基于CEEMDAN和HHT的谐波检测方法

Harmonic detection method based on CEEMDAN and HHT

针对现有谐波检测方法准确度不高的问题,提出一种基于自适应噪声的完全集合经验模态分解(CEEMDAN)和HHT的谐波检测方法。利用CEEMDAN对谐波信号进行分解,通过计算各个固有模态函数(IMF)分量与原始谐波信号的相关性,筛选出真实的谐波分量,再结合HHT变换得到谐波信号的特征参数,包括瞬时幅值、瞬时频率和谐波产生的起止时刻等。对CEEMDAN、EMD、EEMD和FFT 4种谐波检测方法进行仿真,仿真实验表明,CEEMDAN对于常见的暂态谐波信号,可以将平均检测误差控制在1%以下,有效的提高了谐波信号检测的准确度。

To solve the poor accuracy of the existing harmonic detection method,a new method for harmonic detection based on complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)and HHT was proposed.The CEEMDAN was used to decompose the harmonic signals.By calculating the correlation between the local modal function(IMF)components and the original harmonic signals,the real harmonic components were selected,and the characteristic parameters of the harmonic signals were obtained combined with HHT transform,including instantaneous amplitude,instantaneous frequency and the start and end time of the harmonics.The four harmonic detection methods of CEEMDAN,EMD,EEMD and FFT are simulated.The simulation experiments show that CEEMDAN can control the average detection error below 1%for common transient harmonic signals,which effectively improves the accuracy of the harmonic detection.