Tidal waterway earthquake multi-attribute description method and device under gypsum salt-carbonate rock symbiosis background

Abstract

The invention belongs to the technical field of petroleum and natural gas exploration and development, and discloses a tidal waterway earthquake multi-attribute description method and device under the gypsum salt-carbonate rock symbiosis background. According to the method, rock physical parameters are calibrated based on logging data, a parameterized model is established for forward modeling, and the seismic recognition lower limit of a tidal passage is quantitatively determined; adaptive optimization and multi-dimensional fusion are carried out on various seismic attributes by using an integrated learning model, and fine and reliable identification of the tidal tract is realized; in combination with a tidal tract-mudstone shielding reservoir forming mode, favorable lithologic traps are screened and quantitatively evaluated from identification results; the oil-gas possibility is verified through well drilling, and the model is driven to be continuously optimized by actually measured geological data feedback. According to the method, the influence of interference factors such as faults is effectively inhibited, the continuity and boundary definition of tidal passage display are improved, and hidden lithologic trap identification is successfully guided.

Description

Technical Field

[0001] This invention belongs to, but is not limited to, the field of oil and gas exploration and development technology, and particularly relates to a method and apparatus for characterizing seismic properties of tidal channels in a gypsum-carbonate rock symbiotic background. Background Technology

[0002] In a gypsum-carbonate coexisting sedimentary basin in a certain exploration area in western China, tidal channels (hereinafter referred to as "tidal channels") are important oil and gas reservoirs. These tidal channels formed in a platform setting during marine transgression, and often contain high-porosity and high-permeability bioclastic limestone and grain limestone reservoirs. In the later stages of diagenesis, the gypsum-salt rock strata above or on the flanks of the tidal channel can provide a good capping effect, while if the tidal channel itself is filled with mudstone laterally, it can form an effective lateral barrier, thus constituting a "tidal channel-mudstone barrier" type lithological trap. This type of trap has become an important exploration target for increasing reserves and production in this field. However, the accurate identification and trap evaluation of such tidal channels has always been a technical challenge in oil and gas exploration.

[0003] Tidal channels typically exhibit characteristics such as slight downward pull on the phase axis, reduced reflected energy, or discontinuity of the phase axis on seismic profiles. These characteristics are easily confused with the seismic responses of small faults, solution valleys, or other sedimentary microfacies, leading to significant identification uncertainty and making them highly susceptible to being overlooked or misjudged.

[0004] Currently, the industry commonly uses geometric properties such as coherence and variance to characterize channels or faults. However, in gypsum-salt rock backgrounds, salt-related structures (such as salt arches and small faults) are well-developed, generating significant noise interference. Conventional coherence properties are sensitive to these linear structures, but respond poorly to weak-amplitude, nonlinear geological bodies like tidal channels, resulting in discontinuous and unclear channel distributions and unclear morphologies, making it difficult to effectively guide well placement.

[0005] Existing technologies mostly focus on using single attributes for channel identification, lacking a systematic approach that goes from clearly defining response characteristics through forward modeling to finely characterizing optimized attributes, and finally to comprehensively evaluating trap effectiveness. In particular, the failure to effectively combine regional sedimentary patterns (such as tidal channel filling patterns) to establish hydrocarbon accumulation models leads to blind trap identification and low exploration success rates.

[0006] Existing technologies struggle to effectively and accurately identify tidal channels and evaluate their hydrocarbon accumulation potential in gypsum-carbonate rock coexistence settings. To achieve new breakthroughs in this field, effectively expanding exploration areas and steadily increasing oil and gas reserves has become a key challenge. Therefore, a comprehensive identification and evaluation method that overcomes the aforementioned shortcomings is urgently needed in this field. Summary of the Invention

[0007] To address the problems existing in the prior art, this invention provides a method and apparatus for characterizing seismic properties of tidal channels in a gypsum-carbonate rock symbiotic background.

[0008] This invention is implemented as follows: a method for multi-attribute characterization of seismic events in tidal channels within a gypsum-carbonate rock symbiotic background, the method comprising:

[0009] S1. Based on well logging data, calibrate rock physical parameters, establish a parametric model for forward modeling, and quantitatively determine the lower limit for identifying tidal channel earthquakes;

[0010] S2: By using an ensemble learning model to adaptively select seismic attributes and fuse multi-dimensional features, a precise and reliable identification of tidal channels can be achieved.

[0011] S3: By applying the “tidal channel-mudstone shielding” hydrocarbon accumulation model, we screen and quantitatively evaluate the favorable traps available for drilling from the identified tidal channels.

[0012] S4: Confirm the hydrocarbon potential of the trap by deploying wells and drive continuous optimization of the entire approach with feedback from real geological data.

[0013] Furthermore, S1 specifically includes:

[0014] Based on limited drilling data within the work area, the petrophysical parameters of mudstone infill, carbonate reservoir, and surrounding rock were accurately calibrated, thereby constructing high-fidelity, parametric two-dimensional or three-dimensional geological models. By systematically altering the width, cutting depth, and reservoir thickness of the tidal channel, and using wave equations for forward modeling, the seismic response corresponding to different geological scenarios was observed under controlled conditions. This process intuitively verified the physical mechanism by which the tidal channel exhibits phase axis pull-down, reduced reflected energy, and phase axis discontinuity on seismic profiles. Furthermore, it allows for the quantitative determination of the minimum depth (≥3 meters) and minimum width (≥20 meters) that can reliably identify the tidal channel under current seismic data quality. The establishment of this "identification threshold" provides an objective and unified benchmark for subsequent interpretation of actual data, avoiding blindness and arbitrariness in the identification process and laying the scientific foundation for the entire method.

