Intelligent sensorless speed control techniques for brushless DC motor

Abstract

Brushless Direct Current motor is one of the major developments in the newlineevolution of special electrical machines due to its excellent performance, newlinelesser noise and energy efficiency. However the speed control of this motor is newlinestill a concern due to the presence of hall sensor. Many researchers over the newlinepast decade have proposed several approaches for the sensorless operation of newlineBLDC motor. However, those methods involve complex mathematical model newlineand involvement of several logical gate circuits which again increases the newlinesystem complexity. Hence, the sensorless speed control of BLDC motor is newlinestill in the thrust areas of research. This thesis makes an effort to provide newlinesolution to this problem by employing Artificial Intelligence Technologies for newlinesensorless operation of BLDC motor thereby minimizing the system newlinecomplexity. Instead of hall sensor, Adaptive Network based Fuzzy Inference newlineSystem (ANFIS) is trained to generate hall signals from the back emf of the newlineBLDC motor. The generated hall signals are then fed to an Artificial Neural newlineNetwork trained with the switching logic using back propagation algorithm to newlineproduce the pulses required for the inverter. This further reduces system newlinecomplexity by eliminating the need for logical gate design in the switching newlinecircuit. At the same time, the Proportional Integral (PI) controller is used to newlinecontrol the input DC voltage to the three phase inverter that feed the BLDC newlinemotor. newline The gain setting of PI controller plays a major role in effective speed newlinecontrol of BLDC motor. Conventional gain setting may result in reduced newlineperformance in speed control process. This necessitates the optimal setting of newlinegain parameters of PI controller. This research work uses evolutionary newlinealgorithms to optimize the PI controller. newline