UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Heat transfer studies of steam/air mixtures for food processing in retort pouches Ramaswamy, Hosahalli Subrayasastry 1983

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1983_A1 R35.pdf [ 10.58MB ]
Metadata
JSON: 831-1.0095894.json
JSON-LD: 831-1.0095894-ld.json
RDF/XML (Pretty): 831-1.0095894-rdf.xml
RDF/JSON: 831-1.0095894-rdf.json
Turtle: 831-1.0095894-turtle.txt
N-Triples: 831-1.0095894-rdf-ntriples.txt
Original Record: 831-1.0095894-source.json
Full Text
831-1.0095894-fulltext.txt
Citation
831-1.0095894.ris

Full Text

HEAT TRANSFER STUDIES OF STEAM/AIR MIXTURES FOR FOOD PROCESSING IN RETORT POUCHES by HOSAHALLI SUBRAYASASTRY RAMASWAMY B.Sc, Bangalore University, 1970 M.Sc. (Food Technology), Mysore University, 1972 M.Sc. (Food Science), The University of British Columbia, 1979 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Food Science) We accept the thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1983 © Hosahalli Subrayasastry Ramaswamy, 1983 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. (Hosahalli S. Ramaswamy) Department of Food Science The University of British Columbia 1956 Main Mall Vancouver, B.C. Canada V6T 1Y3 Date Ap/Vl 2?,W3 i i ABSTRACT Heat transfer characteristics associated with steam/air mixtures were studied in two pilot scale batch type steam/air retorts: a vertical positive flow retort and a horizontal forced circulation Lagarde retort. A method em-ploying transient heat conduction into rectangular bricks of aluminum and stainless steel was developed to evaluate the surface heat transfer coeffi-cient (h) of steam/air mixtures. A system was designed to facilitate an in-stantaneous drop of the test brick, from an insulated box inside the retort, into a specified steam/air medium after the come-up period. The influences of steam content, temperature, flow rate and flow direction of the heating medium and orientation of test bricks on the associated h values, tempera-ture distribution and pressure stability in the retorts were studied. In ad-dition, thermal processing efficacy was evaluated by measuring the rate of heat penetration into bricks of silicone rubber and rigid nylon which have thermal diffusivities in the range common for foods. In both retorts, steam content (S) of the mixture was found to be the major factor influencing h (p<0.05); however, temperature had no signifi-cant effect (p>0.05). Further, the flow direction and flow rate of the heat-ing media in the positive flow retort, and brick orientation in the Lagarde retort also influenced h (p<0.05). The general relationship between h and S was exponential: h = a exp(bS). In the positive flow retort with the test brick in the vertical orientation, the values of a and b were 153 W/m^ C and 0.0421 respectively, for steam/air media flowing in an upward direc-tion, and were 337 W/m^ C and 0.0355 respectively, for the media flow-ing downward. The surface heat transfer coefficient was also found to increase linearly with the medium flow rate. With the Lagarde retort, i i i steam/air flow was always horizontal and flow rate was not adjustable. In this case, h was influenced by the test brick orientation. For bricks in the vertical orientation, the exponential parameters, a and b, were 1011 W/m^c and 0.0226 respectively, whereas in the horizontal orientation, these were 1669 W/m2C and 0.0132 Temperature distribution studies in the positive flow retort indicated that the overall standard deviation of the medium temperature at several lo-cations during the cook period (excluding come-up) increased (p<0.05) with a decrease in the steam content and flow rate of the heating media. The ef-fects of temperature and flow direction were nonsignificant (p>0.05). In the Lagarde retort, the temperature distribution was not influenced either by steam content or temperature of the steam/air medium. Pressure stability studies indicated that the air content and temperature of the medium in-creased (p<0.05) the standard deviations of retort pressure during the cook period. Based on the temperature and pressure deviations in the two re-torts, steam/air mixtures with 86-90% steam contents were considered to pro-vide satisfactory overriding air pressures for processing of retort pouches at 105-120°C. Heat penetration studies in the positive flow retort using nonpackaged test bricks of silicone rubber and nylon revealed an increase of up to 11% (p<0.05) in the heating rate index (f) of test bricks when the steam con-tent of the media decreased from 100% to 50%. Heating of bricks at 120°C resulted in f values that were 5.5% larger (p<0.05) than those for bricks heated at 105°C. In the Lagarde retort, the effects of temperature and steam content of the media on f values were not significant. Heating bricks in the vertical orientation resulted in higher f values than in horizontal orientation in some tests, while a reverse trend was observed in others. iv The influence of entrapped air (15-30 mL per pouch) in retort pouches containing the bricks on f values was small when using a vertical rack that tightly constrained the bricks, whereas up to 260% higher values of f resulted when using an unconstraining horizontal rack while processing at 105-120°C in media of steam contents above 65%. These increases in f value could be prevented by using overriding air pressures of 70-100 kPa during the retort operation. The lag factor, j , was generally in the range of 0.5-1.0 for test bricks, with or without packaging, in the positive flow retort, and 0.8-1.1 in the Lagarde retort, when evaluated at 42% effectiveness for the come-up time. It was observed that in order for the j values to match the theoretical value of 1.27 for an infinite plate, the effectiveness was in the range of 60-V TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS v LIST OF TABLES ix LIST OF FIGURES xiii NOMENCLATURE xv ACKNOWLEDGEMENTS xviii INTRODUCTION 1 LITERATURE REVIEW 6 Historical Background 6 Factors Affecting Heat Transfer Rates 8 Processing Medium 8 Medium Composition 12 Medium Temperature 13 Medium Flow Rate 14 Medium Flow Direction 15 Package Orientation 15 Residual Gases 16 Temperature Distribution and Heat Penetration Studies 18 Heat Transfer Involving Packaged Foods 20 Theoretical Considerations 20 Thermal Properties of Foods 24 EXPERIMENTAL 26 Surface Heat Transfer Coefficients of Steam/Air Mixtures 26 Estimation Procedure 26 A Quick Release System 26 vi Test Bricks 28 Testing Procedure 32 Factors Affecting Surface Heat Transfer Coefficient 35 The Positive Flow Retort 35 The Lagarde Retort 37 Experimental Design 39 Temperature Distribution in the Steam/Air Retorts 40 Data Acquisition and Analyses 40 Placement of Thermocouples 40 Horizontal Rack 40 Vertical Rack 42 Factors Affecting Temperature Distribution 44 The Positive Flow Retort 44 The Lagarde Retort 44 Pressure Stability in the Steam/Air Retorts 45 Heat Penetration Studies in the Steam/Air Retorts 46 Fabrication of Test Bricks 46 Factors Affecting Heat Penetration 47 RESULTS AND DISCUSSION 49 Surface Heat Transfer Coefficients of Steam/Air Mixtures 49 Estimation Procedure 49 Factors Affecting Surface Heat Transfer Coefficient 54 The Positive Flow Retort 54 The Lagarde Retort 62 An Analysis of the Sources of Error 64 Temperature Distribution in the Steam/Air Retorts 70 Data Analyses 70 v i i Factors Affecting Temperature Distribution 73 The Positive Flow Retort 73 The Lagarde Retort 77 Pressure Stability in the Steam/Air Retorts 81 General Importance 81 Factors Affecting Pressure Stability 81 The Positive Flow Retort 81 The Lagarde Retort 83 Considerations for Satisfactory Overriding Pressures 83 Heat Penetration Studies in the Steam/Air Retorts 87 The Heating Rate Index 87 Factors Affecting Heating Rate Index 88 The positive flow retort 88 The Lagarde retort 96 The Lag Factor.. 102 Theoretical Considerations 106 Heat Transfer in Rectangular Solids 106 The Heating Rate Index 106 The Lag Factor 108 Limiting Surface Heat Transfer Coefficient 113 Pressure/Volume Relationships in Steam/air Processing 117 Effect of Entrapped Air 118 Effect of Water Vapor 119 Effect of Entrapped Air Plus Water Vapor 120 SUMMARY AND CONCLUSIONS 121 LITERATURE CITED 125 Appendix I.... 135 viii Appendix II 138 Appendix III 139 Appendix IV 144 Appendix V 148 Appendix VI 149 Appendix VII 150 Appendix VIII 152 Appendix IX 154 Appendix X 155 Appendix XI 155 Appendix XII 156 Appendix XIII 157 i x LIST OF TABLES Table 1. Factors i n v e s t i g a t e d i n the f r a c t i o n a l f a c t o r i a l experiments f o r the pos i t i v e flow r e t o r t Table 2. T y p i c a l heat t r a n s f e r coefficient r e s u l t s f or steam/air mixtures i n the positive flow retort Table 3. A n a l y s i s of variance i n the surface heat t r a n s f e r c o e f f i -cients of steam/air mixtures i n the positive flow retort Table 4. Surface heat t r a n s f e r coefficient as a function of steam content i n the positive flow retort Table 5. Regression details for the dependence of surface heat t r a n s -f e r coefficients on the medium flow rate i n the positive flow re t o r t Table 6. Surface heat t r a n s f e r coefficient as a function of steam content i n the Lagarde retort Table 7. Estimated e r r o r s i n the evaluation of heat t r a n s f e r c o e f f i -cients of steam/air mixtures while using the aluminum and stainless steel test b r i c k s of di f f e r e n t t h i c k n e s s Table 8. Overall standard deviation in temperature during the cook period under various steam/air conditions in the positive flow retort for the fractional factorial experiments Table 9. Analysis of variance in temperature distribution for studies of the positive flow retort Table 10. Summary of overall standard deviations in temperataure after the come-up period in the positive flow retort Table 11. Analysis of variance in temperature distribution for studies of the Lagarde retort Table 12. Summary of overall standard deviations in temperature after the come-up period in the Lagarde retort Table 13. Analysis of variance in pressure stability for tests in the positive flow retort Table 14. Analysis of variance in pressure stability for tests in the Lagarde retort Table 15. Maximum steam contents in steam/air heating media for pro-viding satisfactory overriding air pressures in the positive flow and Lagarde retorts xi Table 16. Mean f values for silicone rubber and nylon bricks in the positive flow retort for the fractional factorial experiment 89 Table 17. Analysis of variance in f values for various factors tested using nonpackaged thick nylon bricks in the positive flow retort 91 Table 18. Mean f values for silicone rubber and nylon bricks in the positive flow retort... 92 Table 19. Analysis of variance in f values using nonpackaged silicone rubber and nylon bricks in the positive flow retort 93 Table 20. Analysis of variance in f values using packaged and non-packaged silicone rubber bricks in the positive flow retort.. 95 Table 21. Mean f values for silicone rubber and nylon bricks in the Lagarde retort 97 Table 22. Analysis of variance in f values using non-packaged silicone rubber and nylon bricks in the Lagarde retort 98 Table 23. Analysis of variance in f values using packaged and non-packaged silicone rubber bricks in the Lagarde retort 99 xii Table 24. Mean j values for silicone rubber and nylon bricks in the positive flow retort 103 Table 25. Mean j values for silicone rubber and nylon bricks in the Lagarde retort 104 Table 26. Predicted f values for the nonpackaged silicone rubber and nylon bricks in the positive flow and Lagarde retorts 107 Table 27. Limiting surface heat transfer coefficients and associated f values for steam/air processing 115 x i i i LIST OF FIGURES Figure 1. An i n v e r t e d view of the in s u l a t e d box for the quick, r e -lease of the test b r i c k into the heating medium 27 Fi g u r e 2. A schematic representation of the quick release system po-sitioned i n the positive flow r e t o r t 29 Fi g u r e 3. A schematic representation of the quick release system po-sitioned i n the Lagarde ret o r t 30 Fig u r e 4. An exploded view of the test b r i c k showing fabrication details 31 Fi g u r e 5. Flow chart of the i t e r a t i v e computer program for the eva-lua t i o n of surface heat t r a n s f e r coefficient 34 Fi g u r e 6. A schematic representation of the construction details of the positive flow retort 36 Fi g u r e 7. Three orientations of the test b r i c k i n the Lagarde r e t o r t . 38 F i g u r e 8. A diagram showing the c o n s t r u c t i o n a l details of the h o r i -zontal r a c k i n g system for retort pouches 41 Fi g u r e 9. A diagram showing the c o n s t r u c t i o n a l details of the v e r t i c a l r a c k i n g system for reto r t pouches 43 xiv Figure 10. Surface heat transfer coefficient as a function of flow rate for steam/air mixtures in the positive flow retort Figure 11. Surface heat transfer coefficient as a function of steam content for steam/air mixtures in the positive flow retort.. 61 Figure 12. Surface heat transfer coefficient as a function of steam content for steam/air mixtures in the Lagarde retort 65 Figure 13. Temperature histories for all thermocouple locations during the cook, period of a test run in the positive flow retort... ?1 Figure 14. Means and standard deviations of temperature for all thermocouple locations at each recorded time during the cook period of a test run in the positive flow retort 72 Figure 15. Means and standard deviations of temperature at each thermocouple location over the entire cook period of a test run in the positive flow retort 74 Figure 16. Typical heat penetration curves for silicone rubber bricks and cylinders in a steam/air medium containing 85% steam at 120°C 1 0 1 Figure 17. Hypothetical heat penetration curves for a packaged food in a retort with and without come-up periods HO X V NOMENCLATURE a Half-length of a rectangular brick; intercept coefficient in Eg. (14) Bi Biot number (hd/k) b Half-width of a rectangular brick; slope coefficient in Eq.(14) c Half-thickness of a rectangular brick; intercept coefficient in the surface heat transfer coefficient - flow rate relationship d Significant dimension of an infinite plate equal to its thickness or half-thickness depending on it being heated from one side or both sides; slope coefficient in the surface heat transfer coefficient -flow rate relationship Fo Fourier number ( at/d^) F Q Lethality achieved at the center of a package during processing f Negative reciprocal slope of the straight line portion of a heating curve represented by log(Ta-T) vs. t h Surface heat transfer coefficient j Lag factor represented by (Ta-Tpi)/(Ta-Ti) assuming 42% ef-fectiveness for the come-up period jO Lag factor obtained assuming the effectiveness of the come-up time to be zero k Thermal conductivity n Number of observations Pa Total retort pressure due to steam and air Ps Saturated steam pressure at temperature Ta PI Pressure inside a pouch at temperature T l p Probability level for testing statistical significances R Intercept coefficient in Eq.(3) xvi R(pl) R value for an infinite plate of thickness 2a R(p2) R value for an infinite plate of thickness 2b R(p3) R value for an infinite plate of thickness 2c r^ Coefficient of determination S Slope coefficient in Eq.(3); percentage steam in a steam/air mixture S(pl) S value for infinite plate of thickness 2a S(p2) S value for infinite plate of thickness 2b S(p3) S value for infinite plate of thickness 2c T Temperature at x when using Eq.(l) and at the center otherwise Ta Temperature of the heating medium Ti Initial temperature of the material being heated Tpi Pseu do -initial temperature of the material being heated, as obtained from the intercept of the straight line portion of the heating curve extended to the beginning of heating process TpiO Same as Tpi, but calculated at time zero (steam on) when a finite come-up period exists Tpil Same as TpiO, but calculated at time tr T l Temperature of the pouch contents at the time of packaging t Heating time tc Come-up time tr Noneffective portion of the come-up time U Unaccomplished temperature ratio, (Ta-T)/(Ta-Ti) U(b) U for a rectangular brick U(pl) U for an infinite plate of thickness 2a U(p2) U for an infinite plate of thickness 2b U(p3) U for an infinite plate of thickness 2c Va Volume of the headspace air in a pouch at Ta x v i i V I Volume of the headspace a i r i n a pouch at T l x Distance from the c e n t r a l plane of an i n f i n i t e plate a Thermal d i f f u s i v i t y B Root of Eq.(2) Bl F i r s t p o s i t i v e root of Eq.(2) B n nth positive root of Eq.(2) xviii ACKNOWLEDGEMENTS The author wishes to express his sincere appreciation to Dr. Marvin A. Tung for his interest, advice, direction and encouragement throughout the course of this research project and review of the thesis. He is thankful to the members of his research committee: Dr. W.D. Powrie of the Department of Food Science, Dr. K.V. Lo of the Department of Bio-Resource Engineering and Dr. T.D. Garland of East Chilliwack Agri-cultural Co-op., for their constructive criticism and review of the thesis. He is indebted for financial support of this research by the Canadian Commonwealth Scholarship and Fellowship Plan, and the Natural Sciences and Engineering Research Council of Canada Strategic Grants Program and the Agriculture Canada PDR Contract Program. He also extends thanks to Agnes Papke, Trudi Smith and Edwin Kwong for their valuable assistance in this research project. He is extremely grateful to his parents, Smt. Leelavati Subrayasastry and Sri. Hosahalli Subrayasastry, and his wife, Smt. Rajalakshmi Ramaswamy for their understanding and support, without which this study would not have been possible. And finally, he dedicates the thesis to the memories of his grandparents. -1-INTRODUCTION The retort pouch is one of the significant packaging innovations of the past two decades for thermally processed foods. It was described as the "Package of the 80's" and was viewed by some as having the potential to capture a large part of the 34 billion units per year U.S. metal can market (Pinto, 1978). Development of the retort pouch food ranged from the lab work in the early 1950's to use in the Apollo space program beginning in 1968 (Mermelstein, 1976), and commercial feasibility in the early 1960's. Currently, retort pouch foods are being marketed in Canada, Europe, Japan and the U.S.A. The retort pouch, a flexible laminated thin profile food package, is usually made by bonding together films of three materials: an outer film of polyester for strength and durability; a middle film of aluminum foil as a barrier to moisture, light, gas and micro-organisms, and an inner film of polyethylene, polypropylene, or polyolefins as food contact and heat seal material. The retort pouch offers a number of potential advantages over conventional packaging for foods: superior product quality due to faster heat penetration, shelf stability without refrigeration or freezing, easier and faster end-use preparation, better conservation of storage space before and after packing, and easier opening and disposal (Mermelstein, 1976). Studies by Hoddinott (1975), Steffe et al. ( 1980) and Williams et al. (1981) also indicated that the retort pouch packaging system offers possible energy savings in production, processing, transportation and home-preparation. The retort pouch earned the 1977 Gordon Royal Maybee Award of the Cana-dian Institute of Food Science and Technology to Swan Valley Foods, Richmond, B.C.; the Institute of Food Technologists 1978 Industrial -2-Achievement Award and the Food Processing 1978 Top Honors Packaging Award to the U.S. Army Natick R&D Command, (Natick, MA), Continental Flexible Packaging (Chicago, IL)- A Member of the Continental Group, Inc. (formerly Continental Can Co.), and the Flexible Packaging Division (Richmond, VA) of Reynolds Metals Inc.; and the 1978 President's Award of the National Flexible Packaging Association to Specialty Seafoods Inc. (Anacortes, WA) for their smoked fillets of salmon in a retort pouch. Three methods have been used to commercially sterilize foods in flex-ible pouches: steam cook, water/air cook and steam/air cook. The common feature in all these is the necessity to have a positive retort pressure above the internal pouch pressure throughout the retorting operation, particularly during the cooling, in order to protect the package integrity and to retain their original seal strengths. While steam/air processes have been shown to be commercially feasible in Canada, Europe and Japan, there has been some hesitation in the U.S.A. to its acceptance. The reported lower heat transfer coefficient values for steam/air mixtures as compared with water/air heating (Pflug, 1964) appears to be the point of concern. It is important to know the heat transfer characteristics of a heating medium for its efficient application in any processing operation. Steam/air media have been studied by a number of researchers as related to engineer-ing applications in steam power plants, distillation processes, humidifica-tion/dehumidification of air in air conditioning and dehydration operations (Abdul Hadi, 1979; Charm, 1978; Coulson and Richardson, 1977; Hebbard and Badger, 1934; Heldman and Singh, 1981; Kern, 1950; Kreith, 1973; Kusak, 1958; Othmer, 1929). The majority of such studies in the published literature employ steady state methods which maintain the surface tempera-ture of the test material at 5-20 C° below the medium temperature. The -3-results, therefore, may not be directly applicable to thermal processing situations in which the material surface is within a fraction of a degree of the medium temperature for a greater part of the process time. Further, the above applications involve only small quantities of either steam or air in the steam/air mixture and hence differ from thermal processing operations in which appreciable quantities of both steam and air are employed in the media with the dual purpose of achieving heat transfer while simultaneously providing an overriding pressure to prevent pouch bursting. Very few studies have been carried out to evaluate heat transfer from steam/air media under food processing conditions. The mechanisms of heat transfer from steam or steam/air are known to be mainly by means of condensation and convection, as has been well docu-mented in the literature (for example Kreith, 1973; McAdams, 1954). The order of magnitude (<10,000 W/m^ C) of surface heat transfer coefficient (h) for steam or steam/air mixtures reported under food processing conditions (Adams and Peterson, 1982; Blaisdell, 1963; Pflug, 1964; Pflug and Borrero, 1967; Yamano, 1976) indicates the type of condensation of steam to be filmwise. During the thermal processing operations, the efficiency of heat transfer from a steam/air medium may depend on its steam content, temperature, flow rate and flow direction. The product and package char-acteristics may also play a role. No detailed study of the effects of these factors during steam/air processing of foods has so far been carried out. Temperature fluctuations and the presence of cold spots in the retort during the retorting operation can be detrimental to sterilizing of foods in thin profile packaging. Martens (1980) reported that a deviation in the re-tort temperature of 0.5 C° can introduce a change in the resultant process lethality up to 3 min depending on the temperature used for processing of -4-foods in fiat containers. For a similar temperature deviation, Berry (1979) reported differences in the process lethalities up to 14% at a temperature of 121°C. Thus, a reasonable assurance of uniform temperature distribution during the retorting operation is vital for the success of steam/air pro-cessing. Thermal processing of solid foods in flexible or rigid containers basi-cally involves convective heat transfer at the surface of the container and conduction within. These are generally described in terms of three dimen-sionless parameters: temperature ratio, (Ta-T)/(Ta-Ti); Biot number, (hd/k), and Fourier number, ( at/d 2). The temperature ratio is related to the Fourier number through a function of Biot number. These relationships have been recently reviewed and applied to thin profile geometries by Martens (1980) and Monteza (1979), and simplified to be applicable to transient conditions of heat transfer into finite and infinite solids (Ramaswamy et al., 1982). Information on thermal properties of foods and packaging materials is essential in order to use these heat transfer equa-tions. These are: 1) thermal diffusivity, 2) thermal conductivity and 3) surface heat transfer coefficient. In addition, the physical form in terms of product shape and size must be known. These relationships have been generally simplified for use in situations involving thermal processing of foods in steam by assuming the resistance to heat transfer at the container surface to be negligible in comparison with the resistance within the con-tainer. Consequently, the two thermal process parameters, developed initially by Ball (1923) to describe a heat penetration curve, the heating rate index, f, and the lag factor, j, have been considered to be indepen-dent of the heat transfer characteristics of the heating medium. Such independence may not be valid while using steam/air mixtures due to the -5-reduced surface heat t r a n s f e r c o e f f i c i e n t . V ery l i t t l e attention has been g i v e n to s o l v i n g the heat conduction equations under steam/air processing conditions. T h u s , f u r t h e r work appears to be needed to provide a more comprehensive u n d e r s t a n d i n g of the concepts i n v o l v e d i n the application of steam/air mixtures to food p r o c e s s i n g . The objectives of this investigation were: 1. To develop an appropriate transient method to evaluate the surface heat t r a n s f e r coefficient of steam/air mixtures as applied to thermal processing s i t u a t i o n s . 2. To s t u d y the influence of process v a r i a b l e s such as steam content, tem-p e r a t u r e , flow rate and flow d i r e c t i o n of the steam/air mixtures, and orientation of the test b r i c k on the surface heat t r a n s f e r coefficient. 3. To study the effect of the above process v a r i a b l e s on the temperature d i s t r i b u t i o n and p r e s s u r e s t a b i l i t y i n steam/air r e t o r t s . 4. To study the effects of above process v a r i a b l e s on the thermal process parameters, f and j , using rectangular b r i c k s of silicone r u b b e r and nylon materials i n v e r t i c a l and h o r i z o n t a l orientations. 5. To use heat t r a n s f e r models to pred i c t f and j values, and compare these with experimental values. 6. To i d e n t i f y a l i m i t i n g surface heat t r a n s f e r coefficient beyond which f values could be p r e d i c t e d assuming negligible surface resistance to heat t r a n s f e r . -6-LITERATURE REVIEW Historical Background It is well recognized that the U.S. Army Natick R&D Command, in col-laboration with Continental Flexible Packaging and Reynolds Metals Inc., did the exploratory and reliability studies, which began in the mid-1950's and required over two decades, to make the retort pouch concept a commercial reality (Lampi, 1977, 1980; Mermelstein, 1976,1978; Toumy and Young, 1982). The retort pouch concept was conceived in the earlier studies by Hu (1953), Hu et al. (1955), Nelson and Steinberg (1956), Mannheim et al. (1957) and Nelson et al. (1956) at the University of Illinois. While dis-cussing some advantages and problems associated with plastic containers for the purpose of heat processing, the above authors indicated that films with special properties had to be developed for any further work in the area to be meaningful. Wallenberg and Jarnhall (1957) reported that the plastic films, polyester and polyamide (nylon), were capable of withstanding auto-claving conditions. Klis (1961), Mayer and Robe (1961), Morgan et al. (1961) and Tripp (1961) reported the successful development of polyester-foil-polyethylene and vinyl laminates, and the feasibility of processing low acid foods packaged in these laminates. From 1959 to 1966, over two hun-dred film materials were screened at the Natick Laboratories, with respect to strength and durability, bacterial attack and penetration, and migration of extractable substances. This was then followed by a reliability project for about four years (Gould et al., 1962; Kellar, 1959a, 1959b; Rubinate, 1960, 1961). Goldfarb (1970, 1971) conducted a massive development and testing program to demonstrate the feasibility of producing and marketing -7-low acid foods in flexible retort pouches. An account of these early efforts was given by Lampi (1977). Commercial exploitation first surfaced in Europe with limited production of items marketed in Italy in 1960, Denmark in 1966, and Japan in 1969 (Lampi, 1980). Steady growth was reported in Europe up to the mid-1970's (Andrews, 1975). In Japan, the sales reached $70 million in 1973 (Tsutsumi, 1975) and grew to $259 million (460 million pouch packs) in 1978 with an ex-pected production of 800 million pouches in 1980 (Tsutsumi, 1979a). In North America, commercial production began in 1974 with the introduction of four retort pouch products by Swan Valley Foods in Vancouver, B.C. (Peters, 1975). Currently, two companies in Eastern Canada (Magic Pantry Foods Inc., Hamilton, ON; Empaquetage, J.B., Inc., St. Hyacinthe, PQ) have been marketing retort pouch foods. Canada's first Everest expedition, in 1982, used 2500 retort pouch meals prepared by the Canadian companies (Anon. 1982). It is rather strange that although the retort pouch concept was developed in the U.S.A. with over two decades of pioneering work, the major commerialization of the retort pouch foods has been achieved in Canada, Europe and Japan. Schulz (1978) reported that the sole barrier to the initial introduction of the retort pouches into the commercial U.S. mar-ket and to the initial procurement by the military was clearance of the pouch material by the Food and Drug Administration. According to Lampi (1980), the commercial success for the retort pouch in Europe and Japan was motivated by the lack of significant alternative preservation modes, and high costs for alternative packaging materials. In the U.S.A., there has not been the same necessity to aggressively develop and market the pouch. During the early 1970's, interest was high but soon waned during the FDA review of packaging material. However, the development of FDA-approved -8-adhesive systems in the retort pouch construction and the recent energy crisis have been causing renewed commercial interest. There are a number of U.S. companies producing retort pouch foods for retail markets at the present time and for the past several years (Hormel, Inc., Austin, MN; ITT Continental Baking Co., Rye, NY; Kraft Inc., Glenview, IL; Libbey, McNeill & Libby, Inc., Chicago, IL; Specialty Seafoods, Anacortes, WA). The marketing forecast for retort pouched foods indicates a very pro-mising future. Projections for 1980 were 1.5 billion pouches to be sold in Europe, 5 billion in Japan and 1.2 billion in Canada (Mermelstein, 1976). Forecasts by American firms for U.S. pouch sales ranged from a conserva-tive 70 million pouches by 1982 (American Can Co.) to a highly optimistic 210 million pouches (Reynolds Metals Co., Inc.) to be sold annually (Pinto, 1978). But these predictions may not be achieved for another five to ten years due to the slow speed of