The Research On Reactive Power Optimization Based On Improved Genetic Algorithm

With the vigorous development of socialist modernization, electric power system of our country is expanding at an unprecedented scale and speed way. High pressure, long-distance, high-capacity electric power industry begin to occupy the dominant position of the transport, power grids structure and operation mode become more complex, and the system operation constraints are increasingly strengthened. Based on this situation, a lot of new features and requirements are appearing in power system optimization problem. It is very difficult to use the traditional optimization model and the conventional optimization to solve the problem. On the basis of the characteristics of modern power systems and trends, it is an important issue to study the features of modern electric power system technology deeply and improve the modern power system optimization model and algorithm in the current power system research and engineering practice.According to the actual characteristics of modern power and reactive power on power system, a reactive power optimization model which active minimum power loss as the objective function is established; In accordance with the characteristics of reactive power optimization model, an improved genetic algorithm is proposed, and the calculation way of specific adaptive crossover, mutation probability is set out. As a new intelligent optimization algorithm, because of its unique ability of global optimization and robustness, genetic algorithm showed its superiority in the reactive power optimization process. This article prepared a simple genetic algorithm and improved genetic algorithm for reactive power optimization program by the C language in the MATLAB environment. And it combined with reactive power optimization simulation of IEEE14 node and IEEE30 node, two algorithms in the paper are compared and analyzed. The results showed that comparing with the simple genetic algorithm, the improved algorithm had better global convergence and higher computing speed.