微地震信号的参数辨识建模及其Kalman滤波  

Parameter identification modeling of microseismic signals and Kalman filtering

作　　者：夏森[1] 王维波[1] 李树荣[1] 周德山[1] 陈静[1] 

机构地区：[1]中国石油大学(华东),青岛266580

出　　处：《地球物理学进展》2016年第5期2005-2010,共6页Progress in Geophysics

基　　金：中央高校基本科研业务专项基金(13CX02098A);国家自然科学基金(61573377)共同资助

摘　　要：由于微地震信号能量微弱、信噪比低,需要对采集到的微地震数据进行去噪处理,从而提高微地震记录的信噪比,提高震源定位的精度.目前存在许多基于模型的先进滤波方法,如基于机理模型的卡尔曼滤波已经成功应用在微地震信号去噪中.为了建立微地震信号的数学模型,改善卡尔曼滤波效果,本文通过数据辨识方法,对微地震信号建立了ARMA模型,并进一步转化为适用于卡尔曼滤波算法的状态空间模型.在此基础上研究了卡尔曼滤波方法,设计了适用于微地震去噪的卡尔曼滤波实现算法.理论模型和实际微地震监测数据处理结果表明,基于辨识模型的卡尔曼滤波算法能够有效抑制微地震信号中的随机噪声,显著提高微地震监测信号的信噪比,从而验证了该辨识模型的准确性和滤波算法的可行性.The microseismic signals usually have characteristics of the weak energy and low signal-to-noise ratio （SNR）, therefore, microseismic monitoring data must be preprocessed by various filtering methods to improve the SNR and the accuracy of source location. At present, many model-based filtering methods have been used for microseismic data filtering, such as Kalman Filter （KF） which has been successfully applied in microseismic signal denoising based on mechanism model built by mechanism analysis under many assumptions and simplifications. In order to establish a mathematical model of the microseismic signal and improve the effect of KF, this paper builds an ARMA model for a typical microseismic event by identification methods. Then, the ARMA model is converted into a state space model. Based on the research of identification modeling, the KF theory is studied and KF algorithm is designed which is useful for microseismic signal denoising. Through the processing with synthetic signals and microseismic fracturing ground monitoring data of a gas well, the noise is suppressed and the SNR is improved significantly by KF based on identification model, which verifies the feasibility of the identification model and the filtering algorithm.

关 键 词：微地震 辨识建模 ARMA模型 状态方程 卡尔曼滤波 

分 类 号：P631[天文地球—地质矿产勘探] P315[天文地球—地质学]