A Geological Discontinuity Anomaly Detection Method with Adaptive Direction Enhancement

Abstract

The present invention relates to the field of petroleum seismic exploration data processing and interpretation. It involves a method for detecting geological discontinuity anomalies that determines the direction of geological discontinuities and implements adaptive directional enhancement through anisotropic Gaussian iterative filtering of seismic signals, texture parameter calculation, and directional enhancement of discontinuity information detection. While suppressing noise and background interference in different spatial directions present in seismic data, it also highlights geological discontinuities and effective geological structural features in each direction. This method can better extract discontinuity anomaly information about reservoir structure and geological structures, providing reliable technical support for accurate geological structural interpretation, high-precision description of fracture system distribution, and reservoir prediction.

Description

Technical Field

[0001] The present invention relates to the field of petroleum seismic exploration, and specifically to a method for detecting geological discontinuity anomalies with adaptive directional enhancement. This method can extract and highlight discontinuity anomaly information of reservoir structure and geological formation in specific directions while suppressing noise in specific spatial directions of seismic data, thereby providing support for reservoir prediction and geological formation interpretation. Background Art

[0002] Seismic data contain a wealth of geological anomaly information with various directional characteristics. Extracting this anomaly and its directional information is a crucial aspect of oil and gas geophysical exploration. However, this valuable information is often affected by noise contained in the seismic data acquisition process. Therefore, it is crucial to suppress this noise while preserving the seismic anomaly information and its directional characteristics. Traditional noise suppression methods based on Gaussian filtering apply equal filter weights to all directions of the data, lacking the ability to suppress noise in specific directions and highlight the valuable information in those directions. Numerous researchers have addressed this issue. Freeman (1991) proposed a directionally adaptive filter design method. This directional filter is obtained by interpolating the elements of a directional filter basis set, but failed to specify how to correctly obtain this filter basis set and the correct interpolation rules. Van Ginkel et al. (1997) proposed a deconvolution method to improve the angular resolution of Gaussian filtering, but this method uses a computationally complex Fourier deconvolution algorithm. Geusebroek et al. (2003) proposed a fast anisotropic Gaussian filtering method. This filter selects different Gaussian scales in two different directions, so that the filter can better retain important information such as image edges during denoising. The decomposability of the Gaussian function is used to decompose the filter along the major and minor axes into two one-dimensional filters and convolve with the image, simplifying the calculation.

[0003] Seismic curvature attributes are another powerful seismic geometric attribute analysis method after coherence technology. Since the mid-1990s, curvature attributes based on second-order derivatives have been rapidly developed and widely used in structural interpretation, such as detecting the relationship between geological open cracks and Gaussian curvature (Lisle, 1994); Ericsson et al. (1998) demonstrated the relationship between oil and gas production and curvature, and believed that curvature is very helpful for conventional structural and topographic interpretation and improving the imaging quality of faults; Roberts (2001) elaborated on the basic theory of curvature attributes and proposed the first generation of curvature analysis method - surface curvature attributes. The calculation and workflow of curvature attribute showed that curvature attribute is very effective in extracting structural geometric features such as fault and fracture strike, laying the foundation for the promotion and application of curvature attribute in structural interpretation of seismic data. Hart (2002) studied the close relationship between strike curvature and open fractures in northwestern New Mexico. Bergbauer et al. (2003) used kx-ky filtering to calculate curvature at different wavelengths. Al-Dossary and Marfurt (2006) implemented a second-generation curvature analysis method based on 3D seismic data volume, namely volumetric curvature attribute, which utilizes both the amplitude information of seismic data and the geometric characteristics of seismic reflections to reduce the influence of noise and closure errors of horizon tracing on curvature.

