Control of food batch processes based on human knowledge

Batch-to-batch methodologies use the repetitive nature of batch processes in order to determine the optimal operating policy in the presence of uncertainty. In this paper, a general and flexible framework for batch-to-batch control alone or combined with feedback control is proposed: it constitutes a novel multivariable technique for batch processes implemented to obtain a product having a predetermined set of characteristics at the end of the run. This control technique relies on human knowledge used as a tool for evaluating product quality and controlling processes. It was used to control two food processing operations: the chopping operation for manufacturing meat emulsion and the ripening operation applied to obtain the dry sausage. An experimental validation is introduced: the results show that the algorithm is likely to be able to control the process so as to obtain the desired sensory characteristics.