temperatures of four fruit located near the outer surface of the load. ......................................... 132 4 40 Average temperature of the to p layer of boxes and pulp temperatures of four fruit located near the outer surface of the load. ................................................ 132 4 41 Pulp and air temperatures in box B1. ............................................................... 133 4 42 Picture of a channel created by the bulged boxes of fruit inside the aircraft container. .......................................................................................................... 133 4 43 Pulp and air temperatures in box B47. ............................................................. 134 4 44 Pulp and air temp eratures in box B88 .............................................................. 134 4 45 Comparison of the temperatures of the corner and near center box located on the bottom layer of the aircraft container for the three tests. ....................... 135 4 46 Comparison of the temperatures of the corner and near center box located on the fourth layer of the aircraft container for the three tests. ......................... 135 4 47 Comparison of the temperatures of the corner and near cen ter box located on the sixth layer of the aircraft container for the three tests. ........................... 136 5 1 Temperatures inside the aircraft container during the ramp transfer at JFK airport (0429 2007). ......................................................................................... 172 5 2 Pressure and temperatures inside the aircraft container during the JFK GOT DXB flight (04 292007 and 0430 2007). ......................................................... 173 5 3 T emperatures inside the aircraft container during the ramp transfer at DXB airport (0430 2007). ......................................................................................... 174 5 4 T emperatures inside the aircraft container during the ramp transfer at DXB airport (0503 2007). ......................................................................................... 175 5 5 Pressure and temperatures inside the aircraft container during the flight from DXB to NBO (05 03 2007) ............................................................................... 176 5 6 T emperatures inside the aircraft container during the ramp transfer at NBO airport (0503 2007). ......................................................................................... 177

PAGE 14

14 5 7 T emperatures inside the aircraft container during the ramp transfer at NBO airport (0504 2007). ......................................................................................... 178 5 8 Pressure and temperatures inside the aircraft container during the flight from NBO to DXB (05 04 2007 to 05052007). ....................................................... 179 5 9 T emperatures inside the aircraft container during the ramp transfer at DXB airport (0505 2007). ......................................................................................... 180 5 10 T emperatures inside the aircraft container during the ramp transfer at DXB airport (0506 2007). ......................................................................................... 181 5 11 Pressure and temperatures i nside the aircraft container during the flight from DXB to NBO (05 06 2007). ............................................................................... 182 5 12 T emperatures inside the aircraft container during the ramp transfer at NBO airport (0506 2007). ......................................................................................... 183 5 13 T emperatures inside the aircraft container during the ramp transfer at NBO airport (0507 2007). ......................................................................................... 184 5 14 Pressure and temperatures of the inside the aircraft container during the flight from NBO to DXB (05 072007). .............................................................. 185 5 15 T emperatures inside the aircraft container during the ramp transfer at DXB airport (0506 2007). ......................................................................................... 186 5 16 T emperatures inside the aircraft container during the ramp transfer at DXB airport (0512 2007). ......................................................................................... 187 5 17 Pressure and te mperatures inside the aircraft container during the DXB GOT JFK flight (05 122007 and 05132007). ................................................. 188 5 18 T emperatures inside the aircraft container during the ramp transfer at JFK airport (0513 2007). ......................................................................................... 189 6 1 Results of the sensitivity analysis on Zehner's model for the effective thermal conductivity. ...................................................................................................... 215 6 2 Contact surfaces of a product obtained using the CAD tool of the simulation soft ware. ........................................................................................................... 216 6 3 Simulations of the pulp temperatures of fruit F1, F7 and F13 located on the first layer of the box for the first series of tests. ................................................ 217 6 4 Simulations of the pulp temperatures of fruit F30, F34 and F38 located on the second layer of the box for the first series of tests. ........................................... 218

PAGE 15

15 6 5 Simulations of the pulp temperatures of fruit F51, F57 and F63 located on the third layer of the box for the first series of tests. ............................................... 219 6 6 Simulations of the pulp temperatures of fruit F80, F84 and F88 located on the fourth layer o f the box for the first series of tests. ............................................. 220 6 7 Simulations of the pulp temperatures of fruit F101, F107 and F113 located on the f ifth layer of the box for the first series of tests. .......................................... 221 6 8 Simulations of the pulp temperatures of fruit F1, F5 and F6 located on the first layer of the box for the second series of tests. ........................................... 222 6 9 Simulations of the pulp temperatures of fruit F25, F22 and F23 located on the second layer of the box for the second series of tests. ..................................... 223 6 10 Simulations of the pulp temperatures of fruit F29, F33 and F34 located on the third layer of the box for the second series of tests. ......................................... 224 6 11 Simulations of the pulp temperatures of fruit F1, F7 and F12 located on the first layer of the box for the third series of tests. ............................................... 225 6 12 Simulations of the pulp temperatures of fruit F32, F29 and F25 located on the second layer of the box for the third series of tests. ......................................... 226 6 13 Simulations of the pulp temperatures of fruit F37, F43 and F48 located on the third layer of the box for the third series of tests. .............................................. 227 6 14 Simulations of the pulp temperatures of fruit F68, F65 and F61 located on the fourth layer of the box for the third series of tests. ............................................ 228 6 15 Simulation of the average pulp temperature of the top layer of boxes (L6) in the aircraft container. ........................................................................................ 229 6 16 Simulations of the pulp temperatures of four fruit located near the outer surface of boxes B2, B4, B6 and B11 located on fi rst layer (L1). ...................... 229 6 17 Simulations of the pulp temperatures of four fruit located near the outer surface of boxes B27, B29, B32 and B39 located on the third layer (L3). ........ 230 6 18 Simulations of the pulp temperatures of four fruit located near the outer surface o f boxes B75, B77, B80 and B87 located on the sixth layer (AL6). ...... 230 6 19 Simulations of pulp temperatures of fruit F21, F63 and F105, all located on the threedimensional diagonal of box B1 (L1). ................................................ 231 6 20 Simulations of pulp temperatures of fruit F21, F63 and F105, all located on the threedimensional diagonal of box B88 (L6). .............................................. 232

PAGE 16

16 6 21 Simulations of the average temperatures of layers L1, L4 and L7 during the ramp transfer at DXB airport (05032007). ...................................................... 233 6 22 Simulations of the temperatures of RPC5 and RPC6 during t he ramp transfer at DXB airport (05032007). ............................................................................ 234 6 23 Simulations of the temperatures of RPC25 and RPC26 during the ramp transfer at DXB airport (05032007). ............................................................... 235 6 24 S imulations of the temperatures of RPC49 and RPC50 during the ramp transfer at DXB airport (05 032007). ............................................................... 236

PAGE 17

17 LIST OF SYMBOLS A Surface area ( m2) Bi Biot number ( ) CP Specific heat at constant pressure ( Jkg1K1) g Gravitational acceleration ( m s2) k Thermal conductivity (Wm1K1) h Heat transfer coefficient (Wm2K1) Lc Characteristic length ( m ) m Mass (kg) n Number of measurements ( ) D Diameter (m) Nu Nusselt number ( ) Pr Prandtl number ( ) R Radius (m) Ra Rayleigh number ( ) Re Reynolds number ( ) s Standard deviation (population) T Temperature (C) t Time (s) V Volume (m3) Greek Letters Coefficient of volume expansion ( K1) Average surface e missivity ( )

PAGE 18

18 Greek Letters (continued) Ratio of the radius of the contact surface to the radius of the solid particle ( ) Porosity (m3 void m3 total) Kinematic viscosity ( m2s1) Density ( kg m3) Stefan Boltzmann Constant ( 5.670 108 Wm2K4) Subscripts cp Contact between solid particles cond Conduction conv Convection D Diameter based e Effective f Fluid phase (gas or liquid) i Initial m Mean p Particle pk P ackaging r Radiative rs Radiation between solid surfaces rv Radiation through void space s Solid phase sur Surface w Wall

PAGE 19

19 Superscripts 0 Static c ontribution d Dynamic c ontribution

PAGE 20

20 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AIR TRANSPORT OF HORTICULTURAL PRODUCTS: A THERMAL ANALYSIS By William Pelletier August 2010 Chair: Khe V. Chau Cochair: Ray A. Bucklin Major: Agricultural and Biological Engineering Temperature is the most important factor in the postharvest l ife of horticultural products and temperature management is essential to reach the optimum postharvest quality. Preserving the quality of fresh fruits and vegetables requires a proper cold chain from the field to the consumers' home. Too often, transport operations are responsibl e for breaks in the cold chain, particularly in the case of air shipments. Air transport plays a major role in the global food t rade and, even during the recent worldwide economic slump the sector of air freight occupied by food products only experienced a small decline. Maintaining a proper cold chain during air transport operations is challenging since the cargo may spend extended periods of time on the tarmac, where they are exposed, with limited protection, to harsh environmental conditions. Studies o n in flight environmental conditions and their impacts on the temperature distribution in l oad s of fresh fruits and vegetables transported by air have been limited in number. However, such studies are important to develop new handling methods as well as t o predi ct and preserve quality. A thermal analysis was conducted on loads of horticultural products in single boxes and in an aircraft container ex posed to detrimental

PAGE 21

21 temperatures in laboratory conditions. Different fruit sizes and packing arrangements were used for the tests on individual boxes. These tests showed small temperature differences between the pulp and air temperatures within the boxes. Relatively fast rates of change of the temperatures were observed even i n the core region of the boxes. For the aircraft container, a similar thermal behavior was observed in the fruit near the outside surface of the load, particularly for the boxes located on the top row. Even after 8 h of exposure to detrimental conditions the temperatures within the core of the aircraft container remained almost constant. With the exception of the bottom layer, a vertical stratification of the temperatures was observed in the boxes as well as in the container. In addition to laboratory tests, temperatures were al so monitored during several international flights using an instrumented aircraft container loaded with simulated horticultural products. T hermal behavior similar to that of the laboratory tests was observed during the air transport operations. The tests showed that the ramp transfers before and after flights were critical to maintaining a proper cold chain, mainly because of the effect of solar radiation For some shipments, the ramp transfer exceeded 8 h and temperatures above 60C were also measured on the inside surfaces of the aircraft container walls. Even though temperatures were not always optimum within the aircraft cargo compartments, their effects were not as detrimental as those associated with ramp transfer s. Based on the experimental data collected through laboratory and air shipment tests the validity of a heat transfer model based on an effective thermal conductivity was investigated. For individual boxes, simulations showed that an effective thermal

PAGE 22

22 conductivity approach provided accepta ble results in the lower and lateral regions of the load but significantly underpredicted the tem peratures in the core regions. A ircraft container simulations revealed that the temperatures were also underpredicted in the top of the boxes located on the b ottom layer and at the bottom of the boxes located on the top layer. T he model did not provide a good representation of the temperature distribution throughout the load of products but did provide good results in the peripheral region. It appeared that the effect of natural convection must be included in the effective thermal conductivity via a variable dynamic component to improve the overall results of such a modeling approach. However, the res ults indicated that the modeling approach used could still be implemented as a useful tool for air shipments of horticultural products, since temperature abuses are generally observed in the peripheral region of the load, where the model provided useful re sults.

PAGE 23

23 CHAPTER 1 INTRODUCTION 1.1 Temperature and Fresh Horticultural Commodities For most people, the term "living organism" is not the first attribute that comes to mind for fresh horticultural commodities However, fresh fruits and vegetables are intrinsically living organisms; they absorb oxygen and produce carbon dioxide and water generate heat go through a senescence phase and eventually die. T heir complex metabolism depends heavily on temperature, the most important parameter in the postharvest life of fruits and vegetables A p roper temperature is essential to maintain or reach the optimum postharvest quality. A 10C increase of the temperature of a produce generally results in a n increase of the respiration rate by a factor ranging from 2 to 3 depending on the commodity, and proportionally reduce s its shelf life. Temperature also affects the postharvest quality of fresh produce by playing a role in other processes such as the growth of decay organisms and water loss (Kader, 2002 a ) Preserving the quality of fresh fruits and vegetables is important at different levels. Economically, quality is obviously a key factor particularly with the small profit margins growers, distributors and retailers have on these items. Nutri ti onally, produce of good quality have higher levels of vitamins and nutrients essential to the consumer's health. E nvironmentally 5 to 25% of all produce go to waste in developed countries because of low quality and they carry with them a n important carbon footprint accumulated from the field and through the entire distribution chain (Kader, 2002a ) D etrimental temperatures are not the only cause for the quality loss of fresh horticultural products. Other factors such as mechanical damage due to shocks and vibrations during handling and

PAGE 24

24 transport operations may also play a significant role. However, in too many instances poor or improper temperature control is at the core of the problem. Even though proper postharvest temperature management for fresh fruits and vegetables is a challenging task, only three simple rules need to be followed: the proper temperature must be applied 1) to he althy produce, 2) as soon as possible and 3) th roughout the distribution chain, from the field to the consumers' homes. These three rules set the foundations of what is known as the "cold chain", a concept that has also been adapted to most industries dea ling with temperature sensitive products (pharmaceutical, meat seafood, etc.). The expression "cold chain" may however be misleading in some cases since some commodities such as tropical produce, are likely to be damaged by temperatures below 15C. The cold chain is in essence a "proper temperature" chain. Numerous scientific studies have been conducted on the cold chain for fresh horticultural products and several of them have focus ed on transport operations during which the cold chain is often broken. Fresh fruits and vegetables are transported by road, sea, rail and air. However, among these four main modes of transportation, air transportation has v ery unique features that set it apart from the group. 1.2 Air Transport and its Cold Chain From 1970 through the first part of the last decade, the air cargo sec tor show ed good and steady growth, doubling approximately every 10 years ( Sowinski, 2000). However, this all came to a stop in recent years. In the wake of the global economic crisis, the International Air Transport Association (IATA) reports that the industry closed the year 2009 with its largest decline since the post war era Air freight showed a

PAGE 25

25 decline of 10.1% over the year and more importantly only 49.1% of the total available f reight tonn e kilometers were used ( IATA, 2010) This drop in the air transport industry affected most of the world markets including France, Germany, Taiwan, China and the United States. However, even through this major slump, the sector of air freight occupied by food products only experienced a decline i n the low single figure percentages. In the short term, this distinctive sector is also expected to show strong growth and to become of even larger importance for freight forwarders. Africa, a growing exporter of fresh produce and cut flowers, and the Middle East, a major hub for Europebound goods are the regions expected to show the largest growth rates in the coming years ( Anonymous, 2009a; Anonymous, 2009b) For several years now the air cargo industry has played a major role in the global food trade. Time savings associated with air transport make it a mode of transportation essential to provide a consistent year long supply of fresh fruits and vegetables throughout the world. It is impossible fo r produce with short shelf lif e such as berries, to still be marketable after transoceanic or long int ernational shipments using another transportation mode. Fruits and vegetables with a very short shelf life as well as highend produce are often the ones that are carried onboard aircraft. However, the deciding criterion still remains the profit margins, and for produce, profit margins go handto hand with quality ( Sharp, 1998) In the air transport industry, proper cold chain management is very chal lenging but it is still essential to preserve quality (Amos and Bollen, 1998) For almost all situations, fresh fruits and vegetables, as well as luggage, electronic products and other types of freight, are transported using Unit Load Devices (ULDs). ULDs are of two types, aircraft pallets or containers. ULDs are available in different

PAGE 26

26 standardized sizes compatible with airport handling equipment and of course specific commercial aircraft cargo holds. Shape and strength requirements were established by airplane manufacturers and the Federal Aviation Administration (Bye and Bleasdale, 1985). Aircraft pallets are large aluminum sheets, often called "cookie sheets", onto which products are loaded. Product s delivered to the airport already on skids or pall ets can be directly loaded onto the aircraft pallet. O therwise bul k products or packages are manually stacked. In all cases, loaded products are stabilized using a special cargo netting system that is fastened directly to the perimeter of the aircraft pallet. Generally a thick plastic sheet is installed underneath the cargo net, to protect the products from precipitation and other environmental hazards. Height restrictions apply to pallets depending o n the type of aircraft and the position of the palle ts within the lower and upper deck cargo compartments. Aircraft containers are nonhermetic enclosures that can be made of various materials such as aluminum, polycarbonate or fiberglass. Akinaga and Kohda (1992) reported the ventilation rate of an aircraft container to be 0.06 (air change) h1. The door on the container is generally made of sturdy waterproof fabric that is kept closed using multiple straps Similarly to aircraft pallets, their bases are made of thick alumin um. Most containers are manufactured to be used on either upper or lower deck holds and their shape is designed accordingly to maximize the use of the cargo space. P alletized or bulk product s can be loaded inside the containers. Some models of aircraft containers also offer the possibility of using a stabilizing system that fastens the load to their base. Depending on the airline handling procedures, aircraft containers may also be wrapped using a stretchable pla stic film. The film protects the container

PAGE 27

27 against ta mpering, but it restricts even more the air exchange between the container and the environment. Prior to being loaded, all ULDs are weighed to ensure they comply with their prescribed c apacity Also, a loading plan is generated according to the weights of the ULDs to insure that the aircraft is well balanced. Refrigerated and insulated containers have been available for several years now but they still do not occupy a large share of the market when compared to traditional uninsulated ULD s. In general, only highly expensive temperature sensitive goods, such as pharmaceuticals, j ustify the costs associated with their utilization. Almost all fresh fruit s and vegetables are therefore shipped with ULDs that do not provide significant additional thermal protection against the environment. For most air shipments, the time spent in flight corresponds to approximately 50% of the total transit time that includes storage and ramp transfers ( Sharp, 1998). As the cold chai n gains importance in the air transport industry, more facilities with controlled temperature storage are being built. However, in many instances, the storage capacity is limited and restricted to a single temperature. This situation, combined with a lac k of knowledge regarding the handling of temperature sensitive products, results in many claims due t o loads left at room temperature or produce being exposed to chilling temperatures. I n addition, there are often issues associated with the storage of inc ompatible loads. Incompatibilities can be associated with ethylene, odor absorption and potential microbiological contamination. During airport operations, the cargo is most vulnerable when it is on the ramp (tarmac). Prior to being loaded or after being unloaded, the ULDs can sit or transit on

PAGE 28

28 the ramp for a few hours (Amos and Bollen 1998) Some companies may use insulated covers, refrigerated dollies or shaded areas in order to protect the ULDs from the environment, but these are exceptions rather than the norm. On the ramp, the conditions are often extreme A ir temperatures can vary between 40C and 50C depending on the location and the season ( Sharp, 1998) In the summer, solar radiation also has an important effect depending on surf ace physical properties. Surfaces with high solar absorptivity may have their temperature climb close to 70 C within minutes. Cargo is transported on three types of aircraft: full freighter, combi and passenger C ombi type aircraft have both cargo and passenger compartments on their main deck. The majority of all worldwide cargo is carried in the holds of passenger aircraft (Sharp, 1998). O nboard aircraft, temperatures can be quite unpredictable. Depending on the aircraft and the options purchased by t he airline, some cargo compartment s are temperaturecontrolled whereas other s are not. In some cases, aircraft may provide multiple temperature zones whereas others may not even be equipped with a ventilation system for the cargo. Other factors may also affect the temperature. Because of the restricted empty space in the holds, load configurations may strongly influence the airflow and consequently the spatial distribution of the temperature. ULDs located in the vicinity of the door or bleed air duct, m ay be exposed to freezing and soaring temperatures respectively The presence of live animals onboard also requires the crew to maintain a temperature of approximately 20C. In addition to temperature, pressure and relative humidity also exhibit variations specific to air transport. For most commercial flight s, cargo compartments are

PAGE 29

29 pressurized, which implies that the pressure is maintained at approximately 0.8 atm at cruising altitude ( ASHRAE 1995) Relative humidity in the holds is influenc ed by temperature and pressure variations. If live animals and uncovered pallet s of horticultural products are transported, the cargo can also strongly influence the level of relative humidity by releasing water vapor into the ambient air. It is important to reiterate the fact that the volume of free air in a hold is relatively small because the ULDs' design maximizes the use of the cargo space. In nonventilated cargo compartment s relative humidity can therefore reach saturation levels whereas in temp erature controlled compartment s relative humidity can be as low as 5% ( Sharp, 1998; Pelletier, 200 2 ; Pelletier, 2007) These low relative humidity levels are reached when the dry and cold air at high altitudes is used by the air conditioning system. Because of the low air exchange within an aircraft cargo container, the relative humidity of the air surrounding the load can be quite different than the relative humidity of the outside air. 1.3 Statement of the Problem T emperature control has been recognized as a critical factor for the postharvest quality of fresh fruits and vegetables. With new affordable temperature sensor s inundating the market and the progressive integration in to the distribution chain of technologies such as radio frequency ident ification (RFID) tags, there never have been more incentives to actively track and monitor the environmental conditions in which temperature sensitive goods are transported. Studies focusing on temperature and air shipment s of fresh fruits and vegetables have been conducted but they are still in limited number s when compared to other postharvest areas such a s precooling operations Data on inflight conditions are

PAGE 30

30 particularly rare in the literature because of the difficulty of conducting such tests and the reluctance that airlines used to have about documenting this information. S everal airlines did not want to know or publicize what such studies may reveal about their operations In addition, only a few scientists have investigated heat transfer and t he temperature distribution in ULDs loaded with fresh produce. This knowledge is very important in order to eventually develop and effectively use postharvest quality models. The objectives of this study are to : 1) Investigate, in laboratory conditions, the heat transfer within close d boxes of horticultural products 2) Investigate, in laboratory conditions, the heat transfer within an aircraft container loa ded with horticultural products 3) Measure and investigate in flight env ironmental conditions in a loaded ai rcraft cargo compartments 4) Measure and investigate temperatures within a loaded aircraft cargo container during shipment 5) Evaluate the validity of an effective thermal conductivity approach to model the heat transfer in an aircraft container loaded with hort icultural products.

PAGE 31

31 CHAPTER 2 LITERATURE REVIEW 2.1 Air Transport In this section, the important literature regarding air shipments of food products is presented. First, general publications focusing on the air transport operations and logistics as well as packaging and handling techniques are presented. Secondly, publications focusing on the cold chain and the air transport of horticultural products are discussed. 2.1.1 General Publications Sha rp (1998) thoroughly presented the basics of the airfreight industry and discussed cold chain issues due to conditions within the cargo holds of the aircraft and during ground operations. The different types of Unit Load Devices (ULD s: aircraft pallets and containers) were also described by the author Furthermore, Sharp (1998) discussed the use of supplemental cooling, packaging methods and presented recommendations for safer air shipment of perishables. Th e paper published by Villeneuve et al. (2002) completes well the previous document. The authors provided numerous statisti cs and information regarding cold room storage needed for air transport of per ishables, cold room storage available in airports, market previsions, delays as well as airport logistics for perishables. Among other subjects explored by the authors were the time required for the completion of various handling operations and the temperatures encountered in aircraft cargo holds and on the tarmac. The effects of air transport on the quality of perishables were also covered. They conclu ded by discussing the future of perishable airfreight.

PAGE 32

32 Finally, the International Air Transport Association ( IATA ) publishes the Perishables Cargo Regulations (PCR). The PCR is a very detailed guide for shipment s of any perishables including fruits and v egetables. Environmental conditions and cold chain are discussed as well as suitable packaging and handling techniques. This is without a doubt the most complete source of infor mation available on the subject. 2.1 .2 Temperature and Environmental Conditions Harvey et al. (1966) were among the first to conduct a study on shipments of perishables by air. Their interest w as focused on the conditions strawberries are exposed to during their transit from California to wholesale markets in the eastern United States. Test s were conducted in the spring when temperatures are typically milder. The authors studied the effects of several variables such as time, temperature, handling procedures, precooling prior to shipment, modified atmosphere and pallet covers. The y were able to gather temperature measurement s and gas samples during flight. Con sidering the entire transit from the cooling facilities in California to their final destination in New York, 32.2% and 50% of the total time was spent onboard the aircraft and at the airport respectively Average ambient temperatures in the cargo compartment during the two flight segments were 16C and 13C. It was concluded that the pallet cover provided much needed protection against the rough handling associated with airport operations and that cold room storage at the airport would have been beneficial for the final quality of the berries. Oskam et al. (1998) reported the temperature variations of a load of freshcut flower s transit ing from Bogota to Amsterdam In less than 1 h on the tarmac in Bogota, the temperature on the top of the aircraft pallet increase d by approximately 10C

PAGE 33

33 During a ramp transfer in Curaao, the temperature at the same location rose by 15C in 2 h Durin g both of these events, t he center temperature increased by less than 5C In another effort that illustrate d the weakness of the air transport cold chain, Heap (2006) reported that during a shipment from Dublin to Auckland, transiting through Copenhagen and Singapore, the temperature s in the aircraft ranged between 2C and 18C whereas the temperatures on the ground varied between 2C and 30C. H e also showed that ambient temperatures varied between 1C and 26C during a second shipment from Europe to the North America. mond et al. (1999) measured inflight air temperatures in the forward, aft and main deck holds of a Boeing 747400 combi Their objective was to evaluate the suitability of the cargo hold temperatures for the transport of horticultural commodities. In the forward hold, the temperature was maintained around 20C for the first hour before reaching the set point of 10C. Large spatial temperature variations were observed in the aft hold. Two of the temperature loggers installed in that cargo compartment measured temperatures of 15 C and 32C within a distance of 2 m. On the main deck, the cargo compartment was maintained at approximately 20C. The authors explained that such a temperature is to be expected since the same system is used to condition the air in the passenger cabin. It was concluded that special precautions, such as the use of insulated containers, should be taken if horticultural commodities were to be shipped in the aft compartment. The authors also suggested th at the forward hold was best suited for shipment s of tropical and subtropical produce. Villeneuve et al. (2000) studied the heat transfer in an aircraft LD3 containers filled with 1L bottles of water stacked in Reusable Plastic Containers (RPC s). The load was

PAGE 34

34 conditioned to 0C prior the beginning of each experiment. D uring storage inside a cargo terminal maintained at 24C (no solar radiation), the water temperature in the external layer of RPC s increase d by more than 6C in 5 h Similar test s were cond ucted on the ramp with two types of LD3. One LD3 was made of aluminum whereas the other was made of clear polycar bonate plastic. After 2.5 h, the temperature in the external layer of RPCs reached 10.0C and 15.5C in the aluminum and clear plastic contai ner s,, respectively The results showed that solar radiation can have a significant influence on the heat transfer within the load during ground handling operations. Consequently, the material used to manufacture the walls of a container has an important impact During a flight from Sydney and London, Sharp (1989) monitored ambient air temperatures as well as the air temperature in the center of flats of strawberries on the center and lower row of an aircraft container. The container loaded with 1200 kg of fruit was covered with a 10 mm thick expanded polystyrene refl ective shroud. In addition, 20 kg of dry ice (solid carbon dioxide) wrapped in paper was placed on top of the fruit. D uring the 45h trip, 50% of the time was spent in flight. The container was outside of the plane for 30% of the total time and on the ground, onboard the aircraft for the remaining 20%. The results showed that the ambient temperature in the aircraft reached 35C duri ng ramp operations. During flight, the temperature in the cargo compartment decreased slowly at a rate of 5C h1. Once at steady state, the ambient air temperature among the different inflight segments varied between 9C and 16C. Akinaga and Kohda ( 1993) measured the temperature at different locations inside an aluminum aircraft container on the tarmac in Naha Airport (Okinawa, Japan) Over a

PAGE 35

35 period of 240 minutes, the surface temperature of the top of the container averaged around 50C but varied displaying several peaks close to 60C The air t emperature at a distance of 0.1 m below the top of the container stayed mostly above 40C and even reached 45 C. In addition to temperature, relative humidity, pressure, shocks as well as concentrations of carbon dioxide and ethylene were measured. Tests were conducted with shipments of snap beans, okras and chrysanthemums. 2.2 Heat Transfer Modeling In this section, different approaches that can be used to model heat transfer in loads of horticultural products are presented. F undamental studies on heat transfer modeling for packed beds and other porous media are reviewed first Several of these studies were published in chemical engineering journals because of their importan t applications in that field of research. A fter the presentation of main concepts, including the effective thermal conductivity, heat transfer models applied to boxed and bulk horticultural commodities as well as models directly applied to air transport are presented. 2.2.1 Models Applied to Packed Beds and Other Porous Media Packed beds are heterogeneous systems that represent the limiting case of a dense dispersion of solid particles in direct contact within a gaseous or liquid phase (Tsotsas and Martin, 1987). Packed beds may contain thousands of particles, sometimes of different sizes and intricate geometries, combined in various arrangements. Several of these studies are based on arrangements of uniformly sized spheres and cylinders, which are two el emental solid geometries also widely encountered in horticultural products. Catalytic reactors are one of many technical and industrial applications that use packed beds. F luid flow and chemical reactions in

PAGE 36

36 catalytic reactors add to the compl exity of the system. A good understanding and description of heat transfer in catalytic reactors is important since it strongly influences yield, stability, process selectivity and catalyst deactivation (Specchia et al. 1980). It can be very challenging to study transport phenomena in such systems, even using numerical methods and computational fluid dynamics (CFD) Therefore, to simplify the mathematical description of packed beds and their applications, effective properties are often used in combination with analyt ical solutions for the corresponding homogeneous domain. 2.2.1.1 Introduction to effective thermal conductivity The concept of effective thermal conductivity has been introduced to simplify the modeling process of transport phenomena. For heat transfer application it can be defined as the ratio of the total heat flux divided by the corresponding temperature gradient. The effective thermal conductivity is an average transfer parameter, not a true thermodynamic property (Gorring and Churchill, 1961; Bhatt acharyya and Pei, 1975). A few review papers were published on the subject of the effective thermal conductivity for chemical engineering applications. Kulkarni and Doraiswamy (1980) as well as Lemcoff et al. (1990) regrouped, analyzed the var ious correl ations and presented recommendations for calculating transport properties in packed beds. They cover the subjects of one and twodimensional pseudohomgeneous and heterogeneous models. Tsotsas and Martin (1987) presented the parameters influencing the ef fective thermal conductivity of packed beds. Parameters were regrouped into two main categories. The primary parameters were the thermal conductivities of both solid and fluid phases, the porosity and the arrangement of the bed. Particularly for porous bodies with a continuous solid phase, the authors mentioned that systems having

