Underground fracture prediction method and device based on multi-scale information fusion

Abstract

The invention discloses an underground crack prediction method and device based on multi-scale information fusion. The method comprises the following steps: acquiring a construction pattern of a research area; obtaining a fracture analysis result of the research area, and obtaining fracture development characteristics and a fracture density curve of each single well; performing model forward modeling based on the seismic data of the research area to determine a fracture grading standard; based on the post-stack seismic data of the research area, obtaining large-scale and medium-scale fractures according to seismic attributes; according to the post-stack seismic data, obtaining a small-scale crack based on a preset crack prediction model; based on the pre-stack seismic data of the research area and the fracture density curve of each single well, obtaining a micro-scale fracture according to a pre-stack azimuth anisotropic fracture prediction technology; and taking the large and medium-scale fractures, the small-scale fractures and the micro-scale fractures as input variables, and fusing the input variables by adopting a fuzzy neural network to obtain a fracture prediction data volume. A complex network and a multi-information fusion method are combined with a big data technology to carry out high-precision crack identification so as to realize multi-scale characterization of shale cracks.

Description

Technical Field

[0001] The present invention relates to the field of geophysical exploration technology, and in particular to a method and device for predicting underground cracks using multi-scale information fusion. Background Art

[0002] Currently, China's unconventional oil and gas development still lags behind internationally advanced levels in key core technologies, efficiency, and cost. Continuous research is needed to rapidly and accurately identify "sweet spots / intervals," precisely deploy wells, and effectively stimulate reservoirs. This will significantly increase single-well production and recovery rates, enabling large-scale production and effective development of unconventional oil and gas resources under low oil prices. The volumetric fracturing of shale oil and gas reservoirs, and the volume of reservoir stimulation, directly impacts shale oil and gas production. Shale reservoir stimulation involves the use of high-pressure fracturing fluids to create new fractures in the shale reservoir, connecting existing natural fractures and forming a complex fracture network. The connectivity between artificial fractures and natural fractures, as well as between natural fractures, in shale reservoirs significantly impacts reservoir productivity. Research into how shale volumetric fracturing is initiated and the resulting fracture morphologies is crucial.

[0003] Shale reservoirs have well-developed bedding, significant anisotropy, and complex natural fracture systems, making the propagation mechanism of hydraulic fracturing even more complex. Domestic and international researchers have conducted extensive research on shale fracture characterization and propagation mechanisms, numerical simulations, fracturing tests, and microseismic monitoring, achieving a series of research results in shale fracture characterization and the propagation mechanism of hydraulic fracturing. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method and device for predicting underground fractures based on multi-scale information fusion. Through interdisciplinary integration, complex networks and multi-information fusion methods are combined with big data technology to characterize the distribution pattern and development degree of shale fracture networks, as well as to depict the dynamic expansion of shale fracture networks. This overcomes the problems of the current commonly used methods such as lack of systematicity, single information utilization, and relatively one-sided results. It better solves the cross-scale problem of seismic and geological data, and provides a new method for realizing multi-scale characterization of shale fractures and studying the dynamic expansion mechanism of fractures.

[0005] To solve the above technical problems, in a first aspect, an embodiment of the present application provides a method for predicting underground cracks using multi-scale information fusion, comprising the following steps:

[0006] Obtain the structural style of the study area based on geological data;

[0007] Obtaining fracture analysis results of a study area determined based on well logging data, and obtaining fracture density curves of each single well based on the fracture analysis results;

[0008] According to the structural style and fracture analysis of the study area, a model forward modeling is performed based on the seismic data of the study area to obtain a model forward modeling result; the seismic data includes pre-stack seismic data and post-stack seismic data;

[0009] Based on the forward modeling results of the model, a multi-scale fracture classification standard is determined, and the fractures are classified into large-scale fractures, small-scale fractures and micro-scale fractures;

[0010] Based on the post-stack seismic data of the study area and taking the seismic fracture interpretation results as the standard, the large and medium-scale fractures are obtained according to seismic attributes;

[0011] According to the post-stack seismic data, small-scale fractures are obtained based on the preset fracture prediction model;

[0012] Based on the pre-stack seismic data of the study area and the fracture density curves of each single well, micro-scale fractures are obtained according to the pre-stack azimuthal anisotropic fracture prediction technology;

[0013] The large and medium scale cracks, the small scale cracks and the micro scale cracks are used as input variables and fused using a fuzzy neural network to obtain a multi-scale fused crack prediction data volume.

