Computer-Based On-Line Assessment of Sterilizing Value and Heat Distribution in Retort for Canning Process

The global food industry has the largest number of demanding and knowledgeable consumers: the world population of seven billion inhabitants, since every person eats! This population requires food products that fulfill the high quality standards established by the food industry organizations. Food shortages threaten human health and are aggravated by the disastrous, extreme climatic events such as floods, droughts, fires, storms connected to climate change, global warming and greenhouse gas emissions that modify the environment and, consequently, the production of foods in the agriculture and husbandry sectors. This collection of articles is a timely contribution to issues relating to the food industry. They were selected for use as a primer, an investigation guide and documentation based on modern, scientific and technical references. This volume is therefore appropriate for use by university researchers and practicing food developers and producers. The control of food processing and production is not only discussed in scientific terms; engineering, economic and financial aspects are also considered for the advantage of food industry managers.


Introduction
Heating process for food is of importance to the consumers since it is considered to be one of food preservation techniques. Under these techniques food can be stored or edible within a long period of time. One of them which require heat treatment is sterilization process. Thermal sterilization of prepackaged canned foods in retort has been the most widely used during the twentieth century. Typically this method consists of heating food containers in pressurized retorts at specified temperatures for prescribed lengths of times (Teixeira and Tucker, 1997). The process time for canned food is indicated based on the sufficient achievement of bacterial inactivation in each container in order to comply with public health standards or food safety. In addition it will minimize the probability of food spoilage. The traditional methods for thermal process calculations or validation such as Ball and Stumbo methods were developed and widely used ever since. However they required the off-line input of tables and consequently series of calculation steps which might be resulting in too-long or too short heating process. At present there are a lot of commercial software available which could be used either on-line or off-line analysis for sufficient heat treatment or process lethality (F o ) such as CAN-CALC  , and CALSoft™ etc. Balaban (1996cited by Teixeira et al., 1999 described that CAN-CALC software needed to get f h (heating rate factor) and j h (heating lag factor) from heat penetration test prior to be able to predict internal center product temperatures in response to any dynamic boundary temperature for products of any shape and size as shown in figure 1 and 2. Therefore if assumed that the selected can was at the slowest heating point of the retort, simulated system F o for food products that heated by any combination of conduction or convection heat transfer also could be obtained. However, the software performance was emphasizing with its capability to deal with process deviation such as steam shutting off and back on. The CALSoft™ software (Anonymous, 2011) was designed specifically for conducting heat penetration and temperature distribution testing, evaluating the collected data, calculating a thermal process or vent schedule/come-up time, and evaluating process deviations. It was supposed to use with CALPlex™ data logger and claimed for the most widely used commercial thermal processing software.
When comparing to all other methods to calculate F o , the general method has been accepted to be the most accurate. However, the disadvantage of this method is used to be the clumsiness since it has to obtain the lethal rate in every time step. The smaller time step, the more accurate it will be. But when using computer as a tool to perform all these calculation, F o determination become rapid and simple. Fig. 1. Parameters input in CAN-CALC software before simulating for system F o (Balaban, 2004) www.intechopen.com Fig. 2. Graphical display of calculated or predicted and experimented temperature at coldest point in CAN-CALC software (Balaban, 2004) Many researchers (Lappo and Povey ,1986;Ryniecki and Jayas, 1993) had employed the accumulated process lethality to design system process control for batch steam retort. A number of thermocouples were connected to the cans. The mean temperature at the center of those cans was used for calculating process lethality in real time. Datta et al. (1986) used the numerical solution of 2 dimensional heat transfers in a finite cylinder as a part of the decisionmaking software in a computer-based retort control system. Actual retort temperature was read directly from sensors located in the retort and it was continually updated with each iteration of the numerical solution. Heating was continued until the accumulated lethality was reached some designated target value and the process would always end with the desired level of sterilization. However their solution of the model has some limitations since purely conduction-heated canned food was simulated for. Later many research works (Bichier et al.,1995 ;Teixeira, 1992) had been done without these limitations. Visual Basics computer simulation package for thermal process calculation was developed by Chen and Ramaswamy (2007). This graphical user interface (GUI) program was designed for training and testing of artificial neural artwork models and for study of process design or other research purposes. It is applicable to different retort thermal processing with different types of food such as solids, liquids and liquids containing particles in containers of different shapes and size. Temperature in container was solved by using finite difference and a numerical integration method was used for calculating process lethality and quality retention. There have been several attempts to develop control approaches for thermal process operation in food canning. Traditionally it consists of maintaining specified operating conditions that have been predetermined from product or process heat penetration test. The first control strategy was to employ real-time heat penetration data acquisition for intelligent on-line control of thermally processed foods. It was the most effective way to handle process deviation. However prior to start thermal operation, a number of product containers are instrumented with temperature probes then filling and seaming. Connection is made between these containers to data logger through the lead wires. Computer thus have the real-time accessing the data from data logger and perform calculation for accomplished sterilizing value at the coldest spot of container. The calculated accomplished sterilizing value is continually compared with the target value required at the end of heating. This strategy provides very accurate calculation of process lethality and is able to handle the process deviation without operator intervention and without any unnecessary degree of over-processing. The most valuable feature of this control strategy is that it is nearly foolproof since any thing that might have gone wrong earlier in the product preparation is revealed and accounted for. However, the obvious disadvantage for this type of control strategy would be cost prohibitive (Simpson et al., 2007).
Another retort control strategy that many researchers had worked on is about on-line correction of process deviation which is integrating of the real-time data acquisition for retort temperature, on-line correction factor and mathematical heat-transfer model of can temperature (Teixeira and Manson, 1982;Datta et al., 1986;Teixeira and Tucker, 1997). However, the strategy that will be the future trend is microcontroller-based retort control system or simply on-line temperature measurement of retort to lap top computer. When the calculated accomplished lethality reaches the specified target value, computer will automatic shut off or turn on valves (Simpson el al., 2007). Awuah et al. (2007) discussed that Can-Calc process simulation software also was tested for its performance and further integrated into a computer-based on-line control system by Noronha et al. (1995) and Teixeira et al. (1999). As a whole, the purpose of software design and hardware control was based on the fact that foods should not be overheated since it leads to detrimental effect to food quality as well as the waste of energy and water. Thus heat should be minimally applied or applied as necessary as it needs. In order to get the above mentioned process, it is essential to have proper machine or devices associated with analysis method to assess the process efficacy involving heat treatment for any one of canned products and heat distribution in sterilizing device. However, in Thailand most of the hardware and software available now are imported. They are designed basically on either the post assessment (after completely heating foods) or undergo heating. Up to now more efforts have been carried out for developing the intelligent on-line retort control system which is capable of rapid evaluation, on-line correction and printed documentation. The development of local devices or software for such purposes is still rarely found in Thailand. Thus the objectives of the research are to develop visual basic computer software to integrate the on-line data acquisition. The assessment of sterilizing value or process lethality (F o ) as well as heat distribution in retort was performed while heating. The software also can be used as an education tool for thermal processing study.

