Research On Distribution Network Reconfiguration With DGs And EVs

Distribution network,the last link of the power system from generating to sending to people,has a direct influence on the reliability and quality of power supply.As an important measure to optimize the power flow,the distribution network reconfiguration can make system in safe,stable,and economical running by flexibly adjusting the network topology without increasing the additional construction cost.In recent years,with the social economy advancing and the energy demand growing,distributed generation(DG)technology and electric vehicle(EV)booming have brought opportunities and challenges to distribution network.The traditional distribution network reconfiguration strategy is increasingly difficult to meet the operational requirements of an active distribution network.Therefore,it's significant to study the reconfiguration of a distribution network with DGs and EVs.Focusing on the theme of "Considering the reconfiguration strategy of DGs and EVs access to distribution network",the specific work is as follows:Firstly,the coding strategy based on the minimum loop and the topological constraint judgment process based on the connectivity matrix are proposed,according to the characteristics that the distribution network branches are numerous and radial.Minimum loop coding can reduce the occurrence of unreasonable individuals and greatly increase the availability of newly generated topologies in the optimization process.The use of a connectivity matrix to determine topological constraints can simplify the decision process and improve computational efficiency.Secondly,stochastic power flow algorithm based on cumulants and improved models are presented to be used as an important tool to evaluate the randomness of DGs and EVs.This paper first uses the Gram-Charlier series expansion to calculate the probability distribution of state variables,which avoids massive convolution computations.But Gram-Charlier expansion will go wrong in case of a large number of DGs in distribution network.Therefore,it presents an improved discrete-normal stacking algorithm against distorting when dealing with high-permeability DG distribution network power flow.Finally,the correlation is introduced into the stochastic power flow to obtain solutions and summarize the impacts on the system.Thirdly,the single-objective and multi-objective static reconfiguration models of distribution network are established and optimized separately.It first uses the improved gravitation search algorithm(GSA)to solve single-objective reconfiguration with the goal of minimum network loss and voltage insecurity probability,and comprehensively optimizes the network topology and the DGs capacity to explore the different effects of results.Then,non-dominated sorting,crowding distance calculation and elitism retention strategies applied to the gravitational search algorithm is used to obtain Pareto solution set for multi-objective optimization of the minimum network loss,minimum system voltage insecurity rate,and minimum DG input.The optimal solution set provides decision support for the distribution network optimal operation.Fourthly,a dynamic reconfiguration strategy based on the minimum period of information entropy is proposed to solve the disadvantages of information loss brought by probability statistics of DGs and EVs.In the dynamic reconfiguration,the daily load data segmentation will bring about the loss of information while significantly reducing workload.And this segmentation method with the smallest entropy decrease can gain a balance and obtain the optimal section case.The example results show that the dynamic reconfiguration can reduce the more daily load loss than the static reconstruction.Meanwhile,the reconfiguration strategy based on minimum information entropy decrease segmentation has advantages on simplifying the computaion,reducing the system operation loss and improving the voltage level.