Generalized Energy Storage Optimal Allocation Based On Controllable Load Dispatchable Capacity Prediction

Although renewable energy has the advantages of clean and pollution-free,the access of high-permeability renewable energy to power grid will lead to many problems,such as the decline of system stability and flexibility,and energy storage system is the key to improve the stability of renewable energy supply and utilization.However,the deployment cost of traditional battery energy storage system is too high and its service life is limited,which makes it difficult to commercialize.Combining traditional battery energy storage system with virtual energy storage system composed of controllable loads on user side to form generalized energy storage system can reduce deployment and use cost and improve economy.In order to optimize the operation strategy of user-side controllable loads,it is necessary to predict the dispatchable capacity of controllable load at first.Therefore,based on LSTM,a method for predicting the dispatchable capacity of controllable loads is proposed.On this basis,a bi-level programming model combining optimization of operation strategy and configuration capacity of generalized energy storage system is constructed.This disertation focuses on the configuration capacity optimization of general energy storage system of renewable energy with high permeability.The main work and innovations are as follows:1.Modeling controllable load and battery energy storage system,and it is used to simulate the change of dispatching capacity of virtual energy storage system composed of controllable loads.2.LSTM-based forecasting model is established.The historical data of dispatchable capacity simulated by the controllable loads model is used as input to the model,which enables the forecasting model to predict the dispatchable capacity of controllable loads in ultra-short term according to historical data and real-time operation of the system.3.A bi-level programming model of generalized energy storage system in distribution network is established: the outer model uses genetic algorithm to solve the power and capacity configuration problem of generalized energy storage system,that is,to solve the capacity and power configuration scheme with the highest economic benefits of generalized energy storage system in distribution network;the inner model is used to solve the operation optimization problem,that is,to optimize the generalized energy storage operation strategy when the predictive model gives the predictive value of the dispatchable capacity of the controllable loads.4.In view of the distribution network system with high permeability renewable energy at IEEE33 nodes,the proposed method is applied to optimize the configuration capacity of the generalized energy storage system.The optimization results show that the proposed method can improve the economy of the system and reduce the initial investment cost of the energy storage system.