Development of a fuzzy expert system for the selection of batch dryers forfoodstuffs

The selection of dryers is an art in which knowledge, experience and scienc e all play important roles. Historically, dryer selection has been made by experts on the basis of their extensive knowledge. However, in recent years , computer-based techniques have been developed, which have the potential o f at least partially deskilling this process. Of the various possibilities, fuzzy expert systems, in which the selection qualifiers are represented as linguistic rather than numerical variables, are the most promising. This p aper describes the development of a fuzzy system for the selection of batch dryers for food products. It featured a novel modular approach in which in dependent selection goals for dryer type, atmospheric, vacuum or freeze ope ration, and single or multiple units were adopted. This made the system par ticularly flexible and amenable to adaptation. The program starts from a 'd rying process checklist', in which the principal process variables are spec ified, and provides a ranked list of feasible alternative dryers. The algor ithm was extensively tested and provided quite plausible results. Four repr esentative case studies are presented and discussed.