Study On Reactive Power Optimization And Control Strategy Of Distribution Network With PV Nodes

Reactive power optimization can not only improve the voltage distribution of distribution network,and the voltage quality of power users,but also effectively reduce the loss of electricity in distribution network transmission,thus it can reduce the cost of power investment.Therefore,the study of reactive power optimization of distribution network has a strong practical significance.Power flow calculation is the basis of reactive power optimization in distribution network.The traditional forward generation algorithm is often used to calculate the power flow of the distribution network,but with more and more distributed power sources participating in the grid connection,the distribution network has changed from a single source structure to a multi-source structure.This makes it difficult for the traditional pushback algorithm to deal with PV distributed power supply.Based on the network equivalent reactance matrix at PV nodes,an improved forward backstepping algorithm is designed for power flow calculation in distribution networks with PV nodes.It is necessary to select appropriate reactive power compensation points before the reactive power optimization.In this paper,the reactive power margin of the node and the sensitivity of the active power network loss to the change of the nodal injected reactive power are synthesized by the comprehensive evaluation method,and the reactive power compensation points are selected by the comprehensive index.For the shortcomings of traditional ant colony algorithm in finding optimization speed and falling into local optimal solution easily,this paper improves the traditional ant colony algorithm from four aspects: pheromone updating,heuristic factor updating,probabilistic transfer and shutdown.The improved algorithm is applied to the reactive power optimization of distribution network.In order to adjust the voltage of distribution network effectively according to the result of reactive power optimization,the regional partition model of distribution network is established by using branch cutting method in this paper.The fuzzy clustering method and information entropy method are used to give the time interval method of reactive load forecasting curve,and the sub-area control strategy as well as the time-divided control strategy of reactive power optimization.Combined with the sub-area and time-divided control strategy of reactive power optimization,a two-stage reactive power dispatching scheme is proposed in the process of distribution network operation.This scheme can not only reduce the switching frequency of the control equipment,but also deal with the burstvoltage overrun problem.Taking IEEE33 system as an example,the results show that the proposed algorithm and control strategy are feasible and practical in MATLAB R2014 b environment.