单相三电平逆变器谐波检测的小波神经网络算法研究  

Wavelet Neural Network for Single Phase Three Level Inverter Harmonic Detection

作　　者：马辉[1] 王玉华[2] 刘朋朋[1] 贾文超[1] 

机构地区：[1]长春工业大学,吉林长春130012 [2]台州学院,浙江台州318000

出　　处：《变频技术应用》2013年第6期65-69,共5页INVERTER TECHNOLOGIES AND APPLICATIONS

摘　　要：本文采用小波神经网络（WaveletNeuralNetwork，WNN）算法，对时变谐波信号进行检测。利用Harr小波对谐波信号的幅值和相角进行逼近；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速度；给出网络参数的选定方案；确定网络的训练算法。在MATLAB／SIMuLINK环境下对该算法进行仿真，与传统的小波算法比较，该算法不仅可行，而且精确度得到提高。The time-varying harmonic signal was detected by wavelet neural network algorithm. The use of Harr wavelet approximation time-varying amplitude and phase angle; Wavelet on the signal adaptive time-frequency characteristics of neural network is introduced to improve segmentation, neural network approximation and convergence performance; Given the network parameters of the selected scheme; determining the network training algorithm, proving the feasibility and accuracy in the MATLAB / SIMULINK environment simulation.

关 键 词：小波神经网络 时变谐波检测 Harr小波变换 

分 类 号：TM93[电气工程—电力电子与电力传动]