[0015] Furthermore, S2 specifically includes:

[0016] By employing a machine learning-based intelligent attribute optimization method, tidal channel identification has achieved a leap from "human experience-driven" to "data intelligence-driven." This method automatically extracts four types of features from seismic data—amplitude, geometry, texture, and spectrum—using a multi-attribute feature intelligent extraction algorithm. It then uses feature importance assessment and recursive feature elimination techniques to adaptively select the optimal attribute combination. For the core Euler curvature attribute, a patented adaptive optimization mechanism for azimuth parameters was developed, determining the best azimuth parameters through intelligent scanning. An innovative multi-model integrated identification framework was constructed, significantly improving the accuracy and robustness of tidal channel identification in complex geological backgrounds through the collaborative work of a geometric shape recognizer, a reflection feature recognizer, and a texture pattern recognizer, followed by comprehensive decision-making through a multivariate learning fusion processor. Simultaneously, the system introduces a confidence score and uncertainty quantification system, providing a reliability assessment for each identification result by fusing multi-dimensional indicators such as model probability, spatial continuity, and geological rationality. This intelligent improvement not only enhances processing efficiency but also ensures the geological rationality of the technological achievements by encoding geological rules to constrain the algorithm, forming a tidal channel intelligent identification system with continuous learning capabilities.

[0017] The specific implementation process is as follows:

[0018] First, the seismic amplitude data is loaded into 3D seismic data, metadata is extracted, and finally, the standardized 3D seismic data volume is output.

[0019] Next, well logging data is loaded to create training labels, converting oil and gas wells into supervised learning labels.

[0020] Next, geometric feature values ​​are extracted to optimize geometric attributes for depicting the planar distribution of the tidal channel and identify chaotic reflection features within the channel. The model is trained to construct a geometric feature recognition model to learn the recognition rules of geometric attributes such as Euler curvature; an amplitude feature recognition model to learn amplitude anomalies such as weak reflection and pull-down; and a texture recognition CNN model to learn the internal structural features of the tidal channel. The ensemble model is trained to continuously improve the method and clearly depict the tidal channel distribution.

[0021] Then, a geological rule constraint engine is established, the geological rule library is initialized, and geological knowledge is transformed into computable constraint conditions; morphological constraints are applied to ensure that the identification results conform to the morphological characteristics of the tidal channel; spatial constraints are applied to realize the quantitative constraints of spatial relationships such as the perpendicularity of fault strike to the tidal channel distribution direction.

[0022] Finally, sedimentary background knowledge is incorporated into the identification process to improve geological rationality, conduct confidence assessments, and achieve the quantitative requirements for the accuracy of identifying tidal channels through seismic identification.

[0023] Furthermore, S3 specifically includes:

[0024] The innovatively proposed "tidal channel-mudstone blocking" hydrocarbon accumulation model imbues the tidal channels identified by S2 with clear geological significance. By intelligently overlaying tidal channel distribution maps with structural and caprock distribution maps, tidal channel segments located along hydrocarbon migration paths and effectively laterally blocked by tight mudstone are systematically screened. This process essentially performs a rigorous geological rationality filter on the S2 identification results. If the "tidal channels" identified by S2 cannot form an effective hydrocarbon accumulation configuration with structures and caprocks, their value as exploration targets will be significantly reduced. Furthermore, quantitative calculations of key parameters such as area and closure amplitude are performed on each selected potential trap, and their reservoir-caprock combination and migration conditions are comprehensively evaluated. This elevates the qualitative identification of S2 to the level of quantitative target evaluation, verifying that the identification results not only "resemble tidal channels" but also possess the potential to be "good traps."

[0025] Furthermore, S4 specifically includes:

[0026] The favorable traps selected in S3 were deployed as drilling targets. Oil and gas shows, logging interpretations, and test production obtained from actual drilling directly verified whether the "tidal channel-mudstone shielding" traps identified in S2 and evaluated in S3 truly existed and contained oil and gas. Successful drilling provides the strongest proof of the accuracy of S2 identification and the hydrocarbon accumulation model in S3. A continuous optimization feedback mechanism was established. The geological information revealed by drilling was used as new and most reliable constraints, fed back to the starting point of the process: feedback to the forward model in S1 corrected and refined the initial petrophysical parameters and geological model, making the forward response closer to reality; feedback to the attribute interpretation and machine learning model in S2 served as new training samples or constraint rules, further optimizing attribute parameters and improving the prediction accuracy and generalization ability of the machine learning model.

[0027] Another objective of this invention is to provide a device for characterizing tidal channel seismic multiple attributes in a gypsum-carbonate rock symbiotic background, which implements the aforementioned method for characterizing tidal channel seismic multiple attributes in a gypsum-carbonate rock symbiotic background. The device comprises:

[0028] The model building module calibrates rock physical parameters based on well logging data, establishes a parametric model for forward modeling, and quantitatively determines the lower limit for identifying tidal channel earthquakes.

[0029] The attribute optimization module is connected to the model building module. It uses an integrated learning model to adaptively optimize seismic attributes and fuse multi-dimensional features, thereby achieving precise and reliable identification of tidal channels.

[0030] The identification module, connected to the attribute optimization module, uses the "tidal channel-mudstone shielding" reservoir formation model to screen and quantitatively evaluate favorable traps available for drilling from the identified tidal channels.

[0031] The continuous optimization module, connected to the identification module, uses drilling to empirically verify the hydrocarbon-bearing properties of traps and drives the continuous optimization of the entire method with feedback from real geological data.

[0032] Another object of the present invention is to provide a computer device, the computer device including a memory and a processor, the memory storing a computer program, the computer program being executed by the processor causing the processor to perform the steps of the method for multi-attribute characterization of tidal channel seismic events in a gypsum-carbonate rock symbiotic background.

[0033] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method for characterizing tidal channel seismic multiple attributes in a gypsum-carbonate rock symbiotic background.

[0034] Another objective of this invention is to provide an information data processing terminal, which is used to realize a multi-attribute characterization device for seismic events in tidal channels under the symbiotic background of gypsum-carbonate rocks.

[0035] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows:

[0036] Compared with existing technologies, the multi-attribute seismic characterization technology for tidal channels in the context of gypsum-salt-carbonate rock coexistence proposed in this invention has achieved significant technological progress and positive results in terms of tidal channel identifiability determination, seismic attribute interpretation reliability, and exploration target effectiveness.

