Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Temperature and heat transfer studies in a water immersion retort Morello, Gerry F 1987

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

Item Metadata

Download

Media
831-UBC_1988_A7_6 M67.pdf [ 5.49MB ]
Metadata
JSON: 831-1.0097930.json
JSON-LD: 831-1.0097930-ld.json
RDF/XML (Pretty): 831-1.0097930-rdf.xml
RDF/JSON: 831-1.0097930-rdf.json
Turtle: 831-1.0097930-turtle.txt
N-Triples: 831-1.0097930-rdf-ntriples.txt
Original Record: 831-1.0097930-source.json
Full Text
831-1.0097930-fulltext.txt
Citation
831-1.0097930.ris

Full Text

TEMPERATURE AND HEAT TRANSFER STUDIES IN A WATER IMMERSION RETORT by GERRY F. MORELLO B. Sc., University of British Columbia, 1981 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DECREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Food Science We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA November 1987 © GERRY F. MORELLO, 1987 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. Department of f^O&b SUEAlCg The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 DE-6(3/81) ABSTRACT Temperature and heat transfer studies in a pilot-scale water immersion retort were performed. The temperature study investigated the temperature distribution and stability of the retort during the cook period. The investigation of heat transfer uniformity within the retort was based on heating and cooling parameters calculated from the heat penetration curves of food-simulating teflon transducers. The uniformity of sterilizing conditions within the retort was determined from process lethalities calculated for the transducers. Variable retort operating conditions consisted of two retort temperatures (115 and 125°C) and three weir heights (29.2, 31.2 and 34.6 cm). Mean standard deviations of thermocouple readings indicating temperature distribution during the cook period ranged from 0.19 to 0.22 C ° . Slight temperature gradients were found between the upper and lower water channels and between the entrance and exit regions of water channels. The coldest locations (the exit regions of water channels 1 and 2) averaged approximately 0.6 C° lower than the hottest locations (the entrance and exit region of water channel 11 and the exit region of water channel 10). Mean standard deviations of thermocouple readings indicating temperature stability during the cook period ranged from 0.10 to 0.20 C ° . Temperature stability was uniform between all water channels, except channel 11, which was less stable. The entrance and exit regions of water channels displayed similar stability. The existence of heat transfer variability within the water immersion retort was indicated. A retort temperature of 125°C produced smaller f^ and f values ii than 115°C. Variations in weir heipht influenced the distribution of fj values between trays. Weir height 2 (31.2 cm) exhibited uniform values between all trays. Weir height 1 (29.2 cm) exhibited uniform f^ values between all trays, except for a significantly larger value for the very top tray. Although weir height 3 (34.6 cm) created the most variability between tray levels, weir height 1 displayed the widest range of f^ values. More variability in f values between trays was shown during the cool period. Weir height 2 displayed the most uniform f values between trays, however, the range of f values between trays was similar for all three weir heights. Within trays, a gradient of f^ and f values was found between the entrance, exit and middle positions, with the smallest values found in the entrance positions. In comparison, the largest f^ values were found in the exit and middle positions of trays 1 and 10. The largest f values were found in the middle positions of trays 1 and 3 and the middle and exit positions of tray 10. Weir heights 1 and 2 produced smaller values than weir height 3, variations in weir height had no influence on j values. A gradient of values between tray levels was shown, with smaller values associated with upper trays and larger values with lower trays. Smaller and j values were associated with the entrance positions of trays than with the middle and exit positions. A comparison with steam processing indicated larger fj values for the water immersion process and larger f values for the cooling method used with the steam process. Calculation of process lethalities indicated variability of sterilizing conditions iii within the retort. Larger F 0 values were associated with upper trays than with lower trays. Within trays, larger F 0 values were found in the tray entrance positions than the middle and exit positions. The largest F 0 values were exhibited in the entrance positions of the middle to upper trays, while the smallest values were found in the middle and exit positions of the bottom trays. Retort pressure studies indicated pressure stability during the cook period, however, during the initial minutes of the cool period, a significant pressure drop occurred, which the retort corrected. Pressure stability was maintained once the target pressure was re-established. iv TABLE OF CONTENTS Abstract ii List of Tables vii List of Figures x Nomenclature xi Acknowledgements xii I. INTRODUCTION 1 II. LITERATURE REVIEW 4 A. HISTORICAL B A C K G R O U N D 4 B. THERMAL STERILIZATION SYSTEMS 5 1. Development of the Retort 5 2. Conventional Thermal Processing Systems 6 C. THERMAL PROCESSING FLEXIBLE CONTAINERS 8 1. Heating Media 8 2. Media Circulation 11 3. Retort Pressure 12 4. Racking Design 13 D. COMMERCIAL STERILIZERS FOR P O U C H PROCESSING 14 III. EXPERIMENTAL 18 A. RETORT SYSTEM 18 1. FMC 500W Laboratory Sterilizer 18 2. Retort Operating Procedure 20 B. PROCESSING CONDITIONS 21 C. DATA COLLECTION 24 1. Temperature Distribution and Stability 24 2. Heat Transfer Distribution 25 a. Heat Transfer 25 b. Teflon Transducers 28 c. Comparison With Steam Processing 32 3. Process Lethality Calculation 33 4. Retort Pressure 33 IV. RESULTS A N D DISCUSSION 35 A. RETORT TEMPERATURE 35 1. Temperature Distribution 45 2. Temperature Stability 51 B. HEAT TRANSFER 56 1. Heating Rate Index (f^) 56 2. Cool ing Rate index (f ) 62 ° c 3. Heating Lag Factor (j, ) 68 v 4. Cool ing Lag Factor (j^) 75 5. Comparison With Steam Processing 80 C. LETHALITY DISTRIBUTION 83 D. RETORT PRESSURE 89 V. C O N C L U S I O N S ; 93 VI. LITERATURE CITED 97 vi LIST OF TABLES Table 1: Estimated total water flow rates through retort car and corresponding flow through each paired water channel 23 Table 2: Thermocouple locations in the retort 27 Table 3: Transducer locations in the retort 30 Table 4: Some thermophysical properties of teflon 31 Table 5: Sample computer output of retort temperature histories 36 Table 6: Comparison of overall mean retort car temperatures with reference thermometer temperatures 43 Table 7: Range and mean of overall temperature uniformity (26 thermocouples) during the entire cook period 44 Table 8: Analysis of variance for overall standard deviations of temperature 44 Table 9: Range of temperature uniformity at each minute interval, the average uniformity during the cook period, and the range of times for the retort to stabilize 46 Table 10: Analysis of variance for pooled mean thermocouple temperatures 47 Table 11: Duncan's multiple range test comparing pooled mean thermocouple temperatures of different weir heights 49 Table 12: Duncan's multiple range test comparing pooled mean thermocouple temperatures of different water channels 50 Table 13: Range and mean of standard deviations of temperature for each ' thermocouple during the entire cook period 52 Table 14: Analysis of variance for temperature stability 53 Table 15: Duncan's multiple range test comparing mean temperature stabilities of different weir heights 54 Table 16: Duncan's multiple range test comparing mean temperature stabilities of different water channels and the reference thermometer 55 Table 17: Analysis of variance for heating rate indices 58 Table 18a: Duncan's multiple range test comparing f^ values associated with different trays for weir height 1 60 vii Table 18b: Duncan's multiple range test comparing values associated with different trays for weir height 2 60 Table 18c: Duncan's multiple range test comparing f^ values associated with different trays for weir height 3 61 Table 19: Duncan's multiple range test comparing f^ values associated with different tray positions 61 Table 20: Analysis of variance for cooling rate indices . . .63 Table 21a: Duncan's multiple range test comparing f values associated with different trays for weir heig ht 1 65 Table 21b: Duncan's multiple range test comparing f values associated with different trays for weir height 2 65 Table 21c: Duncan's multiple range test comparing f values associated with different trays for weir height 3 66 Table 22: Duncan's multiple range test comparing f values associated with different tray positions 66 Table 23: Analysis of variance for heating lag factors 70 Table 24: Duncan's multiple range test comparing values associated with different weir heights 72 Table 25a: Duncan's multiple range test comparing values associated with different trays for weir height 1 72 Table 25b: Duncan's multiple range test comparing values associated with different trays for weir height 2 73 Table 25c: Duncan's multiple range test comparing values associated with different trays for weir height 3. 74 Table 26: Duncan's multiple range test comparing values associated with different tray positions 74 Table 27: Analysis of variance for cooling lag factors 76 Table 28a: Duncan's multiple range test comparing j values associated with different trays for weir height 1 78 Table 28b: Duncan's multiple range test comparing j values associated with different trays for weir height 2 78 Table 28c: Duncan's multiple range test comparing j values associated with different trays for weir height 3 79 viii Table 29: Duncan's multiple range test comparing j values associated with different tray positions 79 Table 30: Comparison of heating and cooling parameters calculated using pure steam and water immersion processes at 125°C 81 Table 31: Analysis of variance for lethality values calculated using Stumbo's formula method 84 Table 32: Duncan' multiple range test comparing Stumbo's lethality values associated with different weir heights 85 Table 33a: Duncan's multiple range test comparing Stumbo's lethality values associated with different trays for weir height 1 87 Table 33b: Duncan's multiple range test comparing Stumbo's lethality values associated with different trays for weir height 2 87 Table 33c: Duncan's multiple range test comparing Stumbo's lethality values associated with different trays for weir height 3 88 Table 34: Duncan's multiple range test comparing Stumbo's lethality values associated with different tray positions 88 Table 35: Summary of retort pressure data during cook and cool periods 91 ix LIST O F FIGURES Figure 1: Cross section of F M C 500W Laboratory Sterilizer showing the basic retort design 19 Figure 2: Schematic drawing to show thermocouple locations in the retort (not to scale). 26 Figure 3: Schematic drawing to show transducer locations in the retort (not to scale) 29 Figure 4: Sample plot of temperatures monitored at different locations inside the retort during the cook period 39 Figure 5: Sample plot of mean and standard deviations of temperature for all thermocouples at each recorded time during the cook period 40 Figure 6: Sample plot of the mean and its standard deviation at each thermocouple location over the entire cook period 41 Figure 7: Example pressure histories of an experimental run 92 x N O M E N C L A T U R E a Half-length of a rectangular brick-shaped transducer, b Half-width of a rectangular brick-shaped transducer, c Half-thickness of a rectangular brick-shaped transducer. Cp Specific heat capacity. f Cool ing rate index. c ° . . f^ Heating rate index. F 0 Process lethality at the centerpoint of a package, j Cool ing lag factor. Heating lag factor, k Thermal conductivity. p Probability level for testing statistical significances. T. Transducer temperature at start of cool period. T.^ Transducer temperature at start of heating period. Tp - c Pseudo-initial temperature of the transducers at start of cooling period. ^pih Pseudo-initial temperature of the transducers at start of heating period. T^  Retort temperature. T Cool ing water temperature. a Thermal diffusivity. p Specific gravity. xi A C K N O W L E D G E M E N T S The author wishes to express his sincere appreciation to Dr. Marvin A. Tung for his advice throughout the course of this research project. He also wishes to thank the members of his research committee: Dr. William D. Powrie of the Department of Food Science, Dr. K. Victor Lo of the Department of Bio-Resource Engineering and Dr. Allan T. Paulson of Agriculture Canada, for their constructive criticism and review of this thesis. He wishes to especially thank Ian J. Britt for his invaluable advice and technical assistance. Special thanks are also extended to Agnes Papke and R. Alex Speers. Financial support was provided in part by the Strategic Grants Program of the Natural Sciences and Engineering Research Counci l of Canada. xii 1. I N T R O D U C T I O N Commercial sterilization processes apply thermal treatments to foods packaged in hermetically sealed containers in order to inactivate pathogenic and other spoilage organisms, thus attaining food stability at room temperature for extended periods of time. Flexible packages, metal cans and glass jars are available as containers for sterile shelf stable food products and food processors may choose from a variety of retort systems to meet their specific needs. Conventional retort systems utilize saturated steam, steam/air mixtures or water as the heating media. Flexible packages, however, present heat processing problems different from those with rigid containers (Pflug et al., 1963). As the product temperature rises during processing, the internal pressure within the flexible package will increase and if it exceeds the retort pressure by too great a degree, the package may burst. This is particularly a problem during the early stage of the cooling process. Steam/air or water with air overpressure systems are recommended for processing flexible pouches because they have the ability to maintain a retort pressure greater than internal pouch pressures achieved during processing. While steam/air processes have been commercially accepted in Canada, Europe and Japan, several processors in the United States have shown a preference for water immersion/air overpressure systems. The U.S. concern seems to be the lower surface heat transfer coefficient values reported by Pflug (1964) for steam/air mixtures as compared with water immersion heating. Once a retort system has been chosen, the sterilizing efficacy of each retort must be assessed, under the specific operating conditions to be used, before it can be used commercially. Uniformity and stability of heating and cooling within a retort 1 ? is critical to its ability to provide safe, but not excessive thermal processes. Currently, retorts and their operating procedures are usually evaluated by collecting temperature distribution information from thermocouples placed at various positions in the retort. Heat transfer distribution information would be valuable in determining the ability of the heating medium to uniformly release or absorb enthalpy during the process. Several researchers (Pflug, 1964; Ramaswamy, 1983; Tung et al., 1984a; Yamano, 1976) suggested that temperature uniformity within a retort does not necessarily ensure proper distribution of the heating medium to provide a uniform heat transfer distribution. At present, there is no standard methodology to measure heat transfer distribution and stability in thermal processing systems, however, several researchers (Peterson and Adams, 1983; Ramaswamy, 1983; Weintraub, 1986) have used food-simulating transducers to calculate heating rate indices which indicate heat transfer conditions. Two thermal process parameters originally developed by Ball (1923), the heating rate index (f^) and the lag factor (j^), are used to describe a heat penetration curve. The heating rate index is the time required for the straight line portion of the heat penetration curve to change one log cycle of temperature difference, while the lag factor is a measure of the lag in establishing a uniform heating rate. With sufficient heat transfer, for example when processing in pure steam, the surface heat transfer coefficient is very large compared to the thermal conductivity of the product; and therefore only the product's thermal properties and geometric form limit the heat transfer to its centerpoint. However, if heat transfer is reduced below a critical level, the product no longer is the limiting factor and changes in f^ and values are observed which identify changes in heat transfer. 3 The f^ and parameters may also be used to calculate differences in centerpoint lethalities (F 0) experienced by the product in various parts of the retort. A retort providing a large variability in lethal rate distribution may be severely overprocessing some of the product load, resulting in decreases in nutritional, sensor)' and functional properties. A weir type pressurized water retort is one water processing system available for flexible pouches. This system employs water under pressure as the heat transfer medium, with solid aluminum trays for pouch restraint and water flow control. The overall water flow proceeds by gravitational flow and is regulated by an adjustable weir. Other than recent work by McCinnis (1986) studying surface heat transfer coefficients, no other literature has been found discussing the thermal processing capabilities of this particular system. This investigation studied a pilot-scale weir type pressurized water retort with the following objectives: 1. To study the effects of the process variables, temperature and weir height, on the temperature distribution of the retort during heating and cool ing. 2. To study the effects of the above process variables on the heat transfer distribution of the retort during heating and cooling. 