Application of computer simulation and artificial intelligence technologies for modeling and optimization of food thermal processing.

摘要

The major objective of this project was to evaluate the feasibility of artificial neural networks (ANNs) and genetic algorithms (GAs) for modeling and optimization of food thermal processing. The specific objectives were: (1) to develop a comprehensive computer simulation program for thermal processing, (2) to apply ANNs and GAs for modeling and optimization of constant retort temperature (CRT) thermal processing and variable retort temperature (VRT) thermal processing, (3) to develop dynamic models for thermal processing using ANNs, and (4) to explore ANN-model-based analysis of critical control points for deviant thermal processes.; As a preliminary research, neural network models were successfully developed for modeling of residence time distribution (RTD) under aseptic processing conditions. The main configuration parameters of neural networks such as the number of hidden layers and their neurons, learning runs, choice of transfer functions and learning rules were optimized.; In order to provide experimental data needed for developing and testing of ANN models and GA optimization, a comprehensive finite difference computer simulation program for thermal processing was first developed in MS Visual Basic language, which could be used for simulating different thermal processes such as constant retort temperature (CRT) and variable retort temperature (VRT) thermal processing.; The second objective was focused on developing modeling and optimization methods for CRT thermal processing using ANNs and GAs. The ANN models were developed for predicting process time, average quality retention, surface cook value, final temperature difference, lethality ratio, and equivalent energy consumption. Using this optimization program, the effects of process variables on the optimal retort temperature and the maximum average quality retention were investigated.; The final part of the thesis research was focused on applying ANN methods for the analysis of critical control points (CCPs) for deviant thermal processes, one of the important steps required for developing hazard analysis of critical control points (HACCP) program. The results indicated that ANN models could be efficiently used for the analysis of CCPs of thermal processing. Such a concept can be expanded for developing an ANN based HACCP expert system for thermal processing. (Abstract shortened by UMI.)