Research On Reactive Power Optimization Based On Improved Genetic Algorithm

Power system affects national people’s normal life, industry and the protection of the normal production. All aspects of life in society have a very close relationship with power system. The safe operation of power system, maximize profits has a very close relationship all kinds of algorithm. The reactive power optimization is a very important method to ensure the voltage quality and decrease network loss. Reactive power optimization is a complicated problem, because most of the power system data is nonlinear. Therefore, the reactive power optimization is one of the focuses of the researchers. This paper studied the research situation of reactive power optimization.On the basis of an in-depth study the characteristics of electric power system and reactive power optimization intelligent algorithm. Reactive power optimization is for the purpose of minimum network loss, and ensure that the whole system safe and stable operation. Previous reactive power optimization algorithm is mainly based on mathematical model. The model has high requirements of the mathematical model and the accuracy of the optimization function. It is easy to fall into local optimal and not well deal with discrete variables. So many researchers began to optimize traditional optimization method.In this paper, after studying all kinds of intelligent optimization algorithms in the application of reactive power optimization, carefully analyses the characteristics of the traditional genetic algorithm for the operator and coding, and modified, the convergence criterion of make it has higher practicability. The hybrid coding is introduced in this paper, and replaced the previous single encoding, solve the length of the encoded string affect the optimal solution of the problem, but also avoid the continuous variables are dispersed when the value of the error, the search space is bigger, can be more accurate to search the optimal value. Second method in this paper, the previous set of fitness function is optimized, the improved method is based on the speed and precision of algorithm in different stages of optimization the fitness function of different Settings section, which ensures that less calculation time, high calculation precision.With optimized operation this paper, according to the different stage optimization method, the fitness and adopt the hybrid selection operation. Crossover operator in this paper has a lot to do with the number of iterations, so will be linked to the crossover probability and the number of iterations, increase the searching ability of the algorithm. In addition this article is also set to gradually increase the mutation operator, the algorithm is stable. Finally, in order to avoid the computation time is affected by the initial booking the maximum number of iterations, this article set the maximum number of iterations and the best individual optimum combination of judging whether the convergence and the optimal value.This article at the end of the program by using MATLAB genetic algorithm, combined with the current algorithm on IEEE14 and IEEE30 nodes example calculation, the application of the basic algorithm and improved algorithm for comparing the calculation results, verify the improved algorithm, the results show that the improved algorithm has great advantage and better convergence properties, can solve the problem of a lot of mixed modalities, stronger practicability.