Research On Reactive Power Optimization With Classical Algorithm And Genetic Algorithm In Power System

With the rapid development of our country economy, the power grid generally has reactive power to be insufficient, the voltage quality not high and power loss is large and other issues. These issues will have a greater impact for safe and economic operation of the power system. The reactive power optimization is an important means to ensure the system voltage quality and reduce power loss, so the study on reactive power optimization have important theories and practical application values.Reactive power optimization in power system requires continuous use power flow to analyze the reactive power output of generator and load node voltage levels, in order to choose a fast and reliable power flow method is the basis to carry on the reactive power optimization, therefore power flow calculation algorithm is also the research key point in this paper. The paper derive an improved DC power flow method, the method can calculate both the voltage phase angle and voltage amplitude, and through comparison with the Newton-Raphson method found that this method can get a better calculation accuracy, thus makes this DC power flow method be possible to apply in more scopes. The paper also derive a rapid decomposition method based on rectangular coordinate, through using the test system to compare with rapid decomposition, although an increase in the number of iterations, but it has a little advantage in calculation speed. Finally, by using a standard test system to test the Gauss-Seidel method, Newton-Raphson method and rapid decomposition, the example results shows that the fast decomposition method is more suitable as the power flow model of Reactive power optimization based on the classical method and genetic algorithm in this paper.This paper has carefully study the application of the classical method in reactive power optimization, The core of classical method reactive power optimization is the calculation of incremental transmission losses, The method is simple and easy, optimize quickly and other characteristics. The paper also emphasis introduced the application of genetic algorithm, this algorithm suitable for solving the complex nonlinear optimization problem like reactive power optimization, carefully study the encoding method and genetic operators, because of the basic genetic algorithm has slower convergence and easy to fall into local solutions, therefore this paper has made some improvements in the basic genetic algorithm, such as choice strategy, the crossover, the crossover ratio and variation ratio, in order to make this improved genetic algorithm could better apply in the reactive optimization.By testing IEEE14and30node data to verify that the effectiveness of classic method, basic genetic algorithm and improve genetic algorithm, through comparison and analysis of the optimization results, The classical method can apply in reactive power optimization, can effectively reduce the system power loss, although on the decline less than basic genetic algorithm and improved genetic algorithm, but the classical method has the superiority in the calculating speed very much. The improved genetic algorithm has an advantage in the calculation speed than basic genetic algorithm in this paper, can more effectively reduce the system active power loss, shows that the improved genetic algorithm has faster computing speed and better optimization ability.