Method For CNC Machine Tool's Motion Error Abduction Based On Chaotic Characteristic

The precision of the CNC machine tool directly affects the precision of the workpiece,and the error existed in machine body will reflect on the complex machining shape,increase the error of workpiece.CNC machine tool's motion error abduction is particularly important to ensure the quality of the processing in high precision machining process.In recent years,researches on CNC machines tools precision were mostly emphasized on test,and identification.and reducing the reliability of the equipment.Factors affecting the accuracy of the CNC are diverse and complex,specific model is only applicable to corresponding types of machine tools and not common.In this paper,the method for abduction of CNC machine tool's motion error was proposed based on the chaotic characteristic.Chaos theory and fractal geometry was used to study the nonlinear system manifestations of CNC machine tools in case of different errors occurring.One-dimensional time series is generated by the roundness error data in the circumference measuring motion of Double Ball Bar.The wavelet analysis is adopted to denoise the time series.The chaotic characteristic parameters,such as time delay,embedding dimension,are obtained based on the C-C method,and then the chaotic phase space of CNC machine tool is reconstructed.The Wolf method is used to calculate the largest Lyapunov exponent,the G-P method is used to calculate the correlation dimension,and with the power spectrum picture,the chaotic characteristic of the machine tool system is found.Comprehensive study of these characteristic parameters indicate that different parameters influence of different error.So these characteristic parameters can be used as a tool to monitor the state of CNC machine tool and analysis of abduction.In the aspect of error evolution of machine tool,with the increase of circular feed rate,the largest Lyapunov exponent decreases.Finally realized the error of abduction network structures by SVM method,and the result shows that this method has a fast classification speed and a high accurate rate.The method described in this paper can provide greater support for the chaotic prediction of CNC machine tools and follow-up analysis of evolution.This paper mainly includes the following parts:Firstly,test of CNC machine tools roundness error and pretreatment of time series.It has been applied to the time series data of chaotic property analysis by roundness error experiment of machine tool under certain conditions.The wavelet analysis is adopted to denoise the time series.Secondly,the chaotic characteristic is used to analyze the error of machine tool.The main error items of the machine tool is fined,the chaotic characteristics and follow-up analysis of evolution are studied.Acquisition of four important chaotic characterization parameters: time delay,embedding dimension,the largest Lyapunov exponent,the correlation dimension;Power spectrum analysis of timeseries;Feature vectors are constructed to realize the mapping between the machine tool's circular motion error and the feature vector.Finally,study on the method for network error tracing of CNC machine tool.Based on SVM method,with four chaotic characteristic parameters of the feature vectors do as input,the six kinds of typical error sources do as network output,to build the abductive networks.Through training and testing,the abuction of CNC machine tool's motion error based on the chaotic characteristic is realized.And the results show that the recognition rate is higher.