电压波动估计的混沌简化粒子群算法

Chaotic Simple Particle Swarm Optimization Algorithm for Voltage Fluctuation Measurement

针对粒子群算法(particle swarm optimization,PSO)估计电压波动频率收敛速度慢,采样频率和窗口长度选择不当算法易陷入局部最优的问题,采用简化进化公式和混沌理论两项策略加以改进,提出基于混沌简化粒子群(chaotic simple particle swarm optimization,CSPSO)算法的电压波动幅值和频率估计新方法。去掉粒子群算法进化公式的速度项,仅由粒子位置控制进化过程,避免由于速度项引起的粒子发散而导致算法收敛速度慢的问题;并引入混沌理论,利用其随机性、遍历性和规律性提高算法的全局寻优能力,避免简化粒子群算法陷入局部最优。仿真结果验证了该方法的有效性和准确性。

According to the drawbacks of voltage fluctuation measurement based on particle swarm optimization （PSO）, such as slow convergence and getting wrong results for an improper selection of the sampling frequency and window size, a new method was proposed to estimate the amplitude and frequency of voltage fluctuation based on chaotic simple particle swarm optimization（CSPSO）. The method discarded the particle velocity of PSO to avoid slow convergence caused by velocity item and introduced chaos theory using its ergodicity, stochastic property and regularity to solve the problem of local optimum in simple particle swarm optimization. The simulation results show that the proposed method is effective and accurate.