Research On Optimal Allocation Of Energy Storage System Based On Power Prediction Of Renewable Energy Generation

Under the background of "double carbon" target,China is actively developing green and low-carbon new energy power industry,and the rapidly increasing installed capacity of wind and photovoltaic power has put the power system to a severe test.As a flexible regulating resource,the energy storage system has the advantage of good frequency regulation and peak regulation science,which can reduce the power fluctuation of new energy generation.This dissertation takes the optimal configuration of energy storage system based on new energy generation power prediction as the research object,and carries out the work of energy storage battery life estimation,new energy power prediction and scenery storage system research to provide technical support for new energy large-scale grid connection.The main research contents are as follows.(1)Based on the analysis and study of multiple dynamic factors affecting the lifetime of lithium-ion batteries,a nonlinear extended Kalman filter algorithm for battery charge state estimation is established using the Davinan equivalent circuit model.First,the conventional Li-ion battery exchange performance lifetime estimation method and depthof-discharge lifetime estimation method are analyzed and studied,and nonlinear factors such as temperature,battery lifetime and operating point dynamic changes are introduced into the Davinan equivalent circuit model.A nonlinear extended Kalman filter estimation method is proposed to estimate the charge state of a lithium-ion battery in real time under the control scenario of a bidirectional DC-DC converter containing a lithium-ion battery.The results show that the estimation error of the Li-ion battery charge state using the nonlinear expanded Kalman is less than 4%under the test environment of 25℃ with the hybrid power pulse capability characterization,which provides a reference for the Li-ion battery health state and life estimation under the complex environment.(2)To address the problems of large scale and many feature dimensions of new energy generation power prediction data,linear programming and other modules are integrated in the long and short-term memory(LSTM)network algorithm to build an adaptive power prediction model with improved accuracy.Low accuracy and low stability of the LSTM network algorithm in large-scale and multi-dimensional new energy generation scenarios,the principle of integrated learning algorithm is introduced,and the generative adversarial network,support vector machine and linear programming modules are integrated in the scenery power prediction model to build an adaptive LSTM network power prediction model,which improves the accuracy and robustness of power prediction.The proposed model improves 39.27%compared with the traditional LSTM model in the PV simulation scenario and 27.87%compared with the traditional LSTM model in the wind power simulation scenario.It can be seen that the adaptive prediction model improves the wind and light power prediction adaptability and stability problems under different scenarios.(3)On the basis of the improved accuracy of the adaptive power prediction model,the power fluctuation characteristics of the combined wind and storage system are investigated using the AC-DC topology,and the optimal design of the storage configuration under wind and solar complementarity is carried out.Five control strategies for smoothing wind and light power fluctuations are compared and studied,and smooth out the unstable characteristics of new energy power output.The model predictive control algorithm is used to establish a combined AC-DC topology for wind and storage,and to control the fluctuation of new energy grid-connected power within 2%per unit minute.Under the simultaneous consideration of various operating conditions such as different installation types,sampling intervals,and the number of power stations on the energy storage configuration,it is demonstrated that the correlation coefficient and power fluctuation rate of scenic power generation are positively correlated with the energy storage configuration;in addition,when the number of power sources increases,the correlation coefficient of scenic power generation decreases,the complementarity increases,and the capacity of energy storage configuration can be appropriately reduced.This energy storage configuration design study provides a theoretical basis for the optimization of the energy storage configuration of the wind and solar complementary power generation system.(4)Based on the study of the optimal configuration of energy storage in the wind and solar complementary system,a two-layer optimal configuration strategy of electric/thermal hybrid energy storage in the integrated energy system is constructed by taking the carbon emission reduction target as the optimization objective function.The carbon emission factor is characterized as the system operation cost and introduced into the integrated energy system with new energy as the main source,and an integrated energy system model is constructed based on the energy balance of electricity,heat,cooling and gas with the operation constraint of electric/thermal energy storage equipment.The model is controlled optimally in layers,and a two-layer control strategy applicable to electric/thermal hybrid energy storage is established with operating cost as the inner objective function and energy balance as the outer objective function.Simulation measurements of energy enterprise operation data show that the optimal configuration of hybrid energy storage reduces the operating cost of the system by 6.84%without considering the loss caused by lack of load.Meanwhile,it also reflects that the carbon emission of the system is increased to a certain extent when the electric energy storage is profitable,and it is proposed that the optimal configuration of electric/thermal hybrid energy storage is one of the effective ways to effectively reduce carbon emission in the integrated energy system.