Research On CNC Machine Tools' Straightness Error On-Machine-Inertial Measurement

CNC machine tools generally have the problem of poor accuracy retention,which requires long-term precision measurement.The traditional measurement method based on laser interferometer is time-consuming,laborious and expensive,and can only get the space lowfrequency error items of machine tools.It is of great significance to develop efficient and lowcost on-line detection methods for high-efficiency accuracy measurement of machine tools.Therefore,the research on on-machine measurement of CNC machine tools is carried out in this paper.Acceleration sensors are used to measure straightness error,and Kalman filter is used to eliminate part of the test noise.The displacement is obtained by integrating the acceleration signal twice.Zero-phase high-pass filter is considered to remove the integral trend error,and the straightness error of the motion axis is obtained.Firstly,the basic principle of inertial measurement and the application of inertial measurement in straightness error measurement of CNC machine tools are introduced.The influence of the uniform moving speed of the test platform on the test acceleration signal is analyzed.The acceleration signal along the vertical axis of motion is measured under the conditions of high speed,medium speed and low speed,which broadens the bandwidth of the acceleration signal on the basis of the invariable spatial error frequency.A scheme of multispeed test data fusion is proposed,which fuses the measured data at different test speeds to improve the test accuracy and robustness.Secondly,aiming at the problem of zero drift and environmental noise in the measured initial acceleration signal,the data preprocessing algorithm based on Kalman filter is studied.AR model of measurement noise data is established using time series analysis method,and then the system state equation and observation equation in Kalman filter are obtained.The acceleration signal with noise is passed through Kalman filter to reduce the noise of the acceleration signal and improve the stability and accuracy of the data.Then,the process of acceleration integration is deduced,and the problem of error trend term error in acceleration integration is studied.The existing trend term error methods,including polynomial fitting,high-pass filtering,frequency domain low-frequency cut-off algorithms,are deduced and their shortcomings are found.A zero-phase high-pass filtering algorithm based on endpoint continuation is proposed,which ensures the phase stability before and after filtering and restrains the occurrence of ‘endpoint effect' in filtering.The results show that the proposed trend elimination algorithm is more advanced than the reference algorithms.Finally,experimental verification is carried out.The software and hardware platforms for experimental testing are built,and the data testing and processing schemes of this paper are verified by experiments under the conditions of artificial simulation error,sensor installation angle offset,and machine tool error measurement.The results were compared with reference values such as error presupposition value and laser interferometer measurement value.The comparison shows that the test scheme proposed in this paper has good accuracy and can meet the requirement of on-line measurement of straightness errors of NC machine tools.Provide an effective basis for the maintenance of machine tool accuracy...