[0004] Failure to adequately preserve seismic anomaly information and its directional characteristics during seismic data denoising hinders the accuracy of reservoir seismic detection and structural interpretation. If, during seismic data denoising, geological anomalies in specific directions can be retained and enhanced as needed for interpretation, and discontinuity anomaly information in corresponding directions can be further extracted and highlighted in subsequent seismic geometric attribute analysis, reservoir structure and geological structural information (such as indicators such as faults and fracture zones in different directions and their density) can be more accurately and reliably detected. This is precisely what the patented method of this invention achieves. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for detecting geological discontinuity anomaly information with adaptive direction enhancement. The principle of the present invention is to continuously change the direction of the anisotropic Gaussian filter according to the maximum azimuth and the direction of the fracture, so that the data is subject to less abnormal interference while retaining more effective information. The method of the present invention can effectively suppress interference that is different from the direction of the fracture, highlight the effective information of the seismic data, realize the enhancement of discontinuity information in a specific direction and the extraction of high-precision geological anomaly information, and can retain, extract and highlight one or more specific directions of discontinuity anomaly information according to the needs of reservoir and structural interpretation. While improving the signal-to-noise ratio of seismic data, it can more accurately and reliably detect the spatial distribution and internal structure of underground fracture systems and reservoirs. The method of the present invention includes the following main steps:

[0006] (1) Input a three-dimensional seismic dataset with a sampling period of Δt. Set the window length to 2W+1 sampling points. Perform time window processing on the input seismic dataset. Set the time window position τ to construct the corresponding time window seismic dataset S(τ+mΔt,x,y)| m∈[-W,W] , m is the window sampling point number, (x, y) is the plane coordinate;

[0007] (2) Use the following anisotropic Gaussian filter G:

[0008]

[0009] Where,

[0010]

[0011] (σ x ,σ y ) is the adjustment parameter of the anisotropic Gaussian filter, θ is the azimuth angle; in the azimuth angle range of 0° to 180°, K azimuth angles [θ1,θ2,θ3,...,θ K ], for the time window earthquake data set S(τ+mΔt,x,y)| at the time window position τ. m∈[-W,W] , perform anisotropic Gaussian filtering at each azimuth angle as follows:

[0012] F(τ+mΔt,x,y;θ k )| m∈[-W,W],k∈[1,K] =G θ (x,y;σ x ,σ y ,θ k )S(τ+mΔt,x,y)| m∈[-W,W] ;

[0013] (3) Pick up F(τ+mΔt,x,y;θ k )| m∈[-W,W],k∈[1,K]The maximum value of the azimuth angle is recorded as the maximum azimuth angle θ M , which is expressed as follows:

[0014]

[0015] (4) Calculate the maximum azimuth filtering data Z of the layer segment at the time window position τ according to the following formula:

[0016]

[0017] (5) Perform L-level quantization on the obtained data Z(τ,x,y) to form the maximum azimuth filtering quantized data volume Z L (τ,x,y), using Z L (τ,x,y) calculates K azimuth angles [θ1,θ2,θ3,...,θ K ]’s texture parameters: energy E(θ k ), correlation C(θ k );

[0018] (6) Using the texture parameters E(θ k ) and C(θ k ), the weight coefficient of energy and correlation is calculated according to the following formula, and the azimuth angle when the weight coefficient is the largest is taken as the optimal direction of the discontinuous geological anomaly information, which is recorded as θ O :

[0019]

[0020] (7) The data Z(τ,x,y) is transformed into an azimuth angle θ O The anisotropic Gaussian filtering process is as follows:

[0021] F O (τ,x,y)=G θ (x,y;σ x ,σ y ,θ O )Z(τ,x,y);

[0022] (8) Using the processed data F O (τ, x, y), is calculated according to the following formula to obtain the adaptive directional enhanced geological discontinuity anomaly information detection result:

[0023]

[0024] in:

[0025] BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1This is a slice through the target layer of the original 3D post-stack seismic data from the LH region. In the figure, the vertical axis represents the trace number (dimensionless), the horizontal axis represents the line number (dimensionless), and the color scale represents the amplitude of the seismic signal (dimensionless). Darker colors indicate smaller amplitudes, and darker colors indicate larger amplitudes.

