Studies On Hybrid Intelligent Algorithms For Reactive Power Optimization

The topology of power network is more and more complex with the development of power industry and the enlarging size of electric power system, the customer also need higher power quality. As the power market is developed, it is a task that adopting effective measure to reduce power loss and improve voltage quality, which relates the economy benefit of power enterprise.Optimal reactive power dispatch (ORPD) of power system can reduce power loss and guarantee voltage quality effectively, it also can prevent accident of power system or avoid spread of accident, so it can improve the economy, safety and stability of power system. The importance of ORPD has received extensive attention.The ORPD of power system defines that the reactive power is adjusted in order to minimize the real loss of power system and guarantee voltage quality by setting generator bus voltages, shunt capacitors and inductors and transformer taps. It is also called the optimal control of voltage and reactive power.This paper studies thoroughly the method of ORPD, then, proposes two different hybrid intelligent algorithms on the basis of the summary of the case of algorithm research. They are based two different hybrid strategies which consider the characteristic of particle swarm optimization (PSO) algorithm and differential evolution (DE) algorithm and introduce the idea of DE algorithm to PSO algorithm. In the first hybrid intelligent algorithm, each particle of PSO algorithm keeps track of the different third value besides the best previous position found so far by itself and the best previous position among all particles. The third value is created by the mutation operator of DE algorithm. In the second hybrid intelligent algorithm, the velocity is reinitialized and that dimension of the personal best position is mutated by mutation operator of DE algorithm when some dimension of one particle's velocity of PSO algorithm is smaller than the specified value. The two hybrid intelligent algorithms which combine the excellence of PSO algorithm with DE algorithm have better search ability. In this paper, the ORPD of power system is mathematically formulated as a nonlinear constrained optimization problem, the objective is the minimization of the real loss of power system, the constraints are considered by generalizing the objective function using penalty terms. The ORPD of power system is a nonlinear optimization question with multiple constraints and variables that contain continuous and discrete variables. The mixed variables are well solved by mapped code and the rounded discrete variables. The two hybrid intelligent algorithms based two different hybrid strategies are applied to the ORPD of IEEE30 nodes system. The results demonstrate that hybrid algorithms proposed in this paper are effective and possess the excellent value in theory and practice. They have better precision, stability and higher efficiency than PSO algorithm.