PAGE 37

37 different arrangements but the same porosity will usually have different effective thermal conductivities. The secondary parameters were the heat transfer by radiation and fre e convection, solidsolid heat transfer associated with the flattening of the solid particles near contact points, particle shape and size distribution. There are two main modeling approaches to obtain effective properties. One is based on the solution o f the coupled differential equations in both fluid and solid phases, whereas the other is based on thermal resistance network analysis. This study mostly focuses on the later approach. 2. 2 .1.2 Effective thermal conductivity models In some cases, the res istance network used to obtain the effective thermal conductivity is a drastic simplification of the packed bed or porous material structure. However, several models based their analysis on a more accurate representation of the structure of the system suc h as a unit cell. The first step in developing a model for effective thermal conductivity is to analyze how the heat is transferred through the domain. Eight different heat transfer mechanisms that can be encountered i n packed beds they are represented i n Figure 2 1. As described by several authors, the effective thermal conductivity can be represented as the sum of two independent contributions (Kunii and Smith, 1960; Vortmeyer and Schaefer, 1974; Bhattacharyya and Pei, 1975 ; Koning, 2002). The first co ntribution is static and it includes the influence of pure conduction in the fluid and solid phases as well as radiation. In this case, the motionless fluid is considered as a solid with a negligible thermal resistance on any interface. If the fluid is a gas, it is also considered transparent to radiation. The minimum temperature for which radiation heat

PAGE 38

38 transfer within the packed bed or porous media is important varies between 200C and 900C depending on the study (Argo and Smith, 1953; Yagi and Kunii, 1957; Krupiczka, 1967; Kunii and Smith 1960). Some authors mention ed that natural convection should be included in the static effective thermal conductivity, but its influence was neglected most of the tim e because of the relatively small size of the voids and the small temperature gradients in packed beds or other porous material (Woodside, 1958; Krupiczka, 1967, Gorring and Churchill, 1961). The second contribution to the effective thermal conductivity is dynamic and it includes the effect of forced convection on heat transfer. Bauer and Schlnder (1978a) confirmed the independence of the static and dynamic contributions by observing in all their measurements a linear increase of the effective thermal conductivity with the flow rate in packed beds. Bhattacharyya and Pei (1975) showed that the dynamic contribution is the sum of the effect of forced convection on conduction mechanisms and the effect of macroscopic convective heat transfer in the bed (Equ ations 21 and 2 2) = + (2 1) = + + (2 2) In the fifties and early sixties, Kunii, Smith, Yagi and other collaborators published a series of papers on heat transfer and effective thermal conductivity for packed beds and porous media. These papers laid some of the fundamental modeling approaches that were used, modified and improved through the following years (Argo and Smith, 1953; Yagi and Kunii, 1957; Yagi and Wakao, 1959; Kunii and Smi th, 1960; Yagi and Kunii, 1960, Yagi et al., 1960; Kunii and Smith, 1961)

PAGE 39

39 Argo and Smith (1953) presented a model for the radial heat transfer in a packed bed with fluid flow. They assumed that the average temperature of a particle and the gas temperat ure were the same for a given radial position. In their approach, they also assumed that the total radial heat flux is the sum of the heat flux through gas and solid phases. In the gas phase, they summed the thermal conductivity of the gas with the equiv alent radiation conductivity and the conductivity associated with turbulent diffusion. The heat transfer in the solid phase was described as the sum of a series of mechanisms: including conduction in the solid phase as well as convection and radiation heat transfer between solid particles. The authors concluded that the effects of fluid flow, particle diameter, thermal conductivity of the particles and the radial heat transfer in packed beds were satisfactorily predicted by their model. Yagi and Kunii ( 1957) considered the effective thermal conductivity of a packed bed as the sum of two components: one dependent and one independent of fluid flow. At low Reynolds numbers, the heat transfer in the bed was dominated by conduction through the solid phase, c onduction in the film of fluid near the contact surface of two solids and, in the case where the fluid is a gas, radiation between nearby solids and voids. At large Reynolds number s, the total heat transfer was dominated by the lateral mixing of the fluid phase. The authors obtained an expression for the static contribution to the effective thermal conductivity from an approach equivalent to a resistance network analysis. They concluded that their equation adequately predicted published experimental data for packed beds with motionless gases for packings of different geometries and material.

PAGE 40

40 Kunii and Smith (1960) published a paper on heat transfer of porous rocks, a consolidated porous material. Their final results were based on a preliminary model they developed for a packed bed of unconsolidated spherical particles of uniform size. Their preliminary model is the model of interest for this study. Their approach and final equation for the static effective thermal conductivity of unconsolidated particles is almost identical to what was presented by Yagi and Kunii (1957). The main difference is that, Kunii and Smith (1960) considered the conductive heat transfer in the fluid phase of the packed bed Therefore, their general model included conduction and radiation through the fluid phase in parallel with conduction in the solid phase, heat transfer through the contact surface between solid particles as well as conduction and radiation through the stagnant fluid film near that cont act surface. All radiation heat transfer was considered taking place through a no nabsorbing fluid. Equation (23 ) present s the result of their analysis. = 1 + + ( 1 ) 1 1 + + + 2 3 (2 3) In Equation 23 is the ratio of the effective distance, in the direction of heat transfer, between the centers of two solid neighboring particles to their diameter. should vary between 0.9 and 1 for most packed beds. The parameter in Equation 23 accounts for the ratio of the fluid film thickness between touching particles to their diameter. Kunii and Smith (1960) also provided a theoretical equation to calculate The predictions obtained from their equation for static effective thermal conductivity were compared with success against experimental data for a large range of particle sizes and thermal conductivities of both solid and fluid phases.

PAGE 41

41 Krupiczka (1967) presented an empirical correlation for the effective thermal conductivity of granular solids in an immobi le fluid (gas or liquid). The correlation did not take into account convection and radiation heat transfer. To obtain its final correlation, Krupiczka (1967) started by analytically solving the onedimensional heat transfer for two models: one made of cy linders ( = 0.215) and one made of spheres ( = 0.476) B oth cylinders and spheres were analyzed in a simple cubic packing arrangement. The author then studied the same models but for multidimensional heat transfer. He solved the corresponding Laplace equation using nonorthogonal series and obtained a more exact solution. Considering the fact that all solutions displayed a similar character, the author developed a simpler model that matched his theoretical solutions as well as other published experimental data. The result was a general correlation that is a function of the ratio of the thermal conducti vity of the solid and fluid phase as well as the porosity of the medium (Equation 24 ). = ( ) (2 4 ) It is important to mention that Equation 24 is valid for multidimensional heat transfer. Furthermore, Kaviany (2002) suggested that Krupiczka's correlation (Equation 2 4 ) can be used with solidsolid composites. In a series of two papers published in 1978, Bauer and Schlnder (1978a; 1978b) present ed an effective thermal conductivity for packed columns that result s from the addition of two independent parts. The first part is a function of the fluid and flow and takes into consideration the convective heat transfer effect correlated using the Pcl et number. For t he second part of the effective thermal conductivity t he author used

PAGE 42

42 Zehner's model, which handled the contribution of conduction heat transfer in both the fluid and solid phases as well as radiative effects Tsotsas and Martin (1987), Hsu et al. (1994), Hsu et al. (1995), Gupta et al. (2002) also used Zehner type models which they refer ed to as Zehner Bauer or Zehner Schlnder models. The approach used to develop Zehner type models was similar to the one adopted by Kunii and Smith (1960), but the domain of the unit cell was cyli ndrical instead of rectangular. Zehner's model also included a deformation parameter that allowed an adjustment of the solid particles from a spherical to a more elongated of flatten ed geometry In addition, Bauer and Schlnder (1978b) provided parameters for Zehner 's model that include the influence of pressure (Smoluchowski effect), oxide layer, contact surface area and nonmonodispersed packing. Neglecting the Smoluchowski effet, Zehner 's model is given by E quations 2 5 to 2 8 The coefficient accounts for the flattening of the particles and it must be obtained from experimental data. Therefore, is zero for an arrangement of perfect spheres with contact points and no contact surfaces. = 1 1 1 + + 1 + ( 1 ) (2 5) = 2 + 1 ln 1 + ( 1 ) + ( + 1 ) 2 (2 6) = 1 + (2 7 ) = 1 25 1 (2 8 )

PAGE 43

43 Hsu et al. (1994) discussed the Zehner Schlnder model and mentioned that it underpredicts the static effective thermal conductivity for large conductivity ratios (ks/kf). They postulated that the errors in the model prediction were caused by the fact that the model did not include a finite contact surface between solid particles. The authors modified the Zehner Schlnder model by including a "deformed" factor that creates the desired contact surface between the solid particles. They also suggest ed a correction for the correlation of the deformation fact or presented in Equation 2 8 (Equation 29 ) = 1 364 1 (2 9 ) Even though the authors adopted a different approach, the model obtained by Hsu et al. (1994) is in the end very similar to the Zehner type models presented by Bauer and Schlnder (1978b) and Tsotsas and Martin (1987) where particles' deformation and the contact surface between particles are accounted for by B and respectively ( Equation 25) In a subsequent article, Hsu et al. (1995) studied the static effective thermal conductivity using a thermal resistance network approach on twodimensional arrays of in line squares and circular c ylinders as well as threedimensional arrays of inline cubes. For each geometry, cases with and without contact between the objects in the arrays were studied. They found good agreement with experimental data, particularly for the model of inline touching cubes and data on packed bed of spherical particles. Bauer and Schlnder (1978b) also discuss ed the effects of the increase in porosity near a wall on the static thermal conductivity of a system. They highlight ed the fact that the higher porosity near the wall resulted in a decrease of the number of contact points and also in a modification of the geometrical structure of the bed that respectively

PAGE 44

44 contributed to decrease conduction and to increase radiation heat transfer. They concluded that sinc e the opposite effects on conduction and radiation heat transfer were of the same order of magnitude, the global influence of the change in porosity near wall on the effective thermal conductivity could be disregarded. Most studies disagree d with Bauer and Schl nder (1978b) on the subject particularly if the dimensions of the domain were not significantly larger than the particle size. Laguerre et al. (2008) studied transient heat transfer by free convection in a packed bed of spheres using two modeling approaches and experimental data. A packed bed (porous media) approach and a CFD approach were selected for the study. The experimental data was collected using hollow PVC spheres filled with a n aqueous gelatin. For the packed bed approach, the authors used a twotemperature (twophase) model. Equations 2 10 and 211 were used for the convective heat transfer coefficient at the surface of the spheres and at the wall (in presence of spheres) ,, respectively = 2 + 1 09 400 0 71 (2 10) = 1 56 (2 11) Th e effective thermal conductivity used in t he solid phase energy equation was the sum of two terms representing the conductive and radiative contributions (Equations 2 12 to 214 ) The temperature in Equation 2 14 is in Kelvin. = + (2 12) = 2 09 20 8000 (2 13) = 2 15 0 05 (2 14)

PAGE 45

45 For the CFD approach, the Boussinesq approximation was used for buoyancy. Even though both modeling approach provided results in agreement with the experimental data, the authors concluded that the packed bed approach was best. The authors also mentioned that the difference in temperature between the center and the surface of the spheres could not be neglected. According to the example from the packed bed approach presented in the article, that difference was less 0.5C. Such a difference should be negligible, considering the errors associated with convective heat transfer coefficients and the thermal properties of the system. Furthermore, if the results are compared to experimental data, the accuracy of the thermocouple readings and the error associated with their positions are both of the same order of magnitude. 2.2.2 Mode ls Applied to Boxed or Bulk Produce Baird and Gaffney (1976) developed a numerical procedure to study heat transfer in bulk loads of fruits and vegetables during forced air precooling operations. Because of the relatively high cooling rates associated wit h forced air precooling, the authors neglected heat generation and heat transfer through direct contacts between adjacent produce. As part of the model the authors first used a finite difference approach to calculate the transient temperature variations w ithin individual homogeneous spherical produce without heat generation. To calculate the transient temperature variations as a function of the position of in the bed, the authors calculated the total heat transfer between the air flowing through finite layers of independent spherical produce. As the air flows through a layer, it was assumed that its temperature was constant and that the air properties and convective heat transfer coefficient at the surface of each horticultural products was also constant and uniform. The convective heat transfer coefficient used in the model was obtained by optimizing the results of the model to experimental data

PAGE 46

46 obtained from test conducted with oranges and grapefruits. The authors considered the calculated convective heat transfer coefficient as an effective coefficient since it accounted in part for factors not included in the model such as shape factors, evaporative cooling as well as conduction heat transfer between produce. T he authors nondimensionalized the effect ive heat transfer coefficient using the Nusselt number to obtain a correlation in function of the Reynolds number (Equation 215) = 1 17 (2 15 ) Equation 21 5 was based on fruit diameters between 0.072 and 0.107 m and air velocities ranging between 0.05 to 2.03 m/s. Talbot (1987) mainly studied the threedimensional velocity and pressure fields during the forced air precooling of bulk oranges and oranges packed in containers. He found that the change in porosity near the walls of a container was important for the flow analysis. Talbot (1987) used the model developed by Baird and Gaffney (1976) to describe the heat transfer in the studied systems. The main dif ference between the two studies is that Baird and Gaffney (1976) assumed a onedimensional plug flow whereas Talbot (1987) used a threedimensional flow with the velocity and pressure distributions calculated from a finite element porous media flow analysi s. Bellagha and Chau (1985) improved the model introduced by Baird and Gaffney (1976) by including in the numerical procedure the contribution of heat generation and evaporative cooling, due to transpiration at the surface of produce. Bellagha and Chau (1 985) used their model to study heat and mass transfer during the cooling of tomatoes individually and in bulk. The authors assumed constant homogenous thermal properties for the fruit, except for the respiration, which was a function of the temperature. The

PAGE 47

47 heat transfer through contact surfaces between fruit was neglected. The model included the heat transfer within each fruit as well as between the fluid and the fruit. Good agreement was obtained between model simulations and experimental data for bot h individual and bulk tomatoes of different sizes. The authors also found that transpiration and respiration had negligible effects for rapid cooling of tomatoes. To simulate cold room cooling of horticultural products, Bazan (1989) (see also Bazan et al. (1989) ) developed a three dimensional numerical heat transfer model for a simple cubic packing arrangement of spherical fruit in closed boxes A simple cubic arrangement was selected because of the porosity and the number of contact points on a fruit are almost the same as for random packing. The model included the effect the of heat conduction in the air and through the contact surface between fruit as well as the effect of natural convection. The thermal contact resistance between the fruit was neglected. The contribution of evaporative cooling at the surface of the fruit was neglected but the internal heat generation associated with the metabolism of the fruit was included as a function of the temperature. The domain was discretized in cubic volume element s, each containing one fruit. Each fruit was also divided into three concentric spherical shell elements. To model the effect of natural convection the author used the Boussinesq approximation combined with the Darcy flow through porous media. Equation 216 was used to describe the heat transfer between the fruit surface and the surrounding air (Holman, 1986). = 2 + 0 43 (2 16) The author obtained good results with the model. In addition to the simple cubic packing arrangement the model was also used to predict the temperature in a random

PAGE 48

48 packed boxes. Using adjusted values of contact area and packing density, good results were also obtained. Experimental and numerical results corroborated the fact that very sm all temperature gradients exist within the fruit. In the core region of the bin, the surface temperature gradients were found to be small which increased the influence of the conductive heat transfer through the contact surface between fruit. In his dissertation, Beukema (1980) studied the effects of natural convection on heat and mass transfer during the cooling and storage of agricultural products. Beukema's work was also summarized in a series of two papers, Beukema et al. (1982) and Beukema et al. (1983). Beukema (1980) presented different heat and mass transfer models. First, the author dis cussed a general onedimensional model as well as one and twophase models to describe heat transfer in porous medium with pervious boundaries in the direction of the flow. Secondly, the author introduced the heat and mass Storage with Natural Convection model, or SNC model. The heat transfer part of the model is a cylindrical twophase model, twodimensional in the product phase, and onedimensional in the air phase. Among the assumptions that were made, the SNC model neglected the conduction in the air phase as well as the accumulation of energy in the air in comparison to the convective energy transport (pseudo steady state). The heat transfer through the contact surface of two agricultural products was also assumed to be negligible. The effects of heat generation and evaporative cooling at the surface of the produce were included in the model. The author used a static thermal conductivity in the axial direction whereas for the radial direction the thermal conductivity was given as the sum of the sa me static thermal conductivity and a dynamic component function of the air velocity. Very few details were provided

PAGE 49

49 regarding the calculation of the static thermal conductivity used by the author. Even though the heat diffusion term was not present in the energy equation of the air phase, and that the author specifically mentioned otherwise, it seems highly probable based on the available information that the heat conduction in the air was included in the static thermal conductivity used in different models of the study. Equations 2 17 and 218 are correlations from Bird et al. (1960) tha t were used by Beukema (1980) for the convective heat transfer coefficient between the solid and fluid phase. = 2 27 ( 1 ) ( 13 < < 180 ) (2 1 7 ) = 1 27 ( 1 ) ( > 180 ) (2 18) The author solved the transient coupled heat and mass transfer governing equations using finite difference techniques and analyzed the influence of different parameters. Among their findings was the relatively large influence of the porosity on the temperature distribution and the buoyancy driven flow. Model simulations and experimental data with potatoes and artificial products agreed well. In addition to the SNC model, Beukema (1980) developed a transient heat transfer threedimensional natural convect ion model in a porous medium with heat generation for a closed rectangular container. The onephase heat transfer model did not include the effect of produce transpiration. The Boussinesq approximation and the Darcy term were used in the momentum equatio n. The solutions of the model for the temperature and velocity fields were obtained using numerical methods. Again, good agreement with experimental data was obtained.

PAGE 50

50 The solid phase arrangement was not taken into account in any of the models presented by Beukema (1980) since he did not consider any heat transfer through the contact surface between produce. Tanner (1998) and Tanner et al. (2002a; 2002b; 2002c) developed a generalized computer based model to predict cooling rates and mass loss of various horticultural products and packaging systems. The model also predicted local inpack relative humidity and packaging material moisture content. For their model, the domain of the system being investigated was divided into several rectangular zones. Each zone was then represented by a combination of three main components: product, air (cooling medium) and packaging. Submodels were established from energy and mass balances to describe the energy and mass content of components within a zone as well as for intra and inter zone transfers of energy and mass. For their heat transfer model, the contact resistance for product product and product packaging contact was assumed to be 0.015 W/(m K). Regarding nat ural convection, the author mentioned that for typical fruit packages and cooling conditions, the Raleigh number was less than 2000 which indicates that natural convection is small. Consequently, the effect of natural convection could be described as pure conduction in the fluid phase (Tanner, 1998; Holman, 1986). In order to estimate the fluid velocity profiles in different packaging systems, Tanner (1998) used a CO2 trace pulsing technique. The heat transfer model was tested using experimental cooling data for apples and pears with different packaging configurations. The experimental data were collected from different independent sources. For most simulations, the authors defined the zones as volumes that enclosed individual fruit. In general, the model provided good results except for

PAGE 51

51 palletized boxes for which differences between the model predictions and the experimental data were more important. Tanner (1998) concluded that the observed discrepancies were most likely caused by inaccurate model in put data. Alvarez and Flick (2007) wanted to develop a model that could eventually be used to study forced air precooling of fresh fruits and vegetables. The authors chose to use a semi empirical macro porous media approach to describe twodimensional tur bulent flow and heat transfer is stack of spheres. Among the assumptions made by the authors were an isotropic porous medium domain, onedimensional radial temperature variation in the solid spherical products and negligible evaporative cooling. After obtaining the velocity and turbulence intensity in the domain, the authors numerically solved the heat transfer equation for the product and the air using a finite volume method. The experimental and simulated results were in good agreement. 2.2.3 Models Applied to Air Transport Villeneuve et al. 2001 developed a mathematical model to predict transient heat transfer in a LD3 aircraft container. The authors neglected air movement inside the container and model ed the load inside the container as pure conduc tion inside a porous medium. The thermal conductivity was weighted as a function of the respective volumetric fraction the boxes, products and air inside the container. The authors modeled the radiative heat transfer on the outside of the container by considering direct and diffuse solar radiation taking into consideration the geographical latitude and longitude, the elevation above sea level, solar hour angle, solar declination, solar azimuth, wall azimuth and inclination, sky point of cloudiness, optica l and thermal properties of the out side surface of the ULD. The outside convective heat transfer coefficient was calculated from the wind speed and direction and the Nusselt n umber

PAGE 52

52 correlation for laminar flow over a flat plate Results from the model were compared with experimental temperature data obtained from an aluminum and also a clear plastic container The solar radiation component was closely predicted by the model and also recognized to be a major factor during ramp operations. They also concl uded that their conduction model underestimated the heat transfer in the core of the load and that further investigation was needed. Through collaboration with KLM Cargo, Oskam et al. (1998) conducted a study to develop a model for the environmental condit ions influen cing the quality of perishables during flights, particularly fresh cut flowers For their pure conductive finite element transient model with uniform heat generation they used a weighted average for the density and the specific heat as well a s an effective thermal conductivity. Stating symmetry, they selected half of the aircraft pallet of flowers as the domain. They assume d convection and radiation on the top and sides of the aircraft pallet and neglected heat transfer at the bottom and at the plane of symmetry. Neglecting the heat transfer on the bottom surface of a pallet could be considered as a questionable assumption since the boxes of flower s sit directly on a highly conductive aluminum pallet It was reasonable to assume symmetry wi th respect to the center of the pallet because their model was developed for a Boeing 747 aircraft, in which the air was delivered symmetrically from the ceiling of the hold. However, for several Airbus models air delivery and return are positioned on the opposite sides of the cargo compartment, precluding the use of symmetry as a boundary condition (mond et al. 1999) In order to test their model, Oskam et al. (2001) also gathered experimental temperature data in an actual flower shipment They noticed the important effect of

PAGE 53

53 solar radiation during ramp operations. Their results also showed that the model they developed underestimated the heat transfer within the load and that an adjustment on the effective thermal conductivity could be beneficial. Th e authors did not directly provide any information regarding the calculation of the effective thermal conductivity used in their model. However, the authors referred to an article by Wang et al. (1995) in which the effective thermal conductivity was calcu lated from the effective heat conduction area. Wang et al. (1995) mentioned that their calculations took into account the average diameter and length of the stem, the average diameter, thickness and number of leafs per flower as well as the number of flow ers in the box. No further details were provided. Amos and Bollen (1998) developed a mathematical heat transfer model to predict transient temperature variations within an aircraft pallet loaded with crates of asparagus. These crates had a trapezoidal sha pe and therefore create d large air gaps in the load. The authors selected a multi zone approach instead of a simpler conductive model in order to effectively take into account the effect of natural convection in the air gaps between crates. Product zones were considered as lumpedsystems with a n overall heat transfer coefficient based on the thermal conductivity of adjacent zones, geometry and convective heat transfer coefficients. Effective specific heat and density were calculated based on the proporti on of asparagus and air in the product zones. Heat generation was also included in their model. Air zones were also considered as lumpedsystem but without any heat generation. Heat transfer between adjacent air zones was calculated using an empirical e ffective thermal conductivity based on natural convection within enclosures. The model predicted well the temperatures in lower and

PAGE 54

54 center region of the load but underestimated the temperatures in the upper regions. The model showed a high sensitivity wi th respect to the heat generation term. To complement their heat transfer analysis, the authors also discussed the quality and shelf life of the asparagus and the economics associated with the use of different insulated covers and refrigerant s. 2.2.4 Comp utational Fluid Dynamics Instead of using effective properties to represent transport phenomena in a packed beds or porous media, CFD models use a direct approach to numerically solve the governing momentum, energy and mass equation through the entire domain or through a periodic unit cell. CFD models require significant computer power, particularly in the case of packed bed and porous media for which intricate geometries must be meshed. In the case of large systems, CFD models are still not convenient for practical or commercial applications, even considering the latest innovation in computer technologies. However, such models are still important and useful to scientists to get a better understanding of the physics and therefore to calculate and develop bet ter effective property models Several studies focusing on agricultural, food or chemical applications have been published on the subject and among them are those by Logtenberg and Dixon (1998), Logtenberg et al. (1999), Nijemeisland and Dixon (2001), Verboven et al. (2001), N ijemeisland and Dixon (2004), Verboven et al. (2004), Dixon et al. (2005), Gunjal et al. (2005), Verboven and Nicola (2005), Chourasia and Goswami (2006), Smale et al. (2006), Verboven et al. (2006) and Delele et al. (2008).

PAGE 55

55 Figure 21. Heat transfer mechanisms in packed beds.

PAGE 56

56 CHAPTER 3 MATERIAL AND METHODS 3.1 Laboratory Tests The main objective of the laboratory tests was to measure air and pulp temperatures at several locations in loads of refrigerated horticultural products exposed to a stepwise change in ambient temperature. Cold fruit were moved to a warmer environment for a fixed period of time and then moved back to the cold r oom. This was done to simulate situations where products in cold storage are taken out on a loading dock or other nonrefrigerated environments. Oranges were selected for the tests because of their availability, their geometr y and also their extended shelf life, even in harsh testing conditions. Laboratory tests were conducted on two levels: on single boxes of oranges as well as on a loaded aircraft container (with boxes of oranges) D etailed temperature mapping was done at the singlebox level whereas coarser mapping was applied to the aircraft container. 3 .1.1 Monitoring Equipment Air and pulp temperatures were monitored using 3.66 m long 24 AWG Special Limits of Error (SLE) type T thermocouple wire ( Omega Engineeri ng, Inc. Stamford, CT) A CR10X data logging system equipped with a CR10XTCR thermocouple reference thermistor was used to read and record the t emperature measurements U p to three AM416 relay multiplexers, each with 48 single ended thermocouple inputs, and one SM4M storage module (two million data points) were also used to increase the monitoring and data storage capacity of the system ,, respect ively The data logging system as well as the multiplexers and storage module were all manufactured by Campbell Scientific, Inc. ( Logan, UT ). The accuracy of the thermocouples with the data

PAGE 57

57 logging systems was 0.5C. The accuracy of the thermocouples was verified in an agitated ice bath before each experimental set up. 3 .1.2 Single Box Tests For all tests, the experimen tal set up presented in Figure 31 was used. The data logging equipment was placed inside an insulated box to reduce the heat transfer and to minimize temperature gradients on the equipment and ensure good thermocouple readings A plastic film was installed directly on top of the wooden pallet (1.016 m by 1.219 m) as well as on all sides and top of the experimental set up in order to block the convection associated with the ventilation system of the refrigerated rooms. Two refrigerated rooms ( Environmental Growth Chambers, Chagrin Falls, OH) were used for the tests. Each of the room s was programmed using the TC2 microcontroller system to provide a constant temperature. One of the rooms was maintained at a low temperatures whereas the other was maintained at a high temperature. Temperature set points varied slightly between the different series of tests but remained constant throughout a given series. At the beginning of the test, the experimental set up was swiftly moved (approximately 1 min) from the cold chamber to the warmer one, exposing the experimental set up to t he desired stepwise change in temperature. The experimental set up was maintained in the warmer environment for a period of 7 h for the fir st series of tests and 15 h for the second and third series. The box of produce was then cooled down until the pulp temperatures were uniform and steady. Three replications were completed for each series of tests. For all measurements of pulp temperatures, the thermocouple was fi rst f ed entirely through the fruit and then its tip was reinserted at the desire d position in the fruit. This

PAGE 58

58 technique helped to secure the thermocouple in place and also decreased the potential reading errors caused by heat conduction along the thermocouple wire between the tip and the section of the wire exposed to the air. 3 .1.2.1 First se ries of tests Produce and packaging. Commercially harvested Hamlin oranges were used for these tests. A total of 125 oranges were patternpacked in a vented full y telescopic cor rugated fiberboard box (Figure 32). Five layers of 25 fruit were packed in a staggered arrangement as shown in Figure 3 3. The box had a length of 0.429 m, a width of 0.266 m and a height of 0.247 m (interior dimensions). The patternpacked fruit occupied the entire interior volume of the b ox. For this series of tests, the size of all 125 oranges was determined by measuring their diameters in two different orientations with respect to the stem end axis: parallel and perpendicular. All diameters were measured using a digital caliper, model 500196, manufactured by Mitutoyo (Aurora, IL). The accuracy of the caliper was 0.02 mm The mass of each orange was also measured using a Scout Pro SP 2001 digital scale (Oha us Corporation, Pine Brook, NJ ) The scale had an accuracy of 0.0001 kg. The bulk porosity of the oranges was obtained by subtracting the total volume of the 125 oranges (based on their average dimension) from the internal volume of the corrugated fiberboard box. The average diameter of the oranges was calculated at 0.0668 m (s = 0.0025 m) the average mass at 0.1549 kg (s = 0.0158 kg) and the bulk porosity at 0.315. Instrumentation and methodology A total of 100 thermocouples was used to monitor the temperatures during the test s. For each wall of the box, a thermocouple was af fixed using aluminum tape on the internal and external surface. Thermocouples

PAGE 59

59 were positioned in the center of each of the surface. The height of the thermocouple corresponded to the center of the third layer of fruit. In addition to the box surface tem peratures, a thermocouple was also used to measure the air temperature approximatel y 0.03 m from the external surface of each wall of the box. In order to measure the air temperature near the bottom surface, the box was positioned on two blocks of wood wi th height and width of approximate ly 0.09 m and 0.04 m, respectively (Figure 3 1). Pulp temperatures were measured at a half radius distance. Every other orange was instrumented in layers 1, 2, 4 and 5, whereas all oranges in layer 3 were instrumented. There were three fruit (1, 63 and 125) for which the temperatures were not only measured at the half radius, but also at the center of the fruit and in the air surrounding the fruit. Figure 34 identifies the instrumented fruit, shows the location of th ermocouples and also defines the orientation of the box with respect to the x and y axes. The origin of the coordinate system was located at the intersection of the front, bottom and left surfaces (Figure 34). The z axis, whi ch is not represented in Figure 3 4, ha d its positive direction pointing from the bottom surface to the top surface. The identification number of the oranges increased in the positive direction of the x, y and z axes. The layer number also increased from 1 to 5 in the positive z direction, layer 1 being the bottom layer and layer 5 the top layer. Temperatures were recorded at intervals of five minutes. For this series of tests the stepwise temperature change was from 2.5C to 35.0C. 3 .1.2.2 Second series of tests Produce and packaging. For this series of tests, store bought Navel oranges were used. A vented full y telescopic corrugated fiberboard box was filled with a total of

PAGE 60

60 42 fruit distributed in three layers according to the stagger ed arrangement shown in Figure 35. In this case, the interior dimensions of the box were 0.380 m of length, 0.270 m of width and 0.244 m of height. The patternpacked fruit did not occupy the full volume of the box; there was a small gap of 0.03 m between the top of the third layer of fruit an d the interior surface of the lid. The dimensions and mass es of the fruit, as well as the bulk porosity were obtained using the same equipment and method as for the first series of tests (Section 3 .1.2.1). The oranges had an average diameter of 0.0834 m (s = 0.019 m) an average mass of 0.2605 kg (s = 0.0209 kg) and a bulk porosity of 0.425. In strumentation and methodology. A total of 80 thermocouples were used for this series of tests. On the exterior and interior surfaces of each wall, thermocouples w ere again installed and kept in place using aluminum tape. In addition, on each of the four lateral surfaces of the box, a thermocouple was installed in the air gap between the bottom part of the box and the lid (Figure 32). The external air temperature was measured at approximately 0.03 m from each of the six surfaces of the box. Wall surface temperature and external air temperature were all measured in the center of each surface (including bottom surface). Figure 3 6 shows the x and y coordinate axes used to describe the system as well as the identification numbers of the fruit and the positions of the thermocouples located in the pulp of the fruit or in the air surrounding the fruit. As was the case for the first series of test s, the origin of the system of coordinates was located at the intersection of the front, left and bottom surface s T he z axis, which is not shown in Figure 36, pointed in the direction going from layer 1 (bottom layer) to

PAGE 61

61 layer 3 (top layer). As a result, fruit identification numbers increased in the positive dire ction of any coordinate axis. The pulp temperature was measured at the half radius location of each of the 42 fruit. For fruit 1, 23 and 42, pulp temperatures were measured at the half radius lo cation at the center and approximat ely 0.001 m underneath the surface. These three thermocouples were positioned rectilinearly as shown in Figure 36. In addition to the three pulp temperatures, the temperatures of the air surrounding fruit 1, 23 and 42 were measured. Air temperatures just below fruit 35 and 40 and just above frui t 34 were also measured (Figure 36). Temperatures were recorded at intervals of five minutes. For this series of test s the stepwise temperature change was from 3.0C to 25C 3 .1.2.3 Third series of tests Produce and packaging. The corrugated fiberboard box used in the s econd series of tests (Section 3 .1.2.2) was used again but this time a smaller size of Navel oranges was selected. A total of 72 fruit evenly distributed on four layers were needed to fill the box in the staggered arrangement shown in Figure 3 7. The dimension and mass of each orange, as well as the bulk porosity were obtained using the same equipment and method as for the previous series of tests The average diameter, mass and bulk porosity were measured to be 0.0728 m (s = 0.027 m) 0.1879 kg ( s = 0.0178 kg ) and 0.420,, respectively Instrumentation and methodology. Temperatures were monitored by 96 thermocouples. Wall surface temperatures were meas ured on the inside and outside of the box with thermocouples affixed in the center of each surface using aluminum tape. The external air temperature was once again measur ed at a distance of approximat ely 0.03 m from the center of each face of the box.

