Rational analysis and prediction of hydraulic turbine oscillation based on multi-model fusion

As the core mechanical equipment of hydropower plant, hydro-generator set is very important to the safe and stable operation of hydropower station. Under the combined influence of mechanical, hydraulic, electromagnetic coupling and aging of mechanical parts, the faults of hydropower units are mostly manifested in the form of oscillations, so the oscillations and oscillations signal can intuitively characterize the operating state of the units. In this paper, a vibration and pendulum analysis method based on multi-model fusion is studied. The model integrates SVR, LightGBM and XGBoost machine learning algorithms by least square method, and the R2 determination coefficient is used to evaluate the model. Experiments show that the fusion model can predict the oscillation value stably, and the R2 score of the fusion model can reach more than 0.98, which is more than 0.3 higher than that of the single model, verifying the effectiveness of the model.