BACKGROUND OF THE INVENTION A rotary sterilization system is a continuous source processing system with intermittent product agitation. This system is widely used in the canning industry to sterilize a shelf stable food product packaged in containers. It is used most often for sterilizing a food product that benefits from mechanical agitation of the containers.

A rotary sterilization system includes a rotary sterilizer that has one or more cooking shells through which a line of containers {1,..., i,..., I} ne are conveyed. The containers are cooked in the cooking shell (s) at one or more scheduled cooking retort temperatures. The containers are then conveyed in line through one or more cooling shells of the rotary sterilizer. Similar to the cooking shell (s), the containers are cooled in the cooling shell (s) at one or more scheduled cooling retort temperatures.

The containers {1,..., i,..., I} jne are conveyed through each cooking and cooling shell by spiral tracks and a reel. The reel has a scheduled reel speed and imparts movement while the spiral tracks provide the direction for the containers to be conveyed through the shell.

This also provides mechanical agitation of the food product within the containers.

In order for the food product in each container i to be commercially sterilized, a total lethality Fi over a total time interval tf j, td j that satisfies a predefined target total lethality F ; rg must be delivered during the rotary sterilization process to the product cold spot of the container. Here, tfi and tdi are the feed and discharge times when the container is fed into and discharged from the rotary sterilizer. The target total lethality is set by the USDA (U. S.

Department of Agriculture), the FDA (Food and Drug Administration), and/or a suitable food processing authority for destroying certain microorganisms. The reel speed and the cooking and cooling retort temperatures are then scheduled so that each container i will receive a scheduled time-temperature treatment that delivers a total lethality to the container which satisfies the target total lethality.

As is well known, the lethality Fi delivered to the product cold spot of a container i over a particular time interval t, tj is given by the lethality equation: <BR> <BR> <BR> t@<BR> <BR> <BR> Fi = # 10(TC8(t)i - TREF)/zdt<BR> <BR> <BR> t@ where t. and tk are respectively the begin and end times of the time interval t, tj, Tcs (t), is the product cold spot time-temperature profile for the container, z is the thermal characteristic of a particular microorganism to be destroyed in the sterilization process, and TREF is a reference temperature for destroying the organism. Thus, the total lethality F, delivered to the product cold spot over the total time interval tf , td i due to the scheduled cooking and cooling retort temperatures is given by this lethality equation, where t, = tfi and tk = td, i- The total time interval tf j, td and the product cold spot time-time-temperature profile Tcs (t) i must be such that the total lethality F ; over tf ;, td, ; satisfies the target total lethality Fg. In order to ensure that this occurs, various mathematical simulation models have been developed for simulating the product cold spot time-temperature profile based on the scheduled retort temperatures. These models include those described in Ball, C. O. and Olson,

F. C. W., Sterilization in Food Technology: Theory. Practice and Calculations, McGraw-Hill Book Company, Inc., 1957; Hayakawa, K., Experimental Formulas for Accurate Estimation of Transient Temperature of Food and Their Application to thermal Process Evaluation, Food Technology, vol. 24, no. 12, pp. 89 to 99,1970; Thermobacteriology in Food Processing, Academic Press, New York, 1965; Teixeira, A. A., Innovative Heat Transfer Models: From Research Lab to On-Line Implementation in Food Processing Automation II, ASAE, p. 177- 184,1992; Lanoiselle, J. L., Candau, Y., and Debray E., Predicting Internal Temperatures of Canned Foods During Thermal Processing Using a Linear Recursive Model, J. Food Sci., Vol. 60, No. 4,1995; Teixeira, A. A., Dixon, J. R., Zahradnik, J. W., and Zinsmeister, G. E., Computer Optimization of Nutrient Retention in Thermal Processing of Conduction Heated Foods, Food Technology, 23: 137-142,1969; Kan-Ichi Hayakawa, Estimating Food Temperatures During Various Processing or Handling Treatments, J. of Food Science, 36: Manson, J. E., Zahradnik, J. W., and Stumbo, C. R., Evaluation of Lethality and Nutrient Retentions of Conduction-Heating Foods in Rectangular Containers, Food Technology, 24 (11): 109-113,1970; Noronha, J., Hendrickx, M., Van Loeg, A., and Tobback, P., New Semi-empirical Approach to Handle Time-Variable Boundarv Conditions During Sterilization of Non-Conductive Heating Foods, J. Food Eng., 24: 249-268,1995; and the NumeriCAL model developed by Dr. John Manson of CALWEST Technologies, licensed to FMC Corporation, and used in FMC Corporation's LOG-TEC controller.

However, if any of the actual retort temperatures in the cooking and cooling shells drops below a corresponding scheduled cooking or cooling retort temperature, a temperature deviation occurs. Traditionally, when such a deviation occurs, the controller stops the shells' reels and prevents any of the containers {1,..., i,..., I} line from being fed into or discharged from the rotary sterilizer until the deviation is cleared. But, this approach causes numerous problems. For example, significant production down time will result. And, many containers {-loverpr will be over processed since the total lethalities {..., Fi over tfi, tdj,... Ioverpr actually delivered to their product cold spots will significantly exceed the target total lethality F, a All of these problems may result in severe economic loss to the operator of the rotary sterilization system.

In order to prevent such loss, a number of approaches have been discussed and proposed for on-line control of sterilization processes. However, all of these approaches

concern control of batch sterilization processes performed on a batch of containers {1,..., i,..., I} batch. In a batch sterilization process, all of the containers generally receive the same time-temperature treatment whether or not a temperature deviation occurs. Thus, when a deviation does occur, a correction to the process can be made which simultaneously effects all of the containers so that a minimum total lethality Fi over tb, te will be delivered to the product cold spot of each container i, where tb and te are the begin and end times of the batch sterilization process. An example of such an approach is described in concurrently filed and co-pending U. S. patent application serial no. 09/, entitled Controller and Method for Administering and Providing On-Line Correction of a Batch Sterilization Process, filed on November 6,1998, with Weng, Z. as named inventor. This patent application is hereby explicitly incorporated by reference.

In contrast, each container i in a rotary sterilization process will receive a unique time- temperature treatment. Thus, the total lethality Fi over tf j, txj that is actually delivered to each container is different. This makes it difficult to identify, while on-line and in real time, each container that will have a predicted total lethality delivered to it that is below the target total lethality Fg. As a result, the development of a controller that provides on-line handling of a temperature deviation in a rotary sterilization process without stopping the reels of the cooking and cooling shells has been inhibited.

SUMMARY OF THE INVENTION In summary, the present invention comprises a rotary sterilization system, a controller for use in the rotary sterilization system, and a method performed by the controller. The system, controller, and method are used to administer a sterilization process performed on a line of containers and provide on-line handling of a deviation in a scheduled parameter during the process. The containers contain a shelf stable food product that is to be sterilized in the sterilization process. In addition to the controller, the rotary sterilization system includes a rotary sterilizer.

The controller controls the rotary sterilizer in performing the rotary sterilization process according to scheduled parameters. When a temperature deviation below a specific scheduled temperature occurs, the controller identifies those of the containers that will in

response have a total lethality predicted to be delivered to them during the rotary sterilization process that is less than a predefined target lethality. This specific scheduled parameter may be a scheduled retort temperature in a temperature zone of the rotary sterilizer through which the line of containers is conveyed. It also may be a scheduled initial product temperature for the containers or a scheduled reel speed for conveying the containers in line through the rotary sterilizer.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a block diagram of a rotary sterilization system in accordance with the present invention.

