Multi-objective Optimization Allocation Of Wind Farm Hybrid Energy Storage Capacity Based On Wind Power Prediction

Today,the energy-saving and emission-reduction is advocated around the word.In this context,wind power generation technology,as one of the most mature and clean power generation technologies,has been widely used,and its installed capacity has been increasing year by year.However the randomness and intermittence of the wind power cause problems for the power system such as voltage and power fluctuations with large-scale wind power generators connecting to power grid.And these problems will influence the security and stability of power grid operation.In addition,wind power is characterized by anti-peaking,So when the wind powers capacity proportion increases,the grid dispatching department will possibly limit the wind farm's power generation to a certain degree.This will lead to the waste on the clean wind resources and the reduction on revenues.Therefore for the problems,the paper,considering a wind farm in Shanxi province as the actual research object,takes a deep research on these aspects such as prediction of the wind power,the optimal allocation strategy of the hybrid energy storage system,establishment and solution method of the multi-objective optimal allocation model.The main works are as follows:1.The paper analyzes the characteristics of wind farm output power and its influences on power system and then determines the major factors influencing the wind farm's output clearly;The energy storage technology's principle and common energy storage devices are introduced minutely,and then the existing wind farm evaluation index is analyzed.It can provide a theory basis to the following study.2.A dynamic ANFIS method with a regression structure for wind power prediction is given to increase the precision of the prediction model.The paper studies the traditional adaptive neural fuzzy inference system(ANFIS)and then puts forward a dynamic ANFIS method with a regression structure for the time sequence characteristics of wind power and wind speed.The wind power prediction model is got by using the dynamic ANFIS method with a regression structure.And the most excellent prediction model is obtained by training,checking and testing using the historical operating data for several times.On this basis,using the same sample data,three different kinds of wind power prediction models are established by three conventional prediction algorithms,and then compared with the dynamic ANFIS model in prediction accuracy.The results show that the dynamic ANFIS model has a better prediction effect.Finally,the most excellent prediction model can be provided a reliable data source for the following research on the optimal allocation of the energy storage capacity.3.The paper proposes a hybrid energy storage system consisted of battery and pumped-storage station aiming at solving the problems of wind power fluctuation and electricity rationing.And then a multi-objective optimization allocation strategy is designed for the hybrid energy storage system.An multi-objective optimal allocation model and its constraints are obtained whose objective function is to achieve maximum of economic benefit and minimum of wind power fluctuation for the wind joint storage system.The model can be calculated by genetic algorithm(GA)based on the predicted data got by the dynamic ANFIS model for wind power prediction.The capacities of battery and pumped-storage station are got under the condition of the optimal target.The results show the hybrid energy storage system allocated suitable capacity can improve the stability of wind power and increase the power grid's adaptability for wind power.