[0014] In one or some optional implementations of the embodiment of the present application, determining a multi-scale fracture classification standard based on the forward modeling results of the model and classifying the fractures into large-scale and medium-scale fractures, small-scale fractures, and micro-scale fractures includes the following steps:

[0015] According to the structural style and the fracture analysis, geological models with different fault throws are established respectively, and theoretical seismic waveforms are obtained by convolving seismic wavelets with different main frequencies with the formation reflection coefficient;

[0016] The cracks are classified according to the dominant frequency of the seismic data. If the image shows that the seismic phase axis is obviously dislocated or bent and can be identified with the naked eye on the seismic section, it is a large-scale or medium-scale crack; if the image shows that the seismic phase axis is slightly folded or there is a slight change in the lateral energy and there is no effective fault throw on the seismic section, it is a small-scale crack; if the image cannot be identified with the naked eye on the seismic section, it is a microscale crack.

[0017] In one or some optional implementations of the embodiment of the present application, based on the post-stack seismic data of the study area, predicting large and medium-scale fractures according to seismic attributes includes the following steps:

[0018] Based on the post-stack seismic data of the study area, the large and medium-scale fractures are characterized using geometric seismic attributes, including curvature, coherence, dip, and ant body, and verified using the fault interpretation results of the study area, wherein the fault interpretation results of the study area are determined based on the geological data of the study area.

[0019] In one or some optional implementations of the embodiment of the present application, obtaining small-scale fractures based on the post-stack seismic data and a preset fracture prediction model includes the following steps:

[0020] Based on the anisotropic diffusion theory, the post-stack seismic data of the study area are subjected to fault enhancement processing to obtain a seismic profile;

[0021] Based on a preset fracture label database, a U-net convolutional neural network algorithm is used for training to obtain the preset crack prediction model;

[0022] The seismic profile is segmented based on the preset crack prediction model to obtain small-scale cracks.

[0023] In one or some optional implementations of the embodiment of the present application, obtaining microscale fractures based on the pre-stack seismic data of the study area and the fracture density curves of each single well according to the pre-stack azimuthal anisotropic fracture prediction technology includes the following steps:

[0024] Performing azimuth division based on the pre-stack seismic data of the study area to form a plurality of azimuth gathers;

[0025] Extracting frequency, amplitude, and attenuation attributes from each of the azimuth gathers, and performing optimal attribute selection;

[0026] Performing azimuthal ellipse fitting, wherein the major axis direction of the ellipse is the direction of the microscale crack;

[0027] The ellipse flattening is fitted with the single well fracture density curve to obtain the density of the microscale fractures.

[0028] In one or some optional implementations of the embodiment of the present application, obtaining the structural pattern of the study area determined based on geological data includes the following steps:

[0029] Conduct geological simulation based on geological profile and obtain dynamic data;

[0030] By analyzing outcrops and cores, the overall structural pattern, lithology, and stress state of the study area were established;

[0031] The fracture development state is obtained by combining the dynamic data, the fracture parameters are counted, the distribution of the fractures in the formation corresponding to the core is predicted and calibrated, and the structural style is established, which includes normal faults, reverse faults and strike-slip faults.

[0032] In one or some optional implementations of the embodiment of the present application, obtaining the fracture analysis results of the study area determined based on the well logging data, and obtaining the fracture density curve of each single well according to the fracture analysis results, includes the following steps:

[0033] Obtaining the logging data based on imaging logging technology and conventional logging data;

[0034] Performing the fracture analysis based on the well logging data, wherein the fracture analysis includes the inclination, tendency, and filling condition of the fracture;

[0035] Fracture grouping is performed based on the fracture analysis, and fracture densities in different directions are calculated in combination with well trajectory data in the conventional logging data to form the single well fracture density curve.