PAGE 62

62 Fig ure 38 shows the fruit identification numbers and a system of coordinates that was defined as for the first and s econd series of tests (Section 3.1.2.1 and 3 .1.2.2). In addition, the locations where pulp and air temperatures were measured inside the box are indicated in Figure 3 8. The pulp temperature of eac h fruit was measured at the half radius location Fruit 1 was the only fruit for which the pulp temperature was monitored at multiple locations. For this fruit, thermocouples were install ed at the center, at a half radius distance and just under the surface of the fruit ( Figure 38 ) Inside the box, air temperatures were measured near the sur face of fruit 1, 43, 45 and 72. A time interval of five minutes was selected to monitor the temperatures. A cold room at 1C and another at 25C were used to create the stepwise change in ambient temperature for the tests. 3 .1.3 Aircraft Container Tests 3 .1.3.1 Produce and packaging The aircraft container that was used for this series of test s was a LD3 ('LD' stands for lower deck) also known as an AKE (without forklift holes at its base). The different parts of the container and its internal dimensions are presented in Figure 39. The internal volume of an LD3 container is approximately 4. 15 m. The container manufact ured by Alcan (Singen, Germany) had an aluminum frame and all walls except for the front wall, where the door is located, were made of thin aluminum sheet s (0.0005 m). On the front wall, the door was made of sturdy waterproof fabr ic and the remaining part was made of thin, clear polycarbonate plastic. The bottom part of the container was made of a 0.013 m thick sheet of aluminum. This type of aircraft container is widely used because it can be flown on cargo, combi and passenger aircraft and also because it is compatible with widebody aircraft

PAGE 63

63 from different manufacturers including Boeing, Airbus, McDonald Douglas and Lockheed. It is mainly because of its popular ity in the industry and its geometrical compatibility with the experimental load that the LD3 aircraft container was selected. For the tests, the aircraft container was loaded with a total of 88 boxes of oranges. The boxes and the fruit were the same as those described in Section 3 .1.2.1. Each box contained 125 fruit in a staggered arrangement of 5 layers (Figure 3 3). The fruit dimensions and weight s were measured using the same method as for the single box tests (Section 3. 1.2.1). Their average diameter and mass were meas ured to be 0.0683 m (s = 0.0024 m) and 0.15 98 kg (s = 0.0144 kg) ,, respectively The bulk porosity was calculated at 0.290. 3 .1.3.2 Instrumentation and methodology The temperature distribution inside the aircraft container was monitored by a total of 144 thermocouples. Inside the load of fruit, t he pulp and air temperatures were monitored. Inside each of the 88 boxes, the pulp temperature of fruit 63 located in the center of the box (Figure 3 4) was measured. Pulp temperatures were also monitored at 32 additional locations in the load as shown in Figure 3 10. All boxes of oranges were placed such that their orientation (front, back, left and right sides as defined in Figure 3 4) matched the orientation of the aircraft container (Fi gure 310). All thermocouples used to measure pulp temperature s were installed at a half radius distance to the left side (negative xdirection in Figure 34) of the center of the orange. Outside the load, the surface temperature of the inside walls of t he container, the air temperature between the walls of the container and the load as well as the surface temperature of the exterior of the load were also monitored (Figure 3 10). For the bottom interior surface of the container, one thermocouple was located at the center of

PAGE 64

64 the surface underneath the load. For the interior top surface, one thermocouple was again placed at the center of the surface and another in the air gap between the load and the top of the container. Thermocouples used to measured s urface temperatures were maintained in pla ce with aluminum tape. T here were six layers of boxes in the container (Figure 3 10) Because of the unique geometry of a lower deck aircraft container, the first two layers contained 12 boxes whereas there were 16 boxes in the four remaining layers. The space in the bottom left corner was filled with a wooden wedge structure that allowed uniform stacking (Figure 311). The air gaps between the load and the walls of the aircraft container varied from 0.05 to 0. 10 m. The wedge structure was filled with insulated foam to minimize heat transfer at this location. A refrigerated marine container was used as a temperaturecontrolled room to accommodate the large size of the aircraft container. The aircraft container was positioned such as its front side (Figure 310) faced the doors of the marine container. Even though the aircraft container was never moved during the tests, it was set on a standard wooden pallet (1.016 m by 1.219 m) in case of that eventuality. The 20 foot insulated marine container manufactured by Shanghai Reeferco Container C o. LTD (Shanghai, China) was equipped with a t beam floor and a Thermo King Magnum refrigeration unit (Thermo King, Minneapolis, MN). The refrigeration unit can maintain a constant temperature with in the range of 35C to 30C and had a refrigeration capacity of 11.2 kW at 1.7C. The initial goal of this series of tests was to expose the fully loaded container to a stepwise change in temperature as was done for the sing le box tests. However, since the loaded aircraft container could not be moved easily the change in environmental

PAGE 65

6 5 temperature was created by changing the temperature set point of the refrigeration unit from 3C to 30C. For each test, it took 30 min for the supply air of the refrigeration unit to change from 3 C to 30C. For the series of tests conducted with that experimental set up the aircraft container loaded with boxes of oranges was exposed to 30C for a period of 8 h. Three replications we re conducted Between the tests, the fabric door of the aircraft container was opened and a fan was used to help cool down the load to 3C. Tests st arted only when the temperatures in the load were uniform and steady. The data acquisition system was located inside the marine container in an insulated box. A communication cable connected to the data acquisition system was fed outside the marine container through a drain hole to monitor the progress of the experiments and to collect data. 3.2 Air Transpor t Tests The obj ective of the air transport tests was to gather inflight pressure and temperatur e data in a fully loaded LD3 aircraft container to complement the data already obtained in the laboratory with the same type of aircraft container. The tests w ere conducted on six international flights in collaborat ion with a partner airline. As can be imagined, there were several limitations associated with conducting such experiments. First, real horticultural products could not be used because of issues wit h customs clearance and agricultural inspections associated with this type of commodity. Using horticultural product s would have caused delays and put the experiments at risk Also the boxed products were used for multiple flights as being well as loaded and unloaded on several occasions to collect experimental data. Therefore, rough handling conditions associated with air transportation were expected to cause mechanical damage to the

PAGE 66

66 fruit and their typical packaging system (corrugated fiberboard boxes) once again potentially hindering the experiments. 3 .2.1 Product and Packaging System For the in flight tests, it was decided that the best solution was to use commercially available water bottles rather than horticultural products The bottles that wer e selected (Aquapod, Zephyrhills Water, Wilkes Barre, PA) held 325 mL of water and had an orblike shape that resemble a fruit (Figure 312). The mass of 72 randomly selected water bottles was measured and the average was calculated at 0.3445 kg. Their aver age volume was measured at (35 1) 101 mL. The volume of the water bottles was obtained by volume displacement in a 2000 mL graduated cylinder (Nalgene, Rochester, NY). Because of the harsh handling environment associated with air transportation the water bott les were packed in collapsible reusable plastic c ontainers (RPCs) instead of corr ugated fiberboard boxes (Figure 3 13). The RPC s (model Smart Crate GP 6419) were manufactured by IPL ( Saint Damien, Qu bec, Canada). The internal and external length, width and height as provided by the manufacturer were, respectively 0.5746 m, 0.3746 m, 0.1868 m and 0.6000 m, 0.4000 m, 0.2030 m. An interesting feature of the RPC is that they interlock and offer better stability than traditional fiberboard boxes. To further increase the stability of the load and minimize shifting during handling operations, the RPCs in each layer of the load were tied together using cable ties. The mass of an empty RPC was 1.74 kg. Each RPC used in the test contained 72 water bottles distributed in a staggered arrangement with th ree layers of 24 units (Figure 3 14). Each RPC filled with water bottles h ad a total mass of approximately

PAGE 67

67 26.5 kg. Using the internal dimensions of the RPCs and the total volume of the 72 water bottl es, the bulk porosity within a single RPC was calculated at 0.383. The calculations accounted for the presence of air in the water bottles. A total of 50 RPCs (1325.0 kg) were needed to fill the aircraft container. The LD3 aircraft container (Driessen, The Netherlands) used for the tests had the same dimensions as the one used for the tests conducted in laboratory (Section 2.1.3). The maximum net load of the container was 1508 kg. The only difference between the two aircraft containers was that the lef t section of the front side of the container used for the inflight tests was made of aluminum instead of clear polycarbonate (Figure 29). However, because of the small thickness of the of the material and the fact that the container used for the laborat ory tests was not exposed to solar radiation, it is reasonable to state that two aircraft containers had similar thermal behavior. 3 .2.2 Instrumentation To monitor the temperatures two types of sensors were used. The interior surface of the walls of the aircraft container as well as the air temperature inside the load were monitored using temperature probes TMC6HD connected to HOBO data logger s U12 013 ( Onset Computer Corporation, Bourne, MA) The accuracy of the temperature probes (thermistor s) was 0.3C. The HOBO data logger (U12013) also had an internal temperature sensor (thermistor) with an accuracy of 0.4C and a relative humidity (RH) sens or with an accuracy of 2.5% between 10 and 90% RH. The internal sensor of the HOBO was used to monitor the relative humidity and temperature of the air. HOBO data loggers have an operating range of 20 to 70C. To monitor the equivalent pulp temperat ures of fruits or vegetables, sensors had to be installed inside the water bottles and therefore be in direct contact with the water.

PAGE 68

68 Furthermore, to avoid leaks during transport the temperature sensors were confined within the water bottles. It is important to mention that during air transport, pressure variations can contribute significantly to container leaks, which can eventually cause significant damage to the aircraft. To avoid this problem, thermobutton type sensors were used (model M40P85, Alternative Technologie Pharma Inc., Laval, Qubec, Canada). These small battery like cylindrical sensors, with a diameter of 0.0174 m and a thickness of 0.0059 m are water, shock and pressure resistant, which made them perfect for this specific application. The thermobutton sensors used had an accuracy of 0.5C and could operate between 40 and 85C. To keep sensors from sinking or moving freely in the bottle of water, they were individually attached to a small piece of nylon tubing using aluminum tape an d then inserted in the bottle in such a way that the tubing was compressed between the c ap and the bottom of the bottle (Figure 3 15). S mall differences in sensor locations between the different bottles of water had only negligible effects on temperature readings because of the convection inside the bottles. The accuracies of the temperature probes and the thermobutton sensors were verified in an agitated ice bath before the test s were conducted. In each of the 50 RPCs used to load the aircraft container, a water bottle instrumented with a temperat ure sensor as shown in Figure 315, was placed as close as possible to th e center of the RPC Figure 316 shows the orientation of the boxes, the coordinate system and the position of the temperature sensor in bottle 35 in the second layer of the RPC The origin of the coordinate system was located at the intersection of the front, bottom and left surfaces (Figure 316). The z axis, whic h is not represented in Figure 316, had its positive direction pointing fr om the bottom layer to

PAGE 69

69 the top layer of the RPC. In F igure 316, the identification numbers of the water bottles increase in the positive direction of the x y, and z axes. The layer number also increased from 1 to 3 in the positive z direction, layer 1 being the bottom layer and layer 3 the top layer. Inside the aircraft container the 50 RPCs were loaded in seven layers as shown in Figure 317. E ach of the sensors located at the center of a RPC in Figure 317 is in fact located i n bottle 35 as shown in Figure 3 16. T he orientation of each RPC (front, back, left and rig ht sides as defined in Figure 316) matched the orientation of the aircraft container (Figure 317). Figure 3 17 shows on layer 4, the locations of the four temperature probes attached to the inside surface of each of the lateral wall s. These probes w ere attached with aluminum tape in the center of each wall at a height corresponding to the center of the RPCs forming layer 4. Probes were also affixed in the center of the bottom and top surface s inside the aircraft container. A wedge made of layered sheets of expanded polystyrene was used in the bottom left corner of the container to allow a uniform stacking of the load and to minimize heat transfer at that location (Figure 318). The gaps between the front, back, left and right side wall of the container and the load were 0.125 m, 0.125 m, 0.07 m and 0.25 m ,, respectively There was a space of 0.13 m between the top layer of RPCs and the top of the aircraft container. A pressure sensor (HOBO Pressure, Onset Computer Corporation, Bourne, MA ) with an accuracy of 0.01 atm between 0.03 and 1.13 atm was placed in RPC 50 (Figure 3 17). Temperature, relative humidity and pressure data were monitored at synchronized intervals of 10 minutes T o retrieve the data collected by the different sensors, the

PAGE 70

70 entire load had to be broken down and each sensor individually connected to a portable computer. 3 .2.3 Flights Data were collected in 2007 on a total of six international flights. Flights details are presented in Table 31; they were conducted through four different airports. JFK: New York, U.S.A., John F Kennedy International Airport GOT: Landvetter, Sweden, Gothenburg Landvetter Airport DXB: Dubai, U.A. E., Dubai Intern ational Airport NBO: Nairobi, Kenya, Jomo Kenyatta International Airport JFK and DXB are both among the top twenty of the world busiest cargo airports, while NBO airport is a major hub for the transport of fresh produce and cut flowers from Africa. T he re was minimum flexibility regarding the schedule of the flights, the type of aircraft used and the position of the LD3 inside the cargo compartments The position of the LD3 was determined by the airline according to the ir analysis of the payload weight distribution. It was impossible to request a specific position. However, it is important to highlight the fact that having the permission to collect in flight environmental data with a fully loaded aircraft container is something extremely rare and there fore the results should be considered as very valuable experimental data. Figur e 3 19 complements Table 31 and shows the ULD positions for the three models of aircraft that carried the instrumented container during inflight tests. The configurations pr esented in Figure 3 19 vary according to the number of aircraft containers or pallets in the shipments. Between the flights the experimental container was brought inside the cold storage facilities at the DXB and NBO cargo terminal that were both maintaine d at 5C. Data were downloaded after the following flights: JFK GOT DXB (04 292007), NBO DXB

PAGE 71

71 (05 042007), NBO DXB (05 072007) and DXB GOT JFK (05 12 2007). After each of these flights (except for the last one), the RPC's filled with water bottles were laid on the floor to enhance the heat transfer and obtain temperatures as uniform as possible before the next flight.

PAGE 72

72 Figure 31. Experimental set up used for all singlebox tests. Figure 32. Bottom part and lid of a full y telescopic box (vents not shown). Figure 33. Staggered arrangement of the oranges for the first series of test s. A) Top view of layer 1, 2 and 3. B) Top view of layer 2 and 4 (dotted contour) and how they stack with the adjacent layers (solid contour).

PAGE 73

73 Figure 34. Location of the front, back, left and right side of the box, position of the x and y axes, identification of the fruit and location of the thermocouples. Thermocouples located in the pulp of the fruit are repr esented by solid circles; thermocouples located in the air are represented by empty circles (fruit 1, 63 and 125).

PAGE 74

74 Figure 3 5. Staggered arrangement of the oranges for the second series of test s. A) Top view of layer 1 and 3. B) Top view of layer 2 (dotted contour) and how it stacks with the adjacent layers (solid contour). Figure 36. Location of the front, back, left and right sides of the box, position of the x and y axes, identification of the fruit and location of the thermocouples. Thermocouples located i n the pulp of the fruit are repr esented by solid circles; thermocouples located in the air are represented by empty circles (fruit 1, 23, 34, 35, 40 and 42).

PAGE 75

75 Figure 37. Staggered arrangement of the oranges for the third series of test s. A) Top view of layer 1 and 3. B) Top view of layer 2 and 4 (dotted contour) and how they stack with the adjacent layers (solid contour).

PAGE 76

76 Figure 38. Location of the front, back, left and right sides of the box, position of the x and y axes, identification of the fruit and location of the thermocouples. Thermocouples located in the pulp of the fruit are represented by solid circles; thermocouples located in the air are represented by empty circles (f ruit 1, 43, 45 and 72).

PAGE 77

77 Figure 39. Internal dimensions of the LD3/AKE aircraft container with aluminum walls.

PAGE 78

78 Figure 310. Location of the front, back, left and right sides of the container, identification of the boxes and location of the thermocouples.

PAGE 79

79 Figure 311. Experimental set up with the aircraft container (LD3/AKE). A) Wooden wedge and insulation used for stacking the boxes of oranges. B) View of the container being loaded and instrumented with thermocouples. Figure 312. Schematic of the water bottles used for the inflight tests.

PAGE 80

80 Figure 313. Collapsible reusable plastic container (RPC). Figure 314. Staggered arrangement of the water bottles for the in flight tests. A) Top view of layer s 1 and 3. B) Top view of layer 2 (dotted contour) and how it stacks with the adjacent layers (solid contour).

PAGE 81

81 Figure 315. Schematic of the location of the temperature sensor inside a bottle of water used for the in flight tests.

PAGE 82

82 Figure 316. Location of the front, back, left and right sides of the RPC, position of the x and y axes, identification of the water bottles and location of the temperature sensor in bottle 35 (solid circle).

PAGE 83

83 Figure 317. Location of the front, back, left and right sides of the container, identification of the boxes and location of the temperature sensors. Water temperatures are represented by solid circles; air temperatures are represented by empty circles.

PAGE 84

84 Figure 318. Experimental set up with the aircraft container (LD3/AKE). A) Expanded polystyrene wedge used for stacking the RPCs. B) View of the fully loaded container.

PAGE 85

85 Figure 319. ULD position s in the forward (f wd) and aft cargo compartments of three models of aircraft that carried the experimental container during the series of in flight tests. A) B747 400F. B) A330200. C) A310300F.

PAGE 86

86 Table 31. Schedule of the six flights as well as the type of aircraft and the position of the experimental aircraft container during the 2007 inflight tests. Flight Date Departure t Arrival t,u Duration v Aircraft Position JFK w GOT x DXB y 04 29 11:00 11:00 16:00 B747 400F 44R DXB NBO z 05 03 10:05 14:15 5:10 A330 200 41L NBO DXB 05 04 23:00 5:00 5:00 A310 300F 11L DXB NBO 05 06 10:05 14:15 5:10 A330 200 32R NBO DXB 05 07 17:15 23:15 5:00 A330 200 14L DXB GOT JFK 05 12 23:00 8:00 17 :00 B747 400F 43L t: Local time; u: Italic arrival times indicates next day arrival; v: Flight duration hh:mm; w: New York U.S.A., John F. Kennedy International Airport; x: Landvetter, Sweden, Gothenburg Landvetter Airport; y: Dubai, U.A.E., Dubai International Airport; z: Nairobi, Kenya, Jomo Kenyatta International Airport

PAGE 87

87 CHAPTER 4 RESULTS AND DISCUSSION OF LABORATORY TESTS 4.1 Single Box Tests Different combinations of box es orange sizes and staggered packing arrangements were exposed to step changes in ambient temperature to create a break in the cold chain. These procedures simulated scenarios in which a box of produce awaits shipping or is being shipped in detrimental conditions. Table 41 summarizes the information relative to the boxes of oranges used for the three series of tests. The six walls of the corrugated cardboard boxes used in the tests were identified and numbered as shown in Figure 41 The pulp temperatures measured at a half radius distance fr om the center of the fruit are discussed in Sections 4.1.1 to 4.1.3. The locations of these measurements were selected to obtain an approximation of the average temperatures of the fruit. The a verage temperatures of the layers of produce and the average temperature of the box were calculated us ing the temperatures at the half radius distance. Temperatures measured at the center and just below the surface of the fruit as well as air temperatures are discussed in Section 4.1.4. The following notation is used in this chapter. Abox: Average temperature of the all fruit in the box. Lx: Layer x of the box, where x {15} for series 1, x for series 2 and x {14} for series 3 ALx: Average temperature of the fruit in layer x of the box where x {15} for series 1, x Fx : Pulp temperature measured at the half radius distance fr om the center of fruit z, where x for series 1 (Figure 34) x (Figure 3 6) and x (Figure 3 8). Wx : Temperature measured at the center of the inside surface of wall x of the box of produce where x {16} (Figure 41 )

PAGE 88

88 For the sake of simplicity, the results of only one of the test replications of each series are presented in detai l in the f ollowing sections. For each series, comparisons of t he results of the three tests are presented in graphs of the temperature differences between the corner and center fruit of each layer of produce within the box. Even though measurements were taken to position the thermocouples as accurately as possible, there are still uncertainties inherent to this type of temperature monitoring Displacement of the thermocouples during fruit stacking, movement of the fruit and variations of the thermal properties within the fruit as well as with time are all factors that may influence the temperature measurements. These factors must be considered while analyzing the experimental results. 4.1.1 First Series of Tests The box of oranges was stored at 2.5C until a uniform and steady temperatures were obtained and then swiftly transported into a temperaturecontrolled room at 35C for a period of 7 h. The 125 patternpacked fruit were equally distributed over five layers in the box (Figure 34). 4.1.1.1 Analysis of the temperature distribution within the box T he inside surface temperatures of the wall s of the box are presented in Figure 4 2. The temperatures of the lateral walls (W2, W4, W5 and W6) were ver y similar during the test. O n average, they varied within a range of 2.7C. The temperature of the lateral walls corresponded closely to the average between the temperatures of the top (W1) and bottom (W3) walls. The top and bottom walls were the hottest and c oldest respectively The average temperatures of the five layers of fruit are presented in Figure 4 3. Even though the temperatures on all fruit within the third layer were measured, t he

PAGE 89

89 average temperature was calculated using only the temperatures of the odd numbered fruit in order to be consistent with the average temperatures of other layers. Figure 4 3 shows that the second and fifth layers were the coldest and hottest respectively The average temperature difference between these layers throughout the test was 8.9C and it reached a maximum of 10.4C approximately 4 h af ter the beginning of the test. Figure 43 also shows that t he average temperature of the first layer of fruit was closer to that of the third layer. This can be explained by stratification of the air temperature and heat trans fer from the bottom of the box. S upport blocks u nderneath the box allow ed convective heat transfer on most of the bottom surface (Figure 31 ) To illustrate the spatial temperature variations within a layer the average temperature, the temperatures of two corner fruit as well as the temperature of a fruit located near the center of each layer were plotted ( Figures 44 to 4 8 ) For each of the five layers the temperatures of the tw o corner fruit were very similar which indicated a symmetrical temperature distribution with respect to the central yz plane of the box (Figure 3 4). Symmetry was expected because of the similarity of the four l ateral wall temperatures (Figure 4 2) The largest difference between corner and center temperatures was observed in the bottom layer. T he average temperatures of the first and second layers corresponded closely to the arithmetic average of their corner and center fruit temperatures (Figures 4 4 and 45) S mall temperature variations were observed within each of the fourth and fifth layer s (Figures 4 7 and 48) This can be explained by the stratification of the air temperature within the box. Fruit F113 was the closest to t he center of the top layer (L5); however, to better represent the temperature distribution

PAGE 90

90 throughout the layer fruit F107 was used in Figure 4 8. R esults showed that the temperature of fruit F113 was higher than any other f ruit within the fifth row by a minimum average of 4.8C. This difference may have been caused by an incorrectly positioned thermocouple, or the proximity of the fruit to the vent on the top of the box. Figure 49 provides an overview of the temperature distribution throughout the box during the test. It show s the variations of pulp temperatures between the center and corners (top and bottom) the pulp average temperature of the box (Abox) as well as the bottom (AL1) and top (AL5) average layer temperat ure s. Fruit F38, which was located in the center of the second row, was the coldest fruit within the box therefore, it was used as the center t emperature instead of fruit F63. The difference between the bottom (F1) and top corner (F101) was significantly less than the difference between any of the corner temperatures and the center temperature (F38). Over the entire test period, the maximum temperature difference for the pairs F101F38 and F1F38 were 11.5C and 10.5C. 4.1.1.2 Comparisons of the results between test replications The results of the three tests conducted during the first series are compared in Figures 4 10 to 414. Each of the graphs corresponds to a specific layer and presents the temperature difference between fruit located in the corner and th e center of the layer. The pairs of corner center fruit selected were F1F13, F30 F38 F51 F63, F80 F88 and F101F107 The choice of the corner fruit was not critical since the temperatures of corner fruits on a given layer varied si milarly with time. Figures 410 to 414 show th at the results obtained during T ests 2 and 3 were almost identical for all layers whereas the results of Test 1 diverged f rom the others as time progressed. The temperature difference s between the tests were comparable for

PAGE 91

91 the different layers. The m aximum variation between the tests was observed on the fourth layer and calculated at 2.5 C (Figure 4 13) The analysis of the results revealed that the inside wall temperatures were lower during Test 1. This may be explained by a variation of the convective heat transfer on the surface of the box caused by either the different location of the experimental set up inside the temperature controlled room or the displacement of the plastic film surrounding the experimental set up between Tests 1 and 2. For the first three lay ers, Figures 4 10 to 412 show positive temperature differences betw een the corner and center fruit. In all three cases, the positive temperature difference s in creased over the 7h test period. Among the three tests, t he temperature difference observed in the first layer reached a maximum of 9.5 C (Test 3) Within the second and third layer s, the maximum temperature differences between the corner and center fruit were lower at 6.4C (Test 2) and 6.1C (Test 3) ,, respectively The trend was reversed for the fourth and fifth layer s, in which the temperature of the center fruit was higher than that of the corner fruit The temperature difference reached 3.1 C an d 2.3 C within the fourt h and fifth layers respectively revealing again the more uniform temperature distribution within the upper layers of fruit. The significant temperature difference observed within the first layer of produce may be explained by the by the natural convection within the box. Because of the two central slot vents on the top and bottom of the box, t he resulting downward convective flow around the central vertical axis of the box may have enhanced the heat transf er to the neighboring fruit particularly in the top layers This would explain the higher pulp temperature near the center for the fourth and fifth layers. As the air progressed

PAGE 92

92 towards the bottom of the box, it cooled down and resulted in lower heat tr ansfer for the central fruit on the bottom layers. Also, s ince the buoyancy driven flow was restricted at the top of the box, fluid stratification was likely to have contribute d to the increased uniformity of the temperature distribution within the upper layers. 4.1.2 Second Series of Tests For these tests, the two temperaturecontrolled rooms used to create a stepchange in temperature were maintained at 3 C and 25C. The experimental set up remained exposed to high temperatures for a total of 15 h. In the box, 14 fruit were patternpack in each of the three layers (Figure 3 6). 4.1.2.1 Analysis of the temperature distribution with the box Figure 4 15 presents the temperatures measured in the center of the inside surface of each wall of the box. As for the first series of tests, the temperatures of the lateral walls (W2, W4, W5 and W6) were similar. For the test results presented in Figure 4 15, the four lateral wall temperatures varied within an average range of 0.9C over the 15h period. The temperature s of the lateral walls corresponded closely to the arithmetic average of the top (W1) and bottom (W3) wall temperatures. The average temperatures of three layers of fruit are presented in Figures 4 16. The first and third layers displayed the lowest and highest average temperature,, respectively The variations of the average temperatures within the box were smaller than for the first series of test s (Figure 4 3). This may be explained by the smaller number of layers within the box and the higher porosity of the packing arrangement. The difference between the average temperature of the first (AL1) and third (AL3) layer was on average 3.4C and it reached a maximum of 5.7C approximately 2.5 h after the beginning of the test.

