Research On Generator Set Scheduling Optimization Of Wind Farm Based On Power Prediction

Wind power is a green and renewable energy source,and the global wind resource reserves are very rich.The development of wind power can effectively alleviate the increasingly serious energy crisis and environmental problems,and has been widely concerned by the government,industry and academia.However,the randomness and non-stationarity of wind power itself bring difficulties to the power regulation and control of wind farms.A reasonable scheduling strategy can ensure the safe and reliable operation of wind farms,reduce the impact of wind power grid-connected on the grid,and Increase the regulation capacity of wind farms.How to realize the rationalization of wind farm power dispatch has become an important research direction.Based on the wind turbine power forecasting information,the paper studies the more efficient and reliable power scheduling strategy,and studies the short-term wind power forecasting,wind turbine health status assessment and wind farm power dispatching strategy,main tasks as follows:(1)A short-term prediction model of wind power based on modified AMPSO-LSSVM is proposed.The Complementary Ensemble Empirical Mode Decomposition(CEEMD)is used to decompose the wind power time series to reduce the non-stationarity of the signal.Aiming at the low prediction accuracy and poor optimization effect of the prediction model,this paper uses the improved adaptive mutation particle swarm optimization(AMPSO)to optimize the prediction model.By introducing the quadratic decreasing weighting factor and the random variation factor,the possibility that the algorithm falls into local optimum is reduced.The improved particle swarm optimization algorithm has a faster convergence rate by performing simulation tests on two classical performance test functions.Compared with the short-term wind power prediction model based on BP neural network,LSSVM and PSO-SVM,the results show that the prediction accuracy of the short-term wind power prediction model improved by 7%,19%and 8%,with higher prediction accuracy.(2)A health assessment framework for wind turbines based on Mahalanobis distance is proposed.Using the short-term prediction model of wind turbine power,the standard residual set of power prediction and the residual set in actual state are obtained,and the Mahalanobis distance is calculated to represent two residual sets.The degree of similarity between them is obtained by normalization to obtain the real-time health index of the wind turbine.The experimental results show that when the operating state of the wind turbines is abnormal,the evaluation framework can effectively indicate the trend of the health status of the wind turbine.The wind turbine health index can visually indicate the health status of the wind turbine.(3)The wind power prediction information and the health status of the wind turbine are proposed as the important factors of the wind turbine power dispatching strategy.The overall power deviation of the wind farm,the power fluctuation of the single wind turbine and the health status of the wind farm are optimized.A power scheduling optimization model is established.The improved AMPSO algorithm is used to optimize the parameters of the objective function.The experiment proves that the wind farm optimization scheduling strategy of this paper can avoid the overload operation of the wind turbine and ensure the smooth completion of the dispatching demand of the wind farm.