[0037] First, this invention introduces a systematic forward modeling approach to accurately calibrate rock physical parameters based on actual drilling and logging data. Under controlled conditions, it simulates the seismic response characteristics of tidal channels at different scales. This allows for the first time the identification of phenomena such as pull-down reflection, weakened reflection energy, and discontinuity of phase axes in seismic profiles from traditional qualitative understandings relying on interpreters' experience to quantitative identification standards with clear scale constraints. By establishing a lower limit for seismic identification of tidal channels, interpreters have a unified and objective basis for judgment in actual data processing, effectively avoiding over-interpretation or misjudgment caused by insufficient seismic resolution, fundamentally improving the scientific rigor and consistency of tidal channel identification results.

[0038] Secondly, this invention introduces a multi-attribute adaptive fusion mechanism based on identification lower limit constraints at the seismic attribute interpretation level. By systematically screening and integrating multi-dimensional seismic attributes such as amplitude, geometry, texture, and spectrum, it effectively suppresses the problem of single attributes being susceptible to noise and interference from non-target geological bodies. In particular, addressing the impact of strong interference factors such as faults and salt body boundaries on tidal channel identification in the context of gypsum-carbonate rock coexistence, this invention significantly weakens the interference of fault response on attribute results by optimizing the Euler curvature attribute parameter. This makes the tidal channel distribution on the planar attribute map more continuous, the boundaries clearer, and the spatial morphology more consistent with actual sedimentary characteristics, thereby greatly improving the stability and reliability of tidal channel identification in complex geological backgrounds.

[0039] Furthermore, this invention breaks through the traditional research approach that only focuses on seismic geometry identification, creatively establishing a "tidal channel-mudstone blocking" hydrocarbon accumulation model, assigning clear hydrocarbon accumulation significance to the identified tidal channel bodies. By jointly analyzing the distribution of tidal channels with structural conditions and the sealing relationship of caprock mudstone, it achieves effective screening and quantitative evaluation of concealed lithological traps. This transforms areas that originally lacked obvious structural closure characteristics and were difficult to use as exploration targets into new exploration targets with clear hydrocarbon accumulation conditions, thus effectively filling the gaps in traditional structural trap exploration and opening up a completely new exploration field for increasing oil and gas reserves and production.

[0040] Furthermore, this invention introduces a drilling verification and reverse feedback mechanism, using geological information obtained from actual drilling as the most reliable constraint to continuously refine the forward model and seismic attribute identification rules, enabling the entire tidal channel characterization process to have dynamic optimization capabilities. This mechanism not only improves the accuracy of subsequent prediction results but also significantly reduces the uncertainty risks in exploration deployment. Practical applications show that this technical solution can effectively improve the accuracy of trap identification, reduce the probability of empty and inefficient wells, and reduce ineffective drilling investment, demonstrating significant economic benefits and engineering application value.

[0041] This invention achieves a substantial leap from "visible" to "determinable" and from "like a target" to "is a target" through systematic innovation in identification mechanisms, attribute interpretation methods, and hydrocarbon accumulation evaluation ideas. It has significant technical advantages and positive effects in the context of complex gypsum-salt-carbonate rock geological background.

[0042] (1) The expected benefits and commercial value of the technical solution of this invention after transformation are as follows:

[0043] By implementing the technical solution of this invention, an effective gas layer was identified in the target formation. Gas testing results showed that the well produced an industrial gas flow after commissioning, with a considerable daily gas production, confirming the effectiveness of the trap and the reliability of the method of this invention. This signifies that this method not only significantly improves the exploration success rate and reduces risks, but also directly generates economic benefits.

[0044] (2) This invention provides a systematic method from clear response characteristics in forward modeling to fine characterization of preferred attributes, and then to comprehensive evaluation of trap effectiveness. It solves the technical problem of effectively and accurately identifying tidal channels and evaluating their hydrocarbon accumulation potential in the context of gypsum-salt-carbonate rock coexistence. It has important guiding significance for expanding new fields of oil and gas exploration and achieving large-scale reserve increase. Attached Figure Description

[0045] Figure 1 : A schematic diagram of the process for tidal channel identification and trap evaluation based on forward modeling and attribute optimization;

[0046] Figure 2 Schematic diagram of the forward model of the tidal channel (fixed width and variable depth, fixed depth and variable width);

[0047] Figure 3: Conventional coherence properties and curvature properties;

[0048] Figure 3-a Comparison chart of multi-attribute identification in the weak response zone of the tidal channel;

[0049] Figure 3-a (Left): Original seismic multi-attribute composite map of the weak response zone of the tidal channel under the background of gypsum-salt-carbonate rock coexistence;

[0050] Figure 3-a (Right): Schematic diagram of the interpretation results of tidal channel boundaries and orientations obtained by the method of this invention.

[0051] Figure 3-b Comparison of tidal channel identification under fracture interference background;

[0052] Figure 3-b (Left): Multi-attribute composite image of earthquakes affected by faults and tectonic bands;

[0053] Figure 3-b (Right): Schematic diagram of the tidal channel distribution and boundary interpretation results extracted by the method of the present invention under the background of fracture interference.

[0054] Figure 3-c Comparison of detailed depictions of multiple overlapping tidal channels;

[0055] Figure 3-c (Left): Multi-attribute seismic composite map of the multi-stage superimposed tidal channel development zone;

[0056] Figure 3-c(Right): Schematic diagram illustrating the interpretation results of the distribution patterns and bifurcation relationships of multiple tidal channels identified using the method of this invention;

[0057] Figure 4: Multi-azimuth display effect of Euler curvature (preferred 45° azimuth angle); Figure 4-a A comparison diagram of the linear body response of the tidal channel under different azimuth seismic attributes is presented to illustrate that the tidal channel has obvious directional selectivity to different azimuth attributes.

[0058] Figure 4-b The diagram shows a comparison of the tidal channel response under different preferred azimuth attributes, illustrating the adaptive optimization capability of the method of the present invention for the optimal azimuth attribute.

[0059] Figure 4-c The image shows a comparison of the tidal channel recognition effects of the preferred attributes of this invention and the traditional curvature attributes. The left image shows the Euler curvature attribute at an azimuth angle of 45°, and the right image shows the most negative curvature attribute. The red circles mark the areas in the traditional attributes that are prone to misidentification.

