Fault Diagnosis And Fault Tolerant System Of Brushless DC Motor Sensor Based On Neural Network

Brushless DC motors(BLDC)are widely used in various fields due to their excellent characteristics.At present,there are position sensor control for BLDC,and hall sensors are the most commonly used.However,hall sensors are prone to failure under complex and variable conditions.At present,the fault diagnosis and fault-tolerant processing of hall sensor for BLDC have received extensive attention and in-depth research,but there are still many problems.In recent years,the neural network has provided a new solution for fault diagnosis.In this paper,a new method for fault diagnosis of the BLDC sensor using BP neural network and fault-tolerant processing after diagnosis is proposed.The main work and research of this paper are as follows:This paper introduces BP neural network and establishes BP neural network fault diagnosis model in MATLAB/Simulink environment.Several parameters that can reflect the motor fault state are used as the identifieation features of the neural network.The neural network is used to learn the fault feature data.The BP neural network is used for fault detection of sensors.BLDC hall sensor fault-tolerant control system adopts position sensorless control.For the switching system,the motor rotation position sometimes has deviation.This paper adds S function to the hall sensor output value and the output of the line back electromotive force module.The eontrol values are unified to improve the reliability of the system.When the neural network detects that the hall sensor has failed,it switches to the position sensorless control system to achieve fault-tolerant operation of the motor.Finally,this paper designs and constructs a fault-tolerant control system platform for BLDC based on FPGA,and experiments on the proposed motor fault detection and fault-tolerant control system.The experimental results are similar to the simulation results,which proves the correctness of the scheme.Feasibility provides new ideas for follow-up research.