filling and sealing machinery which appears to be a major limitation at the present time. Factors Affecting Heat Transfer Rates Processing Medium Pflug (1964) and Pflug and Borrero (1967) recognized three potential heating media for processing of retort pouches: pure steam, water/air and steam/air. The method employing pure steam would have good surface heat transfer characteristics and the process could be controlled by a single factor, either medium temperature or pressure. The method, essentially, is based on the principle of high temperature-short time processing. Use of high temperature not only reduces the sterilization time, but also eliminates -9-the requirement for added air pressure if the steam pressure outside the pouch remains considerably higher than the pressure inside during the short heating period. However, due to the use of pure steam at higher temperatures, the heating time would have to be very carefully controlled to achieve the desired lethality while optimizing the potential benefits. Air would be reguired in the retort toward the end of the heating period and during cooling to combat the high internal pouch pressures. The "hi-retort" (sterilization temperature, 135°C) and the "u-retort" (sterilization temper-atures up to 150°C) developed by Toyo Seilcan Kaisha Ltd., Japan, belong to this group. When the former process was reported to be commercially available (Tsutsumi, 1979a), the latter was in the developmental stage (Tsutsumi, 1979b). There have been no reports on the use of ultra high temperature (UHT) pouch processing outside Japan. Steam heated water with an overriding air pressure is the most common heating medium currently employed in the U.S. and in some parts of Eur-ope. This method reguires control of water temperature and the total pres-sure. In conventional retorts, additional thermal energy is required at the start of the process to heat the water which will have to be drained prior to cooling. Water hardness may soil pouches and build up scale on the sepa-ration plates, and improper addition of steam to the water in the retort could produce vibrations that might damage or displace the pouches (Pflug et al., 1963). An advantage of water/air methods, on the other hand, is the flexibility in the use of overriding air pressure independent of the opera-tional temperatures. The new Convenience Food Sterilizer (FMC Corp., San-ta Clara, CA) is a horizontal retort which has a unique feature of horizon-tal water flow across the trays carrying pouches. Short come-up times, high heat transfer rates and energy conservation were reported to be special -10-advantages of this retort (Beverly, 1979; Strasser, 1979). The Stock Rotomat (Stock America, Milwaukee, WI) is another example of a retort operating on the water cook principle with overriding pressure provided by steam. Good temperature control, heat utilization and energy savings because of re-use of the heating medium, and fewer corrosion problems due to the presence of minimal amounts of oxygen in the system, were reported to be special features of this retort (Connor, 1979). The use of steam/air mixtures is the third alternative. Pflug and Borrero (1967) found that, if proper precautions were taken to assure con-stant flow of the heating medium to preclude dead spots, steam/air proces-ses could be conveniently used. Proper circulation of the heating medium could be achieved either by a positive flow system (Pflug and Borrero, 1967) or by using a powerful turbo fan as in Lagarde steam/air retorts (J. Lagarde, Montelimar, France). This method seems to overcome some disad-vantages associated the water/air cooks. Yamano and Komatsu (1969) and Yamano et al. (1969a, 1969b, 1975) extensively studied the use of steam/air heating media, and provided a basis for the predominant use of this method in Japan (Lampi, 1977), where conventional horizontal retorts, after piping modifications, have been found suitable for steam/air processing (Lampi, 1979). It has been reported that there are over 300 Lagarde 'air-over-steam' horizontal Turbo Cookers in Europe (Milleville, 1980); however, only a small number of these are used in retort pouch processing. Rexham's Hydrolok Sterilizer (Rexham Corp., Sarasota, FL), a continuous horizontal retort, also operates on the steam/air principle with media circulation being achieved by a fan (Lampi, 1979). Temperature and pressure stability were reported to be ±0.3 C° and ±17 kPa, respectively, with operating temperatures of 132°C at overriding air pressures of 48 kPa (gauge). Pflug et al. (1963), Pflug (1964) and Pflug and Borrero (1967), using both laboratory and commercial size retorts, made a comparative study of steam, steam/air and water/air as processing media for pouches, and reported that none of these media produced cold spots. The steam/air and water/air media were associated with larger come-up times compared to pure steam (Pflug, 1964). In comparing water/air to steam/air, Pflug (1964) re-ported that if the percentage of steam in the steam/air mixture was more than 85%, the steam/air was a better heating medium, if less than 85%, wa-ter was better. Kopetz et al. (1979) reported 5-15% increases in process times while using steam/air media containing 75% steam compared to water at varying flow rates. They concluded that water was more effective than steam/air in heating pouches and other smaller containers. Beverly (1980) reported longer heating times and lower productivity for horizontal retorts using steam/air mixtures compared to those using water/air heating media. These comparisons were, however, based on using a commercial water/air retort and a laboratory model retort containing steam and air. Tsutsumi (1979c) suggested the use of water/air systems to achieve rapid and uni-form heat transfer rates when processing at 120°C or below, or when the pouches contained large amounts of air. Yamano (1976) investigated heat processing of retort pouches in laboratory and industrial scale steam/air retorts and recommended several conditions for practical use of steam/air retorts. Adams and Peterson (1982) compared the surface heat transfer coefficients of steam/air media (80% steam content) and pressurized water at varying flow rates and found 20% higher values to be associated with the former. However, they reported that these differences did not influence the process times to achieve a given lethality (FQ=6.0 min) for tuna products in institutional size retort pouches (pouch thickness, 2.5-3.8 cm). -12-These studies, in general, suggest that all three processing media could be employed for retort pouches, and the choice seems, at this stage, to be a matter of individual preference. Medium Composition This aspect is relevant only to steam/air processing because the other two systems make use of either "pure" steam or "pure" water as heating me-dia. Heat transfer coefficients associated with steam are known to be quite large (Abdul-Hadi, 1979; Ball and Olson, 1957; Othmer, 1929). Hence, when considering the heat transfer during thermal processing of foods in steam, the surface resistance to heat transfer is often neglected and the container surface is assumed to follow the retort temperature instantaneously. In steam/air mixtures, air included in the medium could result in an insulating effect at the package surface, thereby reducing the rate of heat transfer. Othmer (1929) found surface heat transfer coefficient (h) values of 15,900, 7,949, 5,224, 4,145 and 2,725 W/m2C at air contents of 0, 1.07, 1.96, 2.98 and 4.53%, respectively, with the medium at 110°C and the test surface at 114.4°C. Abdul-Hadi (1979) also observed the surface heat transfer coef-ficients of steam condensing on copper, aluminum and nickel surfaces to re-duce considerably in the presence of trace amounts of air. Pflug and Borrero (1967) calculated the h values for steam/air mixtures at 115.6°C for a Reynolds number of 2,100, using the equation of Kusak (1958) as follows: 1,385, 994, 528 and 290 W/m2C at 90, 80, 60 and 40% steam content, re-spectively. Blaisdell (1963) found, using copper transducers, h values of steam/air mixtures at 74-91°C (165-195°F) to range from 193-256 W/m2C at 36% to 414-443 W/m2C at 51% and 732-874 W/m2C at 70.5% steam content. -13-Pflug (1964) compared steam/air media at 100, 90, and 75% steam content using heat penetration into cylindrical aluminum and copper transducers of size 303 x 406 (8.10 cm diameter, 11.1 cm length), and reported h values of 5,053, 1,306 and 852 W/m2C, respectively. Yamano (1976), using a 40% bentonite suspension in pouches, found the h value at 55% steam content (347 W/m2C) to be approximately 25% lower than those at 65-90% (473 W/m2C). Yamano (1976) also found that the f values for food-simulating materials in flexible packages to be almost independent of composition of the steam/air medium at steam contents higher than 70%. Recently, Adams and Peterson (1982) processed bentonite suspensions in institutional size retort pouches using 80% steam and calculated h values of 270-320 W/m2C. Medium Temperature Temperature of the heating medium and the temperature difference (AT) between the heating medium and the test surface have been identified as factors influencing the rate of heat transfer (Coulson and Richardson, 1977; Hebbard and Badger, 1934; Kern, 1950; Kisaalita, 1981; Merrill, 1948). While heating 312 x 708 (9.53 cm diameter, 19.1 cm length) lead cans from ambient temperature, Merrill (1948) found h values of pure steam to be 3,520, 4,429 and 4,997 W/m2C at 104.4, 115.6 and 121.1°C, respectively. Merrill (1948) also found, using the same system, h values for water to range from 852 W/m2C at 65.6°C to 1,079 W/m2C at 104.4°C and 1,192 W/m2C at 121.1°C. Blaisdell (1963) reported h values of 1,442 and 1,703 W/m2C, while heating 303 x 406 copper cylinders in water from 15.6°C to 48.9 and 82.2°C, respectively. Coulson and Richardson (1977) listed average values of h for pure steam condensing on horizontal tubes to range -14-from 10,000-28,000 W/m2C at a AT of 1-11 C°. Kisaalita (1981) found h values of steam/air mixtures at 105-125°C and 50-100% steam content to be related exponentially to the surface temperature or the temperature difference between the surface and the medium. In general, higher surface heat transfer coefficients have been found to be associated with higher medium temperatures having smaller tempera-ture gradients at the test surface. Medium Flow Rate Efficient heat transfer involving convection requires proper circulation of the heating medium to create adequate turbulence at the test surface. Kusak (1958) demonstrated that the heat transfer coefficient was dependent on the associated Reynolds number and composition of the heating medium. Pflug and Borrero (1967) noticed that while using steam/air mixtures it was important to have good circulation of the heating medium to achieve a uni-form temperature distribution and to preclude the formation of cold spots. Blaisdell (1963) reported increased heat transfer coefficient values for steam/air media flowing at faster rates (6.8-32% increase in h for an in-crease in flow rate from 0.48 to 1.8 m/s). Kopetz et al. (1979) reported an 11% increase in process times when flow rates in a water/air retort were re-duced from 75 to 25 gallons per min (gpm). Kisaalita (1981) varied the air flow rate of steam/air mixtures from 4.7 x 10"3 to 14.2 x 10"3 m3/s (calcu-lated Reynolds number of the medium varying from 4,000 to 28,000) and found the effect on the heat transfer coefficient to be nonsignificant. More recently, Adams and Peterson (1982) and Peterson and Adams (1982), stud-ying the effect of flow rates on h values in water/air and steam/air retorts, -15-found that the h value in the water/air system increased from 187 to 278 W/m2C for an increase in flow rate from 10 to 110 gpm (equivalent Reynolds numbers, 3,000-33,000), and in the steam/air system it increased from 270 to 320 W/m2C with flow rates increasing from 27 to 140 kg/h. Further, Peterson and Adams (1982) observed a 10% reduction in process times in the water/air system as the flow rate increased from 10 to 110 gpm. Medium Flow Direction Flow direction of the medium could have significant effects on the associated surface heat transfer coefficients of steam or steam/air mixtures depending on its influence in removing the condensing vapor on the test surface. Kusak (1958) found the flow direction to significantly influence the heat transfer coefficient associated with methanol vapor and air mixtures. Milleville (1980) found the medium flow downward in a vertical steam retort to be effective in reducing come-up time from 18 to 6 min. Although both vertical and horizontal retorts have been employed for retort pouch process-ing, there is not much published information available on their relative effectiveness with respect to transfer of heat. Package Orientation Two general racking systems have been employed to confine the pouch-es, in vertical and horizontal orientations, to a repeatable thickness during processing. The vertical system was the original choice with the idea of keeping the occluded air bubble away from the larger surface of the pouch (Corning, 1979). A pilot scale model of this type was reported by Davis et al. (1972). The majority of commercial retorts however, make use of the -16-horizontal racking system. In their original studies, Pflug et al. (1963) used a vertical rack made from slotted rectangular sheet metal plates sepa-rated by rod-mounted spacers. Provisions for medium circulation were made by placing the pouches in alternate slots accommodating either 0.953 or 1.91 cm thick pouches. Yamano et al. (1969a) and Yamano (1976) compared hori-zontal and vertical systems using steam/air mixtures as heating media. At 104.4°C, the temperature rise in pouches maintained in the vertical orienta-tion was slightly faster, while at 121.1°C, the temperature rise was margin-ally faster for the pouches in the horizontal orientation. From these stud-ies, they concluded that there was no significant difference in the heat transfer rate for the two orientations. Berry (1979) reported significant differences in the F Q values achieved in pouches under vertical and horizon-tal orientations. These were ascribed to the slumping effect resulting in an increased pouch thickness in the vertical orientation due to use of a clear-ance between the vertical plates larger than the thickness of the pouches in the horizontal orientation. Berry (1979) also observed that the vertical orientation offered an advantage for convection heating products. Recently, Roop and Nelson (1981) observed no statistical differences in heating times between vertical and horizontal orientations for pouches processed in conventional sterilizers. Residual Gases Lampi (1977) reported the following reasons for maintaining a low vol-ume of residual gases in retort pouches: product stability, preclusion of pouch bursting during retorting without resorting to high counterpres-sures, assurance of uniform and predictable heat transfer during retorting -17-to guarantee sterility, easier detection of spoilage, and easier cartoning and casing. Wallenberg and Jarnhall (1957) presented a table giving ratios of enclosed air volumes to package surface areas that should not be exceeded if pouch bursting was to be prevented. Davis et al. (1960) pointed out that if only water and air were contained in a pouch, the pouch expansion will be caused mainly by air pressure change within the headspace of the pouch. They also recognized that the volume changes could be predicted according to the ideal gas laws and reported that the maximum difference between the internal and external pressure of a pouch to prevent bursting was 6.67 kPa (internal pressure exceeding the external) during heating in pure steam at 100°C, and 16.0 kPa during cooling. Rubinate (1964) report-ed that conventional steam processing could be used if no more than 10 mL of residual gas were contained in a 1.91 x 11.4 x 17.8 cm package contain-ing 128 to 140 g of product. However, he recommended imposing an excess air pressure of 20-70 kPa during the entire operation. Whitaker (1971) pre-sented a detailed mathematical relationship to determine the maximum volume expansion of a given amount of head space air at several retort tempera-tures and pressures. Yamano (1976) reported that an air pressure of 20 to 30 kPa was required to protect the package integrity while processing at 100-120°C (package size, 16.5 x 20.0 cm; capacity, 250 mL food, 50 mL air). Comparing these with the temperature/pressure relationships and heat transfer rates for steam/air mixtures, Yamano (1976) reported that the de-sirable steam content for retort pouch processing was higher than 70%. Kopetz et al. (1979) found a 35% increase in process times when 150 mL of residual air remained in an institutional size pouch placed horizontally during processing. Further, they reported that an increase in the overrid-ing air pressure from 70 to 172 kPa (gauge) reduced the process times by -18-6-12%. Beverly et al. (1980) identified pouch confinement to be one of the critical factors in sterilizing institutional size pouches and reported that the presence of residual gas (250 mL per container, 30 x 43 cm in size) in an unconfined pouch resulted in a 44% increase in process time as compared to a confined pouch (with no residual gas present in the pouch, the increase was 33%). They also reported that the presence of residual air inside the pouch adversely affected the product quality. Temperature Distribution and Heat Penetration Studies Parcell (1930a, 1930b) found, while using a steam/air retort for pro-cessing of glass containers, that the main problem was the variation of temperature throughout the retort during the come-up period. For a verti-cal retort 0.91 m (3 ft) in diameter and 2.1 m (7 ft) high, he reported that 45-50 min were reguired to achieve a uniform temperature, and that forced circulation of the medium almost halved the come-up time. Pflug and Borrero (1967), after extensive investigation of different media for pouch process-ing, recommended several guidelines for operation of the retort to reduce the come-up time and to achieve adequate temperature stability. In a posi-tive flow vertical retort, they found that the average time required after the venting (retort reaching 104.4°C) to achieve a temperature at all points in the retort within 0.56 C° of the setpoint temperature ranged from 1 min for a medium of 100% steam to 4-5 min at 90%, and 7 to more than 50 min at 75% steam content with air flow rates varying from 2.4-9.4 x 10~3 m^ /s (5-20 ft 3/min). Pflug and Borrero (1967) also found similar variations in the come-up times for the water/air system. In these studies flow rate was identified as an important factor in achieving a faster come-up time and -19-better temperature stability. Yamano (1976) found temperature deviations in a laboratory steam/air retort, after come-up, of ±0.5 C° at 77-100% steam contents and ±0.6- 0.8 C° at steam contents of 59-71%, while processing at 120°C. For an industrial retort, he reported deviations of ±2.0 C° at 10 min, ±1.0 C° at 20 min and ±0.5 C° at 30 min after the come-up period. Milleville (1980) reported that changing the steam inlet from the bottom to the top of a. vertical retort considerably reduced the come-up times. Over the years, significant improvements have been achieved in the design of re-torts for both steam/air and water/air processing, and some commercial retorts that operate on come-up times as low as one minute with a tempera-ture stability of ±0.5 C° during the cook period are available today. Never-theless, factors which affect the temperature distribution under practical loading and operating conditions should be identified. Thermal process parameters, f and j have long been used to charac-terize heat penetration behavior of foods (Ball, 1923). Processing conditions have been generally determined for foods in pouches by methods proven satisfactory over decades of use for metal cans. Large scale production tri-als and marketing studies reported by Duxbury et al. (1970), Goldfarb (1970) and Lampi (1973) indicated that F Q values suitable for commercially canned products were generally adequate for pouches (Lampi, 1977). Pflug et al. (1963), Pflug (1964) and Pflug and Borrero (1967) employed f and j as criteria for their comparisons of different heating media, and concluded that these parameters were adequate to describe the heat transfer process. Yamano et al. (1969a) conducted heat penetration tests using 25% bentonite suspensions and indicated that the heating was most influenced by the con-tainer shape (cylindrical or slab type), with little effect from the heat transfer coefficient. Yamano (1976) presented data showing the effects of -20-varying pouch dimensions on the f values at 110°C. Recently, Spinak and Wiley (1982) reported that process times, for achieving a given lethality during retort pouch processing, calculated by Ball's formula method (Ball and Olson, 1957) were 46-90% longer than those obtained by the general method. They further noticed that inclusion of the lethality contributed due to the come-up period reduced the difference in process times calculated by the two methods to 13-36%, and the use of the actual cooling lag factor, in addition, further reduced the difference to less than 9%. Heat Transfer Involving Packaged Foods Theoretical Considerations Accurate prediction of transient or unsteady state temperature distri-butions in packaged foods during heating or cooling is important both in the design and optimal use of the heating or cooling process. Packaged foods can be assumed to resemble several regular shapes such as slabs, cylinders or spheres. The food in a retort pouch pack takes somewhat the form of a rectangular brick. Heat conduction eguations are generally used to obtain the transient temperatures in foods of various shapes (Ball and Olson, 1957; Charm, 1978; Smith and Nelson, 1969; Smith et al., 1967). Most studies dealing with thermal processing assume negligible resistance to heat transfer at the surface of the objects since heating is commonly done in a pure steam environment. However, as shown by a number of research-ers (Kusak, 1958; Othmer, 1929; Pflug, 1964; Yamano, 1976), the resist-ance cannot be neglected when using steam/air mixtures or water/air as heating media. Under these situations, the solution to the heat transfer -21-equations are more complicated. The equations for heat conduction under a variety of boundary conditions have been summarized by Carslaw and Jaeger (1959). Routine use of these heat transfer concepts has been facili-tated by the development of charts based on numerical solutions to the con-duction equations under several boundary conditions (Clary et al., 1971; Gurney and Lurie, 1923; Heisler, 1947; Newman, 1936; Pflug et al., 1965; Smith and Nelson, 1969). The heat transfer equations as applied to rectan-gular containers have been reviewed by Martens (1980), Monteza (1979) and Yamano (1976). Simplified equations for solving these conduction equations were presented recently by Ramaswamy et al. (1982). The basic equation describing the temperature distribution in an infi-nite plate, initially at a uniform temperature (Ti), when plunged into a me-dium at temperature, Ta, was given by Carslaw and Jaeger (1959), as fol-lows: 2 s i n &n U => , cos(6 nx/d) exp(- B n Fo) , n=l B n + sin B ncos B n where B n is the n th positive root of (1) B tan B= Bi , (2) and U is the unaccomplished temperature ratio, (Ta-T)/ (Ta-Ti), Fo is the Fourier number and Bi is the Biot number. At the center of the infinite plate, x = 0, and when Fo>0.2, the infinite summation series can be approximated by the first term of the series (Heisler, 1947) as U = R exp( -S.Fo) , (3) -22-2 sin 3 i where R= (4) &l + singicc-sgi and S = (5) A finite brick-shaped object may be thought of as being formed by the intersection of three infinite plates (Newman, 1936). If U(b) represents the temperature ratio at the center of a brick, then U(b) = U(pl).U(p2).U(p3) (6) where U(pl), U(p2), and U(p3) are the temperature ratios at the centers of the three constituent infinite plates from which the finite brick is formed. Eg. (6) can be expanded as r , S(pl) S(p2) S(p3) » - i U(b) = R(pl).R(p2).R(p3) exp ^-j — + ^ + ^ } at j (7) where 2a, 2b and 2c are the length, width and thickness of the brick. When the center temperature of the brick is plotted against time as log(Ta-T) vs. t, the equation for the straight line portion of the curve can be written as log(Ta-T) = log[j (Ta-Ti)] - t/f (8) where f is the negative reciprocal slope of the straight line portion of the curve and j is the lag factor defined by j = (Ta- Tpi)/(Ta-Ti). Tpi is the - 2 3 -pseudo-initial temperature obtained from the intercept of the straight line portion of the curve extended to the beginning of the heating process. By rearrangement, Eg. (8) can be written as (Ta-T ) -(t/f) = j .10 =3 .exp (-2.303 t/f ) (9) (Ta-Ti) Recognizing that the temperature ratio in Eq.(9) is the same as U(b) in Eg. (7), and comparing the two equations, it is evident that j = R(pl).R(p2).R(p3) (10) and that the arguments of the exponential functions can be equated, thus the final equation can be written as 2.303 f = (11) . S(pl) S(p2) S(p3), 01 * =2 52 c2 j a' Eg. (11) indicates that for a given material of specified dimensions, f is inversely proportional to the sum of the S/d2 values (d, being the significant dimension) which are dependent on the Biot number associated with the three dimensions of the brick, which in turn depend on the heat transfer coefficient [Eqs.(2) and (5) describe the Biot number - S value relationships]. Ramaswamy et al. (1982) presented simplified relationships for obtaining the R and S values for a given Biot number of specified geometry. For infinite plates, the relationships were: -24-R 0.1138 arctan(Bi)+0.1111 arctan(Bi/3) -0.05142 arctan(Bi/7)+1.0016 (12) S 2.0738 Bi/(Bi+2)+0.2795 arctan(Bi/3) 0.02915 arctan(5Bi)+0.001171 (13) Thermal Properties of Foods Data on the thermal conductivity and thermal diffusivity of the food material are essential in order to use the above heat transfer equations. When the thermal processing situation involves minimal surface resistance to heat transfer, thermal diffusivity of the food material is the predominant thermal property of interest. For this reason greater attention has been given, in recent years, toward establishing reliable methods for determining thermal diffusivity of foods as a function of both composition and tempera-ture. Techniques for measurement of thermal diffusivity of foods have been reported by a number of researchers (Albin et al., 1979; Awbery and Griffiths, 1933; Ball and Olson, 1957; Bhowmik and Hayakawa, 1979; Dickerson, 1965; Gaffney et al., 1980; Harmathy, 1971; Hayakawa, 1972; Hayakawa and Bakal, 1973; Martens, 1980; Mellor, 1979; Nix et al., 1967; Olson and Jackson, 1942; Ramaswamy and Tung, 1982; Thompson, 1919; Wadswortho and Spadaro, 1969). Temperature dependence of thermal diffu-sivity of foods has also been reported in recent years (Choi and Okos, 1980; Heldman and Singh, 1981; Mellor, 1976, 1977, 1979; Ramaswamy and Tung, 1981, 1982). Thermal diffusivity values for some food materials have been reported by other researchers as well (Annamma and Rao, 1974; Bennett et al., 1969; Dickerson, 1968; Ford and Bilanski, 1969; Frechette -25-and Zahradnik, 1968; Gane, 1936; Kostaropoulos et al., 1975; Mohsenin, 1980; Parker and Stout, 1967; Rao et al., 1975; Singh, 1982; Suter et al., 1975; Wratten et al., 1969). Even so, the available information in this area is rather limited and often unreliable due to the nature of the testing method or lack of details regarding the product composition and tempera-ture. Extensive reviews on the other thermal properties of foods, thermal conductivity and specific heat, are available in the literature (Dickerson, 1968; Lentz, 1961; Lentz and van den Berg, 1977; Mohsenin, 1980; Polley et al., 1980; Qashuo et al., 1970, 1972; Woodams and Nowrey, 1968). -26-EXPERIMENTAL Surface Heat Transfer Coefficients of Steam/Air Mixtures Estimation Procedure Estimation of surface heat transfer coefficients of steam/air mixtures was based on data for the heat conduction in a test brick subjected to con-vective heat transfer at the surface. This concept was discussed earlier and an equation [Eq.(ll)] relating the heating rate index to surface heat trans-fer coefficient was derived. Because of the nature of the relationship, Eq.(ll) had to be solved for the surface heat transfer coefficient by an iterative procedure. A Quick Release System Essential requirements for the preceding mathematical analysis were a uniform initial temperature in the test brick and instantaneous exposure of the brick to the heating medium. Both conditions were difficult to achieve in a batch-type steam/air retort. A number of minutes (come-up time) would normally be required to achieve a stable steam/air condition in the retort and the temperature of a conductive test brick would follow the retort tem-perature rise during this come-up period. To overcome this problem, a guick release system was designed (Tung, 1980). The system consisted of an insulated box (Figure 1), approximately 20 x 30 x 35 cm outside dimen-sions made of 5 cm R5 styrofoam insulation material sandwiched between two layers of 1.2 cm plywood, to contain the test brick, within the retort dur-ing the come-up period, in a central chamber measuring 5 x 15 x 20 cm. -27-Figure 1. An Inverted view of the insulated box the test brick into the heating medium. for the quick release of -28-When the desired steam/air condition was achieved, the box was opened by activating a latch on the spring loaded door using a cable extending through a packing gland in the retort wall. The small volume of air (approx. 