Method of non-stationary random vibration reliability of hydro-turbine generator unit

The hydraulic excitation acting on a hydro-turbine generator unit exhibits obvious non-stationary characteristics. In order to account for these characteristics, this study focuses on the non-stationary random vibration reliability of the hydro-turbine generator unit. Firstly, the non-stationary characteristics of the hydraulic excitation are analyzed, and a mathematical expression is constructed using the virtual excitation method. Secondly, a dynamic model of the unit is established to demonstrate the non-stationary random vibration characteristics under hydraulic excitation. Thirdly, an active learning non-stationary vibration reliability analysis method AK-MCS-T-H is proposed combining the Kriging model, the Monte Carlo simulation (MCS) method, and the information entropy learning function H. This method reveals the influence of the non-stationary hydraulic excitation on the random vibration reliability of the hydro-turbine generator unit. Finally, an example is presented to analyze the random vibration reliability. The study shows that the AK-MCS-T-H proposed in this paper can solve the problem of non-stationary random vibration reliability of the Francis hydro-turbine generator unit more effectively.