Experimental Implementation of Robust ANFSI-PI based Hybrid Controller for DC Motor Driver System

Abstract

This paper presents experimental implementation of a robust Adaptive Neuro-Fuzzy Inference System-Proportional Integral (ANFIS-PI) based Hybrid Controller (HC) for dc motor drive system. Conventionally, PI controllers are extensively used for control of dc motor drive system. However, because of inherent working mechanism of PI controllers they suffer from significant overshoot and settling time. Some industrial applications require accurate, fast and stable response of the dc motor drive system. To completely eliminate the overshoot, concurrently, to improve settling time the HC (combination of ANFIS and PI) is presented in this paper. The proposed control scheme is simulated in MATLAB/Simulink. Also, it is experimentally implemented using DS1104 R&D Controller Board from dSPACE. For fair comparison, PI controllers are properly tuned using ControlSystem Designer tool available in MATLAB to get best possible response. Obtained experimental results show that, PI controllers have significant overshoot, i.e., 25.5% and settling time 1.80 sec. Whereas, proposed ANFIS-PI based HC outperforms PI controllers by giving 0% overshoot with 0.08 sec settling time, with identical operating conditions, thus making it suitable for industrial applications.