Speed Control of Brushless DC Motor Using Artificial Neural Network Tuned PID Controller

Abstract

In automotive industry, high performance drives are gaining popularity due to their high efficiency, good dynamic response and low maintenance. The precise rotor movement over a period of time in certain applications such as robotics, guided manipulation and dynamic actuation must be achieved even when the system loads, inertia and other controlling parameters are varying. To do this, the speed control strategy must be adaptive, robust, accurate, and simple to implement. The conventional feedback controllers those are based on linear control theory and are much easier to understand and implement but suffer the disadvantages when the operating points of the process or the plant parameters are changed due to disturbances. Fixed-gain feedback controllers need to be returned to obtain the new optimal settings. For the processes with variable time delays, varying plant parameters, large non-linearties and considerable process noise, the PID controller does not give optimal performance. In view of this an adaptive controller that can modify its behaviour in response to the dynamical changes in the process and the disturbances is developed. Artificial Neural Network (ANN) based intelligent controller can mimic adaptive nature of controller used in non-linear system through its highly parallel and distributed structure. Neural network can generate a nonlinear mapping between the inputs and outputs of a system without the need for a predetermined model. The aim of the thesis is to design a simulation model of brushless dc motor and to control its speed at different values of load torques. The accurate speed control is proposed to achieve through PID controller. The parameters of PID controller are tuned by on line training of the artificial neural network. The performance of the PID type controller with fixed gain, conventional integral controller (PI) and ANN based PID controller have been compared through MATLAB simulation results with focus on feasibility, reliability and accuracy for BLDC permanent magnet synchronous motor drive system. The qualitative and quantitative comparison have shown the superiority of the performance of PID tuned through artificial neural network over integral and PID controller in terms of the reliability and feasibility. The reported percentage overshoot error is well within the permissible limits and rise time is also very low.