Methods Of Large Scale Electric Vehicle To Grid Interaction Strategy

With the outstanding performance of energy saving andenvironment protection,electric vehicle(EV)is the main transportation in the future.In uncontrollable condition,the EV owners tend to charge their EVs at the end of their trips,which would have different affects on load characteristic of residential quarter,voltage quality of distribution grid,frequency quality of area power grid.With vehicle to grid(V2G)technology,interactive strategy of EV could improve the load characteristic,volatage quality and frequency quality in different level of power system and reduce the cost of charging for EV owners.The main contents of the thesis are as follows:1)A charging and discharging dispatch strategy considering load characteristic index in residential quarter is established.Considering driving characteristics and battery characteristics,this paper proposed a dispatching strategy taking account of valley-peak difference,daily load factor,mean-square deviation of load and electricity of users.Since intelligent optimization algorithms could not appropriately solve problem with equation constraints,a bilayer discrete particle swarm optimization is proposed.On the first layer,charging and discharging plan for singleelectric vehicle satisfying all constrains is made with discrete particle swarm optimization;on the second layer,basic discrete particle swarm optimization is applied to optimize model of vehicle to home(V2H)dispatching strategy.Daily load and electricity are analyzed in uncoordinated and coordinated charging modes,V2H mode under different responsiveness of users to V2H dispatching strategy.It is concluded that V2H dispatching strategy could reduce valley-peak difference,load fluctuations and improve daily load factor in most extent.Electricity of V2H mode falls more than 1/3 than that of uncoordinated charging mode;and the higher responsiveness of users to V2H dispatching strategy is,the better load characteristics will be,but the fare per vehicle in V2H dispatching will be higher.2)An electric vehicle aggregator pricing strategy considering voltage out-of-limit probability in distribution network.Considering the effect of fluctuation of electric vehicle charging load guided by time-of-use price on voltage quality of distribution system,this paper proposed a pricing strategy for electric vehicle aggregator in residential area.First,the electric vehicle charging scenarios are established by latin hypercube sampling,and the charging start time in each scenario is calculated based on stackelberg game.Thus,the model of dynamic probability load for electric vehicle is established.Besides,with approach of dynamic probability power flow based on cumulant method,dynamic probability distribution of bus voltage amplitude is analyzed to estimate voltage qualification rate expectation.Considering the objective function of maximizing the expectation profit with the constraints of dynamic price fluctuation limit and expected value of voltage qualification rate,pricing strategy for electric vehicle aggregator is devoleped.Then,bilayer continuous particle swarm optimization is applied to solve pricing strategy.which deal with the constraints of dynamic price in bottom optimization and deal with the constraints of voltage qualification rate in top optimization.Finally,IEEE 33 bus system is used in simulation.Numerical result show:compared with uncoordinated charging,dynamic price optimized by proposed pricing strategy can not only improve the dynamic probability characteristics of bus voltage amplitude significantly in different scale of electric vehicles,satisfying the constraints of expected value of voltage qualification rate,but also ensures the steady growth of profit for electric vehicle aggregator with the increasing number of the electric vehicles.3)An electric vehicle control strategy for frequency regulation considering active unbalance in area power grid is established.Firstly,composition of active power unbalance is analyzed.Secondly,based on the probability distribution of load disturbance and renewable energy source in each area,latin hypercube sampling(LHS)is applied to obtain active power unbalance scenarios.Finally,with minimum of the sum of integral time multiply absolute error(ITAE)under the steps combination vector set from active power unbalance scenarios as function,the droop coefficient of V2G control and PI controller of load frequency control(LFC)is optimized in nonlinear systems by particle swarm optimization.the simulation results show the method of co-optimization of Vehicle-to-Grid Control and Load frequency Control(COVC)has the superior dynamic performance on the various real active power unbalance,meanwhile,COVC could guaranteed the demand of holding SOC from EV owners during frequency regulation.