STUDY OF MODELING AND OPTIMAL CONTROL OF PLUG-IN ELECTRIC VEHICLES AND THE INTEGRATION WITH SMART GRID

The rapid development of plug-in electric vehicles (PEV) brings both challenges and opportunities to the power system. This dissertation studies the modeling and optimal control of vehicle-grid integration (VGI) system with the consideration of both grid-side and PEV owners’ benefits. First of all, a phasor mode single-phase microgrid with PEVs and PV farm is modeled to study the cost effective optimal control of PEV charging / discharging, and cope with the power balancing problem. The linear programming (LP) and genetic algorithm (GA) are used to obtain optimized charging schedules. The time-of-use (TOU) price and discharge incentive are considered for cost minimization. Secondly, the impacts of the bi-directional power flow on battery degradation are investigated. An energy-throughput battery degradation model is applied to simulate the daily lifetime degradation. Case studies with different battery capacities, depth of discharge, temperature, and charging regime are investigated. Thirdly, a grid-tied charging system that enables both grid-to-vehicle (G2V) and vehicle-to-grid (V2G) is designed using SimPowerSystems in Matlab/Simulink. The bi-directional AC-DC and DC-DC converters are designed with PWM-driven bridges. A predictive current control (PCC) is designed to control the AC-DC converter. The phase-lock loop (PLL) technology and band-stop filter is applied to reduce the total harmonic distortion (THD) on grid current. In the previous studies, the control algorithms are designed base on phasor mode model and validated via offline simulation. To further investigate the optimal power management algorithm for real-time simulation, a detailed power electronics level VGI system is modeled with RT-Lab RTE-drive blocks. The real-time simulations with a rule-based algorithm are conducted in Hardware-In-The-Loop (HIL) enabled OPAL-RT simulator. To deal with a large-scale of PEVs charging, a piecewise consensus-based distributed control is presented to minimize the charging power loss and maximize the utilization of PEVs for V2G services. The graph theory is applied to represent the communication network among neighboring PEVs. A metropolis stochastic matrix is applied to specify the communication topology. Case studies with ten PEVs and one hundred PEVs are conducted and analyzed respectively. In the last, due to the rapid response of PEV batteries to the power change demand, PEVs are coordinated with a dual-level consensus-based frequency control method to support the primary frequency control (PFC). The upper-level control aims to minimize the frequency deviation of multi-area power system, and the lower-level control aims to minimize the frequency regulation cost and battery degradation cost for individual PEVs. The consensus-algorithm is applied to specify the information updating and exchange among neighboring individuals. The simulation with different load change are conducted and studied