采用改进CPSO动态搜索时频原子的电能质量扰动信号去噪方法

A Time-Frequency Atoms Dynamic Search Method for Power Quality Disturbance Signal De-Noising Based on Improved Chaotic Particle Swarm Optimization

针对电能质量扰动信号噪声抑制中的难点问题,即有效滤除噪声的同时又能较好地保留信号奇异点信息,提出一种采用改进混沌粒子群(ICPSO)动态搜索时频原子的电能质量扰动信号去噪方法。首先,构建了与电能质量扰动信号时频特征相匹配的过完备原子库,采用正交匹配追踪(OMP)算法求解信号稀疏模型,同时采用ICPSO算法对时频原子匹配过程做进一步优化。然后,以残差比阈值确定迭代终止次数,利用最佳匹配原子和稀疏系数重构原始信号,实现信号去噪的目的。运用文中介绍方法对6种典型的电能质量扰动信号进行去噪处理,并与形态学滤波和小波阈值去噪2种方法进行对比。仿真结果表明,文中方法在有效去除噪声的同时,能完整地保留突变点信息,去噪结果准确性高。

Regarding the difficulty of noise suppressing in power quality(PQ)disturbance signal,i.e.effectively removing noise and well keeping singular points,a de-noising method based on time-frequency atoms,searched dynamically with improved chaotic particle swarm optimization(ICPSO),is proposed in this paper.An over-complete dictionary is built to match time-frequency features of PQ disturbances,and the orthogonal matching pursuit(OMP)algorithm is selected to solve the signal sparse model.Meanwhile,the ICPSO is utilized for further optimization to accelerate atom matching process.Then residual ratio threshold is chosen as iteration terminating criterion.Since no noise disturbance signal can be rebuild with the best matching atoms and its sparse coefficients while the noise cannot do,the goal of de-noising is achieved. Simulation experiments of six common kinds of PQ disturbance signals are performed,and a comparison with the widely used morphological filter and wavelet threshold de-noising methods is carried out.Simulation results indicate that the proposed method can suppress PQ disturbance noise with high accuracy while keeping the singular points well.