Low cycle fatigue damage prediction of steam turbine rotor based on dynamic PLS

This paper establishes the prediction model of low cycle fatigue damage of steam turbine rotor using the rich data from finite element analysis. In order to monitor the damage, a multiple regression analysis of input data/output data with high correlation is made via dynamic PLS. The variation of the process parameters is extracted and it restrains the multiple dependency of the several parameters in different time series. Finally, a simulation of rolling process of a domestic 300MW turbine unit validates the effectiveness and accuracy of the prediction model based on dynamic PLS.