Materials and method 2.1 On-line data acquisition and sterilizing value (F o ) assessment
Quick Basics program was designed and developed to obtain the interfacing data from PCL-812PG card (multifunction data acquisition card) together with PCLD-889 boards (amplifier/multiplexer board with signal conditioning and cold junction sensing circuit) as in figure 3. Up to 8 thermocouples could be instrumented to the loaded cans and hard-wired through a retort (figure 4). They were used for sensing the analog inputs of temperatures from different locations in the retort and then they were transformed to digital temperature data via the interface PCL-812PG A/D card. Thus time-temperature history data from tested cans and some for temperature in the retort were recorded and displayed graphically in every 4.5 second in the developed program as GUI software coded by Visual Basic 6.0. Prior to the test, all the temperature reading probes were calibrated from the temperature range of ambient to 140 o C by comparing to the reading of reliable portable digital thermometer measuring hot oil. The designed computer program is able to access the recorded Quick-Basic data file which provide real-time of time-temperature history in cans and retort, and calculate them for lethal rate in every 4.5 second by Simpson's rule of numerical integration to obtain accomplished F o dynamically during sterilizing. In order to evaluate the accuracy of this program, time-temperature history data was also tested with F-ADDING which is a computer program for calculating F o coded by Rouweler (2000). The minimum accomplished F o among all of them from each probe is obtained as system F o and it is compared simultaneously with the target F o needed to end the process in the minimum of process time. The flowchart of the algorithm for on-line F o assessment is shown in figure 5. This approach was accepted as unquestionably the most effective and most certainly the very safest on-line correction when process deviation occurs (Teixeira and Tucker, 1997;Simpson et al., 2007) since thermocouples were used to on-line measure temperature not only retort but also the cans.

