Study on Operation Rules of Hydropower Stations Based on Random Forests Method
Taking Tankeng Hydropower Station as an example,the random forests regression models are established to simulate the long-term series of power generation and water level respectively by using two different sets of factors as the inputs.The influences of the number of input factors are analyzed and the generalization ability of model is evaluated.The results show that the operation rules extracted from random forests method has higher simulation accuracy,the introduction of derivative factors can slightly improve simulation results,the random forests method does not perform well when it comes to some particular inflow situations,and the long-term low water level is caused by the error propagation in random forests.
国家科技期刊平台
NSTL
万方数据
维普