1000 cm3) within the box chamber was assumed to dissipate quickly upon releasing the test brick so that it would have a minimal effect on the composition of the steam/air mixture. Figures 2 & 3 show the positioning of the box inside a pilot scale positive flow vertical retort and a forced circulation horizontal Lagarde retort, respectively, with the brick released into the heating medium. The cord suspending the brick was accessible through a door at the top of the box in order to pull the brick back into its confined position after the test run. Test Bricks Aluminum (Alcan 1000F) and stainless steel (AISI 317) bricks of the following overall dimensions with centrally located teflon-insulated 24 AWG copper/constantan thermocouples (Omega Engineering, Inc., Stamford, CT) were used in the study: aluminum, 1.93 x 12.1 x 17.8 cm (thin) and 3.88 x 12.1 x 17.8 cm (thick); stainless steel, 1.46 x 12.0 x 17.8 cm (thin) and 2.58 x 12.0 x 17.8 cm (thick). The construction details of the test bricks are shown in Figure 4. The thermocouple, with its soldered tip centrally located, was sandwiched between two plates held together tightly by four threaded bolts. A thin coating of silicone/epoxy resin adhesive was used to secure the thermocouple in the shallow channel from the edge to the center of the brick in order to prevent steam or water leakage along the thermo-couple wire. The following thermal property data for aluminum were obtained from a handbook (Smithells, 1962): thermal conductivity = 239 - 2 9 -Me tol s c r e e n Ve n t m a n i f o l d R e l e a se c a b l e S t e a m s p r e a d e r Figure 2. A schematic representation of the quick release system positioned in the positive flow retort. -30 Figure 3. A schematic representation of the quick release system positioned in the Lagarde retort. -31-Figure 4 . An exploded view of the test brick showing fabrication details. -32-W/mC, specific heat = 938 J/kgC, and density = 2,700 kg/m3. The thermal diffusivity calculated from the above data was 944 x 10"? m2/s. The thermal conductivity of the stainless steel was obtained from another handbook (Metals Hand Book, 1980) as 16.2 W/mC at 100°C, and the density was 7,590 kg/m3 as calculated from the mass and volume of the test brick. The specific heat capacity of stainless steel was experimentally determined using a Perkin-Elmer DSC-2C Differential Scanning Calorimeter (Perkin-Elmer Corp., Norwalk, CT) as 448 J/kgC at 100°C. The calculated thermal diffusivity was 47.6 x 10"? m2/s. All thermo-physical property values were obtained at 100°C which was considered an approximate average temperature of the test bricks during the test runs. Testing Procedure For each test run, the environment temperature around the test brick was monitored using five 24 AWG teflon-insulated copper/constantan thermo-couples attached to .a Kaye Ramp II Scanner/Processor (Kaye Instruments Inc., Bedford, MA). All thermocouples were precalibrated against an ASTM mercury-in-glass thermometer and appropriate corrections were made to the gathered temperature data. The fractional steam content of the mixture was calculated as the quotient of saturated steam pressure (absolute) corres-ponding to the retort temperature, and the total retort pressure (absolute) due to steam and air. The time-temperature data at the center of the test brick after the instantaneous drop from the quick release system were recorded using a Digitec data logger (United Systems Corp., Dayton, OH), at one second intervals. The temperature difference between the medium and the center of the test brick (Ta-T) was plotted on a logarithmic ordinate against time on a -33-linear abscissa using the video display of an Apple II microcomputer. The heating rate index, f, was evaluated by computing the least squares linear fit to the straight line portion of the curve. This f value was then used to determine the heat transfer coefficient. An iterative program for the micro-computer, schematically represented in Figure 5, was developed for this purpose. A first estimate of h was made using the lumped capacity method (Kreith, 1973) assuming the overall Biot number to be less than 0.1: h = 2.303 L.k/( a f ) (12) where L is the characteristic length associated with the brick, obtained by dividing the volume by surface area. Individual Biot numbers for the three dimensions of the brick were then computed and the first roots of the characteristic equations were evaluated through a subroutine. An estimate of the heating rate index, f , was computed and compared with the experi-mental f value. The h values were then incremented or decremented through a loop until the difference between the computed and the experimental f value was small. In the program this difference was chosen to be equal to or less than 0.5% of the experimental value. This computer program was later modified (Appendix I) to employ the simplified relationships developed earlier (Ramaswamy, et al., 1982) for obtaining the roots of the character-istic equations, in order to reduce the computation time. Convergence of the iterative program was based on reducing the difference between the experimental and calculated f values (set to 0.5% of the experimental f value in the original program) to provide the heat transfer coefficient value with a precision of ± 5 W/m2C. -34-/ R E A D / / D A T A / \ L U M P / 2.303kL 5 C O M P U T E Bi. N o t . ftton(/3)-Bi>- ^ R C o o T ) C O M P U T E Figure 5. Flow chart of the iterative computer program for the evaluation of surface heat transfer coefficient. - 3 5 -Factors Affecting Surface Heat Transfer Coefficient The Positive Flow Retort Experiments were carried out in a vertical positive flow steam/air re-tort similar in operation to the one employed by Pflug and Borrero, (1967). The construction details of the retort, schematically represented in Figure 6, have been reported elsewhere (Tung, 1980). The system involved con-tinuous addition of air at a constant flow rate to the retort. Steam was added to the air inlet line through a proportional valve that was controlled to maintain the temperature setpoint. The second setpoint was the maximum pressure in the retort which was regulated through a proportional valve on the vent line. The mixed steam and air were introduced through a cross-spreader either at the bottom or top of the retort with venting through holes in a ring-shaped manifold located at the upper or lower periphery within the cylindrical shell, respectively. The cross-spreader on the top was replaced by a U-shaped spreader around the guick-release system to avoid flow interference by the box. Thus, an homogeneous steam/air environment was created with the general flow direction either upward or downward as desired. This enabled the study of the effect of medium flow direction on the surface heat transfer coefficient. The medium flow rate was controlled by adjusting the air flow rate through a calibrated rotameter at a head pressure of 414 kPa (gauge). The flow rates were expressed in stan-dard cubic feet per minute (scfm), a unit equivalent to 4.7 x 10"^  m3/s at standard conditions of temperature and pressure. The majority of the experiments were carried out at a flow rate of 40 scfm, which was approxi-mately equivalent to five complete changes of the retort environment per minute. In order to study the effect of flow rate on the surface heat trans-fer coefficient, using a fractional factorial design, flow rates of 40 and 60 -36-Legend S Handvalve S Pneumatic control valve V\ Checkvalve Q Pressure regulator T T Filter Taylor fulscope tempera ture and pressure cont ro l le r . Water Rotameter Figure 6. A schematic representation of the construction details of the positive flow retort. -37-scfm were selected. These values were chosen because lower flow rates led to inaccuracies of flow measurement at higher steam contents and higher flow rates could not be achieved at lower steam contents due to venting rate limitations. However, the flow rates were varied from 20 to 75 scfm at selected conditions in order to further investigate the flow rate effects. The Lagarde Retort Experiments were carried out in a pilot scale single car Lagarde retort (J. Lagarde, Montelimar, France), 1.28 m long and 1.12 m in diameter, lo-cated at the Research Station, Agriculture Canada, Kentville, Nova Scotia. A 3 kW (4 HP) turbo fan, located at the back end of the retort, kept the heating medium in constant circulation throughout the retort to provide for uniform distribution of the heating medium in the retort during the come-up and cook cycles. The desired steam/air condition was established in the retort by controlling the medium temperature using steam and total pressure by air introduction or venting through pneumatic-action valves. The retort could be operated manually using the control switchboard located on the front panel, or could be programmed for fully automatic operation, with the desired conditions pre-punched on an aluminum program card. The medium flow pattern (horizontal) and flow rates were characteristic of the retort design. Steam/air composition and temperature of the heating medium were studied as process variables. Experiments were conducted at two different time periods approximately one year apart (summers of 1981 and 1982). Provisions were made in the quick release box to drop the test brick into the medium in one of the three different orientations shown in Figure 7: positions (i) and (ii) indicating vertical orientations of the test brick with its larger face parallel and perpendicular, respectively, to the direction Figure 7. Three orientations of the test brick in the Lagarde retort. -39-of medium flow (which is horizontal) and position (iii) indicating a horizontal orientation for the test brick. Experimental Design A four level fractional factorial design, developed by Taguchi (1957), was employed initially to evaluate the factors influencing the heat transfer coefficient in the positive flow retort. Steam/air composition and temperature were employed at four levels, and medium flow rate and flow direction were studied at two levels (Table 1). Several additional experiments were carried out beyond those required for the fractional factorial design to be used for further analyses. Steam/air composition (40-100%), temperature (105-130°C) and brick orientation (Figure 7) were the factors studied while employing the Lagarde retort. The experiments were designed to study the relation-ship between the surface heat transfer coefficient and steam content of the media under various conditions. Regression techniques were used to identify the relationship between the two variables, while the influences of different factors on the above relationships were analyzed using a covariance test (Snedecor, 1965). Table 1. Factors investigated in the fractional factorial experiments for the positive flow retort. Factor Levels Composition ( % steam ) 50, 65, 85, 100 Temperature ( °C ) 105,110,120,125 Flow rate ( scfm ) 40, 60 Flow direction upward, downward -40-Temperature Distribution in the Steam/Air Retorts Data Acquisition and Analyses Placement of Thermocouples Twenty to thirty precalibrated teflon-insulated 24 AWG copper/constan-tan thermocouples were placed at different locations within the retort to study the uniformity of the heating medium temperature with respect to thermocouple location and heating time. The thermocouple tips were secured at several probable cold spot locations within the racking systems, and in the surrounding volume of the retort. Two types of racks were employed to place the pouch packs in horizontal or vertical orientations. Horizontal Rack The racking system for test packages in horizontal orientation was similar in design to the one recommended by Toyo Seikan Kaisha, Ltd., Japan, as shown in Figure 8. The oblong hole at the center of each aluminum plate was to provide vertical channels for the medium circulation. The other holes in the plates were 2.22 cm diameter drilled on 2.86 cm offset centers resulting in 57% open space in the region where pouches were located. The angle stock was positioned to prevent lateral movement of the pouches during processing and cooling operations, and to control the vertical separation of the plates to 2.5 cm when they were stacked together. The plates were loaded onto an aluminum frame for placing in the retort. -41-O | ogogo O o ° o ° O o ° o ° ! r i 0 o ° o 0  uogogo ogogo 1° 0 1 ogogo ogogo ogogo ogogo ogogo 1 o o - a a-to Figure 8. A diagram showing the constructional details of the horizontal racking system for retort pouches. -42-Vertical Rack Test packages were held upright using a small retort rack similar to those used by Continental Can Co., Chicago, IL (Davis et al., 1972). The rack consisted of a series of stainless steel plates (3.18 mm thick) spaced to hold 2.4 cm thick pouches in one orientation and 1.3 cm thick pouches when inverted (Figure 9). The rack design provided channels for alternat-ing circulation of the heating medium. The temperature data from all the thermocouples were logged at one minute intervals using a Kaye Ramp II Scanner/Processor or a Doric Digitrend 235 (Doric Scientific, San Diego, CA). For a majority of the test runs, data were recorded on magnetic tape using a Columbia 300C Digital Cartridge Recorder (Columbia Data Products Inc., Columbia, MD) connected to the auxiliary I/O board of the data logger through an RS232 cable. The data were then transferred directly to the University of B.C. Amdahl 480/V8 computer for analysis, thus eliminating manual handling of the data. When the recorder facility was not available, data from the paper printout were keyed into the computer manually. A computer program (Appendix II) was developed [based on the ori-ginal program by Tung (1974)] for use with the Amdahl computer, for cal-culating the mean and standard deviation of temperature at the several thermocouple locations within the retort at each time interval. At the end of the heating period, the means and standard deviations of temperature dur-ing the cook period (excluding the come-up time) were computed for each thermocouple location, together with an overall mean temperature and overall standard deviation for the above time period including all thermocouple locations. Thus the temperature uniformity with respect to location and time could be studied for a given processing condition. - 43 -Figure 9. A diagram showing the constructional details of the vertical racking system for retort pouches. -44-Factors Affecting Temperature Distribution The Positive Flow Retort An L-27 three-level five-factor fractional factorial design (Taguchi, 1957) was employed for studying the effects of medium composition (steam contents: 50, 65 and 85%), temperature (105, 110, 120°C), flow rate (20 and 40 scfm) and flow direction (upward and downward), and rack type (horizontal and vertical) on the temperature distribution of the heating medium. The basis of comparison was the overall standard deviation in tem-perature during the cook period (excluding the come-up time). A number of experiments were carried out beyond the factorial design to further inves-tigate these effects. Since the positive flow retort was small [0.6 m in diameter and 0.6 m high (2 ft x 2 ft)], only one rack, providing one package orientation, could be accommmodated at a time. The Lagarde Retort The flow rate and flow direction of the heating medium in the Lagarde retort were characteristic of the retort design and hence could not be varied. The retort temperature and the medium composition were the only variables that could be tested for their influence on temperature distribu-tion. The racking systems developed for the positive flow system were too small for studying orientation effects on the temperature distribution in the Lagarde retort, and therefore were not considered. A full factorial design of experiments involving two levels of temperature (105 and 120°C) and four levels of steam content (50, 65, 85 and 100%) with two replications was employed in this study. -45-Pressure Stability in the Steam/Air Retorts A pressure transducer (Model A-5/1148, Sensotec, Columbus, OH) was used to monitor the retort pressure during the cook and cool cycles of the processing operation. The transducer was installed in the positive flow retort through a port in the retort wall or was mounted on the top of the Lagarde retort through a bleeder port. A constant excitation voltage of 10 V (DC) was provided using a Hewlett-Packard power supply (Model #62148, Hewlett-Packard Co., Rockaway, NJ). The millivolt signals (0-50 mV) from the transducer were recorded at one minute intervals using the Digitec data logger. The pressure transducer was calibrated using a standard dead weight tester (Chandler Engineering Co., Tulsa, OK) over a range of 0 to 350 kPa (gauge) at approximately 35 kPa intervals. This range covered the pressures employed in the retorts. The pressure variations in the retorts were studied by computing the standard deviations in the retort pressure during the cook period. The effects of various influential parameters were studied by using an L-16 fractional factorial design (Taguchi, 1957) for the positive flow retort and a 4 x 2 full factorial design for the Lagarde retort. The process parameters were temperature, steam content, flow rate and flow direction for the posi-tive flow retort, and temperature and steam content for the Lagarde retort. Since the total operating pressures varied considerably for the different conditions, the coefficients of variation (standard deviation in pressure divided by the mean operating pressure, expressed in percentage, during the cook period) were also considered for identifying the potentially important factors. -46-Heat Penetration Studies In the Steam/Air Retorts Fabrication of Test Bricks Test bricks constructed using Dow Corning RTv* 3110 silicone rubber (Dow Corning Corp., Midland, MI) and formed nylon were employed to gather heat penetration data from the steam/air heating media under various conditions. The silicone rubber bricks, with centrally located teflon-insu-lated 24 AWG copper/constantan thermocouples, were formed in molds con-structed of aluminum and glass to give an overall dimension of 1.9 x 12.1 x 17.8 cm. The rubber compound was mixed well with a catalyst in the proportion of 20:1, poured into the mold, and left undisturbed for 3 days during setting. The nylon bricks, with the overall dimensions of 2.1 x 12.1 x 17.8 cm (thin, Nyl-tn) and 2.4 x 12.1 x 17.8 cm (thick, Nyl-tk), also with centrally located 24 AWG copper/constantan thermocouples, were laminations of two rigid nylon slabs of equal thickness secured by means of four threaded bolts. The thermal diffusivities of these test bricks were determined through heat penetration tests in a high steam (>95%) environ-ment, assuming negligible surface resistance to heat transfer, using a modified form of E q . ( l l ) : 0.933 a (13) -47-Factors Affecting Heat Penetration All the factors considered earlier in the temperature distribution studies of steam/air heating media, were treated as potential factors influ-encing the heat penetration rates into test bricks. Experiments were carried out simultaneously for evaluating temperature distribution of the heating medium and heat penetration into test bricks of silicone rubber and nylon. Hence, the design of experiments for this part of the study was essentially the same as the one described earlier for temperature distribution. The heating rate index, f, was used as the basis of comparison to study the influence of different factors on the heat transfer rates. The procedure for evaluating f and j, using an Apple II microcomputer, was described earlier in the testing procedure for evaluating surface heat transfer coefficients. The j values were computed at 42% effectiveness (Ball, 1923) for the come-up period. The test bricks of nylon and silicone rubber were used without pack-aging them in retort pouches, in order to reduce experimental variabilities due to added resistances to heat transfer, while studying the effects of composition, temperature, flow rate and flow direction of the heating medium, and package orientation. In a further experiment, silicone rubber bricks were vacuum-packed in retort pouches (American Can Co., Neenah, WI; polyester/alumium foil/polypropylene) using a vacuum sealing machine (Multivac Model AG5, Algau, W. Germany) to determine the effect of the packaging material and residual air contained in the package on the heat transfer rates. Following the experiments, residual air contents of the pouches were determined by a destructive method (Shappee and Werkowski, 1972) and found to vary from 15 to 30 mL per pouch. The pouches were - 4 8 -equipped with packing glands (O.F. Ecklund, Ltd., Cape Coral, FL) to provide for hermetic entry of the thermocouple leading to the center of the silicone rubber brick. - 4 9 -RESULTS AND DISCUSSION Surface Heat Transfer Coefficients of Steam/Air Mixtures Estimation Procedure The quick release system was effective for introduction of the test brick into the heating medium stabilized at the desired condition. Although no heat flow measurements were made to determine the ability of the insulated box to maintain temperature uniformity in the test brick, continu-ous scanning of the temperature at the center of the brick prior to the actual drop provided a means to check the effectiveness of the box insula-tion. The initial temperature of the test brick was below 60°C at the time of release into the heating medium for all test runs. The actual initial tem-perature was dependent on the order of the test run, being lowest for the first run in a series, due to the gradual absorption by the box of heat which was not completely removed during cooling prior to the next run. In general, the temperature rise in the test brick after the initial equilibration with the chamber environment, during the come-up period (3-5 min), was less than 10 C°. Although initial temperature uniformity was an essential condition in Eq.(l), the final equation employed for the evaluation of sur-face heat transfer coefficient [ Eq.(ll) ] was not dependent on the initial temperature distribution. The results, therefore, were not influenced by the initial temperature when (Ta-Ti) was sufficiently large to provide for an adequate number of time-temperature data pairs for an accurate evaluation of f. -50-Typical surface heat transfer coefficients as evaluated by the procedure outlined, are given in Table 2. The values estimated by the lumped capacity method were considerably lower than those calculated using Eq.(ll). The lumped capacity method was based on the assumption that the internal resistance to heat transfer was negligible when compared to the surface resistance. A Biot number of less than 0.1 is usually taken to represent these situations (Heldman and Singh, 1981). For the aluminum test brick the overall Biot number was below 0.1 when the associated heat transfer coefficient was less than about 3000 W/m2c. The difference between the h values calculated by the two methods was approximately 5% at a surface heat transfer coefficient of about 1000 W/m2C, 10% at 3000 W/m2C and 25% at 10,000 W/m2C (Table 2). These results support the observation made earlier (Ramaswamy et al., 1982) regarding the errors associated with the lumped capacity method. Further, the evaluation of surface heat transfer coefficients involved a double iterative procedure (Figure 5), one in the main program adjusting the heat transfer coefficient value and another in the subroutine ROOT for the calculation of the roots (three times) of the characteristic transcendental function [Eq.(2)] for each main iteration. The computation time increased significantly as the associated heat transfer coefficient values became larger. The simplified relationships developed by Ramaswamy et al. (1982), employed in the computer program for evaluating the first root of Eq.(2) [Appendix I], resulted in substantial savings in computation time (Typical computer output of the program is given in Appendix III). -51-Table 2. Typical heat transfer coefficient results for steam/air mixtures in the positive flow retort Surface heat transfer coefficient Temperature Steam f content Lumped Calculated (°C) (%) (s) (W/m2C) (W/m2C) 105.2 50.3 36.26 1220 1270 115.3 49.6 35.92 1240 1280 121.1 51.0 35.25 1260 1310 125.8 53.2 31.85 1390 1460 105.4 65.3 19.64 2260 2420 114.7 64.4 21.42 2070 2210 120.0 64.7 21.53 2060 2210 125.1 67.3 19.26 2300 2470 105.4 85.4 10.22 4340 4950 114.6 81.5 10.18 4360 4970 120.0 84.9 10.47 4240 4810 125.9 87.6 10.53 4210 4780 104.5 95.2 5.22 8500 10870 115.1 97.6 5.18 8570 10960 119.9 96.9 4.84 9170 11940 125.2 98.2 5.20 8530 10910 1 Medium flow upward at 40 scfm; thin aluminum brick It is difficult to make comparisons of the results from this study with those that have been published. The main reason for this is the use of steady state methods in the majority of previous studies. The values at high steam contents (>95%) were in reasonable agreement with those report-ed by Ball and Olson (1957), Coulson and Richardson (1977) and Othmer (1929). For steam/air mixtures, the coefficients reported by Adams and Peterson (1982), Blaisdell (1963), Pflug (1964) and Yamano (1976) were -52-considerably lower as compared to the results of this study. Blaisdell (1963) and Pflug (1964) were limited to conducting the experiments at atmospheric pressure due to problems in achieving stable retort conditions with come-up times of less than a minute. Pflug (1964) reported values of 5,053, 1,306 and 852 W/m2C at 100, 90 and 75% steam content, respectively, and also cited heat transfer coefficient values obtained by Blaisdell (1963) to range from 193-256 W/m2C at 36% steam to 414-443 W/m2C at 51% steam and 732-874 W/m2C at 70.5% steam content. The temperature of the copper and aluminum transducers employed by Blaisdell (1963) and Pflug (1964) would follow the temperature rise in the retort, during the come-up period, due to high con-ductivities of the test materials. The gradual increase in the retort temperature during the come-up period influences the temperature differ-ence, Ta-T, for a highly conductive test material at each time interval and the resulting f will, therefore, be larger than that for situations involving an instantaneous immersion of the test material into a prestabilized heating medium, unless evaluated far away from the come-up period. The above influence was almost negligible (temperature increase in the test brick less than 10 C°) in the present investigation because the test brick was con-tained in an insulated box which prevented contact with the heating medium during the come-up period. This probably explains why the surface heat transfer coefficient values obtained by Blaisdell (1963) and Pflug (1964) were considerably lower than the values from this study. Yamano (1976) reported values of 473 W/m2C at 65-95% steam and 347 W/m2C at 55% steam content, while Adams and Peterson (1982) reported values of 270-320 W/m2C at 80% steam content. Both of these studies were conducted using bentonite suspensions as test materials. -53-It is extremely difficult to evaluate the heat transfer coefficients precisely using test materials of low thermal conductivity, such as bentonite suspensions, especially when the associated heat transfer coefficients are large. For example, for a material with a thermal diffusivity of 1.5 x 10"? m2/s, and a thermal conductivity of 0.5 W/mC (both being representative thermal properties of food and food-simulating materials like bentonite sus-pensions), and with package dimensions of 2.0 x 12.0 x 18.0 cm, the f value at an infinitely large surface heat transfer coefficient could be cal-culated using Eg. (11) by substituting S(pl) = S(p2) = S(p3) = 2.467. The resulting f was found to be 9.97 min. A variability of 3.7 to 5.7% was found to be associated with the experimental evaluation of f from the heat pene-tration studies reported by Pflug (1964) and Pflug and Borrero (1967), while a coefficient of variation as high as 21% was reported by Tung and Garland (1978) for f values of foods in retort pouches. A deviation of about 5% in the evaluation of f is, therefore, quite conservative. With an f of 10.47 min (5% higher than the calculated value) the heat transfer coeffi-cient, back-calculated using the iterative computer program, was found to be 1,940 W/m2C. Thus, a five percent difference in the evaluated f value could account for a difference in the heat transfer coefficient from infinity to 1,940 W/m2C. Presence of other resistances in the retort pouch contain-ing the test material (such as noncondensible head space gas) could further reduce the overall heat transfer coefficient, thereby masking the effective heat transfer at the surface. On the other hand, a similar analysis for an aluminum brick of similar dimensions, gives a heat transfer coefficient value of about 500,000 W/m2C at an f value 5% greater than that at infinite h. Thus, test bricks of higher conductivity are an important consideration in -54-the evaluation of heat transfer coefficients of heating media such as steam/air mixtures, particularly at higher steam contents. While one might consider the values obtained by Adams and Peterson (1982) and Yamano (1976) using low thermal conductivity test materials to more closely simulate the processing of foods in flexible pouches, the potential ability of the steam/air media to transfer heat is underestimated by those reports. Factors Affecting Surface Heat Transfer Coefficient The Positive Flow Retort Fractional factorial experiments: Analysis of variance results for the fractional factorial design described in Table 1 are presented in Table 3. At the flow rates and steam contents of the medium (flowing upward or downward) employed in the analysis, the effect of temperature on the sur-face heat transfer coefficient was not significant(p>0.05). Similar results were observed when the effect of flow rates (40 and 60 scfm) were compar-ed at different steam contents, temperatures and flow directions (upward or downward) of the heating medium (Table 3, Experiments #3 & #4). The relatively narrow range of medium flow rates (40-60 scfm) may be a reason for the nonsignificance of the flow rate effect. The medium flow direction was found to be a significant (p<0.05) factor when the effects of steam con-tent and temperature were considered at a flow rate of 40 scfm (Table 3, Experiment #5). The medium steam content, however, was found to be the major factor significantly (p<0.01) influencing the surface heat transfer at different flow rates, flow directions and temperatures of the heating medium. -55-Table 3. Analysis of variance in the surface heat transfer coeffi-cients of steam/air mixtures in the positive flow retort experiments. Factor1 Experiment F-ratio #1 #2 #3 #4 #5 Steam content 888** 572** 1213** 456** 755** Temperature 0.4ns 2.1ns 0.6ns 1.5ns 0.6ns Flow rate (scfm) (40) (60) 3.4ns 0.9ns (40) Flow direction (up) (up) (up) (down) 7.3* 1 When factor is not a variable, the level used is indicated in parentheses. * Significant at p<0.05; ** Significant at p<0.01 ns Not significant (p>0.05) Relationship between the surface heat transfer coefficient and steam content: A total of 258 test runs was carried out under varying conditions of steam/air composition, medium temperature, flow rate and flow direction, using test bricks of both aluminum and stainless steel (Appendix IV). Regression of surface heat transfer coefficients from the above test runs on the associated steam contents of the heating media indicated strong hyper-bolic, exponential and polynomial (p<0.05 up to 3rd degree) relationships between the two parameters. The exponential function gave consistently large coefficients of determination (greater than 0.96 for all test conditions shown in Appendix V), and was therefore chosen to represent the heat transfer coefficient - steam content relationship. Kisaalita (1981), Kusak (1958) and Othmer (1929) also found exponential relationships between the two parameters; however, they included other factors such as temperature of the medium, temperature difference between the medium and the test surface, flow rate of the medium, etc., in their equations. The general -56-relationship between the surface heat transfer coefficient (h) and percen-tage steam content (S) was represented as follows: h = a exp( bS ) (14) where a and b were parameters which depended on the experimental condi-tions . Comparisons of the regression equations were made by a covariance test based on pairs of lines transformed to linear functions, and the results are presented in Table 4 (further details included in Appendix VI). The covariance test first compared deviations of the two sets of data from their least squares linear fits as a test for homogeneity of residual variances (if significant, further tests were not needed). The two regression lines were then compared for slopes, levels and overall differences. The results of these analyses indicated that the variance was not significant in any of the test situations; thus useful comparisons could be made of the exponential relationships between surface heat transfer coefficient and steam content under various process conditions. Effect of medium temperature: Results using aluminum (thin) and stainless steel (thin) with the heating medium flowing upward or downward, and at flow rates of 40 or 60 scfm indicated that the exponential relation-ship between the heat transfer coefficient and steam content at 105 and 120°C were not significantly different (p>0.05) from each other (Table 4, Appendix VI). The surface heat transfer coefficient, therefore, could be assumed to be independent of temperature in this range. Accordingly, the data at different temperatures were pooled for testing the effects of other variables. Kisaalita (1981), Kusak (1958), Othmer (1929) and Merrill (1948), -57-using steady state methods, found the surface heat transfer coefficient to depend on the temperature of the medium or the temperature difference between the medium and the substrate. This apparent discrepancy may be due to differences in the nature of the testing methods. Table 4 . Surface heat transfer coefficient as a function 1 of steam content In the positive flow retort. Temperature Flow Regression results 0 o direction'0 n a b r 2 (°C) (W/m2C) (%_1) 105 U 70 153p 0.0421r 0.96 120 U 22 141p 0.0431r 0.98 105 - 125 U 101 153p 0.0421r 0.96 105 D 57 347q 0.0356s 0.93 120 D 14 26 8q 0.0375s 0.98 105 - 125 D 76 337g 0.0355s 0.93 1 h = a exp(bS) 1 U, upward ; ; D, downward ° Coefficients a and b sharing the same letter are not significantly different (p>0.05). Effect of medium flow rate: The effect of medium flow rate (40 and 60 scfm) on the exponential relationship was found to be significant with respect to only the level (Appendix VI) indicating - that the curves were essentially parallel at the two flow rates (thin aluminum brick; temperature, 105-125°C; flow direction, upward). This parallel nature of the relationship indicated that there may be no interaction between steam content and flow -58-rate. The effect of medium flow rate on the heat transfer coefficients was further investigated to cover a broader range of flow rates. The results (Figure 10), at selected steam contents for the media flowing upward and downward, indicated essentially linear relationships (Table 5) between the heat transfer coefficient and flow rate in the 20-75 scfm range. The sensitivity of the heat transfer coefficient to changing flow rate appeared to be greater for media flowing vertically downward. Blaisdell (1963) reported 6.8-32% increases in the heat transfer coefficient when the medium flow rate increased from 0.48 to 1.8 m/s. Peterson and Adams (1982) also found the heat transfer coefficient of steam/air mixtures to increase with the medium flow rate. The greater turbulence at higher flow rates could possibly explain an increased surface heat transfer coefficient at higher flow rates. Table 5. Regression details for the dependence of surface heat heat transfer coefficients on the medium flow rate in the positive flow retort. o Steam Medium n c d r° content flow (%) (W/mzC) (W/mzC scfm) 84 upward 11 5010 19.6 0.86 73 upward 5 2706 24.2 0.97 63 upward 13 1291 26.2 0.98 74 downward 6 3121 51.3 0.99 63 downward 12 1808 56.1 0.98 1 Regression equation: h = c + d(flow rate) M e d i u m flow ra te , scfm Figure 10. Surface heat transfer coefficient as a function of flow rate for steam/air mixtures in the positive flow retort [ steam content, %S; medium flow, upward, + , downward, + ]. -60-Effect of medium flow direction: Medium flow direction, upward or downward, was found to be a significant factor (p<0.05) affecting the heat transfer coefficient - steam content relationship (Table 4, Appendix VI) from all the test bricks (aluminum and stainless steel, thick and thin). In general, the surface heat transfer coefficients associated with steam/air media flowing in the downward direction were higher than those for the media flowing upward at steam contents ranging from 50- 100%. The exponen-tial slope index (b=0.0421 % _ 1 , Table 4) for the medium flowing upward was greater than that for the downward medium flow (b=0.0355 %"*)# which indicated that the heat transfer coefficient was more sensitive to changing steam contents in the upward medium flow. However, higher h values were observed for the downward flow over the entire range of steam contents (Figure 11) due to the larger intercept coefficient (337 W/m2C) for the downward flow compared with the upward flow (153 W/m2C). The qualitative similarity of the relationship while using different test bricks indicates that the effect of the medium flow direction on the surface heat transfer coefficient may be independent of the nature of the testing material. The figure also suggests a minimal interaction between the flow direction and steam content. The higher surface heat transfer coefficients associated with the medium flowing downward may be due to rapid removal of the film or droplets of condensing steam from the test surface (with the medium flow and gravity acting in the same direction), thus exposing new surfaces for condensation of the steam. A heating medium flowing upward, on the other hand, may retard the condensing film or droplets from falling freely under gravity, thereby increasing the resistance to heat transfer at the surface. The associated heat transfer coefficient, therefore, may be considerably lower. The steeper curves for the heat transfer coefficient - flow rate -61-Flgure 11. Surface heat transfer coefficient as a function of steam content of steam/air mixtures in the positive flow retort. -62-relationships (Figure 10) for media flowing downward (the effects of flow rate and gravity being additive) as compared to those flowing upward also support this observation. The Lagarde Retort A total of 182 test runs was carried out under varying conditions of steam/air composition, medium temperature and brick orientation using test bricks of both aluminum and stainless steel (Appendix VII). Regression analyses indicated a strong exponential relationship between the surface heat transfer coefficient and steam content (Appendix VIII). Comparisons of the regression relationships were made using the covariance test. Effect of medium temperature: Results using the aluminum test bricks in the vertical perpendicular orientation indicated that the exponential relationship between the surface heat transfer coefficient and steam content at 105-110°C was not different (p>0.05) from the one at 120-125°C (Table 6, and Appendix IX). The observations were similar in the positive flow retort. The heat transfer coefficient, therefore, can be assumed to be independent of temperature under the various conditions studied. Accordingly, the data at different temperatures were pooled for considering the effects of other variables. Effect of brick orientation: With the heating medium flowing in the horizontal direction, there are two main brick orientations, vertical or horizontal, with the larger surface of the brick aligned in the direction of the heating medium flow. However, when a single test brick was dropped into the medium, the effect of placing the brick in a vertically perpendicular direction was an alternative orientation. These three orientations are described as vertically parallel (referred to as parallel, -63-hereafter), vertically perpendicular (referred to as perpendicular) and horizontal (Figure 7). Table 6 . Surface heat transfer coefficient as a function 1 of steam content in the Lagarde retort. Temperature Brick Regression results 0 position*" n a b r 2 (°C) (W/m2C) (%_1) 105 - 110 120 - 125 V,pa,pe V,pa,pe 48 43 1149w 1247w 0.0208y 0.0197y 0.93 0.86 105 - 130 105 - 130 V,pa V,pe 59 34 1157w 1182w 0.0206y 0.0207y 0.91 0.94 105 H 8 1669x 0.0132z 0.94 1 h = a exp(bS) 2 V, vertical; H, horizontal pa, parallel; pe, perpendicular. Coefficients a and b sharing the same letter are not significantly different (p>0.05). The effect of brick orientation on the resulting heat transfer coefficient was studied using aluminum test bricks (thin and thick). The results of the heat penetration tests carried out in the Lagarde system in the two test periods (1981 and 1982) with the thin aluminum test brick in the parallel orientation (Appendix IX) indicated good reproducibility of the -64-results. The covariance test indicated no significant differences in the variance, slope and overall nature of the heat transfer coefficient-steam content relationship; however, the difference in levels was significant (p<0.05), probably due to the sensitive nature of the covariance test. Com-parison of the effects of parallel and perpendicular brick orientations (Table 6), indicated that the exponential relationships between the surface heat transfer coefficient and steam content were not different (p>0.05). The com-bined relationship is shown in Figure 12. The horizontal brick orientation, however, had a very different exponential relationship compared with either parallel or perpendicular orientations (Figure 12). The covariance test showed significant differences in all comparisons (Table 6). It is possible that lower surface heat transfer coefficient values could be explained by the presence of a comparatively stagnant layer of water (creating an added resistance) from the condensing steam on the upper surface of the test brick. In the other two orientations there may be a rapid removal of the film of condensing steam from the test surface due to gravity, in addition to the medium flow, thus exposing new surfaces for condensation of the steam. An Analysis of the Sources of Error High thermal conductivity of the test material was identified earlier as one of the requirements for successful evaluation of the heat transfer coefficient (Ramaswamy et al., 1983). With a given test brick, the accuracy of the heat transfer coefficient evaluation depends mainly on the accuracy with which the f value could be estimated from the the center temperature history data. With the availability of a large number of data pairs for the Steam content , % Figure 12. Surface heat transfer coefficient as a function of steam content of steam/air mixtures in the Lagarde retort. -66-temperature-time relationships for foods in retort pouches, deviations of about 3-5% have been identified in the estimated f values (Pflug, 1964). Reliable estimation of f was found to be difficult when fewer than 10 data pairs were available for computation in the linear region and variations up to ±10% were observed under these situations. Thus, rapid data acquisition was an important consideration for the evaluation of the surface heat trans-fer coefficient, particularly when using the aluminum test brick, because of its high thermal conductivity. Data logging even at one second time inter-vals during the heat penetration test was inadequate for obtaining an accurate estimate of f using the thin aluminum brick, particularly when the steam contents were higher than 85%. Alternate test bricks of greater thickness and lower thermal conductivity increased the heating time such that more reliable estimates of f could be made. The use of these alternate test bricks improved the confidence in the estimated f values; however, they also contributed to additional sources of error in terms of material variations with respect to their thermal properties, surface characteristics (roughness, wettability, etc.,), and possible differences due to thermo-couple location and contact. The increased thermal conductivity would also reduce the precision of the evaluated h values as discussed earlier. The magnitude of errors in the heat transfer coefficient evaluation computed for a given variability in f values were dependent on the thermal conductivity, size of the test brick and the magnitude of h (Table 7). For a 5% variability in the f and with h values between 1,000 - 15,000 W/m2C, the errors introduced were 3-6%, 3-8%, 3-9% and 4-15% while using test bricks of thin and thick aluminum, and thin and thick stainless steel, respectively. Larger errors were associated with higher values of h, and test bricks of lower thermal conductivity and greater thickness. The -67-standard errors of estimate were in the 8-10% range for the different regression equations. The extent of the errors, when the h values predicted from the regression equation were compared with the experimental values, were dependent on the magnitude of the heat transfer coefficient, and ranged from 4-12% (Table 7). The smaller errors at lower h values should be viewed with some caution because these were characteristic of steam/air media with lower steam contents, which had greater temperature instability and hence, larger compositional variations. Table 7. Estimated errors in the evaluation of heat transfer coefficients of steam/air mixtures while using the aluminum and stainless steel test bricks of different thickness. % Error (±) h Aluminum Stainless steel (w/m2C) Thin Thick Thin Thick Exp 1 Reg 2 Exp Reg Exp Reg Exp Reg 1000 3 6 3 8 3 8 4 4 3000 3 7 3 6 3 6 5 8 5000 5 12 5 6 5 7 7 11 7000 5 12 6 7 6 5 8 9 9000 6 11 7 6 6 6 11 11 11000 6 11 7 9 7 6 12 10 13000 6 7 8 13 12 15000 6 8 9 15 1 Experimental errors assuming a 5% variability in the evaluated f values. Prediction errors associated with the regression equations. -68-In view of the above analysis, differences observed in the heat trans-fer coefficients obtained using the different test bricks were treated as errors contributed due to uncertainties of their thermal property values and surface characteristics. The differences observed in h values due to variations in temperature (105-130°C), flow rates (40-60 scfm) and brick orientation (perpendicular and parallel) were considered relatively small compared to the experimental error magnitudes, and compositional influences. The following equations were derived to relate the surface heat trans-fer coefficient of steam/air mixtures to the steam content in the pilot scale positive flow and Lagarde retorts: Positive Flow Retort Medium flow, upward ; brick orientation, vertical: h = 153 exp(0.0421 S) ( n=101, r2=0.964) (15) Medium flow, downward ; brick orientation, vertical: h = 337 exp(0.0355 S) ( n=76, r2=0.931) (16) Lagarde Retort Medium flow, horizontal; brick orientation, vertical: h = 1011 exp(0.0226 S) ( n=174, r2=0.904) (17) Medium flow, horizontal; brick orientation, horizontal: h = 1669 exp(0.0132 S) ( n=8, r2=0.940) (18) The above studies on the surface heat transfer coefficients for steam/air mixtures in the pilot scale retorts indicated, in general, that medium composition and flow direction were factors of major importance. The -69-h value was also found to increase linearly with the medium flow rate in the positive flow retort. The effect of medium temperature was nonsignificant. While assessing the importance of these results for industrial applications, it must be recognized that these studies were carried out in pilot scale retorts under uninhibited medium flow conditions. Retort type, design of the racking system and product loading pattern could substantially influence the flow behavior of the medium and hence the associated surface heat transfer coefficients. -70-Temperature Distribution in the Steam/Air Retorts Data Analyses Typical output of the computer program for temperature distribution (employing only fifteen thermocouple locations) in a test run of the positive flow retort is given in Appendix X (steam content, 85%; temperature, 120°C; flow direction, downward; flow rate, 40 scfm; racking system, hori-zontal). From the last two columns of means and standard deviations (Mean, S.D.), the come-up time was found to be 4-5 min, and the stabilization time (when the standard deviation in temperature at different locations in the retort, as indicated in the last column becomes steady) was 5 min. The overall mean temperature of the retort during the cook period (excluding the come- up/stabilization time of 5 min) was 120.05°C and the overall stan-dard deviation was 0.15 C°. Figure 13 is a plot of temperatures monitored at different locations inside the retort during the cook period. This plot can be used to observe temperature deviations from the overall mean tem-perature (represented by the horizontal line) for any of the thermocouple locations at any particular time. The temperatures at several locations were 0.1 to 0.3 C° below the setpoint temperature of 120°C; the setting, there-fore, would have to be slightly higher if a greater assurance of maintaining the temperature at all locations during the cook period at >120°C was re-quired. Figure 14, a plot of the means and standard deviations of tempera-ture for all thermocouples at each recorded time during the cook period, was used to check the performance of the different control systems in main-taining the desired steam/air condition (a steady mean and a low standard deviation in temperature). The fluctuations during the cook period -71-122 121 L Figure 13. Temperature histories for all thermocouple locations during the cook period in a test run of the positive flow retort [steam content, 85%; temperature, 120°C; flow direction, downward; flow rate, 40 scfm; rack type, horizontall. -72-122 121 L u 0) 3 O OJ a E 120 b Figure 14. Means and standard deviations of temperature for all thermo-couple locations at each recorded time during the cook period of a test run in the positive flow retort. -73-were found to be ±0.2 C° at 120°C. Figure 15 is a plot of the mean tem-perature and its standard deviation at each thermocouple location over the entire cook period. This plot was used to identify any cold spots among the thermocouple locations developed due to inadequate circulation of the heating medium. In this particular run, thermocouple locations 6-10 and 13 had mean temperatures slightly lower than 120°C. Repeated occurrence of cold spots at some locations might warrant design modifications of the rack-ing system or distribution devices to alter the flow behavior of the media. Thus, considerable information on the uniformity of temperature distribution in a steam or steam/air retort for a given process condition could be obtained from this method of analyzing the temperature distribution and stability data. Factors Affecting Temperature Distribution The Positive Flow Retort Fractional factorial experiments: The overall standard deviations of temperature during the cook period in the positive flow retort under different steam content, temperature, flow rate and flow direction of the steam/air media, and rack type are given in Table 8 and the analyses of their variance with test factors are summarized in Table 9. Among the fac-tors studied, steam content and flow rate of the heating medium were found to be significant (p<0.05) in influencing the temperature uniformity, with the medium at lower steam contents and lower flow rates causing larger tem-perature deviations in the retort. The effects of temperature and flow direction of the heating medium, and rack type were not significant (p>0.05). No specific cold spot location was observed in any of the -74-1 2 2 1 2 1 120 M M * 4) a E 119 118 L-L. 1 5 1 0 Thermocou pie number 15 Figure 15. Means and standard deviations of temperature at each thenno-couple location over the entire cook period of a test run In the positive flow retort. -75-Table 8. Overall standard deviations* in temperature during the cook period in the positive flow retort for the fractional factorial experiments. Run Steam Temperature Flow Rack Flow Standard # content rate type^ direction deviation (%) (°C) (scfm) (C°) 1 50 105 40 H upward 0.36 2 50 105 40 V downward 0.29 3 50 105 20 H downward 0.30 4 50 110 40 V downward 0.23 5 50 110 40 H downward 0.21 6 50 110 20 H upward 0.40 7 50 120 40 H downward 0.25 8 50 120 40 H upward 0.25 9 50 120 20 V downward 0.42 10 65 105 40 H downward 0.17 11 65 105 40 V upward 0.24 12 65 105 20 H downward 0.21 13 65 110 40 V upward 0.21 14 65 110 40 H downward 0.12 15 65 110 20 H downward 0.16 16 65 120 40 H downward 0.27 17 65 120 40 H downward 0.17 18 65 120 20 V upward 0.28 19 85 105 40 H downward 0.16 20 85 105 40 V downward 0.17 21 85 105 20 H upward 0.21 22 85 110 40 V downward 0.19 23 85 110 40 H upward 0.19 24 85 110 20 H downward 0.24 25 85 120 40 H upward 0.12 26 85 120 40 H downward 0.20 27 85 120 20 V downward 0.25 } Data from 30 thermocouples and 30 min cook time 2 Rack type: V, vertical; H, horizontal -76-Table 9. Analysis of variance in temperature distribution for studies of the positive flow retort. Source of variation df MS F-ratio Steam content 2 0.0323 12.42** Temperature 2 0.0019 0.73ns Flow rate 2 0.0125 4.81* Flow direction 2 0.0012 0.46ns Rack type 2 0.0045 1.73ns Error 16 0.0026 * Significant at p<0.05 ; ** Significant at p<0.01 ns Not significant (p>0.05). above test runs indicating good circulation of the heating medium in the retort. A total of 100 test runs was carried out including the runs for the fractional factorial design. Since the effects of temperature, flow direction and rack type were nonsignificant, the results from these tests were pooled and presented for various steam contents and flow rates of the heating media (Table 10). The come-up times generally ranged from 2-6 min, averaging about 3 min for the heating media containing 50-85% steam and 4 min at 100% nominal steam content (actual steam content >97%). Stabilization was achieved 5-10 min after the entry of steam into the retort, with values on the higher side usually associated with steam/air media of lower steam contents. The temperature deviations were slightly higher at a flow rate of 20 scfm (especially at lower steam contents) as compared with those at 40 -77-scfm. The results presented in Table 10 also confirmed the association of larger standard deviations in temperature with the steam/air mixtures having lower steam contents. Pflug and Borrero (1967), using a different procedure, reported slightly higher deviations and longer come-up times in their positive flow retort which also was larger [0.91 m (3 ft) in diameter by 1.82 m (6 ft) high] than to the one employed in this study. Table 10. Summary of overall standard deviations in temperature after the come-up period in the positive flow retort. Standard deviations Steam Flow No. of Come-up content rate replicates time Range Mean (%) (scfm) (min) (C°) (C°) 50 20 6 2 - 6 0.30-0.52 0.42 40 18 2-6 0.18-0.46 0.29 65 20 4 2-4 0.16-0.41 0.27 40 24 2 - 6 0.11-0.44 0.24 85 20 7 2 - 3 0.15-0.45 0.24 40 24 2 - 5 0.12-0.32 0.20 100 17 2 - 5 0.09-0.29 0.17 The Lagarde Retort The analysis of variance results for the 4 x 2 factorial design with two replicates (Table 11) showed no significant differences in standard devia-tions of temperature in the Lagarde retort with respect to steam content, -78-temperature, replication, or their interactions. The pooled data (Table 12) showed come-up periods to range from 2-6 min, averaging 2 min for heating media containing 50-65% steam, 3 min at 85% steam and 4 min at 100% nominal steam content. The standard deviations ranged from 0.20- 0.34 C° for "pure" steam media to 0.24-0.53 C° for steam/air media containing 50% steam. Stabilization times ranged from 3 to 7 min averaging about 4 min for all conditions studied. The influences of steam content and temperature were probably masked due to the forced circulation of the medium by the turbo fan operating throughout the heating period. The temperature devia-tions reported by Yamano (1976) for laboratory and industrial scale steam/air retorts were considerably higher; probably due to construction and operating principle differences of the two retorts. Table 11. Analysis of variance in temperature distribution for studies of the Lagarde retort. Source of variation df MS F-ratio Steam content 3 0.00366 0.25ns Temperature 1 0.00951 0.66ns Replicate 1 0.00051 0.04ns Interactions Steam-Temperature 3 0.00609 0.42ns Steam-Replicate 3 0.00759 0.70ns Temperature-Replicate 1 0.00141 0.10ns Error 3 0.01460 ns Not significant (p>0.05). -79-Table 12. Summary of overall standard deviations in temperature after the come-up period in the Lagarde retort. Standard deviations Steam No. of Come-up content replicates time Range Mean (%) (min) (C°) (C°) 50 4 2 - 3 0.24-0.53 0.33 65 4 2 - 3 0.22-0.35 0.27 85 4 3-4 0.20-0.37 0.28 100 4 2-6 0.20-0.34 0.26 In the present studies, the overall standard deviations in temperature with respect to thermocouple locations (30 in the positive flow and 20 in the Lagarde retort) and heating times (cook period, approximately 30 min) were used to compare the temperature stability in the retorts under several steam/air conditions. Since the overall standard deviations in temperature were not significantly influenced by the operating mean temperatures (105 to 120°C), no further analyses were performed making use of the coefficient of variation in temperature as the basis of comparison. This seems appro-priate because retort operating conditions are normally specified with respect to a specific temperature ± a certain deviation. Furthermore, in this study some factors were found to influence the temperature distribution of steam/air mixtures in the positive flow retort, but were nonsignificant in the Lagarde retort; thus these should be viewed only as potential factors. The temperature distribution in a retort is influenced greatly by the design and operation of the retort and therefore -80-is very specific to the retort under study. The size and shape of the retort, the plumbing details, design and loading of the rack system, the method of achieving medium homogeneity, and possibly other factors, could influence the temperature distribution achieved. The above procedure could, however, be conveniently used to study the temperature distribution in any retort. -81-Pressure Stability In the Steam/Air Retorts General Importance Temperature uniformity and stability of the heating medium have been recognized as factors of major significance in steam/air processing of foods packed in flexible pouches. While a steady retort temperature is important in delivering the desired lethality to the packaged food, pressure stability is important in maintaining the desired steam content of the heating medium (which is one of the major factors governing the rate of surface heat transfer), preventing the headspace air entrapped in the package from interfering with the rate of heat transfer within, and protecting the pack-age integrity. The need to have an overriding air pressure to prevent package bursting has been well recognized; however, no detailed study on the pressure stability during steam/air processing has been reported. Factors Affecting Pressure Stability The Positive Flow Retort The analysis of variance results for the fractional factorial experiments, employing standard deviations in pressure as the basis for comparing the effects of various factors on the pressure stability, are summarized in Table 13. Steam content and temperature of the heating medium were found to be significant (p<0.05). The pressure variability (standard deviation) was found to increase with a decrease in steam content and an increase in temperature of the heating medium, both contributing to increased retort operating pressures. Thus, a greater pressure variability was generally associated with a higher retort pressure. The variability -82-associated with media flowing upward was generally higher than that for media flowing downward ; however, these were not significant trends (p>0.05). Table 13. Analysis of variance 1 in pressure stability for tests in the positive flow retort. Source of variation df MS F-ratio Steam content 3 0.3058 (0.1727) 6.65* (0.29ns) Temperature 3 0.4040 (0.8557) 8.78** (1.42ns) Flow direction 3 0.0982 (0.9523) 2.13ns (1.58ns) Error 6 0.0460 (0.6029) 1 Using standard deviation and coefficient of variation (values in parentheses) as yield parameters. * Significant at p<0.05; ** Significant at p<0.01 ns Not significant (p>0.05) Although the pressure variations were comparatively higher at lower steam contents and higher temperatures, these were not significant (p>0.05) in terms of the coefficients of variation (Table 13, values in parentheses). Thus, the relative variations, as compared to the magnitudes of their operating retort pressures, were not significant (p>0.05). Under the -83-various conditions employed in the study, the standard deviations in pres-sure varied from 0.055 to 1.65 kPa and the coefficients of variation ranged from 0.17 to 0.49%, for retort pressures from 124 to 379 kPa. The L-16 fractional factorial design (Taguchi, 1957) did not provide for the study of interactions among the factors, and hence, these were not considered. The Lagarde Retort Analysis of variance in the results of the factorial experiments em-ploying steam contents at four levels and temperatures at two levels with two replicates, are given in Table 14. The effects due to steam content, replication, and interactions between factors were not significant (p>0.05); however, temperature was found to influence the standard deviations in pressure during the cook period (p<0.05). Higher temperatures resulted in larger pressure variations. Here again, these effects were not significant (p>0.05) when compared with the variations in the operating retort pres-sures as shown by the analysis of variance using coefficients of variation (Table 14, values in parentheses). Among the conditions studied, the stan-dard deviations varied from 0.63 to 3.38 kPa and the coefficients of varia-tion ranged from 0.33 to 1.36% as the retort pressures changed from 124 to 414 kPa. Considerations for Satisfactory Overriding Pressures The maximum standard deviations in pressure found from the above study were 1.65 and 3.38 kPa for the positive flow and Lagarde retorts, respectively. These standard deviations were computed from the pressure data, recorded at one min intervals, for a cook period of approximately 30 -84-Table 14. Analysis of variance 1 in pressure stability for tests in the Lagarde retort. Source of variation df MS F-ratio Steam content 3 0.6307 (2.0932) 2.95ns (0.68ns) Temperature 1 3.0899 (0.3233) 14.44* (0.10ns) Replicate 1 0.0048 (0.8985) 0.02ns (0.29ns) Interactions Steam-Temperature 3 1.2597 (10.3632) 5.89ns (3.35ns) Steam-Replicate 3 0.3312 (3.6541) 1.55ns (1.18ns) Temperature-Replicate 1 0.1426 (0.4516) 0.67ns (0.15ns) Error 3 0.2139 (3.0947) 1 Using standard deviation and coefficient of variation (values in parentheses) as yield parameters. * Significant at p<0.05 ; ns Not significant (p>0.05) min. Assuming the pressure fluctuations followed a normal or t - distribu-tion (because of the finite sample size of 30), the 99% confidence limits for the above were obtained at three standard deviations (valid for both dis-tributions at n>30) as 4.95 and 10.14 kPa, respectively. One of the requirements for protecting the integrity of pouch packs during processing was identified earlier to be a positive external pressure over internal pouch -85-pressure at all times. Overriding air pressures of at least 4.95 and 10.14 kPa were, therefore, desirable in the positive flow and Lagarde retorts, respectively, during the cook period (assuming no appreciable effects from the air entrapped inside the package). Further, since water vapor pressure is temperature dependent, the pressure variations due to temperature fluctuations are also of importance. From the previous study on temperature distribution, maximum standard deviations in temperature during the cook period for the media of steam contents above 85%, where the concern was high due to existing minimal superimposing air pressures, were found to be 0.45 and 0.37 C°, for the positive flow and Lagarde retorts, respectively. Here again, taking three standard deviations for the 99% confidence limits, the variability could be ±1.35 and ±1.11 C°, respectively. The worst situa-tion, with the pouch contents at the maximum possible temperature and the retort steam pressure at the lower limit, should be considered for calculat-ing the satisfactory overriding air pressure required to counteract the in-ternal pouch pressure. The necessary minimum retort pressures (steam pressure plus the required overriding air pressure) so calculated for tem-peratures of 105-120°C (Table 15) indicated the maximum steam contents for the media to vary from 88-90% in the positive flow and 86-89% in the Lagarde retorts in the temperature range studied. Employing the overriding air pressures recommended by Yamano (1976), the maximum steam contents in the heating media were found to be 91% at 120°C and 86% at 105°C, which are in reasonable agreement with the present findings. These results indi-cated that the operating overriding pressures depend on the ability of the retort control systems to maintain the setpoint temperature and pressure. With a more effective control system, higher amounts of steam could be in-corporated into the media, thereby increasing the potential for improved -86-heat transfer. It should be recognized that the above analysis does not account for pressures developed due to the air entrapped within the pouch; therefore, the headspace air volumes would have to be minimal for the above conditions be to be reasonable. However, since the worst situations were considered in the above analysis, the results should provide a considerable safety margin under normal operations. It should also be recognized that Table 15 is based on the temperature and pressure stability data for the two pilot scale retorts described earlier. Thus, the results would not be applicable to other retorts with different pressure/temperature stability characteristics. The desired conditions for other retort systems could be easily calculated using this approach. Table 15. Maximum steam contents 1 in steam/air heating media for providing satisfactory overriding air pressures in the positive flow and Lagarde retorts. Temperature (°C) Steam pressure (kPa) Positive flow retort Lagarde retort Total pressure (kPa) Steam content (%) Total pressure (kPa) Steam content (%) 105 121 137 88 140 86 110 143 161 89 164 87 115 169 189 89 192 88 120 199 221 90 223 89 1 Based on ±5 and ±10 kPa pressure variations, and ±1.35 and ±1.11 C° temperature variations for the positive flow and Lagarde retorts, respectively. -87-Heat Penetration Studies in the Steam/Air Retorts The Heating Rate Index The heating rate indices obtained for several rectangular test bricks made from RTV silicone rubber and rigid nylon were used to compare the relative effectiveness of steam/air heating media under several processing conditions. The heating rate indices ranged from 9-12 min for the silicone rubber bricks, 12-14 min for the thin nylon bricks and 20-26 min for the thicker nylon bricks under the various heating conditions. Normalizing the data from nylon bricks to a thickness comparable to the silicone rubber bricks (by multiplying the f value by the square of the appropriate ratio of thicknesses) the range of f values for the nylon bricks was found to be 10-16 min. Yamano (1976) reported f values of 13-18 min for various foods in retort pouches of a comparable thickness. The values reported by Pflug et al. (1963) ranged from 6-11 min for tomato puree, chicken a' la king and beef slices, under various steam/air processing conditions. Tung and Garland (1978) used a computerized approach for evaluating f values from the heat penetration data for foods in retort pouches and found them to range from 7 to 16 min. Therefore, the test bricks employed in this study were considered to heat at rates comparable to common foods in retort pouches. The thermal conductivity values suggested by the manufacturers for the silicone rubber and nylon materials were 0.21 and 0.23 W/mC, respec-tively. The f values while heating in "pure" steam were 10.26 ±0.66 min for the silicone rubber bricks (six bricks for four runs plus three bricks for seven runs, n=45), 13.12 ±0.59 min for the thin nylon bricks (three -88-bricks for four runs, n=12), and 21.76 ±1.1 min for the thick nylon bricks (two bricks and ten runs, n=20). Variations in f values for different bricks of a given type in a test run were comparable to those of a single brick of the same type tested several times. The coefficients of variation in f values due to different bricks and different test runs were ±6.4% for the silicone rubber bricks and ±4.5-5.1% for the nylon bricks. Thermal diffusivity val-ues, calculated using Eq.