Fuzzy Supervision of PID Controllers' Tuning: Design and Implementation for On-Line Compensation of Nonlinearities

PID Controllers are the most common controller in industrial applications. Their main disadvantage is their loss of performance when the process under control modifies its behavior due to aging or change in operating conditions (raw materials, equipment, operators). This book presents two design alternatives of supervision techniques to on- line change the PID controller tuning parameters to improve the control loop performance. Since fuzzy logic has been used effectively to represent expert knowledge, the supervisors used expert rules in a fuzzy inference system to decide required changes based on process behavior. Two supervisors are designed, named FST1 and FST2 (Fuzzy Supervisor Type 1 and 2). FST1 is designed as a safety measure to act only when the loop becomes unstable. FST2 is intended for loop performance improvement, and it acts when the controller behavior is too agressive or sluggish.