A KNOWLEDGE-BASED SYSTEM USING FUZZY INFERENCE FOR SUPERVISORY CONTROL OF BIOPROCESSES

In this paper, a rule-based, real-time knowledge based system for biop rocess fault diagnosis and control is described. The system was design ed to generate on-line advice for the operators and to supervise autom atic control of bioprocesses, using biotechnical production of lactic acid as an example process. It consists of a real-time data acquisitio n and data processing system linked to a fuzzy expert system written i n Smalltalk V/Mac. The expert knowledge was expressed in the form of a rule-based knowledge network, fuzzy membership functions and control strategies. The fuzzy expert system carries out on-line fault diagnosi ng on the basis of filtered specific rates calculated from process var iable measurements, and provides suitable countermeasures to recover t he process. Fault diagnosis was realized both by backward and forward chaining procedures. The system was constructed to allow three differe nt control strategies (given here in Smalltalk syntax), change of Setp oint, FuzzyAnswer for each discovered fault, employing the fuzzy mean defuzzification method, and linguistic Advice to the operator. The sys tem was successfully tested on-line with a laboratory scale process.