SIMULTANEOUS OPTIMIZATION OF QUALITY RETENTION IN CONDUCTION-HEATED FOODS OF DIFFERENT GEOMETRIES
The complex method, a general and flexible optimization algorithm, was used to determine the variable process temperature profile during thermal processing to maximize (optimize) the volume average or surface retention of a nutrient in conduction heated foods of different geometries. The objective functions were the volume average concentration (VACN) and surface retention (SCN) of a nutrient during the process, the implicit constraints were the target lethality at the cold point, and a threshold below which the center temperature must reach at the end of the process. A process temperature range of 5-150°C was used as an explicit constraint. Another explicit constraint was that the process temperature had to reach 5°C at the end of the process. The control variables for the optimization were variable process temperature profiles (VPTP) at equidistant time steps throughout the process for the optimization of thermal processing for a sphere or finite cylinder for single geometry systems and their combination for a double geometry system. In the double geometry system optimization, a multi-objective optimization problem, the Weighing Method (to connect the objective functions for each geometry) and the Lexicographic Ordering Method (to guarantee the most important objective function preserves its optimal value) were combined with the modified Complex Method. All calculations were performed using a computer program written in Visual Basic V. 6.0. This Windows-based software calculated the optimum variable temperature profiles using the geometry options, objective functions, and constraints chosen by the user. VPTPs for different sizes of individual geometries (sphere, finite cylinder), alone or in combination (sphere and finite cylinder) at different processing conditions, together with their nutrient retentions compared to the constant temperature processes, were presented. Effects of time steps, and reproducibility of the algorithm were shown, and the uniqueness problem of the method was addressed.
The VPTPs resulted in some nutrient retention improvements compared to the constant temperature processes. In a single geometry system, the improvements were less than 1 percentage point in both VACN and SCN for smaller size geometries, while they were up to 2-3 percentage points in VACN and up to 3-4 percentage points in SCN for larger sizes. For double geometry systems, improvements were higher (6 percentage point improvement in both VACN and SCN). This was more obvious in the smaller geometries when the double geometry system consisted of a large and small geometry.