基于连续小波变换和卷积神经网络的无刷直流电机故障诊断

Fault Diagnosis of Brushless Direct Current Motor Based on Continuous Wavelet Transform and Convolutional Neural Network

提出一种能够精确诊断无刷直流电机不同种类轴承故障、转子不平衡、霍尔元件故障和定子绕组三相不平衡故障的方法。将数据采集系统采集到的一维的机械振动信号进行连续小波变换,即可将一维的时域信号转变为二维的时频图像。对不同故障的时频图像利用基于卷积神经网络的深度学习算法进行训练,得到无刷直流电机故障网络模型。利用训练好的模型对验证数据进行推理,即可实现电机故障检测和分类。实验表明,电机8种健康/故障模式的分类精度接近100%。

This study proposes a method that can accurately detect the faults of brushless direct current motor（BLDCM）,including bearing fault,rotor imbalance,Hall sensor fault,imbalance of stator winding resistance.Firstly,the one-dimensional vibration signal acquired from the data acquisition system was transferred to the two-dimensional time-frequency spectrogram by using the continuous wavelet transform.Then,the spectrograms corresponding to different motor faults were trained by applying the deep learning technique based on the convolutional neural network（CNN）.Hence a well-trained model for BLDCM fault diagnosis was established.Finally,motor faults could be detected and classified by applying the simulated model to analyze the testing data.The experimental results show that the classification accuracy is close to 100% when the proposed method is used to motor faults diagnosis.