基于Fast-ICA和对称正交化的谐波检测

提出一种基于Fast-ICA和对称正交化的电能质量谐波检测方法。在简要介绍独立成分分析方法的基础上,把多次谐波信号看作多个独立成分的线性组合,采用基于负熵极大独立性准则的Fast-ICA算法实现一个独立成分的估计。再利用对称正交化方法实现多个独立成分的并行估计,分别分离出基波信号、各次谐波信号和噪声信号,以实现对谐波的检测。仿真结果表明,该方法可较准确地检测出各次谐波成分的频率和幅值且对噪声不敏感,与基于渐进正交化的Fast-ICA算法相比,该方法具有更高的检测精度和更快的收敛速度。

Harmonic detection based on FAST-ICA and symmetrical orthogonalization

A method of harmonic detection based on Fast-ICA and symmetrical orthogonalization is proposed. First, this paper introduces independent component analysis method briefly, treats harmonic signal as a linear combination of multiple independent components, and uses the negative entropy great independence standards based Fast-ICA algorithm to implement the estimations of one independent component. Then, the parallel estimation of multiple components is implemented by using symmetrical orthogonalization method, and the base wave, every harmonic signal and noise signal are obtained separately, thus the harmonic detection is achieved. The simulation results show that this method can accurately detect the frequency and amplitude of every harmonic composition and it is not sensitive to noise. Compared with serial orthogonalization based Fast-ICA algorithm, this method has higher degree of detection precision and faster convergence speed.