Online Monitoring And Fault Diagnosis Of High-end CNC Machine Tool Feed Syste

High-end CNC machine is the main production equipment in China ’s high-end equipment manufacturing industry.As an important part of CNC machine,the feed system’s health status is crucial for ensuring the processing quality and efficiency of CNC machine.At present,the mechanical fault maintenance of high-end CNC machine feed system mainly depends on regular inspection and artificial fault diagnosis,,which can cause untimely fault early warning and high maintenance costs.In this regard,this paper takes the feed system of VMC-850 L vertical machining center as the research object.A variety of signals are selected for monitoring,and the fault diagnosis method based on multi-source monitoring signals is studied to construct an intelligent diagnosis model with high accuracy.In addition,an online monitoring and fault diagnosis system is developed.The specific research contents are as follows :In order to obtain comprehensive information on the health status of the feed system of CNC machine,a feed system online monitoring scheme based on current,vibration,noise and CNC machine feed servo signals is proposed through an in-depth analysis of fault mechanism and multiple signal monitoring principles.In addition,the corresponding data acquisition technology is analyzed.Aiming at the difficulty of artificial fault location of CNC machine feed system,an intelligent fault diagnosis method based on multi-domain feature extraction and ensemble learning model is proposed.Firstly,the monitored signals are preprocessed by wavelet threshold denoising.Then,multiple time-domain and frequency-domain feature indexes of the signal are extracted respectively.Moreover,the signal is decomposed by adaptive noise ensemble empirical mode decomposition,and multiple IMF information entropy is calculated to construct a multi-dimensional mixed domain feature set.Finally,based on the Light GBM model,a Double Ensemble-Light GBM fault diagnosis model is constructed,which can realize sample reweighting training and automatic screening of sensitive features.The effectiveness and advancement of the method are verified by the fault diagnosis experiments using the public dataset and the feed experimental bench self-built datasets and comparing with advanced integrated learning models such as XGBoost.In order to cope with the fault diagnosis problem of CNC machine under complex working conditions,a deep learning-based fault diagnosis method is proposed,combining CNN and Bi LSTM for adaptive feature extraction,and a DARTS-CNN-Bi LSTM fault diagnosis model is established.First,in order to enhance the spatial feature extraction capability of the model to cope with the effects of changing working conditions,the CNN network structure is optimized by DARTS algorithm on the variable working condition data set.Then,CNN and Bi LSTM are combined to further extract the time series features of the signal,and feature dimension reduction is performed by global average pooling.Finally,the softmax classifier is used to complete the fault classification.The effectiveness and advancement of the method are verified by the fault diagnosis experiments using the public dataset and the feed experimental bench self-built datasets and comparing with advanced deep learning models such as Dense Net.Based on the above research,an online monitoring and fault diagnosis system for highend CNC machine is designed.The visual interface development is carried out based on the Pyqt framework,and the functions of user login,real-time equipment status monitoring,intelligent fault diagnosis,database management and so on are realized.After the completion of the system development,the availability of the system is proved by testing the key functional modules.