Research On Online Active Measurement Technology Of Geometric Features For Grinding

Grinding is the key process to determine the quality of the workpiece.On-line active measurement and error feedback compensation control technology can effectively improve the processing accuracy,ensure the quality of the workpiece,and improve the processing efficiency and the degree of automation of the processing system.As a kind of automatic processing control instrument that can be unified with the machine tool,the grinding processing active measuring instrument is an automatic processing instrument that includes intelligent control measurement,error signal feedback and signal acquisition functions.Under the existing conditions,although the active measuring instrument for grinding processing can achieve the automation of online measurement and the human control of the measurement and adjustment module,it cannot estimate the processing flow of the workpiece based on the online measurement,and cannot perform the workpiece in the grinding process.Implement automated control and improve in terms of control,display,operation,reliability,and signal acquisition accuracy.With the goal of high precision,strong real-time and high degree of automation,this subject researches the online measurement and error compensation control technology of grinding processing geometric features,and realizes the closed-loop control of active measurement,online measurement,and error feedback adjustment of the controller.The subject mainly studies the construction of the typical form and position error dynamic measurement model and the form and position error inspection operator used in the grinding process,the optimization setting,and the construction of the grinding process error feedback control system to realize the online workpiece in the process Measurement and compensation control technology;the research results of this project are of great significance to the improvement of active measurement technology,and provide a theoretical basis for the function expansion and production application of the new grinding process active measurement control instrument.The main tasks completed are:(1)According to the new generation of GPS specification,typical shape and position error inspection operators(separation,extraction,filtering,fitting,evaluation)and linear dimension inspection operators are constructed.According to the mechanism of active measurement grinding and the operator of product geometric specification,the scheme of separation and extraction of geometric error is established,the online evaluation of typical geometric error is realized,and the unification of geometric product linear dimension specification design and inspection certification is realized.2)Research on the detection and evaluation method of on-line typical form and position error: the detection and evaluation method of grinding on-line typical form and position error(roundness,cylindricity)is studied.This paper presents an effective method of roundness error evaluation based on Graham algorithm,which can evaluate roundness error by replacing all measuring points with feature points,and improve calculation efficiency.A new method of evaluating cylindricity is introduced,which transforms the solution of 3D cylindricity error into the solution of 2D Flatness Based on the idea of dimension reduction.Combined with the separation technology of form and position errors,the accurate expression of form and position errors under dynamic conditions is realized,thus the dynamic measurement model of form and position errors is constructed,and the online rapid evaluation of typical form and position errors is realized.(3)Research on active incremental learning mechanism and algorithm of support vector regression machine.The standard support vector machine itself can not solve the problems of large-scale data sets and data stream online processing.Therefore,an improved active incremental learning mechanism is proposed.In the training process,the support vector machine and support vector set are obtained by cluster analysis method,and the final support vector machine is judged according to KKT condition.This method can effectively improve the learning speed of SVM.Based on the optimization of standard support vector machine,the fuzzy least square online support vector regression(fllsvm)model is constructed.According to the model,the grinding dimension adjustment value is predicted,and the prediction value is analyzed for different processing factors,so as to improve the grinding efficiency and meet the realtime requirements.(4)This paper analyzes and studies the problems in the actual operation and production of grinding active measuring instrument,enriches the software and hardware functions of the active measuring instrument,and improves the active measuring instrument system,mainly including: realizing the online measurement function of typical roundness error,completing the application of convex hull method to evaluate roundness error in the active measuring instrument,and adopting single and multiple influencing factors to supplement and adjust the value prediction The function of on-line dimension prediction in grinding process is applied in practical production.