Research On Robust Optimization Planning Methods In Active Distribution System

Distribution system planning is one of the main tasks of power system planning, it plays a crucial role to guarantee the supply capacity and power quality. However, with large-scale distributed generation (DG) accessing to the distribution system, the traditional distribution system becomes active distribution systems, and bring the distribution system planning a wide range of uncertainties, which greatly increases the difficulties of distribution system planning. Therefore, it is necessary to study active distribution system optimization planning considering uncertainties reasonably.Firstly, this paper presents the present situation and the development of active distribution system planning, introduces several commonly used methods dealing with the uncertainties, summarizes the development process of the robust optimization method and its advantages, and briefly compares the efficiency of mature commercial software CPLEX solving mixed integer linear programming (MILP) model with traditional meta-heuristic algorithms solving normal model without transformation.Secondly, establishes a joint planning model containing the locating and sizing of DG and distribution system network layout. In order to ensure the radial topology when taking account of DG and several transmission nodes without load, the spanning tree constraints are introduced to achieve acyclic network, then nonlinear power flow calculation models are equivalent to cone programming model and the polyhedral approximation models are used to linearize the cone programming model, thus an active distribution system planning model based on MILP is built. And the CPLEX software is used to solve the MILP model, simulation results demonstrate the high efficiency of the proposed model. Meanwhile, an active distribution system planning software is developed by using MATLAB GUI development environment.Then, considering the uncertainty of wind speed in the distributed wind power planning, robust optimization method is proposed to establish a distribution system robust optimization planning models, whicn can be transformed into a standard MILP model with the use of linearization theory and strong duality theory. In addition, the adjustable robust optimization theory is also put forward to make up deficiencies of conservative robust optimization method. And IEEE 33 node test system verify the effectiveness of the proposed model.Finally, taking the uncertainty of load and renewable DG into account, an adjustable robust two-stage minimax cost programming model is established, in order to solve the model, two single optimization model which are master problem and subproblem respectively are established, then the final results of network planning can be obtain through the iteration calculation of two submodels. The model and algorithm proposed are also verified through the simulation of 23 node distribution system.