基于小波神经网络的时变谐波信号检测

电力系统中存在大量的由于非线性器件对电网电压电流整流、逆变而产生的时变谐波。该文提出使用小波神经网络(wavelet neural network，WNN)算法对这一类谐波进行检测。利用小波变换的时频聚焦特性可得出信号的时变信息；将小波对信号的自适应时频分割特性引入神经网络，提高神经网络的逼近和收敛性能；给出网络参数的选定方案；确定网络的训练算法；分析算法的时效性；并与其它检测方法做出比较。经仿真试验表明，该文所述的方法提高了检测的精度和效率。

A Method of Time-varying Harmonic Detection Based on the Wavelet Neural Network

There are many time-varying harmonics caused by the non-linear instruments that convert or invert the current or voltage of the power system. A novel method based on wavelet neural network (WNN) is proposed. Using the time-frequency focus characteristic, the information of time-varying of harmonics can be gotten. The approach and convergence capability are improved by the wavelet transform characteristic of auto-adapt dividing time-frequency plane. And the precept of the WNN characters selection is given, and the algorithm to train the network is proposed. The efficiency of detection harmonics by WNN has been analyzed, and the performance of different methods is compared. The simulation result illustrates the effectiveness of the presented methods.