PAGE 93

93 The variations of the te mperatures within each of the three layers of fruit are presented in Figures 4 17 to 419 In this series of three graphs, the temperatures of fruit closest to the center as well as the two corner fruit are plotted for each layer. As it was observed in t he first series of tests, the temperatures of the two corner fruit were almost identical. Larger temperature differences were expected since the thermocouples inserted in fruit F1, F25 and F29 were closer to the lateral wall than those installed in fruit F4, F28 and F32 (Figure 3 6). This indicates that the temperature gradients within the fruit were small. The results obtained for the first two layers of f ruit showed very similar trends; however, the first layer displayed a longer initial lag. On the third layer (Figure 4 19), the corner and center fruit temperatures varied within a maximum range 1.7C for the entire duration of the test, which indicates the uniform temperature distribution throughout that layer. Figure 420 presents an overview o f the temperature distribution within the box of produce. In addition to the average pulp temperature of the entire box and the average temperatures of the bottom (AL1) and top (AL3) layers, the temperatures of the fruit located in the bottom (F1) and top (F29 ) corner as well as near the center (F23) of th e box are plotted in Figure 4 20 Over the entire 15 h, a maximum temperature difference of 2.0 C was observed between the bottom corner and center fruit. For the top corner and center fruit, the maximum temperature increased to 5.1C. For this series of tests, it was observed that the temperature of the bottom corner fruit (F1) varied closely with the average pulp temperature of the entire box (Abox) (Figure 4 20); t he absolute value of their difference average 0.5C and reached a maximum of only 1.4C.

PAGE 94

94 4.1.2.2 Comparisons of the results between test replications Comparisons of the results for the three tests of this series are presented in Figures 4 21 to 423. Each of these graphs corresponds to a dif ferent layer and shows the difference in temperature between the corner and center fruit for the three tests conducted. The pairs of fruit F1F6, F25 F23 and F29F34 were chosen for layers 1, 2 and 3, respectively For the first layer (Figure 4 21), the results of the three tests were similar. Among the three tests, t he corner fruit reached a maximum temperature of 3.5 C above that of the center fruit. For Tests 1 and 3, the temperature difference decreased slowly with time after it reached that maximum whereas it remained relatively steady during Test 2. Nevertheless, the maximum difference between the three t ests for the first layer was 1.3C (Figure 4 21). For the seco nd layer of fruit (Figure 4 22), similar results as for the first layer were obtained. The maximum temperature difference between the corner and center fruit of was slightly higher but the trends remained analogous The maximum difference between the three tests was 1.9C From the results presented in Figure 4 19, a negligible temperature difference between the corner and center fruit of the third layer was expected. Tests 2 and 3 corroborated the previous results with average differences of 0.2C and 0.0C respectively over the 15h test period. During the first test, the temperature of the corner fruit was higher than that of the center fruit by an average of 1.3C. The displacement of the corner or center fruit may have caused this small inconsistency between the tests.

PAGE 95

95 4.1.3 Third Series of Tests The box of fruit used f or this series of tests contained a total of 72 fruit that were patternpacked onto four layers of 18 fruit (Figure 3 8) For each test replication, the boxed produce were exposed to a step change in temperature from 1C to 25C. The experimental set up remained in the high temperature environment for a period of 15 h. 4.1.3.1 Analysis of the temperature distribution within the box The change with time of the inside surface temperatures of the six walls of the box are presented in Figure 4 24. Again, the temperatures of the lateral walls of the box ( W2, W4, W5 and W6) were similar. T hey varied within an average range of 1.6C. For the first and second series of tests, it was observed that the temperatures of the lateral walls corresponded to the average of the top (W1) and bottom (W3) wall temperatures. For the third series, the temperatures of the lateral walls particularly walls W5 and W6, were colder and therefore closer to that of the bottom wall. This may be explained by the fact that there were an even number of layer (four) for this series of test s and, therefore, no layer was aligned vertically with the center of the lateral wall. The increased void volume s near the locations of the thermocouples contributed to enhanc ing the convective heat tra nsfer, which may have resulted in colder interior wall surface temperatures. The variations with time of the average temperatures of the four layers are shown in Figure 4 25. The first (AL1) and second (AL2) layer s were the coldest and varied, on average, within 0.3C of each other. Using the average temperature of the first and second layers as a reference, the temperatures of the third (AL3) and fourth (AL4) layers were on average 2.1 C and 4.9 C higher ,, respectively

PAGE 96

96 Figures 4 26 to 429 present the t emperature variations of two fruits located in corners as well as one near the center of each of the layers. The average temperatures of the corresponding layer s are also presented in the graphs. For each layer the temperatures of the t wo corner fruit v aried almost identically except in the case of the first layer for which they diverged by a maximum of 1.4 C Figures 4 26 to 429 also show that the uniformity of the temperature distributions within the layers increased from the bottom (L1) to the top (L4). The average variations between the temperatures of the bottom (Figure 4 26) and top (Figure 4 29) layers were 3.9C and 0.8 C respectively An overview of the temperature distribution throughout the box is presented as a function of time in Figure 4 30. It is interesting to look at the differences between the pulp temperature near the center of the box (F25) and the pulp temperatures at the top and bottom corners (F68 and F1). The average temperature differences over the period of 15 h wer e calculated at 5.9C and 3.0C for the F68F25 and F1 F25 pairs ,, respectively Als o, it can be observed that the change with time of the total average pulp temperature of the box (Abox) varied closely with the temperature of the bottom corner fruit (F1) The maximum temperature difference between the two temperatures in Figure 4 30 was only 0.8C. This interesting detail was also observed for the second series of tests for which the maxi m um temperature difference was calculated at 1.4C (Figure 4 20). For the first series, the difference between the bottom corner fruit and the box average temperatures was more important, it averaged 2.9C and reached a maximum of 3.9 C (Figure 4 9) 4.1.3.2 Comparisons of the results between test replications The resul ts of the three tests conducted in this series are compared in Figures 4 31 to 434, which correspond to the first, second, third and fourth layers ,,

PAGE 97

97 respectively In each of the figures, the pulp temperature difference between the fruit located in a corn er and near the center of the corresponding layer is plotted for the three tests. The pairs of fruit selected for the layers were F1F12 (L1), F32F25 (L2), F37F48 (L3) and F68F61 (L4) (Figure 38). The results of the three tests were c onsistent over the entire 15 h. T he maximum temperature variations between the tests were calculated to be 0.6 C, 1.2 C, 0.7 C and 1.1 C for L1 to L4,, respectively For the first three layers, Figures 4 31 to 433 show that the corner fruit were hotter than the one located near the center. For each of the three layers, the maximum temperature difference was measured to be 4.3C, 3.5 C and 2.9C for layers L1 to L3,, respectively The maximum temperature differences were all reached within the f irst 4 h of the tests. A relatively uniform temperature distribution was observed within the top layer of the two previous series of tests This phenomenon occurred again in this series, as illustrated in Figures 4 29 and 434. The latter shows that the average temperature difference between the corner and center fruit was 0.3 C over a period of 15 h. 4.1.4 Te mperature Differences within the Fruit and with the Surrounding Air Through the previous sections, the results presented were based on the pulp temperatures measured at a half radius distance. This location was selected to obtain an approximation of the mass average temperature of the fruit. In addition to these measurements, the pulp temperature at the center of a small number of fruit as well as the air temperatures near their surfaces were monitored. Experimental data on the temperature gradients within the fruit and the temperature difference between the solid and f luid phase are particularly valuable to establish valid modeling assumptions. These results are presented in this section.

PAGE 98

98 Through a preanalysis of the results, it was found that for most fruit only small differences existed between the temperatures measured just below the sur face of the fruit and in the air surrounding it. For the sake of simplicity and to avoid underestimating the temperature differences presented in this section, the air temperatures were used for the calculations and analysis. 4.1.4.1 First series of tests Table 42 summarizes the maximum and average pulptemperature differences between the half radius distance and the center of the fruit F1 and F63. For the same two fruit, the differences between the air temperature near their surface and the pulp temperature at the half radius distance are summarized in Table 4 3 Fruit F1 and F63 were positioned in a bottom corner and near the center of the box respectively (Figure 3 4). Tables 42 and 43 show that all three tests of the series produced consistent results. For the fruit located near the center of the box (F63) Table 4 2 indicates that the average pulptemperature difference was 0.7 C with a maximum of approximately 1.2 C. The small temperature gradients within the fruit indicate that their thermal behavior is close to that of a lumped system. The t emperature differences were higher for fruit F1. These results may be in part explained by the proximity of the thermocouple (half radius) from the left wall (W4) of the box. Table 4 3 indicates similar maximum and average temperature differences for bot h fruit. An average difference between the air and pulp temperatures of approximately 1.5C can be considered as small for most engineering applications

PAGE 99

99 4.1.4.2 Second series of tests The differences in pulp temperatures of three fruit are presented in T able 44 The instrumented fruit were located at the bottom (F1) and top (F42) corner as well as near the center (F23) of the box. Table 44 shows that the different tests produced consistent results. For the fruit F1 and F23, the average temperature di fferences between the half radius distance and the center did not exceed 0.2 C for any of the three tests. The maximum temperature difference was observed in fruit F1 at 0.6C, which is almost within the accuracy of a single thermocouple reading The differences calculated from the temperatures measured within fruit F42 were also small but negative, which implied that the center temperature was the highest of the two. The presence of an air pocket in the central column of the orange, where the thermocouple was positioned, may be responsible for these results. The smaller thermal diffusivity of the air would cause its temperature to change at a faster rate than the pulp of the fruit Because important vertical temperature stratification was obs erved in all tests, the rotation of the fruit resulting in the displacement of the half radius temperature measurement location below the center of the fruit may also explain these systematic n egative temperature differences (Figure 4 16). Table 45 summ arizes the differences between the air temperature and the pulp temperature at the half radius distance from the center of the fruit. The results for fruit F1 and F23 were very similar to those of the correspon ding fruit for the first series (Table 4 3 ). For all three tests, the average temperature difference did not exceed 1.6 C for any of the fruit F1 and F23 (Table 4 5 ). In the case of the top corner fruit (F42), large differences were calculated between the air and pulp temperatures. For the three t ests, the maximum temperature differences of 9.2C, 9.6C and 9.8C were

PAGE 100

100 measured within the first hour of the experiment. Once that maximum was reached, the temperature differences decreased exponentially and were below 3C for more than half of the tota l test duration (15 h). The temperature stratification within the box combined with the fact that the air temperature was measured just above the fruit instead than on its side contributed to these high temperature differences Table 45 also shows that the results were consistent between the three tests of the series. 4.1.4.3 Third series of tests Contrary to the two othe r test series, pulp temperature differences were measured only for the bottom corner fruit (F1) (Table 4 6 ). Again, the three tests produced consistent results. The maximum temperature difference varied between 2.4C and 2.6 C, whereas the average temperature differences varied between 1.3C and 1.6 C. T he temperature distribution within the fruit was therefo re more uniform than the first series of tests but not as much as the second. Table 4 7 present s the differences between air and pulp temperatures for the fruit located in the bottom corner (F1), center (F43) and top corner (F72) of the box. For fruit F1 and F43 the maximum temperature differences were only 1.9C and 0.7C respectively The air temperatures were measured near the top of fruit F1 and near the bottom of fruit F43 which explains why the average difference of the latter was 0.2 C for each o f the three tests. These results again indicated that the temperature distribution within the fruit was quite uniform. As for the second series of tests, the maximum differences between the air and pulp temperature of the top corner fruit (F72) were high, they varied between 9.6C and 10.3C for the three tests. In all cases, the maximum temperature differences were reached within the first 30 min of the tests and

PAGE 101

101 then decreased exponentially with time. After 10 h the differences between the air and pulp temperatures of fruit F72 were less than 3 C. 4.1.4.4 Theoretical calculations on the uniformit y of the temperature within a fruit According to the results presented previously in Sections 4.1.4.1 to 4.1.4.3, the temperature distribution within several of the fruit was relatively uniform. To evaluate the validity of these observations theoretical calculations can be performed. The Biot number (Bi) corresponds to the ratio of the internal resistance to conduction of an object over the thermal resistance to convection on its surface. A small Bi ot number typically smaller or equal to 0.1, indicates that a lumpedsystem analysis can be performed. For spheres, the Bi ot number can be obtained using Equations 41 and 42 (engel and Ghajar, 2010) = (4 1) = = 6 ( ) (4 2) To calculate the Bi ot number for the three series of tests, the convective heat transfer coefficient (h) around the fruit in t he box must first be estimated. For this purpose, one of the boxed fruit can be considered as an individual sphere exposed to natural convection. Equations 43 to 45 were used to obtain the Nusselt number (Nu) and to calculate the convective heat transfer coefficient. = (4 3) = 2 + 0 589 1 + 0 469 ( 10 ) ( 0 7 ) (4 4)

PAGE 102

102 = (4 5) In order to avoid underestimating the convective heat transfer coefficient, t he difference in temperature between the surface of the fruit and the ambi ent air were assumed to be constant and were based on the first series of test s, for which the largest step change in ambient temperature was used. Therefore, the surface and the ambient air temperatures were assumed to be 2.5C and 35.0C respectively The physical and thermal properties were therefore calculated at the film temperature of 18.8C For the calculation of the Bi ot number the thermal conductivity of the solid (fruit) was taken to be 0.431 Wm1K1 (Singh and Heldman, 1993). T able 48 summa rizes t he results of the Biot number calculations The values obtained for theoretical convective heat transfer coefficient s were 6.23 Wm2K1, 5.78 Wm2K1 and 6.05 Wm2K1 for series 1 to 3, respectively The average value of 6 Wm2K1 was used to calc ulate the corresponding Biot numbers: 0.155, 0.194 and 0.169. These results indicate that the temperatures within the fruit are quite uniform and therefore corroborate the observations made from the experimental results. Even thought the values of the Bi ot numbers are slightly higher than 0.1, considering the fruit as lumped system s would still be a justified assumption for most engineering applications This is particularly true if the uncertainties associated with the location of the thermocouples, the ir measurements and the thermal properties of the fruit are considered. From the experimental and theoretical results that were presented, it can be stated that, in general, the temperature distribution within the fruit and between the fruit and the surrounding air was relatively uniform. Consequently, the temperature measured at

PAGE 103

103 a half radius distance from the center of the fruit can be considered as a good approximation of the mass average temperature of the fruit and the surrounding air. 4.2 Aircraft Container Test The objective of this series of test s was to study the thermal behavior of an aircraft container loaded with horticultural products. The oranges used to conduct the experiments were the same as those used i n the first single box series of tests A total of 88 box es of fruit were required to fill up the aircraft container Each box contained a total 125 fruit evenly distributed on five layers (Figure 3 4). Similar to the single box tests (Section 4.1), the intention was to expose the airc raft container to a step change in temperature to recreate the conditions (without solar radiation) where an aircraft container is taken out of the cold room prior to a shipment. The tests were conducted inside a temperatu re controlled marine container si nce the aircraft container was too large to fit into any of the available temperaturecontrolled room s. The set point of the temperaturecontrolled marine container was changed from 3C to 30 C to recreate the conditions closest to a real stepchange in t emperature. This approach could not provide a sharp change in temperature since 30 min were required for the supply air to reached 30C. However, this approach had the advantage of having the experimental set up fixed during the entire series of tests, w hich helped produce consistent results. Figure 3 10 presents the locations and the identification number of the 88 boxes as well as the locations of the thermocouples throughout the load. Again, layer 1 (L1) and l ayer 6 (L6) correspond to the bottom and t op layer respectively All pulp temperatures were measured at a half radius distance from the center of the fruit. The six main walls of the aircraft container were identified and numbered as shown in Figure 4 35. Throughout this section, the average l ayer temperatures were calculated

PAGE 104

104 using the pulp temperature of fruit F63, which is located near the center of a box (Figure 3 4). The following notation is used in this chapter. Bx out : Pulp temperature measured near the outer surface of the load in box x, where x (Figure 310) Lx : Layer x of boxes in the aircraft container, where x {16 } (Figure 3 10) ALx: Average temperature of layer x of the aircraft container where x {16 } (Figure 3 10) Fx : Pulp temperat ure measured at the half radius dis tance from the center of fruit x, where x 4) Wx : Temperature measured at the center of the inside surface of wall x of the aircraft container, where x {16} (Figure 435) For the sake of simplicity the results of only one of the test replications are presented in d etail Comparisons between the results from the three tests are discussed in Section 4.2.5. 4.2.1 Temperatures of the Walls of the Aircraft Container The temperatures at the center of th e inside surface of the six walls of the aircraft container are presented in Figure 436. Contrar y to what was observed during the series of tests with single boxes of fruit (Section 4.1), heterogeneit ies were expected among the lateral wall temperatures because of the particularity of the air flow inside the marine container as well as the geometry of the aircraft container. In a temperaturecontrolled marine container, the air is sup plied from the back, where the air conditioning unit is installed, and flows through a t beam floor towards the doors During the tests, the floor of the marine container was uncovered and, therefore some of the supplied air escaped the t beam floor along the way. A significant portion of the supplied air also flowed all the way to t he front of the marine container before being recirculated through the air conditioning unit. Consequently the convective heat transfer was not uniform on

PAGE 105

105 different walls of the aircraft container, which was oriented so that its front wa ll faced the door of the marine container (Figure 4 35). The results presented in Figure 4 36 show that after approximately 1.5 h, the temperatures of the top and lateral wall s reached a pseudo steady state. The thermal behavior of the bottom wall (W3) was different since it was in direct contact with the load. Other factors that contributed to the different t hermal behavior were the large thickness of the aluminum floor (0.013 m compared to 0.0005 for the walls) and the fact that the container rested on a wooden pallet to allow, if necessary, any movement of the experimental set up. After 8 h, the temperatures of the walls varied between 21.5C (W2) and 29.1C (W1). Because of the symmetry of the orientations of wall W2 and W4 with respect to the airfl ow, it was expected that the convective heat transfer on their outer surfaces be similar. Consequently, their temperature distribution should also be close to each other The temperature difference observed after 8 h between wall W2 (21.5C) and W4 (23.7C) can be explained by the vertical stratification of the air temperature inside the aircraft container and the fact that the temperature of wall W4 was measured at a location 0.25 m higher than wall W2. 4.2.2 Average Layer Temperatures For layers 1 to 5, the temperatures measured near the center of the box es (fruit F63) remained almost constant over the 8h test period. The center temperatures of the boxes from the top layer (L6) varied more significantly with time. Small temperature v ariations were observed among the boxes of a specific layer, even for the top layer where the temperatures ranged within 3.0C of each other. The refore, the average of all temperatures measured near the center of the boxes on a layer represents the

PAGE 106

106 temper ature variation of the products located in the core of that layer Figure 3 37 shows the variations with time of the average temperature of each of the layers. The average temperature of the top layer (AL6) increased from 3.1C initially t o 12.9C after 8 h. The second and third largest variations were recorded on the fifth (AL5) and fi rst layers (AL1), where the average temperatures increased from 3.2C to 5.1 C and 3.3C to 4.6C respectively T hese results indicate that with the exception to the top layer (L6), the fruit located within the core of the load were not affected significantly by the 8 h break in the cold chain. It was expected that the detrimental conditions had a more severe impact on the temperature o f the fruit located closer to the outer surface of the load. 4.2.3 Temperatures of Fruit Located in the Outer Regions of the Load On layers L1, L3 and L6, the pulp temperatures of several fruit near the outer regions of the load were measured. For the bo x adjacent to the back (W6), left (W4), right (W2) and front (W5) side of the aircraft container, the instrumented fruit were F73, F61, F65 and F55, respectively (Figures 3 4 and 310) F igures 4 38 to 440, which correspond to layers L1, L3 and L6, prese nt the average temperature of the layer as well as the pulp temperatures measured near the outer surface of the load. Figure 438 shows that all four peripheral pulp temperatures measured in boxes B2, B4, B6 and B1 1, varied within a range of 3.4C Also, the average of these four temperatures was higher than the temperature AL1 by an average of 3.8 C over the 8h test. It is important to mention that a n insulated wedge used to achieve proper stacking of the load was placed to the left of the fi rst a nd second layers of boxes (Figure 3 11). The thickness of the wedge was the smallest near the bottom of the first layer. T he

PAGE 107

107 te mperature measured near that location ( B4 out ) indicated that the insulation did not present a large resistance to heat t ransf er. I n Figure 439, the four temperatures measured in the outer region of the load again higher than the average temperature of the third layer (AL3). The difference between the average of the four peripheral temperatures and temperature AL3 was on averag e 2.7C. Throughout the test, the outer surface temperature of box B27, B29, B32 and B39 varied within a range of 2.8C. For this layer, factors that may have affected the pulp temperatures in the outer region of the load were the distance between the load and the walls of the aircraft container as well as the temperatures of the walls of the aircraft container (Figure 436). The distance between the load and the walls of the container was 0.10 m for the front and back walls (W5 and W6), whereas it was 0.05 m for the right and left walls (W2 and W4). Therefore, boxes B29 and B32 were closer to the walls than boxes B 27 and B39 However, the temperatures of the walls adjacent to boxes B29 and B32 were colder than for b oxes B27 and B39 Consequently, the di stribution peripheral temperatures presented in Figure 4 39 are the results of these two opposite effects. Figure 440 illustrates the fact that thermal energy was transferred deeper and more uniformly through the top layer (L6). It can be observed that t he average layer temperature (AL6) corresponded closely ( within 0.7C ) to the average of the four peripheral temperatures. The outer surface temperatures measured in box B75, B77, B80 and B87, varied within a range of 4.6 C during the test, which was slig htly wider than for layer L1 and L3.

PAGE 108

108 The r esults in this section indicate that in addition to the top layer of boxes (L6) the first two layer s of fruit located near the entire out er surface of the load were affected by the detrimental conditions. T hese fruit may first appear as a small percentage when compared to all the fruit located in the core of the load, which according to the results presented in Section 4.2.2, were not affected by the detrimental conditions during the 8 h tests. However, the two layers of fruit near the outer surface of the load, omitting those adjacent to the insulated wedge, add up to 37.9% of the 11,000 fruit of the load Including all the fruit from the top layer, the percentage of affected fruit climbs to 45.7%. This a lso means that 72 of the 88 boxes (81.8%) were affected. During the tests, temperatur es increased to moderate levels. However, the results are likely to be different during real shipments where, in addition to high ambient air temperatures, the aircraft container may be exposed for extended periods to solar radiation. Also, since the air inside an aircraft container reaches very high level s of relative humidity when loaded with fresh fruit and vegetables temperature variations on the scale observed dur ing this series of tests are sufficient to cause condensation on the produce. Condensation contributes to the development of decay organisms that can significantly affect the quality of fresh horticultural products, particularly berries. 4.2.4 Air and Pulp Temperatures within the Load Air and pulp temperatures wer e measured in three locations within box B1, B47 and B88 (Figure 310). The pulp temperature was measured at a half radius distance and the air temperature near the surface of the fruit. In e ach of these box es fruit F 2 1 (bottom corner), F63 (near to the center) and F105 (top corner) were instrumented (Figure 3 4).

PAGE 109

109 The temperatures measured in the bottom corner box, B1, are presented in Figure 4 41. For each fruit, the pulp and air temperatures varied closely as it was observed in the series of tests with a single box of fruit (Section 41) For the entire 8h test, fruit F21 displayed the maximum pulpair temperature difference at 1.8 C. Th e pulp and air temperatures of both fruit F63 and F105 were within 1.0 C of each other. The highest temperatures in box B1 were measured for fruit F21 (16.8C). Fruit F63, located near the center of the box, remained the lowest temperatures among the three fruit (4.9 C) On average, the pulp t emperature differences between fruit F21 and F63, and, fruit F105 and F63 were 8.7C and 1.8 C respectively The increase of the temperatures near top corner fruit (F105) may be attributable to the heat transfer through the vertical air channels formed throughout the height of the aircraft container by the bulged boxes (Figure 442). On the fourth row, box B47, located near the center of the entire aircraft container, was instrumented. The temperatures of f ruit F21 were not available. In Figure 443, f ru it F63 showed the smallest temperature variations; its pulp temperature increased by only 0.8C over 8 h. No significant difference was observed between the temperature of the air surrounding fruit F63 and its pulp temperature. The pulp temperature of fr uit F105 was on average 1.1 C above that of fruit F63. A larger temperature difference between the pulp and air temperature of fruit F105 was observed in box B47 than for box B1. On average, the air temperature was 2.0C above the pulp temperature of fruit F105. L arge temperature variations were observed in box B88 located on the top corner of the load. After 8 h, the pulp temperature of fruit F21, F63 and F105 reached 11.6C,

PAGE 110

110 14.1C and 21.4C respectively (Figure 4 44) For the three fruit, th e differences between the air and pulp temperature were higher t han what was observ ed in box B1. On average, the pulp air temperature difference was 2.8C, 1.7 C and 1.2 C for fruit F21, F63 and F105 respectively The pulp temperatures of fruit F21 and F63 followed a similar trend for the first 2.5 h of the test, but after that time the pulp temperature of fruit F63 was on average 1.3 C higher. The pulp temperature of fruit F105 was on average 6.8 C above than that of fruit F63 located near the center o f box B88. 4.2 .5 Comparisons of the R esults between the Test Replications To compare the results obtained for the three tests of this series the temperature difference between a corner box and a box near the center of layers L1, L4 and L6 were calculated. The pulp temperature measured in the fruit near the center of each box (F63) was used for the calculations. Figure 445 shows the temperature differences between box B3 and B5 for the three tests replications. The results were almost identical for the three tests, they were within 0.3C of each other. Also, f or the three tests, the temperature differences varied between 0.7 C and 0.1 C over 8 h, which indicates that the core temperature of box B3 and B5 were quite similar. Similar results were obtained for box B44 and B47 located on the fourth layer within the aircraft container (Figure 446). For the three replications, the temperature difference between boxes B44 and B47 varied within a range of 0.3 C, indicating that the tests generated consistent results. Again, the core temperatures of the two boxes were very similar since their difference varied between 0.2 C and 0.9C for the entire test duration.

PAGE 111

111 Figure 447 show s the results for the top layer (L6), where the temperature difference between box B76 and B79 was investigated. The results of the thr ee tests were again consistent; they did not vary by more than 0.6C. Because three sides of box B76 were exposed to convective heat transfer, it was expected that the rate of change of its temperature would be higher than that of box 79. This is reflected by the positive temperature differences observed in Figure 447. For the three test replications, the temperature difference between box B76 and B79 varied between 1.0 C and 1.9 C. 4.2.6 Effect o f Heat Generation In situations where heat generation within a load of fresh fruit or vegetables is important, an increase in air and pulp temperatures in t he core region can be measured. In this study, no significant temperature increases were observed f or the fruit located within the core of the third and fourth layer of the load (Figures 4 37 and 443). Also, the fact that initial temperatures were uniform throughout the load for the three test replications is another indication that the effect of heat generation was negligible. This is true particularly since the cooling process was slow compared to other methods such as forced air cooling. This conclusion was expected s ince citrus fruit have a low rate of respiration (Kader, 2002b ) Other products such as asparagus that are often transported by air and have a respiration rate 6 to 12 times higher than citrus (Kader, 2002 b ), may be subjected to quality loss due to heat generation particularly if they are not properly cooled prior to assembly of the load.

PAGE 112

112 4.3 Conclusions from L aboratory Tests For the singleb ox tests, one important finding was the high rate of change of the pulp temperatures in the core of the boxes. The temperature also varied significantly with the position in the box. The largest difference between the average layer temperatures was observed in the first series of tests, for whi ch the bulk porosity was the smallest and the number of layers the highest On average, that difference was 8.9 C, whereas it was 3.4 and 3.9C for the s econd and third series of tests respectively The temperature differences between the fruit near the center of the box and the bottom and top corner fruit were also larger for the first series of tests. For all three series of tests, it was observed that the temperatures within a layer were significantly more uniform in the upper region of the box. Throughout the box, the differences between the air and the pulp temperatures were relatively small. I t was found that the thermal behavior of the fruit a ppr oached that of lumped system and that the pulp temperature measured at a half radius distance was a good approximation of the mass average temperature of the fruit and the surrounding air. For the aircraft container tests, the results showed that with the exception of the top layer of boxes, the fruit located in the core of the load were not significantly affected by the a 8h exposure to detrimental condition. However, the fruit located in the peripheral region of the load, which accounted for 45.7% of al l fruit were affected Consequently, the simulated break in the cold chain had a significant impact on the load. As it was observed for the singlebox tests, the temperature was vertical ly st ratified within the load, except for the bottom layer which was generally at a higher temperature than the second layer because of the heat transfer through the bottom wall.