[0027] Figure 2 is with Figure 1 Correspondingly, the method of the present invention only calculates the geological discontinuity anomaly information detection along the layer slice when the azimuth angle is 0°. In the figure, the ordinate is the track number, dimensionless, and the abscissa is the line number, dimensionless.

[0028] Figure 3 is with Figure 1 Correspondingly, the method of the present invention is used to calculate only the geological discontinuity anomaly information detection results along the layer slice when the azimuth angle is 90°. In the figure, the vertical axis is the track number, dimensionless, and the horizontal axis is the line number, dimensionless.

[0029] Figure 4 is with Figure 1 Correspondingly, the adaptive direction-enhanced geological discontinuity anomaly information detection along the layer slice obtained by the method of the present invention. In the figure, the vertical axis is the track number, dimensionless, and the horizontal axis is the line number, dimensionless. DETAILED DESCRIPTION

[0030] (1) Set the window length to 21 points, the time window position τ is the horizon time of the target layer, and construct the corresponding time window seismic dataset S(τ+mΔt,x,y)| from the 3D seismic dataset. m∈[-10,10] , where the slice along the layer at the time window position τ is as follows Figure 1 As shown;

[0031] (2) In the azimuth range of 0° to 180°, set 10 azimuths with an azimuth increment of 18° and set the adjustment parameters (σ x ,σ y ), for the time-windowed earthquake dataset S(τ+mΔt,x,y)| m∈[-10,10] , perform anisotropic Gaussian filtering at 10 azimuth angles, as follows:

[0032] F(τ+mΔt,x,y;θ k )| m∈[-10,10],k∈[1,10] =G θ (x,y;σ x ,σ y ,θ k )S(τ+mΔt,x,y)| m∈[-10,10] ;

[0033] (3) Pick up F(τ+mΔt,x,y;θ k )| m∈[-10,10],k∈[1,10]The maximum value of the azimuth angle is recorded as the maximum azimuth angle θ M , which is expressed as follows:

[0034]

[0035] (4) Calculate the maximum azimuth filtering data Z of the layer segment at the time window position τ according to the following formula:

[0036]

[0037] (5) The obtained data Z(τ,x,y) is quantized to 16 levels to form the maximum azimuth filtering quantized data volume Z 16 (τ,x,y), using Z 16 (τ, x, y) calculates the texture parameters of 10 azimuth angles [0°, 18°, 36°, ..., 162°]: energy E(θ k ), correlation C(θ k );

[0038] (6) Using the texture parameters E(θ) at 10 azimuth angles [0°, 18°, 36°, ..., 162°] k ) and C(θ k ), the weight coefficient of energy and correlation is calculated according to the following formula, and the azimuth angle when the weight coefficient is the largest is taken as the optimal direction of the discontinuous geological anomaly information, which is recorded as θ O :

[0039]

[0040] (7) The data Z(τ,x,y) is transformed into an azimuth angle θ O The anisotropic Gaussian filtering process is as follows:

[0041] F O (τ,x,y)=G θ (x,y;σ x ,σ y ,θ O )Z(τ,x,y);

[0042] (8) Using the processed data F O (τ, x, y), is calculated according to the following formula, and the result of the adaptive directional enhanced geological discontinuity anomaly information detection is finally obtained:

[0043]

[0044] in:

[0045]

[0046] The embodiment of the present invention is described as follows:

[0047] Figure 1 This is a slice through the target layer of raw 3D post-stack seismic data from a region where multiple phases of tectonic activity have created numerous discontinuous geological anomalies. The target layer is a carbonate reservoir with a well-developed fault system, complex porosity and permeability relationships, and strong heterogeneity. The data contains noise from multiple directions, making it difficult to accurately identify some fault systems.

[0048] Figure 2 and Figure 3 The technology of the present invention is used to calculate the geological discontinuity anomaly information along the layer slices at 0° and 90° respectively. Figure 2 It can be seen from the figure that the detection results mainly highlight discontinuous anomalies and structural features such as faults and cracks distributed laterally, with less noise and background interference; Figure 3 It can be seen that the detection results mainly highlight discontinuous anomalies and structural features such as longitudinally distributed faults and cracks, with very weak noise and background interference.