3. To study the effects of the above process variables on the centerpoint lethality distribution of the retort. 4. To compare the thermal responses of this water system with pure steam environments at the same temperature. II. LITERATURE REVIEW A. HISTORICAL BACKGROUND The discovery of the principle of heat sterilizaton and the initial development of heat processing practices is credited to Nicolas Appert, who did this work in the early 19th century. Although he did not understand the science of the process at this early stage, Appert learned the distinction between acid and non-acid foods in terms of the length of the processing time required (Ball, 1938). Appert's book (Appert, 1810), the first of four editions, was the only major literature on the subject of thermal processing for over 70 years (Coldblith, 1971). Jackson and Benjamin (1948) cited literature by American Can Company (1947), Ball (1938), De Kruif (1926) and Tanner (1932) as providing good coverage of the history of the canning process. A number of reviews (Anonymous, 1960; Bitting, 1937; Deming, 1902; May, 1937; Stare, 1949) are cited by Goldblith (1971) as dealing with the development of the canned foods industry as a whole, canning in certain areas of the United States or the world, or the development of canning machinery. Coldblith (1971, 1972a) concentrated on the development of the science and technology of canning in relation to the social character of the times. The next major stage in the development of thermal processing was the scientific discoveries of Pasteur in 1860. Pasteur was able to prove and establish the basic science behind Appert's processes of food preservation. Science was not applied to the study of thermal processing until 85 years after Appert's first publication, even though millions of cases of food had been packed and consumed during this time (Coldblith, 1971). 4 5 The contributions of Underwood and Prescott in the late 19th century were of major importance. Among their accomplishments, they showed that bacteria are the causative agents of canned food spoilage and heating to temperatures greater than the boiling point of water is needed to achieve sterilization; demonstrated the importance of heat penetration in canned foods and the importance of cooling; and were the first to recommend the use of incubation tests (Goldblith, 1972a). The 20th century has seen major advancements in thermobacteriology by Bigelow and co-workers, process technology by Ball and Stumbo, and improvements in processing equipment (Goldblith, 1972a). B. THERMAL STERILIZATION SYSTEMS 1. Deve lopment of the Retort Goldblith (1972b) gave a detailed account of the early development of the retort and the key people involved. Denys Papin, in 1681, is thought to be the first to use the pressure principle in cooking. Ball (1938) described Papin's "digester" as "an iron pot which had a cover that could be clamped down to withstand moderate pressure". Appert, circa 1831, began using a modified version of Papin's apparatus for cooking canned foods under pressure. This early equipment, quite dangerous because of the difficulty in regulating the fire and resulting steam pressure, did not grow quickly in popularity. Consequently, salt water baths were more common for processing canned food at temperatures above the boiling temperature of water. improvements in safety features and operating methods continued and in 6 1852 Raymond Chevallier-Appert patented (France) a retort equipped with a pressure gauge (Ball, 1938). The retort was able to reach 140°C and was documented as the first designed specifically for food conservation (Goldblith, 1972b). The forerunner of the vertical retorts used today was patented in the United States by A.K. Shriver in 1874. His retort used a steam supply from a separate boiler, was fitted with a safety valve, gauge and thermometer, and handled cans in large iron crates through the open top (Ball, 1938; Goldblith, 1971). This was a major breakthrough for the commercial use of retorts. As microbiological and heat transfer research progressed, the design of newer and improved thermal processing equipment fol lowed. 2. Convent ional Thermal Processing Systems Conventional and aseptic thermal processing systems are available for commercial sterilization of foods, with conventional systems being better suited for conduction heated and viscous foods. There is a variety of conventional systems to meet the specific needs of processors. For example, systems may be batch-type or continuous, employ still or agitation techniques, or use steam, steam/air or water as the heat transfer medium. Direct flame and microwave energy thermal processing systems are also available. Still retorts were used almost exclusively until about 1950, after which continuous agitating systems became more common. The still retorts are either horizontal or vertical and batch operated. These systems generally operate with saturated steam as the heating medium, however, they may be modified to operate with steam/air or water when overpressure is required. 7 Ball and Olson (1957), Brody (1971), Fennema (1975) and Lopez (1981) described and discussed a number of continuous and/or agitating steam retort systems. Bigelow et al. (1920) found that agitation of the container accelerated heat penetration in viscous foods such as cream-style corn. Several agitating and non-agitating water processing systems are also covered by Lopez (1981). One advantage in using water systems is that air overpressure may be applied when processing flexible containers susceptible to bursting. A direct flame sterilization system was developed in France in 1957 and documented by Beauvais et al. (1961). Leonard et al. (1975) reported on the principles of HTST processing in connection with the mechanisms of flame sterilization. Casimir (1975) discussed design parameters and operating characteristics of a flame spin sterilizer. The use of microwave energy in thermal sterilization of food in sealed containers is a relatively new area. The basic principle of large scale sterilization is to let the products pass through a microwave field on a conveyor belt or in a waveguide (Rosenberg and Bogl, 1987). Kenyon et al. (1971) and Ayoub et al. (1974) documented systems for continuous thermal processing of food pouches using microwave energy. They also cited several patents as discussing methods of sterilization by microwave energy in either batch or continuous systems (Jeppson and Harper,' 1967; Landy, 1965; Long et al., 1966). 8 C. THERMAL PROCESSING FLEXIBLE CONTAINERS Flexible packages, rigid metal cans and glass jars are available as containers for commercially sterile shelf-stable foods. The same basic requirements for commercial sterility of the product must be satisfied regardless of the type of container, however, flexible packages present some processing problems not associated with rigid containers. Mermelstein (1976, 1978) discussed the early development of retort pouches and Lampi (1977) gave an excellent in-depth review of the pouch. Areas of concern include: the type of heating medium, media circulation, retort pressure, and racking design. 1. Heat ing M e d i a Steam, steam/air mixtures and water are the potential conventional ' heating media available for thermal processing retort pouches. The use of steam, although superior in terms of come-up time, temperature distribution and heating rate of containers (Pflug, 1964; Pflug and Borrero, 1967), is least feasible, however, because high internal pouch pressures can exceed the retort pressure, especially during the latter stages of the process in the transition from the heating to the cooling cycle. Pure steam would be the ideal medium during the early stages of heating if it was possible to remove all non-condensible gases from inside the pouch before sealing (Pflug et al. , 1963). However, the pressure differentials caused by non-condensible gases within the pouch can be counteracted by processing with steam/air mixtures or water with superimposed air pressure. Some high temperature pure steam processing has been attempted for heating pouches followed by overpressure cooling. 9 It was concluded in the 1960's (Pflug, 1964; Pflug and Borrero, 1967), that water and steam/air could both be effectively used for retort pouches. Processing with water is preferred in the United States, while steam/air is more popular in Japan, Europe and Canada (Lopez, 1981; Milleville, 1980). Beverly (1980) stated there were conflicting views as to whether steam/air or water with air overpressure was the better system to use. Although comparisons have been made, no definitive conclusions can be made since each medium has certain advantages and disadvantages. Processing pouches in water is essentially identical to the well known standard procedure for glass containers, which may be one reason for the American preference. Yamano (1976) used bentonite models to compare heating in steam/air and water, and reported longer come-up times, but smaller f^ values for water. Pflug (1964) reported f^ values were lowest for 100% steam, followed by 90% steam/air, water, and 75% steam/air. Lopez (1981) listed overall heat transfer coefficients based on theoretical calculations by Pflug (1964). Coefficients for steam, water and 75% steam/air were found to be 965, 596 and 497 W / m 2 K . As the percentage of air in steam/air mixtures increases, the overall heat transfer coefficient decreases significantly (Tung et al., 1984b). Tsutsumi (1979b) stated that when heating below 120°C or when pouches have high air contents, water with high overpressures could be used to attain rapid and even heat penetration, thus avoiding low steam/air ratios. Weintraub (1986) reported that processing in water may be more effective than steam/air when large amounts of air are present in pouches. Pflug et al. (1963) listed several disadvantages of processing pouches in 10 water: long process times result from longer come-up and cool -down times for water; water hardness may soil pouches and build up scale on separation plates; many horizontal retorts are not designed for water processing, and improper or rapid steam injection to the water may result in pouch damage because of the resulting mechanical shock. Process water can be pre-heated to process temperatures in a separate reservoir, thus negating the rapid come-up rate advantage of steam/air (Lampi, 1977). Steam/air processing seems to lack most of the disadvantages associated with water processing. The use of steam/air mixtures is not new, as Parcell (1930) reported using it to retort glass containers. The maintenance of proper mixing of steam and air and steam/air distribution to all points in the retort are critical during processing. Milleville (1980) described a procedure which fulfills these requirements. Processing retort pouches with pure steam in Japan was reported by Tsutsumi (1979a). The high temperature-short time and ultra high temperature systems he discussed, employed temperatures of 135°C and 150°C. Tsutsumi (1979b) stated that since the pressure on the outside of the pouch is higher than the inside due to quick heating, no air pressure is necessary during heating, thus the pure steam greatly enhances the heating efficiency and temperature distribution within the retort. Air overpressure is necessary for the cooling cycle. The disadvantages of these high temperature processes include the necessity to closely monitor cook times to ensure the proper lethality and to avoid possible physical changes in foods after prolonged storage. Roop and Nelson (1981) reported successful pure steam processing at 115.6°C of bentonite dispersions in confined pouches. 11 2. M e d i a Circulat ion The objective of media circulation is to attain a uniform temperature and heat transfer distribution throughout the retort load. The media circulation can be manipulated in terms of flow pattern and flow velocity. Pflug and Borrero (1967) noted that the flow pattern in a retort affects the relative rate of heating of containers and the location of the fastest and slowest heating containers. They found that the flow pattern was important when heating with 100% steam, but was even more critical for water or steam/air processes. Some factors affecting the heating medium flow pattern and velocity are the retort type, racking system design, container loading patterns and container size. The velocity of the heating medium can affect the rate of heat transfer to containers. For example, when water is the heating medium, its temperature is reduced during contact with cold containers. Insufficient velocity of water past the containers may result in a temperature gradient and subsequently slower heating of the containers (National Food Processors Association, 1985). Peterson and Adams (1983) studied the effect of water flow rate on heat penetration parameters using a conduction heating model . They determined that the water circulation rate is important not only-i for the maintenance of retort temperature, but also for obtaining consistent, reproducible heat penetration parameters and process times. The f^ values decreased and heat transfer coefficients increased with increasing flow rates. When steam/air is the heating medium, the steam proportion is reduced by condensation of steam from the mixture when releasing heat to containers of food. Insufficient medium velocity results in poor movement and mixing of the steam and 12 air, thus producing composit ion variations in different parts of the retort and consequently differences in the rate of heat transfer. Pflug and Blaisdell (1961) studied the effect of velocities of steam/air mixtures on the heating rate of glass containers and concluded that fj values decreased with increasing velocities. Heat transfer coefficients for steam/air mixtures have been found to increase with the medium flow rate (Adams and Peterson, 1982; Blaisdell, 1963; Tung et al. , 1984b). Pflug and Borrero (1967) recognized the need for induced circulation of steam/air and water and made recommendations to that effect. Adequate circulation of steam/air media can be achieved by a positive flow system or a powerful fan to create turbulent conditions within the retort. Water systems may use circulation pumps or air agitation. 3. Retort Pressure A retort overpressure is essential to maintain the seal integrity of flexible containers during thermal processing, with the critical periods being the latter stages of heating and during the cooling period. Expansion of product and residual gases can create extremely high internal pressures during these periods (Davis et al., 1960), which will result in pouch or seal damage if these internal pressures greatly exceed the retort pressure. Toyo Seikan Kaisha Ltd. (1973b) stated that their RP-F pouch could withstand outer pressure by free expansion, but when inner pressures surpass outer pressures, pouch breakage might occur even at pressure differences of 0.1 kg/cm 2 (1.4 psi). Whitaker (1971) gave a detailed description and example as to the use of overpressure during processing under specific conditions. Various overpressure levels 13 have been suggested by different sources. Pflug et al. (1963) and Coldfarb (1971) recommended 34.47-68.95 kPa (5-10 psi) overpressure. Milleville and Badenhop (1980) stated that in general, an overpressure of 68.95 kPa (10 psi) is specified and should be controlled to within ±6.89 kPa ( ± 1 psi). Toyo Seikan Kaisha Ltd. (1973a) recommended an overpressure of 0.5-0.9 kg/cm 2 (7.1-12.8 psi) when processing at 120°C in order to provide a steam/air mixture containing 85 to 75% steam. Davis et al. (1972) suggested an overpressure of 89.63 kPa (13 psi) to protect seals and prevent rupturing. 4. Racking Design The racking design of a retort system is critical to good thermal processing practices (Wilson, 1980). Good racking designs provide proper pouch support, confinement, media circulation and orientation. Pouches are generally processed in the horizontal position, except for a few products when processing in a vertical plane is advantageous (Berry, 1979; Milleville and Badenhop, 1980). If a vertical design is used, pouches will bulge at the base if not properly supported by the rack. Racking designs should also prevent containers from moving and overlapping during the heating and cooling cycles. Pouch thickness affects the sterilization value of the process, but can be controlled through confinement by the racking design (Beverly et al., 1980; Milleville and Badenhop, 1980). The rack spacing will determine the maximum thickness pouches can reach during processing and a sterilization process can be based on this known thickness. 