PAGE 113

113 Figur e 41. Identification numbers used for the walls of the boxes of produc e. Figure 42. Temperatures of the inside surfaces of the walls of the box for the first series of test s.

PAGE 114

114 Figure 43. Average temperatures of the layers of fruit in the box of the first series of test s. Figure 44. Average temperature (AL1) as well as temperatures of fruit located in corners and near the center of the first layer for the first series of test s.

PAGE 115

115 Figure 45. Average temperature (AL2 ) as well as temperatures of fruit located in corners and near the center of the second layer for the first series of test s. Figure 46. Average temperature (AL3) as well as temperatures of fruit located in corners and near the center of the third layer for the first series of test s.

PAGE 116

116 Figure 47. Average temperature (AL4) as well as temperatures of fruit located in corners and near the center of the fourth layer for the first series of test s. Figure 48. Average temperature (AL5) as well as temperatures of fruit located in corners and near the center of the fifth layer for the first series of test s.

PAGE 117

117 Figure 49. Overview of the temperature distribution throughout the box for the first series of tests. Figure 410. Comparison of the d ifference between the temperatures of the corner and center fruit (F1F13) l ocated on the first layer (L1) of the box for the three tests of the first series.

PAGE 118

118 Figure 411. Comparison of the difference between the temperatures of the corner and center fruit (F30F38) located on the second layer (L2) of the box for the three tests of the first series. Figure 412. Comparison of the difference between the temperatures of the corner and center fruit (F51F63) located on the third layer (L3) of the box for the three tests of the first series.

PAGE 119

119 F igu re 4 13. Comparison of the difference between the temperatures of the corner and center fruit (F80F88) located on the fourth layer (L4) of the box for the three tests of the first series Figure 414. Comparison of the difference between the temperatures of the corner and center fruit (F101F 107 ) located on the fifth layer (L5) of the box for the three tests of the first series.

PAGE 120

120 Figure 415. Temperatures of the inside surface of the walls of the box for the second series of test s. Figure 416. Average temperatures of the layers of fruit in the box of the second series of test s.

PAGE 121

121 Figure 417. Average temperature (AL1) as well as temperatures of fruit located in corners and near the center of the first layer for the second series of test s. F igure 41 8 Average temperature (AL2) as well as temperatures of fruit located in corners and near the center of the second layer for the second series of test s.

PAGE 122

122 Figure 41 9 Average temperature (AL3) as well as temperatures of fruit located in corners and near the center of the third layer for the second series of test s. Figure 420. Overview of the temperature distribution throughout the box for the second series of tests.

PAGE 123

123 Figure 421. Comparison of the difference between the temperatures of the corner and center fruit (F1F6) located on the first layer (L1) of the box for the three tests of the second series. Figure 42 2 Comparison of the difference between the temperatures of the corner and center fruit (F25F23) located on the second layer (L2) of the box for the three tests of the second series.

PAGE 124

124 Figure 423. Comparison of the difference between the temperatures of the corner and center fruit (F29F34) located on the third layer (L3) of the box for the three tests of the second series. Figure 424. Temperatures of the inside surface of the walls of the box for the third series of tests.

PAGE 125

125 Figure 425. Average temperatures of the layers of fruit in the box of the third series of t est s. Figure 426. Average temperature (AL1) as well as temperatures of fruit located in corners and near the center of the first layer for the third series of tests.

PAGE 126

126 Figure 427. Average temperature (AL2) as well as temperatures of fruit located in corners and near the center of the second layer for the third series of tests. Figure 428. Average temperature (AL3) as well as temperatures of fruit located in corners and near the center of the third layer for the third series of tests.

PAGE 127

127 Figure 429. Average temperature (AL 4 ) as well as temperatures of fruit located in corners and near the center of the fourth layer for the third series of tests. Figure 430. Overview of the temperature distribution throughout the box for the third series of tests.

PAGE 128

128 Figure 431. Comparison of the difference between the temperatures of the corner and center fruit (F1F12) located on the first layer (L1) of the box for the three tests of the third series. Figure 432. Comparison of the difference between t he temperatures of the corner and center fruit (F32F25) located on the second layer (L2) of the box for the three tests of the third series.

PAGE 129

129 Figure 433. Comparison of the difference between the temperatures of the corner and center fruit (F37F48) located on the third layer (L3) of the box for the three tests of the third series. Figure 434. Comparison of the difference between the temperatures of the corner and center fruit (F68F61) located on the fourth layer (L4) of the box for the three te sts of the third series.

PAGE 130

130 Figure 4 35. I dentification numbers used for the walls of the instrumented aircraft container. Figure 436. Temperatures of the inside surfaces of the walls of the aircraft container.

PAGE 131

131 Figure 437. Average temperatures of the layers of boxes in the aircraft container. Figure 438. Average temperature of the bottom layer of boxes (AL1) and pulp temperatures of four fruit located near the outer surface of boxes B2, B4, B6 and B22, also located on L1.

PAGE 132

132 Figure 439. Avera ge temperature of the third layer of boxes (AL3) and pulp temperatures of four fruit located near the outer surface of boxes B27, B29, B32 and B39, also located on L3. Figure 440. Average temperature of the top layer of boxes (AL6) and pulp temperatures of four fruit located near the outer surface of boxes B75, B77, B80 and B87, also located on L6.

PAGE 133

133 Figure 441. P ulp temperatures as well as the temperature of the air (a) surrounding fruit F21, F63 and F105, all located on the threedi mensional diagonal of box B1 (L1). Figure 442. Picture of a channel created by the bulged boxes of fruit inside the aircraft container

PAGE 134

134 Figure 443. Pulp temperatures as well as the temperature of the air (a) surrounding fruit F21, F63 and F105, all located on the threedimensional diagonal of box B47 (L4). Figure 444. Pulp temperatures as well as the temperature of the air (a) surrounding fruit F21, F63 and F105, all located on the threedimensional diagonal of box B88 (L6).

PAGE 135

135 Figure 445. Comparison of the differences between the center temperatures of the corner and near center box (B3B5) located on the bottom layer (L1) of the aircraft container for the three tests. Figure 44 6 Comparison of the differences between the center temperatures of the corner and near center box (B44B47) located on the fourth layer (L4) of the aircraft container for the three tests.

PAGE 136

136 Figure 447. Comparison of the difference between the center temperatures of the corner and near center box (B76B79) located on the sixth layer (L6) of the aircraft container for the three tests.

PAGE 137

137 Table 41 Dimensions of the boxes number or fruit and layers as well as characteristic s of the oranges used for the three series of singlebox tests. Box dimensions y Total number of fruit Layers Diameter z Mass z Bulk Porosity Variety m m kg Series 1 0.429 / 0.266 / 0.247 125 5 0.0668 0.1544 0.315 Hamlin Series 2 0.380 / 0.270 / 0.244 42 3 0.0834 0.2605 0.425 Navel Series 3 0.380 / 0.270 / 0.244 72 4 0.0728 0.1879 0.420 Navel y : Length / width / height; z : Average value. Table 42 Maximum and average pulptemperature differences between the half radius distance and the center (c) of fruit F1 and F63 for the first series of tests. Bottom Corner Center F1 F1c ( C) F63 F63c ( C) Maximum Average n y s z Maximum Average n s Test 1 3.2 2.3 86 0.6 1.2 0.7 86 0.3 Test 2 2.9 2.4 86 0.5 1.3 0.7 86 0.4 Test 3 3.1 2.5 87 0.5 1.2 0.7 87 0.3 y : n is the number of measurements; z : s is the standard deviation Table 43. Maximum and average differences between the air temperature near the surface of the fruit (a) and the pulp temperature at the half radius distance of fruit F1 and F63 for the first series of tests. Bottom Corner Center F1a F1 ( C) F63a F63 ( C) Maximum Average n y s z Maximum Average n s Test 1 3.5 1.7 86 0.8 2.0 1.4 86 0.4 Test 2 2.9 1.5 86 0.6 2.5 1.4 86 0.7 Test 3 2.9 1.4 87 0.6 2.2 1.3 87 0.7 y: n is the number of measurements; z: s is the standard deviation

PAGE 138

138 Table 44 Maximum and average pulptemperature differences between the half radius distance and the center (c) of fruit F1, F3 and F42 for the second series of tests. Bottom Corner Center Top Corner F1 F1c ( C) F23 F23c ( C) F42 F42c ( C) Maximum Average n y s z Maximum Average n s Maximum Average n s Test 1 0.6 0.2 173 0.2 0.3 0.1 173 0.1 0.7 0.4 173 0.1 Test 2 0.6 0.2 178 0.2 0.3 0.1 178 0.1 1.3 0.7 178 0.3 Test 3 0.6 0.1 179 0.3 0.3 0.2 179 0.0 1.1 0.7 179 0.3 y: n is the number of measurements; z: s is the standard deviation Table 45. Maximum and average differences between the air temperature near the surface of the fruit (a) and the pulp temperature at the half radius distance of fruit F1, F23 and F42 for the second series of tests. Bottom Corner Center Top Corner F1a F1 ( C) F23a F23 ( C) F42a F42 ( C) Maximum Average n y s z Maximum Average n s Maximum Average n s Test 1 2.9 0.9 173 0.8 3.3 1.1 173 1.1 9.2 2.9 173 1.9 Test 2 2.1 1.0 178 0.5 2.7 1.6 178 0.5 9.6 3.5 178 2.2 Test 3 2.9 1.1 179 0.8 2.9 1.1 179 1.0 9.8 3.4 179 2.1 y : n is the number of measurements; z : s is the standard deviation Table 46. Maximum and average pulptemperature differences between the half radius distance and the center (c) of fruit F1 for the third series of tests. Bottom Corner F1c F1 ( C) Maximum Average n y s z Test 1 2.4 1.3 176 0.5 Test 2 2.6 1.6 177 0.5 Test 3 2.4 1.6 176 0.5 y : n is the number of measurements; z : s is the standard deviation

PAGE 139

139 Table 47. Maximum and average differences between the air temperature near the surface of the fruit (a) and the pulp temperature at the half radius distance of fruit F1, F43 and F72 for the third series of tests. Bottom Corner Center Top Corner F1a F1 ( C) F43a F43 ( C) F72a F72 ( C) Maximum Average n y s z Maximum Average n s Maximum Average n s Test 1 1.8 0.8 176 0.4 0.6 0.2 176 0.2 10.2 2.2 176 2.2 Test 2 1.9 0.8 177 0.5 0.7 0.2 177 0.2 10.3 2.6 177 2.2 Test 3 1.8 0.7 176 0.4 0.5 0.2 176 0.2 9.6 2.6 176 2.2 y : n is the number of measurements; z : s is the standard deviation Table 48. Biot number (Bi) calculations for a single spherical fruit with constant properties exposed to natural convection in air. D Ra D x Nu D x, y h x Bi z m ( ) ( ) (Wm 2 K 1 ) ( ) Series 1 0.0668 1.06 10 6 16.6 6.23 0.155 Series 2 0.0834 2.05 10 6 19.3 5.78 0.194 Series 3 0.0728 1.36 10 6 17.6 6.05 0.169 x: Calculations based on a film temperature of 18.75C; y: Calculations based on Equation 41; z: Calculations based on a average convective heat transfer coefficient of 6 Wm2K1 and a thermal conductivity of 0.431 Wm1K1 for the fruit.

PAGE 140

140 CHAPTER 5 RESULTS AND DISCUSSION OF AIR TRANSPORT TESTS T ests were conducted through four different airports: Dubai International Airport (DXB) in Dubai (United Arab Emirates), Gothenburg Landvetter Airport (GOT) in Landvetter (Sweden), John F. Kennedy International Airport (JFK) in New York (United States) and Jomo Kenyatta International Airport (NBO) in Nairobi (Kenya) The threeletter identifiers presented in parenthesis are the International Air Tra nsport Association (IATA) codes; t hese codes are used throughout this section to identify the airports. Using JFK as a reference, the time difference for each airport when the tests were conducted was + 8 h, + 6 h and + 7 h for DXB, GOT and NBO, respectively Local times, in 24 h notati on, are used throughout this section. All airport local weather conditions were obtained from historical data avail able through the following website: http://www.wunderground.com/history/ To present and analyze the results, each shipment was divided in three sections. The first section corresponds to the period during which the instrumented container was outside on the tarmac during the ramp transfer prior to departure. The second section includes the entire period the container was onboard the aircra ft; on the ground and in flight. The third section corresponds to the time during which the container was outside on the tarmac during the ramp transfer after arrival. Since the tests were conducted through multiple time zones, it was decided that the simplest approach was to use an elapsed time scale for the graphs. The reference time (t = 0 h ) was therefore selected as the moment when the instrumented container was taken out of refrigerated storage to be brought out on the tarmac in preparation for d eparture.

PAGE 141

141 Data analysis showed small temperature variations between the reusable plastic containers ( RPCs ) on a specific layer. This observation can be explained by the loading pattern of the container and the locations of the temperature sensors in the l oad. Consequently, only the average temperatures of the layers of products are presented and discussed in this section. For all figures the prefix "A" identifies an average value. There were seven layers of RPCs in the instrumented aircraft container; l ayer 1 and layer 7 denote the bottom and the top layer respectively The six walls of the aircraft container are numbered as previously shown in Figure 4 35. The following notation is used in this chapter. ALx: Average temperature of the products calcu lated from the experimental data obtained in layer x of the aircraft container, where x Wy: Temperature measured at the center of the inside surface of wall y of the aircraft cont ainer, where y {16} ( Figure 435) P: Atmospheric pressure 5 .1 JFK GOT DXB T he departure time from JFK airport was scheduled at 11:00 on 04292007 and its arrival time at DXB airport was schedule d at 11:00 the next day. A stop at GOT airport was made before the flight reached its final destination. The container w as carried in position 44R in the aft cargo compartment of a B747 400F, a full freighter aircraft (Figure 3 19) 5.1.1 JFK Ramp Transfer At 10:1 0, the container was taken out of refrigerated storage and brought out onto the tarmac in preparation for the loading operations The inst rumented container was on the tarmac for a period of 1 h 50 min, until 12:00. T he average temperatures of the layers of products and the temperatures of the inside wall s of the aircraft container are

PAGE 142

142 presented in Figure 51 During that time the environmental conditions at JFK airport were overcast with an ambient temperature of 14.4C at 09:51 and 16.1C at 11:51. Initially, the average temperatures of the layers or products varied between 7.5C (AL1) and 8.9C (AL3). The average temperatures of the layers of products were constant during the ramp transfer except for the top layer (AL7), which increased slightly from 7.6 C to 8.3 C. Th e environmental conditions during the ramp transfer at JFK airport did not affect sig nificantly the distribution of the product temperatures within the aircraft container. Table 51 complements Figure 51 and presents the initial and final temperatures of the walls of the container U nder overcast conditions during day time the walls ra pidly reached a temperature close to the outside air except for the bottom and left wall s (W3 and W4). The final temperature of the left wall (W4) was 12.6C, a few degrees below ambient temperature. The best explanation for this behavior is the proximi ty of the left wall to another container. Regarding the bottom wall temperature, the situation is different. Since the load sits directly on the bottom wall, that temperature is strongly influenced by the load and this explains why the temperature only reached 10.6C at that location. The larger thickness of the bottom wall also contributes to its different thermal behavior Since most ULDs are not designed to be handled with forkl ift, they rarely si t directly on the ground when fully loaded. D uring ramp transfer operations, unit load devices (ULDs) are placed onto dollies. These dollies are equipped with locking mechanisms to maintain the ULDs in place during transport. On the dollies roller syst ems allow to move and rotate ULDs. Usually the dollies are somewhat open

PAGE 143

143 structures and expose most of the bottom surface of the ULDs to convective heat transfer as well as radiative heat transfer with the tarmac. During cold storage ULDs are also usuall y placed on roller systems (balls or cyl inders) to facilitate handling. 5.1.2 Onboard the Aircraft The conditions during the entire time the instrumented container was on the plane are presented in Figure 52 The p ressure measured inside the aircraft container, is plotted on each graph of Figure 52 to indentify the segments during which the plane was on the ground and in flight. Inside the aircraft, t he instrumented container was loaded near the aft cargo compartment door, at position 44R (Figure 319). The temperature set point of the aft cargo compartment of the Boeing 747F was set by the pilot at 10C for both flight segments (JFK GOT and GOT DXB). The average pressure during the flight segments was 0. 84 atm and 0. 83 atm respectively The average temperatures of the layers in Figure 52 clearly show that the timetemperature conditions to which the container was exposed onboard the aircraft did not significantly affect the temperature distribution within the products In the absence of solar radiation, t he temperatures of the walls of the container provide good indication of the temperature of surrounding air. Figure 52 shows that only moderate temperatures variations were measured during the time the container was onboard the plane. At JFK airport after the container was loaded and the cargo doors were closed, the air temperature spiked just below 20C. At that time, the top wall temperature (W1) was measured at 19.2C while the lateral walls reached an average temperature of 16.8C. Again, the floor of the container (W3) remained colder at 11.0 C Shortly after that sharp increase, the tem perature started to decrease as the air conditioning system was

PAGE 144

144 started with the engines of the aircraft. After 1 h 10 min of taxiing, the temperatures of the wal ls of the container decreased close to the set point of 10 C. Just prior to takeoff the temperatures ranged from 10.1C to 12.5 C. During the first segment of flight, the temperatures of the walls were steady, they ranged between 5 .9 C (W4) and 9.6C (W2) The left wall (W4) was facing the cargo compartment d oor where colder temperatures are usually encountered because of weaker thermal insulation. The right wall (W2) was located near the center of the cargo compartment parallel with the longitudinal axis of the plane and separated by only a few centimeters from the adjacent container. The fact that cargo compartments ventilation systems are designed to move air laterally combined with the small air gap between aircraft container s creates a weaker airflow near the right wall (W2). This explains why the temperature at that location was higher and closer to the temperature of the load. The flight segment between JFK and GOT took 7 h. During the ramp transfer at GOT airport, the instrumented container remained onboard the plane. The plane landed at 2:10 and departed at 4:30. During that 2 h 20 min layover, the ambient temperature at the airport was 6 C. Table 52 presents the wall temperatures of the instrumented container at l anding and takeoff at GOT airport It can be noticed that only small increases in temperatures occurred during that period; the maximum increase was 2.5C and was measured on the front wall of the container (W5). It is difficult to explain the cause of t he small spike in temperature up to 11.8C of the top wall of the container (W1) shortly after landing (Figure 5 2 ). In general, for most temperature sensitive product, the conditions encountered during that stop would not have been detrimental.

PAGE 145

145 During th e second flight segment from GOT to DXB airport the temperatures of the walls of the instrumented container were again steady. T emperatures varied between 5.8C (W4) and 10.0C (W2). These temper atures were almost identical to those observed in the first flight segment. Upon arrival at DXB airport, the wall temperatures increased and converged to an average of 11.1C with the except ion of the bottom wall temperature (W3) which increased only to 7.1C. The flight from GOT to DXB was 5 h 40 min and landed a t 12:10. 5.1.3 DXB Ramp Transfer At 12:30, 20 min after arrival, the instrumented container was unloaded from the aircraft. At 12:00 and 13:00 the ambient temperature at DXB airport was reported at 34.0 C and 33.0C with clear sky conditions. Figure 53 shows the average temperature of the layer s of product and the temperatures of the inside walls of the container during the ramp transfer at DXB airport. It is important to mention that different temperature scales were used for the two graphs of Figure 53 The container remained on the tarmac for a period of 1 h and was exposed to solar radiation for a p eriod of approximately 30 min. Table 53 summarizes the variations of the wall temperatures during that period. The bottom wall temperature (W3) rose from 7.1C to 15.3C. The temperatures of the lateral walls of the contai ner (W2, W4, W5 and W6) i ncreased to an average of 35.0C, which was very close to the ambient air temperature. Due to direct solar radiation, the top wall temperature (W1) increased by 51.0C and reached 62.3C; this is 28.3 C above ambient air temperature. At that time, the air temperature in the top layer of product reached 20.4C. Figure 53 shows that only the temperature of the top layer of product s was influenced during the 1h ramp

PAGE 146

146 transfer. The average temper ature of the top layer (AL7) increased from 8.8C to 10.3C. It must be mentioned that for this specific flight the time spent by the experimental container on the tarmac upon arrival in DXB airport may have been shorter than usual since the ground personnel were aware that it was instrumented Exceptional measures were taken during the ground operations As at JFK airport, the airline refrigerated storage facilities were located onsite at DXB airport directly within the cargo terminal. 5.1.4 Time and Delays In the end, the total transit time from cold room to cold room was 19 h 20 min. A to tal of 2 h 50 min were spent on the tarmac wher e as 16 h 30 min were spent on the plane The actual inflight time for the two segments summed up to 12 h 40 min, which corresponds to 65.5% of the total transit time. The flight departed with a delay of 2 h 10 min and arrived at destination 1 h 10 min behind schedule. Upon arrival, the load was inspected and neither shifting nor mechanical damage were observed. 5.2 First S hipment DXB NBO The flight was scheduled t o depart from DXB airport on 05032007 at 10:05 and land at NBO airport the same day at 14:15 after a 5 h 1 0 min flight. T he instrumented container was loaded onboard an A330200 passenger aircraft. The container was located in position 41L of the aft cargo compartment (Figure 319). 5.2.1 DXB Ramp Transf er At 1:40, 8 h 25 min prior to departure, the container was taken out of the cold room. At 1:00 and 2:00, the ambient air temperature at DXB airport was 32.0 C and 29.0C respectively with clear sky conditions The fact that a unit load device (ULD)

PAGE 147

147 was out of the cold room more than 8 h prior departure is evidence of the challenges associated with maintaining a proper cold chain for air shipments. Figure 54 shows the variations of the average temperatures of the layer s of products as well as the i nside temperatures of the wall s of the container while on the tarmac. T wo different temperatures scales were used on the graphs of Figure 54 t o obtain a better visual representation of the temperature variations. At the moment the container was taken out of the cold room, the inside wall temperatures were quite uniform, they varied between 5.5C and 6.7 C. Once exposed to the outside conditions, the wall temperatures remained relatively steady for the first 4 h and then started to increase until the container was loaded onboard the aircraft. That increase in temperature corresponded precisely with sunrise (5:41, t = 4.02 h) The ambient air temperature at DXB airport 30 min before the container was loaded onboard the aircraft was reported to be 31.0C. The wall temperatures while the instrumented container was out on the tarmac are summarized in Table 54. During that same period, t he average temperatures of the layers of products increased at different rates, but almost linearly (Figure 5 4) Th e initial (t = 0) and final (t = 8 h) temperatures and the corresponding variations are presented in Table 55 Initially, the temperatures of the products in the container were quite uniform, they varied between 6.0C and 7.0C for individual measurement s and between 6.2C and 6.6 C for the layer averages. At the end of the 8h ramp transfer, the top layer (AL7) showed the largest temperature increased reaching 16.1C In Figure 54, a vertical temperature stratification among the layers can be observed with the exception of the bottom layer (AL1) which behaved similarly to the fifth layer. The convective and

PAGE 148

148 radiative heat transfer through the bottom wall of the container contributed to the larger temperature increase of the first layer of products The average temperatures of the bottom core layers of the load ( AL2, AL3 and AL4) remained relatively constant and did not increase by more than 1.0C. 5.2.2 Onboard the Aircraft The instrumented container was onboard the aircraft for a period of 1 h pri or departure at 10:30 (t = 8.83 h) The flight from DXB to NBO airport was 4 h 30 min. Upon arrival to NBO airport at 14:00 (t = 13.33 h) the container remained on the plane fo r 30 min before being unloaded. The temperatures of the inside walls of the container as well as the average temperatures of the layers of products for the entire time the container was onboard the plane, are presented in Figure 5 5 The atmospheric pressure is also plotte d on the graphs to identify the flight segment. T here w ere two pressure plateaus during the flight T he average pressures during the first and second plateau were 0.88 and 0.79 atm respectively Also, the atmospheric pressure upon arrival at NBO airport remained low at 0.84 atm since the city of Nairobi has an elevation of 1624 m with respect to sea level and a standard pressure of 8. 33 k Pa, or 0.822 atm (ASHRAE, 2001 ) The wall temperatures in Figure 5 5 indicate that the ambient temperature within the cargo hold of the passenger plane (A330200) was maint ained around 20C. This is a typical temperature whenever live animals are carried onboard an aircraft ; it is a frequent occurrence on passenger flights because of the presence of pets. The variations in the average temperatures of the layers of product while the container was onboard the aircraft are summarized in Table 5 6 These small increases a re attributable in part to the ambient conditions while the container was within the cargo compartment and in part to the transfer of the heat

PAGE 149

149 accumulated in the outer regions of the load while the container was on the tarmac in DXB. 5.2.3 NBO Ramp Transfer At 14:00 and 15:00 the ambient temperature s at NBO airport were reported at 23.0 C and 24.0C respectively with mostly cloudy conditions. At 14:30 (t = 13.83 h) the instrumented container was unloaded from the plane The temperatures of the inside walls of the container as well as the average temperatures of the layers of products during the period the container was on the tarmac are presented in Figure 5 6 Differen t temperature scales are used for the graphs in Figure 5 6. While on the tarmac the solar radiation combined with the partially cloudy conditions affected the wall temperatures of the container, particularly the top wall (W1). In F i gure 5 6 the temperature of the top wall of the container (W1) peaked at four occasions reaching 55.3C, 57.3 C, 58.3 C and 50.3C within th e first 2 h. Depending on their orientations with respect to the sun at a given time, the lateral surfaces of the container exhibited a similar thermal behavior but with lower magnitudes. Taking into account the fact that the ambient temperature was only 23C, this illustrates clearly the important effect that solar radiation may have on the surface temper atures of an aircraft container or on the outer layer of products loaded onto an aircraft pallet. Again, the bottom wall temperature stayed relatively steady at an average temperature of 13.3C. F or the last part of the ramp transfer, the conditions remained cloudy or the container was placed into a shady area. T he conditions encountered during the ramp transfer did not influence significantly the temperature of the load. This is can be explained in part because the walls of the container reached high temperatures intermittently and in part by the smaller

PAGE 150

150 temperature difference between the wall and the load. During the 4 h th e container was on the tarmac the top layer of the container showed the largest increase on average temperature with 1.3 C ( Figure 5 6 ) A s it is the case for several airports, the airline cold storage facilities were located a few blocks from NBO airport In this case, the ULDs had to be transported for a few minutes over poorly paved roads. S hocks and vibrations associated with that shor t transport operation were severe enough to cause some shifting of the load within the aircraft container. 5.2.4 Time and Delays The total time for the container to travel from the cold room at DXB airport to the cold room in NBO was 17 h 50 min. The total time spent on the tarmac at DXB and NBO airport was 11 h 50 min, which corresponds to 66.4 % of the total time. The container spent 6 h onboard the aircraft and was infl ight for a period of 4 h 30 min, which is 25.2 % of the total travel time. T here were no major delays with respect to the flight schedule. 5.3 First S hipment NBO DXB After its arrival at NBO airport from the previous flight (DXB NBO, 05 032007) the instrumented container was stored in a cold room maintained at 5C for a period of approximately 25 h. Table 57 presents the change on average temperatures of the layer s of products during refrigerated storage T he average temperatures of all layers decreased except for a small increase in the second layer (AL2). The initial temperatures of the second layer was lower than both the first (AL1) and third (AL3) layers of products. The presence of warmer products above and below the second layer caused the slight increase in temperature of the second layer

PAGE 151

151 At the end of the cold storage period, the average temperatures of the seven layers were between 11.1C (AL7) and 8.4 C (AL3), a minimum of 3.4 C above the set point of the cold room. The temperatures of layers 2 to 6 varied by less than 4C over 25 h. This clearly illustrates the small rate of heat transfer within the load of products, particularly for the products located near the center. For this shipment, the instrumented container was scheduled to fly back to DXB airport on a cargo aircraft. The fl ight was schedule d to depart from NBO airport at 23:00 on 05042007 and arr ived the next day at 5:00 after a 5h flight The container was loaded in position 11L within the forward cargo compartment of an A310300F aircraft (Figure 3 19). The position a djacent to the instrumented container (11R) which is located near the door of the cargo compartment was vacant for the f l ight. 5.3.1 NBO Ramp Transfer The container was brought out of the cold room and onto the NBO airport tarmac at 19:30 (05 04 2007), 30 min after sunset. T he ambient air temperatures at NBO airport were 20.0C and 21.0 C at 19:00 and 20:00, respectively S cattered cloud conditions were reported. The container stayed on the tarmac for a period of 2 h 30 min It was loaded onboard the aircraft at 22:00 At that time the ambient air temperature had dropped to 18C and the sky was clear The temperatures of the walls of the container and the average temperature s of the layers of products within the container are s hown in Figur e 57 The average temperatures of the layers of products were not affected by the ambient conditions; they remained steady for the entire ramp transfer. I n the absence of solar radiati on, all wall temperatures reached values close to the ambient air temperature, except for the bottom wall (W3).

