| With the problem of energy shortage and environmental pollution becoming increasingly serious,electric vehicles have attracted more and more attention because of their advantages of energy saving,environmental protection and low noise,which prompt rapid development of electric vehicles.Large-scale electric vehicles will be directly connected to distribution network for energy replenishment,which will bring a significant impact on the safe operation of distribution network.Quantitative evaluation of the hosting capability of distribution network electric vehicles is not only of great significance to ensure the safe and stable operation of distribution network,but also of great significance to guide the development of electric vehicles.In this thesis,the charging characteristics of electric vehicles are analyzed,and distribution network electric vehicles hosting capability is evaluated quantitatively considering the uncertainty of load in distribution network.Besides,this thesis analyzes the influence of charging load on node voltage of distribution network.The main tasks are as follows:Firstly,under the condition of certainty,that is,when the uncertainty of electric vehicles charging load and distribution network load is not taken into account,a hosting capability mathematical model of a single node in distribution network is established with the maximum hosting capability of a single node as the objective function and the node voltage as the constraint condition.Distribution network is divided into three regions,and the maximum hosting capability of regional distribution network is taken as the objective function,and the node voltage is also taken as the constraint condition.Then,by introducing the concepts of delay charge completion margin and cloud model,the disorderly charge model of private electric vehicle is established,and the possibility function of charge delay expressed by cloud model is used to quantitatively analyze whether the electric vehicle is charged or not,thus improving the accuracy of charging load modeling of private electric vehicle.Secondly,considering uncertainty,Cumulants model of electric vehicles charging load is established,and the hosting capability of distribution network electric vehicles is quantitatively analyzed based on Combined Cumulants and Gram-Charlier Expansion.The concept of confidence admission ratio is proposed to quantitatively describe the relationship between charging demand of electric vehicle and the conventional load of distribution network.Thirdly,the relationship between distribution network electric vehicles hosting capability and the possibility of distribution network nodes voltage overrunning the limit of is analyzed and studied.A muti-objective function considering distribution network maximum hosting capability and the minimum penalty for the voltage overrunning limit is constructed,and fast non-dominated sorting genetic algorithm(NSGA-II)is used to analyze and verify the established model.Finally,in order to study distribution network charging station hosting capability,this thesis regard dispersed charging pile as a special charging station,and the concept of simultaneous charging rate of the charging pile is proposed.The charging load models of four charging stations are established,and the capacities of four charging stations which can be connected to distribution network are analyzed and studied. |
PDF Full Text Request