14 It is important that racking designs allow an even medium circulation to ensure uniform temperature and heat transfer throughout the load. Proper designs provide channels on both sides of the containers for the heating medium to circulate through. The design can also orient pouches vertically or horizontally in relation to the heating media flow. Davis et al. (1972) found process times varied depending on this orientation. Tung et al. (1984b) found heat transfer coefficients to be affected by the orientation of a flat brick-shaped test model in steam/air studies. D. COMMERCIAL STERILIZERS FOR POUCH PROCESSING A number of conventional batch and continuous systems exist for thermal processing of retort pouches. Specifically designed or modified retorts are available from American, European and Japanese manufacturers. These systems operate using water or steam/air; however, processors must make large capital investments to buy or modify their equipment. This led Roop and Nelson (1981) to investigate using traditional equipment for both pouches and cans without retort modifications. They found the results were encouraging enough to suggest further study. Traditional batch retorts require significant modifications to convert them to steam/air or water systems. An example of the basic piping and instrumentation for a horizontal steam/air retort was diagrammed by Milleville (1980). This system maintained a temperature uniformity of ±0.56 C° ( + 1 F°) when operated at ( a steam/air mix of 75/25 at 122.2°C (252°F). The Lagarde Company of France produces a steam/air system used by Magic 15 Pantry Foods, Inc. at Hamilton, O N (Anon., 1982; Milleville, 1981; Morris, 1981). Milleville (1981) diagrammed the system which uses a turbo fan and side ducts to circulate the steam/air. Magic Pantry Foods operated at 121.1 °C (250°F) for a 2000 pouch/retort capacity. Lagarde claimed greater energy efficiency and more uniform heat distribution for steam/air than conventional hot-water retorts (Anon., 1982). A rotary retort is also available to provide product agitation. Toyo Seikan Kaisha Ltd. (1973a) of Japan described in detail their "Fully Automatic Overpressure Retort". This steam/air retort was documented to process their RP-F retort pouches at 120°C (248°F) with 1.5 kg/cm 2 (21.3 psig) pressure; and had a temperature distribution within ± 2 C° ( ± 3 . 6 F°) at sterilization temperature. The three car retort had a capacity of 1944 pouches. Lampi (1977) gave further details of a modified version for high temperature processing. Two other steam/air retort systems were available from Specialty Seafoods of Anacortes, W A and Brett & Associates of Delta, BC (Milleville, 1981). Andres and Duxbury (1972), Anon. (1973) and Gee (1973) briefly described a horizontal batch-type water retort used in the United States. This retort circulated water at 1514 litres per minute (400 gpm), operated at an air pressure of 241.3 kPa (35 psig) and had a temperature distribution of within ±1.11 C° ( ± 2 F°). The pouch capacity varied from 2016 to 2688, depending on the product, and a clamp down bar over the pouches in the racks prevented floating. The FMC "Convenience Foods Sterilizer" was designed to circulate water between tiers of pouches using a weir system (Anon., 1979b; Anon. , 1982; Lopez, 1981; Milleville, 1981). FMC claimed the system minimized come-up and cool -down 16 times and achieved efficient uniform heat transfer for temperature variations between containers of less than ±0.56 C° ( ± 1 F°). Pouches were confined between two plates to maintain a constant maximum thickness. The system used a reservoir to pre-heat process water and models were available in four different pouch capacities. The Stock "Rotomat" (Lopez, 1981), a horizontal circulating water retort, has been modified to handle retort pouches. The retort load could be rotated to improve the rate of heating and reduce cook time, however, there was some controversy as to whether pouches could withstand rotational stresses; therefore, rotation of the load was not used (Milleville, 1981). An improved tray feature for pouches was introduced in 1982 (Anon., 1982). A rectangular water retort (Anon., 1981) was also marketed in three production models of 1000, 1500 and 2000 entree-size pouches/load, and Anon. (1975) described a unique horizontal batch retort by Continental which could process pouches in water or steam/air at a capacity of 1200/batch. Continuous retorts may also be suitable for processing pouches. The Rexham "Hydrolock" was designed to process with steam/air circulated by a fan to achieve a temperature distribution within +0.28 C° (±0 .5 F°) (Lampi, 1977). Lawler (1967) was reported by Lampi (1977) to have first used this system. Lopez (1981) and Pinto (1978) described a version of the Hydrolock system which used a carrier system to transport pouches through chambers against the flow of process water. Goldfarb (1971) described a Robins Hydrolock which was very similar to the Rexham steam/air model . A modified Stork "Storklave" system (Anon., 1979a; Lopez, 1981) was 17 reported to be able to handle retort pouches. The system was basically a continuous vertical retort similar to hydrostatic systems, but featured a set of pallet-type carriers which travelled through the system in a continuous movement by a conveying system. A hydrostatic system, specifically for retort pouches and apparently shipped to Japan in 1975 (Gerrish, 1975), was rated at 450 pouches/minute. Gerrish (1975) also reported that FMC was developing a hydrostatic system using hot-water processing with air overpressure and Lampi (1977) mentioned other systems. III. EXPERIMENTAL A. RETORT SYSTEM 1. F M C 500W Laboratory Steri l izer The retort investigated in this study was an FMC Model 500W Laboratory Sterilizer (FMC Corporation, Santa Clara, CA) located at the Department of Food Science pilot plant, University of British Columbia. The FMC 500W retort is a pilot-scale version of a weir type pressurized water processing system. This type of system employs steam-injected hot water as the heating medium with over-riding air pressure. Solid aluminum trays, used for pouch restraint and water flow control, are stacked inside a retort car and the water medium travels through a flow channel which makes up the lower part of each tray. Basically, the system is a plate-type heat exchanger as heat is transferred from the water to the trays by convection and then to the pouches by conduction. In this FMC 500W system, water is initially pre-heated in a reservoir prior to filling the retort car. Hot water enters the top of the car where it travels to one end, is distributed by a perforated plate and flows horizontally through the flow channels of the stacked trays to the opposite end. Water exiting the channels turns 90° and moves upwards to a weir over which it exits the car. The water falls to the bottom of the retort shell where it flows through an outlet and is circulated past a steam injection portal and back to the retort car by a recirculation pump. Figure 1 is a cross sectional diagram showing the basic design of the retort. Cool ing is achieved by simultaneously adding cold water from the main 18 Reference Thermometer Retort Shell Water Inlet -y s. - 4 ~ > 1 —> <-<- -~> 4 - - > 1 1 4 -—> <— —> < ~ - - > 1 1 f ~ - > 1 < ~ —— -*• .... . - > 1 1 < - 1 <^ \ - 4 1 —) 1 Water Distribution Plate \ Water Channel Retort Water Level Water Outlet Figure 1 Cross section of F M C 500W Laboratory Sterilizer showing basic design. 20 water supply and draining the equivalent flow from the retort, while circulating that portion of the total flow that is not being drained. As a means of conserving hot water, a portion of the displaced and partially mixed waters may be shunted back to the reservoir using the recirculation pump during the early stages of cooling. Some specifications for this retort system were provided by the manufacturer. The retort is 610 mm in diameter and 1524 mm long, with a maximum operating pressure of 310 kPa (45 psig). The hot water reservoir located above the retort has the same specifications. The inside dimensions of the retort car are 1048 x 402 x 324 mm. Eleven trays are provided, although only 10 were used in this study due to the transducer thickness, with the dimensions 1041 x 397 x 19 mm. The pouch capacity of 11 trays is 176 (16/tray) for pouches with the dimensions 120 x 184 x 19 mm. There are 11 pairs of water channels when using 10 trays, with individual cross-section dimensions 177.8 x 6.35 mm (7 x 0.25 in). 2. Retort Operat ing Procedure In general, the retort operation procedure provided by the manufacturer was followed. Each experimental run consisted of a 50 minute heating and a 25 minute cooling period. The retort was operated with a steam supply pressure of 365.4 kPa (53 psig). The following is a brief outline of the procedure: 1. Fill reservoir tank 2/3 full with water and pre-heat to 10 C° above the target retort temperature to reduce the come-up time. 2. Pressurize the reservoir tank to 206.8 kPa (30 psig) to aid the transfer of water to the retort at steam-on. 3. Open and close the appropriate valves and set the controls prior to steam-on. 21 The process water and tempernture overshoot controllers were set at 260°F for the target temperature of 125°C and 242°F for the target temperature of 115°C. The process pressure controller was set at 25 psig to maintain a retort pressure of 172.4 kPa . 4. At steam-on, transfer water from reservoir to retort. 5. When water transfer is complete, close retort vent, open retort air supply, and open steam bypass valve. 6. When retort mercury thermometer reaches within 2 C° of the target temperature (approximately 9-10 minutes after steam-on), close steam bypass valve. 7. Prepare reservoir for cool cycle by releasing air pressure. 8. At cool start, close main steam valve, open high flow cold water valve and tranfer some hot process water to reservoir tank. 9. At end of hot water transfer, close high flow cold water valve, open low flow cold water valve 2.5 turns and stop hot water transfer. Turn on level controller and adjust drain valve to maintain constant retort water level. 10. At end of cool cycle, drain retort water and release air pressure. B. PROCESSING CONDITIONS Retort operating conditions were varied when feasible to represent the combinations of conditions which might be utilized in the normal operation of the retort system. During experimentation, U.B.C. Physical Plant operated the main campus steam pressure at 275.79 kPa (40 psig). Practice runs indicated this condition 22 provided insufficient enthalpy to heat a fully loaded retort car. Upon request, Physical Plant increased the steam pressure to 365.42 kPa (53 psig) during experimentation. Ideally, a high pressure steam supply of 413.69 kPa (60 psig) or greater would be more desirable for retort operation. Retort temperatures of 115°C (239°F) and 125°C (257°F) were targeted during experimentation. These temperatures were chosen to represent temperature extremes at which the retort might be operated. Cool ing water temperatures were uncontrollable and therefore ranged from approximately 8 to 15°C. Weir heights of 29.2 cm (11.5 in), 31.1 cm (12.0 in) and 34.6 cm (13.6 in), hereafter referred to as weir height 1, 2 and 3, were set during experimentation. Two weir plates were available, one adjustable and the other non-adjustable. The two lowest settings represent the minimum and maximum heights of the adjustable weir plate. The highest setting represents the only height of the non-adjustable plate. These weir heights were chosen to represent the extremes in available settings. Peterson and Adams (1983) showed that variability in water flow rate affected the heat transfer coefficient and consequently process lethalities. In this system, the water flow rate during heating was governed by the flow capacity of the water re-circulation pump. The flow capacity of the pump was not adjustable, therefore the overall water flow rate could not be varied during heating. The water flow rate during cooling was controlled by the combined re-circulation pump and the cold water inlet valve. The overall cooling flow rate could be varied by manipulating the cold water inlet valve, however, it was kept constant during experimentation. 23 TABLE 1 Estimated total water flow rates through retort car and corresponding flow through each paired water channel. Source Total Flow Rate (liters/minute) Flow Rate/Channel (liters/minute) Recirculation pump 97 8.8 Low flow valve 150 13.6 High flow valve 290 26.4 Pump + Low flow valve 247 22.5 Flow rate meters were not available in this retort system, therefore crude estimations of the heating and cooling water flow rates were calculated by opening valves and measuring the volume of water which collected in an empty retort car during a specific length of time. Table 1 shows the estimated total water flow rate through the retort car and corresponding flow rate through each of the 11 paired water channels (assuming uniform circulation patterns) supplied by the re-circulation pump, the low flow cold water inlet valve (open 2.5 turns), the high flow cold water inlet valve (fully open) and the combined pump and low flow valve. The total heating and cooling flow rates used were estimated to be 97 and 247 liters per minute. A constant retort pressure of 172.37 kPa (25 psig) was set to maintain a consistent head pressure for the re-circulation pump to act against and to provide an overpressure. Weintraub (1986) found no significant difference (p>0.05) in heating rate indices of unpackaged teflon transducers resulting from variable retort pressures. A fully loaded retort car was used during the experimental runs to simulate 24 the worst conditions the retort would be expected to operate under. The ballast load consisted of 100 g cans of Brunswick Canadian Sardines in soya oil (Connors Bros., Ltd., Black's Harbour, NB) purchased locally. Cans of sardines were chosen for several reasons: cans could withstand repeated thermal processing; cans could withstand pressure fluctuations without bursting; two cans placed side by side had approximately the same dimensions as the teflon transducers (or a filled pouch); and the thermal properties of the sardines would be similar to many foods. Rectangular aluminum brackets were used to prevent the pairs of cans from separating. One hundred forty-eight pairs of sardine cans were loaded as ballast, 14 pairs on the 6 trays holding transducers and 16 pairs on the 4 trays holding no transducers. C. DATA COLLECTION 1. Temperature Dis tr ibut ion and Stability The temperature distribution and stability of the fully loaded retort car was studied under different operating conditions. Thermocouple locations were chosen to represent extremes in temperature to which containers might be exposed during the heating and cooling cycles. Twenty-eight teflon-insulated 24 A W G copper/constantan thermocouples (Omega Engineering, Inc., Stamford, CT) with fused sensing junctions were placed in the selected locations. The thermocouples were connected to a Kaye Ramp II Scanner/Processor (Kaye Instruments Inc., Bedford, MA) and the temperature data were logged at one minute intervals. The data were recorded on magnetic tape using a Columbia 300D Digital Cartridge Recorder (Columbia Data Products Inc., 25 Columbia, MD) for subsequent analysis using a microcomputer. All thermocouples were pre-calibrated in a Haake N3-B circulating oil bath (Haake Instruments Inc., Saddle Brook, NJ) against an ASTM mercury-in-glass thermometer and appropriate corrections were made to the temperature data. Figure 2 shows the thermocouple locations in the entrance and exit areas of the water channels. Thermocouples were also located at positions adjacent to where water entered and exited the car . and adjacent to the retort thermometer and temperature control sensor. Table 2 indicates the specific location of each thermocouple with reference to Figure 2. In general, odd numbered thermocouples were located in the entrance regions and even numbered thermocouples in the exit regions of the water channels. Thermocouples positioned in the water channels were sandwiched in hollow pieces of rubber tubing, with the junctions exposed to the water, in order to firmly secure their position and to prevent the sensing junctions from contacting the tray surface. 2. Heat Transfer Distr ibut ion a. Heat Transfer The heat transfer uniformity of the fully loaded retort car was studied under different operating conditions using solid teflon transducers to simulate conduction heating. Eighteen teflon food-simulating transducers (150 x 111 x 21.7 mm) were placed in selected locations to represent extremes in heat transfer to which containers might be exposed during heating and cooling. The temperature histories at the geometric center of the transducers during heating and cooling were logged 26 Retort Car 11 10 Therpiocouple Water Channel 3L 33: Entrance End Side View Exit End © C ™ a t e r , DO Thermocouple Channel ©A Bo Top View FIGURE 2 Schematic drawing to show thermocouple locations in the retort. (not to scale) 27 TABLE 2 Thermocouple locations in the retort. Thermocouple Number Water Channel Position 19 1 A 20 1 B 21 2 A 22 2 B 23 3 C 24 3 D 25 4 C 26 4 D 27 5 A 28 5 B 29 6 A 30 6 B 31 6 C 32 6 D 33 7 C 34 7 D 35 8 A 36 8 B 37 9 A 38 9 B 39 10 C 40 10 D 41 11 C 42 11 D 43 Adjacent to water inlet 44 Adjacent to water exit 45 Adjacent to thermometer 46 Adjacent to thermometer 28 at one minute intervals using the Kaye datalogger and recorded on magnetic tape for further analyses. Two heat penetration parameters, the heating and cooling rate indices (f) and lag factors (j) (Ball, 1923; Ball and Olson, 1957; Olson and lackson, 1942), were calculated from these data using a microcomputer to describe the heating and cooling behavior at the centerpoint of the teflon brick-shaped transducers. The effect of location within the retort car on the ability of the heating medium to transfer enthalpy was compared using these parameters. Figure 3 shows the transducer placements at the entrance, middle and exit positions of selected trays. Table 3 indicates the specific location of each transducer in reference to Figure 3. fa. Teflon Transducers A variety of materials have been used in model systems for comparing factors which affect thermal processing of food products. For example, bentonite suspensions (Peterson and Adams, 1983; Pflug, 1964), bricks of metal and nylon (Ramaswamy, 1983), bricks of teflon (Weintraub, 1986) and silicone rubber and nylon (Ramaswamy and Tung, 1986) have been used. In this study, teflon bricks were used as food-simulating transducers. Table 4 lists some characteristic thermal properties of teflon (Mantell, 1958) and the calculated thermal diffusivity (Leniger and Beverloo, 1975) using these figures. The teflon transducers heat by conduction and their thermal diffusivities are similar to many food products listed by Tung et al. (1984a). Weintraub (1986) reported no apparent thermal degradation or permanent modification to properties other than slight warping of the bricks after repeated high temperature processing. 