[0060] Figure 5 : Plan view of tidal channel identification and trap distribution;

[0061] Figure 6 : Verification diagram of well drilling results;

[0062] Figure 7 : A multi-attribute characterization device for seismic events in tidal channels under the symbiotic background of gypsum-carbonate rocks;

[0063] In the diagram: 1. Model building module; 2. Attribute optimization module; 3. Identification module; 4. Continuous optimization module. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0065] like Figure 1 As shown, this embodiment of the invention provides a method for multi-attribute characterization of seismic events in tidal channels under a gypsum-carbonate rock symbiotic background. The method includes:

[0066] S1. Based on well logging data, calibrate rock physical parameters, establish a parametric model for forward modeling, and quantitatively determine the lower limit for identifying tidal channel earthquakes;

[0067] S2: By using an ensemble learning model to adaptively select seismic attributes and fuse multi-dimensional features, a precise and reliable identification of tidal channels can be achieved.

[0068] S3: By applying the “tidal channel-mudstone shielding” hydrocarbon accumulation model, we screen and quantitatively evaluate the favorable traps available for drilling from the identified tidal channels.

[0069] S4: Verify the hydrocarbon potential of traps by deploying wells and drive continuous optimization of the entire approach with feedback from real geological data.

[0070] S1 specifically includes:

[0071] Based on limited drilling data within the work area, the petrophysical parameters of mudstone infill, carbonate reservoir, and surrounding rock were accurately calibrated, thereby constructing high-fidelity, parametric two-dimensional or three-dimensional geological models. By systematically altering the width, cutting depth, and reservoir thickness of the tidal channel, and using wave equations for forward modeling, the seismic response corresponding to different geological scenarios was observed under controlled conditions. This process intuitively verified the physical mechanism by which the tidal channel exhibits phase axis pull-down, reduced reflected energy, and phase axis discontinuity on seismic profiles. Furthermore, it allows for the quantitative determination of the minimum depth (≥3 meters) and minimum width (≥20 meters) that can reliably identify the tidal channel under current seismic data quality. The establishment of this "identification threshold" provides an objective and unified benchmark for subsequent interpretation of actual data, avoiding blindness and arbitrariness in the identification process and laying the scientific foundation for the entire method.

[0072] S2 specifically includes:

[0073] By employing a machine learning-based intelligent attribute optimization method, tidal channel identification has achieved a leap from "human experience-driven" to "data intelligence-driven." This method automatically extracts four types of features from seismic data—amplitude, geometry, texture, and spectrum—through a multi-attribute feature intelligent extraction network. It then uses feature importance assessment and recursive feature elimination techniques to adaptively select the optimal attribute combination. For the core Euler curvature attribute, a patented adaptive optimization mechanism for azimuth parameters was developed, determining the best azimuth parameters through intelligent scanning. An innovative multi-model integrated identification framework was constructed. Through the collaborative work of a geometric shape recognizer, a reflection feature recognizer, and a texture pattern recognizer, followed by a meta-learning fusion processor for comprehensive decision-making, the accuracy and robustness of tidal channel identification in complex geological backgrounds are significantly improved. Simultaneously, the system introduces a confidence score and uncertainty quantification system, providing a reliability assessment for each identification result by fusing multi-dimensional indicators such as model probability, spatial continuity, and geological rationality. This intelligent improvement not only enhances processing efficiency but also ensures the geological rationality of the technological achievements by encoding geological rules into algorithmic constraints, forming a tidal channel intelligent identification system with continuous learning capabilities.

[0074] S3 specifically includes:

[0075] The innovatively proposed "tidal channel-mudstone blocking" hydrocarbon accumulation model endows the tidal channels identified by S2 with a clear geological function. By intelligently overlaying tidal channel distribution maps with structural and caprock distribution maps, tidal channel segments located on hydrocarbon migration paths and effectively laterally blocked by tight mudstone are systematically screened. This process essentially performs a rigorous geological rationality filter on the S2 identification results. If the "tidal channels" identified by S2 cannot form an effective hydrocarbon accumulation configuration with structures and caprocks, their value as exploration targets will be greatly reduced. Furthermore, the area, closure amplitude, and other key parameters of each potential trap are quantitatively calculated, and their reservoir-caprock combination and migration conditions are comprehensively evaluated. This elevates the qualitative identification of S2 to the level of quantitative target evaluation, verifying that the identification results not only "resemble tidal channels" but also possess the potential of "good traps."

[0076] S4 specifically includes:

[0077] The favorable traps selected in S3 were deployed as drilling targets. Oil and gas shows, logging interpretations, and test production obtained from actual drilling directly verified whether the "tidal channel-mudstone shielding" traps identified in S2 and evaluated in S3 truly existed and contained oil and gas. Successful drilling provides the strongest proof of the accuracy of S2 identification and the hydrocarbon accumulation model in S3. A continuous optimization feedback mechanism was established. The geological information revealed by drilling was used as new and most reliable constraints, fed back to the starting point of the process: feedback to the forward model in S1 corrected and refined the initial petrophysical parameters and geological model, making the forward response closer to reality; feedback to the attribute interpretation and machine learning model in S2 served as new training samples or constraint rules, further optimizing attribute parameters and improving the prediction accuracy and generalization ability of the machine learning model.

[0078] like Figure 7 As shown, this embodiment of the invention provides a device for characterizing tidal channel seismic multiple attributes in a gypsum-carbonate rock symbiotic background, which implements the aforementioned method for characterizing tidal channel seismic multiple attributes in a gypsum-carbonate rock symbiotic background. The device includes:

[0079] Model building module 1 calibrates rock physical parameters based on well logging data, establishes a parametric model for forward modeling, and quantitatively determines the lower limit for identifying tidal channel earthquakes;

[0080] The attribute optimization module 2 is connected to the model building module 1. It uses an integrated learning model to adaptively optimize seismic attributes and fuse multi-dimensional features, thereby achieving precise and reliable identification of tidal channels.

[0081] The identification module 3, connected to the attribute optimization module 2, uses the "tidal channel-mudstone shielding" reservoir formation mode to screen and quantitatively evaluate the favorable traps available for drilling from the identified tidal channels.

[0082] The continuous optimization module 4, connected to the identification module 3, uses drilling to empirically verify the hydrocarbon content of traps and drives the continuous optimization of the entire method with feedback from real geological data.

