Optimal Planning Model And Algorithm Of Distribution Network Of Adapting To The Distributed Generation Access Baesd On DG States And On-line Switch

Distribution network is an important segment for connecting power generation system,transmission system and users effectively. The rationality of the distribution network structure is theconcentrated reflection of the whole power system and there is a direct relationship with the users.Recently, As more and more distributed power merge into the distribution network, a study ofdistribution network planning based on DG has become a difficulty focus and research. At the sametime, distribution network topology is increasingly large and complex. When the switch is on or offwhich caused by network failure, load switching operations and so on, it must make topologychange. As the basis of distribution network analysis, the speed and precision of distributionnetwork must be further improved, so a perfect network topology algorithm is much more needed.This thesis puts forward the way of classifying and layering network structure based on thefeeder model. Using CIM/XML as data source, setting up different ways of data storage withdifferent layers,the computers can save the space for the data and the data can be easily called.What’s more, this thesis uses object-oriented technology to build model of distribution network,classifying and storing all kinds of electrical equipment by different nature.For the sitting and sizing of DG in the existing distribution network, this paper establisheseconomic model of DG based distribution network planning. The model takes the minimization ofthe investment of distributed power cost and maintenance, line loss cost, shadow distributionnetwork investment as objective function, and builds normalized objective function by adding thenode voltage constraints, wire current constraints, distributed power operation constraints that threeinequality constraints through the punishment factor. At the same time, this model considering theother factors such as the distributed generation resource limit. Artificial intelligence algorithm hasdone well in solving all kinds of optimization problem, so this paper use genetic algorithm to settlethe location and constant volume of distributed generation planning problem.This thesis finally analyzes about how to improve the distribution network topology generationrate. Through the combination of global static topology and local dynamic topology, networkstructure can be quickly generated. Thus, the searching for topology after the changing of switchstate can be narrowed, and the speed of topology generation is greatly increased. On this basis, thekey technology such as the flow algorithm of all kinds of distribution network structure werestudied,and Java is used to develop the software of topology drawing in order to realize the goaldistribution automation. Calculation and testing multiple sets of data proves the practicability andfeasibility of the algorithm. Accurate and responsive to switch shift, can completely meet the needsof power distribution network automation.