Research On Optimal Allocation Of Distributed Generation In Active Distribution Network

The raising of power demand and opening of electricity market will make a great development and utilization of high penetration for renewable energy,producing new challenges to the planning,design and operation of distribution network.Distributed generation(DG)widely integrated into distribution network has inspired the evolution of traditional distribution network to active distribution network(ADN)with active control capability.However,the lack of regulating ability in existing ADN limits the distributed generation penetration,and the rapid development of power electronic technology provides an opportunity to solve the problem.Considering the optimization function of power electronic devices,planning of DG for ADN has been studied in this paper.Firstly,influences of DG access to distribution network are detailed and then probability models of intermittent DG are established in the paper.Typical scenarios of DG and load in multi-segment are generated through Latin hypercube sampling and scene reduction technique.Scenodly,work principle and major functions of soft open point have been discussed based on back-to-back voltage source converter,and then the mathematical model is proposed.Thirdly,considering the regulation ability of soft open point,the bi-level planning model for DG is proposed to allocate DG,maximizing the cost-benefit,and optimize system operation,minimizing network losses and voltage deviation.Besides,the modified imperialist competitive algorithm is introduced to improve the convergence speed and the optimization ability.To solve the large-scale mixed integer nonlinear programming problem,a hybrid optimization algorithm combining modified imperialist competitive algorithm and second-order cone programming is proposed by dividing the integer variable and continuous variable and using convex relaxation and linearization in the lower model.Finally,the bi-level model and hybrid algorithm are applied to an active network design.It is verified that the proposed model is capable to get the reasonable planning scheme and the proposed hybrid optimization algorithm presents good convergence and computational efficiency.