The Research Of De-Noising Method On Ground Microseismic Data

Microseismic monitoring is a geophysical technique, which is used to monitor the influence of producing activity and underground state through observing and analysing microseismic events caused by producing activity. Microseismic monitoring is divided into borehole monitoring and ground monitoring. Compared with borehole seismic monitoring, ground monitoring has more advantages, such as no monitoring well, flexible wiring, low cost, all of these makes microseismic monitoring has a much wider foreground. However, the noise of ground microseismic data is complex, microseismic signals are submerged in noise, which lead to the low noise ratio of microseismic data. This brings more difficulty to the data processing and interpretati-on. Therefore, research suitable denoising methods for improve the SNR of ground microseismic data has great significant for microseismic monitoring.At first, the characteristics and distribution regularities of microseismic signal and typical noise are summarized by making a comprehensive analysis to the actual microseismic data. And then, the scope and limitations of methods are summed up by researching the existing microseismic denoising methods. Which lay the foundation for researching denoising method on microseismic data.The small window svd method is an improved svd denoising method, which is proposed aimd at linear interference in microseismic data. At first, the method calcul-ates the standard trace in small window, gets time offsets by using cross-correlation calculation between standard trace and microseismic traces in small window. And then, flatting the microseismic data in small window according to time offsets, making svd decomposition on flatted microseismic data, selecting suitable singular values to refactor microseismic data. At last, accomplishing microseismic data filer by reverse flattening operation. The practice shows that design reasonable small window according to the characteristic of actual microseismic data and iterations of standard trace can suppress slope interference in microseismic data effectively.The single-trace svd method is an improved svd denoising method, which is proposed aimd at periodic interference in single-trace microseismic data. At First, it uses a single-channel seismic records to build the decomposing matrix. And then, selects appropriate singular values to rebulid the matrix through analysing the distribution law of the singular values about matrix. Finally, achieves the purpose of removing noise and highlighting effective signals through rebuliding signals by using svd inverse transform. The practice shows that the method can remove the periodic noise in single-trace microseismic records effectively. It is a denoising method suitable for the ground microseismic data. The improved time-varying skewness/kurtosis is a denoising method based on higher-order statistics, which is proposed aimd at the difference of symmetry or gaussianity between signal and noise. It uses normalized difference of time-varying skewness or kurtosis between small time window and long time window as the filter coefficient. Which weaken the effects of the denoising effect on asymmetry or nongaussianity of noise. To ensure that the signal better outstanding, signal-to-noise ratio of microseismic data is improved obviously.Fractal denoising is mainly used to suppress pulse interference in the microseis-mic data. At first, it constitutes open-close morphological filter and close-open morphological filter by using primary algorithm of fractal theory:morphological opening operation and morphological closing operation. And then, uses junction filter to denoise microseismic data, which is consists of open-close morphological filter and close-open morphological, so as to overcome statistical bias phenomenon in the process of morphological filtering. The practice shows that fractal denoising method can suppress the impulse noise in microseismic data effectively.The analysis of the theoretical model and actual microseismic data shows that the filtering methods researched in subject can suppress the noise in microseismic data effectively, and also improve the signal-to-noise ratio of microseismic data in a large extent. In addition, the noise of microseismic data is complex, it is difficult to obtain ideal denoising effect by using a method alone. So research new methods and combination of existing methods are two development direction for solving the low signal-to-noise ratio of microseismic data.