Fig. 2 is a block diagram of a controller of the rotary sterilization system of Fig. 1.

Fig. 3 is an overall process flow diagram for the controller of Fig. 2 in controlling a rotary sterilization process performed by the rotary sterilization system of Fig. 1.

Fig. 4 is a timing diagram for handling a temperature deviation according to the overall process flow diagram of Fig. 3.

Fig. 5 is a lethality distribution diagram showing the distribution of lethalities for containers affected by the temperature deviation shown in Fig. 4.

Figs. 6 to 9 are detailed process flow diagrams for various steps of the overall process flow diagram of Fig. 3.

DETAILED DESCRIPTION OF THE INVENTION Referring to Fig. 1, there is shown a rotary sterilization system 100 for performing a rotary sterilization process on a continuous line of containers {1,..., i,..., Illite. Each container i contains a food product that is to be sterilized during the process. The system 100 comprises a rotary sterilizer 102, a programmed controller 104, and a host computer 105.

1. Exemplary Embodiment In an exemplary embodiment, the rotary sterilizer 102 includes a cooking shell 106-1 and a cooling shell 106-2 through which the containers {1,..., i,..., I} jne are conveyed in line.

The containers are cooked in the cooking shell 106-1 and cooled in the cooling shell 106-2.

Each of these shells has spiral tracks 108 and a reel 109 to convey the containers through the shell. The reel 109 imparts movement while the spiral tracks 108 provide the direction for the containers to be conveyed through the shell 106-1 or 2.

Furthermore, a feed device 110 of the rotary sterilizer 102 feeds the containers {1,..., line in line to the cooking shell 106-1. The feed device is designed to prevent the escape of steam while loading the containers onto the reel of the cooking shell 106-1. The containers are transferred from the reel 109 of the cooking shell 106-1 to the reel 109 of the cooling shell 106-2 by a transfer device 112. Like the feed device, the transfer device is designed to prevent the escape of steam from the cooking shell while the containers are transferred between the reels of the cooking and cooling shells. The containers are finally off-loaded from the cooling shell's reel by a discharge device 114.

In this exemplary embodiment, the cooking shell 106-1 has multiple temperature zones 115-1,2, and 3. The containers {1,..., i,..., I} ine are pre-cooked in the temperature zones 115-1 and 2 at corresponding scheduled retort temperatures TandT. The zone 115-3 is used to cook the containers at a corresponding scheduled retort temperature TORT3 Similarly, the cooling shell 106-3 has temperature zones 115-4 and 5 in which the containers are cooled at corresponding scheduled retort temperatures TandT-However, as those skilled in the art will recognize and as will be explained later in section 2, other embodiments do exist where fewer or more cooking and/or cooling shells with fewer and/or more temperature zones are used.

At each sample real time tr (e. g., every. 1 to 1 seconds) of the rotary sterilization process, the sensors 116-1,..., 4 of the hydrostatic sterilizer 102 respectively sense the actual retort temperatures TaRTI (),..., TaRT5 () in the corresponding temperature zones 115-1,..., 5 of the cooking and cooling shells 106-1 and 2. Similarly, the rotary sterilizer's sensor 107 senses the actual reel speed va (tj of the reels of the cooking and cooling shells at each each sample real time t,. Finally, the feed device 110 periodically (e. g., every 20 to 30 minutes)

removes a container being fed into the rotary sterilizer and a sensor 117 of the rotary sterilizer senses its actual initial product temperature Ta, p (tr) at that time tir.

The controller 104 administers the rotary sterilization process by controlling the rotary sterilizer 102 and providing on-line handling of any temperature deviations during the process. This is done in response to the actual initial product and retort temperatures Talp (tr) and TaRT,),..., Ts (tJ sensed by the sensors 117 and 116-1,..., 5 at each sample real time t, the actual reel speed va (t,) sensed by the sensor 107, and the actual initial product temperature TaIp (t) sensed by the sensor 117.

The host computer 105 is used to provide input information, namely input parameters and software, used by the controller 104 in administering the rotary sterilization process. The host computer is also used to receive, process, and display output information about the process which is generated by the controller. l. a. Hardware and Software Configuration of Controller 104 Turning to Fig. 2, the controller 104 comprises a main control computer 118 that includes a microprocessor (i. e., CPU) 119, a primary memory 120, and a secondary memory 121. The microprocessor executes an operating system 122, a process control program 123, a process scheduling program 124, and a temperature deviation program 125 of the controller.

The operating system and programs are loaded from the secondary memory into the primary memory during execution.

The operating system 122 and the programs 123 to 125 are executed by the microprocessor 119 in response to commands issued by the operator. These commands may be issued with a user interface 126 of the main control computer 118 and/or the host computer 105 via a host computer interface 127 of the controller 104. The operating system controls and coordinates the execution of the other programs. Data 128 generated by the operating system and programs during execution and data 128 inputted by the operator is stored in the primary memory. This data includes input information provided by the operator with the user interface and/or the host computer via the host computer interface. It also includes output information provided to the user interface or the host computer via the host computer interface that is to be displayed to the operator.

The controller 104 also comprises control circuitry 129. The control circuitry includes circuits, microprocessors, memories, and software to administer the rotary sterilization process by generating control signals that control the sequential operation of the rotary sterilizer 102. As alluded to earlier, the software may be downloaded from the host computer 105 and provided to the control circuitry by the process control program 123. The control signals are generated in response to commands generated by this program and issued to the control circuitry from the microprocessor 119 via a control circuitry interface 130 of the main control computer 118.

Furthermore, at each sample real time tr of the rotary sterilization process, the control circuitry 129 receives sensor signals from the sensors 107,117, and 116-1,..., 5 that represent the actual reel speed va and the actual initial product and retort temperatures Ta, p (tr) and TaRTI (tir).-'. Ts (tr). The control circuitry generates the control signals for controlling the rotary sterilizer 102 in response to these sensed parameters. These sensed parameters are also provided to the microprocessor 119 via the control circuitry interface 130 and recorded by the process control program 123 as data 128 in the primary memory 120. In this way, the process control program compiles and records in the primary memory 120 an actual reel time-speed profile va (t), an actual initial product time-temperature profile Talp (t), and actual retort time- temperature profiles Tg (t),..., T (t)or the corresponding temperature zones 115-1,..., 5.

These profiles are used in the manner described later for providing on-line handling of temperature deviations during the rotary sterilization process.

The sensors 116-1,..., 5 are preferably located in the slowest heating regions of the temperature zones 115-1,..., 5 to provide conservative estimates of the actual retort temperatures TaRTI T (tr)- However, if this is not possible, the process control program 123 may adjust the temperatures provided by the sensors to estimate the actual retort temperatures at the slowest heating regions. This adjustment would be done according to temperature distribution data 128 in the primary memory 120 generated from heating and cooling temperature distribution tests conducted on the temperature zones.

As mentioned earlier, the operating system 122 and the other programs 123 to 125 are normally stored in the secondary memory 121 and then loaded into the primary memory 120 during execution. The secondary memory comprises one (or multiple) computer readable memory (ies) 132 that is (are) readable by the main control computer 118 of the controller 104.

