Research On Prediction And Monitoring Method Of Quality Fluctuation In NC Machining Process For Aerospace Complex

In the process of CNC finishing of complex structural parts in aerospace,the machining quality is very high,and the workpiece is scrapped due to the unqualified dimensional accuracy of the workpiece.The maintenance and stability of the machining quality urgently need to be improved.Machining quality has an inevitable connection with the excessive wear of CNC machine tools,but the relationship between the quality of the workpiece and the wear value of the tool is not clear,and it cannot play a guiding role in the machining process.Therefore,exploring the correlation law between the dimensional accuracy of the workpiece and the amount of tool wear is of great significance for improving the processing quality.In the finishing process of CNC machine tools,the fluctuation law of the workpiece size is not clear.The analysis of the size fluctuation range and the setting of the size detection threshold mainly rely on traditional methods,and there is a lack of methods for analyzing the size of various workpieces and detecting abnormal fluctuations based on big data.Therefore,it is necessary to explore a deep learning method using big data analysis to analyze and detect the quality of the workpiece in the processing process,so as to improve the processing quality of the workpiece.This paper takes the complex structure parts processed in the aerospace numerical control finishing process as the research object.Aiming at the above problems,the following contents are studied:(1)The establishment of a predictive model of workpiece quality.First,establish univariate and multivariate Long Short Memory Neural Network(LSTM)prediction models for the workpiece quality data to predict the workpiece size in real time.Then,in order to improve the prediction accuracy of the model,an improved method of the LSTM prediction model was proposed,and the prediction results of the model were compared.Finally,the model with high prediction accuracy is selected as the final prediction model.Through this model,online real-time prediction of the quality of the workpiece can be used to revise the abnormal processing process in time to maintain the stability of the processing quality.(2)Establishment of correlation analysis model between workpiece quality and corresponding tool compensation value.First of all,based on the workpiece quality data,tool compensation value data or other data that can reflect tool wear generated during the production process of CNC machine tools,establish an association analysis model between them.Then,the approximate matching relationship between the workpiece size and the tool offset value is established through the correlation analysis model,and the associated knowledge of the matching relationship between the workpiece size and the tool offset value is extracted and stored.Finally,the obtained related knowledge is further optimized to obtain the optimal combination of related knowledge.Based on the above results,the combined machining method of tool compensation value and optimal tool compensation value can be recommended.At the same time,the correlation analysis result can also modify the prediction result by analyzing the correlation between the tool compensation value and the prediction result of the LSTM prediction model,and improve the accuracy of workpiece quality monitoring.(3)Based on SPC and machine learning algorithms,a state discrimination model is established for workpiece quality data and an early warning threshold is set.First,establish a classification model to analyze the processing quality interval and the processed workpiece quantity interval under each tool compensation value.Then,SPC is used to mark the quality of the workpiece,and a classification model is established to classify and judge the quality of the workpiece.Finally,the workpiece data is divided into normal values and abnormal values,and the quality of the workpiece is classified based on the above division.A reasonable threshold is set for the accuracy of the workpiece size according to the results of the abnormal value classification and the SPC quality control line,thereby assisting quality early warning.Through the prediction of workpiece quality,the correlation analysis between quality and tool compensation value,the analysis of workpiece quality status and the setting of early warning threshold value,it provides reference for the improvement of workpiece quality during processing.