29 Retort Car Transducer 10 VIA Tray / / 8 / / / / Entrance End Side View Exit End D Transducer E Tray F A B C Top View FIGURE 3 Schematic drawing to show transducer locations in the retort. (not to scale) 30 TABLE 3 Transducer locations in the retort. Transducer Number Tray Number Position 1 1 A 2 1 B 3 1 C 4 3 D 5 3 E 6 3 F 7 5 A 8 5 B 9 5 C 10 6 D 11 6 E 12 6 F 13 8 A 14 8 B 15 8 C 16 10 D 17 10 E 18 10 F The transducers may therefore be used for numerous trials while providing consistent and repeatable measurements. The rectangular teflon bricks were constructed from two teflon slabs (Cadillac Plastics, Montreal, PQ) sandwiching a copper/constantan thermocouple at the geometric center. Working with a solid material provides consistent dimensions and allows thermocouples to be more accurately placed. Detailed construction of the bricks is described by Weintraub (1986). The bricks, averaging approximately 21.7 mm thick, 150 mm long and 111 mm wide, were designed to model the size and 31 TABLE 4 Some thermophysical properties of teflon. Properties of tef lon (Mantell, 1958) Specific Gravity (p) Specific Volume Thermal Conductivity (k) Specific Heat (Cp) Thermal Expansion Heat Resistance Heat Distortion Water Absorption 2.1 - 2.3 g/cm 3 476.2 - 434.8 c m 3 / k g 6 x 10" " cal/s cm C ° 0.25 cal/g C ° 10 x 10" 5 / C ° 500 °F (continuous) 270 °F (66 psi) 0.0 % Calculat ion of thermal diffusivity (Leniger and Beverloo, 1975) k 6 x 1 0 " "cal/s cm C ° a = = p Cp 2.2 g c m " 3 x 0.25 cal/g C ° = 1.09 x 10" 7 m 2 /s shape of a filled retort pouch. Each brick was calibrated using pure steam in a pilot-scale vertical retort. The reproduction of thermal behavior between each teflon brick is dependent on the brick construction. Since f^ and f values are dependent upon brick dimensions and brick heating approximates an infinite slab, correction factors for conduction heating were used to adjust the f values of each brick to f values for a standard thickness of 19 mm (0.75 in) (Ball and Olson, 1957; Stumbo, 1973). Since j, and j are dependent upon the position in the brick, correction factors 32 were used to adjust the j values of each brick to j values for a standard distance from the brick surface of 9.5 mm (Olson and lackson, 1942; Ball and Olson, 1957). c. Comparison With Steam Processing The thermal responses of the teflon transducers processed in pure steam were studied using the same FMC 500W Laboratory Sterilizer as modified by Britt (1987) to operate with pure steam or steam/air mixtures. Two meshed racks (Britt, 1987) were centered in the retort with nine transducers placed on each rack such that their surfaces were openly exposed to the retort environment. An attempt was made to expose the transducers to uniform heating conditions by randomly placing them on the racks for each experimental run and by using no ballast. It was hoped that this would reduce any variability in the transducer responses resulting from retort influences. The main steam supply pressure was 365.42 kPa (53 psig) and a target temperature of 125°C was selected. The total heating period of 50 minutes consisted of a 5 minute venting period and a 6 minute come-up time. At steam off, the retort was f looded with cooling water from the low flow cold water valve (opened 2.5 turns) so that the transducers were submerged. The cooling water was circulated using the re-circulation pump while simultaneously adding and draining cold water. 33 3. Process Lethality Calculat ion Using microcomputer software (Pro Calc Associates, Surrey, BC), Stumbo's formula method (Stumbo, 1973) was used to calculate centerpoint lethality values (F 0 ) . Calculations were based on heating and cool ing parameters determined for the specifically located teflon transducers. Smith and Tung (1982) discussed and compared formula methods available for calculating thermal process lethality. Stumbo's method accounts for the lethality contributed by the cooling cycle through the use of variable j values. Come-up times were not taken into consideration for these lethality calculations. 4. Retort Pressure An electronic pressure transducer (Model A-5/1148, Sensotec, Columbus, OH) was mounted in the top of the retort shell to record the retort pressure during each experimental run. The transducer was calibrated using a deadweight pressure calibrator (Chandler Engineering Co. , Tulsa, OK) and an excitation voltage was supplied by an HP 6214A power supply (Hewlett Packard, Palo Alto, CA). The pressure histories of the experimental runs were logged at one minute intervals using the Kaye datalogger and recorded on magnetic tape for further analyses. The retort pressure control needle was set at 25 psig (172.37 kPa) during all experimentation. This setting was chosen to provide a consistent retort head pressure against which the water recirculation pump had to work and to supply a typical overpressure that would be appropriate for retort pouch processing. Retort pressures under pure steam conditions do not produce an overpressure condition, for example, pure steam at 115 and 125°C create retort 34 pressures of only 70.58 and 132.67 kPa gauge, respectively. Atmospheric pressure is approximately 101.33 kPa. A high pressure air line was used to supply the necessary air to create a steam/air mixture in the retort headspace which produced the target pressure. The target retort pressure provided a theoretical overpressure of 101.79 kPa (14.76 psig) while processing at 115°C and 39.70 kPa (5.76 psig) while processing at 125°C. IV. RESULTS A N D DISCUSSION A. RETORT TEMPERATURE The efficacy of a retort system is traditionally assessed by determining the temperature distribution and stability of the retort under specific operating conditions. Temperature distribution describes the temperature uniformity at different locations within the retort at a specific time. Temperature stability describes the temperature uniformity of a specific location over a period of time. A series of locations within the retort was monitored with thermocouples during experimentation. Table 5 is a sample computer output showing the temperature histories of 26 selected locations in the retort car during the come-up and cook period for an experimental run. Temperature distribution is recorded in the bracketed right hand column in the form of mean temperature and standard deviation of the combined locations at each one minute interval. Temperature stability over the cook period, 1 8 - 5 0 minutes, is recorded at the bottom of the table in the form of mean temperature and standard deviation of each individual location. The grand mean temperature and standard deviation for the combined locations during the entire cook period is recorded at the bottom of the table. Figure 4 is a sample 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 temperature (represented by the horizontal line) for any of the thermocouple locations at any particular time. Figure 5 is a sample plot of mean and standard deviations of temperature for all thermocouples at each recorded time during the cook period. This plot can be used to observe the 35 36 TABLE 5 Sample computer output of temperature histories. TEMPERATURE DISTRIBUTION WATER PROCESS RUN C00K10 T h e r m o c o u p l e number and c o r r e c t i o n f a c t o r MEAN S.D. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 r 1me 0 4 0 3 0 4 0 2 0 2 0 3 0 2 0 2 -0 1 -0 1 -0 1 -0 1 -0 1 0 2 0 2 ( Mean. S D. ) 0 0 0 2 0 2 0 1 -0 1 0 2 -0 6 -0 1 0 2 0 2 0 2 0. 16 1 16 1 16 0 15 8 15 8 15 8 15 9 16 0 15 9 15 7 15 7 15 7 15 8 15 7 15 9 15 6 15 8 15 8 15 7 15 7 16 3 16 1 17 7 17 3 18 8 18 1 ( 16 18. 0 83) 1 . 46 1 62 1 52 0 56 0 83 6 57 1 95 1 60 9 100 1 62 0 95 e 61 7 98 6 72 3 106 1 82 9 107 3 69 7 10B 7 88 0 112 3 97 1 1 16 8 107 6 119 3 81 3 ( 84 6 3 , 22 40) a. 64 3 74 8 61 6 57 2 67 0 72 0 80 9 74 9 106 3 69 4 109 0 73 3 108 8 82 4 109 2 95 7 1 14 1 90 2 112 9 106 8 1 1 1 8 1 13 6 112 S 116 5 109 7 95 5 ( 91 94. 20 03) 3. 83 7 74 5 61 1 66 9 99 0 56 2 102 3 74 2 106 5 93 0 106 0 84 9 107 8 98 3 107 9 100 3 108 1 102 8 1 10 3 103 3 110 3 106 1 111 8 107 7 1 13 4 98 4 ( 95 95 , 16 60) 4. 88 6 81 8 76 3 69 1 108 6 68 3 109 7 88 2 1 12 1 103 7 1 12 8 94 6 112 7 106 2 1 12 6 107 3 113 1 109 5 1 14 1 1 10 0 1 14 6 1 1 1 8 1 16 1 1 14 2 1 16 4 106 6 ( 103 04, 14 83) 5. 94 8 90 6 92 3 76 1 1 14 4 86 .2 1 16 6 98 9 1 17 6 109 8 1 18 0 107 6 1 18 5 112 4 117 6 1 12 9 1 18 5 1 14 9 1 19 5 1 15 0 119 0 1 16 6 120 7 1 18 4 121 2 113 3 ( 110 0 5 . 12 25) 6. 101 4 97 4 106 5 81 3 120 0 100 1 121 0 107 5 122 1 1 15 3 122 6 1 14 6 123 3 117 4 123 1 1 18 3 123 0 120 2 124 1 120 0 123 9 121 2 124 9 122 8 125 4 118 0 ( 115 98 . 10 71) 7. 106 5 99 4 1 16 6 78 1 120 6 108 5 121 6 115 4 122 7 1 18 6 122 6 116 7 122 9 120 2 122 9 120 4 122 9 121 8 123 5 121 6 123 3 122 1 123 7 123 1 123 8 120 7 ( 117 70. 10 0 3 ) 8. 1 10 0 100 9 118 7 91 1 122 9 1 14 4 123 5 1 18 7 123 3 120 6 123 7 120 0 124 0 121 3 124 0 121 7 123 9 122 6 124 2 122 3 124 0 122 9 124 6 123 7 124 5 120 8 (119 7 0 , 7 85) 9 . 117 5 106 7 121 7 104 1 123 9 1 18 0 124 3 121 0 124 3 121 8 124 4 122 0 124 3 122 8 124 6 122 8 124 7 123 4 125 0 123 1 124 6 123 7 125 0 124 3 125 0 121 6 ( 121 72 . 5 20) 10. 120 3 1 13 3 122 2 1 1 1 0 125 0 120 2 125 1 122 3 125 2 123 0 125 0 123 0 125 1 123 2 125 3 123 6 125 0 123 9 125 5 123 8 125 .2 124 2 125 6 124 9 125 3 121 6 ( 122 9 9 , 3 56) 11 . 124 . 1 1 17 6 125 1 1 16 2 126 3 121 .6 126 4 123 6 126 5 124 0 126 3 124 1 126 3 124 3 126 4 124 7 126 3 124 9 126 7 124 7 126 .4 125 2 126 9 126 0 126 6 123 1 (124 6 3 , 2 64) 12. 126 1 121 3 125 7 1 19 3 126 0 123 . 1 126 1 124 5 126 2 124 9 126 1 124 8 126 0 125 1 126 1 125 5 126 1 125 5 126 3 125 4 126 .0 125 7 126 4 126 4 126 1 124 4 ( 125 2 0 . 1 65) 13. 125 7 124 4 125 5 122 4 125 5 123 7 125 6 124 8 125 7 124 9 125 61 124 9 125 5 125 1 125 6 125 2 125 5 125 4 125 7 125 2 125 5 125 3 125 9 125 9 125 6 124 1 ( 125 16, 0 78) 14 . 126 2 124 5 125 8 123 2 126 1 124 2 126 2 124 6 126 2 124 6 126 1 124 7 126 0 124 9 126 3 124 9 126 1 124 9 126 5 125 0 126 2 125 1 126 6 125 8 126 4 124 6 ( 125 4 5 , 0 89) 15. 126 3 124 9 126 2 124 0 126 1 125 2 126 3 125 2 126 4 125 4 126 3 125 4 126 1 125 5 126 1 125 7 126 3 125 9 126 4 125 6 126 1 125 8 126 5 126 4 126 1 125 3 ( 125 8 3 , 0 59) 16. 126 2 125 2 126 1 124 7 126 0 125 3 126 2 125 5 126 2 125 7 126 2 125 6 126 2 125 8 126 3 125 8 126 2 125 9 126 4 125 8 126 1 126 0 126 6 126 4 126 1 125 6 ( 125 9 3 . 0 42) 17. 126 3 125 2 126 2 124 8 126 0 125 5 126 3 125 7 126 2 125 7 126 1 125 7 126 2 125 9 126 3 125 9 126 2 125 9 126 5 125 8 126 . 1 126 0 126 5 126 4 125 9 125 7 ( 125 9 6 . 0 39) 18. 126 . 1 125 4 126 1 125 1 126 1 125 7 126 4 125 9 126 3 125 8 126 1 125 7 126 3 125 9 126 3 125 9 126 1 126 0 126 4 125 9 126 0 126 0 126 6 126 5 126 1 125 8 ( 126 02 . 0 33) 19. 126 1 125 5 126 2 125 2 126 2 125 9 126 3 126 0 126 5 125 9 126 3 125 9 126 2 126 0 126 4 126 1 126 3 126 2 126 6 126 0 126 1 126 2 126 6 126 5 126 1 126 0 ( 126 13, 0 3 D 20. 126 2 125 6 126 2 125 4 126 1 126 0 126 3 126 0 126 2 126 0 126 2 125 9 126 1 126 1 126 3 126 2 126 1 126 1 126 3 126 1 126 1 126 2 126 5 126 5 126 1 126 0 ( 126 1 1 . 0 23) 21 . 126 1 125 6 126 2 125 4 126 1 126 .0 126 3 126 0 126 2 126 0 126 1 125 9 126 2 125 9 126 3 126 1 126 1 126 1 126 4 126 0 126 . 1 126 1 126 5 126 5 126 3 126 0 ( 126 10, 0 24) 2 2 . 126 2 125 7 126 2 125 6 126 1 126 .0 126 3 126 1 126 5 126 1 126 3 126 0 126 2 126 1 126 3 126 . 1 126 2 126 3 126 4 126 1 126 .3 126 3 126 6 126 5 126 3 126 0 ( 126 18. 0 22) 2 3 . 126 . 1 125 6 126 2 125 6 125 9 126 . 1 126 1 126 1 126 5 126 1 126 3 126 0 126 1 126 0 126 1 37 Table 5 Continued 126 2 126 3 126 3 126 .4 126 .0 126 . 1 126 2 126 5 126 5 126 . 1 126 1 ( 126 13. 0. 22) 2 4 . 126 .2 125 .7 126 2 125 .7 126 . 1 126 .2 126 3 126 . 1 126 .3 126 . 1 126 3 126 .0 126 1 126 1 126 2 126 .2 126 .3 126 3 126 .3 126 . 1 126 .3 126 4 126 .6 126 .5 126 .4 126 0 ( 126 19. 0. 20) 2 5 . 126 . 1 125 .8 126 2 125 .7 126 .2 126 .2 126 4 126 . 1 126 3 126 . 1 126 2 126 .0 126 2 126 1 126 3 126 .2 126 .2 126 3 126 .3 126 . 1 126 .3 126 3 126 .6 126 .5 126 .3 126 1 ( 126 20. 0 19) 2 6 . 126 .2 125 .8 126 2 125 .8 126 1 126 . 1 126 3 126 2 126 3 126 .2 126 2 126 . 1 126 1 126 1 126 1 126 .3 126 .3 126 3 126 .3 126 . 1 126 . 1 126 4 126 5 126 6 126 .5 126 2 (126 21 . 0. 18) 2 7 . 126 1 125 8 126 3 125 .8 126 2 126 .2 126 5 126 2 126 2 126 . 2 126 2 126 . 1 126 2 126 1 126 4 126 3 126 3 126 3 126 5 126 2 126 . 1 126 5 126 6 126 6 126 .3 126 2 ( 126 25, 0. 20) 2 8 . 126 2 125 8 126 3 125 .9 126 2 126 .3 126 4 126 3 126 3 126 .2 126 2 126 1 126 2 126 3 126 3 126 3 126 3 126 4 126 .5 126 2 126 .4 126 5 126 .7 126 6 126 .7 126 2 (126 30, 0 20) 2 9 . 126 1 125 .9 126 1 125 .9 126 . 1 126 .4 126 3 126 .3 126 4 126 .2 126 2 126 .2 126 1 126 3 126 1 126 3 126 3 126 4 126 .4 126 2 126 .2 126 5 126 6 126 7 126 .4 126 3 ( 126 27. 0 19) 3 0 . 126 2 125 9 126 2 125 .9 126 1 126 .3 126 4 126 .3 126 3 126 .2 126 2 126 .2 126 3 126 3 126 4 126 3 126 3 126 4 126 3 126 2 126 . 1 126 5 126 6 126 7 126 .6 126 3 ( 126 29. 0 19) 31 . 126 1 125 9 126 2 125 9 126 1 126 .4 126 3 126 3 126 2 126 .2 126 1 126 . 1 126 2 126 2 126 3 126 3 126 2 126 3 126 3 126 2 126 . 1 126 5 126 6 126 6 126 .0 126 2 ( 126 22. 0 18) 3 2 . 126 2 126 0 126 3 125 9 126 1 126 3 126 5 126 3 126 3 126 .2 126 2 126 .2 126 2 126 3 126 4 126 3 126 .2 126 3 126 .3 126 2 126 .3 126 5 126 .6 126 6 126 .2 126 2 ( 126 27, 0 16) 33. 126 1 126 .0 126 2 126 .0 126 1 126 .4 126 3 126 3 126 2 126 .2 126 2 126 .2 126 2 126 3 126 3 126 3 126 3 126 4 126 .3 .126 2 126 . 1 126 5 126 6 126 6 126 .7 126 2 ( 126 28, 0 18) 34. 126 2 126 O 126 1 126 .0 126 3 126 .4 126 4 126 .3 126 3 126 .2 126 2 126 .2 126 3 126 3 126 3 126 3 126 3 126 4 126 .4 126 2 126 .2 126 5 126 7 126 6 126 .6 126 3 ( 126 31 . 0 17) 35. 126 2 126 0 126 3 126 0 126 3 126 4 126 3 126 3 126 2 126 .2 126 2 126 .2 126 3 126 3 126 5 126 3 126 3 126 4 126 4 126 2 126 .2 126 5 126 7 126 7 126 .3 126 3 ( 126 31 , 0 16) 3 6 . 126 1 126 0 126 1 126 0 126 1 126 4 126 5 126 4 126 5 126 .2 126 2 126 .2 126 3 126 3 126 3 126 3 126 3 126 4 126 .5 126 2 126 .2 126 6 126 7 126 7 126 . 1 126 3 ( 126 30 , 0 19) 3 7 . 126 2 126 . 1 126 3 126 . 1 126 0 126 .4 126 4 126 4 126 4 126 .2 126 2 126 .2 126 1 126 3 126 3 126 4 126 3 126 4 126 4 126 2 126 . 1 126 6 126 5 126 7 126 .4 126 3 ( 126 30. 0 16) 3 8 . 126 2 126 1 126 2 126 0 126 1 126 .4 126 4 126 3 126 1 126 .2 126 1 126 .2 126 2 126 3 126 3 126 3 126 2 126 3 126 4 126 2 126 . 1 126 5 126 6 126 6 126 .2 126 3 ( 126 26 . 0 15) 3 9 . 126 2 126 1 126 2 125 9 126 1 126 .4 126 4 126 3 126 3 126 .2 126 2 126 .2 126 2 126 3 126 4 126 3 126 3 126 4 126 4 126 2 126 3 126 4 126 7 126 6 126 .3 126 2 ( 126 2 9 . 0 )6) 4 0 . 126 2 126 0 126 2 126 0 126 0 126 4 126 1 126 3 126 3 126 2 126 2 126 .2 126 0 126 3 126 1 126 3 126 3 126 4 126 3 126 2 126 1 126 5 126 6 126 6 126 5 126 3 ( 126 2S, 0. 17) 41 . 126 3 126 1 126 3 126 0 126 1 126 .4 126 4 126 4 126 3 126 .2 126 2 126 .2 126 3 126 3 126 3 126 3 126 3 126 4 126 4 126 2 126 .3 126 5 126 7 126 7 126 .5 126 4 (126 3 3 , 0 16) 4 2 . 126 1 126 0 126 2 126 .0 126 0 126 .4 126 3 126 4 126 3 126 .2 126 2 126 .2 126 2 126 2 126 3 126 3 126 3 126 4 126 4 126 2 126 1 126 5 126 6 126 6 126 . 1 126 3 ( 126 26, 0. 17) 4 3 . 126 1 126 0 126 2 125 9 126 1 126 4 126 4 126 4 126 4 126 2 126 3 126 .2 126 1 126 3 126 2 126 3 126 3 126 4 126 3 126 2 126 1 126 5 126 6 126 6 126 4 126 3 ( 126 28, 0. 17) 4 4 . 126 2 126 2 126 3 126 0 126 1 126 4 126 4 126 4 126 4 126 3 126 3 126 .2 126 2 126 3 126 4 126 4 126 3 126 5 126 4 126 2 126 3 126 6 126 6 126 7 126 7 126 3 ( 126 35, 0. 17) 4 5 . 126 2 126 .2 126 2 126 .0 126 1 126 .4 126 3 126 4 126 3 126 .3 126 1 126 .3 126 3 126 3 126 3 126 3 126 .3 126 4 126 4 126 2 126 . 1 126 5 126 7 126 6 126 .0 126 4 ( 126 29, 0. 16) 4 6 . 126 1 126 1 126 2 126 0 126 1 126 4 126 4 126 4 126 3 126 2 126 2 126 2 126 1 126 3 126 1 126 3 126 3 126 4 126 3 126 2 126 0 126 5 126 6 126 6 126 .5 126 3 ( 126 27, 0. 17) 4 7 . 126 2 126 2 126 3 126 0 126 1 126 4 126 4 126 4 126 3 126 .2 126 1 126 2 126 2 126 3 126 3 126 3 126 2 126 4 126 4 126 2 126 1 126 5 126 7 126 6 126 4 126 3 ( 126 30, 0. 16) 4B. 126 1 126 1 126 2 125 9 126 1 126 4 126 3 126 3 126 3 126 2 126 2 126 2 126 1 126 3 126 1 126 3 126 2 126 4 126 3 126 2 126 0 126 5 126 6 126 6 126 1 126 2 ( 126 24, 0. 17) 4 9 . 126 2 126 1 126 2 125 9 126 1 126 4 126 3 126 3 126 2 126 .2 126 1 126 2 126 2 126 3 126 1 126 3 126 3 126 3 126 4 126 1 126 0 126 4 126 6 126 6 126 1 126 3 ( 126 24, 0. 16) 50. 126 2 126 1 126 2 126 0 126 0 126 4 126 3 126 4 126 2 126 .3 126 2 126 .2 126 2 126 3 126 1 126 3 126 3 126 4 126 4 126 2 126 2 126 5 126 7 126 6 126 . 1 126 4 ( 126 28, 0. 17) Mean 126 2 125 9 126 2 125 8 126 1 126 3 126 3 126 3 126 3 126 2 126 2 126 1 126 2 126 2 126 3 126 3 126 3 126 3 126 4 126 2 126 2 126 4 126 6 126 6 126 3 126 2 S.O. 0.06 0.21 0.06 0.25 0.08 0. 18 0.09 0. 14 0. 10 0. 1 1 0.06 0. 13 0.08 O. 13 0. 11 0. 10 0. 07 0 . 1 1 0. 07 0. 08 0. 1 1 0. 14 0. 06 0. 