A Time-Frequency Domain First Arrival Detection Method for Microseismic Signals

Abstract

The invention provides a micro seismic signal time-frequency domain first arrival detection method. The method includes the following steps that: when the ICEEMDAN (improved complete ensemble empirical mode decomposition with adaptive noise) is adopted to perform data processing, microseismic signals are used as initial data, a specific white noise is added at each stage of decomposition, and a unique residual is calculated, so that each IMF (intrinsic mode function) can be obtained; the ICEEMDAN can adaptively decompose a complex signal into a series of IMF components which satisfy high frequency-to-low frequency series distribution; with the IMF components adopted as input, noise dominant modalities are directly removed, and other modalities are subjected to DFA (detrended fluctuation analysis) de-noising, and residual noises are removed through an interval hard threshold method; denoised results of different scales are fused and reconstructed, so that denoised seismic records, namely, effective signals can be obtained; and the first arrival information of the effective signals is obtained through high-precision time-frequency analysis detection. Compared with an empirical mode decomposition result, the method can significantly reduce modal aliasing phenomena, realize accurate reconstruction of original signals and has better convergence.

Description

technical field [0001] The invention relates to the field of exploration geophysics, in particular to a microseismic signal detection method based on the combination of overall average empirical mode decomposition ICEEMDAN and detrended fluctuation analysis (DFA) in seismic data processing. Background technique [0002] The energy of microseismic event signals is usually weak, and there is energy loss in the propagation process, which leads to the shortcomings of weak energy and low signal-to-noise ratio in the effective signal of seismic data received on the ground. Therefore, improving the signal-to-noise ratio of microseismic data is the The primary task of data processing and interpretation. With the improvement of positioning accuracy and the increasing demand for full-moment tensor inversion of focal mechanisms, the requirements for denoising technology are also gradually increasing. There are many traditional methods for suppressing random noise, which can be divided...

AI Technical Summary

Problems solved by technology

Surface microseismic data have the characteristics of strong noise and weak effective signal. Conventional denoising methods are often difficult to obtain better denoising results. Therefore, it is very important to study denoising methods specifically for ground microseismic data.

In view of the instability of the Empirical Mode Decomposition (EMD) method and the phenomenon of mode mixing (Huang, 1998), the overall Empirical Mode Decomposition (EEMD) uses the statistical characteristics of the uniform distribution of the Gaussian white noise spectrum to add different White noise makes the signal have continuity at different scales, but the calculation efficiency of this method is not high (Wu, 2009); the complete overall empirical mode decomposition method (CEEMD) can effectively Eliminate the residual auxiliary noise in the reconstructed signal, and the computational efficiency can also be improved (Yeh, 2009), but the accuracy of reconstruction will be lacking; Torres (2011, 2014) uses an improved complete overall empirical mode decomposition method ( ICEEMDAN) can accurately reconstruct the original signal, effectively reduce false modes and noise in the modes, and the computational cost is also reduced

[0003] However, the above methods have not been able to effectively remove noise and achieve the purpose of accurately detecting microseismic signals

Images