The computer readable memory (ies) is (are) therefore used to direct the controller in controlling the rotary sterilization process. The computer readable memory (ies) may comprise a PROM (programmable read only memory) that stores the operating system and/or the other programs. Alternatively or additionally, the computer readable memory (ies) may comprise a magnetic or CD ROM storage disc that stores the operating system and/or the other programs. The computer readable memory (ies) in this case is (are) readable by the main control computer with a magnetic or CD ROM storage disk drive of the secondary memory.

Moreover, the operating system and/or the other programs could also be downloaded to the computer readable memory (ies) or the primary memory from the host computer 105 via the host computer interface 127.

The controller 104 controls the rotary sterilization process according to the flow and timing diagrams of Figs. 3 to 9. In doing so, a finite difference simulation model is used by the process scheduling program 124 to simulate a scheduled product cold spot time- temperature profile Tcs (t) ° that applies to all of the containers {1,..., i,..., I} ine- Similarly, the temperature deviation program 125 uses the model to simulate corresponding product cold spot time-temperature profiles {..., Tcs (t) lu,...) for corresponding selected containers {..., i,...} sel at each sample real time tr during a temperature deviation. This model may be the earlier mentioned NumeriCAL model and used for both conduction heated food products and convection heated food products. Or, it may be one of the models described in the Teixeira et al., 1969 and Manson et al., 1970 references and used for conduction heated food products.

As will be evident from the foregoing discussion, the novelty of the invention described herein is not in which model is used, but in the manner in which it is used according to the flow and timing diagrams in Figs. 3 to 9. l. b. Overall Process Flow In the first step 134 for controlling the rotary sterilization process according to the overall process flow of Fig. 3, the input parameters for the rotary sterilization process are defined and provided to the controller 104. The input parameters include a predefined sampling time period At, for each real time increment t,-At, t, from the previous sample real time to the current sample real time tr during the process. The input parameters also

include a initially scheduled product temperature T, lp for the food product in the containers being processed. The input parameters further include the traditional heating and cooling factors ih7 fh, Xbh7 2, ic, and fc to be used in the simulation model. The heating factors ih, fh, xbh, and f2 are respectively the heating time lag factor, the heating curve slope factor, the broken heating time factor, and the broken heating curve slope factor that are pre-defined for the food product. Similarly, the cooling factors je and fc are respectively the cooling time lag factor and the cooling curve slope factor that are also pre-defined for the food product. The input parameters additionally include the earlier discussed thermal characteristic z for destroying a particular microorganism in the food product and the associated reference temperature TREF.

Also included in the input parameters is the earlier discussed target total lethality Ftarg and earlier discussed scheduled retort temperatures TSRTI,..., TsRTS Finally, the input parameters include the minimum and maximum reel speeds vmin and vma, and reel step information S for the reels 109 and spiral tracks 108 of the cooking and cooling shells 106-1 and 2 and length and location information L1, ...., L5 for the corresponding temperature zones 115-1,..., 5 in the shells.

In order to perform step 134, the operator issues commands with the user interface 126 and/or the host computer 105 to invoke the process control program 123. Then, the operator enters the input parameters Tlp, jh7 fh, xbh, 2, jc, fcS Frg TSRTI07-7 TSRT507 Vmin7 VmaÇ7 S, and L, L5 with the user interface 126 and/or the host computer 105. The process control program 123 loads the entered input parameters into the primary memory 120 for use by the programs 123 to 125. The execution of these programs is controlled and coordinated by the process control program in the manner discussed next.

The process control program 123 first invokes the process scheduling program 124.

In step 135, the process scheduling program simulates the entire rotary sterilization process to be administered to a container i to define an initially scheduled reel speed vS° for the reels of the cooking and cooling shells 106-1 and 2. This also results in an initially scheduled time- temperature treatment TsRT (t) i° that is to be given to each container i. This treatment includes pre-cooking portions at the scheduled retort temperatures TSRTl0 and TSRT20 over corresponding initially scheduled time durations At, ° and At2°. The treatment also includes a cooking portion at the scheduled retort temperature TSRT30 over a corresponding initially scheduled time duration lvt3°. Finally, the treatment includes cooling portions at the scheduled retort

temperatures TSRT40 and TSRT50 over corresponding initially scheduled time durations At4 and Ats". The precise manner in which step 135 is performed is discussed in greater detail in section l. c., but will be briefly discussed next.

The initially scheduled reel speed vS° and the initially scheduled total time-temperature treatment TsRT (t) i° are defined by using the simulation model mentioned earlier. Specifically, the process scheduling program 124 uses the simulation model to iteratively and incrementally simulate an initially predicted product cold spot time-temperature profile Tcs (t) iO that is predicted to occur at the product cold spot of each container i during the rotary sterilization process. This simulation is based on the input parameters TSIP, ih, fh, Xbh, f2, jc, fcs <BR> <BR> 0 0<BR> and TSRTl,!"-)sRT5' The process scheduling program 124 also iteratively and incrementally computes an initially predicted lethality Fi° that is predicted to be delivered to the product cold spot of each container i during the rotary sterilization process. In doing so, the program iteratively and incrementally computes a predicted total lethality Fi° that satisfies the target total lethality F"rg and is predicted to be delivered to the product cold spot over a simulated total time interval [0, At, °+... +At5°. This computation is made based on the product cold spot time-temperature profile Tcs (t) i° over this total time duration and the input parameters z and TREF. Furthermore, the lethality equation described earlier is used to make this computation, where tm = 0, tk =At+... +At Tcs (t) = Tcs (t) °, and F. = Fj°.

The initially predicted total lethality Fi'ove [0, Ivt°+... +At5° is iteratively and incrementally computed until the initially scheduled reel speed vo is determined for which this lethality satisfies the target total lethality F"rg. Moreover, the initially scheduled time durations/vt, °,...,/t5° are determined from reel speed speed v,', the reel step information S, and the temperature zone length and location information L,,..., L5. Thus, definition of the reel speed vs° also includes definition of the pre-cooking, cooking, and cooling portions of the initially scheduled total time-temperature treatment TsRT (t) ° on which the portions of the profile Tcs (t) ° over the time durations At, °,..., Ats are based.

The process control program 123 controls the administration of the rotary sterilization process in steps 136 to 149. In doing so, it first sets a counter j to zero in step 136. This counter is used to count each time that the currently scheduled reel speed vs is adjusted during the rotary sterilization process.

Then, at the current sample real time tr, the process control program 123 causes the control circuitry 129 in step 137 to administer the rotary sterilization process at the currently scheduled reel speed vs and at the scheduled retort temperatures TsRT10, ..., TsRT50 in the corresponding temperature zones 115-1,..., 5. In doing so, the control circuitry appropriately controls the rotary sterilizer 102 and monitors the actual retort temperatures TaRTI (),... I TaRT5 (tr) in the corresponding temperature zones 115-1,..., 5 at the time tr to verify that they are at least equal to the corresponding scheduled retort temperatures T,..., Tr-1 this embodiment of the controller 104, the scheduled retort temperatures will remain the same throughout the rotary sterilization process regardless if temperature deviations occur in the temperature zones. Thus, if such a temperature deviation does occur in a particular temperature zone 115-n, then the control circuitry administers corrections at the time tr so that the actual retort temperature TARTON (tr) in the temperature zone 115-n will eventually be brought up to at least the corresponding temperature TsRTn°.

