Systematic error separation and compensation for complex surface in OMM

The emergence of on-machine measurement (OMM) technology has opened up new possibilities for error compensation in complex surface machining. Due to the harsh working environments of machine tools, the error data obtained from OMM often contains a considerable degree of random errors. To separate the systematic errors from the random errors and improve the effectiveness of the measurement results and the accuracy of surface parts machining, a method known as ICEEMDAN-VMD is proposed. This method, based on improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), permutation entropy, correlation coefficient, power spectral entropy, and variational mode decomposition (VMD), is entirely data-driven and does not necessitate manual intervention, making it well-suited for OMM. The accuracy of error separation using the ICEEMDAN-VMD method was assessed through simulation, utilizing root mean square error and signal-to-noise ratio as evaluation metrics. Application of this method in machining experiments resulted in a substantial reduction of 73.58% in the average error, 64.33% in the maximum error, and 67.86% in the standard deviation on surfaces. This implementation notably enhanced surface consistency and effectively avoided over-cutting. The results conclusively demonstrated the theoretical feasibility, high accuracy, and practical effectiveness of the proposed method.