Research On Data Mining Technology Of CNC Lathe Cutting Parameters

CNC lathe,one of the most widely used CNC machine tools,is mainly used for the cutting machining of shaft or disk parts,like the inner and outer cylindrical surfaces,conical surfaces with arbitrary taper angles,complex curved surfaces and cylindrical and conical threads.There are two major problems in the current cutting process.The first problem is that the level of professional programming personnel engaged in numerical control processing is uneven,so that the processing quality cannot be guaranteed.The second problem is that as the amount of cutting data continues to increase,the Machining Manual alone cannot be used for large-scale reuse of cutting parameters.The cutting database combining mechanical processing technology and computer technology can effectively solve the problems and make the cutting data more applicable.This article first systematically summarizes the research situation of well-known database systems at home and abroad,and proposes this article's research content.Second,taking the shaft parts as the processing object,a correlation database for various turning parameters is established by My SQL database.The collected cutting data are put into the database as the original data,and then the Apriori algorithm is used to mine frequent cutting data,which provides initial data for subsequent data processing.The CBR(Case-based Reasoning)algorithm can match design examples with historical examples,which can provide the most similar examples for process makers.In order to find the historical example that is most similar to the design example,a similarity calculation method using the CBR algorithm and hierarchical strategy is proposed.This strategy is different from the traditional one because it classifies the refined attributes and performs the calculation sequentially from the first level to the last level.At the same time,the value of the weight coefficient is considered,the subjective weight and objective weight are combined by the multiplicative synthesis method to reflect the importance of each instance.The grading strategy was applied to the reasoning process of cutting cases,and the most similar example was retrieved from historical examples.Compared with the results of general case reasoning,the refined strategy can improve the overall similarity,and make the similarity difference of each historical instance more obvious.At the same time,in order to supplement the shortcomings that the CBR algorithm cannot retrieve instances that do not exist in the database or does not meet the threshold requirement of similarity,an ANN(Artificial Neural Network)model is established using Tensor Flow machine learning framework.It takes three cutting parameters as the input variables,including cutting depth ap,cutting speed vc and feed amount f,and chooses three indicators that characterize the surface roughness of the workpiece as output variables,which includes the average deviation Ra of the contour arithmetic,the maximum height Ry of the contour and the ten-point height of the micro unevenness Rz.On this basis,the neural network is trained using historical processing data,and the trained model is used to predict the surface roughness of the workpiece.The prediction results show that the surface roughness prediction method based on Tensor Flow framework is more accurate and more convenient for modeling.Finally,a CNC lathe cutting parameter database system using the B/S architecture is set up on the Pycharm IDE(Pycharm Integrated Development Environment)development platform,which realizes the functions of data mining,artificial neural network reasoning and case matching.The work done in this paper has a certain reference value for the intelligent programming of turning process.