07 O. 21 O. 14 G r a n d mean t e m p e r a t u r e > 126 24 C S t a n d a r d d e v l a t I o n • 0 20 c S t a b l 1 I z a t l o n o r R e t o r t come - up t i m e • 18 rain 38 performance of the retort in maintaining a stable cook temperature. Figure 6 is a sample plot of the mean temperature and its standard deviation at each thermocouple location over the entire cook period. This plot can be used to identify cold spots among the thermocouple locations. Temperature fluctuations and the presence of cold areas in the retort can be detrimental to the sterilization of foods. Berry (1979) reported that a temperature deviation of 0.56 C° can result in differences in process lethalities of up to 14% at a temperature of 121.1 °C. Annotated British guidelines (Milleville and Badenhop, 1980) stated that the general tolerance adopted for a uniform temperature distribution is within +1 C° and — 0 . 5 C° of the process temperature. The U.S. reviewers felt tolerance of temperature variation during the cook should not be expressed as a mean deviation, but as a range where the lowest temperature must be equal to or greater than the specified process temperature. To ensure all food containers in a retort load are sufficiently processed, all points in the retort must be at or above the target temperature within one minute after the cook begins. National Food Processors Association (1985) stated that in addition, after the first minute, all thermocouple readings should have a maximum range of 1.67 C ° . ( 3 F°) and should be within 0.83 C° (1.5 F°) of the reference temperature device. In this water retort system, the reference thermometer is located in the drain piping, external to the retort car. Table 6 compares the overall mean retort car temperatures and the reference thermometer temperatures for the two target process temperatures. The table indicates that the range of reference thermometer temperatures fell below the target temperatures, while the range of the retort car temperatures was above the target temperatures. The average thermometer TEMPERATURE DISTRIBUTION WRTER PROCESS RUN COOK10 ° 19A 20 + 21* 22 * 23 * 24* 25 2 26 Y 27 * 28 " " 34 8 35 y 36 6 37 ' 38 1 39" 40 9 41 1 42 " 43 v 29 x 30 1 31 ' 32- 33 44 TEMPERATURE DISTRIBUTION WATER PROCESS RUN COOK10 MEAN AND S.D FOR RLL CHANNELS AT EACH TIME f t | T | 1 4 - | | | { t I 18.0 22.0 26.0 30.0 34.0 38.0 TIME (MIN) 42.0 46.0 50.0 54.0 58.0 FIGURE 5 Sample plot of mean and standard deviations of temperature for all thermocouples at each recorded time during the cook period. o TEMPERATURE DISTRIBUTION WATER PROCESS RUN COOKIO MEAN TEMP I S.D FOR DIFFERENT CHANNELS 00 19 23 27 31 35 39 43 20 22 24 26 28 30 32 34 36 38 40 42 44 21 25 29 33 37 41 <->°. UJ or ZD I— d o L U t \ l . 4—4- + * I •I- t * _4 4) 1)1 CNJ . FIGURE 6 Sample plot of the mean temperature and its standard deviation at each thermocouple location over the entire cook period. 42 temperatures were 1.5 C° below the average retort car temperatures for the 115°C target, and 2.0 C° below for the 125°C target. It should be noted that the setpoint of the retort temperature controller was 1.67 C° (3 F°) above both target temperatures. Table 6 indicates these higher setpoints were necessary to ensure that target temperatures were reached in the retort car. Retort temperatures are monitored using the reference thermometer. The fact that the reference thermometer temperatures were lower than the actual retort car temperatures safely ensured that target process temperatures were exceeded in the retort car. This is a positive feature of the retort design. The large discrepancy between the setpoint temperature and the reference thermometer observed for this specific retort, however, is not advisable and warrants some adjustment. To compensate for temperature variations between the minimum and average retort temperature, it is common practice to adjust the controller setpoint so that slightly higher temperatures than the target retort temperature are produced. This practice ensures that the minimum temperature in the retort will remain above the process temperature. The overall standard deviations of temperature during the cook period for each experimental run were calculated as an indication of the overall variability of temperature in the retort car with respect to time and location. Table 7 compares the range and mean overall standard deviations of temperature observed for different temperature and weir height conditions. Analysis of variance results in Table 8 indicate temperature was significant (p<0.05) in influencing the overall temperature uniformity of the retort car. Overall uniformity was observed to be slightly better at a retort temperature of 115°C (±0.21 C°) than at 125°C (±0 .26 C°) during the 43 TABLE 6 Comparison of overall mean retort car temperatures with reference thermometer temperatures. Overall Mean Car Reference Thermometer Temperature (°C) Temperature (°C) Target 115°C Range 115.6 - 116.4 114.0 - 115.0 Average 115.9 114.4 Target 125°C Range 125.5 - 126.5 123.5 - 124.4 Average 126.0 124.0 cook period. The effect of weir height was not significant (p>0.05). 44 TABLE 7 Range and mean of overall temperature uniformity (26 thermocouples) during the entire cook period. Temperature Weii- Overall Standard Deviation (°C) Height 1 Range (C°) Mean (C°) 115 1 0.18 -• 0.22 0.20 115 2 0.21 • • 0.23 0.22 115 3 0.18 • • 0.25 0.21 125 1 0.25 • • 0.30 0.27 125 2 0.21 • • 0.38 0.28 125 3 0.21 • • 0.26 0.22 'We i r Height 1=29.2 cm, 2 = 31.1 cm, 3 = 34.6 cm TABLE 8 Analysis of variance for overall standard deviations of temperature. Source of df Mean Square F-Ratio Variation Temperature 1 0.10272E-01 5.50 * Weir height 2 0.18500E-02 0.99 ns Temp x Weir 2 0.13389E-02 0.71 ns Error 280 0.18667E-02 ns * - not significant (p>0.05) significant (p<0.05) 45 1. Temperature Distr ibut ion The standard deviations of temperature for all thermocouple locations at each one minute interval during the cook period were used to compare the temperature distribution. Analysis of variance results indicated the effects of temperature, weir height and their interaction were not significant (p>0.05) in influencing the temperature distribution. Table 9 compares the range of temperature uniformity at each minute interval, the average uniformity during the cook period and the range of times for the retort car to stabilize below a standard deviation of 0.30 C ° . During experimental runs, the time for the final thermocouple location to reach the target retort temperature was approximately 18 - 22 minutes. For a greater than 99.5% confidence level, a temperature distribution with a standard deviation of 0.30 C° will have a temperature range of 1.8 C° between thermocouples. Table 9 indicates the range of times for the retort to achieve a distribution below this. Although most locations in the retort car reached the target temperature in 18 minutes, in some cases, several more minutes were required for the temperature distribution to stabilize. During the come-up period, temperatures were observed to increase more rapidly in upper water channels and in the channel entrance regions (Table 5). This is an indication of circulation patterns during the come-up period, where it appears better circulation and a faster temperature rise occurred. In general, experimental runs at 115°C appeared to achieve temperature uniformity more quickly than at 125°C. The range of standard deviations (Table 9) in the raw data indicated that although the poorest uniformity occurred during the initial minutes of the cook, there was significant improvement through the remainder of this period. The average 46 TABLE 9 Range of temperature uniformity at each minute interval, the average uniformity during the cook period, and the range of times for the retort to stabilize. Temperature (°C) Weii-Height 1 Range of Std. Dev. (C°) Mean Std. Dev. (C°) Time to Stabilize Below 0.30 C ° (min) 115 1 0.13 - 0.42 0.19 16 - 20 115 2 0.13 - 0.32 0.21 18 - 19 115 3 0.12 - 0.31 0.20 18 - 19 125 1 0.13 - 0.47 0.22 20 - 22 125 2 0.16 - 0.33 0.21 20 - 20 125 3 0.14 - 0.28 0.20 19 - 21 'Wei r Height 1=29.2 cm, 2 = 31 .1 cm 3 = 34.6 cm standard deviations indicated satisfactory overall temperature distribution during the cook. Although study of standard deviations indicates the uniformity of temperature between thermocouple locations, it does not identify specific areas exhibiting the maximum and minimum extremes in temperature. A comparison of temperatures at individual thermocouple locations is therefore necessary. The analysis of variance presented in Table 10 was based on the pooled mean temperatures of thermocouple locations from all experimental cook periods. Temperature was a significant factor (p<0.05), since the two target temperatures were well spaced (115 and 125°C). The effects of weir height were significant (p<0.05, Table 11) where all 47 TABLE 10 Analysis of variance for pooled mean thermocouple temperatures. Source of Variation df Mean Square F-Ratio Temperature 1 127.84E + 02 228.63E + 03 * Weir height 2 8.5286 152.53 * Water Channel 13 7.9527 142.23 * Position 1 0.77002 13.77 * Interactions Temp x Weir 2 1.1512 20.59 * Temp x Channel 13 0.21156 3.78 * Temp x Position 1 0.51607E-01 0.92 ns Weir x Channel 26 0.57904E-01 1.04 ns Weir x Position 2 0.26389E-02 0.05 ns Channel x Position 13 0.25241 4.51 * Error 429 0.55914E-01 ns - not significant (p>0.05) * - significant (p<0.05) 48 three weir heights were significantly different from each other with respect to overall mean temperature. Weir height 1 exhibited the lowest overall mean temperature, whereas weir height 3 showed the highest overall mean. Water channels significantly influenced temperature (p<0.05), with Table 12 indicating there was some stratification of temperatures within the retort car. The top water channel (11) in the retort car had significantly higher mean temperatures. The mean temperatures were significantly lower at the reference thermometer. In general, within the retort car the lowest water channels exhibited slightly lower mean temperatures than the higher channels. Thermocouple position within the water channels, the entrance and exit regions, was significant (p<0.05). In general, the entrance region showed slightly higher mean temperatures than the exit region. Figure 6 illustrates graphically the slight difference between temperatures at the entrance (odd numbers) and exit (even numbers). The temperature x weir height interaction was significant (p<0.05). Generally, for both experimental temperatures, mean temperatures were largest when the tallest weir height was used. Temperature x water channel interaction was significant (p<0.05). In general, the same pattern in water channel mean temperatures, already discussed, was observed for both experimental temperatures. Water channel x position interaction was significant (p<0.05). Within the retort car, Duncan's multiple range test indicated the coldest locations on average (p<0.05), were the exit regions of water channel 1 (120.6°C) and 2 (120.6°C). In general, the hottest locations (121.2°C) appeared to be the entrance and exit regions of water channel 11 and the exit region of channel 10. Physically, these two exit regions are closest 49 TABLE 11 Duncan's multiple range test comparing pooled mean thermocouple temperatures of different weir heights. Weir Pooled Mean Duncan's He ight 2 Temperature (s.d.) Test 1 (°C) 1 120.6 (5.0) a 2 120.8 (5.1) b 3 121.0 (5.1) c 'Values with the same letter are not significantly different (p>0.05). 2 W e i r height 1=29.2 cm, 2=31.1 cm, 3 = 34.6 cm to where the hot water first contacts the retort car. During experimentation, the coldest locations in the retort car averaged approximately 0.6 C° lower than the hottest locations. Although there was a slight temperature gradient between the top and the bottom water channels, as well as between the entrance and exit regions of the water channels, the overall temperature distribution appeared to be satisfactory. 50 TABLE 12 Duncan's multiple range test comparing pooled mean thermocouple temperatures of different water channels and the reference thermometer. Water Pooled Mean Duncan's Channel Temperature (s.d.) Test 1 (°C) Thermometer 119.2 (4.9) a 2 120.7 (5.1) b 1 120.8 (5.1) b,c 3 120.8 (5.1) b,c,d 6ab 120.9 (5.1) c,d,e 6cd 120.9 (5.1) c,d,e 5 120.9 (5.1) c,d,e 9 120.9 (5.2) d,e,f 7 120.9 (5.2) d,e,f 8 120.9 (5.1) e,f 4 121.0 (5.1) e,f 10 121.1 (5.1) f 11 121.2 (5.1) g 'Values with the same letter are not significantly different (p>0.05). 51 2. Temperature Stability The standard deviations of temperature with respect to each thermocouple location during the entire cook period were used to compare the temperature stability. Table 13 compares the range and means of standard deviations observed for the thermocouple locations during the cook. Temperatures increased during the cook period for unknown reasons. Table 14 summarizes the results of the analysis of variance for the temperature stability comparison. The effect of temperature was significant (p<0.05) with respect to temperature stability. Mean stability was slightly better at a retort temperature of 115°C (±0.11 C°) than at 125°C (±0.16 C°). The effect of weir height was significant (p<0.05, Table 15), where mean temperature stability was significantly better at each temperature when using weir height 3. The effect of water channel on temperature stability was significant (p<0.05). Table 16 indicates mean temperature stabilities in water channels 1 - 10 and at the reference thermometer were not significantly different (p>0.05) from one another. Water channel 11, however, was significantly less stable than the others. During the cook period, steam was injected into the circulating water in order to maintain the temperature. As mentioned, the retort temperature increased slightly during the cook. Since water channel 11 was located just below the top plate where hot water first enters the car, the temperatures at that channel were likely influenced to a greater extent by this temperature increase than at the other water channels. 52 TABLE 13 Range and mean of standard deviations of temperature for each thermocouple during the entire cook period. Temperature Weir Standard Deviation ( ° Q Height ' Rang e (C°) Mean (C°) 115 1 0.05 - 0.62 0.12 115 2 0.04 - 0.42 0.11 115 3 0.05 - 0.23 0.10 125 1 0.06 - 0.70 0.16 125 2 0.05 - 0.39 0.20 125 3 0.05 - 0.25 0.12 Weir Height 1=29.2 cm, 2 = 31.1 cm, 3 = 34.6 cm The position within the water channel was indicated to be a significant factor (p<0.05). In general, the entrance region (±0.12 C°) of the channels exhibited a better mean temperature stability than the exit region (±0 .14 C°). Temperature x weir height interaction was significant (p<0.05). The two significantly worst mean temperature stabilities were observed at 125°C for weir height 1 (±0 .16 C°) and weir height 2 (±0 .20 C°). Temperature x position interaction was significant (p<0.05). The two significantly worst mean temperature stabilities were observed in the entrance (±0 .14 C°) and exit (±0 .18 C°) regions at 125°C. Weir height x water channel interaction was significant (p<0.05). The significantly worst mean temperature stability was observed for water channel 11 with weir height 1 (±0 .32 C°). Water channel x position interaction was significant (p<0.05). In general, the 53 TABLE 14 Analysis of variance for temperature stability. Source of df Mean Square F-Ratio Variation Temperature 1 0.33842 74.18 * Weir height 2 0.99829E-01 21.88 * Water Channel 13 0.37946E-01 8.32 * Position 1 0.10573 23.18 * Interactions Temp x Weir 2 0.50485E-01 11.07 * Temp x Channel 13 0.10245E-02 0.22 ns Temp x Position 1 0.38763E-01 8.50 * Weir x Channel 26 0.16929E-01 3.71 * Weir x Position 2 0.69109E-02 1.51 ns Channel x Position 13 0.33888E-01 7.43 * Error 429 0.45620E-02 ns - not significant (p>0.05) * - significant (p<0.05) 54 TABLE 15 Duncan's multiple range test comparing mean temperature stabilities of different weir heights. Weir Mean Duncan's Height 2 Std. Dev. Test 1 (C°) 3 0.12 a 1 0.14 b 2 0.15 b 1 Values with the same letter are not significant (p>0.05). 2 W e i r height 1=29.2 cm, 2 = 31.1 cm, 3 = 34.6 cm mean temperature stability was not different (p>0.05) between the entrance and exit regions of all channels, except for water channels 9 and 11. The worst stability was observed in the exit region of channel 11. The results indicated satisfactory temperature stability in the retort car during the cook period. 55 TABLE 16 Duncan's multiple range test comparing mean temperature stabilities of different water channels and the reference thermometer. Water Mean Duncan's Channel Std. Dev. Test 1 (C°) 1 0.11 a 6cd 0.11 a 5 0.11 a 9 0.12 a 7 0.12 a 6ab 0.12 a 4 0.12 a 10 0.13 a 2 0.13 a Thermometer 0.13 a 11 0.19 b 'Values with the same letter are not significant (p>0.05). 56 B. HEAT TRANSFER The influence of temperature, weir height, tray level and tray position on the heating and cool ing parameters, f^, i^, and j were studied. These parameters were used to evaluate the heat transfer distribution in the water processing system. 