[0036] In a second aspect, an embodiment of the present application provides a device for predicting underground cracks using multi-scale information fusion, comprising:

[0037] The first acquisition module is used to obtain the structural style of the study area determined based on geological data;

[0038] The first analysis module is used to obtain the fracture analysis results of the study area determined based on the well logging data, and obtain the fracture density curve of each single well according to the fracture analysis results;

[0039] A model forward modeling module is used to perform model forward modeling based on the structural style and fracture analysis of the study area and the seismic data of the study area to obtain model forward modeling results and obtain multi-scale fracture classification standards; the seismic data includes pre-stack seismic data and post-stack seismic data;

[0040] The first fracture determination module is used to obtain large and medium-scale fractures based on the post-stack seismic data of the study area and the seismic fault interpretation results according to the seismic attributes;

[0041] The second fracture determination module is used to obtain small-scale fractures based on the post-stack seismic data and a preset fracture prediction model;

[0042] A third fracture determination module is configured to obtain microscale fractures based on the pre-stack seismic data of the study area and the fracture density curves of each single well according to the pre-stack azimuthal anisotropic fracture prediction technology;

[0043] The fusion module is used to take the large and medium scale cracks, the small scale cracks and the micro scale cracks as input variables, and fuse them using a fuzzy neural network to obtain a multi-scale fused crack prediction data body.

[0044] In a third aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the multi-scale information fusion method for predicting underground cracks.

[0045] In a fourth aspect, an embodiment of the present application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0046] Memory for storing computer programs;

[0047] The processor is used to implement the underground crack prediction method based on multi-scale information fusion when executing the program stored in the memory.

[0048] The beneficial effects of the above technical solutions provided by the embodiments of the present application include at least:

[0049] The embodiment of the present application provides a method for predicting underground fractures by multi-scale information fusion, which obtains the structural style of the study area determined based on geological data; obtains the fracture analysis results of the study area determined based on well logging data, and obtains the fracture density curve of each single well according to the fracture analysis results; performs model forward modeling based on the seismic data of the study area according to the structural style and fracture analysis of the study area, and obtains the model forward modeling results; the seismic data includes pre-stack seismic data and post-stack seismic data; determines the multi-scale fracture classification standard based on the model forward modeling results; obtains large and medium-scale fractures based on post-stack seismic attributes; obtains small-scale fractures based on the post-stack seismic data and a preset fracture prediction model; obtains micro-scale fractures based on the pre-stack seismic data of the study area and the fracture density curves of each single well according to the pre-stack azimuth anisotropic fracture prediction technology; uses large and medium-scale fractures, small-scale fractures and micro-scale fractures as input variables, and adopts fuzzy neural networks for fusion to obtain a fracture prediction data body. The present invention aims to achieve multi-scale characterization of shale fractures through the multi-information, cross-disciplinary integration of geology, seismology, and well logging, and the use of complex networks and multi-information fusion methods combined with big data technology to carry out high-precision fracture identification, including the distribution pattern and development degree of shale fracture networks, and the characterization of the dynamic expansion of shale fracture networks. It overcomes the problems of the current commonly used methods such as lack of systematicity, single information utilization, and relatively one-sided results, and better solves the cross-scale problems of seismic and geological data, promotes the study of the dynamic expansion mechanism of fractures, and provides technical support for the large-scale exploration and development of shale oil and gas.

