Control Strategy Design Of Electrified Railway Energy Storage System Based On Traction Load Forecast

Railway is a major infrastructure in China,and China has become the first country in electrified railway operation.The peak traction load power not only causes technically negative-sequence-based power quality problems,but also directly affects user benefits economically.Focusing on the problem of excessive peak power of traction load,this article introduces super capacitor energy storage devices in the electrified railway system to achieve peak-shaving and valley-filling of traction loads,so as to reduce the maximum demand,reduce the cost of electricity,increase the capacity utilization of traction transformers,and improve the economics of the railway sector.Based on a large number of related papers,this paper conducts an in-depth study on the peak-shaving and valley-filling control strategy of energy storage devices,and proposes an optimal control strategy for energy storage systems based on traction load prediction,which aims to improve the effect of traction load peak-shaving and valley-filling.The main research contents and innovations of this article are as follows:The optimal control strategy of energy storage peak shaving and valley filling is proposed on the basis of traction load forecasting.Therefore,the traction load forecasting technology is studied first.In order to overcome the shortcomings of the existing traction load forecasting methods,on the basis of in-depth analysis of the traction load characteristics,this paper combines the idea of deep learning to establish a traction load forecasting model based on LSTM.First,the principle of the LSTM algorithm and the analysis of its applicability for traction load forecasting modeling are studied,and then the traction load forecasting model framework and forecasting flowchart based on the LSTM algorithm are proposed.Finally,the effectiveness of the proposed forecasting model is verified by experiments.In order to further improve the accuracy of traction load prediction and improve the training efficiency of the prediction model,a traction load prediction model based on the CNN-TCN joint model was established.First,the principle of CNN algorithm and TCN algorithm and the analysis of their applicability for traction load prediction modeling are studied.According to the network structure analysis of CNN and TCN,CNN is used for traction load local feature extraction,and TCN is used for traction load prediction,and then a traction load prediction model based on CNN-TCN is built.Finally,based on experiments,it is verified that the model has better performance than the traction load prediction model based on LSTM,with higher prediction accuracy and less training time.On the basis of traction load forecasting,in view of the shortcomings of the current electrified railway energy storage peak shaving control strategy,coordinated consideration of the traction load peak shaving and valley filling effect and energy storage SOC operation status,and proposed an energy storage system control strategy based on traction load forecasting.The strategy is based on the traction load forecast data,divides the SOC of the energy storage device by introducing new decision variables,and determines the charge and discharge model of the energy storage in combination with the traction load power.Then an optimization model is established with the goal of the best traction load peak shaving effect,and the dynamic adaptive particle swarm optimization algorithm is used to obtain the current charging and discharging power of the energy storage peak shaving and valley filling.Finally,the validity and superiority of the scheme proposed in this paper are verified by establishing an optimization model in the MATLAB platform.