Heat distribution performance in retort
A small vertical retort with diameter of 38.8 cm and electric boiler were constructed for the test as shown in figure 4. The interfacing devices was assembled -interface cards, thermocouples, connectors, computer and peripheral equipments-vertical retort and electric boiler. One probe of thermocouples (probe # 1) were connected to the end tip of mercury thermometer in the retort and one (probe # 8) connected to the center of can which was hot filled with distilled water and then seamed. The rest of them (6 probes) were distributed appropriately inside the retort as shown in figure 6. The on-line graphical display of temperature from 8 thermocouple probes was shown while heating. In addition, the probes # which provides minimum and maximum temperature, as well as maximum temperature difference were indicated through out the heating process. The sterilization temperature at 110 and 121 o C were chosen to investigate the heat distribution in the retort.

Process design and minimum heat accumulation in canned products
Heat accumulation in canned products during sterilizing could be investigated either from their heat penetration profiles or accomplished F o values. Thus three probes of thermocouples were connected to the cans and located those (3 cans) in the basket since these 3 locations tend to be the cold points of system -probes # 3, 4, and 5 attached to the cans located at the positions 1, 3, and 5 in the basket respectively (figure 7) and probe # 6 exposed directly to the temperature of the heating medium in the retort. The cans were hot filled with concentrated pineapple juice then seamed and put into the basket at specific locations in the retort as mentioned above. The retort was full loaded with the rest of the cans. Specify target sterilizing value was chosen according to product characteristics (table 1) in GUI window. The information about target sterilizing value could be added to the file by pressing updated button. Subsequently sterilization was commenced by removing air in retort by replacing it with steam. Start button in F o determination GUI window was pressed to begin recording time-temperature data via the interfacing devices until the minimum value of accomplished F o (system F o ) reached specified target F o . Thus process schedule was recorded automatically and displayed graphically. Even though concentrated pineapple juice is acid food (pH > 4.0), mild sterilizing is usually sufficient and therefore could be applied. Specified target F o would be chosen as 10 121.1 F = 0.6-0.8 minutes (Rouweler, 2000) in this case. However in order to demonstrate the process design with this educational tool, the experiments were carried out with the holding temperature during sterilization selected to be at 110 o C and 120 o C.  Table 1. The characteristics of tested product.

Coldest spot in container
To validate the capability of program when it was used to find the coldest spot in a container, the interfacing devices were assembled as before. One of the cans was hard-wired with 2 thermocouple probes: probe# 4 at 1/4 of central axis from bottom of can, and probe# 5 at half of central axis as shown in figure 8 while probe# 6 was for measuring temperature of medium in the retort. This can was filled up with baby corns in saline solution which was solids in liquid type of canned food, and then seamed. The position to put this can in the retort was the slowest heating point of this equipment. After the retort was full loaded with cans, sterilization process was carried out at 121 o C for some certain period of time as before, minimum accomplished F o of which could be obtained.

Portable educational tools for computer-based off-line assessment of sterilizing unit
The objective of this part was to design the computer program for assessing process lethality from interfacing data system via USB-A/D board. Thus the driver and interfacing program for data logger, National Instrument USB-9211A, 4 Channel 24 bit (figure 9) was installed to a note book computer. The commercial interfacing software was sensing voltage signal through 1 to 4 type-T thermocouples and transforming to digital data stored in a note book computer (figure 10).  located in the basket at the center of bottom layer as the most probably slowest heating point. However one thermocouple (probe # 1) was exposed in the autoclave indicating medium temperature measurement during sterilizing. Temperature of 121 o C for 15 min was chosen for demonstrating sterilizing condition. Temperatures from 4 channels was recorded and stored in text file (*.txt) for every 2 seconds after running autoclave until finishing cooling process. QuickCalFo (Chamchong et al., 2008) was software designed to perform the off-line process lethality assessment by using Visual Basic 6.0 program. Input data of temperature and time during sterilization was retrieved from stored text file (figure 12) while target F o for each product with specific can size was pre-entered and saved into the program or selected from the list of available data before starting analysis for system F o . The result could display the temperature and time record in a spread sheet as well as heat penetrating curves and lethal rate profiles. F o values from each temperature-time profile were calculated by Simpson's rule general method and the minimum value was shown in the combo box nearby as system F o or accomplished F o . The accuracy of F o calculation from this software was validated as before by comparing it with that obtained from F-ADDING program coded by Rouweler (2000).  www.intechopen.com

