The Study Of Multi-Objective Optimization On Passive Power Filter Parameter Design

Wide-range application of nonlinear load has brought serious harmonic pollution to power grid, as well as a series of power quality problems. However, some sensitive load has tougher and tougher requirement on power quality. Therefore, Passive power filter (PPF) is most widely used as harmonic-eliminating device in power system, because of simple configuration, low investment, high reliability, and low operation cost. This paper penetrates on a profound and detailed study on the parameter design of PPF.First, this paper analyses the configuration and working principle of PPF. Moreover, a thorough analysis is conducted on the factors which are related to the filter's characteristics, including the effects of equivalent frequency offset, quality factor, reactive power capacity of the single tuned filter (STF) and high pass filter (HPF), the system equivalent harmonic impedance etc. As the foundation of designing filter, this step would supply the design principles for the coming engineering design methods and multi-objective optimization.What is more, this paper analyzes the decision strategy of setting the filter group, and deduces the theory of engineering methods used to design the filter parameters, such as minimal capacitor installed capacity method, reactive power allocating method and over-voltage limitation method. All these methods are realized by programming under MATLAB. The parameters by the engineering methods can be used as the initial value for the coming multi-objective optimization of filter parameters.Next, it is proposed a randomized coefficient genetic algorithm (GA) to solve the PPF parameters multi-objective optimization. The multi-objective of the optimization is set by considering the filtering efficient, invest index, and reactive power compensation. Based on the GA toolbox in MATLAB, the filter parameter optimization was realized through programming. What is new in this paper is that a new index, named"filtering efficient"is proposed to measure the harmonic-eliminating effect. Through the comparison to the traditional harmonic-eliminating method, and simulation, the feasibility of the"filtering efficient"served as a measuring index is to be proved.