Research On State Estimation And Optimal Management Methods For The Energy Storage System In Micro-grid

As the quantity and quality of global energy demand continue to increase,the energy shortage and environmental pollution problem are becoming more and more serious.Moreover,the problem of low intelligence of centralized power generation systems is becoming more and more prominent.To address these issues,a micro-grid system consisting of renewable energy systems,energy storage systems,loads,energy conversion devices,monitoring devices and protection devices is proposed.As an important part of the micro-grid system,energy storage systems play an important role in promoting the full consumption of the renewable energy,enhancing the system inertia of the micro-grid,improving the power quality of the micro-grid and increasing the economic benefits of the micro-grid.Due to the consideration of the impact of complex dynamic characteristics and randomness of the energy storage system on micro-grid energy management,this thesis focuses on the dynamic characteristics modeling,state estimation of the energy storage batteries,as well as the optimal power scheduling of the micro-grid.Taking into account the complex dynamic characteristics of the energy storage system,a state estimation method based on particle filter and a remaining dischargeable time prediction method based on open-loop algorithm have been proposed to solve battery mileage anxiety problem.Moreover,a Markov chain Monte Carlo-based energy storage stochastic modeling method has been designed,and a global online optimization framework based on model predictive control technology has been constructed to realize energy management of the residential micro-grid by considering the randomness of the electric vehicle.The main works and contributions of this thesis can be summarized as follows:1)Aiming at the modeling and state estimation of the energy storage batteries,Characterization effects of nonlinear characteristics of the state-of-charge,such as recovery effect,based on two different state-of-charge definitions has been discussed and analyzed firstly.Then,a battery discrete state space model with open circuit voltage as the state variable has been established,and an online open circuit voltage estimation algorithm based on particle filter has been designed.Considering the influence of the battery polarization effect on the state-of-charge,the state-of-charge estimation that minimizes the polarization effect has been achieved based on the online estimation of open circuit voltage and the open-circuit voltage and state-of-charge relationship curve.Experiments under different conditions have verified the accuracy of the proposed battery model and algorithms.2)Aiming at the problem for predicting the remaining discharge time of the battery,a definition of the remaining dischargeable time based on the terminal voltage threshold has been given first.Then,an open-loop based battery terminal voltage curve prediction method has been constructed by relying on the model,the state-of-charge estimation and the prediction of the future operating conditions of the battery.Finally,the prediction of the remaining dischargeable time can be obtained by the terminal voltage prediction curve.Experiments under different current and temperature conditions have been performed to verify the effectiveness of the proposed battery prediction algorithm.3)Aiming at the economically optimal energy scheduling problem for the micro-grid with random energy storage subsystem,a Markov chain Monte Carlo-based parameter identification method has been proposed firstly to realize the establishment of the probability distribution function for stochastic energy storage systems.Then,the optimization goal and constraint of the optimization problem have been designed according to the nonlinear characteristics of the energy storage systems and the probability distribution function of the electric vehicle,so that the micro-grid energy scheduling problem can be transformed into an optimization problem.Based on the above research,a global online optimization framework has been built based on model predictive control technology.The proposed optimization framework can fully consider the energy storage property of the stochastic energy storage under the premise of satisfying the demand of stochastic energy storage energy,so as to improve the economic benefits of the micro-grid while ensuring the normal use of the stochastic energy storage.