1. Heat ing Rate Index (f^) Heating rate indices of the teflon transducers were calculated to determine heat transfer efficiency within the retort car. Stumbo (1973) defined f^ as the time, in minutes, required for the straight line portion of a heat penetration curve to traverse one log cycle. Ball and Olson (1957) essentially defined a heat penetration curve as a time-temperature curve for a specific point in an object, which is undergoing changes in heat energy. The rate of temperature change at the center of a transducer during heating is dependent on the surface heat transfer coefficient, the thermal diffusivity of the teflon and the temperature gradients which cause heat transfer. Ramaswamy (1983) pointed out the mathematical dependence of the heating rate index on the heat transfer coefficient, where f^ is determined by the Biot number, which in turn is determined by the heat transfer coefficient. The relationship between thermal diffusivity (a) and heating rate index (f^) for conduction heating of a brick geometry was reported by Olson and Jackson (1942): 0.933 a = • (1) f h (1/a2 + 1/b2 + 1/c 2 ) 57 where a, b and c are one-half the brick thickness, length and width, respectively. Analysis of variance results in Table 17 indicate temperature, tray level, tray position and several interactions significantly (p<0.05) influenced heating rate indices. The influence of weir height was not a significant main effect (p>0.05). A smaller mean f^ value was observed at a retort temperature of 125°C (16.58 +1.07 min) than at 115°C (16.73 +1.17 min), suggesting better heat transfer conditions occurred at the higher temperature. Jackson and Olson (1940) reported fj values were independent of the retort temperature and the initial product temperature when extremely high heat transfer coefficients were present. A limiting heat transfer coefficient exists for the transducers, below which, changes in the heat transfer coefficient will affect fj values. Therefore, increased surface heat transfer coefficients at the higher temperature may be responsible for the difference. McCinnis (1986), using a similar retort, reported larger heat transfer coefficients ac water temperatures increased during heating. Heat transfer conditions between tray levels are shown for the three experimental weir heights in Tables 18a, 18b and 18c. Table 18b suggests heat transfer conditions between trays were very uniform for weir height 2. Weir height 1 exhibited a significantly (p<0.05) larger f^ value for tray 10 than the other trays, indicating poor heat transfer in comparison. This observation provides evidence for the concern that, with weir height 1 (the lowest weir height), the overall water flow rate might be insufficient to sustain an adequate water flow through the uppermost water channels. Temperature distribution data did not identify any significant temperature deviations in this region. Weir height 3 had the most variability between trays, while weir height x tray level interaction indicated the 58 TABLE 17 Analysis of variance for heating rate indices V-Source of Variation df Mean Square F-Ratio Temperature 1 1.7956 5.75 * Weir height 2 0.33532 1.07 ns Level 5 2.8608 9.16 * Position 2 47.841 153.19 * Interactions Temp x Weir 2 0.28676 0.92 ns Temp x Level 5 0.81050 2.60 * Temp x Position 2 0.84373 2.70 ns Weir x Level 10 1.1803 3.78 * Weir x Position 4 0.8922 2.86 * Level x Position 10 18.497 59.23 * Error 280 0.31229 ns - not significant (p>0.05) * - significant (p<0.05) 59 widest range of f| values between trays occurred with weir height 1. Differences in mean fj values may be explained by differences in flow rates through the water channels caused by non-uniform circulation patterns in the retort car. As mentioned, slower water flow rates result in smaller heat transfer coefficients and consequently slower heat transfer rates. Table 19 indicates the variability of f^ values among the tray positions studied. The entrance position exhibited significantly (p<0.05) smaller mean heating rate indices indicating the most efficient heat transfer was at the tray entrance region, fol lowed by the tray exit and tray middle regions. This corresponds with the results of McCinnis (1986), who found heat transfer coefficients measured at the exit end of each tray during heating to be significantly lower than values measured at the entrance. McGinnis (1986) cited Deissler's (1955) study of flow behavior in pipes and channels having an abrupt entrance, in which the thickness of the local laminar sub-layer in the entrance region of the flow channel was found to be less than in the remaining length of channel. Similar less severe conditions may occur in the exit region of the trays, which would explain its smaller mean f^ value compared to the middle region. The baffled distribution plate at the entrance end of the retort car may create turbulent conditions producing better heat transfer at the entrance region. In general, tray level x tray position interaction suggested the most efficient heat transfer conditions occurred in the entrance positions of all tray levels, and the least efficient occurred in the exit and middle positions of trays 1 and 10. 60 TABLE 18a Duncan's multiple range test comparing values associated with different trays for weir height 1. Tray Mean f h (s .d.) Duncan's (min) Test 1 8 16.39 (0.49) a 5 16.42 (0.45) a 1 16.59 (0.67) a 3 16.60 (2.06) a 6 16.63 (0.56) a 10 17.51 (1.51) b 1 Values with the same letter are not significantly different (p>0.05). Duncan's multiple range weir height 2. test TABLE 18b comparing fj values associated with different trays for Tray Mean f h <*.d.) Duncan's (min) Test 1 8 16.52 (0.58) a 5 16.54 (0.41) a 6 16.66 (0.80) a 1 16.76 (1.24) a 3 16.79 (2.09) a 10 16.80 (0.74) a 'Values with the same letter are not significantly different (p>0.05). 61 Duncan's multiple range weir height 3. TABLE 18c test comparing f^ values associated with different trays for Tray Mean f| (s.d.) Duncan's (min) Test 1 8 16.12 (0.63) a 5 16.53 (0.37) b 6 16.56 (0.85) b 1 16.59 (0.58) b 10 16.64 (0.77) b 3 17.12 (2.20) c Values with the same letter are not significantly different (p>0.05). TABLE 19 Duncan's multiple range test comparing f^ values associated with different tray positions. Position Mean fj (s.d.) Duncan's (min) Test 1 Entrance 16.02 (0.60) a Exit 16.59 (1.02) b Middle 17.35 (1.22) c 1 Values with the same letter are not significantly different (p>0.05). 62 2. C o o l i n g Rate Index (f ) ° c Cool ing rate indices of the teflon transducers were calculated to determine the cooling efficiency within the retort car. Stumbo (1973) defined f as the time, in minutes, required for the straight line portion of a cooling curve to traverse one log cycle. The concepts are essentially identical to those of heating rate indices, except heat energy is removed. Faster cooling rates (smaller f ) reduce product temperatures more quickly, thus exposing foods to less time at quality damaging temperatures. Conversely, more efficient cooling periods contribute less to product lethality. Cool ing rates are generally considered to be greater than or equal to heating rates during thermal processing. Increased water flow rates during cooling for this water system possibly could have produced smaller f values than f^ values. Pooling all collected data, the grand mean f^ (16.65 +1.12 min) and f (16.57 + 0.83) values were nearly identical. Analysis of variance results in Table 20 indicate temperature, tray level, tray position and several interactions significantly (p<0.05) influenced cooling rate indices. Weir height did not significantly (p>0.05) affect overall cooling rate indices. Transducer temperatures were allowed to stabilize near the experimental retort temperature before the cooling period was started. A smaller mean f value was observed at a retort temperature of 125°C (16.46 ±0.73 min) than at 115°C (16.67 ±0.90 min), suggesting better heat removal conditions occurred when cooling began at the higher temperature. Cool ing water temperatures were fairly consistent between experimental runs, however, the higher retort temperature may influence cooling 63 TABLE 20 Analysis of variance for cooling rate indices (f c). Source of Variation df Mean Square F-Ratio Temperature 1 3.6354 23.26 * Weir height 2 0.65595E-01 0.42 ns Level 5 4.2638 27.28 * Position 2 21.461 137.29 * Interactions Temp x Weir 2 0.56723E-01 0.36 ns Temp x Level 5 2.2977 14.70 * Temp x Position 2 0.96011E-01 0.61 ns Weir x Level 10 0.45089 2.88 * Weir x Position 4 0.60125 3.85 * Level x Position 10 9.1098 58.28 * Error 280 0.15632 ns - not significant (p>0.05) * - significant (p<0.05) 64 water temperatures enough to create differences in surface heat transfer coefficients. Tung et al. (1984a) found the thermal diffusivity of certain test materials to be dependent on temperature. Changes in transducer thermal diffusivities caused by different initial temperatures before cooling should be considered. It may be possible that thermal diffusivity changes have contributed to some extent t o . the observed differences in f values; however, this possible effect was not studied in detail. Tables 21a, 21b and 21c show the variability of f values observed between tray levels for the three weir heights. Weir height 2 exhibited the most uniform f values between tray levels, although not as uniform as during the cook period. In general, the three weir heights show very similar patterns of f variability between trays. Excluding tray 1, the most efficient cooling seems to have been in the middle trays (5, 6, 8), with less efficient cooling in the very top tray (10) and the lower tray (3). The increased water flow rate during cool ing may have created increased circulation in the middle and upper/middle trays in comparison to the very top and bottom regions. Tray 1, being on the bottom of the retort car, probably experienced further heat removal by water flow in the bottom of the retort shell, which was in contact with the retort car. The range of f • values among tray levels was the same for all three weir heights. In general, the tray level x tray position interaction suggested cooling conditions were most efficient in the entrance region of most trays and least efficient in the middle positions of trays 1 and 3 and the middle and exit regions of tray 10. Table 22 indicates cooling was most efficient overall in the entrance region of trays, followed by the exit and middle regions. These are similar to the results TABLE 21a Duncan's multiple range test comparing f values associated with different trays for weir height 1. Mean f (s.d.) Duncan's Tray c Test 1 (min) 5 16.14 (0.62) a 8 16.40 (0.23) a,b 1 16.58 (0.70) b 6 16.60 (0.51) b 3 16.89 (1.01) c 10 16.98 (0.89) c 1 Values with the same letter are not significantly different (p>0.05). TABLE 21b Duncan's multiple range test comparinj g f values c associated with different trays for weir height 2. Mean f (s.d.) Duncan's Tray (min) Test 1 8 16.29 (0.29) a 5 16.33 (0.56) a 1 16.36 (0.79) a 6 16.43 (0.68) a 10 16.85 (0.77) b 3 17.08 (1.41) b Values with the same letter are not significantly different (p>0.05). 66 TABLE 21c Duncan's multiple range weir height 3. test comparing f values associated with different trays for Mean f (s.d.) Duncan's Tray c (min) Test 1 1 16.12 (0.84) a 8 16.42 (0.42) b 5 16.47 (0.50) b 6 16.64 (0.85) b 10 16.68 (0.77) b 3 16.99 (1.42) c Values with the same letter are not significantly different (p>0.05). TABLE 22 Duncan's multiple range test comparing f values associated with different tray positions. Position Mean f (s.d.) Duncan's c (min) Test 1 Entrance Exit Middle 16.11 (0.50) a 16.60 (0.64) b 17.00 (1.01) c Values with the same letter are not significantly different (p>0.05). 67 observed for heating rate indices, but represent heat removal instead. The increased flow rate of the cooling water, compared to the heating water flow rate, did not create uniform cooling rates across the length of the trays. Weir height x tray position interaction suggested that the effect of position is not the same at each weir height. 68 3. Heating Lag Factor (j^) Heating lag factors were calculated as an indication of the retort system's effectiveness in establishing a constant heat transfer rate and for subsequent lethality calculations. These factors, when multiplied by the difference between the retort temperature and the initial product temperature, locate the intersection of the extension of the straight line portion of a heat penetration curve and a vertical line representing beginning of process or zero time (Stumbo, 1973). Heating lag factors are mathematically defined as: r ih where T r is the retort temperature, T.^ is the initial temperature of the transducers and T is the pseudo-initial temperature for heating. The pseudo-initial temperature is generally calculated using the 4 2 % come-up effectiveness criteria of Ball (1923). Calculation of jj is extensively covered by Ball and Olson (1957), who indicated that the could be strongly influenced by the initial temperature difference between the retort heating medium and the centerpoint of the heated object; and the come-up time. The heating lag factor is a function of the position of the measured point, the shape of the object being heated, and the initial temperature distribution within the object (Ball and Olson, 1957). The range of initial temperatures between transducers varied from 0.7 - 5.1 C° for individual experimental runs. Initial transducer temperatures ranged from 11.6 - 20.8°C, with the largest difference between transducers for an experimental run being 5.1 C ° . Although average come-up times were approximately 18 minutes, there was variability 69 in come-up times with respect to specific locations within the retort car. Smaller values were associated with faster establishment of a constant heating rate for a specific retort temperature. Analysis of variance results in Table 23 indicate temperature, weir height, tray level, tray position and several interactions significantly (p<0.05) influenced heating lag factors. A larger mean value was observed at a retort temperature of 125°C (0.625 ±0.107) than at 115°C (0.584 ±0.093). This difference is deceiving in terms of comparing the effectiveness of the retort temperatures in establishing a constant heating rate, since Equation 2 shows a higher retort temperature will generate a larger value. The different fj values observed for the two temperatures will have also influenced the calculation. It is conceivable, even with pre-heated water, that the retort would be able to establish constant heat transfer conditions sooner for a lower retort temperature. Table 24 indicates weir height 3 was not as effective in establishing a constant heat transfer rate as the two shorter weir heights. During the come-up period, the time for water to fill the retort car and begin flowing through the circulation system was different for each weir height used. The longest amount of time occurred for weir height 3. As a result of this longer delay, steam injected water would not have been returned to the retort car as soon as for the shorter weir heights, thus creating a slightly longer come-up time for weir height 3. Temperature x weir height interaction indicated the combination of weir height 1 and 115°C produced the smallest j ^ values, while weir height 3 and 125°C produced the largest values. 70 TABLE 23 Analysis of variance for heating lag factors <in>. Source of Variation df Mean Square F-Ratio Temperature 1 0.13290 136.88 * Weir height 2 0.37876E-01 39.01 * Level 5 0.38103 392.44 * Position 2 0.16978 174.87 * Interactions Temp x Weir 2 0.17512E-01 18.04 * Temp x Level 5 0.56048E-02 5.77 * Temp x Position 2 0.67311E-02 6.93 * Weir x Level 10 0.11343E-01 11.68 * Weir x Position 4 0.15360E-02 1.58 ns Level x Position 10 0.25211E-01 25.97 * Error 280 0.97093E-03 ns - not significant (p>0.05) * - significant (p<0.05) 71 Tables 25a, 25b and 25c show the general pattern of variability between tray levels for the three weir heights. A gradient of values is apparent, with small values associated with upper trays and larger values associated with lower trays. During temperature distribution tests, retort temperatures rose more quickly in the upper tray regions, which this gradient reflects. Weir height x tray level interaction showed weir height 1 had the most uniform values between trays, whereas weir heights 2 and 3 had the widest range of values. Table 26 shows the entrance position had a significantly smaller mean than the middle and exit positions. During the come-up period, the entrance side of the retort car was exposed first to changing heating conditions. Temperature distribution data recorded definite temperature gradients between the entrance and exit regions of the retort car, which showed that slower temperature increases occurred in the exit regions. Temperature x tray position interaction indicated similar results occurred for both experimental temperatures. Tray level x tray position interaction suggested the middle and exit regions of tray 1 produced the largest values, whereas all of tray 10 and the entrance regions of trays 3 - 8 exhibited the largest j , values. 72 Duncan's multiple range heights. test TABLE 24 comparing values associated with different weir Weir Mean (s.d.) Duncan's He ight 2 Test 1 1 0.594 (0.078) a 2 0.594 (0.093) a 3 0.626 (0.119) b 'Values with the same letter 2 W e i r Height 1=29.2 cm, 2-are not significantly different = 31.1 cm, 3 = 34.6 cm (p>0.05). Duncan's multiple range weir height 1. test TABLE 25a comparing values associated with different trays for Tray Mean (s.d.) Duncan's Test 1 10 0.551 (0.037) a 8 0.558 (0.043) a 5 0.559 (0.045) a 6 0.572 (0.052) a 3 0.612 (0.076) b 1 0.710 (0.069) c 'Values with the same letter are not significantly different (p>0.05). 73 Duncan's multiple range weir height 2. TABLE 25b test comparing values associated with different trays for Tray Mean (s.d.) Duncan's Test 1 10 0.513 (0.019) a 8 0.543 (0.027) b 6 0.565 (0.043) b,c 5 0.569 (0.047) c 3 0.630 (0.089) d 1 0.743 (0.075) e Values with the same letter are not significantly different (p>0.05). 