A Hybrid Model For Predicting The Degradation Trend Of Hydropower Units Based On Deep Learning

The degradation of hydropower units affects the safe and stable operation of hydropower stations and even the power system. Obtaining the degradation trend of hydropower units and predicting it accurately is a problem to be solved. In this paper, a framework based on healthy state model and degradation trend prediction model is developed to obtain and predict the degradation trend of hydropower units. Firstly, the healthy state model of hydropower unit considering the effect of working water head and active power was established based on particle swarm optimization and artificial neural network. Secondly, according to the healthy state model, the degradation trend of hydropower unit was deduced. Then, the degradation trend was decomposed by variational mode decomposition to obtain several modal components, and long short-term memory neural network model was constructed for each modal component for prediction. Finally, the results of prediction models were superimposed to obtain the prediction of degradation trend. In this paper, the actual data of a certain hydropower station was used for verification, and the results show that the proposed model is suitable for predicting the deterioration trend of hydropower units.