Then the process control program 123 waits for the next sample real time tr = tr+Atr in step 138. In step 139, this program records the actual retort temperatures TaRTI (), T TaRT5 (tr) in the temperature zones 115-1,..., 5 at each sample real time t,. By doing so, the program 123 compiles the corresponding actual retort time-temperature treatments TaRT1(t), ..., TaRT5(t).

Similarly, the program records the actual initial product temperature Ta, p (tr) periodically sensed by the sensor 117 to compile the actual initial product time-temperature profile Ta, (t).

Furthermore, the program also records the currently scheduled reel speed vs at each time t.

This is done to compile a time-reel speed profile v (t) for the rotary sterilization process to provide a record of the changes in the reel speed vsj.

Then, in step 140, the process control program 123 determines whether any temperature deviations are occurring at the time tr in the temperature zones 115-1,..., 5. In doing so, the program 123 monitors each temperature TaRTnI) to determine if it is less than the corresponding scheduled cooking or cooling retort temperature TsRTn°.

If no deviation is occurring, then the process control program 123 proceeds to step 141. Any of the under processed containers {..., i,...} underpr that were identified in step 148 for segregation and are being discharged by the discharge device 114 at the current sample real time tr are then segregated in step 141 by the discharge device. The process control program causes the control circuitry 129 to control the discharge device 114 in performing this

segregation in the manner discussed later. In step 149, the process control program sets the currently scheduled reel speed vs to the initially scheduled reel speed vS° if all of the containers {..., i,...} aff affected by a temperature deviation have been discharged. Both steps 141 and 149 are discussed later in more detail. The process control program then administers the rotary sterilization process in step 137 and waits for the next sample real time tr = tr+Atr in step 138 to repeat the steps 139 to 149.

However, if the process control program 123 does determine in step 140 that a temperature deviation is occurring in a temperature zone 115-n at the current sample real time tr then the process control program invokes the temperature deviation program 125. In the example shown in Fig. 4, the temperature deviation occurs in the temperature zone 115-3. In step 142, the program 125 identifies the container i that currently at the time tr has the minimum total lethality Fj predicted to be delivered to its product cold spot over its currently scheduled total time interval tf,, t. This minimum lethality container i is identified from among the containers {..., i,...} aff that are currently affected by the temperature deviation.

These affected containers are those of the containers {1,..., i,..., I} line that are at the time tr currently in the temperature zone 115-n in which the temperature deviation is occurring. This is determined using the reel step information S, the reel time-speed profile v (t) compiled in step 139, and the length and location information L,,..., L5 for the temperature zones 115-1, ..., 5.

In one approach for identifying the minimum lethality container i from among the affected containers {..., i,...} aff, the temperature deviation program 125 may use an optimization search technique, such as the Brendt method disclosed in Press, W. H., Teukolsky, S. A., Vettering, W. T., and Flannery, B. P., Numerical Recipes in Fortran : The Art of Scientific Computing, Cambridge University Press, 1992. In this case, the program iteratively computes predicted total lethalities {..., Fj over tfj, td,...} sel for containers {..., i, selected to be evaluated. Based on these lethalities, the program iteratively bisects the list of affected containers to select the selected containers from among the affected containers until the minimum lethality container i is identified.

In a variation of the approach just described, the temperature deviation program 125 may initially use predefined intervals to initially select containers {..., i,...}, nt at the intervals for evaluation. Then, around those of the initially selected containers that have the lowest

predicted total lethalities {..., F,'over tf,, t,...}, nt, the optimization search technique just described is used.

In still another approach for identifying the minimum lethality container i, the temperature deviation program 125 may select all of the affected containers {..., i,...} a. as the selected containers {..., i, ...}se1 for evaluation. In doing so, the program computes at each sample real time tr the predicted total lethality Fij over [tf,i, td,ij] for each container i. From the computed lethalities {..., fui over t,, t,...} se) for the selected containers, the minimum lethality container i is identified.

In each of the approaches just described, the predicted total lethality Fi over tf j, td, ii for each selected container i is computed in the same way. Specifically, the temperature deviation program 125 first computes an actual current lethality Fij delivered to the container's product cold spot over the actual time interval tf j, tr that the container has been in the rotary sterilizer 102. This is done by simulating the portion of the rotary sterilization process that was actually administered over this time interval. In doing so, the simulation model mentioned earlier is used to iteratively and incrementally simulate the actual portion of the product cold spot time-temperature profile Tcs (t) j over this time interval for the container i.

This is done based on the input parameters jh, fh, xbh, 2, jc, and fcX the actual initial product temperature TaIP(tf,i) for the container i, and the portions of the actual retort time-temperature profiles TaRTI(t),..., TaRTn (t) respectively over the actual time intervals tfi, tl, ijl, that the container was in the temperature zones 115-1,..., n. Here, n identifies the temperature zone 115-n in which the temperature deviation is occurring. As mentioned earlier, in the example of Fig. 4, this is the temperature zone 115-3.

The actual initial product temperature Tp (tf,) for the container i is obtained from the actual initial product time-temperature profile Talp (t) compiled in step 139. The actual time intervals [tf,i, t1,ij], ..., (tn-1,ij, tr] for the selected container i are determined by the temperature deviation program 125 from the reel time-speed profile v (t), the reel step information S, and the temperature zone length and location information L,,..., Ln.

In the example of Fig. 4, the temperature deviation occurs in the temperature zone 115-3. Thus, the portion of the product cold spot temperature profile Tcs (t) li that actually occurred over the actual time interval tf,, r is based in this case on the portions of the actual retort time-temperature profiles TaRTOt), TaRT2 (t), and TaRT3 (t) respectively over the actual time

intervals tf., t,.', (t ;/, t,, and (t2, jj, tr The time intervals [tf,i, ti,ij] and (t1,i, t2,ij] have the initially scheduled time durations Atl° and/t2° since the temperature deviation began at the deviation begin time te while the container i was in the temperature zone 115-3. If, however, this container was in another temperature zone 115-1 or 2 when the deviation began, then the time intervals tf, i, tl jj and/or (t1,ij, t2,ij,] would have different time durations #t1j and/or #t2j because the reel speed vs would have been changed while the container was in that temperature zone.

From the actual portion of the product cold spot time-temperature profile Tcs (t) i over tf, i, and the input parameters z and TREK, the temperature deviation program 125 iteratively and incrementally computes the actual current lethality Fij that has been delivered to the product cold spot of the selected container i over the actual time interval tf j, tr. This is done using the lethality equation described earlier, where tm = tuf,, tk = tr, Tcs (t) = Tcs (t) ijX and Fi Fij. The precise manner in which the actual current lethality is computed in step 142 is discussed in greater detail in section l. d.

Then, the temperature deviation program 125 simulates the remaining portion of the rotary sterilization process that is predicted to be administered to the selected container i over the scheduled remaining time interval (tr, td,ij assuming that the temperature deviation ends after the time tr. In performing this simulation, the simulation model mentioned earlier is used to iteratively simulate the predicted remaining portion of the product cold spot time- temperature profile TCS(t)ij based on the input parameters j, fh, Xb, f2, Jc, and f, the actual product cold spot temperature Tcs (tir) ; at the time t,, and the scheduled retort temperatures TSRTnOX TSRT50 over the currently scheduled remaining time intervals (tr, t,..., (t, t,'.

