Design And Realization Of Industrial Equipment Fault Diagnosis System Based On Neural Network

With the development of technology and industrial intelligence,more and more data is collected in the field of mechanical equipment health management,and the diagnosis analysis of equipment data and intelligent fault warning have also become a trend.However,the traditional methods cannot meet the needs of massive data processing in the context of big data,and the working condition information in the massive data is complex and changeable.The manual marking training method is contrary to the intelligent guiding ideology.In this context,it is necessary to adopt intelligent processing algorithms such as deep learning to design a new fault diagnosis model suitable for complex working conditions to improve the accuracy and robustness of equipment fault diagnosis.Therefore,this article focuses on the design and implementation of industrial equipment fault diagnosis system oriented to complex working conditions.The main research contents are as follows:1.Fault feature learning based on convolutional neural network: aiming at the problem of low reusability of fault high-dimensional features,propose a feature extraction algorithm based on onedimensional convolutional neural network to process data and information for the final classification mission.This method improve the mining performance and anti-noise ability of the model through structural improvement.Through comparative experiments,it is proved that the feature extraction method proposed in this paper has a significant improvement in the feature extraction ability of industrial equipment compared with mainstream neural networks.2.Auto-encoder-based domain mapping method: aiming at the problem of reduced diagnostic accuracy caused by domain drift caused by different working conditions,this paper proposes a domain mapping method based on auto-encoder,which maps the features extracted by the feature extraction network to the public space to eliminate the interference of fluctuations in operating conditions.Through comparison with the results of the methods proposed in the literature in recent years,the method is more robust and the accuracy rate is maintained at a higher level.3.Construction of industrial equipment diagnosis platform: The platform is consisted of the user module,the equipment management module,the equipment fault management module,and the monitoring data module,which can realize equipment fault warning and maintenance,equipment historical data storage and other functions to meet actual production needs.