Energy Management Strtegy Optimization In Regional Energy Internet

With the development of new energy technologies and the concept of the Internet of everything,there are many energy-generating,energystoring and energy-using devices with different scales and businesses in the energy market,and their energy flow situation is very complicated.Under the background of regional energy Internet,this paper studies the energy management strategy of distributed energy resources by taking the charging station integrated with photovoltaic,energy storage and charging as the research subject.Accurate generation load prediction information provides data information basis for energy management of charging station.In order to predict photovoltaic power generation accurately under different meteorological conditions,a model selection method based on random forest is proposed in this paper.This method firstly uses the correlation between photovoltaic power generation and meteorological factors to perform cluster analysis on the original data,then establishes a model pool and matches the optimal model of each cluster from the model pool,records the corresponding model labels of the original data,and finally trains the random forest with the original data set with model labels to complete the learning task of model selection.In the prediction stage,the adaptive model is generated by the random forest,and the prediction result is obtained from the model.The effectiveness of the proposed method is demonstrated by simulation experiments on the actual photovoltaic dataset.MAE,RMSE and MAPE are reduced by 23.83%,21.71%and 2.49%,respectively,compared with the single prediction model.As the main energy consumption group of charging stations,electric vehicles can bring benefits to charging stations,and their disorderly charging behavior will have an impact on charging stations and power grids.In this paper,an EV charging scheduling strategy based on optimal time allocation is proposed to guide users’ charging behavior.This method comprehensively considers all kinds of time costs of electric vehicles,and introduces queue weight coefficient to evaluate the impact of electric vehicles on the overall time allocation.Monte Carlo simulation experiment proves that this strategy can reduce the travel time and queuing time,improve the travel experience of EV users,and promote regional power balance,which proves the superiority of this method.Finally,based on photovoltaic power generation prediction method and the results of EV charging load demand guided by optimal time allocation strategy,this paper proposes a charging station energy management strategy based on deep deterministic policy gradient for complex energy flow between distribution network,charging station and EV.This strategy aims to improve the operating income and guarantee the energy supply capacity of the charging station.The deep deterministic policy gradient algorithm is adopted to solve the energy transaction volume between the charging station and distribution network,which not only ensures the reliable operation of the charging station,but also improves the economic benefits.The performance of this strategy is verified by the simulation results.In conclusion,this paper optimizes the distributed energy resources management strategy under the background of regional energy Internet.By planning reasonable energy scheduling,this strategy promotes regional electricity consumption,improves the utilization rate of clean energy and brings higher economic benefits.