Sensorless PMBLDC Motor Control Strategies by Artificial Neural Network (ANN) with PI Controller

Abstract

Due to benefits in terms of cost, complexity, and reliability, sensorless control of BLDC motors has attracted a lot of interest. In order to obtain effective and precise control of a sensorless BLDC motor, a control strategy that combines the skills of an artificial neural network (ANN) and a proportional integral (PI) controller is needed. Traditional BLDC motor control techniques call for sensors to measure things like rotor position or speed. Sensor use, however, raises system complexity and expense. As opposed to this, sensorless control systems estimate the rotor position and speed using the motor’s back-EMF voltage or current. This eliminates the requirement for additional sensors. In this paper, the rotor location and speed of the BLDC motor are estimated using an ANN. An electrical motor signal dataset and a dataset of known rotor locations are used to train the artificial neural network (ANN). The ANN may learn the correlation between the electrical data and the rotor position during the training procedure. Without the use of physical sensors, the ANN can estimate the rotor location and speed in real time once it has been trained. A PI controller is coupled with the ANN-based estimator to improve control performance. The PI controller generates control signals to modify the phase currents in the motor by comparing the estimated rotor position and speed to the intended values. With minimum inaccuracy, the motor can track the intended speed or position, thanks to excellent motor control. On a prototype BLDC motor, the suggested control strategy is put into practice and tested. The results of the experiments show how successful the sensorless control approach is based on an ANN with PI controller.