一种小波网络逆变换的谐波检测算法研究

提出一种小波网络逆变换,在小波对原始信号进行高频、低频段分解时,分别用小波系数模极大值进行奇数和偶数抽取,得到信号的特征向量。再利用BP神经网络对抽取后的信号逼近训练,得到新的小波系数,然后用逆变换进行信号重构。通过比较重构和原始信号。得到谐波值,再反相注入电网进行补偿。实验结果表明,该算法可以满足电力系统谐波检测的要求。

A harmonic detection algorithm study based on wavelet network inverse-transform

In this paper,a wavelet network inverse-transform is designed.Odd and even extraction of high-frequency and low-frequency band signal are extracted the signal by wavelet decomposition.It trains and approaches wavelet signal by BP neural network.Then it reconstructs signal by wavelet inverse-transform.We can get harmonic wave by subtracting the reconstructed signal from the original,which are inverted into the power grid to compensate.Experiments show that the algorithm can satisfy the requirements of the system harmonic detection.