(13), were 1.31 x 10~7 m2/s for the silicone rub-ber bricks, 1.25 and 0.97 x 10~7 m2/s for the thin and thick nylon bricks, respectively. The property variation between the thin and thick nylon bricks was probably due to differences in the nature of the nylon material (obtained at different times from two different sources). This variation should not be of concern because these bricks were employed in different experimental designs; the thicker ones in the positive flow retort and thinner ones in the Lagarde retort. The thermal diffusivities of all the test bricks were within the range of values reported for foods (Dickerson, 1968). Factors Affecting Heating Rate Index The positive flow retort The mean f values for the nylon and silicone rubber bricks under the 27 conditions of the fractional factorial design are given in Table 16 (individual values given in Appendix XI). The f values obtained for the silicone rubber bricks in the vertical orientation were found to be quite large and variable. The clearance between the vertical plates in the rack was 2.4 cm, which was larger than the thickness of the silicone rubber bricks (1.9 cm). Thus, these bricks were not well confined in the rack, -89-Table 16. Mean f values 1 for nonpackaged silicone rubber and nylon bricks in the positive flow retort for the fractional factorial experi-ments. Run Steam Temperature now Package Flow Mean f value # content rate orienta- direction (min) tion 3 (%) (°C) (scfm) Rubber Nylon 1 50 105 40 H upward 9.95 20.75 2 50 105 40 V downward 13.95 21.73 3 50 105 20 H downward 10.26 22.85 4 50 110 40 V downward 13.64 22.09 5 50 110 40 H downward 10.48 22.71 6 50 110 20 H upward 10.26 22.62 7 50 120 40 H downward 10.20 22.48 8 50 120 40 H upward 10.20 22.48 9 50 120 20 V downward 15.69 21.15 10 65 105 40 H downward 10.11 21.90 11 65 105 40 V upward 14.58 21.97 12 65 105 20 H downward 10.40 21.94 13 65 110 40 V upward 13.88 21.69 14 65 110 40 H downward 10.64 22.46 15 65 110 20 H downward 10.16 21.95 16 65 120 40 H downward 10.85 23.09 17 65 120 40 H downward 10.85 22.94 18 65 120 20 V upward 14.28 22.42 19 85 105 40 H • downward 10.16 22.06 20 85 105 40 V downward 12.14 21.49 21 85 105 20 H upward 10.19 22.30 22 85 110 40 V downward 12.43 21.02 23 85 110 40 H upward 10.23 21.43 24 85 110 20 H downward 10.24 22.41 25 85 120 40 H upward 10.55 23.03 26 85 120 40 H downward 10.68 22.95 27 85 120 20 V downward 15.47 21.50 1 Mean value from five silicone rubber and two nylon bricks 2 Silicone rubber bricks measured 1.9 x 12.1 x 17.8 cm and nylon (thick) bricks measured 2.4 x 12.1 x 17.8 cm; The clearance between the plates in the vertical rack was 2.4 cm. 3 Package orientation: V, vertical; H, horizontal -90-whereas the nylon bricks, 2.4 cm thick, were properly sandwiched between the plates. The f values for the nylon bricks were observed to be more uniform. In the horizontal orientation, although none of the bricks were constrained, this was not a problem because of the large open space (57%) on the separation plates (Figure 8) for media circulation. Due to the above reason, the data from silicone rubber bricks from this set of experiments were not considered in the analyses of factors influencing the heating rate index. In all subsequent experiments with the vertical rack, care was taken to ensure proper contact of the separation plates with the test bricks by adjusting the spacers between the plates. Analysis of variance in f values for nonpackaged nylon bricks (Table 17) indicated only the package orien-tation to be a significant factor (p<0.05). Steam content, temperature, flow rate and flow direction of the heating medium did not have an appreciable influence on the f values (p>0.05). The mean f value in the horizontal orientation (22.37 min) was about 3.2% higher than that in the vertical orientation (21.67 min). The fractional factorial tests were based on 27 test runs as compared with 128 runs required for a full factorial design with no replicates; hence, more data were considered necessary to substantiate the findings. In a different study, employing a full factorial design (mean f values obtained for the design are given in Table 18, individual values listed in Appendix XII) with steam content and temperature at four levels and two package orientations, all these factors were found to influence (p<0.05) the f value of nylon bricks (Table 19). The mean f values at steam contents of 50, 65, 85 and 100% steam content were 23.12, 22.99, 22.59 and 22.20 min, respectively. These appear to be consistent with the previously established association of higher rates of heat transfer with higher steam contents. The -91-Table 17. Analysis of variance in f values for various factors tested using nonpackaged thick nylon bricks in the positive flow retort. Source of variation df MS F-ratio Steam content 2 0.132 0.51ns Temperature 2 0.874 3.39ns Flow rate 2 0.161 0.62ns Flow direction 2 0.689 2.67ns Package orientation 2 1.563 6.06* Error 16 0.258 * Significant at p<0.05; ns Not significant (p>0.05) mean f value for bricks heated at 120°C was 23.48 min, about 5.5% higher than the mean f for the bricks heated at 105°C (22.25 min). Volume expan-sion, resulting in an increased thickness of the test bricks or reduced thermal diffusivity [reported by Ramaswamy and Tung (1982) for silicone rubber] at higher temperatures could be reasons for this finding. Orientation effects of the test bricks were found to be reversed in this study, compared with the earlier fractional factorial design, with the mean f value in the vertical orientation, (23.35 min), being 5.7% higher than for the horizontal orientation (22.09 min). This reversal of effects may be due to the large experimental variability associated with the evaluation of f values. These results also suggest that the resistance of the racking plates to the transfer of heat from the medium to the bricks is small in comparison with the resistance offered by the brick material. The high thermal conductivity of steel (16.2 W/mC) and aluminum (239 W/mC) compared with that of silicone rubber (0.21 W/mC) and nylon (0.23 W/mC), and their relative thicknesses, support the above observation. -92-Table 18. Mean f values 1 for silicone rubber and nylon bricks^ in the positive flow retort. Mean £ value (min) Steam Temperature Horizontal Vertical content orientation orientation Rubber Rubber Nylon Rubber Rubber Nylon (%) (°C) packaged packaged 50 105 11.11 11.60 22.54 11.93 10.85 23.39 50 110 10.69 10.52 21.97 11.83 10.48 23.77 50 110 10.98 11.32 22.24 11.43 11.20 24.34 50 120 11.39 11.62 23.34 11.39 11.62 23.34 65 105 10.50 19.28 21.57 11.56 10.01 22.64 65 110 10.57 18.32 21.20 10.99 10.16 23.45 65 110 10.90 17.38 22.70 11.07 10.52 23.89 65 120 10.92 11.22 23.85 11.86 11.47 24.63 85 105 10.06 20.58 21.11 10.70 10.13 23.17 85 110 10.88 20.80 22.28 10.52 9.66 23.28 85 110 9.97 20.80 21.48 10.81 10.10 22.78 85 120 10.34 18.98 22.23 10.98 10.35 24.39 100 105 10.19 22.34 21.39 10.19 9.62 22.21 100 110 10.21 22.67 22.36 10.26 9.95 22.64 100 110 10.49 19.13 20.68 10.88 9.88 22.26 100 120 10.69 20.81 22.58 10.90 10.41 23.44 } Mean value from four silicone rubber and two nylon bricks. 2 Silicone rubber bricks measured 1.9 x 12.1 x 17.8 cm, while nylon (thick) bricks measured 2.4 x 12.1 x 17.8 cm, with appropriate clearance in the vertical orientation. Further experiments were also carried out to investigate the effects of steam content and temperature of the heating media, package orientation, and packaging of the silicone rubber bricks in retort pouches (residual air content, 15-30 mL) on the f values. The mean f values under different -93-conditions for this factorial experiment are also included in Table 18 (individual values in Appendix XII). Analysis of variance in f values for the nonpackaged silicone rubber bricks (Table 19) indicated that the steam content of the medium and orientation of the test bricks were significant (p<0.05) factors. As found with the nylon bricks, lower f values were found to be associated with higher steam contents (11.34, 11.05, 10.53 and 10.48 min at steam contents of 50, 65, 85 and 100%, respectively). The mean f value in the vertical orientation was 11.08 min, about 4.3% higher than in the horizontal orientation (10.62 min). Medium temperature and all two-way interactions were found to be nonsignificant (p>0.05). Table 19. Analysis of variance in f values using nonpackaged silicone rubber and nylon bricks in the positive flow retort. Source of variation Rubber brick Nylon brick df F-ratio df F- ratio Steam content 3 12.37** 3 5.02* Temperature 3 1.44ns 3 7.96* Brick orientation 1 15.23* 1 45.37** Interactions Steam-Temperature 9 0.75ns 9 1.69ns Steam-Orientation 3 0.85ns 3 0.68ns Temperature-Orientation 3 0.30ns 3 0.43ns Error 9 9 * Significant at p<0.05 ; ** Significant at p<0.01 ns Not significant (p>0.05) -94-Large increases were observed when the f values of packaged silicone rubber bricks were compared with those for the nonpackaged ones of the same type, while heating them in an unconstrained horizontal orientation in media with steam contents above 85%. Differences were also observed in the f values for the packaged silicone bricks in horizontal orientation at steam contents higher than 65% compared with those at 50% steam content, and at temperatures 105-110°C (65% steam content) compared with 120°C. These variations caused appreciable nonhomogeneity within the cell blocks for the factorial analyses, except at 50% steam content. The analysis of variance test performed for the combined data, therefore, showed all factors and most of their two-way interactions to be highly significant (p<0.01). In order to make a more meaningful interpretation, the data from Table 18 were analyzed further for the significance of the effects of different factors in each of the two orientations, separately. Analysis of variance results with the test bricks in the horizontal orientation (Table 20), indicated the steam content, packaging and steam content-packaging interactions to be significant (p<0.05). Contrary to the earlier finding, larger f values were found to be associated with media hav-ing higher steam contents (particularly for the packaged bricks), indicating a larger resistance to heat transfer at higher steam contents. The f values for the silicone rubber bricks in retort pouches were considerably higher (up to 120%) at steam contents of 65% (and above) at temperatures of 105-110°C, and at 85% (and above) at 120°C, which may indicate that super-imposing air pressures under these situations were inadequate to counteract the effects of entrapped air on the overall heat transfer. With steam con-tents of 50% at 105-110°C and <65% at 120°C, there were no appreciable differences in the f values obtained for the silicone rubber bricks with or -95-Table 20. Analysis of variance in f values using packaged and non-packaged silicone rubber bricks in the positive flow retort. Horizontal Vertical Source of variation orientation orientation df F-ratio df F-ratio Steam content 3 19.30** 3 19.88** Temperature 3 1.32ns 3 6.03* Packaging 1 197.61** 1 36.93** Interactions Steam-Temperature 9 1.00ns 9 0.78ns Steam-Packaging 3 25.77** 3 0.24ns Temperature-Packaging 3 2.12ns 3 1.51ns Error 9 * Significant at p<0.05; ** Significant at p<0.01 ns Not significant (p>0.05) without packaging. This indicated that an overriding air pressure up to 100 kPa may be necessary to minimize the resistance to heat transfer due to entrapped air (15-30 mL per package in this study) while processing pouch packs in an unconstrained horizontal orientation. External means of providing proper contact between the contents and the pouch material at the larger heat transfer surfaces could minimize the need to have high overriding pressures as evidenced by the results under the vertical orientation (Table 20) for which steam content, temperature and packaging were found to be significant (p<0.05) factors. Packaged silicone rubber bricks in the vertical orientation had a mean f value of 10.40 min, even lower than that for the nonpackaged bricks (11.08 min). This could be due to the possibility of a slightly better sandwiching effect for the packaged -96-silicone rubber bricks in the vertical rack, or due to experimental varia-tions. It was observed earlier (but not included in the analysis) that when the silicone rubber bricks were loosely placed in the vertical rack, the f values obtained were significantly higher and also were characterized by a greater variability. Thus, adequate contact of the heat transfer surfaces was essential for establishing lower f values. The mean f values for the bricks in vertical orientations were found to decrease with increasing steam contents (11.34, 10.95, 10.41 and 10.26 min at 50, 65, 85 and 100% steam, respectively), while they increased with temperature (10.62 min at 105°C to 11.12 min at 120°C). These observations were similar to those obtained for the nylon bricks. The Lagarde retort The mean f values for the full factorial analysis involving steam con-tent (four levels), temperature (two levels), orientation (two levels), packaging (two levels) and replication (two levels) are given in Table 21 for the silicone rubber and nylon bricks (individual values given in Appendix XIII). An analysis of variance in the f values obtained (Table 22) for the silicone rubber bricks (nonpackaged) in vertical or horizontal orientation and the nylon bricks in the horizontal orientation showed neither steam con-tent nor temperature of the heating media to be significant (p>0.05). Although not indicated in Table 22, an analysis of the data from Table 21 for silicone rubber bricks for orientation effects indicated the resulting f values to be significantly influenced (p<0.05) by the orientation (horizontal orientation being associated with f values about 5% higher compared to -97-Table 21. Mean f values 1 for silicone rubber and nylon bricks 2 in the Lagarde retort. Mean f value (min) Steam Temperature Horizontal Vertical content orientation orientation Rubber Rubber Nylon Rubber Rubber (%) (°C) packaged packaged 50 105 10.32 11.63 13.62 10.34 10.29 50 105 10.54 11.48 12.53 9.82 10.26 50 120 10.58 11.33 12.42 10.16 9.99 50 120 10.68 11.74 12.24 10.08 10.44 65 105 10.16 25.93 12.94 10.90 10.07 65 105 10.39 19.84 12.82 9.80 9.96 65 120 10.72 13.20 12.20 10.37 10.26 65 120 10.68 11.73 12.92 9.59 10.32 85 105 9.98 28.55 12.49 10.06 9.70 85 105 10.51 22.47 12.81 10.05 10.22 85 120 10.35 13.76 12.81 9.48 10.01 85 120 10.91 12.57 12.24 9.69 10.53 100 105 10.21 36.58 12.74 9.81 10.23 100 105 10.66 19.92 13.37 10.36 11.21 100 120 10.88 20.55 13.82 10.22 11.32 100 120 10.36 28.28 12.54 9.63 10.47 1 Mean value from two or three bricks of each type. 2 Silicone rubber bricks measured 1.9 x 12.1 x 17.8 cm, while nylon (thin) bricks measured 2.1 x 12.1 x 17.8 cm, with appropriate clearance in the vertical orientation. vertical). Here again, when analysis of variance was performed to include the effect due to packaging, all experimental factors and most of their two-way interactions were found to be significant (p<0.05), possibly because of the nonhomogeneity introduced due to the large f values of packaged silicone rubber bricks in the horizontal orientation under heating conditions involving steam contents above 65%. Hence, these were considered separately. -98-Table 22. Analysis of variance in f values using nonpackaged silicone rubber and nylon bricks in the Lagarde retort. Horizontal Vertical orientation orientation Source of variation : Rubber Nylon Rubber df F-ratio F-ratio F-ratio Steam content 3 Temperature 1 Replicate 1 Interactions Steam-Temperature 3 Steam-Replicate 3 Temperature-Replicate 1 Error 15 0.15ns 0.45ns 0.66ns 7.30ns 0.60ns 1.69ns 2.99ns 0.32ns 2.47ns 0.31ns 0.28ns 0.38ns 1.27ns 0.32ns 1.58ns 2.26ns 0.15ns 0.01ns ns Not significant (p>0.05) In the vertical orientation, the analysis of variance (Table 23) indi-cated only the packaging effect to be significant (p<0.05). The mean f value for the packaged silicone rubber bricks (10.33 min) was about 3% higher than the mean value (10.02 min) for the silicone rubber bricks with-out packaging. In the horizontal orientation, however, the effects of steam content, temperature, packaging, and their two-way interactions were found to influence (p<0.05) the f value. Significantly higher f values were ob-served for the packaged silicone bricks in horizontal orientation, while heating in media at steam contents above 65%. These results were similar to those obtained in the positive flow retort. The f values obtained for pack--99-aged silicone bricks while heating under the above conditions were up to 260% higher compared with those for the nonpackaged silicone bricks in vertical or horizontal orientations. Slight discrepancies in the observed f values for nonpackaged silicone bricks in the two retorts may be due to differences in the construction and operating principles of the two retorts or due to experimental variations. The relatively larger variations in the f values of the packaged silicone bricks in the two retorts may be due to differences in the actual volume of air contained in the pouches. Table 23. Analysis of variance in f values using packaged and non-packaged silicone rubber bricks in the Lagarde retort. Source of variation Horizontal orientation Vertical orientation df F-ratio df F-ratio Steam content 3 5.23* 3 2.02ns Temperature 1 5.72* 1 0.07ns Packaging 1 38.29** 1 5.92* Replicate 1 1.07ns 1 0.15ns Interactions Steam-Temperature 3 1.12ns 3 0.02ns Steam-Packaging 3 5.23* 3 1.95ns Steam-Re plicate 3 0.16ns 3 1.68ns Temperature-Packaging 1 6.84* 1 2.70ns Temperature-Replicate 1 2.42ns 1 0.44ns Packaging- Replicate 1 1.39ns 1 3.64ns, Error 13 * Significant at p<0.05 ; ** Significant at p<0.01 ns Not significant (p>0.05) -100-In general, an increase of up to 11% (p<0.05) in the f values was ob-served as the steam content of the media in the positive flow retort de-creased from 100% to 50%. Heating of the bricks at 120°C also resulted in an increase in the f value (up to 5.5%) as compared with heating at 105°C. In the Lagarde retort, the effects of both steam content and temperature on the f values were nonsignificant (p>0.05). No conclusion could be drawn with respect to orientation effect; some results indicated a higher f value in the vertical orientation while others showed the reverse trend. Entrapped air in pouches did not appear to influence the f in the vertical constrained orientation while they resulted in up to 260% higher values of f in the unconstrained horizontal orientation when processing at 105-120°C in media © of steam contents above 65%. The heat penetration curves for the packaged silicone rubber bricks in horizontal orientation were uniquely different when employing heating media of high steam contents as compared with similar bricks, packaged or not, under other conditions. The former ones possessed a characteristic broken heating behavior (Figure 16). The f values for the packaged silicone bricks calculated from the first straight line portion of these curves were com-parable to the f values for the other silicone bricks. The f value reported in this study for the above situation was the slope index calculated from the latter part of the broken heating curve which represented a larger span of the heating curve involving the region of higher center temperatures. Another noteworthy feature of the heat penetration curves was the relative lack of the characteristic lag period for the bricks as compared with the cylinder. The f value of the silicone rubber in a 300 x 401 can (7.3 cm inside diameter and 10.1 cm length) was approximately four times that of the brick (1.9 x 12.1 x 17.8 cm). -101-110 T i m e , m in Figure 16. Typical heat penetration curves for silicone rubber bricks and cylinders In a steam/air medium containing 85% steam at 120°C I a, for a nonpackaged brick (1.9 x 12.1 x 17.8 cm); b, for a similar brick in a retort pouch (entrapped air, 15-30 mL), and c, in a 300 x 401 cylindrical can (time scale doubled for this curve) 1. -102-The Lag Factor Although the heating rate index, f, was considered important while com-paring the different media with respect to their heat transfer capabilities, the lag factor, j, should be coupled with it in order to fully describe the heat penetration curve and to facilitate thermal process calculations. The lag factor could be influenced only to a small extent due to differences in the surface heat transfer coefficient in the range found for the steam/air mixtures (Yamano, 1976). On the other hand, the initial temperature difference, Ta-Ti, and the come-up time could have strong influences on the j value (Ball and Olson, 1957). With the initial temperatures ranging from 30-7(J°C and the come-up times varying from 2-6 min, the j values computed using the 42% effectiveness criteria of Ball (1923) for various conditions in the positive flow retort are given in Table 24. Similar results in the Lagarde retort, with initial temperatures ranging from 30-40°C and come-up times of 2-6 min, are presented in Table 25. The j values were generally in the range of 0.5-1.0 for the test bricks of silicone rubber and nylon in the positive flow retort and 0.8-1.1 in the Lagarde retort. The lower values in the above range were generally associated with longer come-up times and higher initial temperatures. Batch to batch variation in the retort come-up times and initial brick temperatures may be reasons for the observed variability in the j values. The j values for packaged silicone rubber bricks in the horizontal orientation at high steam contents showed much larger variations among the replicates, lower values being usually associated with situations involving higher f values. The theoretical values for situations involving no come-up time and high surface heat transfer coefficients were reported to be 1.273 -103-Table 24. Mean j values 1 for silicone rubber and nylon bricks z in the positive flow retort. Mean j value Steam Temperature H o r i z o n t a l V e r t i c a l content orientation orientation R ubber Rubber Nylon Rubber Rubber Nylon (%) (°C) packaged packaged 50 105 0.632 0.629 0.915 0.770 0.670 0.964 50 110 0.875 0.889 0.944 0.712 0.626 0.941 50 110 0.725 0.700 0.990 0.781 0.687 0.974 50 120 0.415 0.428 0.715 0.623 0.560 0.957 65 105 0.568 0.404 0.851 0.444 0.370 0.724 65 110 0.701 0.462 1.027 0.674 0.602 0.970 65 110 0.639 0.475 0.951 0.650 0.591 0.948 65 120 0.483 0.504 0.825 0.461 0.400 0.760 85 105 0.540 0.433 0.910 0.674 0.616 0.930 85 110 0.453 0.345 0.830 0.705 0.636 0.969 85 110 0.691 0.447 0.997 0.673 0.590 0.892 85 120 0.562 0.392 0.865 0.587 0.522 0.892 100 105 0.579 0.378 0.946 0.736 0.677 0.995 100 110 0.626 0.312 0.986 0.731 0.611 0.951 100 110 0.623 0.415 0.897 0.583 0.541 0.906 100 120 0.496 0.377 0.818 0.502 0.429 0.807 1 Mean value from three silicone r u b b e r and two nylon b r i c k s . 2 Silicone r u b b e r b r i c k s measured 1.9 x 12.1 x 17.8 cm, while nylon ( t h i c k ) b r i c k s measured 2.4 x 12.1 x 17.8 cm, with appropriate clearance i n the v e r t i c a l o r i e n t a t i o n . -104-Table 25. Mean j values 1 for silicone rubber and nylon bricks^ in the Lagarde retort. Mean j value Steam Temperature content H o r i z o n t a l orientation V e r t i c a l orientation R ubber Rubber Nylon Rubber Rubber (%) (°C) packaged packaged 50 105 0.848 0.768 0.815 1.124 0.864 50 105 0.800 0.929 0.946 0.978 0.800 50 120 0.908 0.638 0.902 1.175 0.935 50 120 0.906 0.803 1.004 1.099 0.959 65 105 0.821 0.600 0.862 1.020 0.871 65 105 0.742 0.700 0.870 0.870 0.779 65 120 0.653 0.430 0.722 0.956 0.723 65 120 0.579 0.671 0.672 0.640 0.547 85 105 0.905 0.984 1.033 1.081 1.001 85 105 0.815 0.909 0.955 0.964 0.864 85 120 0.753 0.674 0.879 0.961 0.790 85 120 0.849 0.725 0.954 1.026 0.891 100 105 0.849 0.655 0.947 1.002 0.898 100 105 0.849 0.654 0.947 0.923 0.780 100 120 0.646 0.624 0.747 0.744 0.654 100 120 0.903 0.633 0.993 1.053 0.931 1 Mean value from three b r i c k s of each type. 2 Silicone r u b b e r b r i c k s measured 1.9 x 12.1 x 17.8 cm, while nylon (thin) b r i c k s measured 2.1 x 12.1 x 17.8 cm, with a p p r o p r i a t e clearance i n the v e r t i c a l o r i e n t a t i o n . -105-for an infinite plate and 2.064 for a finite brick (Ball and Olson, 1957). The length and width of the test bricks used in this study were 5-8 times larger than their thicknesses. The temperature rise at the central plane was, therefore, predominantly determined by heat transfer through the two larger surfaces, except near the edges of the brick. Hence, these bricks could be considered to heat like an infinite plate and values of j between 1.2 and 1.3 should be expected, if the come-up times were essentially zero. The j values found in the present study for the bricks were considerably lower than even these values. The j values calculated based on no come-up time (or 100% as effectiveness of the come-up time instead of 42% used earlier) varied from 1.3-1.8. These values were comparable to (or higher than) the j value for an infinite plate, suggesting that a different come-up time correction factor (higher than 42%) might be applicable for situations involving processing of foods in thin brick-shaped configurations. In the present investigation, the effectiveness of the come-up time, in order that the calculated j values match the expected value of 1.27 for an infinite plate, varied from 60-90% depending on the length of the come-up time and the temperature gradient, Ta-Ti. A further analysis of the lag factor as related to the come-up time, f and Ta-Ti is presented in a later section. -106-Theoretical Considerations Heat Transfer in Rectangular Solids A detailed discussion of the theoretical background for heat transfer in rectangular solids under convective surface heat transfer was made earlier leading to the derivation of Eq.(ll) which relates the heating rate index, f, and the surface heat transfer coefficient, h. Eq.