摘要
针对水轮机在线监测振摆信号降噪处理等问题,本文提出经验模态分解(EMD)与小波阈值降噪处理方法。该方法首先将振摆信号进行EMD自适应分解,得到一系列本征固有模态函数(IMF),然后结合小波阈值函数对每一个IMF进行阈值降噪处理,最后将经过小波阈值函数降噪处理的IMF进行重构,构成降噪后的信号。并通过仿真信号验证了处理方法的正确性,采用该方法对在线监测系统采集的振摆数据进行降噪处理,结果表明该方法为能有效地对在线监测振摆数据进行降噪。本文提出的EMD与小波阈值降噪方法适用于水轮机振动信号的降噪处理这一领域,该方法将EMD与小波阈值相结合,避免单一小波阈值降噪中分解层数和小波基选择的局限性,相比传统小波阈值降噪方法,有较高的信噪比。
The empirical mode decomposition(EMD) and wavelet threshold de-noising method are proposed to deal with the problem of vibration pendulum signal de-noising for on-line monitoring of hydraulic turbines. Firstly, the vibration signal is decomposed into a series of intrinsic mode functions(IMF) by EMD, and then the wavelet threshold function is used to de-noise each IMF. Finally, the IMF after denoising by wavelet threshold function is reconstructed to form the denoised signals. The reconsitution of the processing method is verified by the simulation signals, and the method is used to denoise the data collected by the online monitoring system, the results show that the method can effectively de-noise the on-line vibration data. The EMD and wavelet threshold de-noising method proposed in this paper is suitable for the field of hydraulic turbine vibration signal de-noising, to avoid the limitation of decomposition level and wavelet base selection in single wavelet threshold de-noising, compared with traditional wavelet threshold de-noising method, it has higher signal-to-noise ratio.
作者
杨峰雄
付向涛
袁翔
YANG Fengxiong;FU Xiangtao;YUAN Xiang(Guazhi Power Plant of Wuling Power Co.,Ltd.,Changsha 410000,Hunan,China)
出处
《电力大数据》
2022年第1期18-25,共8页
Power Systems and Big Data
关键词
水轮机
在线监测
小波阈值
振动
摆度
hydraulic turbine
online monitoring
wavelet threshold
vibration
swing
作者简介
袁翔(1992),男,硕士,助理工程师,主要从事水电站自动装置检修等方面的工作。