Computer-based on-line assessment of sterilizing value
The software package for process design was divided into 3 parts: (1) the main window of the GUI to receive the input parameter which is target sterilizing value. The user can choose this value from pull down combo box or add/delete and update to have more choices for later use ( figure 12).
(2) Graphical window of temperature and time profiles with 8 corresponded text boxes to display accomplished sterilizing values from maximum 8 probes ( figure 13). There is one text box at the bottom to display system sterilizing value which is the minimum value among all of accomplished sterilizing values from each probe. System sterilizing value increases while the process is underway heating and cooling and ultimately reaches the designated target sterilizing value. Program then displays text message at the bottom of GUI for the operator to stop steaming and total process time during heating is shown in upper right corner text box. The temperature and time record can be used for process design or as documentation in quality assurance system. (3) The spread sheet of temperature-time recorded from 8 thermocouple probes which could display minimum and maximum temperature, as well as maximum temperature difference (max-min) at each time interval through out the heating process ( figure 14).

Heat distribution in retort
Practically heat distribution in a retort should be carried out before performing the assessment of sterilizing value of the process in order to validate the slowest heating point. Therefore heat distribution in a retort (as in part 3 from mentioned above) was observed from on-line temperature record obtained from different locations of this equipment. For sterilizing at 110 o C in a small retort unit distributed heat could be indicated by temperature values at positions 1-8 in the retort corresponding to probe # 1-8 ( figure 6). In addition minimum heating reading from thermocouple probe which was connected to the can located at the slowest heating point was able to be quantified as the accomplished F o for the system. Then it could be used as an indicator for stopping steaming. Therefore the display was able to assure the minimum heat distribution occurring while heating. The coldest point of the system should be coming from the can which had the thermocouple connected with and was located at position 8 or at the upper layer of cans and at the center of the basket. To get enough heat treatment for the products, heat distribution test must be carried out once the machine was installed or process/product was modified. For the same retort, it was found that when holding temperature of sterilization was moderate at 110 o C heat distribution was more uniform than that at 121 o C. Temperature difference from max and min at any holding time or temperature deviation was between 1.8 to 3.1 o C (1.6-2.8%) when sterilizing temperature was at 110 o C but it was between 5-14 o C (4.1-11.6%) at 121 o C sterilizing temperature. This was possible since heating at 121 o C required higher heating rate. However more of stagnant point or dead legs would appear. According to the steam flow pattern in this retort, the probe located at positions 3 (on top of the can which was at the center of basket) and 6 (upper layer and in between cans) were found to be the stagnation points and shown minimum convective heat transfer in each run at higher sterilizing temperature. However the minimum heating point for lower sterilizing temperature was found to get changed to be either position 5 or 2 (top of the retort) at the early stage of holding temperature in sterilizing period and then changed to location 3 for the rest of holding time. This was possible because the more amount of steam used during heating at 121 o C, the narrower the stagnation area would be. In addition, when lower amount of steam used at 110 o C sterilizing temperature, the probes at position 5 and 2 in the top layer of cans initially would get contacted with steam slower than any other locations. After heating at this temperature for a while, heating was up to the top of retort. Temperature at position 5 and 2 would be no longer the minimum. However the exit point of coming steaming was from the bottom. Whenever steam valve was not fully opened the thermocouples in the lower layer would be affected or heated first. Therefore accomplished F o obtained from the can at the center upper layer of the basket was suitable to be system F o because the temperature from the probe nearby (probe 3) had demonstrated minimum heat received.