[0049] Figure 4 The technology of the present invention is used to calculate 10 directions of the input seismic data and then perform adaptive direction-enhanced geological discontinuity anomaly information detection along the layer slice. As can be seen from the figure, while greatly suppressing noise and background interference, the spatial distribution of discontinuous anomalies such as faults and cracks in various directions and at different scales, as well as complex structural features, are extracted and highlighted. Some small folds, sinkholes and other geological features are also clearly visible.

[0050] The advantages of the method of the present invention are that it can effectively suppress noise and interference information by combining the spatial direction or strike of geological discontinuity information (such as fracture systems, river channels, specific geological bodies, etc.). While significantly suppressing noise, it can also retain effective geological discontinuity information, better highlighting geological discontinuity anomalies and effective features in that direction, and better reflecting the details and discontinuities in a specific direction. At the same time, according to the needs of seismic interpretation, based on the main distribution direction of geological discontinuity information in a specific work area, this method can select one or more fixed directions, thereby extracting discontinuity anomaly information in one or several main directions. This method can provide reliable technical support for accurate geological structure interpretation, high-precision description of fracture system distribution, reservoir prediction, etc.

[0051] The above embodiments are only used to illustrate the present invention, and the various implementation steps of the method can be changed. Any equivalent transformations and improvements based on the technical solution of the present invention should not be excluded from the scope of protection of the present invention.

Claims

1. A method for detecting geological discontinuity anomalies with adaptive directional enhancement, comprising the following steps: (1) Input a three-dimensional seismic dataset with a sampling period of Δt. Set the window length to 2W+1 sampling points. Perform time window processing on the input seismic dataset. Set the time window position τ to construct the corresponding time window seismic dataset S(τ+mΔt,x,y)| m∈[-W,W] , m is the window sampling point number, (x, y) is the plane coordinate; (2) Use the following anisotropic Gaussian filter G: Where, (σ x ,σ y ) is the adjustment parameter of the anisotropic Gaussian filter, θ is the azimuth angle; in the azimuth angle range of 0° to 180°, K azimuth angles [θ1,θ2,θ3,...,θ K ], for the time window earthquake data set S(τ+mΔt,x,y)| at the time window position τ. m∈[-W,W] , perform anisotropic Gaussian filtering at each azimuth angle as follows: F(τ+mΔt,x,y;θ k )| m∈[-W,W],k∈[1,K] =G θ (x,y;σ x ,s y ,i k )S(τ+mΔt,x,y)| m∈[-W,W] ; (3) Pick up F(τ+mΔt,x,y;θ k )| m∈[-W,W],k∈[1,K] The maximum value of the azimuth angle is recorded as the maximum azimuth angle θ M , It is expressed as follows: (4) Calculate the maximum azimuth filtering data Z of the layer segment at the time window position τ according to the following formula: (5) Perform L-level quantization on the obtained data Z(τ,x,y) to form the maximum azimuth filtering quantized data volume Z L (τ,x,y), using Z L (τ,x,y) calculates K azimuth angles [θ1,θ2,θ3,...,θ K ]’s texture parameters: energy E(θ k ), correlation C(θ k ); (6) Using the texture parameters E(θ k ) and C(θ k ), the weight coefficient of energy and correlation is calculated according to the following formula, and the azimuth angle when the weight coefficient is the largest is taken as the optimal direction of the discontinuous geological anomaly information, which is recorded as θ O : (7) The data Z(τ,x,y) is transformed into an azimuth angle θ O The anisotropic Gaussian filtering process is as follows: F O (τ,x,y)=G θ (x,y;σ x ,s y ,i O )Z(τ,x,y); (8) Using the processed data F O (τ, x, y), is calculated according to the following formula to obtain the adaptive directional enhanced geological discontinuity anomaly information detection result: in:

Images