期刊名称：IOP Conference Series: Earth and Environmental Science
印刷版ISSN：1755-1307
电子版ISSN：1755-1315
出版年度：2019
卷号：300
期号：4
页码：1-8
DOI：10.1088/1755-1315/300/4/042097
出版社：IOP Publishing
摘要：The load spectrum is primarily used to provide a dynamic raw input of a basic load change to the component for simulation calculations or fatigue tests of fatigue life. In the load spectrum compilation of engine crankshaft, in order to preserve the influence of load sequence effect on fatigue damage during extrapolation and improve the accuracy of time domain extrapolation, this paper proposes a cyclic time domain extrapolation method based on SVR model. The research results show that the machine learning model has good learning ability and generalization ability, and the time domain extrapolation method can better realize the expansion of the measured samples of the crankshaft.