Process design or schedule and minimum heat accumulation in canned products
The process design or schedule of acid food like concentrated pineapple juice was shown in table 2. The obtained process time of sterilizing acid food at 110 and 120 o C was 10 and 4.5 minutes excluding cooling period respectively. At higher sterilizing temperature (120 o C), it was shorter than that at lower temperature (110 o C) since both were calculated based on the same specified target F o (0.7 minutes) or having the same area under the heat penetration curve before stopping steaming. Although the specified target sterilizing value for such a product was chosen to be 0.7 minutes but F o of the system obtained was 1.29 and 0.94 minutes for sterilizing at 110 and 120 o C, respectively. This was because the calculation was including come-up time, holding and cooling period. Thus a little over-process could occur in each run since there was slow removal of heat during cooling. Improved and proper design of cooling system in the retort would provide better product quality in terms of organoleptic properties. Heat accumulation in canned products can be observed from all accomplished F o values obtained from different locations in retort. For sterilizing at 110 o C in figure 13 (a), accumulated heat which could be indicated by accomplished F o at positions 1, 3, and 5 were 1.29, 1.91 and 2.29 minutes, respectively, while the one outside the can was 3.53 minutes. Thus this was assuring that the coldest point was from the position 1 or at the bottom layer of cans and at the center of the basket. To get enough heat treatment for the products, heat distribution test must be carried out once the machine was installed or process/product was modified.   According to the steam flow pattern of this retort shown in figure 15, the cans located at positions 1 and 5 were supposed to be at the stagnation points and had minimum convective heat accumulation in each run. However the minimum accumulated heating www.intechopen.com point was found to get changed from position 1 located at the bottom center of the basket to position 5 located at the upper center of the basket while sterilizing at higher temperature. This was possible because the more amount of steam used during heating at 120 o C, the narrower the stagnation area would be. In addition, while cooling, the can at position 5 (top layer of cans) would get contacted with blowing air into retort during balancing pressure right after stop steaming. Therefore system F o would be obtained from the can at the center upper layer of the basket due to smaller heat accumulation. However cold water could significantly enhance heat removal when it was leveled up to that position in retort.

Coldest spot in a container
The sterilizing values at 2 different points could be used as indicated tool to validate the coldest spot in a container while undergo sterilizing process. As shown in figure 8 the can was instrumented with 2 thermocouple probes and hard-wired through a retort. Probe# 4 was at 1/4 of central axis from bottom of can, and probe# 5 was at the half of central axis while probe# 6 was for measuring temperature in the retort. It was found that temperature rising and dropping from probes #4 and 5 were almost identical and hard to distinguish. Exposed to heating medium in retort 9.54 Table 3. Validation of coldest spot in the can by sterilizing value

Portable educational tools for computer-based off-line assessment of sterilizing unit
The temperature and time logging on-line was displayed as in figure 16 but they were retrieved off-line as a text file shown in figure 17. Then F o was calculated from these data which were coming from thermocouple probes connected to 3 cans and one exposed to heating medium in the autoclave. From QuickCalFo as shown in figure 18, it showed that the slowest heating point in the retort was from thermocouple probe # 3 which was attached to the can located at the center bottom layer of the basket in the autoclave. Thus the accomplished sterilizing value or system F o was chosen from the minimum of 12.81 minutes among all from 4 thermocouple probes. In analysis F o frame box at the bottom right corner, the message after analyzing indicated whether the product had enough heat treatment or not. To display temperature and time as a spread sheet on the left side of this figure, time interval of data was selected first at the bottom between in every 1, 2 or 5 minutes. Then the temperature and time only at this specific interval was shown in the spread sheet of GUI. Heat penetration and lethal rate profiles from all 4 probes were also displayed graphically. The x and y range of these 2 graphs were adjusted automatically according to process temperature and time span used. In addition, this visual basic form could be printed out for food safety documents.

Conclusion
A computer program for on-line data acquisition and accomplished F o assessment was developed by MS Visual Basic 6.0 language. This computer-based on-line device was able to evaluate the coldest point of the can in the retort and calculate process lethality or system F o dynamically while sterilizing. Too much over or under process was avoidable for process design or schedule with the integration of such a device for on-line accomplished F o determination during preprocessing. The setup of hardware and software for computerbased on-line assessment of sterilizing unit would be needed for the cases of new products, processes or equipments. Non-uniform heat distribution in retort always exists. The designed program was able to perform heat distribution evaluation by recording and displaying maximum/minimum temperature deviation at different locations in retort during holding temperature in sterilization process. Lower sterilization temperature at 110 o C had lower temperature deviation (1.6-2.8%) in the retort during the holding temperature comparing to 4.1-11.6% at 121 o C. Thus lower temperature tended to lower the deviation.