Compiling the Programmed Load Spectrum of Machine Tools Based on ACER-Weibull Extrapolation Method and Clustering Algorithm

Abstract The programmed load spectrum is an important foundation for reliability tests, reliability design, and life prediction. Because of the various working conditions and limited load information of machine tools, this characteristic needs to be fully considered when preparing the programmed load spectrum. In this article, a framework for compiling the programmed load spectrum of machine tools is established based on time-domain extrapolation and a clustering algorithm. First, the average conditional exceedance rate (ACER)-Weibull time-extrapolation method is proposed for load extrapolation. This method uses the average conditional exceedance rate method and the Weibull distribution to extrapolate the load above the maximum threshold and below the minimum threshold, respectively. Secondly, the Ckmeans.1d.dp clustering algorithm is used to divide the intervals of the programmed load spectrum, and the joint probability density function of the mean and amplitude is established using the Copula function. Finally, this framework is applied to compile the programmed load spectrum of computer numerical control lathes, and the cutting force spectrum is compiled, demonstrating the superiority of the proposed method.