The Application Of Improved Genetic Algorithms Of The Optimal Design Of Insulation Structure Of Power Tranformers

The aim of this paper is to conclude the accomplishment realized by author recently. With the rapid development of computer technology and the perfect progress of optimal method, it is possible to implement the optimization of power transformer insulation structure.There are four parts in this paper.The first part consists of Chapter 2.The reliability evaluation method of the whole insulation system are studied for the optimal arrangement of insulation parts, the whole region scan method are used to find the weakest path in the insulation system, the test results and the computation results of some models show it is an effective method and has acceptable precision.The second part consists of Chapter 3 and 4.It introduces the principle of genetic algorithms. It sets up the genetic algorithms by means of taking some effective measures on selection operator, crossover operator, mutation operator and some parameters setting. Finally, the genetic algorithms model is verified by its application on engineering project.The third part consists of Chapter 5.The application of improved genetic algorithms is represented by the optimization of shielding ring at the end region of tapped winding of power transformer. The optimal arrangement of insulation barriers angle rings and caps are also discussed to improve the insulation reliability of the whole system. A study presented here is carried out to establish an optimal mathematical model for graded insulation design of the large power transformer. By using genetic algorithm, an optimization model is worked out for the design of the intershielded-continuous winding insulation structures. The results show that the proposed model is effective for practical use.The fourth part consists of Chapter 6. Genetic algorithm has the ability of doing a global searching quickly and stochastically. But it can't make use of enough system output information. It has to do a large redundancy repeat for the result when solving to certain scope. So the efficiency to solve precision results is reduced. Ant algorithm converges on the optimization path through information pheromone accumulation and renewal. It has the ability of parallel processing and global searching. The speed at which the ant algorithm gives the solution is slow, because there is little information pheromone on the path early. The algorithm in this paper is based on the combination of genetic algorithm and ant algorithm. First, it adopts geneticalgorithm to give information pheromone to distribute. Second, it makes use of the ant algorithm to give the precision of the solution. Finally, it develops enough advantage of the two algorithms. The simulation results show that very nice effects are obtained.The Chapter 7 is the conclusion of the whole paper. It gives the summary of main contributions and the work should be continued later.