The actual product cold spot temperature TCS(tr)ij for the selected container i is obtained from the actual portion of the product cold spot time-temperature profile TCS(t)ij over tfj, tr that was just described. Moreover, the currently scheduled time intervals (tr, tn,ij], ..., (t4,ij,td,ij] for the container i are determined by the temperature deviation program 125 from the reel time-speed profile v (t), the reel step information S, and the temperature zone length and location information L,,..., L5.

In the example of Fig. 4, the temperature deviation occurs in the temperature zone 115-3. Thus, the predicted remaining portion of the product cold spot temperature profile TCS(t)ij is based on the scheduled retort temperatures TsRT30, TsRT40, and TsRT50 respectively over

the currently scheduled remaining time intervals (t, t3, ;, (t3, ;, t4, ;, and (t4,ij, td,ij. In this example, the time intervals (t2,ij, t3,ij], (t3,ij, t4,ij] and (t4,ij, td,ij] respectively have re-scheduled time durations #t3j, #t4j, and At that are different than the initially scheduled time durations/vt3°, tt4°, and lit5° since the currently scheduled reel speed v5j at the current sample real time tr has been re-scheduled from the initially scheduled reel speed vs0.

The temperature deviation program 125 iteratively and incrementally computes the total lethality Fij predicted to be delivered to the product cold spot of the selected container i over the scheduled total time interval tf j, td ji. This is done based on the predicted remaining portion of the product cold spot time-temperature profile TCS(t)ij over (t,, t, the actual current lethality Fjj over tfj, tr that was just described, and the input parameters z and Trek.

This is also done using the lethality equation described earlier, where tm = tr, tk = td,ij, TCS(t) = TCS(t)ij,(t)ij, and Fi = Fij. The predicted total lethality is the sum of the actual current lethality and a predicted remaining lethality Fij that is predicted to be delivered to the container's product cold spot over the time interval tr td, j. The precise manner in which the predicted total lethality is computed in step 142 is discussed in greater detail in section l. e.

Then, in step 143, the temperature deviation program 125 determines at the current sample real time tr if the container i with the minimum predicted total lethality F ; over tf j, td ij is less than the target total lethality Ftarg. If it is not, then this means that all of the affected containers {..., i,...} a also have predicted total lethalities {..., F,'over tf,, t,...} that are at least equal to the target total lethality. In this case, the process control program 123 proceeds to step 141 and causes any of the previously identified under processed containers {..., i,...} underpr that are being discharged at the time tr to be segregated. Then, in the manner discussed earlier, the process control program 123 administers the rotary sterilization process in step 137 and waits for the next sample real time tr = tr + #tr in step 138 to repeat the steps 139 to 148.

In this embodiment, if it is determined in step 143 that the minimum total lethality Fi over tf j, td ij is less than the target total lethality Ftarg, then the temperature deviation program 125 determines in step 144 if the currently scheduled reel speed vs is set to the minimum reel speed vu,,. If it is not, then the program increments the counter j in step 145 and defines a re- scheduled (or adjusted) reel speed vs in step 146.

In step 146, the re-scheduled reel speed vs is defined in a similar manner to the way in which the initially scheduled reel speed vS° is defined in step 135. But, in this case the actual product cold spot temperature Tcs (tr) i at the time tr and the actual current lethality Fi over tf, i, for the minimum lethality container i are used in simulating the remaining portion of the rotary sterilization process in order to compute a predicted total lethality Fi over t This is done in a similar manner to that described earlier for computing the predicted total lethality for a container in step 142. But, similar to step 135, this is done iteratively and incrementally until the reel speed is determined for which the predicted total lethality satisfies the total target lethality Fag or the reel speed equals the minimum reel speed vu,,. The precise manner in which step 146 is performed is discussed in greater detail in section l. f., but will be briefly discussed next.

The definition of the re-scheduled reel speed therefore also results in the definition of a re-scheduled remaining time-temperature treatment TSRTIt) i. The treatment includes a remaining cooking portion at the scheduled retort temperature TSRT30 over a corresponding re- scheduled time duration At. Similarly, the treatment also includes cooling portions at the scheduled retort temperatures TSRT40 and TSRT50 over corresponding re-scheduled time durations At4i and At.

Ideally, it is desired that the minimum predicted total lethality Fji over tf j, td ji for the minimum lethality container i will satisfy the target total lethality Ftarg But, as just mentioned, the re-scheduled reel speed vs may be limited to the minimum reel speed vmjn. In this case, the minimum predicted total lethality will not satisfy the target total lethality Fg.

If the temperature deviation program 125 determines this to be the case in step 147, then this means that under processed containers {..., i,...} undetpr from among the affected containers {..., i,...} affwill have predicted total lethalities {..., Fj over tfj, td ji,...} underpr that are less than the target total lethality. The minimum lethality container i is of course one of the under processed containers. The under processed containers are to be segregated and are identified at the current real sample time tr in step 148 by the program.

Fig. 5 shows the distribution of the affected containers {..., i,...} aff and the under processed containers {..., i,...} underpr to be segregated at the time t, In identifying the under processed containers in step 148, the program 125 uses a similar approach as that used in step

142 to identify the minimum lethality container i. But, in this case, the additional criteria of the target total lethality Fats ils used to expand the search.

Once the under processed containers {..., i,... underpr have been identified at the current real sample time , the process control program 123 then proceeds to step 141. As discussed earlier, this program causes the control circuitry 129 to control the discharge device 114 in segregating any of the under processed containers that are being discharged at the current sample real time t. In order to segregate the under processed containers, the process control program tracks these containers to determine when they will be discharged. This is done using the reel time-speed profile v (t), the reel step information S, and the temperature zone length and location information L,,..., L5.

The steps 137 to 149 are repeated until the temperature deviation is cleared. In this way, at each sample real time tr during the deviation, the list of under processed containers {..., i,...} underpr at the time tr is combined with the list from the previous sample real time t,..

As a result, the list of under processed containers is dynamically updated and maintained.

Since these under processed containers are segregated when discharged in step 141, this will ensure that only those of the containers {1,..., i,... I} ne that are adequately processed are released for distribution.

The list of affected containers {..., i,...} aff is also dynamically updated and maintained in the same manner as the list of under processed containers {., i,} underpr-When the temperature deviation is cleared, this list will remain the same and the process control program 123 tracks the containers in this list until they have all been discharged. This tracking is done in the same manner in which the under processed containers are tracked. The process control program 123 will then set the currently scheduled reel speed vs back to the initially scheduled reel speed vs° in step 149.

Furthermore, the controller 104 has the unique feature of being able to handle multiple temperature deviations. For example, if another temperature deviation does occur, then the steps 137 to 149 are repeated during this deviation. Therefore, even if a selected container i is exposed to multiple temperature deviations, the predicted total lethality F over tf js td jj that will be delivered to it can be accurately determined based on those of the actual retort temperature profiles T) (t),..., T (t) that it has been treated with over the rotary

sterilization process. Moreover, this results in the list of under processed containers {..., i,} t derpr being further updated and expanded.

I. c. Detailed Process Flow for Step 135 of Fig. 3 Fig. 6 shows the detailed process flow that the process scheduling program 124 uses in step 135 of Fig. 3 to define the initially scheduled reel speed v5°. In doing so, this program iteratively performs a simulation of the rotary sterilization process that is predicted to be administered to each container i in sub-steps 150 to 160 of step 135.