(ll) was employed for evaluating the surface heat transfer coefficient making use of experimental f values from rectangular test bricks of high thermal conductivity (aluminum and stainless steel). Test bricks of rectangular geometry, with thermal pro-perties in the range common for food materials, were later employed to com-pare the effectiveness of the steam/air heating media. Theoretical computa-tion of the f values and the lag factors, j, for the silicone rubber and nylon bricks under the various steam/air processing conditions will be discussed in this section. The Heating Rate Index The surface heat transfer coefficients associated with steam/air media at different steam contents in the positive flow and Lagarde retorts can be obtained using Eqs. (15)-(18). These values may then be used to compute the f values for the silicone rubber and nylon bricks using Eg. (11), since the thermal property values and dimensions of these bricks are known. The predicted f values (Table 26) at steam contents ranging from 50-100% in the positive flow and Lagarde retorts under different flow conditions compared well with the experimental values under the appropriate conditions (Tables 16, 18 and 21). Slight differences between the experimental and calculated -107-Table 26. Predicted f values for the nonpackaged silicone rubber and nylon bricks 1 in the positive flow and Lagarde retorts. Positive flow retort Steam Medium flow: upward Medium flow: downward content — h f value (min) h f value (min) RubberNyl-tn Nyl-tk RubberNyl-tn Nyl-tk (%) (W/m2C) (W/m2C) 50 1260 10.62 13.58 22.47 1990 10.49 13.42 22.24 65 2360 10.45 13.38 22.17 3390 10.39 13.30 22.07 85 5480 10.34 13.24 21.98 6890 10.33 13.22 21.95 100 10310 10.31 13.19 21.91 11730 10.30 13.19 21.90 Lagarde retort Brick: horizontal Brick: vertical h f value (min) h f value (min) Rubber Nyl-tn Nyl-tk Rubber Nyl-tn Nyl-tk (W/m2C) (W/m2C) 50 3290 10.40 13.31 22.07 3130 10.40 13.32 22.09 65 4010 10.37 13.28 22.03 4390 10.36 13.27 22.01 85 5220 10.35 13.24 21.99 6900 10.33 13.22 21.95 100 6360 10.33 13.23 21.96 9690 10.31 13.20 21.91 1 Nylon bricks: Nyl-tn, thin (2.1cm), and Nyl-tk, thick (2.4cm) 2 Brick: vertical ° Medium flow: horizontal -108-values may be due in part to the experimental variability in measuring the f values. The differences in computed f values for the test bricks at 50% steam were only about 3% higher than those at 100% steam. Similar differences with an added variability of ±5% at each level of steam content may or may not produce statistical differences in f values with respect to steam content. These observations confirmed the earlier experimental findings, with the steam content to be significant in some cases while not significant in others. This theoretical analysis could be applied only to the nonpackaged test bricks, since no study has been made to determine the magnitude of the interfacial resistances in the packaged bricks. Experimental results indicated the f values of the bricks to increase by about 3% due to packaging, when properly confined, thereby slightly decreasing the overall rate of heat transfer. Presence of entrapped air along the larger surface of the test brick could, however, result in up to 260% higher values of f. The overall rate of heat transfer will be considerably reduced under these situations. These limitations must be recognized when using the theoretical approach. The Lag Factor The lag factor, j , is normally evaluated through a heat penetration curve for use in thermal process calculations: j = (Ta-Tpi)/(Ta-Ti) (19) where Tpi is the pseudo-initial temperature obtained from the intercept of -109-the straight line portion of the log(Ta-T) vs. t curve extended to the be-ginning of the process. For situations involving a zero come-up time, the j value calculated as above should be comparable to the j calculated using Eq.(4) for an infinite plate (R value) and Eq.(10) for a finite brick-shaped solid. As discussed earlier, thin rectangular bricks of materials like silicone rubber, nylon or food, which have relatively low thermal conductivities, could be considered to behave like infinite plates while heating or cooling. The predicted j values at several surface heat transfer coefficients, in the range common for steam/air mixtures (1,000-10,000 W/m2C), for the silicone rubber and nylon bricks were essentially the same, and equal to 1.27. In situations involving non-zero retort come-up times, the j values calculated using Eq.(19) could be different from values calculated from Eq.(4). A correction factor of 42%, suggested by Ball (1923), for the effectiveness of the come-up period, has been generally employed in thermal process calculations involving cylindrical containers. In practice, this would mean shifting the X-axis of the heating curve from the actual beginning (time zero, corresponding to "steam on" in a batch retort) to a time, tr, equivalent to 58% of the come-up time, tc (Ball and Olson, 1957). These are illustrated in Figure 17 which is an inverted plot of log(Ta-T) vs. t. This new time (tr) would then be the beginning of the theoretical or effective process and the pseudo-initial temperature for the process is obtained from the Y-intercept at time = tr for use in calculating the j. Theoretically, this would mean shifting an imaginary heating curve, having the same f value but no come-up period [j for which can be calculated using Eg. (4) ], along the X-axis to match the actual heating curve. The j value for this new curve originating at time, tr, would be the same as for the imaginary curve -110-Figure 17. Hypothetical heat penetration curves for a packaged food in a retort with and without come-up periods. - I l l -originating at time zero. The effectiveness of the come-up time can be calculated using the relationship: Effectiveness(%)= [1 - (tr/tc) ] x 100 (20) An expression for tr/tc can be obtained in terms of known parameters as follows: From Figure 17, it can be seen that the two points [0, (Ta-TpiO)] and [tr, (Ta-Tpil)] lie on the straight line portion of the heating curve (for which the slope is equal to -1/f), TpiO and Tpil being the pseudo-initial temperatures (Tpi) at time, zero, and at time, tr, respectively. Recognizing that the relationship between Ta-T and t is semi-logarithmic, [log(Ta-Tpil) - log(Ta-TpiO)] 1 = - - , (21) [tr-0] f (Ta-Tpil) tr or log { i = , (22) 1 (Ta-TpiO)' f tr , (Ta-TpiO) or — = log \ \. (23) f { (Ta-Tpil) > Dividing and multiplying the temperature terms on the right hand side by (Ta-Ti), Eq. (23) can be written as, tr i (Ta-TpiO) (Ta-Ti) ) — = l o g ] — • [ (24) f ( (Ta-Ti) (Ta-Tpil) ) -112-In terms of j values, [ (Ta-Tpil)/(Ta-Ti) ], represents the corrected j value assuming a time, (tc-tr), of the come-up period, tc, to be effective. Theoretically, this j value is equal to 1.273 for an infinite plate. The term [ (Ta-TpiO)/(Ta-Ti)], represents the experimental j value for the heating curve assuming zero retort come-up time. Representing this by jO, and rearranging Eq.(24), an expression for tr can be obtained (for an infinite plate): Dividing both sides of Eq.(25) by tc, the final equation can be obtained: Thus the effectiveness of come-up period could be obtained using Eq.(20). It is evident that the come-up time effectiveness is dependent on the jO which depends on initial temperature difference, Ta-Ti, [and f value, which influences (Ta-TpiO)], come-up time (tc) and the heating rate index (f). The j values from this study, calculated at 42% effectiveness for the come-up period, varied from 0.5-1.1, considerably lower than the expected value of 1.273. At 100% effectiveness, the j values were generally above 1.3 indicating the come-up time effectiveness to be somewhere in between. The results presented in Tables 24 and 25 show considerable variation in the j values probably because of the differences in the initial temperatures and f values for the different test bricks, and retort come-up times for the different test runs. tr = f log[ jO /1.273] (25) tr f — = — log tc tc (26) -113-Considering the maximum f and jO values found in the heat penetration studies (f = 23 min, jO = 1.8, thick nylon bricks), the high-side value of tr from Eq.(25) was found to be 3.5 min. The minimum effectiveness of the come-up period (2-6 min) was, therefore, equal to [ 1.0-(3.5/6)] x 100 or 42%. Similar calculations for the silicone rubber bricks gave a value of 75% minimum effectiveness. In a recent study, Spinak and Wiley (1982) report-ed the effectiveness of come-up time for retort pouch processing to range from 35-77%, while Yamano (1976) found the 42% effectiveness suggested by Ball (1923) to be acceptable. It is beyond the objectives of this research to explore this concept in detail; however, further research related to the effect of controlled come-up times involving different time-temperature patterns, different (Ta-Ti) and f values, on the resulting process lethali-ties should provide useful information for thermal processing operations. Limiting Surface Heat Transfer Coefficient Experimental variability associated with the evaluation of f for foods packaged in flexible pouches were reported to be as high as ±21% by Tung and Garland (1978), while ±3.7-5.7% variations were found from the f values reported by Pflug (1964). In the present investigation, standard deviations of 5-6% in f values were found to be common. It appears, therefore, that a variability of ±5% for experimental evaluation of f values in foods is a reasonable assumption to make. From Table 26, it can be observed that the differences between the f values become smaller at larger surface heat transfer coefficient values. Traditionally, steam has been assumed to pos-sess an infinitely large surface heat transfer coefficient (Ball and Olson, 1957), and most thermal processing calculations are based on relationships -114-similar to Eq.(13). If a variability of ±5% can be assumed for the f values, then a surface heat transfer coefficient value can be identified, by using Eq.(ll), to give an f value no greater than 5% more than the f value ob-tainable at infinite surface heat transfer coefficient [Eq.(13)]. This h value can then be considered the "limiting surface heat transfer coefficient", beyond which f values could be predicted using Eq.(13) with less than 5% error involved. An analysis of the calculated f values at several surface heat transfer coefficient levels for rectangular test bricks of various dimensions and thermal properties indicated that the limiting surface heat transfer coeffi-cient value increased with thermal conductivity, decreased with package thickness and remained independent of the thermal diffusivity. The magni-tude of the f value, however, was influenced by all of the above variables. Typical detailed information on the limiting surface heat transfer coefficient for various package dimensions and the respective f values for a range of thermal diffusivities common to foods, and at a thermal conductivity value of 0.5 W/mC, are given in Table 27. The f values for infinite surface heat transfer can be obtained either by using Eq.(13) or by dividing the values in Table 27 by 0.95. Similar results at other thermal conductivity values could be obtained by using the computer program given in Appendix I with slight modification. Approximate minimal surface heat transfer coefficient values, h", at conductivities, k", can be obtained from any value h' at k' (from Table 27), using the relationship, h" = h'(k"/k'). The value of k1 employed in Table 27 is 0.5 W/mC. At thermal conductivity values of 0.21 and 0.23 W/mC, the silicone rubber and nylon bricks had a minimal heat transfer coefficient value of about 700 W/m2C. Since all test conditions in Table 26 had an h value T a b , e 27. Limiting surface heat transfer c o . f M C . n t . and associated f v a . e s for steam/air processing. Dimensions LIml ting h 1 .0 1. 1 1 .2 1.3 cm 0.5 0.6 1.2 1 .4 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 cm 12.0 12.0 12.0 12 .0 12 .0 12.0 12.0 12 12 12 12 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 15.0 15.0 15.0 15.0 15.0 18.0 21 .0 24.0 27 .0 30.0 cm W/m2C 18.0 18.0 18.0 18.0 18.0 18.0 18.0 18 18 18 18 18 18 18.0 18.0 18.0 18 .O 18 18 18 18 21 24 27 .0 30.0 30.0 30.0 30.0 30.0 30.0 8030. 6690. 5720. 5000. 4440. 3990. 3310. 2820. 2460. 2170. 2050. 1940. 1530. 1250. 1050. 900. 790. 700. 630. 570. 570. 580. 590. 590. 590. 600. 610. 620. 620. 630. Thermal d i f f u s i v i t y , m'/s (X 10') 1.4 1.5 1.6 1.7 f value (min) 1 .02 0. 93 0. 85 0. 78 1 .47 1 . 33 1 . 22 1 . 13 1 .99 1 . 81 1 . 66 1 . 53 2.60 2. 36 2. 16 2. 00 3.28 2. 98 2. 73 2 . 52 4 .04 3. 68 3. 37 3. 1 1 5.80 5. 27 4 . 83 4 . 46 7 .85 7 . 14 6. 54 6 . 04 10. 19 9. 27 8. 49 7. 84 12.82 1 1 . 65 10. 68 9. 86 14.23 12 . 94 1 1 . 86 10. 95 15.71 14. 28 13. 09 12 . 08 24 .02 21 . 84 20. 02 18. 48 33.72 30. .65 28. 10 25. 94 44 .57 40 .52 37. . 14 34 28 56.33 51 .21 46 .94 43 .33 68 .75 62 .50 57 .29 52 .89 81 .65 74 .22 68 .04 62 .80 94 .79 86 . 17 78 .99 72 .91 108.03 98 .21 90 .03 83 . 10 115.76 105 .24 96 .47 89 .05 118.48 107 .71 98 .73 91 . 14 120.28 109 .35 1O0 .23 92 .52 121.63 110 .57 101 .35 93 .56 122.61 1 1 1 .46 102 . 17 94 .31 127.81 116 . 19 106 .51 98 .32 131.15 119 .22 109 .29 100 .88 133.38 121 .26 111 . 15 102 .60 135.04 122 .76 112 .53 103 .88 136. 14 123 .77 1 13 .45 104 .73 0.73 1 .05 1 .42 1 .86 2.34 2 .89 4 . 14 5.61 7 . 28 9. 16 10. 16 11 .22 17 . 16 24.09 31 .83 40.24 49. 1 1 58.32 67.70 77. 17 82.69 84 .63 85.92 86.87 87.58 91 .29 93.68 95.27 96.46 97.25 0.68 0.98 1 .33 1 .73 2 . 19 2.70 3.87 5.23 6.80 8.55 9.49 10.47 16.01 22.48 29.71 37 .55 45.83 54.43 63. 19 72.02 77. 18 78 .98 80. 19 81 .08 81 .74 85.21 87.43 88.92 90.03 90.76 0.64 0.92 1 .24 1 .62 2.05 2.53 3.62 4.91 6.37 8.01 8.89 9.82 15.01 21 .07 27.85 35.21 42 .97 51 .03 59.24 67 .52 72.35 74 .05 75. 18 76.02 76.63 79.88 81 .97 83.36 84 .40 85.09 0.60 0.86 1 . 17 1 .53 1 .93 2.38 3.41 4 .62 6.00 7.54 8.37 9.24 14 . 13 19.83 26.22 33. 14 40.44 48.03 55.76 63.55 68. 10 69.69 70.75 71 .54 72 . 12 75. 18 77. 15 78.46 79.43 80.08 0.57 0.81 1.11 1 .44 1 .82 25 22 36 66 12 91 2 3 4 5 7 7 8.73 13.35 18.73 24.76 31 .29 38 .20 40.36 52.66 60.02 64.31 65.82 66.82 67 .57 68 . 1 1 71.01 72.86 74. 10 75.02 75.64 0.54 0.77 1 .05 1 .37 1 .73 2.13 3.05 4. 13 5.37 6.75 7.49 8.27 12.64 17.75 23.46 29.65 36. 19 42.97 49.89 56.86 60.93 62.36 63.31 64 .01 64.53 67.27 69.02 70.20 71 .07 71 .65 0.51 0.73 1 .00 1 .30 1 .64 2.02 2.90 3.93 5. 10 6.41 7 . 12 7.85 12.01 16.86 22.28 28. 17 34.38 40.82 47.39 54.02 57 .88 59.24 60. 14 60.81 61 .30 63.91 65.57 66.69 67.52 68.07 Calculated at a thermal conductivity value of 0.5 W/mC. T. and L are the thickness, width and length of the brick. -116-greater than 700 W/m2C, the variability in the calculated theoretical f values should be less than 5%. This is in agreement with the 3% value re-ported earlier. The experimental f values also matched the predicted values from Table 27 within reasonable error (±5%). These results indicate that steam/air mixtures with an associated heat transfer coefficient as low as 1260 W/m2C could be as effective as pure steam, which has a considerably higher h value, in processing rectangular bricks (thickness >1.9 cm) of low thermal conductivity materials like silicone rubber and nylon. For food materials with a common conductivity value of 0.5 W/mC, at the commonly used pouch thickness of 1.9 cm (0.75 in), the minimum surface heat trans-fer coefficient value could be about 2000 W/m2C (Table 27). The limiting h will be larger for packages with thickness lower than 1.9 cm. Steam/air media with a minimum steam content of 60 and 50% in the positive flow and Lagarde retorts, respectively, are capable of delivering the desired h value of 2000 W/rrTC. Under situations involving h values above the minimal level, thermal diffusivity of the food material governs the overall rate of heat transfer into a brick of specified dimensions. Since the packaging method and the volume of entrapped air could significantly affect f values, the above analysis is valid mainly for ideal situations involving nonpackaged bricks. However, the results of this study indicated that, with proper orientation and confinement, the f values obtained for the bricks contained in pouches were only 3% higher than those for the nonpackaged bricks of similar type, and therefore, the experimental values for the packaged bricks were comparable to the above theoretical values. In the present study, it was also observed that presence of headspace air in the pouches could result in large increases (up to 260%) in the f -117-values in the unconstrained horizontal orientation while being heated at 105-120°C in media of steam contents above 65%, suggesting the need to maintain headspace air at a minimal level and to properly confine the pouches in order to obviate the use of high superimposing air pressures. It is important to recognize these factors because, in some commercial steam/air retorts for food products, the retort pouches are processed in the horizontal orientation in a medium containing 75% steam. The nonhomogeneity of food samples, which might influence the effective thickness of the pouch, should also be considered. Further research is required to identify the critical factors involved in these operations. Pressure/Volume Relationships in Steam/air Processing Davis et al. (1960) reported a maximum pressure difference of 6.67 kPa between the inside and outside of the pouch (internal pressure exceed-ing the external) while heating in steam at 100°C to prevent damage to the food package. This is of major concern only while heating in pure steam because, in steam/air mixtures, the presence of air contributes to a super-imposing external pressure thereby maintaining a positive outside pressure over the internal. However, if the retort pressure was subjected to large variations, the possibility of the internal pressure exceeding the external cannot be overlooked, especially while employing steam/air mixtures at high steam contents and low temperatures. Yamano (1976) recommended that an overriding air pressure of 20-30 kPa was desirable for steam/air processing at 100-120°C. -118-Effect of Entrapped Air An approximate analysis of the influence of noncondensible gases within a food package can be made assuming that the ideal gas laws apply to this situation. The volume, VI, of the entrapped air at the packaging temperature, T l , and pressure, PI, would expand to a volume, Va, at the retort temperature, Ta, and pressure, Pa, according to the ideal gas law: If the expansion were to be prevented, the retort pressure would have to be selected such that The required pressure, Pa, is therefore dependent on the initial pressure and temperature, and the processing temperature. Considering room temperature (T1=293°K) and atmospheric pressure (Pl=101 kPa, absolute) at the time of packaging, the minimum retort pressure required to prevent expansion of the entrapped air could be calculated using Eq.(28) at different retort temperatures. For example, at 105, 110, 115 and 120°C, pressures of 131, 132, 134 and 136 kPa, respectively, would be required. Comparing these values with the saturated steam pressures (Ps) at the retort temperatures, it was found that at temperatures of 110°C and above, the steam pressure in the retort was higher than the required pressure to prevent the expansion of the entrapped air. At 105°C, however, the steam pressure of 121 kPa was lower than the required 131 kPa to prevent air expansion. A superimposing air pressure was, therefore, essential. Since (Pl.VD/Tl = (Pa.Va)/Ta. (27) Pa > Pl(Ta/Tl). (28) -119-the fractional steam content of a steam/air mixtures was calculated by Ps/Pa, the maximum steam content of the medium to give a retort pressure of 131 kPa at 105°C was found to be 92%. The entrapped air content within the packages was, therefore, not critical with respect to package integrity during steam/air processing at temperatures of 110°C and above. However, the presence of a large volume of air could interfere with the heat transfer rates even when expansion was prevented, as evidenced by the results presented earlier for silicone rubber bricks packaged in retort pouches with an entrapped air content of 15-30 mL when heated in an unconstraining horizontal rack. Overriding air pressures of 70-100 kPa were found neces-sary in order for these bricks to heat at rates comparable to similar bricks without packaging or with adequate constrainment. Therefore, while preven-ting volume expansion protects package integrity, the ability to keep the entrapped air away from the larger surfaces determines the rate of heat transfer into the package. Effect of Water Vapor The water vapor pressure inside a pouch due to food moisture could be considered to be the same as that of saturated steam outside (neglecting the lowering of the partial pressure of water vapor due to the effect on wa-ter activity of the food material) when the contents of the pouch approach the retort temperature. Therefore, in order to protect package integrity, the outside pressure should be at least slightly above the saturated steam pressure at the operating temperature. Pressure and temperature variations in the retort could result in the internal pressure exceeding the external when operating at pure or high steam conditions. The necessary super-imposing conditions for this situation were worked out earlier (Table 15). -120-Effect of Entrapped Air Plus Water Vapor When considering entrapped air and water vapor in combination, the internal pressure would be equal to the water vapor pressure plus the partial pressure of the entrapped air. Knowledge of the volumes of entrapp-ed air, the pouch and pouch contents, their initial conditions and relative expansions at processing temperatures, would be necessary to analyze this situation. Whitaker (1971) and Yamano (1976) have arrived at some useful relationships for idealized situations making a number of assumptions and approximations. Further research in this area is necessary in order to better understand the pressure/volume relationships with respect to pouch integrity and the'effects on heat transfer rates. -121-SUMMARY AND CONCLUSIONS A method employing transient heat conduction was developed to evaluate surface heat transfer coefficients associated with steam/air mixtures using data on heat penetration into rectangular bricks of high thermal conductiv-ity. A quick release system was designed to contain the test brick in an in-sulated box inside the retort during the come-up period and to release it into the heating medium when the desired conditions were achieved, thereby facilitating an instantaneous immersion of the brick into the steam/air heating medium. This condition was an essential requirement in the heat transfer model but difficult to achieve in batch-type retorts due to heating of the test brick during the come-up period. A detailed analysis of the theoretical background, experimental methodology and discussion of the results were given and some limitations of the existing methods for evaluating the surface heat transfer coefficient were identified. Using the developed procedure, the influence of factors such as steam content, temperature, flow rate and flow direction of the heating medium, and brick orientation on the associated surface heat transfer coefficients of steam/air mixtures were studied in two pilot scale batch retorts. In the positive flow retort, steam content, flow rate and flow direction were identified as significant (p<0.05) factors influencing the h values, whereas in the Lagarde retort, the medium steam content and brick orientation were found to be significant (p<0.05). The steam/air temperature was found to be nonsignificant (p>0.05) in both retorts. In the positive flow retort, the h value was also found to increase linearly with the medium flow rate. In general, the relationship between h and the percent steam content (S) of the steam/air media could be accurately expressed by an exponential -122-function, h = a exp(bS). Regression coefficients for this exponential relationship were reported for various steam/air conditions in the positive flow and Lagarde retorts. Temperature distribution studies in the positive flow retort indicated that the overall standard deviation in temperature at several locations during the cook period were significantly (p<0.05) influenced by the steam content and flow rate of the heating medium, while the effects of tempera-ture, flow direction and rack type were nonsignificant (p>0.05). In the Lagarde retort both factors studied, medium composition and temperature, were found to be nonsignificant (p>0.05). Pressure stability studies indi-cated that the standard deviations in the retort pressure during the cook period increased significantly (p<0.05) with the air content and temperature of the medium in the positive flow retort, and with the medium temperature in the Lagarde retort. In both retorts, these effects were nonsignificant (p>0.05) when the coefficients of variation were considered instead of standard deviations. Based on the maximum standard deviations in tempera-ture and pressure (±0.45 C° and ±1.65 kPa, respectively, in the positive flow retort, and ±0.37 C° and ±3.38 kPa, respectively, in the Lagarde re-tort), the maximum steam contents in the media to give a satisfactory over-riding air pressure (effect of entrapped air considered to be minimal) were considered to vary from 88-90% in the positive flow retort, and 86-89% in the Lagarde retort for the temperature range of 105-120°C. Heat penetration studies in the positive flow retort using rectangular test bricks of silicone rubber and rigid nylon indicated an increase of up to 11% (p<0.05) in the f values as the steam content of the media decreased from 100% to 50%. Heating the bricks at 120°C also resulted in an increase of up to 5.5% (p<0.05) in the f value as compared with heating at 105°C in -123-both retorts. In the Lagarde retort, the effects of both steam content and temperature of the medium on the f values were nonsignificant (p>0.05). No conclusions could be drawn with respect to the orientation effects on the f value; some studies indicated a higher f value in the vertical orientation while others showed the reverse trend. Entrapped air inside the pouches (15-30 mL per pouch) did not appear to influence the f value of the silicone rubber bricks (increase in f value about 3%) in the vertical constrained orientation, whereas up to 260% higher f values resulted when using an unconstrained horizontal orientation while processing at 105-120°C using steam contents above 65%. The differences could be considerably reduced using superimposing air pressures of 70-100 kPa during the heating period. The j values were generally in the range of 0.5-1.0 for the test bricks in the positive flow retort, and 0.8-1.1 in the Lagarde retort, when calcu-lated at 42% effectiveness for the come-up time. Batch to batch variations in the retort come-up times and initial temperature of the test bricks were thought to be reasons for the above variability in the j values. Further-more, the come-up time effectiveness of 42% was considered conservative for situations involving processing of foods in thin profile brick-shaped con-figurations. It was observed that in order for the j values to match the theoretical value of 1.27 for an infinite plate, the effectiveness of the come-up time would be in the range of 60-90% depending on the length of the come-up time and initial temperature of the test brick. A detailed theoretical treatment was used to predict the f and j values for the silicone rubber and nylon bricks under the associated surface heat transfer coefficients of steam/air mixtures. The predicted f values compared well with the experimental f values. A theoretical relationship was derived for calculating the effectiveness of come-up time and to assess its influence -124-on the evaluated j values. The effectiveness was found to depend on the f value, the length of retort come-up time and the initial temperature differ-ence between the retort and the test brick. A theoretical analysis of the heat transfer equations was also made to identify a 'limiting h value' beyond which the h value could be assumed to be infinite, with less than 5% error introduced in the evaluated f value. Assuming the effect of entrapped air inside the package to be minimal, the limiting h for processing common foods in conventional pouches (thickness >1.9 cm) was identified to be 2000 W/m2C. Steam/air mixtures containing over 60% steam were capable of providing this h value. Thus, the useful range of steam/air mixtures for thermal processing of foods in thin profile packages would be more than 60% steam (with respect to delivering the desired h value) and less than 85% (with respect to protecting the package integrity). Although these studies have been carried out in pilot scale retorts, and the results may not be directly transferable to commercial processing operations involving large retorts, they should contribute to a better understanding of the concepts involved in the application of steam/air mixtures to food processing. -125-LITERATURE CITED Abdul-Hadi, M.I. 1979. An analytical investigation into dropwise condensation of different steam/air mixtures on substrates of various materials. Can. J. Chem. Eng. 57(4) :459. Adams, J.P. and Peterson, W.R. 1982. Recent studies in processing of institutional-sized retort pouches. Presented at the 42nd Annual Meeting of the Institute of Food Technologists, Las Vegas, NV. June 22-25. Albin, F.V., Badari Narayana, K., Srinivasa Murthy, S., and Krishna Murthy, M.V. 1979. Thermal diffusivities of some unfrozen and frozen food models. J. Food Technol. 14(4) :361. Andrews, W.P. 1975. Food Packaging: The successful story of flex-pack worldwide use. Activities Rep. of the R&D Associates for Military Food and Packaging Systems, Inc. 27(1): 147. Annamma, T.T. and Rao, C.V.N. 1974. Studies on thermal diffusivity and conductivity of fresh and dry fish. Fishery Technol. (India), 11(1):28. Anon. 1982. We noticed that. Food in Canada 42(3) :46. Awbery, J.H. and Griffiths, E. 1933. Thermal properties of meat. J. Soc. Chem. Ind. 52:326. Ball, CO. 1923. Thermal process time for canned food. Bull. 37. National Research Council, Washington, DC. Ball, CO. and Olson, F.C.W. 1957. Sterilization in Food Technology. McGraw-Hill Book Co., Inc., New York, NY. Bennett, A.H., Chace, W.G. and Cubbedge, R.H. 1969. Heat transfer properties and characteristics of Appalachian area 'Red Delicious' apples. ASHRAE Trans. 75(2): 133. Berry, R., Jr. 1979. The sterilization of food in pouches - critical parameters for still processing. Proc. of the Conference, held in Indianapolis, IN. Using Retort Pouches Worldwide - Focus on the Present with a Look to the Future. Sponsored by Food Sciences Inst., Purdue Univ., March 14-15, p.7. Beverly, R.G. 1979. Sterilization methodology applied to pouches and trays. Proc. of the Conference, held in Indianapolis, IN. Using Retort Pouches Worldwide - Focus on the Present with a Look to the Future. Sponsored by Food Sciences Inst., Purdue Univ., March 14-15, p.102. Beverly, R.G. 1980. Retort pouch in the '80s. Food Eng. 52(3) :100. -126-Beverly, R.G., Strasser, J. and Wright, B. 1980. Critical factors in filling and sterilizing of institutional pouches. Food Technol. 34(9) :44. Bhowmik, S.R. and Hayakawa, K.-I. 1979. A new method for determining the apparent thermal diffusivity of thermally conductive food. J. Food Sci. 44(2):469. Blaisdell, J.L. 1963. Natural convection heating of liquids in unagitated food containers. Ph.D. Thesis, Michigan State Univ., East Lansing, MI. Carslaw, H.S., and Jaeger, J.C. 1959. Conduction of Heat in Solids, 2nd ed. Oxford Univ. Press, London, England. Charm, S.E. 1978. The Fundamentals of Food Engineering, 3rd ed. The AVI Publishing Co., Inc., Westport, CT. Choi, Y. and Okos, M.R. 1980. Thermal properties of tomato juice. ASAE Paper No.80-6515. Clary, B.L., Nelson, G.L. and Smith, R.E. 1971. The application of geometry analysis technique in determining the heat transfer rates from biological materials. Trans. ASAE 14(3) :586. Connor, S.F. 1979. Sterilization methodology applied to pouches and trays. Proc. of the Conference, held in Indianapolis, IN. Using Retort Pouches Worldwide - Focus on the Present with a Look to the Future. Sponsored by Food Sciences Inst., Purdue Univ., March 14-15, p.132. Corning, A. 1979. Sterilization methodology applied to pouches and trays. Proc. of the Conference, held in Indianapolis, IN. Using Retort Pouches Worldwide - Focus on the Present with a Look to the Future. Sponsored by Food Sciences Inst., Purdue Univ., March 14-15, p.92. Coulson, J.M. and Richardson, J.F. 1977. Chemical Engineering, 3rd ed. Vol. I. Pergamon Press, Oxford, England. Davis, E.G., Karel, M. and Proctor, B.E. 1960. The pressure- volume relation in film packages during heat processing. Food Technol. 