基于改进的软阈值小波包网络的电力信号消噪

Power signal denoising based on improved soft threshold wavelet packet network

为了克服传统软、硬阈值在信号消噪中存在的不足,首先构造了一种新的软阈值函数,并将小波包变换和神经网络相结合;然后,基于改进的软阈值小波包网络对实际采集电力信号进行消噪。MATLAB仿真结果表明:改进的软阈值小波包网络比传统硬、软阈值小波包消噪的信噪比高,且均方根误差小,是一种非常有效的电力信号消噪方法。

In order to overcome the traditional soft and hard threshold in signal denoising presence, a new soft threshold function is constructed, and wavelet packet transform with neural network is combined which to eliminate the actual electric power signal noise. The MATLAB simulation shows that the improved soft threshold wavelet packet network is more effective than the traditional hard and soft threshold wavelet packet denoising method, very effective in noise elimination.