Study On Charge/Discharge Optimization Control Satategy Of Electric Vehicles In Distribution Network

The application of Electric vehicles (EVs) will become a critical path of realizing energy saving and emission reduction at home and abroad. When large-scale EVs plug in grid, it can produce effects on the grid that cannot be ignored. However, as the access port, the effects on distribution network is direct, which threat the satety operation of distribution network and deteriorate the power quality of users. At the same time, the well developed V2G (Vehicle to Grid) technology make EVs to be energy storage units that can transport power to grid. This is also an opportunity for distribution grid. To assess the effect on distribution network when large-scale EVs access to the network and study how to coordinate and control the charge/discharge behavior can better respond to these challenges and opportunities, and weaken the influences on the distribution network, even to a certain extent, assist the operation of distribution network. Considering that EVs will universally plug in resident life distribution network in the future, in this paper, we assess the influence on distribution network and study charge/discharge optimization strategy considering that there is an EVs aggregator at the resident life distribution transformer. Partitioning agent will be responsible to guide and monitor the charge/discharge behavior of every EV on its own jurisdiction.First of all, account of the randomness of charging on distribution network level, the charging starting time and charging time are uncertainty, a charging load probability model is established to evaluate the charging demand of the EVs aggregator. Study the influence of the uncontrolled random charging imposed on distribution network has a certain guiding significance for distribution network planning and operation.Then, on the premise of meeting the charging demand of aggregators, a charging optimizaton model considering distribution network frame structure is proposed to achieve the smallest active network loss by optimizing the EVs aggregator’s charging power, which refer to constraints of the node voltage, line power flow, capacity of distribution transform, dynamic climbing constraint of aggregators’charging power and charging energy balance. Compare the effects on system load curve, node voltage and active network loss imposed by uncontrolled random charging and controlled orderly charging respectively, and analyze the situation of system voltage off-limit, branch overload and distribution transformer overload. The result indicate that the controlled orderly charging can reduce the power loss effectively, smooth the fluctuated load, improve the voltage level and invoid the voltage quality declining, branch overload and distribution transformer overload that may be caused by uncontrolled random charging in certain period of time. It increase the economy and safety of the distribution network operation.Based on the above optimization model, consider the EVs can serve as energy storage unit and participate in charge/discharge bi-directional dispatch, a model is destablished with discharging to optimize the charge and discharge power of the aggregators. The results show that the charge/discharge bi-directional optimization dispatch can bring more additional benefits for distribution system with the assurance of driving demand.At last, as the distributed photovoltaic (PV) power play a more and more important role in the residents’life distribution network, we study the influence that PV access location and capacity imposed on the system. Compare the voltage and the aggregator’s optimized charge/discharge results of node15when accessing different capacities of PV. It indicate that the accessing of PV change the original optimized charge/discharge strategy dramatically.The original-dual interior point method is implied to solve those large-scale nonlinear programming problem with many time sections in this article. The simulation results of IEEE-33, PG&E69and an actual119-bus distribution system show that the algorithm has good convergence, high universality and better optimization performance.