74 Duncan's multiple range weir height 3. TABLE 25c test comparing values associated with different trays for Tray Mean jj (s.d.) Duncan's Test 1 10 0.536 (0.018) a 8 0.566 (0.041) b 5 0.577 (0.043) b 6 0.583 (0.056) b 3 0.665 (0.098) c 1 0.830 (0.098) d V a l u e s with the same letter are not significantly different (p>0.05). TABLE 26 Duncan's multiple range test comparing values associated with different tray positions. Position Mean (s.d.) Duncan's Test 1 Entrance 0.559 (0.073) a Middle 0.627 (0.115) b Exit 0.628 (0.089) b Values with the same letter are not significantly different (p>0.05) 75 4. C o o l i n g Lag Factor (j^) Cool ing lag factors were calculated as an indication of the retort system's effectiveness in establishing a constant cool ing rate and for subsequent lethality calculations. These factors, when multiplied by the difference between the product temperature when cooling starts and the cooling water temperature, locate the intersection of the extension of the straight-line portion of the cooling curve and the vertical line representing the beginning of cooling (Stumbo, 1973). Mathematically, cooling lag factors are defined as: T - T . = w p|c ( 3 ) T - T. w ic where T w is the cooling water temperature, T. is the temperature of the transducers when cooling started and T . c is the pseudo-initial temperature for cooling. The principles associated with j values are similar to those for values. Larger j values can be an indication of extended periods of time in establishing a constant rate of cooling. Analysis of variance results in Table 27 indicate temperature, tray level, tray position and several interactions significantly (p<0.05) influenced cooling lag factors. Cool ing lag factors were slightly larger for retort temperatures of 115°C (1.441 ±0.077) than of 125°C (1.428 ±0.088). This is deceiving, since Equation 3 shows that a higher transducer temperature at the start of the cool will generate a slightly smaller j ' c value. The different f values observed for the two temperatures will have also influenced the i calculation. 'c 76 TABLE 27 Analysis of variance for cooling lag factors (ic>. Source of Variation df Mean Square F-Ratio Temperature 1 • 0.14029E-01 3.99 * Weir height 2 0.10045E-01 2.86 ns Level 5 0.72355E-01 20.59 * Position 2 0.18146 51.64 * Interactions Temp x Weir 2 0.54530E-01 15.52 * Temp x Level 5 0.32927E-01 9.37 * Temp x Position 2 0.46876E-02 1.33 ns Weir x Level 10 0.18489E-01 5.26 * Weir x Position 4 0.43835E-02 1.25 ns Level x Position 10 0.29544E-01 8.41 * Error 280 0.35138E-02 ns - not significant (p>0.05) * - significant (p<0.05) 77 Tables 28a, 28b and 28c indicate the general patterns of j variability between tray levels observed for each weir height. Complete uniformity of j values did not occur with use of any weir height. A slight gradient of values appeared between the upper and lower trays for weir heights 1 and 2, whereas weir height 3 exhibited a mixed pattern. Temperature distribution data indicated that during the initial part of cooling, temperatures decreased slightly faster in the bottom tray regions. Co ld water has a greater density than hot water, and tends to sink when the two are mixed. Weir height x tray level interaction showed the range of j values to be greatest for weir height 3, whereas the range for weir heights 1 and 2 were equal. The entrance region of trays is shown in Table 29 as having the smallest j values. The entrance regions will be exposed to cooling conditions before the other positions as a result of the retort design. Temperature distribution data showed temperature gradients existed between the entrance and exit regions of water channels, particularly during the early part of the cooling period. In general, tray level x tray position interaction showed the exit and middle positions of tray 10 exhibited the largest j values, while the entrance positions of trays 3 - 8 exhibited the smallest. 78 Duncan's multiple range weir height 1. TABLE 28a test comparing j values associated with different trays for Tray Mean j (s.d.) X Duncan's Test 1 1 1.375 (0.017) a 3 1.392 (0.048) a;b 6 1.417 (0.056) a,b,c 8 1.430 (0.058) b,c,d 10 1.460 (0.038) c ,d 5 1.474 (0.094) d 'Values with the same letter are not significantly different (p>0.05). Duncan's multiple range weir height 2. TABLE 28b test comparing j values associated with different trays for Tray Mean j (s.d.) X Duncan's Test 1 3 1.387 (0.061) a 1 1.400 (0.037) a,b 8 1.441 (0.063) b,c 6 1.448 (0.044) c 5 1.462 (0.092) c 10 1.473 (0.053) c Values with the same letter are not significantly different (p>0.05). 79 Duncan's multiple range weir height 3. test TABLE 28c comparing j values associated with different trays for Tray Mean j (s.d.) 'c Duncan's Test 1 6 8 1.383 (0.170) 1.398 (0.144) a a,b 3 1.432 (0.061) b,c 1 1.443 (0.029) c 5 1.460 (0.122) c 10 1.548 (0.067) d 1 Values with the same letter are not significantly different (p>0.05). Duncan's multiple range positions. test TABLE 29 comparing j values associated with different tray Position Mean i (s.d.) 'c Duncan's Test 1 Entrance 1.391 (0.056) a Middle 1.441 (0.094) b Exit 1.472 (0.091) c 1 Values with the same letter are not significantly different (p>0.05). 80 5. Compar i son W i t h Steam Processing Pure steam experiments were performed to compare heat transfer efficiency with the water system. Heating and cooling parameters were calculated for a retort temperature of 125°C. Table 30 compares the overall means and standard deviations calculated from the pure steam experiments with the parameters calculated from the water experiments. The mean heating rate index was approximately 5 .5% smaller for pure steam than for water, suggesting heat transfer was more efficient using the steam environment. When the heat transfer coefficient is the limiting factor in heat transfer, changes in this value will affect heat transfer and consequently f^ values. Pflug (1964) reported water to have a smaller heat transfer coefficient than pure steam. Peterson and Adams (1983) found surface heat transfer coefficients increased with higher water flow rates. The fj values calculated for the water system may possibly be improved if an increased water flow rate could be used. The mean cooling rate index was approximately 7.2% smaller for the water system than the procedure used for the steam experiments. Water was the cool ing medium in both cases, however, the results suggest that cooling was more efficient with the water system. Cool ing water did not flow past the transducers as fast in the steam experiments because the retort shell contained a larger volume of water than the retort car. At 125°C, the water system's f value was smaller than the value, signifying that heat removal had been more efficient than heat input. Water flow rates were observed to be more than double during the cooling period. The mean heating lag factor was smaller for water than for steam 81 TABLE 30 Comparison of heating and cooling parameters calculated using pure steam and water immersion processes at 125°C. Steam Water Heating Rate Index Mean (min) 15.71 16.58 Std. Dev. (min) 0.71 1.07 Cooling Rate Index Mean (min) 17.65 16.46 Std. Dev. (min) 1.47 0.73 Heating Lag Factor Mean (min) 0.691 0.623 Std. Dev. (min) 0.102 0.107 Cool ing Lag Factor Mean (min) 1.276 1.428 Std. Dev. (min) 0.081 0.098 82 experiments. This difference is a reflection of the difference in come-up times (water - 18 min, steam - 6 min) and the heating rates observed. The mean cooling lag factor was smaller for the steam experiment cooling method than the water retort cooling method. The results reflect the difference in cooling rates observed and the time for each method to establish a constant cooling rate. Assuming constant cooling conditions were established within similar time frames, smaller j values indicate less efficient turnaround from heating into cooling when using the water immersion heating. 83 C. LETHALITY DISTRIBUTION The effectiveness of the water system in providing uniform sterilizing conditions throughout the retort car was determined by calculating theoretical centerpoint lethality values for the food-simulating transducers. Normally, process lethalities are a function of food product and retort variabilities. However, since the transducers were standardized with respect to geometry, size and thermal characteristics, differences in F 0 values at various retort locations will identify non-uniform sterilizing conditions resulting from retort performance. Sterilizing uniformity throughout a retort will prevent over-processing in some areas of the retort, since process times are developed to compensate for the worst sterilizing conditions. Stumbo's formula method was chosen for the F 0 calculations, which accounts for the sterilizing effect of heating and cool ing. Sterilizing effects of the come-up time were not considered in the calculations. It is interesting to note that experimental j values were very similar to the assumption of Ball's formula method, that j c = 1.41. Both methods assume f^ is equal to f . For these reasons, no appreciable differences in results were found using Ball's method. The F 0 values calculated were based on an arbitrary process time of 30 minutes, a retort temperature of 121.1 °C (250°F), an initial transducer temperature of 15°C (59°F) and z = 1 0 C° (18 F°). Analysis of variance results in Table 31 indicate temperature, weir height, tray level, tray position and several interactions significantly (p<0.05) influenced the sterilizing conditions in the retort. 84 TABLE 31 Analysis of variance for lethality values (F 0 ) calculated using Stumbo's formula method. Source of Variation df Mean Square F-Ratio Temperature 1 5.4886 22.08 * Weir height 2 1.2868 5.18 * Level 5 35.049 141.01 *' Position 2 98.043 394.43 * Interactions Temp x Weir 2 0.09897 0.40 ns Temp x Level 5 0.80635 3.24 * Temp x Position 2 0.84518 3.40 * Weir x Level 10 2.5780 10.37 * Weir x Position 4 0.71653 2.88 * Level x Position 10 20.951 84.29 * Error 280 0.24857 ns - not significant (p>0.05) * - significant (p<0.05) 85 TABLE 32 Duncan's multiple range test comparing Stumbo's lethality values (F 0 ) associated with different weir heights. Weir Mean F 0 (s.d.) Duncan's Height 2 (min) Test 1 1 9.68 (1.32) a 3 9.71 (1.55) a 2 9.88 (1.51) b 1 Values with the same letter are not significantly different (p>0.05). 2 W e i r Height 1=29.2 cm, 2 = 31.1 cm, 3 = 34.6 cm Overall sterilizing conditions were apparently better during retort temperatures of 115°C (9.88 ±1.42 min) than of 125°C (9.62 ±1.51 min) as evidenced by the F 0 values. This is interesting, since f^ calculations showed heat transfer was slightly better during processing at 125°C. O n the other hand, f calculations showed heat removal was better during cooling at 125°C and consequently a smaller sterilizing contribution was made than for cooling at 115°C. The total contribution of the heating and cooling parameters to F 0 calculations seems to have been slightly greater using parameters obtained at the lower temperature. Weir height influenced the overall sterilizing conditions of the retort. Mean sterilizing conditions were best using weir height 2, while there was no difference between weir heights 1 and 3 as Table 32 indicates. Weir height was shown to have had no significant (p>0.05) influence on mean f^, and j values, however, values were significantly (p<0.05) affected. In this case, the value is an indication of the overall ability of the retort to establish constant heating conditions. 86 Table 24 showed that weir heights 1 and 2 were most efficient at establishing this condition. Considering all parameters, weir height 2 provided more effective overall sterilizing conditions than weir height 1. Non-uniform sterilizing conditions between tray levels are indicated by Tables 33a, 33b and 33c. Sterilization patterns differed somewhat between weir heights, but generally the upper trays exhibited higher lethality conditions than the lower trays. Of particular interest is tray 10, which showed a significantly (p<0.05) lower F 0 value with weir height 1 than with the taller weirs. Considering the range of F 0 values for each weir height, the overall uniformity between trays was equally good when using weir heights 1 and 2. Weir height x tray level interaction indicated tray 1 with weir height 3 resulted in the poorest sterilizing conditions (p<0.05), while the top two trays (8 and 10) with weir heights 2 and 3 equally exhibited the highest lethality conditions. Table 34 indicates sterilizing conditions were greatest in the entrance region of trays, followed by the exit regions and the middle regions. This is not surprising, since similar patterns were observed for heat transfer conditions during heating. Temperature x tray position interaction indicated the entrance position for both experimental temperatures provided similar conditions, while the lowest lethality conditions were observed for the middle position at 125°C. Weir height x tray position interaction indicated all three weir heights had the same entrance/exit/middle pattern. In general, locations with the highest F 0 values appeared to have been the entrance regions of the upper trays, with the lowest in the middle and exit regions of the bottom trays. 87 TABLE 33a Duncan's multiple range test comparing Stumbo's lethality values (F 0 ) associated with different trays for weir height 1. Tray Mean F 0 (s.d.) (min) Duncan's Test 1 1 8.54 (1.07) a 10 9.44 (0.91) b 3 9.62 (2.38) b,c 6 9.88 (0.68) c 8 10.29 (0.43) d 5 10.32 (0.65) d 'Values with the same letter are not significantly different (p>0.05). Duncan's multiple range different trays for weir test height TABLE 33b comparing Stumbo's 2. lethality values (F 0 ) associated wit! Tray Mean F 0 (s.d.) (min) Duncan's Test 1 1 8.42 (1.46) a 3 9.35 (2.54) b . 5 10.22 (0.63) c 6 10.22 (0.98) c 8 10.48 (0.46) c 10 10.59 (0.69) c Values with the same letter are not significantly different (p>0.05). 88 TABLE 33c Duncan's multiple range test comparing Stumbo's lethality values (F 0 ) associated with different trays for weir height 3. Tray Mean F 0 (s.d.) Duncan's (min) Test 1 1 8.04 (0.96) a 3 8.85 (2.56) b 5 9.99 (0.67) c 6 10.11 (0.79) c 8 10.55 (0.52) d 10 10.70 (0.76) d 'Values with the same letter are not significantly different (p>0.05). TABLE 34 Duncan's multiple range test comparing Stumbo's lethality values (F 0 ) associated with different tray positions. Position Mean F 0 (s.d.) Duncan's (min) Test 1 Middle 8.88 (1.66) a Exit 9.61 (1.08) b Entrance 10.77 (0.84) c Values with the same letter are not significantly different (p>0.05). 89 D. RETORT PRESSURE The efficacy of a retort system is dependent upon its ability to attain a target retort pressure and to maintain pressure stability throughout the thermal process. This is critical during processing with flexible containers, such as retort pouches, where a consistent target overpressure is necessary to maintain compression on the non-condensible gases within the package and to improve heat transfer and protect the package integrity. As described earlier, the target retort pressure of 172.37 kPa produced theoretical overpressures of 101.79 kPa at 115°C and 39.70 kPa at 125°C. These overpressures were consistent with those cited in the literature. The retort pressure and stability were evaluated during the cook period, 1 8 - 5 0 minutes, and during the cool period, 50 - 75 minutes. Table 35 summarizes the retort pressure data collected during the cook and cool periods. The grand mean retort pressure during the cook period was 174.0 kPa (25.2 psig) with a standard deviation of 1.29 kPa and coefficient of variation of 0.74% for the combined experimental runs. The minimum and maximum mean pressures recorded were 171.50 and 176.09 kPa. Statistical analysis indicated no significant difference (p>0.05) in retort pressure resulted from retort temperature, weir height or temperature x weir height interactions. The results indicate the effectiveness of the retort system in attaining the target pressure during the cook period. The grand mean retort pressure during the cool period was 171.6 kPa (24.9 psig) with a standard deviation of 1.14 kPa and coefficient of variation of 0.66% for the combined experimental runs. The minimum and maximum mean pressures 90 recorded were 169.36 and 173.06 kPa. Statistical analysis indicated no significant difference (p>0.05) in retort pressure resulted from retort temperature, weir height or temperature x weir height interactions. The results indicate the effectiveness of the retort system in attaining the target pressure during the cool period. The retort pressure stabilities during the cook and cool periods were evaluated using the standard deviation of pressures during these periods and the range of minimum and maximum observed pressures. The grand mean standard deviation during the cook period was 0.95 kPa with a range of 0.79 - 1.07 kPa for the combined experimental runs. The minimum and maximum pressures, for any run, during this period were 170.7 and 178.7 kPa. No significant differences (p>0.05) in pressure stability resulted from retort temperature, weir height or temperature x weir height interactions. The results indicate good pressure stability during the cook period. The grand mean standard deviation during the cool period was 2.59 kPa with a range of 0.93 - 4.61 kPa for the combined experimental runs. The minimum and maximum pressures, for any run, during this period were 150.9 and 177.2 kPa. No significant differences (p>0.05) in pressure stability resulted from retort temperature, weir height or temperature x weir height interactions. These results indicate pressure stability during the cool period was acceptable but lower than during the cook period. A significant pressure drop of as much as 20 kPa was consistently observed during the initial minutes of all cool periods and is illustrated in the form of a spike in Figure 7. This pressure drop occurred when steam in the steam/air mixture occupying the retort headspace collapsed as cool ing water entered the retort and 91 TABLE 35 Summary of retort pressure data during I cook and cool periods. Cook Coo l Grand Mean Pressure (kPa) 174.0 171.6 Standard Deviation (kPa) 1.29 1.14 Coefficient of Variation (%) 0.74 0.66 Minimum Observed Mean (kPa) 171.50 169.36 Maximum Observed Mean (kPa) 176.09 173.06 Grand Mean Standard Deviation (kPa) 0.95 2.59 Range (kPa) 0.79 - 1.07 0.93 - 4.61 Minimum Observed Pressure (kPa) 170.7 150.9 Maximum Observed Pressure (kPa) 178.7 177.2 hot water was being shunted back to the reservoir. The air make-up supply was unable to compensate quickly enough for the void created by the steam collapse and water transfer, thus resulting in a brief but significant pressure drop. Once the retort system had totally compensated for this pressure drop, pressure stability was good through the remainder of the cool period as evidenced by the small range of standard deviations. This pressure drop is potentially troublesome, since package damage would result if the pressure was to fall substantially below the internal pouch pressure. With reference to the target overpressure of the thermal process, it is advisable to slightly increase the retort pressure just prior to initiating the cool period to ensure the pressure will not fall below the internal pouch pressure. 