[0083] This instrument uses a gypsum-carbonate coexisting sedimentary system as its geological background and conducts multi-attribute intelligent characterization and evaluation of the seismic identifiability and hydrocarbon accumulation effectiveness of tidal channels. First, the model building module calibrates petrophysical parameters based on well logging data and core interpretation results, constructing parametric geological models for different tidal channel sizes, infill types, and overlying shielding conditions. Forward modeling is then used to generate corresponding seismic response templates, quantitatively determining the minimum identifiable scale and seismic response characteristics of the tidal channel under this complex background. Subsequently, the attribute optimization module uses forward modeling samples and actual seismic data as input, employing an ensemble learning model to adaptively optimize and weight multiple seismic attributes, extracting the most sensitive and stable feature combinations for tidal channel identification, thus achieving a refined characterization of the tidal channel. Building upon this, the identification module introduces a "tidal channel-mudstone shielding" hydrocarbon accumulation model, performing spatial overlay and hydrocarbon accumulation constraint analysis on the identified tidal channels, screening and quantitatively evaluating favorable traps with effective shielding and reservoir conditions. Finally, the continuous optimization module verifies the oil and gas content of the trap through drilling and feeds the actual drilling results back to the aforementioned model and attribute optimization process to dynamically correct parameters, weights, and criteria, thereby achieving continuous self-learning and iterative optimization of the device performance.

[0084] The seismic multi-attribute characterization device for tidal channels in a gypsum-carbonate rock coexistence background described in this invention is based on the synergistic utilization of seismic and drilling data. Through the sequential linkage of modules such as model building, attribute optimization, trap identification, and continuous optimization, it achieves a systematic characterization of tidal channels, from identifiability constraints to hydrocarbon accumulation targets. First, the model building module calibrates the petrophysical parameters corresponding to different lithologies based on existing well logging data, constructs a parametric geological model, and conducts forward modeling. By comparing and analyzing the differences in seismic response caused by variations in the geometric parameters of the tidal channel, it quantitatively determines the lower limit of seismic identification of the tidal channel under the current seismic data quality conditions, providing objective constraints for subsequent interpretation. Subsequently, the attribute optimization module extracts multi-dimensional attributes from the seismic data under the aforementioned lower limit constraint and adaptively filters and fuses attribute combinations through ensemble learning, suppressing noise interference and the influence of non-geological anomalies to achieve precise identification of the tidal channel. Building upon this foundation, the identification module incorporates the hydrocarbon accumulation configuration relationship between tidal channels and mudstone plugging. It jointly analyzes the identified tidal channels with structural conditions and caprock distribution to screen and quantitatively evaluate favorable traps with effective plugging conditions. Finally, the continuous optimization module uses drilling verification results as feedback, applying the real geological information revealed by actual drilling to the model building and attribute identification process. This corrects and updates rock physical parameters and identification rules, thus forming a continuously iteratively optimized tidal channel seismic characterization mechanism, ensuring the reliability and engineering applicability of the characterization results in complex geological contexts.

[0085] Example 1: Implementation of Establishing the Lower Limit for Seismic Identification in Tidal Channels Based on Forward Modeling

[0086] To address the issues of ambiguity and subjectivity in seismic identification of tidal channels due to unclear geological scales, this embodiment provides a method for quantitatively determining the lower limit of seismic identifiability. Within a gypsum-carbonate rock coexistence block, well logging data from existing wells are collected, including P-wave velocity, S-wave velocity, and density curves. Based on lithological interpretation results, the petrophysical parameters corresponding to the mudstone infill, carbonate reservoir, and surrounding rocks of the tidal channel are calibrated. On this basis, a two-dimensional or three-dimensional geological model incorporating velocity, density, and stratigraphic sequence structure is constructed. Multiple controlled geological scenarios are formed by artificially setting tidal channel bodies of different widths, cutting depths, and reservoir thicknesses within the model. For each geological scenario, a corresponding synthetic seismic profile is generated using wave equation numerical simulation (finite difference method).

[0087] By comparing the synthetic seismic responses under different geological scenarios, the variation law of reflection characteristics of tidal channels on seismic profiles was systematically analyzed. The results show that when the cutting depth of a tidal channel is less than 3 meters or its width is less than 20 meters, its seismic response gradually weakens and becomes difficult to distinguish from background reflections. When the cutting depth is not less than 3 meters and the width is not less than 20 meters, the tidal channel on the seismic profile exhibits stable characteristics of phase axis pull-down, weakened reflection energy, and local phase axis discontinuity. Therefore, a lower limit for the identifiability of tidal channels was established under the seismic data quality conditions of this work area, providing a unified and objective criterion for subsequent interpretation of actual data. This method quantitatively correlates geological scale with seismic response, establishing an objective and repeatable identification threshold, fundamentally avoiding ineffective interpretation of seismically indistinguishable geological bodies.

[0088] Example 2: Implementation of Seismic Multi-Attribute Adaptive Fusion Identification under Identification Lower Limit Constraint

[0089] Based on the tidal channel identification lower limit determined in Example 1, the geological scale corresponding to the identification lower limit is used as the spatial filter parameter to preprocess the seismic data volume, remove false anomalies caused by high-frequency noise, and perform multi-attribute calculations on the processed data to generate multi-dimensional seismic attribute data volumes such as amplitude, geometry, texture, and spectrum. Furthermore, a multi-attribute feature sample set is constructed to ensure that the features involved in the identification only correspond to the spatial scale range supported by the seismic resolution capability, avoiding the introduction of indistinguishable small-scale anomalies into the identification process.

[0090] An ensemble learning approach (including random forest, AdaBoost, or a combination thereof) is employed to adaptively filter and fuse multidimensional attributes. Different attribute combinations generate tidal channel identification results, and a collaborative decision-making mechanism (probability-weighted averaging) outputs the final spatial distribution of tidal channels. Compared to single-attribute interpretations, this implementation significantly reduces the number of false anomalies under conditions of strong interference from gypsum and carbonate rock reflections. The identification results demonstrate superior spatial continuity and geological plausibility, validating the effectiveness of multi-attribute fusion identification under the constraint of the identification lower limit.