14(3):165. Davis, R.B., Long, F.E. and Robertson, W.F. 1972. Engineering considerations in retort processing of flexible packages. Food Technol. 26(8):65. Dickerson, R.W., Jr. 1965. An apparatus for the measurement of thermal diffusivity of foods. Food Technol. 19(5) :198. Dickerson, R.W., Jr. 1968. Thermal properties of foods. In: The Freezing Preservation of Foods, 4th ed. Vol. II. D.K. Tressler, W.B. Van Arsdel and M.J. Copley, eds. The AVI Publishing Co., Westport, CT. -127-Duxbury, D.D., Sams, P.F., Howeler, W.F., Gee, J.H. and Miller, W.N. 1970. Reliability of Flexible Packaging for Thermoprocessed Foods under Production Conditions. Phase I: Feasibility. Tech. Rep. 72-77-GP. U.S. Army Natick Laboratories, Natick, MA. Ford, R.M. and Bilanski, W.K. 1969. Thermal diffusivity of alfalfa stems. Trans. ASAE 12(2) :249. Frechette, R.J. and Zahradnik, J.W. 1968. Thermal properties of Mcintosh apple. Trans. ASAE 11(1) :21. Gaffney, J.J., Baird, CD. and Eshleman, W.D. 1980. Review and analysis of the transient method for determining thermal diffusivity of fruits and vegetables. ASHRAE Trans. 86(2) :261. Gane, R. 1936. Thermal conductivity of the tissue of fruits. Ann. Rep., Food Investigation Board, Great Britain, p.211. Goldfarb, P.L. 1970. Pouch for low-acid foods. Part I. Mod. Packaging 43(12):70. Goldfarb, P.L. 1971. Pouch for low-acid foods. Part II. Mod. Packaging 44(1):70. Gould, W.A., Geisman, J.R., Weiser, H.H., Bash, W.D., Moore, W.J., Salzer, R.H. and Long, F.E. 1962. Heat processing of vegetables in flexible films. Ohio Agric. Exp. Stn., Res. Bull. 905. Gurney, H.P. and Lurie, J. 1923. Charts for estimating temperature distributions in heating or cooling of solid shapes. Ind. Eng. Chem. 15:1170. Harmathy, T.Z. 1971. Peak-time method for measuring thermal diffusivity of small solid specimens. AIChE J. 17(1): 198. Hayakawa, K.-I. 1972. A new method for the computerized determination of apparent thermal diffusivity of food. Proc. of the First Pacific Chem. Eng. Congr., Kyoto, Japan, Oct. 10-14. Part 11:129. Hayakawa, K.-I. and Bakal, A. 1973. New computational procedure for determining the apparent thermal diffusivity of a solid body approximated with an infinite slab. J. Food Sci. 38(4) :623. Hebbard, G.M. and Badger, W.L. 1934. Steam-film heat transfer coefficients for vertical tubes. Ind. Eng. Chem. 26(4) :420. Heisler, M.P. 1947. Temperature charts for induction and constant temperature heating. Trans. ASME 69:227. Heldman, D.R. and Singh, R.P. 1981. Food Process Engineering, 2nd ed. The AVI Publishing Co. Inc., Westport, CT. -128-Hoddinott, R.I. 1975. The retortable pouch: Advantages to processor, retailer, consumer. Package Devel. 5(2) :25. Hu, K.H. 1953. Investigation of the use of plastic containers for heat processed foods. Ph. D. Thesis, Univ. of Illinois, Urbana, IL. Hu, K.H., Nelson, A.I., Legault, R.R. and Steinberg, M.P. 1955. Feasibility of using plastic film packages for heat processed foods. Food Technol. 9(5):236. Kellar, R.G. 1959a. Flexible Packages for Processed Foods. QMFCIAF Rep. No. 31-59. Quartermaster Food and Container Inst. for the Armed Forces. Chicago, IL. Kellar, R.G. 1959b. Flexible packages for processed foods. Mod. Packaging 33(1):145. Kern, D.O. 1950. Process Heat Transfer, McGraw-Hill Book Co., Inc., New York, NY. Kisaalita, W.S. 1981. A study of heat transfer from steam/air mixtures to a retort pouch substrate. M.A.Sc. Thesis, Univ. of British Columbia, Vancouver, BC. Klis, J.B. 1961. Canning without cans. 2. Food Process. 22:38. Kopetz, A.A., Prange, CA. and Flessner, R.J. 1979. The future in our hands: Critical factors in retort pouch thermal process assurance. Activities Rep. of the R&D Associates for Military Food and Packaging Systems, Inc. 31(2) :49. Kostaropoulos, A.E., Spiess, W.E.L. and Wolf, W. 1975. Reference values for the thermal diffusivity of food-stuffs. Lebensm. - Wiss. u. Technol. 8(3): 108. Kreith, F. 1973. Principles of Heat Transfer, 3rd ed. Harper and Row, New York, NY. Kusak, L.J. 1958. The condensation of vapors from noncondensing gases. Ph.D. Thesis, Cornell Univ, Ithaca, NY. Lampi, R.A. 1973. The Reliability Program. Rep. T-7310. Packaging Inst., St. Louis, MO. Lampi, R.A. 1977. Flexible packaging for thermoprocessed foods. Adv. Food Res. 23:305. Lampi, R.A. 1979. Flexible packaging for thermoprocessed foods. In: Fundamentals of Food Canning Technology. J.M. Jackson and B.M. Shinn, eds. AVI Publishing Co., Inc., Westport, CT. Lampi, R.A. 1980. Retort pouch: The development of a basic packaging concept in today's high technology era. J. Food Process Eng. 4:1. -129-Lentz, CP. 1961. Thermal conductivity of meats, fats, gelatin gels and ice. Food Technol. 15(5) :243. Lentz, C P . and van den Berg, L. 1977. Thermal conductivity data for foods: their significance and use. ASHRAE Trans. 83(3) :533. Mannheim, H.C, Nelson, A.I. and Steinberg, M.P. 1957. Film packaging of heat processed orange juice products. Food Technol. 11(8) :417. Martens, T. 1980. Mathematical model of heat processing in flat containers. Ph. D. Thesis, Catholic Univ., Louvain, Belgium. Mayer, P.C. and Robe, K. 1961. Canning without cans. 1. Food Process. 22:36. McAdams, W.H. 1954. Heat Transmission. 3rd ed. McGraw-Hill Book Co. Inc., New York, NY. Mellor, J.D. 1976. Thermophysical properties of foodstuffs. 1. Introductory review. Bull. Int. Inst. Refrig. 56(3):551. Mellor, J.D. 1978. Thermophysical properties of foodstuffs. 2. Theoretical aspects. Bull. Int. Inst. Refrig. 58(3) :569. Mellor, J.D. 1979. Thermophysical properties of foodstuffs. 3. Measurements. Bull. Int. Inst. Refrig. 59(3):551. Mermelstein, N.H. 1976. An overview of the retort pouch in the U.S. Food Technol. 30(2) :28. Mermelstein, N.H. 1978. Retort pouch earns 1978 I FT Food Technology Industrial Achievement Award. Food Technol. 32(6) :22. Merrill, D.G. 1948. Heating rates of glass containers. Ind. Eng. Chem. 40:2263. Metals Hand Book. 1980. Properties and Selection: Stainless Steels, Tool Materials and Special Purpose Metals, 9th ed. Vol. 3. American Society for Metals, Metals City, OH. Milleville, H.P. 1980. Steam/air retorting of pouches. Food Process. 41(3):98. Mohsenin, N.N. 1980. Thermal Properties of Foods and Agricultural Materials, Gordon and Breach Science Publishers, New York, NY. Monteza, P.F.C. 1979. Prediction of nutrient retention in foods packaged in retortable pouches and sterilized by thermal conduction. Ph. D. Thesis, Louisiana State Univ. and Agricultural and Mechanical College, Baton Rouge, LA. Morgan, B.H., Long, F.E. and Bock.J.H. 1961. Problems in Assuring Sterility in Thermally Preserved Flexibly Packaged Foods. R&D Associates, Quartermaster Food and Container Inst., Chicago, IL. -130-Nelson, A.I. and Steinberg, M.P. 1956. Retorting foods in plastic bags. Food Eng. 28(1) :92. Nelson, A.I., Hu, K.H. and Steinberg, M.P. 1956. Heat processible food films. Mod. Packaging 29(10): 173. Newman, A.B. 1936. Heating and cooling rectangular solids. Ind. Eng. Chem. 28:545. Nix, G.H., Lowery, G.W., Vachon, R.I. and Tanger, G.E. 1967. Direct determination of thermal diffusivity and thermal conductivity with a refined line-source technique. Progress in Aeronautics and Astronautics: Thermophysics of Spacecraft and Planetary Bodies 20:865. Olson, F.C.W. and Jackson, J.M. 1942. Heating curves- theory and practical applications. Ind. Eng. Chem. 34:337. Othmer, D.F. 1929. The condensation of steam. Ind. Eng. Chem. 21(6) :576. Parcell, J.W. 1930a. Investigation on retorting of glass containers. Part I. Canning Age 11(7): 475. Parcell, J.W. 1930b. Investigation on retorting of glass containers. Part II. Canning Age 11(8):535. Parker, R.E. and Stout, B.A. 1967. Thermal properties of tart cherries. Trans. ASAE 10(4) :489. Peters, J.W. 1975. Retail debut of retort pouch earns consumer acceptance. Food Prod. Devel. 9(2) :22. Peterson, W.R. and Adams, J.P. 1982. Water velocity effect on heat penetration parameters during institutional size retort pouch processing. Presented at the 42nd Annual Meeting of the Institute of Food Technologists, Las Vegas, NV. June 22-25. Pflug, I.J. 1964. Evaluation of heating media for producing shelf stable food in flexible packages. Phase I. Final Rep., Contract DA19-AMC-145 (N). U.S. Army Natick Laboratories, Natick, MA. Pflug, I.J. and Borrero, C. 1967. Heating media for processing food in flexible packages. Phase II. Tech. Rep. 67-47-GP. U.S. Army Natick Laboratories, Natick, MA. Pflug, I.J., Bock, J.H., and Long, F.E. 1963. Sterilization of food in flexible packages. Food Technol. 17(9) :87. Pflug, I.J., Blaisdell, J.L. and Kopelman, I.J. 1965. Developing temperature time curves for objects that can be approximated by a sphere, infinite plate or infinite cylinder. ASHRAE Trans. 71:238. -131-Pinto, A. 1978. Retort pouch: moving to close the materials and machinery gap. Mod. Packaging 51(3):23. Polley, S.L., Snyder, O.P. and Kotnour, P. 1980. A compilation of thermal properties of foods. Food Technol. 34(11):76. Qashou, M.S., Nix, G.H., Vachon, R.I. and Lowery, G.W. 1970. Thermal conductivity values for ground beef and chuck. Food Technol. 24(4):189. Qashou, M.S., Vachon, R.I. and Touloukian, Y.S. 1972. Thermal conductivity of foods. ASHRAE Semi-Annual Meeting, New Orleans, LA. Ramaswamy, H.S. and Tung, M.A. 1981. Thermophysical properties of apples in relation to freezing. J. Food Sci. 46(3) :724. Ramaswamy, H.S. and Tung, M.A. 1982. Modified Dickerson method to study temperature dependence of thermal diffusivity of foods. Presented at the 25th Annual Meeting of the Canadian Institute of Food Science and Technology, Montreal, PQ., May 30 - June 3. Ramaswamy, H.S., Lo, K.V., and Tung, M.A. 1982. Simplified eguations for transient temperatures in conductive foods with convective heat transfer at the surface. J. Food Sci. 47(6):2042. Ramaswamy, H.S., Tung, M.A. and Stark, R. 1983. A method to measure surface heat transfer for steam/air mixtures in batch retorts. J. Food Sci. (In press). Rao, M.A., Barnard, J. and Kenny, J.F. 1975. Thermal conductivity and thermal diffusivity of process variety squash and white potatoes. Trans. ASAE 18(6):1188. Roop, R.A. and Nelson, P.E. 1981. Processing retort pouches in conventional sterilizers. J. Food Sci. 47(1) :303. Rubinate, F.J. 1960. Flexible Containers for Heat-Processed Foods, Proc. of the Conference on Flexible Packaging for Military Food Items, National Academy of Sciences/Quartermaster Food and Container Inst, for the Armed Forces, Chicago, IL. p.21. Rubinate, F.J. 1961. Flexible Packages for Heat Processed Foods, Proc. 13th Res. Conference of the Am. Meat Inst. Foundation, Chicago, IL. Rubinate, F.J. 1964. Army's "obstacle course" yields a new look in food packaging. Food Technol. 18(11):71. Schulz, G.L. 1978. Trends in food packaging for foodservice. J. Food Prot. 41(6):464. -132-Shappee, J. and Werkowski, S.J. 1972. Study of a nondestructive test for determining the volume of air in flexible food packages. Tech. Rep. 73-4-GP. U.S. Army Natick Laboratories, Natick, MA. Singh, R.P. 1982. Thermal diffusivity in food processing. Food Technol. 36(2):87. Smith, R.E. and Nelson, G.L. 1969. Transient heat transfer in solids: theory versus experiment. Trans. ASAE 12(6) :833. Smith, R.E., Nelson, G.L. and Henrickson, R.L. 1967. Analyses on transient heat transfer from anomalous shapes. Trans. ASAE 10(2):236. Smithells, C.J., ed. 1962. Metals Reference Book, 3rd ed. Vol. II. Butterworths, London, England. Snedecor, G.W. 1965. Statistical Methods, 5th ed. The Iowa State Univ. Press, Ames, IA. Spinak, S.H. and Wiley, R.C. 1982. Comparisons of the general and Ball formula methods for retort pouch process calculations. J. Food Sci. 47(3):880. Steffe, J.F., Williams, J.R., Chhinnan, M.S. and Black, J.R. 1980. Energy requirements and costs of retort pouch vs. can packaging systems. Food Technol. 34(9) :39. Strasser, J.H. 1979. Up-date on retort pouch fillers and sterilizer. Activities Rep. of the R&D Associates for Military Food and Packaging Systems, Inc. 31(2) :54. Suter, D.A., Agrawal, K.K. and Clary, B.L. 1975. Thermal properties of peanut pods, hulls and kernels. Trans. ASAE 18(2) :370. Taguchi, G. 1957. Experimental Designs. Maruzen Publishing Co., Tokyo, Japan. Thompson, G.E. 1919. Temperature-time relations in canned foods during sterilization. Ind. Eng. Chem. 11:657. Toumy, J.M. and Young, R. 1982. Retort-pouch packaging of muscle foods for the armed forces. Food Technol. 36(2) :68. Tripp, G.E. 1961. Conventional retorting of flexibly packaged products. Activities Rep. of the R&D Associates for Military Food and Packaging Systems, Inc. 13:222. Tsutsumi, Y. 1975. The growth of food packed in retortable pouches in Japan. Activities Rep. of the R&D Associates for Military Food and Packaging Systems, Inc. 27(1): 149. -133-Tsutsumi, Y. 1979a. Producing pouches and trays to suit particular needs. Proc. of the Conference, held in Indianapolis, IN. Using Retort Pouches Worldwide - Focus on the Present with a Look to the Future. Sponsored by Food Sciences Inst., Purdue Univ., March 14-15, p.24. Tsutsumi, Y. 1979b. New technology applied to pouches and future trends. Proc. of the Conference, held in Indianapolis, IN. Using Retort Pouches Worldwide - Focus on the Present with a Look to the Future. Sponsored by Food Sciences Inst., Purdue Univ., March 14-15, p.67. Tsutsumi, Y. 1979c. Sterilization methodolody applied to pouches and trays. Proc. of the Conference, held in Indianapolis, IN. Using Retort Pouches Worldwide - Focus on the Present with a Look to the Future. Sponsored by Food Sciences Inst., Purdue Univ., March 14-15, p.86. Tung, M.A. 1974. Temperature distribution in a steam/air retort for thermally processed foods in flexible pouches. Univ. of British Columbia, Vancouver, BC, 9pp. Tung, M.A. 1980. Thermophysical studies for improved food processes. A First Quarter Rep. (16th June to 15th Sept.). DSS File No. 35SZ.01804-9-0001, prepared for the Agriculture Canada PDR Program, Ottawa, ON. Tung, M.A. and Garland, T.D. 1978. Computer calculation of thermal processes. J. Food Sci. 43(2):365. Wadsworth, J.I. and Spadaro, J.J. 1969. Transient temperature distribution in whole sweetpotato roots during immersion heating. I. Thermal diffusivity of sweetpotatoes. Food Technol. 23(2) :85. Wallenberg, E. and Jarnhall, B. 1957. Heat sterilization in plastics. Mod. Packaging 31(7):165. Whitaker, W.C. 1971. Processing flexible pouches. Mod. Packaging 44(2) :83. Williams, J.R., Steffe, J.F. and Black, J.R. 1981. Economic comparison of canning and retort pouch systems. J. Food Sci. 47(1) :284. Woodams, E.E. and Nowrey, J.E. 1968. Literature values of thermal conductivities of foods. Food Technol. 22(4) :150. Wratten, F.T., Poole, W.D., Chesness, J.L., Bal, S. and Ramarao, V. 1969. Physical and thermal properties of rough rice. Trans. ASAE 12(6):801. Yamano, Y. 1976. Studies on thermal processing of flexible food packages by steam-and-air retort. Ph.D. Thesis, Kagawa Univ., Kagawa-Ken, Japan. Yamano, Y., and Komatsu, Y. 1969. Sterilization of foods in flexible packages. Part I. Heat characteristics of a pilot retort for sterilization of flexible packages. J. Food Sci. Technol. (Japan), 16(3): 113. -134-Yamano, Y., Komatsu, Y., and Ikegami, Y. 1969a. Sterilization of foods in flexible packages. Part II. Thermal characteristics and storage stability of various film-wrapped foodstuffs. J. Food Sci. Technol. (Japan), 16(3):119. Yamano, Y., Yamaguhi, K., Nagata, H., Komatsu, Y., and Kishimoto, A. 1969b. The sterilization of foods in flexible packages. Part III. Temperature distribution and heat characteristics of an industrial scale retort. J. Food Sci. Technol. (Japan), 16(3): 124. Yamano, Y., Ejiri, K., Endo, T., and Senda, M. 1975. Heat penetration in to flexible packages with food simulated materials. J. Food Sci. Technol. (Japan), 22(5): 199. -135-Appendix I. Computer program for the evaluation of surface heat transfer c o e f f i c i e n t s of steam/air mixtures. ,*»,^*»»»»**••*»*»**»»«*»*»****»»********" , ,****** ,** 1 (;.«..»«.««»»»*»«•*.»»»***«»»»«»«»..««»»»»»**»««»*»»»»*»»*»** 2 C • 3 C This is a program to predict the surface heat transfer • 4 C c o e f f i c i e n t by using the ava i l a b l e data on thermal • 5 C and physical properties of a rectangular test brick of * 6 C known dimensions and the measured f-sub-h value from a * 7 C heat penetration tes t . * 8 C » 9 C The program 1s based on so l u t i o n to conduction heat * 10 C transfer involving convection at the surface and makes • 11 C use of the s i m p l i f i e d r e l a t i o n s h i p s developed by * 12 C Ramaswamy et al.(1982) published in d.Food Sci.47:2042. * 13 C 14 C**»* 15 1G 17 C READ DATA 18 19 C T i t l e for the output (LABEL) upto 80 characters 20 21 C Product dimensions: th1ckness(A1), width(A2) and 1ength(A3) 22 C in cm (Format F7.3) 23 24 C Thermal properties: conductivity (COND) in W/mC (F7.3) and 25 C d i f f u s i v i t y (DIFF). in m.sq/s X 10»*7 (F7.3). 2G 27 C Data logger c o r r e c t i o n factor (CORR) strokes/min (F7.3) 28 29 30 C Sample I.D. (Up to 32 characters) followed by f-sub-h values 31 C in seconds (F7.3) for which h value is to be evaluated. 32 33 34 INTEGER LABEL(20). ID(8) 35 36 1 READ(5.5) LABEL 37 WRITE(6,6) LABEL 38 READ(5.10,END*1000) A1.A2.A3 39 READ(5,20) COND,01FF,COR 40 4 1 5 FORMAT(20A4) 42 6 FORMAT('1',10X,20A4,///) 43 7 F0RMAT(10X, ' '. 44 & ' ',/ 45 S 25X,'Sample ID'.13X.'f-sub-h',/10X. 46 8' ./11X. 47 &9X, ' input c a l c . h './/11X, 48 & 9X,' s s W/m2C'./10X, 49 &' - './) 50 8 FORMAT( 10X. ' ' , 51 & /) 52 10 F0RMAT(3F7.3) 53 20 F0RMAT(3F7.3) 54 WRITE(6,50)A1,A2.A3,COND,DIFF,COR 55 50 F0RMAT(2OX,* PRODUCT INFORMATION',/20X. 56 & ' - './15X, -136-& 'Thickness « '.F7.3.' cm'./15X. ft 'Width - '.F7.3.' cm'./15X, 57 S B ft 'Wiam - . r. . w„. — , 59 & 'Length • '.F7.3.' cm',/15X, 60 ft /15X.'Thermal conductivity • '.F7.3,' W/mC',/15X. 61 ft 'Thermal d i f f u s i v i t y • '.F7.3.' X 10**-7 m sq/s'./IOX. 62 ft /15X.'Dlgltec c o r r e c t i o n factor • '.F7.3.' s/stroke',//) 63 64 65 WRITE(6.7) 66 67 A1-A1/20O. 68 A2»A2/200. 69 A3«A3/200. 70 71 DIFF«DIFF/(10.**7) 72 73 C Read sample 1.0. and f-sub-h data 74 75 100 READ(5.110,END=500) ID 76 110 FORMAT(8A4) 77 READ(5,120) FH 78 120 FORMAT(F7.3) 79 80 FH=FH*C0R 81 82 C 83 C 84 85 86 87 88 C 89 C 90 C 91 92 93 300 HTOHTC+500.0 94 BI1=HTC*A1/COND 95 BI2«HTC*A2/C0ND Compute the f i r s t estimate of h value using the lumped capacity method. HTC-2.303*((A1*A2*A3)/(A1*A2+A2*A3+A3*A1)) «C0ND/(DIFF*FH) Increment h values by 500 and compare the actual and predicted f-sub-h values: go through the loop u n t i l the the di f f e r e n c e between the two h values i s less than 500. H S 5r=210?3^Bl2;(BI2.2.)*0.2795*ATAN(BI2/3.)-0.02915.ATAN(5..BI2) ?rJ20?3^Br3;(BI3*2.)*0.2795.ATAN(BI3/3.)-0.02915.ATAN(5..BI3) 101 1 0 2 S3*S3*0.0O1171 \H FSUBH=2.303/(DIFF.(S1/A1-2.*S2/A2-2.*S3/A3"2.)) IF(FH.LT.FSUBH) GOTO 300 C Loop around again to reduce the difference to 100. 104 105 106 107 108 109 350 HTC-HTC-100. 1 1 0 BI1»HTC*A1/COND 1 t 1 BI2«HTC*A2/C0Np I ^ O T ^ I ^ !1J I;:210?38^B;72;(B12 + 2.)*0.2795.ATAN(BI2/3.)-0.02915.ATAN(5..B12) 115 11 S S2-S2+0.001171 -137-117 S3 = 2.0738*BI3/(BI3+2 .)+0.2795*ATAN(BI 3/3.)-0.02915*ATAN(5 . *BI3) 118 53=53+0.001171 1 19 120 FSUBH=2 303/(DIFF*(Sl/A1**2.+S2/A2**2.+S3/A3**2. )) 121 IF(FH.GT.FSUBH) GOTO 350 122 123 C Loop around again to f i n e tune the difference to <10. 124 125 400 HTOHTC+10. 126 BI1*HTC*A1/COND 127 BI2=HTC*A2/C0ND 128 BI3*HTC*A3/C0ND 129 S1=2.0738*BI1/(BI1 + 2.)+0.2795*ATAN(BI1/3.)-0.02915*ATAN(5.*BI 1 ) 130 S1=51+0.001171 131 S2=2.0738*BI2/(BI2+2. )+0.2795*ATAN(BI2/3.)-0.02915*ATAN(5.*BI2) 132 S2=S2+0.001171 133 S3=2.0738*B13/(BI3+2.)+0.2795*ATAN(BI3/3.)-0.02915*ATAN(5.*BI3) 134 S3=S3+0.001171 \H FSUBH=2.303/(DIFF*(S1/A1**2.+S2/A2»*2.+53/A3**2.)) 137 IF(FH.LT.FSUBH) GOTO 400 C Loop around again to further tune the difference to <1. 138 139 140 141 425 HTC«HTC-1. 142 BI1=HTC*A 1/COND 143 BI2=HTC*A2/C0ND 144 BI3=HTC*A3/COND 145 S1=2.0738*BI1/(BI1 + 2. )+0.2795*ATAN(B11/3.)-0.02915*ATAN( 5 . *BI 1 ) 146 S1=S1+0.001171 147 S2=2.0738*BI2/(BI2+2. )+0.2795*ATAN(BI2/3.)-0.02915*ATAN(5.*BI2) 148 S2=S2*0.001171 149 S3=2.0738*BI3/(BI3+2.)*0.2795*ATAN(BI3/3.)-0.02915*ATAN(5.*BI3) 150 S3*S3+0.001171 151 152 FSUBH = 2.303/(DIFF*(Sl/A1**2.+S2/A2**2.+S3/A3**2. )) 153 IF(FH.GT.FSUBH) GOTO 425 154 155 C Truncate the h value to the nearest 10 giv i n g a 156 C p r e c i s i o n of plus minus 5.0 W/m2C. 157 158 HTC-AINT(HTC/10.+0.5)*10. 159 160 161 C Print out the f i n a l h value 162 163 164 WRITE(6.450) ID.FH.FSUBH,HTC 165 450 F0RMAT(11X.BA4.2F7.2.2X.F7.0) 166 GOTO lOO 167 168 500 WRITE(6,8) 169 GOTO 1 170 171 10O0 STOP 172 END -138-Appendix H . Typical output of the computer program for heat transfer c o e f f i c i e n t evaluation. HEAT TRANSFER COEFFICIENT: LAGARDE: THIN AL BRICK PRODUCT INFORMATION Thickness • 1.930 cm Width « 12.000 cm Length « 17.800 cm Thermal conductivity « 239.000 W/mC Thermal d i f f u s i v i t y - 944.000 X 10**-7 m sq/s Di g l t e c c o r r e c t i o n factor * 1.053 s/stroke Sample ID f-sub-h Input c a l c . h s s W/m2C RUN #1 : 45. 4 119 . 7 C 16. 03 16. 03 3040. RUN 2 : 44 . 2 %S. 118. 8 C 16 . 03 16. 03 3040. RUN 3 : 59. 2 %s. 119 . 3 C 13 . 37 13. 37 3710. RUN 4 : 52 . 9 %s. 120. 3 C 13. 38 13. 39 3700. RUN 5 : 51 . 5 %s. 119. 0 C 12. 49 12. 49 4000. RUN 6 : 57 . 4 %s. 119. 2 C 13. 38 13. 39 3700. RUN 7 : 57. 5 %s. 119 2 C 12 66 12, .66 3940. RUN 8 : 60. .9 %s. 119 . 1 C 11 , .51 11 , .51 4370. RUN 9 : 60 .5 %s. 118 .8 C 12 .24 12 .24 4090. RUN 10: 65 6 y.s. 119 .0 C 10 .52 10 .52 4830. RUN 1 1 : 66 .3 %s. 1 19 .7 C 10 .93 10 .93 4630. RUN 12 : 34 . 2 %s, 103 .3 c 19 .90 19 .90 2410. RUN 13: 33 .9 %s. 103 . 1 c 19 .20 19 .20 2500. RUN 14 : 40 .8 %s. 104 .7 c 17 .64 17 .64 2740. RUN 15 : 41 .6 %s. 105 .3 c 16 .45 16 .45 2960. RUN 16: 44 .5 %s. 104 .7 c 19 .82 19 .82 2420. RUN 17 : 44 .7 %s. 104 .8 c 18 .89 18 .89 2550. RUN 18 : 49 .0 %s. 105 . 1 c 14 .32 14 .32 3440. -139-Appendix I I I . A computer program for studying the temperature d i s t r i b u t i o n 1n a steam/air r e t o r t . 1 £.*»»**.»«»•**.»*»•»»**»********»**»•***********************••*•**•** 2 C 3 C * 4 C This 1s a program to study the temperature d i s t r i b u t i o n in a * 5 C r e t o r t . The program Is designed to handle data from up to 30 * 6 C thermocouples and 200 time-steps. It can be e a s l i l y modified * 7 C to accommodate other requirements. The computer program 1s * 8 C designed to output a table of means and standard deviations * 9 C of temperature at each time step, and for each location * 10 C during the cook period (excluding come-up time), and a grand * 11 C mean and standard deviation in temperature for a l l locations * 12 C during the cook period. These are also graphically presented * 13 C for a better understanding of the temperature d i s t r i b u t i o n . * 14 c * 15 c******************************************************************* 16 17 18 19 C READ DATA 20 21 22 C TITLE : T i t l e for the output up to 80 characters. 23 24 C NTC (number of thermocouples). NTIME (number of time steps) 25 C CUT (come-up time) 26 27 C TCID (thermocouple l . d . ' s ) , TCCF (thermocouple correction factors) 28 29 C TIME (time) and TEMP (temperatures) 30 31 32 DIMENSION TI ME(200).TEMP(200,30),TITLE(20).TT(6000),ST(200). 33 1TM(200).TU(200),TL(200),CM(30),CU(30),CL(30),TCID(30). 34 2SDT(200),SDC(30).ST2(200),SC2(30),TCCF(30),SC(30) 35 INTEGER CUT 36 INTEGER F0RM(19) 37 1 FORMAT(20A4) 38 2 FORMAT(313 ) 3 9 3 F0RMAT( ' ', 40 1 ' -',/) 41 4 FORMAT(1H1,10X,20A4,///) 42 5 FORMAT (6X , ' , 43 & ' ',/. 44 & 30X,'Thermocouple number and correction factor',//, 45 & ' ',15(2X,A4).5X,/.5X.15(2X,A4)) 46 6 F0RMAT(F5.2, 15F6.2,/,5X, 15F6.2) 47 8 F0RMAT(' Mean',15F6.1./.5X,15F6.1) 48 9 FORMAT(' S.D. ', 15F6.2./.5X.15F6.2) 49 10 F0RMAT('Grand mean temperature » '.F6.2.' C ) 50 11 F0RMAT('Standard deviation « '.F6.2,' C ) 51 12 F0RMAT(/) 52 14 FORMAT(IH1) 53 15 F0RMAT(15F5.2./,15F5.2) 54 16 F0RMAT(' T1me'.15F6.1.5X,'( Mean, S.D.)'./.5X,15F6.1) 55 17 F0RMAT(15A4,/15A4) 56 18 FORMAT('Stab 11izat1 on or'/'Retort come - up time • ',12,' min') 57 19 F0RMAT(F5.2.15F6.2) 58 20 FORMAT('Grand mean temperature » '.F6.2,' F') 59 21 F0RMAT('Standard deviation « '.F6.2,' F' ) 60 READ(5,1.END«1000) TITLE 61 READ(5.2) NTC,NTIME.CUT 62 NCUT-CUT+1 63 CALL MAKEFT(FORM,NTC) 64 READ(5.17) (TCID(I).I«1.NTC) 65 READ(5.15) <TCCF(I ) .I«1.NTC) 66 PRINT 4. (TITLE) 67 PRINT 5.(TCID(I),1-1,NTC) 68 PRINT 16,(TCCF(I),I-1.NTC) 69 PRINT 3 70-140-71 72 C I n i t i a l i z e the variables for summation 73 74 75 DO 30 I*1.NTIME 76 TM(I)=0. 77 ST(I)=0. 78 ST2(I)»0. 79 DO 30 d=1.NTC 80 SC(d)"0. 81 SC2(d)-0. 82 30 TEMP(I,d)"0. 83 STT»0. 84 STT2=0. 85 RT=(NTIME-NCUT+1) 86 RC'NTC 87 RTT=RT*RC 88 NTT=RTT 89 90 91 C Read m the time-temperature data 92 93 DO 40 1 = 1,NT IME 94 IF(NTC.GT.15) GOTO 35 95 READ(5.19) TIME(I),(TEMP(I.d),d=1.NTC) 96 GOTO 40 97 35 READ(5,6) TIME(I).(TEMP<I.d).d«1.NTC) 98 40 CONTINUE 99 DO 55 1 = 1 .NT I ME 100 DO 45 d=1.NTC 101 TEMP(I,d)=TEMP(I,J)+TCCF(J) 102 45 ST(I)=ST(I)+TEMP(I,d) 103 TM(I)=ST(I)/RC 104 DO 50 d=1.NTC 105 50 ST2(I)=ST2(I)+(TEMP(I,d)-TM(I))**2 106 SDT(I)=SQRT(ST2(I)/(RC-1.)) 107 WRITE(6,F0RM) TIME(I),(TEMP(I.d).d=1.NTC),TM(I),SDT(I) 108 55 CONTINUE 109 PRINT 3 110 DO 60 I-NCUT.NTIME 111 60 STT = STT+ST(I ) 112 TTM=STT/RTT 113 DO 75 J=1.NTC 114 DO 65 I=NCUT,NTIME 115 65 SC(J)=SC(J)+TEMP(I.0) 116 CM(J)=SC(J)/RT 117 DO 70 I=NCUT.NTIME 118 70 SC2(d)=SC2(J)+(TEMP(I.J)-CM(J))**2 119 SDC(d)=S0RT(SC2(d)/(RT-1.)) 120 75 CONTINUE 121 PRINT 8,(CM(d),d«1.NTC) 122 PRINT 9,(SDC(d) ,d«1,NTC) 123 PRINT 12 124 DO 80 I'NCUT.NTIME 125 DO 80 d=1.NTC 126 80 STT2=STT2+(TEMP(I,d)-TTM)«*2 127 SDTT=S0RT(STT2/(RTT-1.)) 128 IF(TTM.LT.200.) GOTO 85 129 PRINT 20. (TTM) 130 PRINT 21. (SDTT) 131 GOTO 90 132 85 PRINT 10,(TTM) 133 PRINT 11.(SDTT) 134 90 PRINT 18.(CUT) 135 136 137 C P l o t t i n g section 138 139 CALL PLOTS 140 -141-141 C 142 143 144 145 146 147 148 149 150 151 152 100 153 154 110 155 156 157 158 159 160 130 161 162 163 164 165 140 166 167 168 169 150 170 17 1 C 172 C 173 174 175 176 177 160 178 170 179 180 181 182 183 184 185 186 187 188 189 180 190 191 192 193 194 195 196 197 190 198 200 199 205 200 201 202 203 204 210 205 220 206 Seal l i n g the axes IXKEL»((FL0AT(NTIME-NCUT)/10.)+0.9) XSKEL«FLOAT(IXKEL) IMEAN=TTM IF(IMEAN.GT.200) GOTO 100 YMIN=FLOAT (IMEAN)-2.0 YSKEL-0.5 GOTO 110 YMIN= F LOAT(IMEAN)- 10.0 YSKEL-2.0 XMIN=TIME(NCUT) DO 130 I=NCUT,NTIME TIME(I)=(TIME(I)-XMIN)/XSKEL TU(I)=((TM(I)+SDT(I))-YMIN)/YSKEL TL(I)=((TM(I)-SDT(I))-YMIN)/YSKEL TM(I)=(TM(I)-YMIN)/YSKEL CONTINUE DO 140 J =1.NTC CU(J ) = ( (CM(Jl+SDC(J))-YMIN)/YSKEL CL(J)=(( CM( J)-SDC(J) )-YMIN)/YSKEL CM(J)»(CM(J)-YMIN)/YSKEL CONTINUE DO 150 I=NCUT,NTIME DO 150 J =1,NTC TEMP(I, J) = (TEMP(I.J)-YMIN)/YSKEL CONTINUE Plot *1. Time - temperature data from a l l thermocouples a f t e r the come-up period. IF(IMEAN.GT.200)GOTO 160 CALL AXIS(0..0.,13HTEMPERATURE,C.13,10.0.90.,YMIN,YSKEL) GOTO 170 CALL AXIS(0..0..13HTEMPERATURE,F,13,10.0,90..YMIN,YSKEL) CALL AXIS(0..0.,12HTIME.MINUTES.-12,10..0..XMIN.XSKEL) TTM=(TTM-YMIN)/YSKEL CALL PLOT(0..TTM.+3) CALL PLOT(10..TTM.+2) CALL PL0T(1..10..+3) CALL SYMBOL(1..10., . 14,TITLE.0.,80) DO 180 d=1.15 IF(TEMP(1,u).EO.O.) GO TO 200 RJ=10.*d/15. RSJ=RJ-.1 CALL SYMB0L(RSJ,9.6,.07,0,0..-1) CALL SYMBOL(RJ.9.5,.14,TCID(J),0.,4) IF(NTC.LE.15) GOTO 205 DO 190 J-16.NTC IF(TEMP(I.J).EO.O.) GO TO 200 RJ=10.*(J-15)/15. RSJ=RJ-.1 CALL SYMBOL (RSJ.9.1..07.J.O.,-1) CALL SYMBOL(RJ.9.,.14,TCID(J).0.4) CONTINUE CONTINUE DO 210 J-1.NTC IF(TEMP(NCUT,J) EO.O.) GO TO 220 CALL SYMBOL(TIME(NCUT),TEMP(NCUT,J)..07.J.0..-1) K=NCUT+1 DO 210 I«K.NTIME CALL SYMBOL(TIME(I).TEMP(I.J)..07,J.O..-2) CONTINUE -142-207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 230 240 250 260 Plot H2. Mean and standard deviations for a l l channels at each time Interval a f t e r the come-up period. CALL PL0T(15..0..-3) IFUMEAN.GT.200) GOTO 230 CALL AXIS(0..O..13HTEMPERATURE.C.13.10.0.90..YMIN,YSKEL) GOTO 240 CALL AXIS(0.,0..13HTEMPERATURE.F.13.10.0.90.,YMIN,YSKEL) CALL AXIS(0.,0.,12HTIME.MINUTES,-12,10.,0..XMIN.XSKEL) CALL PLOT(0..TTM.+3) TITLE,0.,80) 'MEAN AND S.D FOR ALL CHANNELS AT EACH TIME' CALL PL0T(10..TTM.+2) CALL SYMBOL(1.,10.,.14 CALL SYMBOL(1.,9.5,.14 1.0..42) CALL SYMBOL(TIME(NCUT ) ,TM(NCUT),.07,2,0..-1) DO 250 I«K.NTIME CALL SYMBOL(TIME(I).TM(I),.07,2,0.,-2) DO 260 I=NCUT.NTIME CALL PL0T(TIME(I).TL(I).+3) CALL PL0T(TIME(I).TU(I).+2) CONTINUE CALL PLOT(15..0..-3) Plot #3. Mean and standard deviations for a l l channels during the cook period (afte r the come-up). CALL SYMBOL(1..10. , . 14 .TITLE,0. ,80) CALL SYMBOL(1.,9.5,.14,'MEAN TEMP & S.D FOR DIFFERENT CHANNELS',0. IF(IMEAN.GT.200) GOTO 270 CALL AXIS(0..0..13HTEMPERATURE,C,13,10.0,90..YMIN,YSKEL) GOTO 280 270 CALL AXIS(0.,0..13HTEMPERATURE,F.13,10.0.90..YMIN.YSKEL) 280 CALL PL0T(0.TTM.+3) CALL PLOT(10..TTM.+2) DO 290 d=1,NTC XPJ*10.*J/NTC XJ=XPJ-.25 PI=3.141592654 YY«8.0*0.5*SIN(PI*J/2.) CALL SYMBOL(XJ,YY,.14,TCID(u).0.,4) CALL SYMBOL(XPJ.CM(J),.07,0.0.,-1) CALL PL0T(XPJ.CL(J).+3) CALL PL0T(XPd.CU(J),+2) 290 CONTINUE XNP=XL*5. CALL PL0T(25..O..-3) C **END OF PLOTTING SECTION** PRINT 14 CALL PLOTND 1000 STOP END ,38) -143-259 2G0 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 SUBROUTINE MAKEFT(FMT.CHNLS) C This routine sets up the format for a prin t statement * C 1n the temperature d i s t r i b u t i o n program. The format changes * C according to the number of channels, since two l i n e s are * C needed If more than 15 channels are used. * C***************************************** ************************* INTEGER FMT(19),CHNLS INTEGER F0RM(19)/'(F5. ' . ' 1. ' '.'F6.1',' '.' '.*F6.1',*.'.' 1 ' (6X) '5X, " ' . ' ( * ' ,F ', '6.2, , "'.'.'.'F5.2'. 2 ' . " ) " ' . ' ) ' / INTEGER NUMB(15)/'1','2','3'.'4','5','6','7','8'.'9'.'10'. 1 ' 11 ' , ' 12','13','14','15'/ INTEGER BLANK/' '/.COMMA/','/,SKIP/'/5X,'/ DO 100 1 = 1 . 19 100 FMT(I) = FORM(I) IF (CHNLS.GT.15) GOTO 300 FMT(3) = NUMB(CHNLS) FMT(5) « COMMA DO 200 1=6.8 200 FMT(I) « BLANK IF (CHNLS.EQ.15) GOTO 400 FMT(9) = NUMB(15-CHNLS) RETURN C Come here i f > 15 channels 300 FMT(3) • NUMB(15) FMT(5) = SKIP FMT(6) = NUMB(CHNLS-15) IF (CHNLS.EO.30) GOTO 400 FMT (9) = NUMBOO-CHNLS) RETURN C Come here i f exactly 15 or 30 channel 400 DO 500 1=9,11 500 FMT(I) • BLANK RETURN END Appendix IV. Surface heat transfer c o e f f i c i e n t s associated with steam/air mixtures in a po s i t i v e flow r e t o r t . Run Steam Temperature Flow Flow Brick h Run Steam Temperature Flow Flow Brick h n content rate d i r e c t i o n type * content rate d i r e c t i o n type % C (scfm) W/m2C % C (scfm) W/m2C 1 50. 3 105. 2 40 up al -tn 1270 2 51 . 4 105. 7 60 up al -tn 1550 3 49 . 6 1 15. 3 40 up a 1 -tn 12B0 4 50. 9 1 16 . 1 60 up al -tn 1 150 5 51 . 0 121 . 1 40 up al -tn 1310 6 54 . 3 121 . 1 60 up al -tn 1640 7 53. 2 125. 8 40 up al -tn 1460 B 65 3 105. 4 40 up a 1 -tn 2420 9 63 5 105 . 1 60 up al -tn 2590 10 64 4 1 14 . 7 40 up al -tn 2210 1 1 62 9 1 14 . 0 60 up a 1 -tn 2460 12 64 7 120 O 40 up al -tn 2200 13 64 6 1 19 8 60 up al -tn 2760 14 67 3 125 1 40 up al -tn 2470 15 65 5 125 2 60 up al -tn 3000 16 85 4 105 4 40 up al -tn 4950 17 85 2 105 2 60 up al -tn 4640 18 81 5 1 14 6 40 up al -tn 4970 19 82 O 1 14 8 60 up al -tn 5550 20 84 9 120 0 40 up a -tn 4810 21 82 7 1 19 7 60 up a -tn 5140 22 87 .6 125 9 40 up a -tn 4780 23 87 .0 125 5 60 up a -tn 5340 24 95 . 