摘要:Load extrapolation is the paramount step in compiling load spectra of the mechanical components. To avoid the limitation of the stationary processes in parametric extrapolation methods, the non-parametric extrapolation method is widely investigated in recent years. However, the accuracy of kernel density estimation of the large load cycles in existing non-parametric methods should be still improved. Aiming at this issue, a non-parametric rain-flow extrapolation method based on load extension is presented. In this method, combining with non-parametric extrapolation model and Markov chain Monte Carlo method, the measured load–time histories for semi-axle in different operating sections are extended and extrapolated. The extrapolated load spectra curves and the relative damage ratio are compared with the results obtained by the existing rain-flow extrapolation methods. The results show that the proposed method can realize the reasonable extrapolation of load–time histories.