The roundness error is separated by Fourier transform and hypothesis test of mathematical statistics.For high-precision measurement point data of circular contours,the minimum area method is used to fit the center of the circle.Fourier transform is applied to transform the circular contour signal into harmonic signal.Starting from the ideal circular contour model,Fourier components are gradually added.In the iterative process,the dominant model containing the unknown systematic form error is constructed through partial F-test.Residual significance Fourier components are obtained through autocorrelation test,chi square test,and J-B test,in order to further construct a significant model and achieve the goal of separating systematic errors from roundness errors.Through the analysis of experimental data on 16 circular holes processed by CNC,it has been proven that this method can effectively separate roundness errors.
维普
万方数据