Research On Optimized Charging Of Electric Vehicles Based On Multi-agent Technology

With the development of economy and people's living standard, the global problems such as energy crisis and climate warming is becoming more and more serious. To reduce the environment pollution made by the traditional fuel vehicles, electric vehicles develop rapidly. With the electric vehicle technology development and national policy support, the number of electric vehicles is bound to maintain rapid developing speed to meet the environmental protection need. The plug-in of large-scale electric vehicles will be a great impact on the load curve. Therefore, the study of electric vehicle charging optimization model has practical significance.This thesis makes a thorough research on user's travel habits and establishes a micro-grid system model with electric vehicles. It proposes a hierarchical dynamic flexible charging control scheme based on multi-agent technology, and make analysis and discussions through intelligent algorithms. The contents of each part of the study are as follow:First of all, it introduces the factors that affect the electric vehicle charging load,including related parameters of electric vehicle battery and user travel characteristics, such as electric vehicle type, charging mode, charging time, daily travel mileage, battery consumption and so on. Based on the basic assumptions, it uses the direct Monte-Carlo method to simulate the use and charging data of single electric vehicle, then the electric vehicle charging daily load curve is obtained after superposition.Secondly, it forecasts the daily load curve based on the PJM's historical daily load data. The EV load curve and the daily load curve are superimposed to analyze the effect caused by large-scale electric vehicles under the disorder charging condition. And the IEEE 30 node model is used to research the node voltage deviation under different penetration rate of electric vehicles. It provides the theoretical basis for the charging optimization scheme.Then, a dynamic flexible optimization scheme for EV charging based on multi-agent technology is proposed. A multi-agent charging control architecture is established. The distribution network is divided into three layers according to the voltage level. Based on the multi-agent technology: The Charging Agents are responsible for the data collection,processing and communication; The Scheduling Agents are responsible for data classification and the charging plan allocation scheme; The Global Agent gives the charging margin for each period based on the improved PSO algorithm. The optimized charging scheme can smooth the load curve of the power grid, realize the peak load shifting and alleviate the dimension disaster phenomenon.Finally,it develops a set of practical electric vehicle charging optimization system using Qt5. This system includes parameter settings,optimization,database-related functions.