摘要
本文采用小波神经网络(Wavelet Neural Network,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.Using 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;By giving the network parameters of the selected scheme,and determining the network training algorithm,the feasibility and accuracy in the MATLAB/SIMULINK environment simulation are proved.
出处
《节能技术》
CAS
2012年第6期516-520,共5页
Energy Conservation Technology