180 Cook period 92 175 o- o. o- o. o o- , o . 0 o < > o o ' ° o . 0 . 0 o-o o ° S. 170 D W V) Q_ 165-160 155 150 15 20 25 30 35 40 45 50 55 Time (minutes) Cool period 180-i 175-S. 170 07 Z3 V) 0) 165-160-O. . .o- -o-155-150 50 55 60 65 70 Time (minutes) 75 FIGURE 7 Example pressure histories of an experimental run. V. C O N C L U S I O N S Temperature studies in the water immersion retort indicated temperature distribution and stability to be quite satisfactory. Overall temperature uniformity throughout the cook period was slightly more favorable at a retort temperature of 115°C than at 125°C. Variation of weir height did not affect the overall temperature uniformity. Temperature distributions based on standard deviations of thermocouple readings were not influenced by retort temperature or weir height. Mean standard deviations during the cook period ranged from 0.19 to 0.22 C ° . A comparison of mean thermocouple readings identified regions displaying variable temperatures. A slight temperature gradient between warmer top and cooler bottom water channels was indicated during the cook period. Very small temperature differences were also indicated between the entrance and exit regions of water channels. The coldest regions in the retort, identified as the exit regions of water channels 1 and 2, averaged approximately 0.6 C° lower than the hottest regions, identified as the entrance and exit regions of water channel 11 and the exit region of water channel 10. Temperature stability based on standard deviations of thermocouple readings during the cook period indicated slightly better stability using a retort temperature of 115°C and using the tallest weir height. Mean standard deviations of temperature ranged from 0.10 to 0.20 C° indicating the magnitude of temperature increased during the cook period. Temperature stability was relatively uniform throughout the retort, except for water channel 11, which displayed less stable conditions. 93 94 Heat transfer studies indicated that the retort temperature influenced the rate of heat transfer during both the cook and cool periods. Smaller f^ and f values, indicating more efficient heat transfer, were found for a retort temperature of 125°C than of 115°C. The f, and f values were not affected by variations of weir h c 7 height. Heat transfer uniformity between trays varied slightly during the cook period depending on the weir height used. Weir height 2 produced uniform f^ values between all trays. Weir height 1 showed uniform fj values between trays, except for a significantly larger value which occurred in tray 10. Although weir height 3 created the most variability between trays, weir height 1 displayed the widest range of f^ values between trays. During the cool period, more variability in heat transfer between trays was indicated than during the cook period. Weir height 2 gave the most uniform f values between trays. Generally, the most efficient cooling occurred in trays 1, 5, 6 and 8. The range of f values between trays was similar for all three weir heights. Heat transfer uniformity within trays varied depending on the position in the tray during both cook and cool periods. Smaller f^ and f values were found at the entrance position of trays than at the exit and middle positions. Generally, the most efficient heat transfer conditions occurred in the entrance positions of all trays for both cook and cool periods. During the cook period, the least efficient locations were the exit and middle positions of trays 1 and 10. During the cool period, the least efficient locations were the middle positions of trays 1 and 3 and the middle and exit regions of tray 10. 95 Heating r.nd cooling lag factors indicated differences in the establishment of constant heat transfer conditions. An influence of retort temperature on and j values was indicated, however, the unknown degree of influence of variable f^ and f values on the lag factor calculations did not make a conclusion possible. Smaller values resulted using weir heights 1 and 2 than weir height 3, indicating quicker establishment of a constant heat transfer rate. Weir height did not influence j values. A gradient of j| values between trays was indicated, with smaller values associated with upper trays and larger values with lower trays. Patterns of j values between trays were not clear, however, there appeared to be better uniformity of ]'c values between trays. Within trays, and j values were smaller in the entrance region than the middle and exit regions. A comparison of the water immersion processes with a pure steam process at similar retort temperatures indicated larger f^ values were produced by the water process. Larger f values were produced by the cooling method used for the steam process. Distribution of F 0 values within the retort was variable. Non-uniform sterilizing conditions between trays occurred using all weir heights, however, the smallest range of F 0 values resulted using weir height 2. Larger F 0 values were associated with upper trays, while the smallest values occurred in the bottom trays. Within trays, larger F 0 values were found at the entrance position than at the exit and middle tray positions. Generally, the largest F 0 values were exhibited in the entrance positions of the middle to upper trays, while the smallest values were found in the middle and exit positions of the bottom trays. Areas in the retort 96 showing the lowest F 0 values will require attention when performing heat penetration studies on food products for future process determinations. Retort pressure studies demonstrated that stable pressures were maintained regardless of the retort temperature or weir height. Stable pressures were maintained during the cook period, however, a significant pressure drop occurred during the initial minutes of the cool period. Once the target pressure was re-established, stability was maintained for the remainder of the cool period. It should be noted that these studies were performed in a pilot-scale retort, thus the results may not apply directly to a commercial retort of the same type. They give an indication, however, of the potential temperature and heat transfer variability which might be encountered. The methodologies developed in this study to evaluate retort performance could be applied to assess processing condition efficacy of a wide variety of batch-type food sterilizing systems. VI. LITERATURE CITED 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. American Can Company. 1947. Canned Food Reference Manual. Third edition., Rogers-Kellogg Stillson, Inc., New York, NY. Andres, C. and Duxbury, D. 1972. Automatic line produces thermally processed foods in flexible packages. Food Processing 33(11 ):58. Anon. 1960. 150 years of canning. Food Mfg. 35(8):332. Anon. 1973. H o w close are we to canning food in flexible pouches. Canner Packer (December) p.19. Anon. 1975. Continental launches commercial scale pilot line for retortable pouch foods. Food in Canada 35(6):52. Anon. 1979a. A close look at sterilization. Food Processing Industry 48:577. Anon. 1979b. Reduces process cycle time for pouches and trays, keeps energy usage at minimum. Food Processing 40(7):76. Anon. 1981. Rectangular retort for rectangular pouch saves time, energy. Food Processing 42(8):62. Anon. 1982. What's happening with the retort pouch? Food Eng. 54(4):97 . Appert, N. 1810. Le Livre de Tous les Menages ou L'Art de Conserver, Pendant Plusieurs Annees, Toutes les Substances Animales et Vegetables . Chez Patris et Cie. , Paris. Ayoub, J.A., Berkowitz, D., Kenyon, E.M. and Wadsworth, C.K. 1974. Continuous microwave sterilization of meat in flexible pouches. J. Food Sci. 39:309. 97 98 Ball, C O . 1923. Thermal process time for canned food. Bulletin 37. National Research Counci l , Washington, DC. Ball, C O . 1938. Advancement in sterilization methods for canned foods. Food Research 3(1)and(2):13. Ball, C O . and Olson, F.C.W. 1957. Sterilization in Food Technology. McGraw-Hil l Book Co . , Inc., New York, NY. Beauvais, M. , Thomas, C , and Cheftel, H. 1961. A new method for heat-processing canned foods. Food Technol, 15(4):5. Berry, M.R., )r. 1979. The sterilization of food in pouches - critical parameters for still processing. Proceedings of the Conference Using the Retort Pouch World Wide - 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. 1980. Retort pouch in the '80's. Food Eng. 52(3):100. Beverly, R.G., Strasser, J. and Wright, B. 1980. Critical factors in filling and sterilizing of institutional pouches. Food Technol. 34(9):44. Bigelow, W.D . , Bohart, G.S., Richardson, A.C. and Ball, C O . 1920. Heat penetration in processing canned foods. N C A Bull. 16-L. Bitting, A .W. 1937. Appertizing or the Art of Canning: Its History and Development. The Trade Pressroom, San Francisco, CA. Blaisdell, J.L. 1963. Natural convection heating of liquids in unagitated food containers. Ph.D. thesis, Michigan State Univ., East Lansing, M l . Britt, I.J. 1987. Extending the processing capabilities of a pilot scale retort. M.A.Sc. thesis, University of British Columbia, Vancouver, BC. Brody, A. 1971. Food canning in rigid and flexible packages. Crit. Rev. Food Technol. 2:187. Casimir, D.J. 1975. Flame sterilization. CSIRO Food. Res. Q. 35:34. 99 Davis, E.G., Karel, M. and Procter, 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. Deissler, R.G. 1955. Turbulent heat transfer and friction in the entrance region of smooth passages. Trans. A.S.M.E. 77:1221. De Kruif, P. 1926. Microbe Hunters. Harcourt, Brace and Co. , New York, NY. Deming, O.L. 1902. Science and Experiment as Applied to Canning. Sprague Canning Machinery Co. , Chicago, IL. Fennema, O.R. (ed.). 1975. Principles of Food Science - Part II: Physical Principles of Food Preservation. Marcel Dekker, Inc., New York, NY. Gee, J.H. 1973. Retort pouch moves ahead. Food Eng. 45(1):68. Gerrish, S.L. 1975. Retortable pouches get their chance here. Modern Packaging 48(2):17. Coldbl i th, S.A. 1971. A condensed history of the science and technology of thermal processing - Part 1. Food Technol. 25(12):44. Coldbl ith, S.A. 1972a. A condensed history of the science and technology of thermal processing - Part 2. Food Technol. 26(1 ):64. Coldbl i th, S.A. 1972b. Controversy over the autoclave: Who first used the autoclave for the preservation of foods? Food Technol. 26(12):62. Goldfarb, P.L. 1971. Pouch for low-acid foods. Part II. Modern Packaging 44(1 ):70. Jackson, J .M. and Benjamin, H.A. 1948. Sterilization of foods. Ind. Eng. Chem. 40:2241. 100 Jackson, J.M. and Olson, F.C.W. 1940. Thermal processing of canned foods in containers. IV. Studies of the mechanisms of heat transfer within the container. Food Res. 5(4). Jeppson, M.R. and Harper, J.C. 1967. Microwave heating of substances under hydrostatic pressure. U.S. Pat. 3,335,253. Kenyon, E.M., Wescott, D.E., La Casse, P. and Gould, J.W. 1971. A system for continuous thermal processing of food pouches using microwave energy. J. Food Sci. 36:289. Lampi, R.A. 1977. Flexible packaging for thermoprocessed foods. Adv. Food Res. 23:305. Landy, J.J. 1965. Method of sterilizing food in sealed containers. U.S. Pat. 3,215,539. Lawler, F.K. 1967. New sterilizer made in France. Food Eng. 39(7):73. Leniger, H.A. and Beverloo, W.A. 1975. Food Process Engineering, pg. 62, D. Reidel Publishing Co. , Dordrecht, Holland. Leonard, S., Merson, R., Marsh, G., York, G., Heil, J. and Wolcott , T. 1975. Flame sterilization of canned foods: An overview. J. Food Sci. 40:246. Long, F.E., Shaw, F.B. and Lisle, H.C. 1966. Microwave sterilization and vacuumizing of products in flexible packages. U.S. Pat. 3,261,140. Lopez, A. 1981. A Complete Course in Canning. Eleventh edition. The Canning Trade, Baltimore, M D . Mantell , C.L. (ed.) 1958. Engineering Materials Handbook, 1st edition, McGraw-Hil l Book Co . , New York, NY. May/ E.C. 1937. The Canning Clan. The Macmillan Co . , New York, NY. 101 McCinnis , D.S. 1986. Surface heat transfer distribution in a weir type pressurized water retort for processing foods in flexible retort pouches. Can. Inst. Food Sci. Technol. J. 19:45. 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. Milleville, H.P. 1980. Steam/air retorting of pouches: simpler, less expensive method for processing with overriding pressure. Food Processing 41(3):98. Milleville, H.P. 1981. Retorts for processing pouches conserve heat. Food Processing 42(1):94. Milleville, H.P. and Badenhop, A.F. 1980. Guidelines on good manufacturing practice for sterilizable flexible packaging operations for low-acid foods. Technical Manual No. 4, Campden Food Preservation Research Association. Dept. of Food Sci. and Technol. Oregon State University, Corvallis, OR. Morris, C.E. 1981. Inside North America's new retort pouch plant. Food Eng. 53(6):56. National Food Processors Association. 1985. Guidelines for thermal process development for foods packaged in flexible containers. National Food Processors Association, Research Laboratories, Washington, DC. Olson, F.C.W. and Jackson, J .M. 1942. Heating curves: Theory and practical application. Ind. Eng. Chem. 34:337. Parcell, J.W. 1930. Investigations on the retorting of glass containers. Part 1. Canning Age 11(7):475. Peterson, W.R. and Adams, J.P. 1983. Water velocity effect on heat penetration parameters during institutional size retort pouch processing. J. Food Sci. 48:457. 102 Pflug, I.J. 1964. Evaluation of Heating Media for Producing Shelf Stable Food in Flexible Packages. Phase I. Final Report, Contract No. DA19-AMC-145(N) U.S. Army Natick Laboratories, Natick, MA. Pflug, I.J. and Blaisdell, J.L. 1961. The effect of velocity of steam/air mixtures on the heating of glass containers. Mich. State Univ. Agr. Expt. Sta. Quart. Bull. 44(2):235. Pflug, I.J., Bock, J.H. and Long, F.E. 1963. Sterilization of food in flexible packages. Food Technol. 7(9):87. Pflug, I.J. and Borrero, C. 1967. Heat Media for Processing Foods in Flexible Packages. Phase II. Technical Report 67-47-CP. U.S. Army Natick Laboratories, Natick,MA. Pinto, A. 1978. Retort pouch: Moving to close the materials and machinery gap. Modern Packaging 51(3):23. Ramaswamy, H.S. 1983. Heat transfer studies of steam/air mixtures for food processing in retort pouches. Ph.D. thesis, University of British Columbia, Vancouver, BC. Ramaswamy, H.S. and Tung, M.A. 1986. Model l ing heat transfer in steam/air processing of thin profile packages. Can. Inst. Food Sci. Technol. J. 19:215. Roop, R.A. and Nelson, P.E. 1981. Processing retort pouches in conventional sterilizers. J. Food Sci. 47:303. Rosenberg, U. and Bogl, W. 1987. Microwave pasteurization, sterilization, blanching, and pest control in the food industry. Food Technol. 41(6):92. Smith, T. and Tung, M.A. 1982. Comparison of formula methods for calculating thermal process lethality. J. Food Sci. 47:626. Stare, F.A. 1949. The Story of Wisconsin's Great Canning Industry. The Canning Trade, Baltimore, M D . 103 Stumbo, C.R. 1973. Thermobacteriology in Food Processing. Second edition. Academic Press, New York, NY. Tanner, F.W. 1932. Microbiology of Foods. Twin City Printing Co . , Champaign, IL. Toyo Seikan Kaisha, Ltd. 1973a. Fully Automatic Overpressure Retort for Retortable Pouches. Brochure. Revised edition I. Tokyo, Japan. Toyo Seikan Kaisha, Ltd. 1973b. Opening up tomorrow's markets. RP-F Guide. Tokyo, Japan. Tsutsumi, Y. 1979a. New technology applied to pouches and future trends. Proceedings 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. 1979b. Sterilization methodology applied to pouches and trays. Proceedings 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., Ramaswamy, H.S. and Papke, A . M . 1984a. Thermophysical Studies for Improved Food Processes. Final Report. DSS File No. 35SZ.01804-9-0001, prepared for the Agriculture Canada PDR Program, Ottawa, O N . Tung, M.A., Ramaswamy, H.S., Smith, T. and Stark, R. 1984b. Surface heat transfer coefficients for steam/air mixtures in two pilot scale retorts. J. Food Sci. 49:939. Weintraub, S.E. 1986. Residual gas effects on heat transfer in overpressure processing of flexible packages. M.Sc. thesis, University of British Columbia, Vancouver, BC. Whitaker, W.C . 1971. Processing flexible pouches. Modern Packaging 44(2):83. Wilson, D.C. 1980. Theoretical problems in pouch processing. Proceedings of the Winter Meeting of American Society of Agricultural Engineers. Chicago, IL. December 2-5. 104 Yamano, Y. 1976. Studies on thermal processing of flexible food packages by steam-and-air retort. Ph.D. thesis, Kyoto University, Kyoto, Japan. 

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.831.1-0097930/manifest

Comment

Related Items