PAGE 152

152 5.3.2 Onboard the Aircraft The temperatures as well as the pressure measured while the container was onboard the aircraft are presented i n Figure 58 The container was onboard the plane fo r a period of 1 h prior to departure. T he flight from NBO to DXB airport was 4 h 30 min and during that time the av erage pressure was 0.82 atm. The pressure did not plateau at two levels as it was observed in the pr evious flight (DXB NBO 05032007). A t departure (23:00, t = 3.5 h) the top wall of the container (W1) was at 19.3C whereas the lateral walls (W2, W4, W5 and W6), varied between 15.5C and 16.7C. Upon arrival at DXB airport (4:30, t = 8 h) the top wall temperature had decreased to 15.5C and the temperatures of lateral walls ranged between 12.4C and 13.9C The temperature of the bottom wall of the container was not included in Figure 5 8 since the temperature probe appeared to have been damaged by the harsh handling conditions during t he transport of the container from the cold storage facilities to the airport. Because of the small temperature difference s between the walls of the container and the load the average temperatures of the layers of products were relatively steady while t he container was onboard the aircraft. The maximum increase was measured at the top layer, where the average temperature (AL7) rose by 0.9C over a period of 6 h. Again, w ith the exception of the bottom layer of product (AL1), a vertical stratification o f the average layer temperatures can be observed in Figure 58 5.3.3 DXB Ramp Transfer The instrumented container was taken out of the plane 30 min after landing (5:00, t = 8.5 h) at DXB airport. At that time, the conditions were partly cloudy and the am bient temperature of 25.0 C was reported at the airport The container was left on the tarmac

PAGE 153

153 until 9:20, a period of 4 h 20 min. From 9:00 to 10:00, the ambient temperature increased from 30 C to 34C and the sky remained clear. Figure 5 9 presents the temperatures of the walls of the container as well as the average temperatures of the layers of products during the ramp transfer. Different temperature scales are used for the two graphs of Figure 59 The temperatures of the walls of the container increased for first 2 h, and then sharply decreased and remained relatively steady until the container was brought into the cold room This indicates that after 2 h, the container was placed in a shady area. If the container had been exposed to solar radiation during the entire ramp transfer previous results and observations showed that the top wall (W1) would have reached temperatures above 50C. On that day (05 052007), the sun rose 40 min after the container was unloaded from the aircraft (5:40, t = 9.17 h) From Figure 59 it can be observed that the product temperatures increased only slightly since the container was not exposed to extreme ambient temperatures or peak solar radiation level s. The variations of the average temperatures of the layers of products are summarized in Table 58. Again, the top layer (AL7) showed the largest variation; its average temperature increased from 12.2C to 15.8C. 5.3.4 Time and Delays For this flight, the total transit time between the col d rooms at NBO and DXB airports was 12 h 50 min. A total of 6 h 50 min was spent on the tarmac, which corresponds to 53 .2 % of the entire transit time. The instrumented container spent the remaining 6 h of its journey (46.8%) within the cargo compartment of the aircraft and was in flight for a period of 4 h 30 min ( 35.1 % ). The flight departed on time and even arrived 30 min ahead of the scheduled arrival time.

PAGE 154

154 5.4 Second S hipment DXB NBO The instrumented container was scheduled to ship from DXB airport on 05062007 at 10:05, on the same flight as the first shipment to NBO airport (DXB NBO, 05032007) The aircraft was scheduled to reach destination at 14:15 after a flight of 5 h 10 min. T he container was again loaded in the aft cargo compartment but in position 32R this time (Figure 319). 5.4.1 DXB Ramp Transfer As it was the case for the first DXB NBO shipment, the container was taken out of refrigerated storage (5 C) at 1:40, 8 h 25 min before the scheduled departure time. It is dif ficult to confirm that this is a standard practice, but this shows that ULDs are likely to spend extended period of time on tarmac during ramp transfers. At 2:00, 20 min after the container was bought out onto the tarmac, the ambient temperature was 30C. The sun rose 4 h later at 5:39. The instrumented container was left on the tarmac until 9:40 when it was loaded onboard the aircraft. At 10:00, the ambient temperature had increased to 34C. During the entire ramp transfer, the sky remained clear. The temperatures of the walls of the container as well as the average temperatures of the layers of products during the 8h period the container was on the ramp are presented in Figure 510. As for the first DXB NBO flight, the temperatures of the walls of the container were relatively steady until sunrise. With the increase of solar radiation, wall t emperatures rose significantly until the container was loaded onboard the aircraft. Table 59 provides the temperatures of the walls of the container when it was taken out of the cold room, at sunrise and when it was loaded. For this entire shipment, t he bottom wall temperature (W3) was unavailable since the temperature probe was damaged in the previous flight.

PAGE 155

155 From Figure 510 it can be observed that the initial temperatures of the product were not as uniform as for the first flight to NBO. The average temperatures of the seven layers of products varied between 9.1 C (AL3) and 11.4C (AL7). During the ramp transfer, the average temperatures of the layers increased steadily and displayed the same vertical stratification that was observed in previous flights. The variations of the average temperatures of the layers of products during the ramp transfer are presented in Table 5 10 5.4.2 Onboard the A ircraft At 9:40, after 8 h on the tarmac, the instrumented container was loaded in the aft cargo compartment The plane departed at 11:10 (t = 9.5 h) and landed at NBO airport at 14:40 (t = 14.0 h) The flight duration was 4 h 30 min. The container stay ed on the plane for 1 h after arrival. The pressure and temperatures measured while the container was onboard the aircraft are shown in Figure 511. Similarly to observations during the previous flight to NBO airport the pressure successively pla teaued at 0.86 atm a nd 0.79 atm during the flight. The temperature of the walls did not var y drastically during the flight. The temperatures of the lateral walls of the container (W2, W4, W5 and W6) were quite uniform; they were within 3.2 C of each other. The average temperature of the four lateral walls at departure (t = 9.5 h) and arrival ( t = 14 h) was 23.2C and 17.7C respectively The temperature of the top wall (W1) at departure and arrival was measured at 32C and 23 C respectively These temperatures indicate again that the temperature set point of the aft cargo compartment was close to 20C. The temperature of the product increased slightly while the container was onboard the aircraft. The variations of the average temperatures of the l ayers of products are

PAGE 156

156 presented in Table 511 The results showed again the vertical stratification of the product temperatures 5.4.3 NBO Ramp Transfer At 15:40 (t = 15 h), the container was unloaded from the aircraft. At 16:00, t he ambient temperature at NBO airport was 23.0 C with scatteredcloud conditions. The sporadic exposure to solar radiation created three successive peaks in the temperatures of the container walls ( Figure 512). During those events, t he top wall (W1) reached the highest temper atures at 31.2C, 45 .1C and 32.8C. After the last peak, the wall temperatures decreased and converged progressively as the sun was setting. At sunset (18:30, t = 17.83 h), the temperature of the walls averaged 19.8C and ranged between 18.9C (W4) and 21.1C (W1). Just b efore the container was brought inside the cold room, the temperatures of the wall had decreased slightly. They averaged 18.5C and ranged between 17.5C and 19.7C. Figure 512 also shows that the average temperatures of the layer of products did not vary significantly during the ramp transfer at NBO airport. 5.4.4 Time and Delays After the container was initially taken out of the cold room at DXB airport, it took 21 h 50 min before it was again stored in a controlled temperature environment at NBO airport. From the tot al transit time, 14 h 50 min (67.9%) was spent on the tarmac during ramp transfer whereas the container was i n flight for only 4 h 30 min (20.6%). No major delays were observed during the shipment. The flight departed 1 h 5 min late but arrived only 25 min behind schedule at NBO airport.

PAGE 157

157 5.4.5 Comparison with the Previous DXB NBO Flight It is interesting to compare the results of the two flights from DXB to NBO since the container was carried on the same flight and the same t ype of passenger aircraft (A330200). It can be noted by comparing the temperature variations between the first ( Figures 54 to 56 ) and second ( Figure 510 to 512) flight that very similar results were obtained. On e important thing to notice is that even though the container was loaded in different positions in the aft cargo compartment the temperatures of the walls of the container indicated that it was exposed to very similar conditions. More discrepancy in the results would have been expected if the container was located near a cargo door for one of the two flights. The logistics of the handling operations were also very similar between the shipments. In both cases, the instrumented container was taken out of refrigerated storage for appr oximately 8 h before departure. Luckily, half of that period occurred before sunrise, which may be the reason why such an ex tended break in the cold chain was allowed. Nevertheless, the temperature of the products, particularly those in the top layers of the load increased significantly during both ramp transfers at DXB airport. The temperature of the aft cargo compartment during the flights was also too high (about 20C) to provide any cooling to the load of products. It can be concluded that for most temperature sensitive products, the two shipments from DXB to NBO airport can be considered as 15 h 40 min and 21 h 50 min c old chain breaks. 5.5 Second S hipment NBO DXB Prior to this shipment, the container was stored for 17 h in a cold room set at 5C. The initial and final average temperatures of the layers of products are presented in Table 512.

PAGE 158

158 Unlike the previous flig ht from NBO to DXB, the container was flown onboard a passenger aircraft (A330200), the same model of aircraft than the two previous flights from DXB to NBO. The container was loaded in the forward cargo compartment in position 14L (Figure 319). The fl ight was planned to leave NBO airport at 17:15 (05 072007) and arrived the same day at 23:15, after a 5h flight. For this entire shipment, the bottom wall temperature (W3) was unavailable since the temperature probe was damaged in a previous flight. 5.5 .1 NBO Ramp Transfer The container was taken out of cold room and brought onto the tarmac at 15:30, 1 h 45 min prior to the schedule departure. At that time, the average temperatures of the seven layers of products varied between 9.8C (AL1) and 14.3 C (A L7). The temperatures were stratified from the bottom to the top layer. The container stayed on the tarmac for 50 min before being loaded onboard the aircraft. During the ramp transfer, the conditions at NBO airport were mostly cloudy and the ambient temperature was 24.0C. Figure 5 13 presents the variations of the temperatures within the container d uring that period. Even though the temperature of the top wall of the container (W1) reached 46.9C and was above 41C most of the ramp transfer, the temperature distribution within the load remained steady. D ifferent temperature scales were used in Figure 513. 5.5.2 Onboard the Aircraft The container was loaded onboard the plane at 16:20 (t = 0.83 h), 1 h before the a ctual departure (17:20, t = 1.83 h). After a flight of 4 h 20 min, the plane landed at DXB airport at 22:40 (t = 6.17 h). The i nstrumented container stayed onboard the aircraft for an additional 30 min before being unloaded at 23:10. The pressure as wel l

PAGE 159

159 as the temperatures measured while the container was onboard the aircraft are shown in Figure 514. Shortly after takeoff, the temperatur es of the wall of the container decreased. For most of the flight, the lateral walls remained between 10C and 15C. The top wall of the container (W1) remained at higher temperatures. Its temperature was measured at 24.1 C and 15.7C at departure and arrival respectively and it reached a minimum of 14.7C during the flight. Since the temperature differences between the products and the walls of the container were small, the average temperatures of the layers of products were relatively constant while the container was onboard the aircraft. As it was the case for the two previous shipment s onboard the A330 200 ai rcraft between DXB and NBO airports, two pressure plateaus were measured during flight. In this case, the plateaus were measured at 0.81 atm and 0.85 atm. This phenomenon may be due to the model of aircraft or to the flight crew who for different reasons may decide to regulate the pressure differently depending on the cruising altitudes associated with that route. 5 .5.3 DXB Ramp Transfer It was 23:10 (t = 6.67 h) at DXB airport when the container was unl oaded and placed on the tarmac It stayed there until 10:00 (t = 17.5 h) the next day, a period of 10 h 50 min. During the ramp transfer, there were clear sky conditions and the ambient temperature varied between 28C and 31C. The temperature variations in the instrumented container are presented in Figure 5 15 The wall temperatures measured during the ramp transfer were relatively constant and did not exhibit the sharp increase associated with a direct exposure to solar radiation that was observed during the ramp transfers of previous shipments Except for the first and last segment of the ramp

PAGE 160

160 transfer the wall temperatures varied between 20.3C and 27.7C. T his indicates that the container was stored in a shady area or within a nonrefrigerated warehouse. Until the arrival at DXB airport, the t emperature distribution within the load of products had remained relatively constant. For that shipment, i t is during the ramp transfer at DXB airport that the temperature of the products increased the most ( Figure 5 15 and Table 513). The average temperatures of the layers increased almost linearly. With the exception of the bottom layer (AL1), the temperature distribution of the layers of products was again vertically stratified. The difference in temperature between adjacent layers increased from th e bottom to the top layers. 5.5.4 Time and Delays The total duration of the shipment from the cold room at NBO airport to the cold room at DXB airport was 17 h 30 min. The instrumented container was on the tarmac for 11 h 40 min, which corresponds to 66. 7% of the total transit time. It is important to mention that 93% of the time the container spent on the tarmac was at DXB airport during the extended overnight ramp transfer. T he flight time was 4 h 20 m in and accounted for only 24.8% of the total trans it time. No delay s were encountered during that shipment. 5 .6 Shipment DXB GOT JFK The flight was scheduled to depart from DXB airport at 23:00 on 05122007 and to arrive the next day at 8:00 at JFK airport after a 17h transit time that included a stop at GOT airport. Similar to the JFK GOT DXB flight, the container was transported onboard a B747400F. It was loaded again in the aft cargo compartment, but this time it was located in position 43L (Figure 319). For this entire shipment, the bottom wall

PAGE 161

161 temperature (W3) was unavailable since the temperature probe was damaged in a previous flight. 5.6.1 DXB Ramp Transfer Since the flight was delayed, the container was brought out of the 5 C cold room onto the tarmac at 22:40. At 23:00, the ambient temper ature was still at 32.0C with clear sky conditions. The container was loaded onto the plane 1 h 20 min later at 00:00. The temperatures of the walls of the container and the average temperatures of the layers of products during the time the container was on the tarmac are pres ented in Figure 516. After a period of 20 min on the tarmac, the top wall of the container (W1) reached 27.9C whereas the lateral walls (W2, W4, W5 and W6) varied between 21.7 C and 24C. These temperatures decreased slowly throughout the ramp transfer After 1 h 20 min, when the container was loaded inside the aft cargo compartment, the temperature of the top wall of the container (W1) had dropped to 26.5 C whereas the lateral walls (W2, W4, W5 and W6) varied between 18.4C and 20.0 C. 5.6.2 Onboard t he Aircraft The flight departed at 1:40 (t = 3 h), approximately 1 h 40 min after the loading operation was completed. Figure 517 displays the pressure and the temperatures measured during the entire time the container was onboard the aircraft. While the plane was still on the ground, the temperature of the top wall container increased up to 29.9C. The temperatures of the lateral walls increased between 19.7C and 21.5C. At departure, the air conditioning system was started with the engines and the temperature in the cargo compartment started to decreas e. For the two flight segments, the pilot set the temperature of the aft cargo compartment at 20C. Once in flight, the temperatures of the walls did not vary significantly. The top wall temperature (W1) was

PAGE 162

1 62 again the highest; it averaged 19.5C. Three of the lateral walls (W2, W4 and W5) were almost identical throughout the first flight segment; their average was calculated at 14.7C. The average temperature of the back wall of the container ( W6) was 16.8C. That first segment took 6 h and the aircraft landed at GOT airport at 5:40 (t = 9 h). The average inflight pressure was 0.84 atm. The ambient temperature at GOT airport was 10C, which explains the decrease of the wall temperatures measured during the stop as the aft cargo door was opened. After the 2 h stop, t he aircraft departed from GOT airport at 7:40 (t = 11 h) for the second flight segment leading to JFK airport The temperatures of the walls of the container were similar to the previous segment, but laid within a slightly narrower range. The top (W1) and back (W6) wall temperatures averaged 18.1C and 17.3 C. Again, the three other lateral walls (W2, W4 and W5) displayed almost identical te mperatures during the flight. T heir average was calculated at 15.4C The average pressure was calculated at 0.83 atm during the flight. The aircraft landed at JFK airport on 05132007 at 9:10 (t = 18.5 h), 7 h 50 min after its departure from GOT airport. Upon arrival, the container stay ed onboard for 50 min before being unloaded onto the tarmac. For the entire period of time the container was onboard the aircraft, the average temperatures of the layers of products increased at different rates but they all followed almost linear trends The only exception is the temperature of the top layer of products (AL7). It displayed a faster rate of change while the aircraft was on the ground at DXB airport and that the top wall temperature reached 29.9C ( Figure 517 ) The variations of the ave rage layer temperatures are summarized in Table 514.

PAGE 163

163 5.6.3 JFK Ramp Transfer At JFK airport, the instrumented container stayed on the tarmac for a period of 2 h 20 min, between 10:00 and 12:20 (0513 2007). Throughout the ramp transfer, the container was exposed to clear sky conditions and the ambient temperature varied between15.6C and 17.2C. For the first 30 min, the effects of solar radiation were observed on the temperatures of the walls of the container and t he top wall temperature (W1) reached a maximum of 36.3C (Figure 5 18) For the last part of the ramp transfer, the container was placed in a shady area and the wall temperatures were steady. The top (W1) and the lateral (W2, W4, W5 and W6) wall temperatures averaged 18.4C and 15.7C respectively Figure 518 also shows that the temperatures of the product were not affected during the ramp transfer at JFK airport. 5.6.4 Time and Delays The total transit time between the cold rooms at D XB and JFK airports was 21 h 40 min. Only 16.9% ( 3 h 40 min) of the total transit time was spent on the tarmac during ramp transfers. The container was onboard the plane for the remaining 18 h and spent 13 h 30 min inflight, which corresponds to 62.3 % of the total journey. Important delays were encoun tered during the shipments. The flight departed from DXB airport 2 h 40 min behind schedule and consequently was 2 h late upon arrival at JFK airport. It is important to mention that the ground crew at DXB airport were aware of the delays and accordingly adjusted the tim e at which they brought out the ULDs prior to the loading operations. This contributed to the short period of time the instrumented container spent on the tarmac when compared to previous flights of this series of tests.

PAGE 164

164 5.7 Air Transport Logistics and Detrimental Temperatures during Tests The results of the inflight tests confirmed that air transport may contribute to the decrease of the quality o f fresh horticultural products. A combination of the air transport logistics and environmental conditions can be identified as the principal cause of the problems. Looking at the air transport logistics, flight delays are often identified as a major issue (Villeneuve et al. 2002). Any delays are important for temperature sensitive cargo but delays on departure times are critical since the ULDs may have to stay for a longer period of time on the tarmac. T able 515 presents th e delays encountered during the six in flight tests. Looking at departures times the longest delays were as sociated with flights between JFK GOT DXB (2 h 10 min) and DXB GOT JFK (2 h 40 min) In both occurrences, the time spent by the instrumented container on the tarmac during the ramp transfer was less than 1 h 50 min. These reasonable times indicate that t he ground crew adjusted their operations to limit the actual ramp transfer times. Also, a B747400F aircraft was used for both flights. This freighter has the largest cargo capacity among the three aircraft used for the tests. The cargo capacity of the B747400F was 117,000 kg compared to 36,500 kg and 17,000 kg for the A310300F and A330200 respectively This indicated that the cargo capacity and therefore the time associated with the loading and unloading operations may have an impact on the flight d elays. Among the four other flights, delays were less than 25 min for three of them, and 1 h 5 min for the other The impacts of the departure delays on arrival times were mitigate d by variables such as shorter taxi or flight times. Delays exceeding 2 h can affect the quality of fresh fruit s or vegetables if they result in longer ramp transfer times

PAGE 165

165 in detrimental ambient conditions The analysis of Figures 5 1 to 518 showed that the temperature distribution of the products within the aircraft container was mostly influenced by the environmental conditions on the tarmac, during ramp transfer not by the conditions onboard the aircraft. To get a better insight on the cold chain logistics associated with the tests, the total transit time from cold room to cold room is broken down into different segments Table 516 summarizes the total duration of each flights and the corresponding fractions spent in flight, on the tarmac and onboard the aircraft T he results of the flights between JFK and DXB ai rport s and those between DXB and NBO airports should be analyzed independently because of their different transit distances. For the flights between JFK and DXB airports more than 60% of the total time was spent in flight and more than 80% was spent onboard the aircraft. Only 14.7% and 16.9% of the total transit time was spent outside on the tarmac. These percentages correspond to average ramp transfers of 1 h 25 min and 1 h 50 min for the JFK GOT DXB and DXB GOT JFK flight respectively Such durations ar e reasonable and did not result in large increases of the product temperatures (Figures 51, 5 3, 5 16 and 518). For the four flights between DXB and NBO airports the percent ages of the total time spent in flight were significantly less. I t ranged between 20.6% and 35.1%. T he sum of the time spent on the tarmac prior departure and after arrival ranged between 53.2% and 67.9% of the total time. For three of the flights the instrumented container was on the tarmac for more than 10 h. Such ramp transf ers in combination with detrimental ambient temperatures inevitably result in important temperature variations

PAGE 166

166 within the load of products ( Figures 5 4, 5 10 and 515). This is true for both low and high ambient temperatures. Solar radiation is also a critical factor during ramp transfers as it can increase the surface temperature of exposed objects several degrees above the ambient air temperature. On the tarmac at DXB airport, the temperature of the top wall of the container reached 62.3C, which was 28.3C above ambient air temperature. The impact of solar radiation can be even larg er for products transported on aircraft pallets since they are often directly exposed. For aircraft containers, the opaque walls as well as the internal air gaps add resistances to heat transfer. Clear wall aircraft containers behave differently when exposed to solar radiation. Sola r radiation is not entirely reflected and absor bed at the surface of the wall and a significant portion is transmi tted through the clear material A portion of the radiation absorbed by the load and reemitted in the infrared portion of the spectrum is tra pped within the container. Even though the resulting greenhouse effect is not as strong as that created by a glas s enclosure, it contributes to the accumulation of energy and raises the te mperatures within the container. 5. 8 Impacts of the Results for Fresh Horticultural Products 5.8.1 Effect of Heat Generation Heat generation can be an important factor for heat tr ansfer analysis in systems composed of fresh fruits and vegetables. All horticultural products generate a certain amount of thermal energy because of their physiological metabolism. This is an important factor that could not be reproduced using an experimental load composed of water bottles. The rate of heat generation is linked to respiration rate of the produce, which is influenced by both internal and environmental factors. Some of the internal

PAGE 167

167 factors that influence the respiration rate are t he type of product, its degree of maturity as well as preharvest factors such as temperature, humidity and water supply E nvironmental factors include temperature, gas concentrations ( oxyge n, carbon dioxide and ethylene) and physical stress (shocks, vibrations, mechanical damage). Depending on these factors as well as on the volume to surface ratio of the products, heat generation can influence the temperature distribution within a single product. T he temperature near the center of a large stack of produce exposed to low air flow and heat transfer such as a fully loaded aircraft container, can also increase significantly because of heat generation. The results of the in flight tests indicate d that the quality of produce shipped in an aircraft container is likely to be affected by heat generation because of the low rate of heat transfer in the core region of the load. Results showed that the temperatures of the products (water bottles ) within layers 2 to 5 had a slow rate of change, hence a low rate of heat transfer. Consequently it can be assumed that for horticultural products with a relatively high rate of heat generati on, the temperature of products in the core layers of the load would have exhibit ed an increase in temperature during transit. It is expected that the temperature of the products located near the center of the load, where the heat transfer is minimum, would have increased even during the period of refrigerated storage. 5.8.2 Effect of Gas Concentrations As discussed in the Section 5.8.1, the results of the inflight tests, showed a low rate of heat transfer within the container, particularly near the core region of the load. This observation also indicates that the gas diffusion would also be small in the core

PAGE 168

168 region particularly for produce with higher bulk density and poor or incompatible (mixed load) packaging systems. The concentration of oxygen and carbon dioxide in the air surrounding fruits and vegetables influence s the ir respiratio n rate and consequently their quality and shelf life. Depending on the type of products (nonclimacteric or climacteric), ethylene may also influence directly or indirectly t he respiration rate. For some produce, respiration rate can be slow ed down and s helf life prolonged by low oxygen concentrations and high carbon dioxide concentrations. Depending on the type of produce these concentration can range between 1 to 5% for oxygen and 2 to 15% for carbon dioxide ( Kader, 2002b). The se gas concentrations al so have other benefits such as the inhibition of decay organisms and the decrease of sensitivity to ethylene. However, lower oxygen concentration and higher c arbon dioxide concentration than recommended can cause irreversible physiological damage, off fla vors and off odors as a result of anaerobic respiration (fermentation). Conditions propitious to anaerobic respiration are likely to be observed within an aircraft container. The results of the in flight tests indicated low rates of heat transfer in the core layers of the load, which could lead to high temperatures and high respiration rates due to heat generation. During the tests, water vapor also accumulated within the container and cause d condensation, which indicates a low gas exchange rate between the container and the ambient air The gas exchange rate would have been even smaller if the container would have been wrapped in a plastic sheet, as it is regularly done for both containers and aircraft pallets Also, the supply of oxygen is limited by the small air volum e in the load and the fact that its concentration is reduced by about 20% due to

PAGE 169

169 the lower atmospheric pressure in flight T hese conditions indicate that fresh fruits and vegetables transported by air are exposed to multiple factors promoting the onset of anaerobic respiration, particularly in the core region of the load. Asparagus, peas, snap bean and cut flowers are among the products prone to generate anaerobic conditions because of their high respiration rate (Kader, 2002b). 5.8 .3 Mechanical Damage After all shipments through NBO airport, the load of RPCs inside the instrumented aircraft container had shifted. This phenomenon was not observed after the flights between JFK and DXB airports. F actor s such as the vibrations and shocks associated with the poor conditions of the road between the different storage facilities and the airport as well as the lack of suspension on the dollies may be identified as the main causes for the load movements. The shifting that was noticed in th e load, did not results in major damage to the RPCs or the load. The only observed damage to the RPCs was that two of them located on the bottom layer had their front side panel dislodged from one of the hinges. T he RPCs were still able to bear the load with the three remaining sides; however, the results may have been different if cardboard boxes would have been used instead or RPCs. T he performance of corrugated cardboard boxes was not directly investiga ted during this series of tests. Nevertheless it is still interesting to discuss some of the potential effects of the environmental condit ions intrinsic to air transport on cardboard boxes. Th e heavy condensation that was noticed inside the instrumented container did not affect the RPCs, but the ex posure of unwaxed cardboard boxes to high level of humidity and condensation results in a significant decrease in their nominal strength.

PAGE 170

170 High humidity, combined to the shocks and vibrations during ground handling operations, leads to mechanical damage to the boxes and consequently to the products inside. The improper stacking of boxes that is widespread in the air transport industry and motivated by the restrictive dimensions of the ULDs also contributes to mechanical damage during handling and transport operations Boxes are generally designed to support, for a given orientation, the weight associated with a maximum number of boxes. Improper stacking orientation or height result in crushed boxes. Spoiled products and claims are not the only negative cons equences to mechanical damage. D amaged load s of fruits and vegetables, perishable products or hazardous materials can potentially leak out of their ULD. These leaks can cause contamination and also accumulate within the aircraft fuselage engendering c orrosion and resulting in major damage to the aircraft. 5.9 Conclusions from Air Transport Tests The results of the air transport tests showed that the ULDs can be exposed to outside ambient conditions for several hours during ramp transfers. It was obser ved that the temperatures of the wal ls of the aircraft container can increase by more than 20C above the ambient air temperature because of the effect of solar radiation during ramp transfer s. Under cloudy conditions or in a shady area, the temperatures of the walls of the container approached that of ambient air. Therefore, it is recommended that airlines use reflective breathable (for gas exchange) ULD covers to protect against solar radiation and that airport authorities set up shaded ar eas on the ramp for temperature sensitive cargo. It was found that the environmental conditions during ramp transfer s had a larger impact than inflight conditions on the temperature distributi on within the load of

PAGE 171

171 products. Even though important temper ature variations are be expected according to the position of the container within aircraft cargo compartments (mond et al. 1 999), such variations were not observed during the experimental shipments. T ests indicated that the core region of the load was exposed to small rates of heat transfer. Consequently, cold room storage of a fully loaded aircraft container proved to be a very inefficient cooling method. For products exposed to detrimental conditions prior to a shipment, it is recommended to break down the load or to circ ulate cold air through the products in order to enhance heat transfer. In addition to detrimental temperatures, the result s indicated that the quality of horticultural products transported by air is likely to be influenced by the eff ects of heat generation, anaerobic respiration, high relative humidity (condensation) and mechanical damage.

PAGE 172

172 Figure 5 1 Average temperatures of the layers ( A L) and temperatures of the inside wall s of the aircraft container (W) during the r amp transfer at JFK airport (04 292007).

PAGE 173

173 Figure 5 2 Pressure (P) as well as average temperatures of the layers (AL) and temperatures of the inside wall s of the aircraft container (W) during the JFK GOT DXB flight (04 29 2007 and 04302007).

PAGE 174

174 Figure 5 3 Average t emperatures of the layers (AL) and temperatures of the inside wall s of the aircraft container (W) during the ramp transfer at DXB airport (04 302007).

PAGE 175

175 Figure 54 Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the ramp transfer at DXB airport (05 032007).

PAGE 176

176 Figure 55 Pressure (P) as well as average temperatures of the layers (AL) and temperatur es of the inside walls of the aircraft container (W) during the flight from DXB to NBO (05 032007).

PAGE 177

177 Figure 5 6 Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container during the ramp transfer at NBO air port (05 032007).

PAGE 178

178 Figure 5 7 Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container during the ramp transfer at NBO airport (05 042007).

PAGE 179

179 Figure 58 Pressure (P) as well as average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container during the flight from NBO to DXB (05 04 2007 to 05052007).

PAGE 180

180 Figure 5 9 Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container during the ramp transfer at DXB airport (05 052007).

PAGE 181

181 Figure 510. Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the ramp transfer at DXB airport (05 062007).

PAGE 182

182 Figure 5 11. Pressure (P) as well as average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the flight from DXB to NBO (05 062007).

PAGE 183

183 Figure 5 12. Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the ramp transfer at NBO airport (05 0 6 2007).

PAGE 184

184 Figure 513. Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the ramp transfer at NBO airport (05 072007).

PAGE 185

185 Figure 514. Pressure (P) as well as average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the flight from NBO to DXB (05 072007).

PAGE 186

186 Figure 5 15. Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the ramp transfer at DXB airport (05 062007).

PAGE 187

187 Figure 5 16. Average temperatures of the layers (AL) and temperatures of the inside wall s of the aircraft container (W) during the ramp transfer at DXB airport (05 122007).

PAGE 188

188 Figure 5 17. Pressure (P) as well as average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the DXB GOT JFK flight (05 122007 and 05132007).