In step 150, the process scheduling program 124 first defines the initially scheduled reel speed vs° as the maximum reel speed vma. Then, in step 151, the program defines the time durations At°,..., lit5° for how long each container i is scheduled to be in the respective temperature zones 115-1,..., 5. This is done based on the initially scheduled reel speed, the reel step information S for the reels 109 and spiral tracks 108 of the cooking and cooling shells 106-1 and 2, and the length and location information L,,..., L5 for the temperature zones.

In step 152, the current sample simulation time ts is initially set to zero by the process scheduling program 124. This is the begin time of the simulated rotary sterilization process for the container i. The program also initially sets the predicted product cold spot temperature Tcs (ts) io of the container's product cold spot at this time to the scheduled initial product temperature Tslp. Similarly, the lethality F*° predicted to be delivered to the product cold spot over the current simulation time interval [0, tj is initially set by the program to zero.

Steps 153 to 157 are then performed by the process scheduling program 124 in each iteration of the simulation. In step 153 of each iteration, the program increments the current sample simulation time ts by the amount of the sampling period jazz This results in a new current sample simulation time tS.

Then, in step 154 of each iteration, the process scheduling program 124 simulates the portion of the product cold spot time-temperature profile Tcs (t) io predicted to occur at the product cold spot of the container i over the current simulation time increment ts-At, ts.

This is done using the simulation model discussed earlier and is based on the predicted

product cold spot temperature Tcs (ts-Atr) iO for the product cold spot at the previous sample simulation time tus-au and the heating and cooling factors j,"f,"Xb,"fi, J, and f. In the first iteration, this product cold spot temperature will be the scheduled initial product temperature Ts, p from step 152. However, in each subsequent iteration, the product cold spot temperature is obtained from the portion of the product cold spot temperature profile predicted over the previous simulation time increment tS-2Atr, tS-/tr that was simulated in step 154 of the previous iteration. Moreover, the simulation is also based on the respective scheduled retort temperatures T, Trs"when the current sample simulation time ts is within the corresponding simulation time intervals [0, #t10], ..., [#t10+...+#t40, #t50]. These time intervals indicate how long the container i is scheduled to be in the respective temperature zones 115-1,..., 5.

The lethality FiO that is predicted to be delivered to the product cold spot of the container i over the current simulation time increment tS-Atr, tS is then computed by the process scheduling program 124 in step 155 of each iteration. This is done based on the portion of the product cold spot time-temperature profile Tcs (t) i' predicted over this time increment and the input parameters z and TREF. This is also done in accordance with the lethality equation described earlier, where t, = tS-At, tk = ts, Tcs (t) =Tcs (t) i0 and Fj = Fi In step 156 of each iteration, the process scheduling program 124 computes the lethality Fj° predicted to be delivered to the product cold spot of the container i over the current simulation time interval [0, tS. This is done by adding the predicted lethality F,"over the current simulation time increment t,-At,, t in step 154 to the lethality Fio predicted to be delivered to the product cold spot over the previous simulation time interval [0, tS-Atr. In the first iteration, the predicted lethality over the previous simulation time interval is zero from step 152. In each subsequent iteration, this lethality is computed in step 156 of the previous iteration.

Then, in step 157 of each iteration, the process scheduling program 124 determines whether the current simulation time ts has reached the end time At, °+... +/\t5° of the simulated rotary sterilization process for the container i. If it is not, then the program returns to step 153 for the next iteration. In this way, steps 153 to 157 are repeated in each subsequent iteration until it is determined that the end time for the simulated rotary sterilization process has been reached. When this finally occurs, the program sets in step 158

the lethality Fj over the current simulation time interval [0, tj to the total lethality F ; predicted to be delivered to the container's product cold spot over the total simulation time interval [0, At, 0+... +At5o.

When this finally occurs, the process scheduling program 124 determines in step 158 whether the predicted total lethality F° over [0, Ot, °+... +Ot5 is at least equal to the target total lethality Ftarg. If it is not, then the program decrements in step 160 the initially scheduled reel speed v5° by a predefined reel speed offset Av. This results in the re-definition of this reel speed. Steps 151 to 160 are then repeated until step 159 is satisfied. The reel speed for which step 159 is satisfied is then used in steps 136 to 148 of Fig. 3 in the manner discussed earlier. l. d. Detailed Process Flow for Computing Lethality Fjj over tfj, tr in Steps 142 and 148 of Fig. 3 Fig. 7 shows the detailed process flow that the temperature deviation program 125 uses in steps 142 and 148 of Fig. 3 to compute the actual current lethality Fj delivered to the product cold spot of the container i over the actual time interval tf j, tr that the container has been in the rotary sterilizer 102. This is done by iteratively performing sub-steps 161 to 168 of steps 142 and 148 to simulate the actual portion of the rotary sterilization process that has been administered to the container's product cold spot over this time interval. Here, steps 161 to 168 are respectively similar to steps 151 to 158 of Fig. 6 and discussed in section l. c., except for the differences discussed next.

In step 161, the temperature deviation program 125 defines the actual time intervals tf ;, tl, ',..., (tn_,, ;, t that the container i has actually been in the respective temperature zones 115-1,..., n up to the current sample real time t. In this step, the definition of these time intervals is based on the accumulated reel time-speed profile v (t).

In step 162, the temperature deviation program 125 initially sets the product cold spot temperature Tcs (ts) for the product cold spot of the container i at the initial sample simulation time ts to the actual initial product temperature TaIP (tfi). This temperature is obtained from the actual initial product time-temperature profile T,,, (t). Moreover, the program initially sets the

actual lethality Fj delivered to the product cold spot over the current simulation time interval tf ;, ts to zero.

In step 164 of each iteration, the process scheduling program 124 simulates the portion of the product cold spot time-temperature profile Tcs (t) ; that actually occurred at the product cold spot of the container i over the current simulation time incrernent tS-Atr, tS.

This simulation is based on the respective actual retort temperatures TaRT1(ts), ..., TaRTn(ts) when the current simulation time tu is within the corresponding simulation time intervals [tf,i, t1,ij], ..., (tn-1,ij, tr]. These actual retort temperatures are obtained from the corresponding actual retort time-temperature profiles TaRTI(t),..., TaRTOt).

The actual lethality FJ that was delivered to the product cold spot of the container i over the current simulation time increment t-At,, tj is then computed by the temperature deviation program 125 in step 165 of each iteration. This is done based on the actual portion of the product cold spot time-temperature profile Tcs (t) j that was simulated over this time increment. In this case, Tcs (t) =Tcs (t) ; and Fi = Fj in the lethality equation described earlier.

In step 166 of each iteration, the temperature deviation program 125 computes the actual lethality Fj delivered to the product cold spot of the container i over the current simulation time interval tf j, tS. This is done by adding the actual lethality Fj over the current simulation time increment tS-lvtr, tS in step 164 to the actual lethality Fj over the previous simulation time interval tf , tS-Atr.

Then, in step 167 of each iteration, the temperature deviation program 125 determines whether the current simulation time ts has reached the current sample real time t,. If it is not, then the program returns to step 163 for the next iteration. In this way, steps 163 to 167 are repeated in each subsequent iteration until it is. determined that the current sample real time has been reached. When this finally occurs, the temperature deviation program 125 sets in step 168 the lethality Fij over the current simulation time interval tfj, tS to the actual current lethality Fij over the actual time interval tf j, tr and the product cold spot temperature Tcs (ts) for the container at the current sample simulation time to the actual product cold spot temperature Tcs (tir) ; at the current sample real time.

