基于WTSD-EMDLPF的低频振荡信号降噪

De-noising of Low-frequency Oscillation Signals Based on WTSD-EMDLPF

针对电力系统低频振荡信号监测易受环境、设备等噪声污染问题,提出联合小波阈值软降噪和经验模态分解低通滤波的降噪方法,称之为WTSD-EMDLPF。首先,通过小波阈值软降噪滤除大部分白噪声,再利用经验模态分解低通滤波去除经验模态分解的首个固有模态分量以进一步减弱残余噪声,最后合成余下固有模态分量,重构低频振荡信号。运用数值信号算例、IEEE四机两区域系统产生的有功功率低频振荡信号算例仿真以及北美电网实测数据的降噪实验分析,证明WTSD-EMDLPF对低频振荡信号降噪的有效性。

Aiming at the problem that the low-frequency oscillation signals of the power system is vulnerable to noise pollution such as environment and equipment noises,the method of WTSD-EMDLPF combining with wavelet threshold soft de-noising(WTSD)and empirical mode decomposition low pass filter(EMDLPF)was proposed.Firstly,majority of white noise was filtered by WTSD.Then,the first intrinsic mode function(IMF)of empirical mode decomposition(EMD)was removed by EMDLPF to further reduce the residual noise.Finally,the remaining intrinsic mode components were synthesized to reconstruct the low-frequency oscillation signals.The simulation experiments were carried out utilizing the cases of the numerical signals,the active power low-frequency oscillation signals generated by a 4-machine and 2-area system and the actual oscillation data of North America power grid.The results show that the WTSD-EMDLPF has the effectiveness in noise reduction for low-frequency oscillation signals.