Wind Power Forecasting And The Study Of Wind Power, Hydropower Coordinately

With the scale of wind generation enlarging, the researches on wind generation are becoming more and more active both at home and abroad. The wind generation has disadvantages of unsteadiness and indeterminacy, which hydro generation can avoid, so in recent years the study of wind-hydro coordinated generation has become the best technological way and there have been some research achievements on it. The present thesis, from the beginning of researching on wind speed forecasting, makes further study on coordinated operations on wind-hydro generation.Firstly, three different wind prediction models, namely, Time series forecasting model, BP neural network forecasting model and improved PSO-BP neural network forecasting model are studied in the thesis. According to specific examples, Time series forecasting model is better than improved PSO-BP neural network forecasting model and improved PSO-BP neural network forecasting model is better than BP neural network forecasting model in the same situation. Secondly, the author builds model of the wind-hydro coordinated generation based on Time series forecasting model. In the model, daily load is supplied by both wind generation and hydro generation, and it is supposed that wind generation is used up and the insufficient part is supplied by hydro generation, taking into account the wind energy cannot be stored, the target function can only choose the variables related to the hydro generation when the model of the wind-hydro coordinated generation is established, wind generation is considered into the constraints. For the soulution of target function, this paper uses hybrid algorithm of improved PSO and BPSO, adopts improved PSO algorithm of the units load distribution and uses BPSO algorithm of start and stop. The example analysis shows the hybrid algorithm is effective.In the end, the paper made a conclusion and prospected the next work.