Improved Genetic Algorithms And Its Application In The Optimal Reactive Of Power System

The reactive powero ptimization of a power system can efficientlyminimize the real power losses and improve voltage level of it,so that theyearly running charge will be reduced and quality of power energy can beincreased. Study the problem of reactive power optimization has the greatsignificance in theory and practical application. In this paper, the fullunderstanding of the reactive power optimization of the currentsituation,analyzes the advantage and disadvantage and applicable scope andpresents the researching direction of reactive power voltage optimization.This disseriation adopts genetic algorithm as optimization algorithmand makes some developments to genetic algorithm. The developments canbe divided: The popular hybrid encoding mode method is used in thereactive power optimization. It can conquer error of binary code method fordiscrete variable and resolves solution accuracy dependent on the length ofcoded string. This method has high accuracy and the advantage of globalsearch. Selection operation uses the method of combining fitness value withsmall group competition. Meanwhile, the choosing operation is based on theelitist strategy. It makes the initial poputation not only keeps optimum butalso keeps diversity. In order to get global optimum solution more rapidlyand better, this paper improves the choosing, crossing and variationoperation. Finally, in order to ensure the accuracy of the ultimate result,termination criterion and local judge optimal operation are used to judge theresult. In order to prove the superiority of improved genetic algorithm, theproposed algorithm is evaluated on test function.Reactive power optimization problem is a large-scale nonlinearoptimization problem with a large number of variables and uncertainparameters, the operating variables include continuous and discretevariables， so the optimization beeomes very eomplex. In this paper,theobjective funetion is to minimize the active power loss. Power flow is thebase and tool of reaetive power optimization, whose eonvergence and computation speed is very important to the efficiency of optimization. Thispaper deeply analyses the power flow method of a power system， andchooses fast decouple load flow containing the merits of simplenes, highcompution speed, P-Q memory save and good convergence to compute loadflow.In order to prove the effect of improved genetic algorithm, thisdissertation uses this algorithm to calculating the IEEE-30bus system. Theresults verify that the proposed algorithm advance the speed of calculateand improve theastringency of algorithm.