PAGE 189

189 Figure 5 18. Average temperatures of the layers (AL) and temperatures of the inside walls of the aircraft container (W) during the ramp transfer at JFK airport (05 132007).

PAGE 190

190 Table 51. Initial (10:10, t = 0 h) and final (12:00, t = 1.83 h) temperatures of the inside walls of the instrumented container while outside on the tarmac at JFK airport (04 292007). Wall z Initial temperature ( C) Final temperature ( C) T final T initial ( C) W1 6.8 15.3 8.5 W2 6.7 15.9 9.2 W3 6.6 10.6 4.0 W4 5.7 12.6 6.9 W5 6.6 16.2 9.6 W6 6.8 15.6 8.8 z: Refer to Figure 435 Table 52 Initial (2:10, t = 10 h) and final (4:30, t = 12.33 h) temperatures of the inside walls of the instrumented container while in the plane during ramp transfer at GOT airport (0430 2007). Wall z Initial temperature ( C) Final temperature ( C) T final T initial ( C) W1 8.8 11.1 2.3 W2 9.8 11.2 1.4 W3 6.8 7.6 0.8 W4 7.9 9.9 2.0 W5 9.0 11.5 2.5 W6 9.2 10.9 1.7 z: Refer to Figure 435 Table 53. Temperatures of the inside walls of the instrumented container when unloaded of the aircraft (12:30, t = 18.33 h) and after a 30 min exposure to solar radiation on the tarmac at DXB airport (04302007). Wall z Initial temperature ( C) Final temperature ( C) T final T initial ( C) W1 11.3 62.3 51.0 W2 11.1 37.1 26.0 W3 7.1 15.3 8.2 W4 10.9 32.4 21.5 W5 11.4 35.9 24.5 W6 11.0 34.5 23.5 z: Refer to Figure 435

PAGE 191

191 Table 54. Initial (1:40, t = 0 h) and final (9:30, t = 7.83 h) temperatures as well as temperatures at sunrise (5:41, t = 4.02 h) of the inside walls of the instrumented container while outside on the tarmac at DXB airport (05 032007). Wall z Initial temperature ( C) Temperature at sunrise ( C) Final temperature ( C) W1 6.7 23.4 51.9 W2 6.4 20.4 28.5 W3 5.8 11.5 14.3 W4 5.5 18.6 36.7 W5 6.4 20.0 30.5 W6 6.4 19.8 35.6 z: Refer to Figure 4 35 Table 55. Initial (1:40, t = 0 h) and final (9:30, t = 7.83 h) average temperatures of the layers of products in the instrumented container while outside on the tarmac at DXB airport (05032007). Layer x Initial temperature ( C ) n y s z Final temperature ( C ) n s T final T initial ( C ) AL1 6.3 5 0.2 8.6 5 0.2 2.3 AL2 6.3 5 0.2 6.6 5 0.2 0.3 AL3 6.2 5 0.2 6.8 5 0.4 0.6 AL4 6.4 8 0.2 7.4 8 0.5 1.0 AL5 6.5 8 0. 3 8.6 8 0.5 2.1 AL6 6.6 7 0.2 11.1 7 0.8 4.5 AL7 6.6 8 0.3 16.1 8 0.5 9.5 x : Layer 1 and 7 correspond to the bo ttom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation.

PAGE 192

192 Table 56. Initial (10:30, t = 7.83 h) and final (14:30, t = 13.83 h) average temperatures of the layers of products in the instrumented container between the time when the products were loaded onboard the plane at DXB airport and unloaded on the tarmac at NBO airport (0503 2007). Layer x Initial temperature ( C) n y s z Final temperature ( C) n s T final T initial ( C) AL1 8.6 5 0.2 9.6 5 0.2 1.0 AL2 6.6 5 0.2 7.3 5 0.4 0.7 AL3 6.8 5 0.4 8.4 5 0.6 1.6 AL4 7.4 8 0.5 9.6 8 0.5 2.2 AL5 8.6 8 0.5 11.3 8 0.9 2.7 AL6 11.1 7 0.8 13.4 7 0.5 2.3 AL7 16.1 8 0.5 18.8 8 0.7 2.7 x : Layer 1 and 7 correspond to the bo ttom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation. Table 57. Initial and final average temperatures of the layers of products in the instrumented container during the 25 h overnight cold room storage at NBO airport (05 032007 to 0503 2007). Layer x Initial temperature ( C) n y s z Final temperature ( C) n s T final T initial ( C) AL1 10.4 5 0.2 8.7 5 0.6 1.7 AL2 8.1 5 0.2 8.6 5 0.4 0.5 AL3 9.3 5 0.5 8.4 5 0.4 0.9 AL4 11.0 8 0.4 8.5 8 0.4 2.5 AL5 12.4 8 0.9 9.1 8 0.3 3.3 AL6 14.1 7 0.5 10.1 7 0.4 4.0 AL7 20.1 8 0.7 11.0 8 0.7 9.1 x : Layer 1 and 7 correspond to the bo ttom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation.

PAGE 193

193 Table 58. Initial (5:00, t = 8.50 h) and final (9:20, t = 12.83 h) average temperatures of the layers of products in the instrumented container while outside on the tarmac at DXB airport (05042007). Layer x Initial temperature ( C) n y s z Final temperature ( C) n s T final T initial ( C) AL1 9.4 5 0.4 10.4 5 0.4 1.0 AL2 8.6 5 0.4 8.9 5 0.2 0.3 AL3 8.6 5 0.2 9.2 5 0.2 0.6 AL4 9.0 8 0.4 9.9 8 0.3 0.9 AL5 9.8 8 0.3 10.9 8 0.4 1.1 AL6 10.9 7 0.2 12.6 7 0.3 1.7 AL7 12.2 8 0.3 15.8 8 0.4 3.6 x : Layer 1 and 7 correspond to the bo ttom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation. Table 59. Initial (1:40, t = 0 h) and final (9:40, t = 8 h) temperatures as well as temperatures at sunrise (5:39, t = 3.98 h) of the inside walls of the instrum ented container while outside on the tarmac at DXB airport (05062007). Wall z Initial temperature ( C) Temperature at sunrise ( C) Final temperature ( C) W1 9.6 23.6 51.4 W2 8.4 21.9 33.3 W3 W4 7.6 20.2 31.7 W5 8.7 20.2 31.1 W6 8.5 19.5 33.6 z: Refer to Figure 435

PAGE 194

194 Table 510. Initial (1:40, t = 0 h) and final (9:40, t = 8 h) average temperatures of the layers of products in the instrumented container while outside on the tarmac at DXB airport (05062007). Layer x Initial temperature ( C) n y s z Final temperature ( C) n s T final T initial ( C) AL1 9.7 5 0.2 11.6 5 0.2 1.9 AL2 9.3 5 0.2 9.4 5 0.2 0.1 AL3 9.1 5 0.2 9.9 5 0.4 0.8 AL4 9.4 8 0.3 10.8 8 0.4 1.4 AL5 10 8 0.3 11.9 8 0.5 1.9 AL6 10.6 7 0.3 14.1 7 0.5 3.5 AL7 11.4 8 0.4 18.1 8 0.6 6.7 x : Layer 1 and 7 correspond to the bo ttom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation. Table 511. Initial (9:40, t = 8 h) and final (15:40, t = 15 h) average temperatures of the layers of products in the instrumented container between the time when the products were loaded onboard the plane at DXB airport and unloaded on the tarmac at NBO airport (0506 2007). Layer x Initial temperature ( C) n y s z Final temperature ( C) n s T final T initial ( C) AL1 11.6 5 0.2 12.6 5 0.2 1.0 AL2 9.4 5 0.2 10.6 5 0.2 1.2 AL3 9.9 5 0.4 11.4 5 0.4 1.5 AL4 10.8 8 0.4 12.6 8 0.5 1.8 AL5 11.9 8 0.5 14.1 8 0.7 2.2 AL6 14.1 7 0.5 15.1 7 0.6 1.0 AL7 18.1 8 0.6 20.9 8 0.4 2.8 x : Layer 1 and 7 correspond to the bottom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation.

PAGE 195

195 Table 512. Initial and final average temperatures of the layers of products in the instrumented container during the 17 h overnight cold room storage at NBO airport (05062007 to 0507 2007). Layer x Initial temperature ( C) n y s z Final temperature ( C) n s T final T initial ( C) AL1 13.5 5 0.0 9.8 5 0.5 3.7 AL2 12.0 5 0.3 10.6 5 0.4 1.4 AL3 12.9 5 0.4 10.8 5 0.2 2.1 AL4 14.0 8 0.4 11.0 8 0.7 3.0 AL5 14.7 8 0.6 11.6 8 0.8 3.1 AL6 16.4 7 0.5 13.5 7 0.8 2.9 AL7 21.1 8 0.3 14.3 8 0.9 6.8 x : Layer 1 and 7 correspond to the bo ttom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation. Table 5 13. Initial (23:10, t = 6.67 h) and final (10:00w, t = 17.5 h) average temperatures of the layers of products in the instrumented container while outside on the tarmac at DXB airport (05062007 and 0507 2007). Layer x Initial temperature ( C) n y s z Final temperature ( C) n s T final T initial ( C) AL1 10.5 5 0.4 13.4 5 0.2 2.9 AL2 10.2 5 0.4 12.0 5 0.4 1.8 AL3 10.6 5 0.2 12.7 5 0.4 2.1 AL4 10.9 8 0.6 13.8 8 0.4 2.9 AL5 11.6 8 0.6 14.9 8 0.8 3.3 AL6 13.3 7 0.6 16.9 7 0.4 3.6 AL7 15.1 8 0.6 19.3 8 0.4 4.2 w : Next day, 05 072007 ; x : Layer 1 and 7 correspond to the bottom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation.

PAGE 196

196 Table 514. Initial (00:00, t = 1.33 h) and final (10:00w, t = 19.33 h) average temperatures of the layers of products in the instrumented container while onboard the aircraft for the DXB GOT JFK flight (05 122007 and 05 132007). Laye r x Initial temperature ( C) n y s z Final temperature ( C) n s T final T initial ( C) AL1 7.1 5 0.2 9.8 5 0.3 2.7 AL2 7.6 5 0.2 8.1 5 0.2 0.5 AL3 7.6 5 0.2 8.2 5 0.2 0.6 AL4 7.6 8 0.3 8.8 8 0.3 1.2 AL5 7.9 8 0.4 9.9 8 0.4 2.0 AL6 8.4 7 0.3 11.8 7 0.5 3.4 AL7 9.7 8 0.2 14.6 8 0.2 4.9 w : Next day, 05 072007 ; x : Layer 1 and 7 correspond to the bottom and top layer respectively ; y: n is the number of measurements; z: s is the standard deviation. Table 515. Scheduled and actual departure and arrival times as well as corresponding delays for all flights. Flights Departure date Departure Arrival Scheduled time Actual time z Delay (hh:mm) Scheduled time z Actual time z Delay (hh:mm) JFK GOT DXB 04 29 2007 11:00 13:10 2:10 11:00 12:10 1:10 DXB NBO 05 03 2007 10:05 10:30 0:25 14:15 14:00 NBO DXB 05 04 2007 23:00 23:00 5:00 4:30 DXB NBO 05 06 2007 10:05 11:10 1:05 14:15 14:40 0:25 NBO DXB 05 07 2007 17:15 17:20 0:05 23:15 22:40 DXB GOT JFK 05 12 2007 23:00 1:40 2:40 8:00 10:00 2:00 z: Italic times indicates the event took place t he day after the departure date.

PAGE 197

197 Table 516. Total transit time of all flights as well as the corresponding time and percentage of the total time spent by the instrumented container in flight, on the tarmac and onboard the aircraft. Flights Departure date Total timey (hh:mm) Flight Tarmac Onbo ard the aircraft x Time (hh:mm) % z Time (hh:mm) % z Time (hh:mm) % z JFK GOT DXB 04 29 2007 19:20 12:40 65.5 2:50 14.7 16:30 85.3 DXB NBO 05 03 2007 17:50 4:30 25.2 11:50 66.4 6:00 33.6 NBO DXB 05 04 2007 12:50 4:30 35.1 6:50 53.2 6:00 46.8 DXB NBO 05 06 2007 21:50 4:30 20.6 14:50 67.9 7:00 32.1 NBO DXB 05 07 2007 17:30 4:20 24.8 11:40 66.7 5:50 33.3 DXB GOT JFK 05 12 2007 21:40 13:30 62.3 3:40 16.9 18:00 83.1 x: This period include the time the instrumented container was onboard the plane on the ground as well as in flight; y: Total time from the cold room at the origin airport to the cold room at destination airport; z: Perc entage of the total time

PAGE 198

198 CHAPTER 6 EVALUATION OF HEAT TRANSFER MODELS BASED ON AN EFFECTIVE THERMAL CONDUCTIVITY For practical applications in the transport industry the approach use d to predict the temperature within a load of horticultural products must be simple and efficient Real time monitoring could also be possible using r adio frequency identification technology (RFID). The selected approach should be based on the thermal properties of the load, its initial temperature as well as variations with time of the inside temperatures of the walls of the box or container (boundary conditions). To achieve this, an effective property model is the best suited. H eat transfer model s based on an effective thermal conductivi ty are solved using a simple approach based on a pure diffusion over the entire domain. In addition to the flu i d and solid thermal conductivities these models may also include the effects of factors such as the contact surface area between solid s, fluid flow as well as radiative heat transfer Effective thermal conductivities have been used for several years to describe the heat transfer in packed beds (Kunii and Smith, 1960; Bauer and Schlnder, 1978a; 1978b; Tsotsas and Martin, 1987; Hsu et al., 1994; Hsu et al 1995; Gupta et al., 2002). For air transport applications, Villeneuve et al. (2001) and Oskam et al. (1998) used p ure diffusion models to describe the heat transfer in loads of simulated or real horticultural products. H owever, the effective thermal conductivity used was not based on an analysis as complete as th ose presented for packedbed applications and both models underestimated the heat transfer in the core of the load. The objective of this chapter is to evaluate the validity of a heat diffusion model based on an effective thermal conductivity to model the heat transfer within loads of

PAGE 199

199 horticultural products. Comparisons between simulated results and those obtained in Chapters 3, 4 and 5 were analyzed. 6.1 Model and General Assumptions Zehner's static effective thermal conductivity model was selected to sim ulate the experimental results ( Equations 25 to 2 8, Bauer and Schlnder, 1978b; Tsotsas and Martin, 1987) Zehner's model was preferred to Kunii and Smith's (1960) because it included the heat transfer through the finite contact area between particles which is an important factor for several horticultural products. The model by Hsu et al. (1994) was also considered, but its added complexity appeared to provide advantages over Zehner's model only for values of ks/kf > 103, which was far higher than the present range of interested. The decision to use a static model for the effective thermal conductivity which accounts for the conduction in the fluid phase but neglects any convective heat transfer, was based on the results obtained by two groups of authors. The first group was Tanner (1998) and Tanner et al. (2002a; 2002b; 2002c) that assumed pur e conduction in the fluid phase and obtained good results with their multi zone model. The second group was Bazan (19 8 9 ) and Bazan et al. (1 9 89). The authors found that natural convection in a closed box caused the location of the maximum temperature to move upwards but also concluded that low convective heat transfer coefficient existe d at the surface of the fruit. In addition, Bazan (1989) fo und a maximum temperature difference of 2C between bott om and top corner fruits. The author explained that this result could no t be observed in pure conduction models with uniform boundary conditions (symmetry wi th respect to each direction). However in this study the boundary conditions used for the static effective thermal conductivity model vary

PAGE 200

200 according to the walls of the box or container and therefore accounts for some of the effect s of natural convection within the enclosure In addition to considering only conduction heat transfer in the fluid phase, the following general assumptions were made for the model (the term "products" is used for both fruit and water bottles): 1) H eat generation from the fruit is neglected. 2) Thermal and physical p roperties of the air and the products are assumed constant and uniform. 3) The relative humidity inside the box of fruit and the aircraft container is near saturation, therefore the evaporative cooling due to transpiration on the surface of the fruit is as sumed to be negligible. 4) The contact resistances between products are negligible. 5) There are small temperature differences between the surfaces of adjacent products and therefore radiation heat transfer between products is negligible. 6) Radiative heat transfer between the products and the inside s urfaces of the walls of a box or reusable plastic container (RPC) is negligible. 7) Differences between pulp and air temperatures within the load are small. From these assumptions, Zehner's model can be simplified to Equations 61 to 6 3. The deformation factor B can be approximated by Equation 64 or calculated from E quation 65. = 1 1 + 1 + ( 1 ) (6 1 ) = 2 1 ln 1 ( 1 ) ( + 1 ) 2 (6 2) = 1 (6 3)

PAGE 201

201 = 1 364 1 (6 4) = 1 ( 3 4 + + 2 ln ) ( 1 ) (6 5) The effective thermal conductivity obtain ed by Equations 61 to 65 was used in combination with the transient threedimensional heat conduction equation that was simplified according to the discussed assumptions = + + (6 6) For the simulations corresponding to the series of test s with a single box of fruit, the effective density (e) and the effective specific heat (CPe) were calculated using a bulk porosity weighted average (Equations 67 and 68). = + ( 1 ) (6 7) = + ( 1 ) (6 8) For the simulations conducted on loads inside aircraft containers, the contribution of the corrugated boxes or the RPCs was included in the calculations for the effective density and specific heat ( Equations 69 and 6 10) = + ( 1 ) + + (6 9 ) = + ( 1 ) + + (6 10) 6.2 Sensitivity Analysis on the Effective Thermal Conductivity Model A sensitivity analysis was conducted on Zehner's model with respect to the three main parameters (factors): the thermal conductivity of the solid (ks), the bulk porosity () and the contact surface ratio (). For the analysis, each parameter was varied on three

PAGE 202

202 levels: minimum, average and maximum. The thermal conductivity of the solid phase (ks) varied between 0.2 and 0.6 Wm1K1; this range includes the thermal conductivities of a ll fresh fruits and vegetables tabulated in ASHRAE (2006) and engle (2010) The bulk porosity () was varied between 0.26 and 0.476, which correspond to cubic or hexagonal close packing and simple cubic packing arrangements respectively These arrangeme nts have to the minimum and maximum bulk porosities for packing with contact between spherical particles. The contact surface () was varied between 0 and 0.6, which correspond to the spherical particles with no deformation (contact point) and particles with important deformations (contact surface) Table 6 1 and Figure 6 1 present the details of the sensitivity analysis. F rom Table 6 1, columns 1, 2 and 6 respectively indicate that the effective thermal conductivity increased as the surface ratio increased, decreased as the porosity increased and increased as the thermal conductivity of the solid increased. These observations were all expected for a physically sound model. F or the parameter ranges used in the analysis, columns 4 and 5 show that the in fluence of the contact surface ratio was more important than the influence of the porosity on the effective thermal conductivity. Columns 8 and 9 show that the combined influence of the porosity and contact surface ratio was also more important than the i nfluence of the thermal conductivity of the solid on the effective thermal conductivity. The presence of i nteractions between the three paramet ers is obvious according to Equations 6 1 to 63 ; columns 3 and 7 in Table 6 1 also show the interactions and how the y increase with the thermal conductivity of the solid. Even though Table 6 1 provides detail ed information on the sensitivity analysis, Figure 6 1 simplifies the presentation of the results.

PAGE 203

203 The linear dependence of the effective thermal conductivity on the contact surfac e ratio is obvious on Figure 6 1 In addition, t he interactions between parameters are clearly illustrated On each of the graphs, t he spread between the three curves illustrates the effect of the porosity, which increases with the thermal conductivity of the solid Also, t he higher slopes of the curves as the thermal cond uctivity of the solid increases illustrate its influence on the effect of the contact surface ratio The small variations between the slopes of the curves, for a given thermal conductivity of the solid, show that the interaction between the porosity and the contact surface ratio was minimal. 6.3 Simulation Parameters and Boundary Conditions A simulation was completed for each of the singlebox test series as well as for the laboratory aircraft container tests In the case of the air transport tests, the extended ramp transfer at Dubai international airport (DXB) on 050 3 2007 was simulated. The five simulations were run using effective thermal conductivities calculated from Zehner's model (Equations 61 to 65) in combination with the transient threedimensional heat conduction equation (Equation 66). The simulations were performed using Comsol Multiphysics (version 3.5a) with the GMRES solver (linear system solv er) and a geometric multigrid preconditioner. The number of nodes for each of the three series of single box tests as well as the laboratory and air transport aircraft container tests was 53,880, 86,296, 87,977, 489,182 and 406,207 respectively For all simulations, the time interval was 300 s. The parameters and the resulting effective thermal conductivities used for the simulations are presented in Table 62. The contact surface ratios () which is the ratio

PAGE 204

204 of the radius of the contact surface between products to the radius of the product were obtained using the computer aided design (CAD) tool of the simulation software. For each series of tests, the packing arrangement was reproduced in the CAD environment using an averagesized product. Th en, a single product was isolated from the arrangement to expose its contact surfaces (Figure 62 ). The c ontact surface ratio () was calculated from the average radius of the contact surfaces The effective densities and specific heat were calculated based on Equations 67 to 6 10. The initial temperatures (T0) corresponded to the initial average temperature of the products in the box or the aircraft container. The parameters as well as the thermal and physical properties used for the simulations are s ummarized in Table 6 3. The boundary conditions were defined differently for singlebox and aircraft container simulations. However, in all cases the boundary conditions were based on the temperatures measured on the inside surfaces of the wall s of the b ox or aircraft container To implement these transient temperatures in the simulation software, the data sets were first thoroughly approximated with exponential or piecewise linear regressions (functions of time) using SigmaPlot (version 11.0) For the t hree series of singlebox tests, t he inside wall temperatures were measured at the actual boundaries of the porous domain. Since t he results of the test s showed that the temperatures within the boxes were stratified vertically (z direction) the temperatures at the top and bottom walls were assumed to be uniform and equal to the corresponding transient temperature measured at their center. For the lateral walls, the temperatures were assumed to be a function of time as well as the z direction; to account for the vertical stratification. For each lateral wall two functions (in z) were

PAGE 205

205 defined: one for the bottom half and one for the top half. It was assumed that the temperature varied linearly between the bottom and the center as well as between the center and the top of the lateral walls. This assumption was supported by the results presented in Chapter 4; for several tests the lateral wall temperatures were approximately equal to the average of the bottom and top wall temperatures. Even thoug h natural convection was neglected in the governing equations, the approach used to define the boundary conditions integrated in the simulation some of the effect s of the natural convection that were revealed through the analysis of the experimental data. During the aircraft container tests, a vertical stratification of the temperatures was also observed. Therefore, for the temperatures of bottom and top boundaries of the load were again assumed to be uniform and equal to the transient temperature measured at the center of the corresponding inside wall surface. For t he lateral walls, the temperatures were assumed to be a function of time as well as the z direction. They were defined by two linear functions, as it was the case for the singlebox simulation s. However, the temperatures of the lateral walls of the aircraft container could not be used as the boundary conditions. In this case, natural convection in the air gaps that separated the walls of the container and the load had to be considered. The convective heat transfer coefficients associated with natural convection on a vertical wall were calculated using Equations 6 11 to 613, where the characteristic length (Lc) was the height of the load (engel and Ghajar 2 010). = (6 11)

PAGE 206

206 = 0 825 + 0 387 1 + 0 492 (6 12) = (6 13) Since the height of the side of the load was shorter on the left side of the container (Figure 4 35), a different convective heat transfer coefficient was obtained for that boundary. For the simulations, it was assumed that the air temperatures far from the surfaces of the load were equal to the temperatures of the lateral walls of the aircraft container, which were defined by t w o linear functions as previously described. Since the boundary conditions varied with time, air properties at an average temperature as well as an average temperature di fference were used to calculate the Rayleigh number s and the convective heat transfer coefficient s. In circumstances where natural convection is an important heat transfer mechanism, radiative heat transfer also need s to be considered. For the simulations, the radiative effect on the boundaries of the load was i n cluded in the form of a radiative heat transfer coefficient (Equation 6 14) which was added to the convective heat transfer coefficient (Equation 6 15). = ( + ) ( + ) ( ) ( ) ( 6 14 ) = + (6 15 ) It was possible to use a combined heat transfer coefficient in this case since the air temperature was assumed to be equal to the temperature of the surroundings (inside

PAGE 207

207 surfaces of the aircraft container). Table 6 4 presents the convective and radiative heat transfer coefficient use d for the aircraft container simulations. For the aircraft container tests, the properties of expanded polystyrene were used for the insulated wedge. The density, thermal conductivity and specific heat of expanded polystyrene are 16 kg m3, 0.04 Wm1K1 a nd 1200 Jkg1K1, respectively (engel and Ghajar, 2010) 6.4 Simulation Results 6.4.1 Single Box Tests For each of the three series of tests, the results of the simulations were compared to the experimental temperatures for three fruit on each of the layers within the boxes. The locations of the fruit varied slightly depending on the series, but in all three cases a corner, intermediary and center fruit were selected on each l ayer. Figures 34, 3 6 and 38 show the locations of the fruit for the simulations of series 1 to 3, respectively 6.4 .1 .1 First series of tests Figure 63 presents the results for the first layer of fruit in the box. The model provided good results for t he intermediary (F7) and center fruit (F13) ; the simulated results were within 1.9 C and 1.2C of the experimental data, respectively For the corner fruit, the model underestimated the temperature. The difference between the experimental and simulated data increased with time. A maximum temperature difference of 3.7C was reached over the 7h test period. For the second layer, Figure 6 4 shows that the model predicted the temperature of the corner fruit (F30) within 1.2C of the experimental data. R esults were not as good for the intermediary and center fruit for which the simulated temperatures were colder than the experimental data by a maximum of 3.7C and 4.6C respectively

PAGE 208

208 For third and fourth layers (Figures 6 5 and 66), the correlations bet ween the simulations and the experimental data were poor, particularly for the intermediary and center fruit. For the fourth layer, the simulation underpredicted the temperatures by up to 15.8 C for the center fruit (F88) For those two layers, the temperatures of the corner fruit (F51 and F80) were predicted within 2.5 C and 3.4C respectively which are still acceptable deviations. Figure 6 7 presents the result of the fifth layer. The temperatures of the corner and center fruit (F101 and F113) w ere significantly over and underpredicted, respectively As mentioned in Section 4.1.1.1, the temperature of fruit F113 was higher than any other fruit within that layer; therefore, an experimental error may contribute to that important difference between the measured and simulated temperatures. For the intermediary fruit (F107), the simulated temperatures were within 3.7C of the experimented data. These results indicate the effective thermal conductivity modeling approach failed to adequately predict the temperatures within the box of fruit. T emperatures were significantly underpredicted particularly in the core of the load (third and fourth layers). 6 .4 .1.2 Second series of tests Figures 68 to 610 present the simulation results for the three layer s of fruit. In this case, t he model predicted the experimental data with more accuracy than for the first series of tests. For the first layer of fruit, the corner, intermediary and center fruit were predicted within 2.8C, 1.5 C and 1.8C over the 15h test period. For the second layer of fruit, the model predicted within 2.1C the temperature of the corner fruit (F25). However, the simulated temperatures were lower than the

PAGE 209

209 experimental measurements for the intermediary and center fruit. The maximum temperature differences were 4.2C and 6.3 C respectively On the third layer, the simulation overpredicted the corner fruit (F29) temperature by an average of 2.1 C. The maximum difference between the simulated and experimental results was calculated at 4.5 C. For the intermediary and center fruit, the simulated temperatures were more accurate; they were, respectively within 1.9C and 3.1 C of the experimental data. Again, the effective thermal conductivity model failed to accurately predict the temperatures near the center of the box. However, the model provided good results for most fruit adjacent to a wall of the box. 6 .4 .1.3 Third series of tests Figure 611 shows that the simulated temperatures corroborated the experimental results for the f ruit located within the first layer. The maximum temperature differences for the corner, intermediary and center fruit were 2.2C, 1.1 C and 1.3 C respectively The results of the simulation for the second, third and fourth layers are presented in Figure s 6 12 to 614. Again, poor agreement between simulated and experimental temperatures was observed. Nevertheless, the temperatures of the corner fruit located within the second and third layer were predicted within 3.1C and 2.8C respectively As for the two previous series of tests, the diffusion based model underpredicted the heat transfer in the core region of the box. 6 .4 .2 Aircraft Container Tests 6.4.2.1 Laboratory t ests Figures 34 and 310 show the locations of the fruit and the boxes for the laboratory aircraft container simulation. As presented in Chapter 4, the temperature of

PAGE 210

210 the fruit located near the center (F63) of the boxes did not vary significantly during the experiment except for the top layer of the aircraft container (sixth layer ). Consequently, Figure 615 presents the results of the simulation for the average of the center temperatures of the 16 boxes located on the sixth layer of the aircraft container (AL6). T he simulation model failed to correctly predict the temperature in the upper region of the load. Figure s 6 16 to 618 present the temperature s of the four fruit located near the outer surface of the load for the first, third and sixth layers respectively The correlations between simulated and experimental temperatures were particularly g ood for boxes B11, B39 and B87, for which the model predictions were within 1.8C, 2.6 C and 0.7C of the measured temperatures. All three boxes were all located near the front wall of the aircraft container (W5). Simulated results were not as good for the other peripheral fruit. For the first layer, the maximum difference between simulated and experimental temperatures were 4.5 C, 2.6 C and 4.2C for the fruit in boxes B2, B4 and B6, respectively For the third layer, the maximum tem perature differences were 3.9C, 6.5 C and 4.3C for the fruit in boxes B29, B32 and B39 respectively Results were slightly better for the sixth layer where the temperature differences were 2.2C, 3.8 C and 4.1C for the fruit in boxes B75, B77 and B80, respectively Figures 619 and 620 presents the simulation results for boxes B1 and B88, where the temperatures were measured in three fruit located on the threedimensional diagonal of the box. For box B1, the simulation results were excellent f or the fruit F21 and F63 located at t he bottom and center of the box; temperatures were predicted

PAGE 211

211 between 1.9C and 0.6 C respectively The simulation underpredicted the temperature near the top of the box (fruit F105). Results were different for box B88 wher e the temperature of fruit F105 located with in the top layer of the box was predicted within 2.5C; however, the model again underpredicted the heat transfer away from the boundary Th e bottom and center f ruit temperatures (F21 and F63) were lower than the measured temperatures by an average of 3.5C and 4.3C respectively 6.4.2.2 Air t ransport t ests To evaluate the simulation model with the experimental data collected during air shipment s, the extended ramp transfer at DXB airport on 05032010 was selected. Figures 316 and 317 show the locations of the water bottles and the RPCs for the air transport simulation. Figure 621 presents the simulated average temperatures for first, fourth and seventh layers of RPCs within the aircraft container. Ex perimental data showed the average temperatures of the first and fourth layers (AL1 and AL4) varied by only 2.3C and 1.1 C respectively For the same layers, t he model predicted changes in the average temperatures of 1.9 C and 0.2C. For the seventh layer, the average simulated temperature was lower than the experimental data, but the difference between the two remained within 3.6C. In addition to the average layer temperatures, the simulated temperatures near the center of two RPCs located on the same layers ( L1, L4 and L7) are shown i n Figures 6 22 to 624. The RPCs in Figures 6 22 to 624 were selected to present a corner as well as a core layer temperature. The results of the simulation were good for the first and fourth layers where very little temperature variations were observed. As it was the case for the laboratory tests, the simulated temperatures were below the

PAGE 212

212 experimental data for the RPCs located within the top layer of the container. The maximum differences between simulated and experimental temperatures were 3.2C and 4.0C for RPC49 and RPC50, respectively 6.5 General Discussion on the Simulation Results It is difficult to set a general criterion of acceptability for model s since it can vary greatly depending on their application. In the case of air transport of horticultural products, a model that could predict within 5 C the temperatures within a load would be a useful tool. The model presented in this chapter failed to meet this criterion. In general, the model underestimated the rate of heat transfer in the core of the load. The temperatures in the top layer of products were also underpredicted, even though the boundary condition for the top surface (temperature equal to the top wall temperature) conveyed t he assumption of a large heat transfer. The use of a physically justifiable and higher effective thermal conductivity for the simulation is also insufficient to account for the rate of change of the temperatures in the core region of the load. In additi on, a higher effective thermal conductivity would overestimate the temperatures of the products located in the peripheral layers of the load. The locations of the thermocouples in the load are also a source of error for the simulations. Creep compression of the fruit within the boxes as well as the boxes themselves, during the laboratory aircraft container tests, may have affected the initial locations used for the simulations. This may in part explai n the better simulation results obtained for the botto m layers of the load, since the incertitude of the thermocouple locations was less than for the upper layers. However, because of the small

PAGE 213

213 temperature variations measured within the fruit, these errors cannot explain the large discrepancies observed betw een the simulations and the experimental data. According to simulation results it seems that the effects of natural convection on the temperature distribution were not negligible. For singlebox tests the presence of vents on the top and bottom wall appeared to have been an important factor, facilitating the air flow and increasing the effect of natural convection. For aircraft container tests, the influence of the vents on natural convection may explain the better simulation results obtained for the boxes or RPCs located at the bottom of the load compared to those located at the top. The effect of natural convection could be included in the model by adding a dynamic component to the static effective thermal co nductivity (Equations 2 1 and 22). The dynam ic component would have to be obtained from an analysis of the buoyancy driven flow in the load and would likely be a function of position as well as time. This will results in a significantly more complex approach. 6. 6 Conclusions on the Effective Thermal Conductivity Modeling Approach A heat transfer model based on a static effective thermal conductivity did not adequately predict the temperatures in all regions within l o ads of horticultural products. For the product size s and packing arrangements studied, in order to improve the temperatur e prediction in the core region, the effects of natural convection within the loads must be included by adding a dynamic component to the effective thermal conductivity. However, this additional component would result in a highly complex analysis and would therefore lack the simplicity needed for most commercial applications.

