DEVELOPMENT OF A MODEL REFERENCE ADAPTIVE SYSTEM WITH A FUZZY-PID REGULATOR FOR SPEED CONTROL OF A SINGLE-PHASE INDUCTION MOTOR

Abstract

Accurate speed control of Single-Phase Induction Motors (SPIMs) is difficult to achieve without physical speed sensors, yet sensor-based schemes add cost, complexity, and vulnerability to noise and mechanical wear. Most sensorless control research has concentrated on three-phase machines, leaving SPIM sensorless control comparatively underexplored, and existing estimators such as the Model Reference Adaptive System (MRAS) lose accuracy under parameter variation and load disturbance while conventional PID regulation alone struggles to keep pace. This thesis addresses this gap by presenting the design, implementation, and validation of a sensorless speed control system for a SPIM that integrates MRAS speed estimator with a Fuzzy Logic-Proportional Integral Derivative (Fuzzy-PID) controller. Its novelty lies in combining MRAS's sensorless adaptability with a Fuzzy-PID regulator that dynamically retunes its gains from the estimated speed error, so that estimation and control adapt together rather than as separate stages. The control model was developed in MATLAB/Simulink and evaluated under varying load conditions. The MRAS estimated rotor speed by comparing reference and adjustable models, eliminating the need for physical sensors; the fuzzy logic controller handled system nonlinearities through linguistic rules and membership functions; and the PID controller continuously adapted its gains in response to the error signal. Simulation results showed that the integrated MRAS-Fuzzy-PID controller outperformed both MRAS-Fuzzy controller and a reference MRAS-only system on every metric tested. It achieved the fastest rise time, at 0.36 seconds, against 0.43 seconds for MRAS Fuzzy and 0.45 seconds for the reference system; settling time was reduced to 1.3 seconds, a 50% improvement over the initial setup; and steady-state error was minimized to 0.03%, compared with 0.12% and 0.22% respectively. The final settled speed reached 1429 rpm, closely matching the reference and exceeding the accuracy of both comparative controllers. These results confirm the superior speed-tracking accuracy, faster dynamic response, and robust adaptability of the MRAS-Fuzzy-PID approach under load variation. By removing the physical speed sensor while retaining accurate, adaptive regulation, the proposed system offers strong potential for low-cost, energy-efficient, intelligent motor control in residential and industrial applications, contributing an integrated estimation-and-control framework for future sensorless single-phase drives.