基于深度学习的家用电器电机故障诊断方法研究

Research on Fault Diagnosis Method of Electrical Motor Noise for Household Appliances Based on Deep Learning

针对目前家用电器电机采用人工听诊方式判断电机故障的现状,设计基于深度学习的电机故障诊断方法,旨在实现电机生产线的自动化与智能化。文章设计一个二分支的一维卷积神经网络,并在该基础上优化混合切片二分支卷积神经网络模型。经实验验证,该网络使用后,训练集准确率能达到99.67%,测试集准确率能达到98%,采用该方法进行电机故障噪声诊断准确率高且实用性好。

Aiming at the current status quo of using manual auscultation to determine motor faults in household electrical appliances motors,a deep learning-based motor fault diagnosis method is designed,aiming to realize automation and intelligence of motor production line.The article designs a two-branch one-dimensional convolutional neural network,and optimizes the hybrid sliced two-branch Convolutional Neural Network model on that basis.After experimental verification,the accuracy of the training set can reach 99.67%and the accuracy of the test set can reach 98%after the use of this network,and the method is used for motor fault noise diagnosis with high accuracy and good practicality.