An Auto-Tuning Fuzzy PI Controller Using Subtractive Clustering For Spherical Tank Process – A Real-Time Performance Evaluation

Abstract

In this paper, a simple and effective method for tuning of fuzzy PI controller based on fuzzy logic with subtractive clustering which minimizes the number of rules is proposed. The tuning of input scaling factors is done by gain updating factors determined by rule base which is designed based on the performance measures such as peak overshoot, rise time and amplitude. The subtractive clustering method is used to reduce the fuzzy inference rules of the three fuzzy reasoning blocks, namely rule base for α, rule base for β, and control rule base. The real-time performance comparison of auto tuned fuzzy PI with and without subtractive clustering is done in terms of several performance measures such as settling time, rise time and integral square error. To show the effectiveness and robustness of the proposed auto tuning mechanism, the analysis is carried out on a spherical tank process, which operates over a wide range and non-linear in nature. Also, the results show that the computational time and the memory requirements are reduced to a great extent for the subtractive clustering approach.