PAGE 214

214 While the modeling approach presented had some flaws, it did predict with good agreement the temperatures of the peripheral products. As discussed in Chapter 4, peripheral products accounted for approximately 50% of all product s within the aircraft container and are the most susceptible to temperature abuse. The ability to predict the temperature of the peripheral products is an important contribution because these products are usually the basis for the rejection of a load. From that perspective, the modeling approach used, even without a dynamic component, can still be a useful tool for air shipments of horticultural products. In addition, t he effective thermal conductivity approach presented in this study may provide better results for smaller size products such as berries. First, the heat transfer through the smaller air gaps associated with the packing arrangement of the fruit may be more effectively described by a diffusion model. Secondly, the smaller air gaps may also lower the effects of the natural convection.

PAGE 215

215 Figure 61 Results of the sensitivity analysis on Zehner's model for the effective thermal conductivity.

PAGE 216

216 Figure 62. Contact surfaces of a product obtained using the CAD tool of the simulation software.

PAGE 217

217 Figure 63. Simulation s of the pulp temperatures of fruit F1, F7 and F13 located on the first layer of the box for the first series of tests.

PAGE 218

218 Figure 64 Simulation s of the pulp temperatures of fruit F30, F34 and F38 located on the second layer of the box for the first series of tests.

PAGE 219

219 Figure 65 Simulation s of the pulp temperatures of fruit F51, F57 and F63 located on the third layer of the box for the first series of tests.

PAGE 220

220 Figure 66. Simulation s of the pulp temperatures of fruit F80, F84 and F88 located on the fourth layer of the box for the first series of tests.

PAGE 221

221 Figure 67 Simulation s of the pulp temperatures of fruit F101, F107 and F113 located on the fifth layer of the box for the first series of tests.

PAGE 222

222 Figure 68 Simulation s of the pulp temperatures of fruit F1, F5 and F6 located on the first layer of the box for the second series of tests.

PAGE 223

223 Figure 69 Simulation s of the pulp temperatures of fruit F25, F22 and F23 located on the second layer of the box for the second series of tests.

PAGE 224

224 Figure 61 0 Simulation s of the pulp temperatures of fruit F29, F33 and F34 located on the third layer of the box for the second series of tests.

PAGE 225

225 Figure 611. Simulation s of the pulp temperatures of fruit F1, F7 and F12 located on the first layer of the box for the thir d series of tests.

PAGE 226

226 Figure 61 2 Simulation s of the pulp temperatures of fruit F32, F29 and F25 located on the second layer of the box for the third series of tests.

PAGE 227

227 Figure 61 3. Simulations of the pulp temperatures of fruit F37, F43 and F48 located on the third layer of the box for the third series of tests.

PAGE 228

228 Figure 61 4. Simulations of the pulp temperatures of fruit F68, F65 and F61 located on the fourth layer of the box for the third series of tests.

PAGE 229

229 Figure 61 5 Simulation of t he average pulp temperature of the top layer of boxes (L6) in the aircraft container Figure 616. Simulations of the pulp temperatures of four fruit located near the outer surface of boxes B2, B4, B6 and B11 located on first layer (L1).

PAGE 230

230 Figure 617. Simulations of the pulp temperatures of four fruit located near the outer surface of boxes B27, B29, B32 and B39 located on the third layer ( L3) Figure 618. Simulations of the pulp temperatures of four fruit located near the outer surface of boxes B75, B77, B80 and B87 located on the sixth layer (AL6)

PAGE 231

231 Figure 619. Simulations of pulp temperature s of fruit F21, F63 and F105, all located on the threedimensional diagonal of box B1 (L1).

PAGE 232

232 Figure 620. Simulations of pulp temperature s of fruit F21, F63 and F105, all located on the threedimensional diagonal of box B88 (L6 ).

PAGE 233

233 Figure 621. Simulations of the average temperatures of layers L1, L4 and L7 during the ramp transfer at DXB airport (05 032007).

PAGE 234

234 Figure 622. Simulations of the temperature s near the center of RPC5 and RPC6 located on the first layer (L1) of the aircraft container during the ramp transfer at DXB airport (05032007).

PAGE 235

235 Figure 623. Simulations of the temperatures near the center of RPC25 and RPC26 located on the fourth layer (L4) of the aircraft container during the ramp transfer at DXB airport (05032007).

PAGE 236

236 Figure 624. Simulations of the temperatures near the center of RPC49 and RPC50 located on the seventh layer (L7) of the aircraft container during the ramp transfer at DXB airport (05032007).

PAGE 237

237 Table 61 Sensitivity analysis of Zehner's model for effective thermal conductivit y ke using the thermal conductivity of the solid (ks), the bulk porosity () and the contact surface ratio () as factorsz. 1 2 3 4 5 6 7 8 9 k s k e A (k e |k s ) s (k e |k s ) s (A (k e |k s )) A ( s (k e |k s )) A (k e |k s ) s k e |k s ) s (A (k e |k s )) A ( s (k e |k s ) 0.2 0.26 0 0.0 0.108 0.149 0.00109 0.000144 0.00136 0.134 0.00150 0.00780 0.0127 0.2 0.26 0 0.3 0.149 0.2 0.26 0 0.6 0.189 0.2 0.368 0.0 0.087 0.133 0.00141 0.2 0.368 0.3 0.133 0.2 0.368 0.6 0.179 0.2 0.476 0.0 0.071 0.119 0.00157 0.2 0.476 0.3 0.119 0.2 0.476 0.6 0.168 0.4 0.26 0 0.0 0.163 0.274 0.00821 0.000568 0.00919 0.244 0.00976 0.4 0.26 0 0.3 0.274 0.4 0.26 0 0.6 0.385 0.4 0.368 0.0 0.124 0.243 0.00952 0.4 0.368 0.3 0.243 0.4 0.368 0.6 0.363 0.4 0.476 0.0 0.094 0.216 0.00984 0.4 0.476 0.3 0.216 0.4 0.476 0.6 0.337 0.6 0.26 0 0.0 0.202 0.393 0.02419 0.001177 0.02575 0.350 0.0269 3 0.6 0.26 0 0.3 0.393 0.6 0.26 0 0.6 0.583 0.6 0.368 0.0 0.148 0.349 0.02680 0.6 0.368 0.3 0.349 0.6 0.368 0.6 0.549 0.6 0.476 0.0 0.11 0 0.309 0.02627 0.6 0.476 0.3 0.309 0.6 0.476 0.6 0.507 z : A() are averages and s are variances (population)

PAGE 238

238 Table 62. Parameters used with Zehner's model to calculate the effective thermal conductivity (ke) for the different simulations. k s k f k e Wm 1 K 1 Wm 1 K 1 Wm 1 K 1 Series 1 0.431 0.02505 0.315 0.356 0.22 Series 2 0.431 0.02469 0.425 0.493 0.22 Series 3 0.431 0.02461 0.420 0.328 0.18 LD3 Fruit 0.431 0.02487 0.290 0.348 0.23 LD3 Bottles 0.589 0.02487 0.383 0.427 0.28 Table 63. Parameters as well as thermal and physical properties used for the different simulations. k e s f pk e C Ps C Pf C Ppk C Pe T 0 Wm 1 K 1 kg m 3 Jkg 1 K 1 C Series 1 0.22 0.315 1005 1.209 687 3830 1007 2941 2.4 Series 2 0.22 0.425 865 1.229 498 3830 1007 2630 3.0 Series 3 0.18 0.420 930 1.233 540 3830 1007 2644 1.5 LD3 Fruit 0.23 0.290 999 1.219 930 y 719 3830 1007 1340 y 2942 6.4 LD3 Bottles 0.28 0.383 999 1.225 910 z 636 4185 1007 1925 z 2900 3.3 y : Properties of paper (engel and Ghajar, 2010) ; z : Properties of polypropylene (engel and Ghajar, 2010) Table 64. Convective and radiative heat transfer coefficients (Wm2K1) used for the aircraft container simulations. Walls Laboratory Fruit Air Transport Bottles h h rad h h rad W2 W5 W6 3.05 5.00 2.94 5.08 W4 3.13 5.00 3.06 5.08

PAGE 239

239 CHAPTER 7 CONCLUSIONS R esults of the thermal analysis on single boxes of fruit showed that the temperature distribution within the fruit and between the fruit and the surrounding air was relatively uniform. Consequently, the pulp temperature measured at a half radius distance can be considered as a good approximation of the mass average temper ature of the fruit and the surrounding air. Results also indicated that t emperatures were more uniform within the upper layers of fruit; the largest temperature variations within a layer were observed at the bottom of the boxes. Relatively fast rates of change of the temperatures were observed even in the core region of the boxes. The temperatures of the products located in the core of the aircraft container did not vary significantly during extended exposure to detrimental conditions. However, the pro ducts located near the outside surface, which account ed for close than 50% of the load, were significantly affected. Air transport tests s howed that the temperature distribution of the products within the aircraft container was mostly influenced by the env ironmental conditions during ramp transfer s, not by the conditions onboard the aircraft. Ramp transfers exceeding 8 h were observed during the tests. The effect of solar radiation during ramp transfer s was significant; it increased the temperatures withi n the aircraft container significantly above that of the ambient air. Results indicated that for produce with a relatively high rate of heat generation, the temperature in the core layers of the load would have increase d during transit and even during per iod s of refrigerated storage. In addition, the results suggested that the quality of fresh horticultural products transported by air is likely to be affected by factors

PAGE 240

240 such as anaerobic respiration, high relative humidity (condensation) and mechanical damage. Heat transfer simulations were run using a n effective thermal conductivity model. The approach provided acceptable res ults in the outer regions of the load for individu al box and aircraft container simulations However, the model significantly underpredicted the temperature in the core regions of individual boxes. Aircraft container simulations also underpredicted the temperatures at the top of the boxes located within the bottom layer and at the bottom of the boxes located within the top layer Consequently, it was established that the model did not adequately predict the temperature distribution throughout the loads of horticultural products that were studied. T o improve the results of such modeling approach, the effect of the natural convect ion must be included in the effective thermal conductivity v ia a variable dynamic component This new and more complex approach would not be appealing for the development of commercial temperature and quality monitoring solutions for shipments of horticul tural products However, since that most of the temperature variations within an aircraft container was observed in the peripheral region of the load, the modeling approach used, even without a dynamic component, could still be implemented as a decisional tool for air shipment of horticultural products.

PAGE 241

241 LIST OF REFERENCES Akinaga, T., Kohda, Y., 1992. Problems in the air shipment of horticultural products. In: Advances in Food Engineering. CRC, Boca Raton, FL, USA, 575587. Akinaga, T., Kohda, Y., 1993. Environmental condition during air shipment of horticultural products from Okinawa to T okyo. In: Proceedings of the International Conference for Agricultural Machinery & Process Engineering, Seoul, Korea, 413422. Alvarez, G., Flick, D., 2007. Modelling turbulent flow and heat transfer using macroporous media approach used to predict cooling kinetics of stack of food products. Journal of Food Engineering 80, 391401. Amos, N.D., Bollen, A.F., 1998. Predicting the deterioration of asparagus quality during air transport. In: Refrigerated Transport, Storage and Retail Display. Proceedings of the Conference of Commission D2/3, with D1, Cambridge, United Kingdom, 163170. Anony mous 2009a. Africa provides capacity relief? Air Cargo World 99 (7), 3234. Anonymous 2009b. Slow growth forecast for forwarders. Air Cargo World 99 (12), 4850. Argo, W.B., Smith, J.M., 1953. Heat transfer in packed beds prediction of radial rates in gas solid beds. Chemical Engineering Progress 49 (8), 443451. ASHRAE, 2006. Thermal properties of food. In: ASHRAE Handbook Refrigeration. American Society of Heating, Refrigeration and Air Conditioning Engineers, Atlanta, GA, USA, 9.1 9.31. ASHRAE, 1995. Aircraft. In: ASHRAE Handbook HVAC Applications. American Society of Heat ing, Refrigeration and Air Conditioning E ngineers, Atlanta, GA, USA, 9.1 9.8. ASHRAE, 2001. Climatic design information. In: ASHRAE Handbook Fundamentals. American Society of Heating, Refrigeration and Air Conditioning Engineers, Atlanta, GA, USA, 27.1 27 .71. Baird, C.D., Gaffney, J.J., 1976. A numerical procedure for calculating heat transfer in bulk loads of fruits or vegetables. ASHRAE Transactions 82 (2), 525540. Bauer, R., Schlnder, E.U., 1978a. Effective radial thermal conductivity of packings in gas flow. part I. convective transport coefficient. International Chemical Engineering 18 (2), 181188.

PAGE 242

242 B auer, R., Schlnder, E.U., 1978b. Effective radial thermal conductivity of packings in gas flow. part II. thermal conductivity of the packing fraction without gas flow. International Chemical Engineering 18 (2), 189204. Bazan, T., Chau, K.V., Baird, C.D., 1989. Heat transfer simulation of the bulk cooking of fruits. In: Proceedings of the 1989 International Winter Meeting of the American Society of A gricultural Engineering, New Orleans, LA, USA, Bazan, T., 1989. Mathematical modeling of heat transfer in the cooling of fruit in closed containers. PhD thesis. University of Florida, Gainesville, FL, USA. Bellagha, S., Chau, K.V., 1985. Heat and mass transfer during the cooling of tomatoes individually and in bulk. In: Proceedings of the 1985 Summer Meeting of the American Society of Agricultural Engineers, Michigan State University, East Lansing, MI, USA, Paper 856002. Beukema, K.J., 1980. Heat and m ass transfer during cooling and storage of agricultual products as influenced by natural convection. PhD thesis. Wageningen University, Wageningen, The Netherlands. Beukema, K.J., Bruin, S., Schenk, J., 1982. Heat and mass transfer during cooling and stor age of agricultural products. Chemical Engineering Science 37 (2), 291298. Beukema, K.J., Bruin, S., Schenk, J., 1983. Threedimensional natural convection in a confined porous medium with internal heat generation. International Journal of Heat and Mass Transfer 26 (3), 451458. Bhattacharyya, D., Pei, D.C.T., 1975. Heat transfer in fixed bed gas solid systems. Chemical Engineering Science 30, 293300. Bird, B., Stewart, W.E., Lightfoot, E.N., 1960. Transport Phenomena. John Wiley & Sons, New York, NY, USA. Bye, J.H., Bleasdale, M., 1985. Shipping perishables by air: 1985 and 1990. In: Technology Advances in Refrigerated Storage and Transport. Proceedings of Meetings of Commissions D1, D2 and D3, Orlando, FL, USA, 251256. engel, Y.A., Ghajar, A.J., 2 010. Heat and Mass Transfer : Fundamentals & Applications. McGraw Hill, New York, NY, USA. Chourasia, M.K., Goswami, T.K., 2006. Simulation of transport phenomena during natural convection cooling of bagged potatoes in cold storage, part I: Fluid flow and heat transfer. Biosystem Engineering 94 (1), 3345. Delele, M.A., Tijskens, E., Atalay, Y.T., Ho, Q.T., Ramon, H., Nicola, B.M., Verboven, P., 2008. Combined discrete element and CFD modelling of airflow through random

PAGE 243

243 stacking of horticultural products in vented boxes. Journal of Food Engineering 89, 3341. Dixon, A.G., Nijemeisland, M., Stitt, E.H., 2005. CFD study of heat transfer near and at the wall of a fixed bed reactor tube: Effect of wall conduction. Industrial & Engineering Chemistry Research 44 (16), 63426353. mond, J. P. Mercier, F., Nunes, M.C.N., 1999. In flight temperature conditions in the holds of an widebody aircraft. In: Proceedings of the 20th International Congress of Refrigeration, Sydney, Australia, 1924. Gorring, R.L., Churchill, S.W., 1961. Thermal conductivity of heterogeneous materials. Chemical Engineering Progress 57 (7), 5359. Gunjal, P.R., Ranade, V.V., Chaudhari, R.V., 2005. Computational study of a singlephase flow in packed beds of spheres. AIChE Journal 51 (2), 365 378. Gupta, M., Yang, J., Roy, C., 2002. Modelling the effective thermal conductivity in polydispersed bed systems: A unified approach using the linear packing theory and unit cell model. The Canadian Journal of Chemical Engineering 80, 830839. Harvey, J.M., Couey, H.M., Harris, C.M., 1966. Air transport of california strawberries: Factors affecting market quality in spring shipments 1966. Agricultural Research Service, United States Department of Agricul ture, Marketing Research Report No. 774. Heap, R.D., 2006. Cold chain performance issues now and in the future. In: Proceedings of the IIR IRHACE Conference, University of Auckland, New Zealand, 7587. Holman, J.P., 1986. Heat Transfer. McGraw Hill, New York, NY, USA. Hsu, C.T., Cheng, P., Won g, K .W., 1994. Modified zehner schlu nder models for stagnant thermal conductivity of prous media. International Journal of Heat and Mass Transfer 37 (17), 27512759. Hsu, C.T., Cheng, P., Wong, K.W., 1995. A lumped parameter model for stagnant thermal con ductivity of spatially periodic porous media. ASME Journal of Heat Transfer 117, 264269. IATA 2010. 2009: Worst demand decline in history encouraging year end improvements. 2010 (01/28), http://www.iata.org/pressroom/pr/2010012701.htm Kader, A.A., 2002a. Modified atmospheres during transport and storage. In: Postharvest Technology of Horticultural Crops. University of California, Agriculture and Natural Resources, Oakland, CA, USA, 135144.

PAGE 244

244 Kader, A.A., 2002b. Postharvest biology and technology: An overview. In: Postharvest Technology of Horticultural Crops. University of California, Agriculture and Natural Resources, Oakland, CA, USA, 3947. Kaviany, M., 2002. Principles of Heat Transfer. Wiley Interscience, New York, NY, USA. Koning, B., 2002. Heat and mass transport in tubular packed bed reactors at reacting and nonreacting conditions. PhD thesis. T wente University, Enschede, The Netherlands. Krupiczka, R., 1967. Analysis of thermal conductivity in granular materials. International Chemical Engineering 7 (1), 122144. Kulkarni, B.D., Doraiswamy, L.K., 1980. Estimation of effective transport properties in packed bed reactors. Catalysis Reviews, Science and Engineering 22 (3), 431483. Kunii, D., Smith, J.M., 1960. Heat transfer characteristics of porous rocks. AIChE Journal 6 (1), 7177. Kunii, D., Smith, J.M., 1961. Heat transfer characteristics of porous rocks: II. thermal conductivities of unconsolidated particles with flowing fluids. AIChE Journal 7 (1), 2934. Laguerre, O., Ben Amara, S., Alvarez, G., Flick, D., 2008. Transient heat transfer by free convection in a packed bed of spheres: Comparison between two modelling approaches and experimental results. Applied Thermal Engineering 28, 1424. Lemcoff, N.O., Pereira Duarte, S.I., Martinez, O.M., 1990. Heat transfer in packed beds. Reviews in Chemical Engineering 6 (4), 229292. Logtenberg, S.A., Dixon, A.G., 1998. Computational fluid dynamics studies of the effects of temperaturedependent physical properties on fixedbed heat transfer. Industrial & Engineering Chemistry Research 37, 739747. Logtenberg, S.A., Nijemeisland, M., Dixon, A.G., 1999. Computational fluid dynamics simulations of fluid flow and heat transfer at the wall particle contact points in a fixed bed reactor. Chemical Engineering Science 54, 24332439. Nijemeisland, M., Dixon, A.G., 2001. Comparison of CFD simulations to experiment for convective heat transfer in a gas solid bed. Chemical Engineering Journal 82, 231246. Nijemeisland, M., Dixon, A.G., 2004. CFD study of fluid flow and wall heat transfer in a fixed bed of spheres. AIChE Journal 50 (5), 906921.

PAGE 245

245 Oskam, N.A., Sillekens, J.J.M., Ceton, C., 1998. Validated models of the thermodynamic behaviour of perishables during flight. In: Refrigerated Transport, Storage and Retail Display. Proceedings of the Conference of Commission D2/3, with D1, Cambridge, United Kingdom, 171178. Pelletier, W., 2002. Study of fog formation in non ventilated aircraft cargo compartment. Master thesis. Universit Laval, Qubec, QC, Canada. Pelletier, W., 2007. Study of the performance of different ULD protective covers for temperature sensitive shipments. Center for Food Distribution and Retailing, University o f Florida, Gainesville, FL, USA. Sharp, A.K., 1989. Air transport of strawberries from australia: Requirements and possibilities. Food Australia 41 (5), 755760. Sharp, A.K., 1998. Air transport of perishables. In: Food Transportation. Blackie Academic & Professional, London, UK, 97110. Singh, R.P., Heldman, D.R., 1993. Introduction to Food Engineering. Academic Press, San Diego, CA, USA. Smale, N.J., Moureh, J., Cortella, G., 2006. A review of numerical models of airflow in refrigerated food applicati ons. International Journal of Refrigeration 29 (6), 911930. Sowinski, L.L., 2000. Entering the century of air cargo. World Trade 13 (10), 110111. Specchia, V., Baldi, G., Sicardi, S., 1980. Heat transfer in packed bed reactors with one phase flow. Chem ical Engineering Communications 4, 361380. Talbot, M.T., 1987. Pressure and velocity distribution for air flow through fruits packed in shipping containers using porous media flow analysis. PhD thesis. University of Florida, Gainesville, FL, USA. Tanner D.J., 1998. Mathematical modelling of heat and mass transfer and water vapour transport in apple coolstores. PhD thesis. Massey University, Palmerston North, New Zealand. Tanner, D.J., Cleland, A.C., Opara, L.U., 2002a. A generalised mathematical modell ing methodology for design of horticultural food packages exposed to refrigerated conditions: Part 2, heat transfer modelling and testing. International Journal of Refrigeration 25 (1), 4353. Tanner, D.J., Cleland, A.C., Opara, L.U., Robertson, T.R., 200 2b. A generalised mathematical modelling methodology for design of horticultural food packages exposed to refrigerated conditions: Part 1, formulation. International Journal of Refrigeration 25 (1), 3342.

PAGE 246

246 Tanner, D.J., Cleland, A.C., Robertson, T.R., 2002 c. A generalised mathematical modelling methodology for design of horticultural food packages exposed to refrigerated conditions: Part 3, mass transfer modelling and testing. International Journal of Refrigeration 25 (1), 5465. Tsotsas, E., Martin, H., 1 987. Thermal conductivity of packed beds: A review. Chemical Engineering and Processing 22, 1937. Verboven, P., Flick, D., Nicola, B.M., Alvarez, G., 2006. Modelling transport phenomena in refrigerated food bulks, packages and stacks: Basics and advances. International Journal of Refrigeration 29 (6), 985997. Verboven, P., Hoang, M.L., Baelmans, M., Nicola, B.M., 2004. Airflow through beds of apples and chicory roots. Biosystems Engineering 88 (1), 117125. Verboven, P., Hoang, M.L., Schenk, A ., Nicola, B.M., 2001. Effect of pallet box stacking on cooling of horticultural products in cool stores. Model IT II, Acta Horticulturae 566 239244. Verboven, P., Nicola, B.M., 2005. CFD: A fresh breeze in engineering and design of food cooling pr ocesses. 2007. Villeneuve, S., Mercier, F., Pelletier, W., mond, J. ., 2000. Effect of environmental conditions on air shipment of perishables during ground operations. In: Proceedings of the American Society of Agricultural International Meeting, Milwaukee, WI, USA, Paper 006058. Villeneuve, S., Vzina, J., mond, J. ., 2002. La logistique du froid dans le transport arien des denres prissables. Revue Gnrale du Froid (1029), 3238. Vortmeyer, D., Schaefer, R.J., 1974. Equivalence of oneand twophase models for heat transfer processes in packed beds: One dimensional theory. Chemical Engineering Science 29, 485491. Wang, H., Ceton, C., Touber, S., 1995. Modelling the microclimate of packed cut flowers during precooling. In: Proceedings of the 19th International Congress of Refrigeration Commission C2, Hague, The Netherlands, 707714. Woodside, W., 1958. Calculation of the ther m al conductivity of porous media. Canadian Journal of Physics 36, 815823. Yagi, S., Kunii, D., 1957. Studies of effective thermal conductivities in packed beds. AIChE Journal 3 (3), 373381. Yagi, S., Kunii, D., 1960. Studies on heat transfer near wall surface in packed beds. AIChE Journal 6 (1), 97104.

PAGE 247

247 Yagi, S., Kunii, D., Wakao, N., 1960. Studies on axial effec tive thermal conductivities in packed beds. AIChE Journal 6 (4), 543546. Yagi, S., Wakao, N., 1959. Heat and mass transfer from wall to fluid in packed beds. AIChE Journal 5 (1), 7985.

PAGE 248

248 BIOGRAPHICAL SKETCH William Pelletier was born in Saint Pascal de Kamouraska, in Qubec (Canada), in 1977. He completed his highschool degree in Saint Pascal before starting in 1994 a two year degree in S ciences at the Cgep de RivireduLoup. In 2000, he received his Bachelor D egree in F ood E ngineering at Unive rsit Laval in Qubec City. In 2002, he e arned his Master of Science in A grifood E ngineering from the same institution ; h is thesis was on the formation of fog in nonv entilated aircraft cargo compartments. The project was supported by a grant from the Nat ural Sciences and Engineering Research Council of Canada. In 2003, he started his doctoral studies at the University of Florida supported by a grant from the Fonds Qubecois de Recherche sur la Nature et les Tech nologies and the Walter H. Johnson, Jr. Scholarship from The International Air Cargo Association (2004) Also, since 2007, he was the instructor for the course Heat and Mass Transfer in Biological S ystem s (ABE 3612C) provided through the Agricultural and Biological Engineering D epartment In 2 010, he received the Jack L. Fry Award for teaching excellence by a graduate student in the College of Agricultural and Life Sciences.