Study on water level prediction method of Shaping Hydropower Station based on BP neural network
Wenwen Feng, Xiaohui Lei, Chao Wang, Haocheng Huang · 2021
Runoff Hydropower Stations are widely present in cascade reservoirs. Its inflow and water level prediction are affected by many hydraulic and unit characteristic control factors, and the change trend has strong nonlinearity and randomness, which is difficult to accurately simulate with traditional hydrodynamic models. The problem of high-precision prediction of water level has become a major obstacle restricting the refined operation of hydropower stations. A new method of reservoir water level prediction based on BP neural network is proposed in this paper. A multi-layer BP neural network water level prediction model was established and applied to the water level prediction of Shaping Hydropower Station. Its results are compared with the simulation results of the water balance model. The results show that the BP neural network model has high accuracy. At the same time, the data is used to drive the operation of the model, thereby avoiding the influence of the hydropower station’s own static curve on the prediction results. This method has great practical value. This research can provide an important reference for the purpose of realizing water level prediction, output adjustment, unit scheduling optimization, and intelligent gate control of a run-off reverse regulating Runoff Hydropower Stations.