Research on Automatic Fault Identification Method for Calligraphy and Painting Mounter Based on Machine Learning
A fault diagnosis model for mounting machine motors based on convolutional neural networks is proposed to address the issue of possible motor faults during operation.Attention mechanism and multi-scale modules are introduced,and the processing abil-ity of long and short term memory networks is utilized to enable the model to identify faults in high noise environments.The results show that when the number of iterations reaches 25,the accuracy of the hybrid model,one-dimensional convolutional wavelet neural network model,long-term and short-term memory model,and recurrent neural network model are 0.99,0.82,0.78,and 0.72,re-spectively.In larger datasets,the operation time of the four models is 0.56 s,0.64 s,0.62 s,and 0.58 s,respectively.This indi-cates that the proposed hybrid model can maintain a high level of performance even when facing a large validation set,and requires relatively few iterations.At the same time,it has been proven that the model can accurately judge the operating status of the motor in a large amount of noise,providing a new improvement idea for identifying faults in calligraphy and painting mounting.
calligraphy and painting mounting machinewavelet transformfault diagnosisMotorconvolutional neural network
万方数据
维普
