Researches On Optimal Site Selection And Capacity Determination Of Distributed Generation Considering The Correlation

The current electric power construction is developing rapidly in the direction of the smart grid.The free access of distributed generation(DG)is the basic requirement of smart grid,DG can make power apply form efficient and varied,and it can also conserve energy and protect the environment.But DG can also completely change the power flow,the network loss at the same time,and the majority of DG(like the wind power generator,photovoltaic cell,etc.)has the nature of the uncertainty output random.If a number of different types of DG units access to multiple position,it can cause amazing difference on the network loss,planning,cost and power supply reliability of distribution network directly,and the impact is difficult to predict.So this article does related research about kinds of distributed generation optimal site selection and capacity determination and its influence on the distribution network.Because of the intermittent and randomness of distributed generation,the objective function or constraint condition is difficult to determine in optimal site selection and capacity determination,usually to solve the problem by chance constrained programming method within a certain confidence interval.With many intermittent and randomness output devices(such as distributed generation and electric vehicle charging station)access to the distribution network,it may cause the number of discrete chance constrained model scale rise as exponential form,using the multiplication principle or random sampling can hardly obtain satisfactory treatment effect.This article chooses two typical random output devices blower and photovoltaic cells as representation.According to the output density curve,making probability statistics for period of time,and forming correlation relationship in a specific time period in the light intensity,wind and general load.Getting reference correlation coefficient matrix according to the history of planned standard samples to make sure that they have approximate relationship by using Nataf transform.In other period,using multiplication principle to combine the discrete model.With a multi-object function including comprehensive consideration of the whole life cycle cost management in investment cost index,network loss index and voltage deviation index and giving objective weight coefficient by weight matrix as far as possible using 9/9-9/1 scale method and illustrating the rationality by consistency inspection.Solving the multi-objective programming to a single target in IEEE33 node example by bat algorithm to achieve the quickness and accuracy.Comparing the optimal decision of simulation with several representative solution in the Pareto solution concentration to illustrate the significance of simulation decision-making.At the same time,the article also makes a detailed instruction in the artificial intelligence algorithm performance comparison,the sensitivity change of chance constrained programming and the siting selection result effect in positive and negative correlation between DG resources and distribution network load.Taking above-described distributed generator into simulation calculation can show that considering the correlation of distributed generator optimal site selection and capacity determination can not only greatly reduce the number of discrete model,but also affect the investment cost,network loss and voltage deviation index to some extent.It can verify that the factors considering the correlation between DG resources and distribution network load could affect planning result,which has guiding significance for engineering application.