| The simultaneous access of a high proportion of renewable energy and large-scale electric vehicles to the power grid poses new challenges to the economic operation,safety and stability of the distribution network.With the development of vehicle to grid(V2G)technology,electric vehicles have the dual characteristics of load and storage.They can participate in the supply and demand balance of active distribution network,respond to the demand of active distribution network in real time on the demand side,restrain the load fluctuation of active distribution network and absorb intermittent energy power economically and efficiently.Therefore,considering V2G mode to optimize the operation of active distribution network is of great significance.First,considering the power uncertainty of distributed generation,a time series modeling method based on state number decision model is proposed.Considering the membership error of state division caused by noise aliasing,the moving average filtering method is used to filter the original sequence.Aiming at the low calculation efficiency and accuracy of Markov chain Monte Carlo(MCMC)simulation method.The Metropolis-Hastings algorithm is used to generate the state sequence,and two evaluation indexes are used to construct the state number decision model to obtain the optimal success rate.The example simulation shows that this method can generate the power sequence with similar characteristics to the original sequence,which provides the data basis for the optimal operation of active distribution network.Second,in order to structure an accurate model of charging load of electric vehicles,a modeling method of charging load considering trip path decision is proposed.Aiming at the low precision of random sampling in Monte Carlo simulation,Latin hypercube sampling is used to extract the initial trip time and initial charge state,considering the uncertain charging demand of users.The fuzzy comprehensive evaluation method is used to establish the charging behavior model with three evaluation indexes.Considering the impact of traffic conditions on path planning,the dynamic Floyd algorithm is proposed to determine the trip path,and the temporal and spatial distribution of electric vehicles charging load is obtained by Monte Carlo simulation.The example simulation shows that this method can dynamically simulate the actual trip of users according to the real-time road conditions,has advantages in path planning,the impact of passenger capacity and model accuracy,and can provide model for the regulation of electric vehicles under the coexistence of disorder and order in the active distribution network.Finally,aiming at the voltage fluctuation of large-scale electric vehicles and distributed generation connected to distribution network,a double-layer optimization operation model of active distribution network considering V2G mode is established.The upper optimization model schedules the electric vehicle agent(EVA),and optimizes the charge and discharge power of each EVA as the input of the lower optimization model.The lower level optimization model optimizes each voltage regulation mode,and proposes an adaptive differential evolution and biogeography-based optimization(SaDEBBO)algorithm.The simulation analysis shows that under different control strategies,the coordination and interaction between V2G mode and various voltage regulation modes have significant advantages in reducing the operation cost of each EVA,stabilizing load fluctuation and the safe and economic operation of active distribution network.Compared with other optimization algorithms,SaDEBBO algorithm has better solution and better convergence. |
