Application of CEEMDAN algorithm in roundness error evaluation

Abstract Roundness error is an important evaluation criterion for evaluating machining errors. Some advanced algorithms have been applied to roundness assessment, such as the empirical mode decomposition (EMD) algorithm. However, due to the issue of mode mixing in the EMD algorithm, this paper uses the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) algorithm to evaluate roundness error. First, the same set of simulated circular hole radius measurement data was decomposed by EMD and CEEMDAN, respectively. Then, the direct component (DC) of each intrinsic mode function (IMF) generated after decomposition was used to calculate the cutoff wavenumbers of each IMF. The IMF was screened according to JB/T9924-2014, and the remaining IMF components were reconstructed (summed) after interfering signals were eliminated. The results of the roundness error calculated by the EMD and CEEMDAN algorithms were obtained and compared. The center of the circle fitted by the least square method was obtained by using the data points of the actual roundness contour coordinates, and the radius value of each sampling point was calculated. By comparing the experimental results, it was found that the accuracy of the CEEMDAN algorithm is higher than that of EMD, demonstrating the effectiveness of the CEEMDAN algorithm in roundness error evaluation.