2 104 5 40 up a -tn 10870 25 99 .2 105 7 60 up a -tn 1 1380 26 97 .6 1 15 1 40 up a -tn 10960 27 98 .2 1 15 .4 60 up a -tn 10650 28 96 .9 1 19 .9 40 up a 1 -tn 1 1940 29 97 .9 120 . 3 60 up a 1 -tn 11890 30 98 .2 125 .2 40 up a 1-tn 10910 31 97 .6 125 .0 60 up a l- t n 12 140 32 50 .2 104 .8 40 up a 1-tn 1 180 33 49 .2 109 .6 40 up a 1-tn 1280 34 49 .5 1 14 .6 40 up a 1-tn 1390 35 50 .9 120 . 4 40 up a 1-tn 1470 36 62 2 105 O 40 up al -tn 2290 37 67 3 1 10 9 40 up al -tn 2330 38 65 3 1 14 8 40 up al -tn 2180 39 62 4 120 0 40 up al - tn 2040 40 81 8 105 0 40 up al -tn 4970 41 83 6 109 8 40 up al - tn 4300 42 84 3 1 14 6 40 up al -tn 4340 43 B2 9 1 19 6 40 up al - tn 4430 44 99 7 105 2 40 up al -tn 12810 45 99 8 1 10 1 40 up al -tn 1 1580 46 100 0 1 15 5 40 up al - tn 1 1080 47 100 O 120 5 40 up a 1 - tn 1 1320 48 50 0 104 9 40 down a 1 - tn 1530 49 48 8 109 4 40 down al - tn 1530 50 49 9 1 14 7 40 down a 1 - tn 1590 51 49 0 1 19 1 40 down al -tn 1620 52 64 7 104 8 40 down al -tn 2950 53 62 8 1 10 0 40 down al - tn 3230 54 64 9 1 15 1 40 down al - tn 31 10 55 63 2 120 0 40 down al - tn 2980 56 84 4 104 9 40 down al - tn 5570 57 83 9 109 8 40 down al - tn 5500 58 85 4 1 15 1 40 down al - tn 5430 59 86 0 120 2 40 down a 1 -tn 5630 60 95 6 105 0 40 down al - tn 12080 61 96 1 1 10 0 40 down al -tn 13160 62 98 6 1 14 8 40 down a 1 - tn 12970 63 100 0 120 3 40 down al -tn 1 1880 64 48 7 104 3 40 up al - tn 1 160 65 54 9 105 0 40 down al -tn 1820 66 60 6 105 2 40 up al -tn 1790 67 63 5 104 6 40 down al - tn 2870 68 69 8 105 1 40 up al - tn 2680 69 74 0 104 9 40 down a 1 - tn 3560 70 79 .8 105 3 40 up a 1 - tn 4090 Note symbols used: Aluminum ( a l ) . steel ( s t ) , thin (tn) and thick (tk) Appendix IV. cont i nued Run Steam Temperature Flow Flow Brick h Run Steam Temperature Flow F'°« B r ^ k h 0 content rate d i r e c t i o n type » content rate d i r e c t i o n type % C (scfm) W/m2C % C (scfm) W/m2C 71 84 .9 105. 1 40 down al -tn 72 89.3 105. 1 40 up al -tn 73 94 .4 104.9 40 down al -tn 74 98 .0 104 . 7 40 up al -tn 75 55.4 120. 2 40 up a l - t n 76 60.4 120. 1 40 down al -tn 77 64.2 1 19 .6 40 up a l - t n 78 70. 2 120.0 40 down a l - t n 79 75.4 120.0 40 up al -tn 80 79.7 119.8 40 down a l - t n 81 84 .9 120.0 40 up a l - t n 82 89.4 119.7 40 down a l - t n 83 94 .1 119.7 40 up a l - t n 84 99.3 119.9 40 down a l - t n 85 49.7 104 .7 40 up st - t n 86 54 .9 104 .8 40 down st-t n 87 59.0 104 . 5 40 up st - t n 88 65. 1 105 . 3 40 down st-t n 89 69. 1 104 .8 40 up st - t n 90 74 .4 104 . 7 40 down st - t n 91 79.7 104 .9 40 up st - t n 92 85 .8 105 .4 40 down st - t n 93 89.6 105.2 40 up st - t n 94 92.4 104 .6 40 down st - t n 95 100.0 105.3 40 up s t - t n 96 56.0 120.5 40 up st - t n 97 59.4 119.8 40 down st - t n 98 65.3 120. 1 40 up st - t n 99 70.0 119.8 40 down st - t n 100 74.3 1 19.8 40 up st - t n 101 80.6 120. 3 40 down st - t n 102 85.7 120.3 40 up st - t n 103 89.7 120. 2 40 down st - t n 104 96.3 120.3 40 up st - t n 105 98.9 120.0 40 down st - t n 5470 5630 9320 1 1880 1580 2350 2030 3610 3300 5140 5000 71 10 8270 12660 1270 24 10 2060 3990 2810 4650 4 130 7 160 6800 9660 9310 1570 2880 2490 4020 3590 5200 5480 8590 9090 10270 106 107 108 109 I 10 II 1 112 113 1 14 115 1 16 1 17 118 1 19 120 121 122 123 124 125 126 127 128 129 130 13 1 132 133 134 135 136 137 138 139 140 99. 7 105. 2 40 up s t - t n 10600 99. 0 105. 0 40 down st - t n 13130 95. 0 105. 1 40 up s t - t n 8000 94 . 1 104 . 8 40 down st - t n 8470 91 . 3 105 . 0 40 up s t - t n 7630 89. 9 104 . 7 40 down st - t n 7910 88 . 9 105. 0 40 up s t - t n 5860 87 . 9 104 . 8 40 down st - t n 7120 84 . 4 104 . 9 40 up s t - t n 5130 83 . 8 104 . 7 40 down st - t n 6930 80. 0 105. 0 40 up s t - t n 4 160 79. 5 104 . 8 40 down st - t n 6200 76 . 1 105. 2 40 up s t - t n 3370 73 . 2 104 . 7 40 down st - t n 5120 77 . 2 105 0 40 up s t - t n 3740 76. 4 104 . 7 40 down st - t n 5650 72 8 105 . 1 40 up s t - t n 3310 70 . 1 104 .9 40 down s t - t n 41 10 73 . 4 105 .4 40 up s t - t n 3540 7 1 .4 104 .7 40 down st - t n 4820 66 . 1 105 . 2 40 up s t - t n 2760 64 . 7 105 .0 40 down st - t n 4220 60 . 4 105 . 1 40 up s t - t n 2 160 59 .2 104 . 7 40 down st - t n 3320 56 . 7 104 .6 40 up s t - t n 2030 56 .8 104 .5 40 down st - t n 3280 54 . 1 105 . 1 40 up. s t - t n 1700 53 .9 104 .8 40 down s t - t n 2410 50 .9 105 .2 40 up s t - t n 1520 49 .2 104 .5 40 down s t - t n 2360 80 . 2 104 .8 40 up al -tk 3990 68 .5 104 . 1 40 up al -tk 2660 59 .6 104 .8 40 up al -tk 2050 49 .2 104 9 40 up al -tk 1300 98 .0 104 . 7 40 up al -tk 8420 Note symbols used: Aluminum ( a l ) . steel ( s t ) . thin (tn) and thick (tk) Appendix IV. contInued "Ron""steam"Te mper ature Flow Flow Brick h Run Steam Temperature Flow ^ "type h , content rate d i r e c t i o n type * content rate d i r e c t i o n type % C (scfm) W/*2C % C (scfm) 141 97. 6 104 . 6 40 up al-tk 8230 142 78 . 8 105. 1 40 up al-tk 3700 143 87 . 7 104 . 9 40 up al-tk 5410 144 48. 8 104 . 5 40 up al-tk 1430 145 49. 3 104 . 7 40 down al-tk 1640 146 59. 8 104 . 9 40 up al-tk 2180 147 59. 4 104 . 7 40 down al-tk 3140 148 68. 8 104 . 7 40 up al-tk 2890 149 68. 8 104 . 7 40 down al-tk 3350 150 78. 8 104 . 7 40 up al-tk 3730 151 78. 9 105. 0 40 down al-tk 5440 152 88. 9 105. 0 40 up al-tk 6080 153 86. 5 104 . 2 40 down al-tk 6860 154 98. 4 105. 0 40 up al-tk 9400 155 98 . 8 105 1 40 down al-tk 9600 156 54 . 4 105 0 40 up al-tk 1640 157 53 7 104 .6 40 down al-tk 2050 158 49 .9 104 .8 40 up al-tk 1940 159 64 .9 104 .9 40 up al-tk 2610 160 62 4 104 .4 40 down al-tk 3610 161 75 .5 105 .0 40 up al-tk 3760 162 74 .3 104 .9 40 down al-tk 4790 163 78 .9 104 .6 40 down al-tk 5630 164 85 .2 105 .2 40 up al-tk 4940 165 84 .4 104 .9 40 down al-tk 74 10 166 87 .0 104 .5 40 down al-tk 7510 167 94 .0 105 . 1 40 up al-tk 7440 168 93 .4 104 .9 40 down al-tk B350 169 99 .0 105 .0 40 down al-tk 9460 170 49 .7 104 .7 40 up st-tk 1040 171 50 .6 105 .2 40 down st-tk 2600 172 53 .7 104 .5 40 up st-tk 1 100 173 54 .4 104 .7 40 down st-tk 27 10 174 59 .2 104 .9 40 up st-tk 1470 175 60 .4 105 .0 40 down st-tk 4 130 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 .3 .2 .4 .7 .4 .9 65.4 64 .9 70. 1 71.1 75.8 74 .9 80 79 84 84 88 86 95.0 93.7 94.4 100.0 100.0 lOO.O 100.0 lOO.O 100.0 100.0 100.0 84.4 84 . 1 84.4 84 .6 82 .9 72 72 73 72 72 73 .9 . 2 1 .5 1 .9 1 104 105 104 105. 105 . 104 105 105. 1 104 .9 105.3 105. 1 104 .9 105.4 105.0 105.2 105.2 105.2 ,105 . 2 105.2 105. 1 71.3 105. 1 105 . 1 105 . 1 104 .9 104 .8 104 .9 104 104 104 104 105 104 104 104 .8 104.3 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 38 25 25 13 12 58 58 50 47 29 30 14 up down up down up down up down up down up down up up down up up up up down down down down up up up up up up up up up up up up st-tk st-tk st-tk st-tk st-tk st-tk st-tk St-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk st-tk al-tk al-tk a l - t k a l - t k al-tk a l - t k a l - t k a l - t k al-tk a l - t k al-tk al-tk 1870 44 10 2110 4420 3460 4940 3500 7010 5000 7870 5620 910O 9120 9130 10900 12620 12620 12620 12620 14930 14930 14930 14930 5840 5210 5690 4130 3230 4150 4140 3870 3780 3530 3340 2220 Note symbols used: Aluminum ( a l ) . steel ( s t ) . thin (tn) and thick (tk) Appendix IV. contInued Run # Steam content % Temperature C Flow rate (scfm) F 1 ow d 1rect1 on Brick type h W/m2C Run » Steam content % Temperature C F low rate (scfm) Flow di rectIon Brick type h W/m2C 211 212 70. 62 . 3 8 103. 104 . 8 6 13 72 up up al-tk al-tk 2 1 10 3060 235 236 84 . 83 . 1 9 1 19. 119 . 5 6 25 24 up up al - t k a l - t k a l - t k a l - t k 5380 5100 . 3480 4080 213 214 63. 63. 9 3 105 . 104 . 1 6 74 73 up up al-tk al-tk 3260 3240 237 238 83 . 85. 6 2 119. 119. 3 9 12 13 up up 215 63. 1 104 . 5 64 up al-tk 2970 239 63. 0 104 . 9 69 down al-tk 5500 216 63. 1 104 . 6 64 up al-tk 3150 240 63 . 3 105 . 0 70 down al- t k 5540 217 63. 7 104 . 8 54 up al-tk 2470 24 1 63 . 3 104 . 8 64 down al-tk 5340 218 63. 3 104 6 54 up al-tk 2680 242 63. 7 105 . 0 65 down al-tk 5600 219 63. 1 104 . 6 43 up al-tk 2460 243 63. 7 104 . 9 54 down al-tk 4990 220 63. 3 104 7 43 up al-tk 2500 244 63. 9 105 . 1 55 down al-tk 5350 221 63. 3 104 .6 32 up al-tk 2230 245 63 . 0 104 . 9 43 down al- t k 4060 222 63. 1 104 .6 32 up al-tk 21 10 246 64 . 2 105 , .2 44 down al-tk 4240 223 62 . 9 104 .4 21 up al-tk 1780 247 62 8 104 .8 32 down al-tk 3500 224 63. 3 104 .5 21 up al-tk 1840 248 63 .7 105 . 1 33 down al-tk 3580 225 63. 1 104 .5 1 1 up al-tk 1 190 249 62 .2 104 .5 21 down al- t k 3240 226 63 6 104 .5 1 1 up al-tk 1460 250 63 .5 104 .9 22 down al- t k 2840 227 81 8 1 18 .8 65 up al-tk 5860 251 62 .4 104 . 4 10 down al-tk 2690 228 83 .9 1 19 . 7 73 up al-tk 6450 252 63 . 7 105 .0 1 1 down al-tk 2570 229 84 . 1 1 19 .7 62 up al-tk 6370 253 73 .9 105 .0 53 down al-tk 5940 230 84 .7 1 19 .9 64 up al-tk 6510 254 73 .9 105 O 53 down al- t k 5790 231 84 .4 1 19 .6 51 up al-tk 6260 255 73 .9 105 .0 30 down al- t k 4490 232 83 .9 1 19 .6 49 up al-tk 6150 256 73 .8 104 . 9 30 down al-tk 4690 233 83 .3 1 19 .4 35 up al-tk 5370 257 73 .2 104 .7 15 down al- t k 3910 234 84 .4 1 19 .8 38 up al-tk 5860 258 73 .2 104 . 7 15 down al- t k 3960 Note symbols used: Aluminum ( a l ) , steel ( s t ) , thin (tn) and thick (tk) -148-Appendix V. Regression d e t a i l s for the surface heat transfer c o e f f i c i e n t 1n the p o s i t i v e flow r e t o r t . P a r t i c u l a r s Medium flow n Slope Intercept r* Brick-type Temp D i r e c t i o n Rate C (scfm) Alumi num Thin Thick 105 Up 40 14 0. 0458 4 . 78 0. 980 120 Up 40 13 0. 0435 4 . 92 0. 978 105 Down 40 9 0. 04 11 5. 26 0. 974 120 Down 40 9 0. 0390 5. 45 0. 986 105 Up 60 4 0. 0393 5. 30 0. 970 120 Up 60 4 0. 0439 5. 03 0. 990 105-125 Up 40 43 0. 0438 4 . 90 0. 976 105-125 Down 40 26 0. 0404 5. ,35 0. 972 105-125 Up 60 15 0. .0430 5. .06 0. 972 105-125 Up 40 20 0 .0355 5, .53 0. 978 105-125 Down 40 14 0 .0339 5 .90 0. 962 105 Up 40 21 0 .0384 5 .33 0 988 120 Up 40 5 0 .0426 5 .OO 0 . 99B 105 Down 40 21 0 .0319 6 . 17 0 .968 120 Down 40 5 0 .0334 5 .96 0 .982 105-125 Up 40 26 0 .0391 5 .28 0 .988 105-125 Down 40 25 0 .0318 6 . 16 0 .964 105-125 Up 40 12 0 .0525 4 . 15 0 .974 105-125 Down 40 11 0 .0314 6 .28 0 .962 Regression equation: ln(h) - Intercept + [Slope X Steam content (%)] -149-Appendlx VI. Covariance analyses of the exponential r e l a t i o n s h i p between heat transfer c o e f f i c i e n t and steam content under various conditions 1n the p o s i t i v e flow r e t o r t . P a r t i c u l a r s of Comparison Level Slope Overall Block Temp D i r e c t i o n Rate C (scfm) A1um i num Thin 105 vs 120 Up 40 ns ns ns Thin 105 vs 120 Down 40 ns ns ns Thin 105 vs 120 Up 60 ns ns ns Thin 105-125 Up 40 vs 60 * * ns ns Thin 105-125 Up vs Down 40 • • ns * * Thick 105-125 Up vs Down 40 * * ns Thin vs * * Thick 105-125 Up 40 ns * * Thin vs Thick 105-125 Down 40 * * * * * * Thin/thick 105-125 Up vs Down 40 • * ns * * Steel Thin 105 vs 120 Up 40 ns ns ns Thin 105 vs 120 Down 40 ns ns ns Thin 105-125 Up vs Down 40 * * * * * * Thick 105 Up vs Down 40 • * * * * * Thin vs * * Thick 105-125 Up 40 * * * * Thin vs Thick 105-125 Down 40 » * ns ns Th1ck/th1n 105-125 Up vs Down 40 * * » » • * Aluminum vs Steel 105-125 Up 40 ns ns ns 105-125 Down 40 • * • • * * Overa11 105-125 Up vs Down 40 * * • • » * ns not s i g n i f i c a n t at p=0.05 ** s i g n i f i c a n t at p< 0.05 -150-AppencHx VII. Surface heat transfer c o e f f i c i e n t s associated with steam/air mixtures in the Lagarde r e t o r t . Run Steam Temperature h Run Steam Temperature h * content * content % C W/m2C % C W/m2C 1 88. 4 110. 6 7590 2 87. 6 1 10. 0 7960 3 87 . 3 109. 9 6740 4 56. 0 1 10. 2 3200 5 56. 0 110. 2 3220 6 55. 8 110. 1 3430 7 39. 2 109 . 9 2230 8 39. 8 110. 3 2640 9 41 . 0 111. 2 2790 10 74 . 2 110. .5 5820 1 1 72. 8 1 10. 1 4980 12 73. ,5 110. 4 4950 13 75. 1 119. 9 5060 14 75. 6 120. 1 5230 15 75 . 1 119. 9 4870 16 97. 2 120. 8 9490 17 95. 9 120. 4 10120 18 95. 9 120. 4 8180 19 56. 8 120. 2 3220 20 58. 2 120. .5 3630 21 56 . 8 120. 2 3980 22 91 . 0 128. .7 8260 23 88 . 8 127. 9 6060 24 90 .4 128 . 7 7810 25 70. 3 127. 8 5920 26 71 . 6 128. .2 4470 27 71 . 8 128. .3 4630 28 70. 3 128. .9 5120 29 61 . 3 127 . 7 4610 30 62 .2 128 .2 5320 31 79 . 6 128 . 7 6060 32 74 .0 126 .2 5450 33 46 . 6 1 18 9 2760 34 65 .6 119 .9 4150 35 65 . 6 119, .9 3740 36 80, .7 119 .4 5580 37 79. 9 119. . 1 4680 38 74 .4 124 .3 5200 39 74 . 2 123 .8 5540 40 77 O 1 14 .8 5270 41 77. .0 1 14 , .8 5560 42 45 .4 119 .7 3520 43 44 . 2 1 18 , .8 3040 44 59 .2 1 19 .3 3700 45 52 . 9 120 .3 3700 46 51 .5 119 .0 3990 47 57 . 4 119 .2 3700 48 57 .5 119 .2 3930 49 60. 9 1 19 . 1 4370 50 60 .5 118 .8 4080 51 65 .6 119 .0 4830 52 66 .3 119 .7 4630 53 34 . 2 103 . 3 2400 54 33 .9 103 . 1 2500 55 40 8 104 .7 2740 56 41 .6 105 .3 2950 57 44 .5 104 .7 2410 58 44 .7 104 .8 2540 59 49 .0 105 . 1 3430 60 48 .4 104 .7 2880 61 52 .0 104 . 2 3390 62 51 .8 104 .0 3530 63 54 .9 104 .5 3310 64 55 .0 104 .6 3520 65 65 .0 104 .8 5450 66 74 .7 104 .8 4780 67 75 . 2 104 .6 5610 68 86 .8 104 .8 6380 69 84 . 7 104 .6 6880 70 98 .0 104 .4 9140 71 75 .0 120 .8 5670 72 71 .4 119 . 1 5470 73 82 .2 1 19 .6 6480 74 80 .7 1 19 .0 6580 75 73 .6 1 10 .7 5440 76 55 .3 104 .7 3400 77 55 .0 104 .6 3500 78 51 .8 104 .6 2980 79 51 .5 104 .5 3210 80 46 .7 104 .2 2850 81 47 .4 104 .6 3320 82 43 .8 104 .4 2930 83 43 .7 103 .9 3160 84 40 .4 104 .6 3610 85 40 . 5 104 .6 3290 86 35 .8 104 .3 2480 87 35 . 1 104 .2 2530 88 56 .4 118 .4 4250 89 73 .6 1 10 .7 5440 90 63 .9 106 .9 3800 -151-Appendlx VII. .continued Run Steam Temperature h H content % C W/m2C Run Steam Temperature h * content % C W/m2C 91 70. 0 106. 8 4840 92 78. 3 106. 1 6720 93 94 . 2 106. 1 9060 94 59. 8 105. 9 3420 95 61 . 7 105. 7 3970 96 65. 7 105. 7 4180 97 73 . 0 105. 5 4330 98 77. 5 105. 5 4510 99 81 . 1 105. 5 4920 100 89. 6 105. 8 5390 101 90. 0 105. 6 5560 102 44 . 3 118. 8 3160 103 44 . 0 1 18 . 5 3440 104 49 . 3 11B. 5 3470 105 49 . 1 1 18 . 3 3250 106 54 . 4 118. 3 3660 107 54 . 1 118. 2 3450 108 62. 2 119. 6 4220 109 61 . 2 1 19 . 1 3310 110 64. 9 118 . 9 4190 1 1 1 65 . 0 118. 9 4820 112 34 . 9 104 . 2 2180 113 34 . 9 104 . 2 2230 114 37. 2 104. 4 2270 115 37 . 1 104 . 3 2340 1 16 43 . 8 104 . 0 2370 1 17 44 . 3 104 . 4 2620 118 47 . 6 104 . 1 2720 1 19 48 . 0 104 . 4 2700 120 51 . 8 104. 2 2820 121 51 . 8 104 . 1 3090 122 53. 9 104 . 3 3090 123 54 . 0 104 . 3 3150 124 96. 8 103. 5 10500 125 97 . 7 103 . 8 12460 126 81 . 0 104 . 5 7460 127 82 3 104 . 8 6790 128 72. 9 104. 8 5520 129 73 . 4 104 . 4 4920 130 63 .5 104 , .7 5050 131 63 .9 104 . 6 3960 132 60 .0 104, .9 3880 133 98 . 2 119. .7 11 1 10 134 96 .6 1 18 .9 13030 135 84 . 3 119 .7 7580 136 84 .3 119 .3 7340 137 75 .2 119 .6 5640 138 74 .5 119 .3 6000 139 53 .9 104 .3 2520 140 51 .5 104 .0 2740 141 47 .8 104 . 1 2130 142 36 .7 104 .0 2440 143 44 . 4 104 . 1 3010 144 40 .0 104 .5 2930 145 34 .8 104 . 1 2710 146 39 . 1 104 .2 2550 147 75 .6 104 .8 5190 148 86 .0 105 .0 9460 149 85 .6 104 .9 10210 150 99 .5 104 .3 10860 151 97 .6 103 .8 10070 152 36 . 1 104 .9 2080 153 36 . 2 104 .9 1940 154 40 .5 104 .5 1920 155 40 . 2 104 .4 2370 156 43 .7 104 .4 2520 157 44 . 5 104 .7 2300 158 47 .5 104 .2 2560 159 47 . 5 104 . 2 2370 160 51 .6 104 . 1 31 10 161 51 .3 104 .0 3050 162 54 .9 104 .4 2780 163 54 . 4 104 . 1 2910 164 63 .9 118 .6 4020 165 64 .5 1 18 .8 4330 166 62 .6 118 .7 4350 167 60 . 1 1 18 .7 4270 168 59 .9 118 .6 4340 169 54 .4 1 18 .7 3610 170 54 .7 1 18 .7 3280 171 48 .9 118 .3 3000 172 49 .0 118 .5 3380 173 43 .8 1 18 .3 2900 174 43 .7 118 .3 2790 175 74 .9 119 .2 4770 176 75 .4 1 19 .4 4370 177 84 .6 119 .8 5580 178 84 .6 119 .4 5290 179 99 .0 1 19 .7 7440 180 97 .9 119 .3 8080 181 60 .8 105 .O 3060 182 98 .6 104 . 1 7510 -152-Appendix VIII. Regression d e t a i l s for the surface heat transfer c o e f f i c i e n t in the Lagarde r e t o r t . P a r t i c u l a r s Brick-type Temp Brick o r i e n t a t i o n n C O Slope Intercept A 1um1num Thin 105-110 120-125 Perp+Parallei 1 Perp+Parallei 1 33 42 0. 0. 0214 0199 6. 7 . 99 1 1 0. 0. 953 895 105-130 105-130 105-130 Perpendlcular' P a r a l l e i ' P a r a l l e i ' 34 18 41 0. 0. 0. 0207 0200 0223 7. 7 . 6 06 13 91 0. 0. 0. 937 888 911 105 Hor i z o n t a l 1 8 0 0132 7 42 0. 935 105-130 Perp.+para1lei 1 93 0 ,0205 7 07 0. 915 Thick 105-110 120-125 Perpendicular' Perpendicular' 21 16 0 0 .0268 .0248 6 6 .67 .86 0 0 973 932 105-130 Perpendicular' 37 0 .0263 6 .72 0. .956 105-130 P a r a l l e i ' 13 0 .0254 6 .79 0 .903 105-130 Perp.+parallei' 50 0 .0260 6 .75 0 .937 Steel 105-130 Perpend i c u l a r ' 31 0 .0204 6 .95 0 .921 Regression equation: ln(h) « Intercept + [Slope X Steam content (%)] 1 combined data from 1981 and 1982 studies ' data from 1981 study ' data from 1982 study -153-AppencHx IX. Covariance analyses of the exponential r e l a t i o n s h i p between surface heat transfer c o e f f i c i e n t and steam content under various test conditions in the Lagarde r e t o r t . P a r t i c u l a r s of Comparison Block Temp C O Brick o r i e n t a t i o n Variance Level Slope Overall A1um i num Thin 105 vs 1 2 0 Perp. ns ns ns ns Thin 105 - 125 Perp. i vs P a r i . s ns • ns ns Thin 105 - 125 Pari . t vs P a r i . ] ns * ns ns Thin 105 - 125 Perp. i vs Horz. 1 * • * • Thin 1 0 5 - 125 Pari . i vs Horz. 1 * * * * Thick 105 vs 125 Perp. 3 ns ns ns ns Thick 105 - 125 Perp. J vs Pa r i . 3 ns ns ns ns Thin vs Thick 105 - 125 Perp. 1 ns ns * * 105 - 125 Pari . 1 ns ns ns ns 105 - 125 Perp. 1 + Pari . 1 1 ns ns ns ns Steel 105 vs 1 2 0 Perp. 3 ns ns * • A 1um i num vs Steel Thin 105 - 125 Perp. 1 ns • ns • Thick 105 - 125 Perp. 1 ns * ns * ns not s i g n i f i c a n t at p • 0 . 1 0 * s i g n i f i c a n t at p < 0 . 0 5 ' combined data from 1981 and 1982 studies ' data from 1981 study ' data from 1 9 8 2 study Perp. pependlcular brick o r i e n t a t i o n P a r i , p a r a l l e l brick o r i e n t a t i o n Appendix X. Typical computer output of the temperature d i s t r i b u t i o n program. TEMPERATURE DISTRIBUTION RESULTS: RUN #95. 120 C. 85% S. UP. HORIZONTAL (PAR) Thermocouple number and correction factor Ch01 Ch02 Ch03 Ch04 ChOS Ch06 Ch07 Ch08 Ch09 Ch10 Ch12 Ch13 Ch14 Ch15 Ch16 Time -0. 9 -0. 9 -0. 9 -0. 9 -0. 9 - 1 . 0 -0. 9 -0. 9 -0. 9 -0. 9 -0. 9 -0. 8 -0. 9 -0. 9 -0. 9 ( Mean, S. D.) 0 0 35. 0 32 2 36. 7 40. 5 31 . 5 35. 9 32. 4 32 . 2 29. 3 34 . 7 37 . 4 32 . 4 37. 7 47. 0 61 4 ( 37. 09. 8. 00) 1 0 88 . 6 70. 9 89. 3 88. 0 85 9 86. 8 85. 4 88. 4 82. 4 85. 5 88. 3 90. 9 75. 1 82 2 93 8 ( 85. 43. 5. 91) 2 0 108. 9 104 9 109. 0 108. 8 108 5 108 . 4 1 10. 0 109. 0 109. 3 110. 6 111. 4 112. 0 107 2 107. 5 112 0 ( 109. 17. 1. 89) 3 0 117. 0 116. 3 1 17 . 7 117. 5 1 17 5 117. 4 117. 4 1 17 6 117 3 117. 4 117. 8 118 0 1 16 9 117 0 117 9 (117. 38. 0.44) 4 0 119. 7 1 19 4 1 19 7 119. 6 1 19 6 1 19. 4 1 19. 5 1 19 5 119 4 119. 4 1 19. 7 1 19 8 1 19 3 119 3 119 7 (119. 53. 0. 16) 5 0 120. 0 1 19 8 120 0 119. 9 1 19 9 1 19. 8 1 19. 8 1 19 8 119 7 119. 8 120 0 120 1 1 19 7 119 8 120 0 (119. 87. 0. 12) 6 0 120. 0 1 19 9 120 0 119 9 119 9 1 19 8 1 19 9 1 19 9 1 19 8 119 8 120 1 120 1 1 19 8 1 19 9 120 1 ( 119 93. 0. 1 1 ; 7 0 120 0 1 19 9 120 o 120 0 1 19 9 119 8 1 19 9 1 19 9 119 8 1 19 8 120 1 120 1 1 19 8 1 19 9 120 0 ( 119 93. 0. 10) 8 0 120 1 120 0 120 1 120 1 120 0 1 19 9 120 0 120 0 119 9 119 9 120 2 120 3 1 19 9 120 0 120 2 ( 120 04. 0. 12) 09 0 120 3 120 2 120 2 120 2 120 2 120 O 120 1 120 1 120 0 120 0 120 3 120 3 120 0 120 1 120 2 ( 120 15, 0. 11 ) 10 O 120 0 1 19 9 120 0 120 0 1 19 9 119 8 1 19 9 1 19 9 1 19 8 119 9 120 1 120 1 119 9 1 19 9 120 0 (119 94, 0 09) 1 1 0 120 0 1 19 9 120 o 120 0 1 19 9 1 19 8 119 9 119 9 119 8 119 8 120 1 120 1 119 9 1 19 9 120 0 ( 119 93. 0 10) 12 O 120 2 120 O 120 2 120 1 120 0 119 9 120 0 120 0 119 9 119 9 120 2 120 3 120 0 120 0 120 2 ( 120 06. 0 13) 13 0 120 2 120 1 120 2 120 2 120 2 120 0 120 0 120 1 120 0 120 0 120 3 120 4 120 1 120 1 120 3 ( 120 15. 0 12) 14 O 120 3 120 2 120 3 120 3 120 3 120 1 120 2 120 2 120 1 120 2 120 4 120 4 120 2 120 2 120 4 ( 120 25. 0 10) 15 .0 120 2 120 2 120 2 120 2 120 2 120 0 120 1 120 1 120 0 120 0 120 3 120 3 120 1 120 1 120 2 ( 120 15. 0 10) 16 .0 120 0 120 0 120 O 120 0 120 0 1 19 8 119 9 119 9 1 19 8 1 19 9 120 1 120 1 1 19 9 1 19 9 120 0 ( 1 19 95. 0.09) 17 .0 120 0 1 19 9 120 0 120 0 119 9 119 8 119 9 1 19 9 119 8 1 19 8 120 1 120 1 119 9 119 9 120 0 (119 93. 0 10) 18 .0 120 0 1 19 9 120 0 120 0 1 19 9 119 8 119 9 119 9 119 8 119 8 120 1 120 1 119 9 119 9 120 0 ( 119 93. 0 10) 19 .0 120 0 119 9 120 0 120 0 119 9 119 8 119 9 119 9 119 8 119 9 120 1 120 1 1 19 9 1 19 9 120 0 ( 119 94. 0.09) 20 .0 120 0 120 0 120 0 120 0 1 19 9 1 19 8 119 9 119 9 119 8 119 9 120 1 120 2 119 9 120 0 120 1 ( 119 97, O 11) 21 .0 120 0 1 19 9 120 0 120 0 120 0 119 8 1 19 9 1 19 9 1 19 8 119 9 120 1 120 2 119 9 120 0 120 0 ( 119 96. 0 11) 22 .0 120 1 120 0 120 1 120 0 120 0 119 9 119 9 119 9 119 9 119 9 120 2 120 3 120 0 120 0 120 1 ( 120 02. 0 12) 23 .O 120 2 120 .0 120 2 120 1 120 0 119 9 120 0 120 0 119 9 119 9 120 2 120 3 120 0 120 0 120 2 ( 120 06. 0 13) 24 .0 120 2 120 1 120 2 120 1 120 1 119 9 120 0 120 0 119 9 120 0 120 2 120 3 120 0 120 1 120 2 ( 120 09. 0 12) 25 .0 120 2 120 . 1 120 2 120 2 120 1 119 9 120 0 120 0 120 0 120 0 120 3 120 3 120 1 120 1 120 2 ( 120 1 1 . 0 12) 26 .0 120 2 120 . 1 120 2 120 2 120 . 1 120 0 120 0 120 0 120 .0 120 .0 120 3 120 3 120 1 120 1 120 2 ( 120 12, 0 11) 27 .0 120 .2 120 .2 120 .2 120 2 120 .2 120 0 120 1 120 . 1 120 .0 120 0 120 3 120 4 120 2 120 2 120 3 ( 120 17. 0 12) 28 .0 120 2 120 .2 120 2 120 .2 120 .2 120 .0 120 1 120 . 1 120 .0 120 .0 120 3 120 4 120 .2 120 2 120 .3 ( 120 17. 0 12) 29 .0 120 .2 120 .2 120 .2 120 .2 120 .2 120 .0 120 .O 120 .0 120 .0 120 .0 120 .3 120 .3 120 .2 120 .2 120 .3 (120 . 15. 0 12) 30 .0 120 2 120 . 1 120 .2 120 2 120 . 1 1 19 .9 120 0 120 .0 120 .0 120 .0 120 3 120 3 120 . 1 120 . 1 120 .2 (120 .11. 0 12) 31 .0 120 .2 120 . 1 120 .2 120 .2 120 . 1 120 .0 120 .0 120 .0 120 .0 120 .0 120 .3 120 .3 120 . 1 120 . 1 120 .2 (120 . 12. 0 .11) 32 .O 120 .2 120 .2 120 .2 120 .2 120 .2 120 .0 120 . 1 120 . 1 120 .0 120 .0 120 .3 120 .4 120 .2 120 .2 120 .3 ( 120 . 17. 0 12) 33 .0 120 .2 120 .2 120 .2 120 .2 120 .2 120 .0 120 . 1 120 . 1 120 .0 120 .0 120 .3 120 .4 120 .2 120 .2 120 .3 ( 120 . 17. o 12) Mean 120 . 1 120 .0 120 . 1 120 . 1 120 . 1 119 .9 120 .0 120 .0 119 .9 119 .9 120 .2 120 .3 120 .0 120 .0 120 .2 S. 0. 0. 11 0. 13 0. 10 0. 11 0. 13 0.09 0.09 0. 10 0. 10 0.09 0. 10 0. 12 0. 14 0. 12 0. 12 Grand mean temperature = 120.05 C Standard deviation » 0.15 C St a b l 1 i z a t i o n or Retort come - up time • 5 min -155-Appendlx XI. f values for s i l i c o n e rubber and nylon bricks' in the p o s i t i v e flow r e t o r t for the f r a c t i o n a l f a c t o r i a l experiments. f value. m1n Run Steam Temp Flow Rack Flow S i l i c o n e rubber nylon H cont. rate type' d i r ' ('/.) C O (scfm) 1 50 105 40 H U 9. 05 9. 81 9. 77 10. 78 10. 35 20 .53 20. 97 1 50 105 40 V D 16 . 38 11 . 69 18. 07 12. 50 1 1 . 13 22 .85 20 .61 3 50 105 20 H D 9. 13 10. 06 9. 91 1 1 . 52 10. 67 23 .71 21 .99 4 50 110 40 V D 15. 76 13. 03 17. 42 10. 64 11. 35 22 .03 22 . 15 5 50 110 40 H D 9. 28 10. 42 10. 29 11 . 14 1 1 . 27 23. .30 22 . 12 6 50 110 20 H U 9. 16 9. 79 10. 08 11 . 70 10. 58 22 .25 22 .99 7 50 120 40 H D 9. 37 10. .22 10. .68 10. .07 10. 68 20. .35 24 .60 8 50 120 40 H U 10. 22 10. 68 10. 07 9. 37 10. 68 20. .35 24 .60 9 50 120 20 V D 17 . 46 17. 31 18. 01 11 . 65 14 . 01 21 . ,37 20, .93 10 65 105 40 H D 9. 59 9 94 9. 89 10. ,79 10. 33 22. ,58 21 , .21 11 65 105 40 V U 14 . 99 10 .98 17 . 98 16. .28 12 . 66 22. .89 21 . 05 12 65 105 20 H D 9. .90 10. .30 10. .26 11 . , 15 10. 37 22. .36 21 . 52 13 65 110 40 V U 13. .54 10 .65 18. .27 13. .07 13. 88 21 .81 21 . 57 14 65 110 40 H D 10 95 10 .37 10 .31 10. .84 10. 73 22. 61 22, .30 15 65 110 20 H D 9. .03 10 .29 10 .31 10. .79 10. 38 21 . 52 22 . 38 16 65 120 40 H D 10 93 10 .97 10 .25 11 . 49 10. 61 21 . 47 24 .71 17 65 120 40 H D 10 85 10 .61 10 .67 11 . 50 10. .61 21 . 62 24 . 26 18 65 120 20 V U 15 .24 11 .70 17 .91 13 .89 12 . 67 21 . 75 23. .08 19 85 105 40 H D 9 .84 9 .77 10 . 17 10 .75 10. 25 22 .63 21 , .48 20 85 105 40 V D 1 1 .85 10 .97 13 .61 12 .01 12. .28 20, .39 20 .58 21 85 105 20 H U 10 . 19 9 .38 9 .87 10 .80 10. 69 22 88 21 .71 22 85 1 10 40 V D 12 .28 10 .39 14 .61 11 .49 13 .36 20 .82 21 .22 23 85 1 10 40 H U 10 .87 9 .40 9 .99 10 .59 10. .32 21 , .03 21 .83 24 85 1 10 20 H D 10 . 18 9 .69 10 .29 10 .61 10. .45 22 .65 22 . 16 25 85 120 40 H U 10 . 12 10 .25 11 . 18 10 .55 10 64 22 .05 24 .01 26 85 120 40 H D 11 .29 9 .77 10 . 15 11 .84 10 .37 22 .35 23 .55 27 85 120 40 V D 15 .29 10 .49 16 .64 17 .46 17. .47 20 .82 22 . 19 1 S i l i c o n e rubber bricks measured 1.9 x 12.1 x 17.8 cm and nylon (thick) bricks measured 2.4 x 12.1 x 17.8; The clearance between the plates 1n the v e r t i c a l rack was 2.4 cm. ' Rack type: V. v e r t i c a l ; H, horizontal ' Medium flow: U, upward; D. downward -156-Appendix XII. f values for s i l i c o n e rubber and nylon b r i c k s ' in the p o s i t i v e flow r e t o r t . Horizontal racking system f va1ue, min Steam Temp. " C O n t . Rubber bricks Rubber bricks packaged Nylon bricks (%) C O 50 105 11 . 41 1 1 . 12 10. 94 10. 98 50 110 11 . 47 1 1 . 32 10. 29 9. 67 50 110 11 . 77 1 1 . 27 1 1 . 03 9. 85 50 120 12 . 10 11 . 42 1 1 . 54 10. 50 65 105 11 . 28 10. 81 10. 44 9. 48 65 110 11 . 04 10. 92 10. 37 9. 94 65 110 11 . 30 11 . 27 10. 55 10. 48 65 120 11 . 52 1 1 . 26 10. .59 10. .29 85 105 10. .45 10. 62 9. .86 9. .31 85 110 11 . 05 11 . 29 10 .96 10 .23 85 110 10 .55 10 .41 10 . 11 8 .80 85 120 10 .97 10 .78 10 .21 9 .40 100 105 10 .57 10 .49 10 . 16 9 .54 100 1 10 10 .64 10 .49 10 .30 9 .42 100 110 11 .24 10 .93 10 . 16 9 .61 100 120 11 . 15 11 .02 10 .92 9 .68 12. 01 1 1 . 05 11 . 32 12. 02 23. 93 21 . 15 1 1 . 04 10. 29 9. 72 1 1 . 02 24. . 14 19 . 79 11 . 74 11. 02 10. 20 12. 32 23. 14 21 . 34 11 . 36 12. 46 11 . 13 11 . ,51 24. .55 22. . 16 22. . 15 24. 01 19. ,74 11 . 20 22. ,39 20 ,74 21 . 06 22. . 11 17. ,95 12. . 14 22. .34 20. .05 19 .84 20 . 11 17. .91 11 . 66 24. ,06 21 . ,33 11 , .54 10 .83 10. .84 11. .67 25. . 16 22. .54 22 .41 27 .06 21 . 68 11 . 17 22 .51 19. 71 22 .07 26 .02 20. .68 14 .42 23, .64 20. .92 23 .04 28 .06 21 .36 10 .74 22 .90 20. .05 21 .66 23 . 16 20 . 14 10 .94 23 .48 20 .97 18 .03 34 .45 25 .54 11 .32 22 .56 20 .21 19 .40 33 .34 22 .74 15 .20 23 .53 21 . 19 20 .01 22 .05 22 .56 11 .90 22 .58 20, .78 21 .43 28 .26 22 .00 11 .55 23 .91 21 .24 V e r t i c a l racking system f value. m1n Steam Temp. C O n t . Rubber bricks Rubber bricks packaged Nylon bricks (%) C O 50 105 13 64 50 1 10 13 , 22 50 1 10 12 .78 50 120 12 . 10 65 105 12 .96 65 1 10 12 .72 65 1 10 12 .84 65 120 13 .20 85 105 12 , .28 85 110 11 .86 85 110 12 .30 B5 120 12 .56 100 105 12 . 13 100 110 12 .01 100 110 12 .50 100 120 12 .04 11 . 24 11 . 29 11 . 55 11 . 45 11 . 52 11 . , 1 1 11 . 11 11 . 29 10. 53 11 . 42 11 . 54 10. 50 11 . 36 11 . 01 10 .89 10. 89 10. 76 9 .58 11 . 03 10. 93 9 .47 11 . 76 11 . 33 11 , . 15 10 .81 10. . 17 9 .53 10 .46 10 . 15 9 .61 11 .06 10 ,48 9 .39 11 . 11 10 .75 9 .51 10 .04 9 .66 8 .92 10 .39 9 .26 9 .38 11 .37 10 .41 9 .22 10 .95 10 .65 9 .95 10. 38 10. 29 11 . 58 10. 35 10. 12 10. 84 10. 85 10. 38 11 . 89 11 . 36 12. 46 11. 13 10. 55 10. 15 10. 49 10. 27 9. 69 9. 93 10. .57 10. 11 10. .57 11 . 18 10. 62 12. 16 io .20 9. .58 10. . 18 9 .68 9. ,25 9 .66 10 .45 9. 56 10 .04 10 65 10. .09 10 . 15 9 .66 9 .41 9 .69 10 .35 9 .67 9 .87 9 .80 9 .53 10 .08 10 .54 9 .90 10 .60 11 . 14 23. 15 23 62 10. 61 23. .53 24. .00 11 . 67 23. .66 25 .02 11 . 51 24. .55 22 . 16 10. 76 22 .76 22 .51 10. . 13 23 . 14 23 .75 10. 83 23. .49 24 .28 11 . 92 24. .86 24 .39 10. 55 22 .68 23 .66 10 .03 23 .22 23 .34 10. .36 22 .93 22 .62 10 .51 23 .96 24 .82 9 .71 22 . 19 22 .22 9 .92 23 . 18 22 . 10 10 . 10 23 . 13 21 .38 10 .61 23 .96 22 .92 1 S i l i c o n e rubber bricks measured 1.9 x 12.1 x 17.B cm. while nylon (thick) bricks measured 2.4 x 12.1 x 17.8 cm, with appropriate clearance m the v e r t i c a l o r i e n t a t i o n . Appendix XIII. f values for s i l i c o n e rubber and nylon bricks' In the Lagarde retort Steam Temp, cont. <y.> c o f value, min Horizontal racking Rubber Rubber packaged Nylon V e r t i c a l racking Rubber Rubber packaged 50 50 105 105 10.08 10.38 10.55 10.70 11. 12. 79 84 11. 11. 63 36 11. 10. 46 25 14.24 12.28 13. 12. 36 75 13. 12. 26 55 11. 10. 36 49 9. 10. 83 00 9. 8. 83 97 10. 10. 21 59 10.36 9.93 50 50 120 120 10.32 10.40 10.83 10.95 11. 12. 56 68 11. 11. 37 85 11. 10. 07 68 12.88 12.37 12. 12. 85 22 11. 12. 55 13 10. 10. 62 22 9. 10. 96 86 9. 9. 91 17 10.07 10.51 9.90 10.36 65 65 105 105 9.94 10.20 10.38 10.57 11. 24. 60 66 37. 11. 62 80 28. 23. 58 05 12.93 12.57 12. 12. 78 89 13. 13. 12 01 11 . 9. 43 96 10. 9. 68 88 10. 9. 59 55 9. 10. 93 08 10.21 9.83 65 65 120 120 10.44 10.55 10.99 10.80 11. 11 . 36 72 14. 12. 40 .38 13. 11. 84 .07 12.55 13.09 12. 13. 23 .02 11. 12. 82 .65 11 . 9. 57 .98 9 9 .54 .57 10. 9. 01 23 10 10 .09 . 13 10.43 10.50 85 85 105 105 9.86 10.23 10. 10 10.79 10 23 .80 .70 43 22 .49 .32 31 21 .36 .59 12.67 13. 15 12 12 .46 .58 12 12 .33 .70 8. 9 .75 .57 10 10 .71 .88 10 9 .72 .71 9 10 .99 .34 9.45 10. 10 85 85 120 120 10. 11 10.61 10.58 11.20 11 14 .63 .46 11 12 .02 .55 18 10 .63 .70 12.76 12.36 12 12 .74 .34 12 12 .92 .02 9 10 .52 .27 9 9 .66 .68 9 9 .25 . 13 10 10 . 15 .58 9.86 10.47 100 100 105 105 10.03 10.43 10.39 10.89 44 20 .87 . 15 27 20 .70 .06 37 19 . 17 .54 12.87 13.70 12 13 .99 .24 12 13 .37 . 16 9 9 .20 .82 10 11 .68 .20 9 10 .55 .05 10 11 .04 .68 10.04 10.74 100 100 120 120 10.66 10.04 11 10 10.68 20 44 .71 .45 19 20 .91 .87 21 19 .04 .52 14.06 12.65 13 12 .88 .85 13 12 .52 . 13 9 9 .34 .38 11 10 .27 .04 10 9 .05 .48 11.66 10.79 10.99 10. 15 S i l i c o n e rubber bricks measured 1.9 x 12.1 x 17.8 cm while nylon ( t h i c k ) bricks measured 2.1 x 12.1 x 17.8 cm. with appropriate clearance in the v e r t i c a l orientation. PUBLICATIONS Ramaswamy, H.S., Ranganna, S. and Govindarajan, V.S. 1980. A non-destructive test for determination of optimum maturity of French (Green) beans (Phaseolus vulgaris). J. Food Quality. 3:11. Ramaswamy, H.S. and Richards, J.F. 1980. A reflectance method to study the green - yellow changes in fruits and vegetables. Can. Inst. Food Sci. Technol. J. 13(3) :107. Ramaswamy, H.S. and Ranganna, S. 1981. Thermal inactivation of peroxidase in relation to quality of frozen cauliflower (var. Indian Snowball). Can. Inst. Food Sci. Technol. J. 14(2) :139. Ramaswamy, H.S. and Tung, M.A. 1981. Thermo-physical characteristics of apples in relation to freezing. J. Food Sci. 46(3) :724. Ramaswamy, H.S. and Richards, J.F. 1982. Flavor of poultry meat - a review. Can. Inst. Food Sci. Technol. J. 15(1) :7. Ramaswamy, H.S. and Ranganna, S. 1982. Maturity parameters for okra (Hibiscus esculentus (L) Moench var. Pusa Sawani). Can. Inst. Food Sci. Technol. J. 15(2): 140. Ramaswamy, H.S., Tung, M.A. and Lo, K.V. 1982. Simplified equations for transient temperatures in conductive foods with convective heat transfer at the surface. J. Food Sci. 47(6):2042. Ramaswamy, H.S., Lo, K.V. and Staley, L.M. 1982. Air drying of shrimp. Canadian Agric. Eng. 24(2) :123. Ramaswamy, H.S. and Lo, K.V. 1983. Simplified mass diffusion relationships for drying of solids. Canadian Agric. Eng. (In press). Ramaswamy, H.S., Tung, M.A. and Stark, R. 1983. A method to measure surface heat transfer from steam/air mixtures in batch retorts. J. Food Sci. (In press). 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0095894/manifest

Comment

Related Items