[0091] Example 3: Implementation method for screening and evaluating tidal channel traps based on mudstone blocking conditions

[0092] To address the challenges of identifying numerous geological bodies along tidal channels but lacking clear hydrocarbon accumulation potential and posing difficulties in selecting optimal drilling targets, this embodiment provides a rapid screening method that integrates mudstone obstruction conditions. After seismic identification of the tidal channels, the spatial distribution results of the tidal channels are jointly analyzed with tectonic interpretation results and mudstone distribution information. Through spatial overlay, the method prioritizes screening channel segments located along hydrocarbon migration paths that exhibit continuous mudstone distribution in the lateral or updip direction of the tidal channel, ensuring that the selected targets possess the geological basis for effective sealing.

[0093] Quantitative evaluations were conducted on the selected tidal channel segments, calculating their planar distribution area and structural closure amplitude, and a comprehensive analysis of the spatial configuration relationship between the reservoir and mudstone was performed. The evaluation results indicate that only some tidal channels meet the structural conditions while possessing effective mudstone shielding; the remaining channels, although exhibiting geometrical characteristics, lack reservoir-forming conditions. This implementation method further transforms seismic identification results into exploration targets with clear geological significance, avoiding the direct application of all identification results to drilling decisions.

[0094] This method directly transforms seismic interpretation results into graded exploration targets, improving drilling decision-making efficiency and avoiding exploration investment in ineffective geological bodies (having shape but no traps).

[0095] Example 4: Implementation of Drilling Feedback-Driven Model and Iterative Optimization of Identification Process.

[0096] Based on the drilling results, the geological model and rock physical parameters were revised on the one hand, and the attribute identification rules and weights were optimized on the other hand.

[0097] Tidal channel traps with superior evaluation results from Example 3 were selected as drilling targets. Drilling operations were carried out, and well logging interpretation results and hydrocarbon show information generated during the actual drilling process were obtained. The drilling results show that some tidal channel traps contain carbonate reservoirs with good physical properties, and the hydrocarbon shows correspond well with the seismic prediction results, verifying the reliability of the tidal channel identification and hydrocarbon accumulation screening process.

[0098] The actual lithological distribution, reservoir thickness, and fluid information revealed by drilling are fed back into the previous geological model and seismic interpretation process to correct the original rock physical parameters, and a new forward modeling is then performed based on this. The corrected forward modeling results show a significantly improved match with the actual seismic response, providing a model basis that more closely reflects the actual geological conditions for the identification of tidal channels in adjacent blocks, demonstrating the crucial role of drilling feedback in the overall methodology.

[0099] The locations of tidal and non-tidal waterways confirmed in the drilling results are used as new constraint samples and introduced into the seismic attribute fusion and identification process. By comparing the differences in attribute responses corresponding to successfully drilled targets and unsuccessfully drilled targets, the original attribute combinations and their weights are adjusted to make the identification rules more closely match the actual geological response characteristics.

[0100] After multiple rounds of feedback optimization, the results were reapplied to the interpretation of seismic data in un-drilled areas. The results showed a significant improvement in the stability and predictive consistency of the tidal channel identification results. This implementation method demonstrates that by continuously introducing real geological information to reverse-correct the identification process, the accuracy and generalization ability of tidal channel characterization can be gradually improved, making the entire technical process have sustainable optimization value for engineering applications.

[0101] This embodiment transforms the technical solution from a static method into a dynamic system with learning capabilities. As exploration progresses, the system's accuracy and generalization ability can continuously improve, making it particularly suitable for exploration in new areas where geological understanding is gradually deepening.

[0102] This invention provides a method for identifying tidal channels and evaluating traps in gypsum-carbonate rock coexisting zones based on geological model forward modeling and seismic attribute optimization theory. The theoretical basis and implementation path of the forward modeling and attribute optimization method are explained first. Seismic geological comprehensive prediction methods are processes for quantitatively characterizing and effectively identifying complex geological targets. Their core lies in transforming geological models into interpretable information through geophysical methods. Based on differences in technical approaches, they can be divided into two main categories: model-based forward modeling and data-driven attribute analysis. Specifically, this includes forward modeling methods that verify the seismic response characteristics of geological bodies through numerical simulation, and attribute analysis methods that extract geometric and dynamic features from seismic data to identify target bodies. Forward modeling methods are mainly used to establish a quantitative relationship between geological bodies and seismic responses, while attribute analysis focuses on extracting effective information from actual data.

[0103] First, by comprehensively utilizing limited drilling and logging data, a geological model of tidal channels in a gypsum-carbonate rock symbiotic environment is constructed to approximate the actual underground geological structure. Under known model parameters, the propagation characteristics of seismic waves in different lithological combinations are simulated to clarify the seismic response patterns of tidal channels and their mudstone infill. Second, based on actual 3D seismic data and identified seismic anomalies, multi-attribute calculations and optimization analyses are conducted. Simultaneously, a technical approach guided by forward modeling and centered on Euler curvature attribute optimization is proposed. Through continuous iterative optimization of attribute parameters and model settings, the accuracy and reliability of tidal channel identification are gradually improved until a tidal channel distribution model and trap evaluation results with the highest degree of consistency with the actual underground conditions are formed.

[0104] Figure 2This paper presents geological models of tidal channels under different widths, depths, and reservoir thicknesses, along with their forward modeling seismic responses. The calibration of rock physical parameters based on well data serves as the basis for the forward modeling. Due to the complexity of actual subsurface geological structures, parametric modeling and systematic forward modeling analysis are the main methods for clarifying the seismic identification characteristics and quantification thresholds of tidal channels. Model 1 (fixed width, variable depth and reservoir thickness): The tidal channel width is fixed at 200 meters, the depth increases from 1 meter to 10 meters, and the reservoir thickness increases from 10 meters to 40 meters. Forward modeling results show that the tidal channel response exhibits a downward pull phenomenon, which is most pronounced when the depth is ≥3 meters and the reservoir thickness is 30 meters. Model 2 (fixed depth, variable width): The tidal channel depth is fixed at 5 meters, and the width increases from 10 meters to 700 meters. Forward modeling results show that when the width is ≥20 meters, the downward pull and weak reflection characteristics of the tidal channel are clearly discernible. Identification threshold: Combining the two models, it is clear that current seismic data can reliably identify tidal channels with a depth ≥3 meters and a width ≥20 meters.

