Decentralized Charge Control Of Electric Vehicle Bace On Alternate Direction Method Of Multiplier

Electric vehicles have got rapid development because of its transportation energy use efficiency, less pollution. A large number of electric vehicles connected to the electricity grid have become a challenge in the future. But the electric vehicles randomly access to distribution network will cause negative effects such as transformer and circuit overload, voltage deviation. Electric vehicle charging immediately is the car owner’s behavior. Optimizing the electric vehicle charging regulation can weaken the negative influence of the distribution network, even more can auxiliary power grid operation. So the electric vehicles orderly charging become particularly important. However, with the development of smart grid, centralized dispatching control of electric car gradually shows great communication bandwidth demand, shortcomings single point of failure. In this background, the research on distributed scheduling control has great meaning.This paper aim the distribution network of household routine EVs charging mode. The nonlinear constraint convex mode contain distribution network. The objective function considering power flow, variable capacity, the electric vehicles’charging demand which making it considering the maximization of the EV owners’benefit with the minimization of the active power loss. And the paper analyzed three kinds of cases the user of the system total pay and distribution network loss. The results showed that the orderly charging control effectively reduce the system network loss, slow load fluctuations, reduce the users pay cost. At the same time the orderly charging control can avoid circuit, transformer overload caused by the disorderly charging which improved economy and safety of distribution network operation.Under the basis of the above model, this paper introduced a new Alternating Direction distributed Alternating Direction Method of Multipliers algorithm. Using the alternate direction method of multiplier (ADMM), the centralized optimization model of charging can be converted into the individual sub-problems in the decentralized optimization model based on the each device. The sub-problems’ objective function and constraints are convex which guarantee the convergence of the algorithm. During each iteration of ADMM, only a few information exchanged between the device and the adjacent interactive information points. As a result, the EV owners’ information security can be protected. Meanwhile, it can overcome some disadvantages by the centralized optimization, such as the high communication requirements and the high computational overhead. The simulations for IEEE33 and 119-bus network systems show that results of the decentralized optimization model and the centralized model are the same. The proposed algorithm shows high computing efficiency.Finally, by considering the users using car uncertainty and the forecasting of distribution network load randomness. This paper studied the real-time rolling control of the EVs. Comparing the before implementing rolling control and after, the results show that the rolling dynamic real-time scheduling can make better results.By implementing scrolling before and after rolling the calculation results of comparison show that the rolling dynamic real-time scheduling can make better results. The proposed algorithm shows low communication cost and good applicability to the rolling scheduling schema in real-time.