I. e. Detailed Process Flow for Computing Lethality Fj over Itfi, td, in Steps 142 and 148 of Fig. 3

Fig. 8 shows the detailed process flow that the temperature deviation program 125 uses in steps 142 and 148 of Fig. 3 to compute the lethality Fij predicted to be delivered to the product cold spot of a selected container over the total time interval tf ;, td that the container is in the rotary sterilizer 102. In this case, the program iteratively performs a simulation of the predicted remaining portion of the rotary sterilization process to be administered to this container using sub-steps 169 to 176 of steps 142 and 148. Like steps 161 to 168, steps 169 to 176 are respectively similar to steps 151 to 158 of Fig. 6 and discussed in section l. c., except for the differences discussed next.

In step 169, the temperature deviation program 125 defines the remaining time intervals (tj, t,..., (t, t that the container i is predicted to be in the respective temperature zones 115-n,..., 5 after the current sample real time t,. The definition of these time intervals in step 169 is based on the currently scheduled reel speed vs.

In step 170, the temperature deviation program 125 initially sets the initial sample simulation time tS to the current sample real time t,. The program also initially sets the predicted product cold spot temperature Tcs (ts) i3 for the product cold spot of the container i at this sample simulation time to the actual product cold spot temperature Tcs (tr) i3 obtained from step 168 of Fig. 7. Moreover, the program initially sets the predicted lethality Fj3 to be delivered to the product cold spot over the current simulation time interval tf j, tS to the actual lethality F ; over the actual time interval tf,, tj also obtained from step 168.

In step 172 of each iteration, the temperature deviation program 125 simulates the portion of the product cold spot time-temperature profile Tcs (t) ; that is predicted to occur at the product cold spot of the container i over the current simulation time increment [ts-#tr, ts].

The simulation is based on the respective scheduled retort temperatures T--"'TsRT5"when the current simulation time tS is within the corresponding simulation time intervals (tr, tn i3, (t4 j, td j3.

The lethality Fj that is predicted to be delivered over the current simulation time increment t,-A, t. is then computed by the temperature deviation program 125 in step 173 of each iteration. This is done based on the predicted portion of the product cold spot time- temperature profile Tcs (t) j that was simulated over this time increment in step 172.

In step 174 of each iteration, the temperature deviation program 125 computes the lethality F ; predicted to be delivered to the product cold spot of the container i over the current simulation time interval tf,, tj. This is done by adding the predicted lethality F ; over the current simulation time increment tS-Atr, tj from step 173 to the predicted lethality Fi over the previous simulation time interval tfi, ts-A.

Then, in step 175 of each iteration, the temperature deviation program 125 determines whether the current sample simulation time tu has reached the predicted discharge time tu 3 for the container i. If it has not, then the program returns to step 171 for the next iteration. In this way, steps 171 to 175 are repeated in each subsequent iteration until it is determined that the predicted discharge time has been reached. When this finally occurs, the program sets in step 176 the lethality F ; over the current simulation time interval tf j, tS to the predicted lethality F ; over the currently scheduled total time interval tf ;, t ;. l. f. Detailed Process Flow for Step 146 of Fig. 3 Fig. 9 shows the detailed process flow that the temperature deviation program 125 uses in step 146 of Fig. 3 to define the re-scheduled reel speed vs. This program uses sub- steps 178 to 187 to iteratively perform a simulation of the remaining portion of the rotary sterilization process predicted to be administered to the minimum lethality container i identified in step 142 of Fig. 3 and discussed in section l. b. Steps 178 to 187 are respectively similar to steps 159 and 151 to 159 of Fig. 6 and discussed in section l. c., except for the differences discussed next.

In step 178, the temperature deviation program 125 first decrements the currently scheduled reel speed vs by the predefined reel speed offset Av. If the decremented reel speed is greater than the minimum reel speed vin, té re-scheduled reel speed is defined as the decremented reel speed. However, if the decremented reel speed is less than or equal to the minimum reel speed, then the re-scheduled reel speed is defined as the minimum reel speed.

Since a re-scheduled reel speed vs is defined in step 178, the re-scheduled remaining time intervals (t, tD jj,..., (t43, td 3 that the minimum lethality container i is predicted to be in the respective temperature zones 115-n,..., 5 after the current sample real time tr need to be defined. This is done in step 179.

Step 180 to 186 are the same as steps 170 to 176 of Fig. 8 and discussed in section l. e. Thus, these steps are used to compute a total lethality F ; predicted to be delivered to the product cold spot of the minimum lethality container i over the re-scheduled total time interval tfi, tdii. It should be noted here that this is done using the actual current lethality F ; over tf j, tr and the actual product cold spot temperature Tcs (tory for the minimum lethality container i computed in steps 161 to 168 of Fig. 7.

Then, in step 187, the temperature deviation program 125 determines if the predicted total lethality Fj over tfi,. satisfies the target total lethality F"rg If it does not, then the program determines in step 188 whether the re-scheduled reel speed vs equals the minimum reel speed Vmjn. If it does not, then steps 181 to 188 are repeated until it is determined in step 187 that the target lethality has been satisfied or it is determined in step 188 that the minimum reel speed has been reached. In this way, the reel speed is re-scheduled.

2. Alternative Embodiments As indicated earlier, the embodiment of controller 104 associated with Figs. 3 to 9 and described in section 1. is an exemplary embodiment. Alternative embodiments that utilize the principles and concepts developed in Figs. 3 to 9 and section 1. do exist. Some of these embodiments are discussed next.

2. a. Scheduling and Re-Scheduling Variations The operator of the rotary sterilization process 100 may want to keep the initially scheduled reel speed v5° and retort temperatures T,,..., T constant throughout the entire rotary sterilization process. Thus, in this embodiment, the temperature deviation program 125 is simply used to identify the under processed containers {..., i,...} underpr in the manner discussed earlier in section l. b. when a temperature deviation occurs. More specifically, the steps 145 to 147 would be eliminated from the flow diagram of Fig. 3.

In another embodiment, the initially scheduled retort temperatures T, RTIO,..., TsRT5 0 may be re-scheduled when a temperature deviation occurs. In this case, the temperature deviation program 125 would define a re-scheduled retort temperature T,,..., or TsR in a

similar manner to which it defined a re-scheduled reel speed vs in step 146 of Fig. 3 and steps 178 to 188 of Fig. 9. In this embodiment, the initially scheduled reel speed v, l may be kept constant or a re-scheduled reel speed vs may be defined in conjunction with the re-scheduled retort temperature.

2. b. Identifying and Segregating Over Processed Containers Since re-scheduled reel speed vs may be defined when a temperature deviation occurs, it is possible that some of the containers {1,..., i,..., I} may be over processed due to the slower re-scheduled reel speed. In this case, a maximum total lethality FmaX may be pre- defined and included as one of the input parameters. Then, the over processed containers {..., i,...} o,, e, 1 with predicted total lethalities {..., F3 over tfj, td3,} overpr over this maximum total lethality would be identified in a similar manner to that way in which the under processed containers {..., i,...} underpr are identified in step 148 of Fig. 3 and discussed in section l. b. These containers would be segregated in the same way that the under processed containers are segregated in step 141 of Fig. 3. As a result, the remaining containers that are not under or over processed would have a uniform quality food product using this technique.