[0105] Figure 3 shows the planar plots of conventional coherence attributes and the most positive and most negative curvature attributes extracted along the top surface of the target layer. Seismic geometric attributes are the basis for characterizing the boundaries and discontinuities of geological bodies. Since the response of a single attribute to complex geological features is limited, comparative analysis of multiple attributes is a necessary method to evaluate its characterization effect. Coherence attribute: Shows a significant response to NW-SE trending faults, only displaying a discontinuous, limited "M"-shaped anomaly in the eastern part of the well, and cannot effectively characterize the continuous distribution of the tidal channel. Most negative curvature attribute: Sensitive to the incised valley floor of the tidal channel, revealing a linear feature approximately 70 km long and trending NW-SE, but its distribution is affected by faults. Most positive curvature attribute: Reflects the uplift of the shoulders on both sides of the tidal channel, indicating a tidal channel width of approximately 350-450 meters.

[0106] Figure 4 shows slices of Euler curvature attributes at different azimuth angles (-90° to 60°). The sensitivity of Euler curvature to the azimuth of surface curvature is the basis for attribute optimization. Since different geological bodies have different optimal response azimuth angles, multi-azimuth scanning and comparison are key methods to select the parameters that best highlight the target body and suppress interference features. Azimuth angles of 30° and 60°: Tidal channels are more clearly displayed than conventional curvature, and have a certain suppressive effect on NW-SE trending faults. Azimuth angle of 45°: The tidal channel has the most continuous linear features and the clearest boundaries, and has the best suppressive effect on fault interference, making it the final selected parameter.

[0107] Figure 4-a Comparison of linear body response of tidal channel under seismic attributes at different azimuth angles;

[0108] Azimuth angle −90°: Schematic diagram of the weakest lateral response of the tidal channel

[0109] Azimuth angle −60°: Schematic diagram of the weak oblique response of the tidal channel

[0110] Azimuth -30°: Schematic diagram of the oblique moderate response of the tidal channel

[0111] Azimuth 0°: Schematic diagram of enhanced near-vertical response in the tidal channel

[0112] Azimuth angle 30°: Schematic diagram of optimal direction response enhancement in the tidal channel ( (indicated)

[0113] Azimuth angle 60°: Schematic diagram of the suboptimal direction response of the tidal channel

[0114] Figure 4-b Comparison of tidal channel response before and after azimuth attribute optimization

[0115] Azimuth angle 30°: Schematic diagram of weak tidal channel response at non-optimal azimuth angles

[0116] Azimuth angle 45°: Schematic diagram showing the clearest tidal channel response at the optimal azimuth angle ( (indicated)

[0117] Azimuth angle 60°: Schematic diagram of reduced response after deviating from the optimal azimuth angle

[0118] Figure 4-c Comparison chart of the method of this invention and traditional curvature properties

[0119] Figure 4-c Comparison of the tide channel recognition effects of the preferred attributes of this invention and traditional curvature attributes

[0120] Figure 4-c (Left): Schematic diagram of linear volume response under Euler curvature property at an azimuth angle of 45°.

[0121] Figure 4-c (Right): Schematic diagram of linear volume response under the most negative curvature property

[0122] Red circle: Areas that are easily identified by breakage, construction noise, or texture misidentification in traditional properties.

[0123] Figure 5The study presents the preferred Euler curvature attribute (azimuth 45°) superimposed on the structural map of the target layer, and the distribution of newly identified traps based on this attribute. The combination of preferred seismic attributes and geological accumulation models serves as the basis for trap identification. Since the identification of lithological traps requires the integration of multiple information sources, overlay analysis of structural, attribute, and geological models is the primary method for effectively delineating targets. Tidal channel distribution: The Euler curvature attribute clearly delineates a main tidal channel extending approximately 110 km in a northeast-southwest direction. New traps: A total of 27 lithological traps, shielded by mudstone, were identified on the flanks of the tidal channel (updating direction), with a total area of ​​80.66 square kilometers, the largest trap area being 11 square kilometers, and an amplitude of 20-75 milliseconds.

[0124] Since earthquake prediction has multiple solutions, well verification is the ultimate method to verify the correctness of hydrocarbon accumulation models and the effectiveness of technical methods. Figure 6 The paper demonstrates the logging interpretation results of wells deployed using this research method: an effective gas layer was interpreted in the target interval. Gas testing results show that the well achieved industrial gas flow after production, with a considerable daily gas production, confirming the effectiveness of the trap and the reliability of the method of this invention. It also verifies the hydrocarbon accumulation model jointly controlled by tidal channels and lateral mudstone blockage, indicating that this oil and gas trap has good exploration prospects.

[0125] Figure 7 The overall structure of a multi-attribute seismic characterization device for tidal channels in a gypsum-carbonate rock coexistence background is shown. This device includes a model building module 1, an attribute optimization module 2, an identification module 3, and a continuous optimization module 4. These modules are connected sequentially according to data flow and functional logic, forming a closed-loop optimized tidal channel identification and evaluation system.

[0126] Among them, the model building module 1 is used to calibrate rock physical parameters based on well logging data and geological knowledge, establish parametric geological models of tidal channels, mudstone shields and surrounding rocks, and obtain seismic response characteristics under different tidal channel widths, thicknesses and filling states through forward modeling, thereby quantitatively determining the seismic identifiable lower limit of tidal channels, providing a physical constraint basis for subsequent attribute optimization and identification.

[0127] The attribute optimization module 2 is used to perform integrated learning analysis on various seismic attributes (including different azimuth attributes, curvature attributes, coherence attributes, etc.). By comprehensively evaluating the attribute response consistency, signal-to-noise ratio and sensitivity to the geometric features of the tidal channel, it adaptively optimizes the attribute combination that is most sensitive to the tidal channel and least sensitive to faults and tectonic noise, thereby achieving optimal fusion of multiple attributes.

