Research On The Model And Algorithms Of Reactive Power Optimization In Power Systems

Reactive power optimization is an important means to ensure the system operated safely and economicly, to improve the system voltage level and to reduce the power consumption. With developing of power industry, expanding of power system, and increasing complexity of network structure, the users put forward higher requirements of the quality of power provided. Therefore, the research of reactive power optimization has theoretical and actual value.This thesis has carried out a new explore and try in the electrical power system calculation and reactive power optimization. For solving linear equations is an important part of power flow calculation, the paper introduces Gaussian Elimination and Jordan Elimination firstly, and makes the relevant summary and comparisons. In addition, various flow calculation methods are summarized systematicly. Improved algorithm is also proposed based on Gauss-Seidel algorithm, and examples demonstrate the improved algorithm has less iteration number.The regulation of the transformer plays an important role in optimizing the system reactive power, and this paper gives non-automatic and automatic adjustment methods of transformer tap. Using non-automatic adjustment method to regulate the taps is simple, so it is used in classical method. The calculation of reactive incremental loss is the base of Classical method. In this paper, the methods of admittance matrix in both polar and rectangular forms are derived, and the form of polar coordinate is given examples. Based on this reserch, it takes a deep study of reactive power optimization in classical method, including the mathematical model, variable constraint and algorithm flow.In this paper, based on the summary of the research various algorithms and some pitfalls of the particle swarm optimization considered, a modifyed particle swarm algorithm has been formed. The reactive power optimization is mathematieally formulated as a nonlinear constrained optimization problem, and its objective function is the minimization of the real loss of power system, besides the constraints are considered by generalizing the objective funetion using penalty terms. Reactive power optimization of power system is a nonlinear optimization question with multiple constraints and variables that contain continuous and diserete variables. The mixed variables are well solved by the rounded discrete variables.At the end of the thesis, the classical method and the modifyed particle swarm algorithm are applied to the IEEE-14 and IEEE-30 bus system, and some of the optimization results are presented. The results demonstrate that classical algorithm for reactive power optimization is feasible and holds certain advantages in the calculation of time. It also shows that the modifyed particle swarm algorithm has better precision and higher efficiency than PSO algorithm, and has excellent value in theory and pratice.