2. c. More Conservative Approaches In steps 142 and 148 of Fig. 3 discussed in section l. b. and in steps 161 to 168 of Fig.

7 discussed in section l. d., an aggressive approach was discussed for simulating the actual portion of the product cold spot time-temperature profile TCS (t) that occurs over the actual time interval tf j, tr that a container i has been in the rotary sterilizer 102. Specifically, this portion of the product cold spot time-temperature profile is based on the actual retort time- temperature profiles TaRTI t)t),..., TaRTn (t) over the corresponding time intervals t,, t,/,....

However, a more conservative embodiment could be employed which uses only the portion of the actual retort time-temperature profile TaRTn (t) over the time interval from the time when the container is first affected by the temperature deviation to the current sample real time t,. Specifically, the portion of the product cold spot time-temperature profile Tcs (t) lu

over the time intervals [tf,i, t1,ij], ..., (tn-2,ij, tn-1,ij] would be based on the corresponding scheduled retort temperatures 1) °,..., Tn- tor the temperature zones 115-1,..., n-1 in which the temperature deviation is not occurring.

Thus, if the container enters the temperature zone 115-n while the temperature deviation is occurring, the portion of the product cold spot time-temperature profile Tcs (t) i over the time interval (tn l ji, tr would still be based on the portion of the actual retort time- temperature profile TaRTn (t) over this time interval. But, if the temperature deviation begins at the deviation begin time td while the container is in this temperature zone, then the portion of the product cold spot time-temperature profile over the time interval (t_,, ;, te would be based on the scheduled temperature TsRTn°. In this case, only the portion of the product cold spot time-temperature profile over the time interval (td, tr would be based on the portion of the actual retort time-temperature profile TaRTn (t) over this time interval. In either case, this results in the actual lethality Fj delivered over the time interval tf j, tr being computed more conservatively in steps 142 and 148 of Fig. 3 and in sub-steps 161 to 168 of Fig. 7.

Similarly, the actual initial product temperature Ta, p (tf j) for a container i was used in steps 142 and 148 of Fig. 3 and in sub-steps 161 to 168 of Fig. 7 of Fig. 7 for computing the actual lethality F3 over tf j, tr. However, rather than using this actual initial product temperature, the scheduled initial product temperature T, lp may be used. This also results in the actual lethality being more conservative.

2. d. More Aggressive Approaches A more aggressive approach than that described earlier in section l. c. can be taken for defining the initially scheduled reel speed Vs". In this approach, a first additional step could be added after step 159 of Fig. 6 to determine whether the predicted total lethality Fi'ove 0, Ato+... +lit4o is within the target total lethality Fag by a predefined lethality tolerance AF.

If this is the case, the reel speed obtained in step 160 in the last iteration is used as the initially scheduled reel speed. However, if this is not the case, then the reel speed from the last iteration is overly conservative. As a result, a second additional step may be added to increase this reel speed by, for example,. 5Av. Steps 151 to 159 and the two additional steps

are then repeated until the first additional step is satisfied. In this way, the initially scheduled reel speed is further refined in an aggressive manner.

Similarly, a more aggressive approach can also be taken for defining the re-scheduled reel speed v, j. In this case, the steps 178 to 188 of Fig. 9 discussed in section l. f. would also include the two additional steps just described.

2. e. Deviations in Scheduled Initial Product Temperature and/or Reel Speed In addition to temperature deviations in the scheduled retort temperatures TSRTIO...

TSRTSO, there may be deviations in other scheduled parameters of the rotary sterilization process. For example, there may be deviations in the scheduled initial product temperature T.,, and/or the currently scheduled reel speed vs. Thus, the controller 104 may be configured to handle these deviations as well in order to identify any under and/or over processed containers {..., i,...} underpr and/or {..., i, -} overpr resulting from the deviation. This is done in a similar manner to that described earlier in sections l. b. to l. e. for temperature deviations in the scheduled retort temperatures.

2. d. More Aggressive Approaches A more aggressive approach than that described earlier in section l. c. can be taken for defining the initially scheduled reel speed v5°. In this approach, a first additional step could be added after step 159 of Fig. 6 to determine whether the predicted total lethality Fi'ove 0, Atlo+... +At4o is within the target total lethality F, arg by a predefined lethality tolerance AF.

If this is the case, the reel speed obtained in step 160 in the last iteration is used as the initially scheduled reel speed. However, if this is not the case, then the reel speed from the last iteration is overly conservative. As a result, a second additional step may be added to increase this reel speed by. 5Av. Steps 151 to 159 and the two additional steps are then repeated until the first additional step is satisfied. In this way, the initially scheduled reel speed is further refined in an aggressive manner.

Similarly, a more aggressive approach can also be taken for defining the re-scheduled reel speed vs. In this case, the steps 178 to 188 of Fig. 9 discussed in section l. f. would also

include the two additional steps just described. However, another additional step would also have to be added.

2. e. Deviations in Scheduled Initial Product Temperature and/or Reel Speed In addition to temperature deviations in the scheduled retort temperatures TA-, TSRT40, there may be deviations in other scheduled parameters of the hydrostatic sterilization process. For example, there may be deviations in the scheduled initial product temperature T,,, and/or the currently scheduled reel speed vs. These deviations would be detected by monitoring the actual initial product time-temperature profile Talp (t) and the actual reel time- speed profile va (ty. Thus, the controller 104 may be configured to handle these deviations as well in order to identify and segregate any under and/or over processed containers (.... i,... underpr and/or {..., i,...} ovetpr resulting from the deviation. This is done in a similar manner to that described earlier in sections l. b. to l. e. for temperature deviations in the scheduled retort temperatures.

2. f. Different Combinations of Cooling and Cooking Shells and Temperature Zones The rotary sterilizer 102 of Fig. 1 was described as having one cooking shell 106-1 with three temperature zones 115-1,..., 3 and one cooling shell 106-2 with two temperature zones 115-4 and 5. Correspondingly, the flow and timing diagrams of Figs. 3 to 9 were described in this context as well. However, those skilled in the art will recognize that the rotary sterilizer may have more than one cooking shell and more than one cooling shell with more or less temperature zones. For example, in a simple case, the cooking and cooling shells may each have just one uniform temperature zone. As those skilled in the art will recognize, the flow and timing diagrams of Figs. 3 to 9 would have to be correspondingly adjusted for the specific combination of cooking and cooling shells and temperature zones used.

2. g. Other Continuous Source Sterilization Systems The present invention has been described in the context of a rotary sterilization system 100. However, as those skilled in the art will recognize, the invention can be similarly practiced in any other continuous source sterilization system in which containers or carriers of containers are conveyed in line through the system's sterilizer. For example, the invention may be used in a hydrostatic sterilizer, as described in concurrently filed and co-pending U. S. patent application serial no. 09/, entitled Controller and Method for Administering and Providing On-Line Handling of Deviations in a Hydrostatic Sterilization Process, filed on November 6,1998, with Weng, Z. as named inventor. This patent application is hereby explicitly incorporated by reference.

3. Conclusion While the present invention has been described with reference to a few specific embodiments, the description is illustrative of the invention and is not to be construed as limiting the invention. Various modifications may occur to those skilled in the art without departing from the true spirit and scope of the invention as defined by the appended claims.