Algorithm Study On Network Reconfiguration In City Distribution System

Network reconfiguration is not only important methods in circulate and control of distribution system, but also important part of distribution management system. In theory, network reconfiguration is a complex and many object combination optimization problem. From 1980's, people had extensive study on network reconfiguration of distribution system, so many relatively mature methods and theory of network reconfiguration had formed and developed, but all methods have many shortcoming. The ability of global searching and extensive application in other scopes of genetic algorithm had proved that it is attracting goal for us to study how to apply genetic algorithm to network reconfiguration.The thesis discussed the home and foreign methods of distribution network reconfiguration. In theory , the method of math optimization can assure the global optimization result, but in fact it will result in the combination blast when the dimension increased. The method of on the base of simulated annealing can get the global optimization result, but it have shortcoming that it depend on the parameter and calculation is big. The method of on the base of artificial neural network can get the result in the shortest time and don't calculate the flow, but the precision depended on the stylebook, it is difficult that get the full stylebook and it need train the stylebook with the long time. The fuzzy math and the expert system depended on the development of the other technology. Optimal flow pattern and branch changing can't assure the global optimization result , but it can get the result in short time if it combined with the elicitation method . Genetic algorithm has many specialty in network reconfiguration, it will get more application in network reconfiguration if it can combine with the distribute system peculiarity and improve the convergent speed and astringency. So the thesis applied the genetic algorithm on the research of the network reconfiguration.On the base of the research of the theory of genetic algorithm and the method of genetic operation, improved on the method of genetic operation, put forward one formula of the adaptive crossover and adaptive mutation, proved the genetic algorithm ability of the global optimization result in non- linear combination optimization by the math example.On the base of analysis the specialty of distribute system, put forward memory the data with structure array, it is easiness to identify the structure of network and calculatethe power flow, explained the method of identify the structure of network. Indicated that the method of backward and forward update is good for power flow calculation, explained the method and step of calculation. Adopted the math model of network reconfiguration that is the minimization of loss reduction. On the base of research of theory and algorithm, discussed improved genetic algorithm about distribute network. Initial population don't repeat forming in order to guarantee diversification of population; the thesis put forward that crossover adopted part mapping crossover and mutation adopted gemination mutation, All of improved method can easily content with radialization and connectedness. Put forward a new convergent criterion, so it can avoid immaturity convergence. The thesis also primarily discussed that combine genetic algorithm with fuzzy math .