Hydro-Turbine Coordination Power Predictive Method of Improved Multi-Layer Neural Network Considered Adaptive Anti-Normalisation Strategy

Due to the limitation of economics and time cost, the data obtained from hydro-turbine coordination field test are insufficient to fully guide the setting of unit operating parameters. To enlarge the amount of data, realise power point tracking, and avoid the problems of high non-linearity with hydro-turbine physical model which is difficult to simulate in actual field, a mathematical prediction model is proposed based on an improved multi-layer neural network. Using the rule activation function, L2 regularisation, Adam optimiser and its gradient parameters are optimised by PSO algorithm in the prediction model. It is found that lacking true value in the process of anti-normalisation leads to difficulty for actual forecast of neural network. Therefore, an adaptive anti-normalisation strategy is proposed to improve the actual prediction accuracy, which can judge the value of the interval. According to the analysis of examples with hydro-turbine coordination and non-coordination test, the results show that the proposed prediction model and interval strategy can effectively forecast the coordination operating conditions of the turbine with high accuracy under small samples.