Research On Fault Diagnosis Method Of Intelligent Manufacturing Equipment Based On Equipment Electrocardiogram

Intelligent manufacturing equipment is the main carrier of intelligent manufacturing and has become the competitive target of industrial countries.However,the current routine operation and maintenance methods of intelligent manufacturing equipment such as timing monitoring and post-repair cannot meet the purpose of real-time monitoring.Based on equipment electrocardiogram(ecg)as the center,the candy packaging production line of blanking robot fault prediction and diagnosis of study,combined with the latest stack noise reduction since coding network,was proposed based on equipment ecg,a new method of intelligent manufacturing equipment real-time fault diagnosis,ecg and equipment applied to fault diagnosis of the robot,the experimental results show the effectiveness of the proposed method.The main work of this paper is as follows:(1)Analyzed and studied the key technical issues of realizing ecg of intelligent manufacturing equipment.The ecg technology was applied to the mechanical equipment,and the key technologies to realize the ecg of the equipment,such as the definition of equipment state,data preprocessing and baseline modeling,were studied.Based on the analysis of a large amount of data,four states in the operation process of intelligent manufacturing equipment are defined,namely,Good,Watch,Warning and Abnormal value.An improved Rindar criterion is proposed to preprocess the data.The improvement is that the maximum value of each group of data is 2.1* Abnormal value,which can eliminate obvious gross errors in the data.During the operation of intelligent manufacturing equipment,whether the duration of each sub-action meets the time requirement needs to be measured and the baseline value is the execution time standard of each sub-action.Through a large number of data analysis,this paper established a baseline model.(2)Realize the visualization of device ecg through Python.Based on the Python legibility,good scalability and is rich in third party libraries,through Pyserial libraries to connect a serial port,sensor,data were collected through Numpy to deal with data collected,data preprocessing methods using in the second chapter puts forward the improvement of the Rhine criterion,the last of the target parameter by Mayplotlib library equipment real time visual display of ecg.(3)Equipment ecg was combined with improved SDAE algorithm for equipment fault diagnosis.Combining the equipment ecg with the deep neural network,the stacking de-noising selfcoding network can extract the features from the high-dimensional original data layer by layer,and the method based on the equipment ecg and the deep neural network is applied to the robot fault diagnosis.However,when the original time domain signal was input,the network had a large number of nodes and a complex structure.However,when the original time domain signal is input,the network node number is large and the structure is complex.In order to prevent overfitting in the network,dropout is used to randomly set the input node of SDAE to zero,so as to improve the generalization performance of the network.Using the Robot Execution Failure data set,the LP1 data set was analyzed in the time and frequency domain,and the feature vector with the contribution rate of >0.95 was retained.The multi-source features were weighted and fused with KPCA,and the fused features were input into the improved SDAE network to judge the robot's fault state.(4)The proposed fault diagnosis method is verified on the production line of small batch candy packaging,and the experimental results show the effectiveness of the proposed method.