光纤旋转地震记录中的谐波噪声压制

Harmonic noises removal in rotational optical fiber seismic record

随着旋转地震仪的日趋成熟,天然地震旋转分量信号的观测工作进入大规模的试验阶段.我们使用一款基于光纤陀螺的三分量旋转地震仪在福建泉州台进行了为期28天的天然地震观测.室内处理数据时发现旋转信号中存在大量能量很强且频率成一定规律的谐波噪声,本文针对此类具有明显特征的谐波噪声分别采用了低通滤波、时频滤波、奇异值分解以及单频滤波4种不同的方法对其进行压制试验,从而更好地识别有效地震事件.经结果对比表明奇异值分解法对此类谐波的去噪效果较为理想.分析去噪处理后的数据发现谐波掩盖下存在一些有效的旋转观测信号,经偏振分析应为近震源的人工干扰.

With the development of rotational seismometers,the observation of earthquake rotational motions entered into large-scale experimental stage.We used a three-component rotational seismometer based on fiber optic gyroscope to observe natural earthquakes for 28 days at Quanzhou Station.During data processing,it is found that there are a lot of harmonic noises with strong energy and regular frequency in the rotational signal.In order to remove these harmonic noises and better identify effective seismic events,this paper introduces four methods to suppress harmonic noises:low-pass filtering,time-frequency filtering,singular value decomposition and mono-frequency interference elimination,among which singular value decomposition method has a better effect.After further filtering,it is found that there are some effective rotational observation signals under the cover of harmonics.After analysis of polarization,they should be artificial interference near the source.