[0128] Based on the optimized attribute combination, the identification module 3 identifies and depicts the tidal channels in the study area, extracts the planar morphology, direction, width variation and bifurcation characteristics of the tidal channels, and analyzes the internal structure and lateral shielding relationship of the tidal channels in conjunction with the "tidal channel-mudstone shielding" hydrocarbon accumulation model, thereby screening out favorable lithological traps with sealing and reservoir conditions.

[0129] The continuous optimization module 4 is used to receive drilling empirical data and new geological data, and feed back the actual drilling results (including well logging interpretation, oil testing conclusions, etc.) to the model building module 1 and the attribute optimization module 2 to dynamically correct the original model parameters, attribute weights and identification thresholds, thereby forming a continuous optimization closed loop driven by real geological results, and improving the adaptability and reliability of the method in different blocks and under different geological conditions.

[0130] In summary, the technical solution provided by this invention, based on forward modeling quantitative calibration and seismic attribute optimization, achieves accurate characterization of the spatial distribution of tidal channels and effective evaluation of lithological traps in the context of gypsum-salt-carbonate rock coexistence. The identification and reliability evaluation of tidal channel-mudstone-shrouded traps have significant guiding significance for expanding new areas of oil and gas exploration and achieving large-scale reserve increases.

[0131] For seismic identification and trap prediction of tidal channels, this study utilizes seismic forward modeling technology to analyze the seismic response characteristics of tidal channels with different widths, depths, and reservoir thicknesses. A research approach of "quantitative forward modeling, attribute optimization, and model integration" is proposed. By systematically adjusting geological model parameters and attribute calculation parameters, seismic identification models for tidal channels under different geological conditions are established and compared with actual drilling and oil and gas discoveries for verification. This gradually forms a method for tidal channel identification and trap evaluation that conforms to subsurface geological laws. By constructing a seismic forward model and a multi-attribute fusion analysis process, the seismic response laws of tidal channels under different geometric scales and different surrounding rock contact relationships are clarified. The distribution range of tidal channel-mudstone-shrouded traps is comprehensively judged and quantitatively predicted, ultimately aiming to improve the success rate of exploring hidden oil and gas reservoirs and effectively guide drilling deployment.

[0132] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0133] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the core ideas and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for multi-attribute characterization of seismic events in tidal channels under a gypsum-carbonate rock coexistence background, characterized in that, The method includes the following steps: Based on well logging data, the petrophysical parameters corresponding to different lithologies are calibrated, a parametric geological model is constructed, and a seismic forward modeling is performed to quantitatively determine the lower limit of identifiability of tidal channels under the current seismic data conditions. Under the constraint of the recognition lower limit, multi-attribute feature extraction and fusion are performed on the seismic data to complete the seismic identification of the tidal channel; The identified tidal channels are analyzed in conjunction with structural conditions and mudstone plugging conditions to screen and quantitatively evaluate favorable traps that can form effective plugging. The favorable traps are verified by drilling, and the real geological information obtained from drilling is fed back to the aforementioned model construction and attribute identification process, forming a closed-loop optimized tidal channel characterization process.

2. The method as described in claim 1, characterized in that, The parametric geological model is constructed by changing the width, cutting depth, and reservoir thickness of the tidal channel, and the seismic response under different geological scenarios is simulated using the wave equation forward modeling to determine that the minimum identifiable width of the tidal channel on the seismic profile is 20 meters and the minimum identifiable cutting depth is 3 meters.

3. The method as described in claim 1, characterized in that, The quantitative evaluation of favorable traps includes at least the calculation and comprehensive judgment of the trap area, closure amplitude, and the relationship between the reservoir and mudstone sealing, so as to distinguish tidal channel traps with drilling value.

4. A method for identifying multiple attributes of seismic events in tidal channels under a gypsum-carbonate rock coexistence background based on identification lower limit constraints, characterized in that... The method includes: The lower limit for identifying earthquakes in tidal channels, determined by forward modeling, is used as a constraint condition. Under the aforementioned constraints, multi-dimensional seismic attribute features are extracted from seismic data; The seismic attributes are adaptively filtered and fused using ensemble learning to output the spatial distribution results of tidal channels that meet the lower limit of recognition constraints.

5. The method as described in claim 4, characterized in that, The earthquake attribute features include at least amplitude features, geometric features, texture features and spectral features, and the integrated learning method completes the tidal channel identification based on the collaborative decision-making of multiple identification results.

6. The method as described in claim 4, characterized in that, While outputting the tidal channel identification results, a confidence score is generated for each identification result to characterize the reliability of the results in terms of spatial continuity and geological rationality.

7. A continuous optimization method for seismic characterization of tidal channels in a gypsum-carbonate rock coexistence background, characterized in that, The method includes: Tidal channel traps identified based on seismic multi-attribute identification and hydrocarbon accumulation analysis will be deployed as drilling targets. To obtain well logging interpretation results and oil and gas display information generated during the drilling process; Using the real geological information obtained from the drilling, the previously constructed forward model and seismic attribute identification process are reverse-corrected to continuously improve the accuracy of tidal channel characterization.

8. The method as described in claim 7, characterized in that, The reverse correction involves using the lithological distribution and reservoir characteristics revealed by drilling to correct the rock physical parameters and simultaneously update the forward simulation model.

9. The method as described in claim 7, characterized in that, The reverse correction also includes inputting drilling results as new constraint samples into the seismic attribute identification process to optimize attribute combinations and identification rules.

10. A device for characterizing seismic properties of tidal channels in a gypsum-carbonate rock coexisting background for implementing the method according to any one of claims 1 to 9, characterized in that, include: The model building module is used to calibrate rock physical parameters based on well logging data and perform forward modeling to determine the lower limit of seismic identification in tidal channels. The attribute recognition module, connected to the model building module, is used to complete the screening, fusion, and tidal channel identification of multiple seismic attributes under the recognition lower limit constraint. The trap evaluation module, connected to the attribute recognition module, is used to screen and quantitatively evaluate the identification results based on mudstone plugging conditions. The feedback optimization module, connected to the trap evaluation module, is used to continuously correct and optimize the aforementioned modules based on the real geological information obtained from drilling.

Images