[0050] Other features and advantages of the present application will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purposes and other advantages of the present application can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0051] The technical solution of the present application is further described in detail below through the accompanying drawings and examples. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings are used to provide a further understanding of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings:

[0053] Figure 1 A schematic diagram of the process of underground crack prediction method based on scale information fusion provided in an embodiment of the present application;

[0054] Figure 2 Another schematic diagram of the process of underground crack prediction using scale information fusion provided in an embodiment of the present application;

[0055] Figure 3 A schematic diagram of the results of the forward modeling provided in the embodiment of the present application;

[0056] Figure 4 A structural diagram of a convolutional neural network algorithm based on U-net provided in an embodiment of the present application;

[0057] Figure 5 A flow chart of the pre-stack anisotropic fracture prediction technology provided in an embodiment of the present application;

[0058] Figure 6 A fuzzy neural network structure diagram provided in an embodiment of the present application;

[0059] Figure 7 A schematic diagram of the structure of a device for predicting underground cracks using multi-scale information fusion provided in an embodiment of the present application;

[0060] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0061] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0062] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0063] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0064] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0065] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0066] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0067] It should be understood that the size of the serial numbers of the steps in the following embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0068] The inventors discovered that, in the prior art, domestic and international scholars have conducted extensive research on shale fracture characterization and expansion mechanisms, numerical simulations, hydraulic fracturing tests, and microseismic monitoring. While a series of research results have been achieved in shale fracture characterization and the study of hydraulic fracture expansion mechanisms, deficiencies remain in fracture characterization. The challenge of achieving cross-scale fusion characterization by utilizing multidisciplinary geological engineering data and research results at different scales remains urgent. Multi-information fusion is precisely the theoretical approach of processing multi-source information to obtain comprehensively usable information. Introducing complex networks can characterize both the macroscopic characteristics of the network and the analysis of network topology at different scales. This approach can effectively characterize the distribution patterns and development level of shale fracture networks, as well as depict the dynamic expansion of shale fracture networks.

[0069] In order to solve the problems of common methods such as lack of systematicity, single information used, and relatively one-sided results, and to solve the cross-scale problems of seismic and geological data, the inventors have made this application after further research and development, providing a method and device for underground crack prediction with multi-scale information fusion.

[0070] Example 1

[0071] The embodiment of the present application starts with a comprehensive analysis of geological cores, analyzes and identifies the fractures in the study area, clarifies the fracture development characteristics and their matching relationship with tectonic activities, and determines the fracture development and planar distribution characteristics; based on the geological fracture identification mode and structural style, seismic fracture prediction is carried out, and fractures are divided into identifiable large and medium-scale fractures and difficult-to-identify small and micro-scale fractures. With the fracture description results of well logging as a constraint, a study of seismic-based pre-stack and post-stack multi-attribute fracture prediction methods is carried out; in order to effectively improve the credibility of the data and reduce the multi-solution of fracture prediction, multi-information fusion technology is carried out by integrating multiple disciplines such as geology, seismology, and well logging to finely characterize the distribution pattern and development degree of the underground fracture network.

[0072] See Figure 1 and Figure 2 The specific implementation steps are:

[0073] S1. Obtain the structural style of the study area based on geological data.

[0074] In step S1, a geological simulation is first performed based on the geological overview to obtain dynamic data. Then, by analyzing outcrops and cores, the overall structural pattern, lithology, and stress state of the study area are established. Combined with the dynamic data, the fracture development state is determined, and fracture parameters are calculated. The distribution of fractures in the strata corresponding to the cores is then predicted and calibrated, and structural patterns are established. These patterns include normal faults, reverse faults, and strike-slip faults.

[0075] Furthermore, the interpretation and detailed characterization of faults are crucial for oil and gas development. Numerous studies have shown that fault systems significantly influence the formation and preservation of shale gas reservoirs: large faults can cause some degree of damage to oil and gas reservoirs, small faults can effectively improve reservoir formations, and micro-scale fractures are important reservoir spaces. Therefore, after establishing the structural pattern, we further analyzed the fault development pattern and combined it with high-precision seismic data to conduct fault interpretation and determine the fault distribution characteristics within the study area.

[0076] S2. Obtain fracture analysis results of the study area determined based on well logging data, and obtain fracture density curves of each single well based on the fracture analysis results.

[0077] Logging data is obtained using imaging logging technology and conventional logging data. Fracture analysis is performed based on this logging data to identify fracture inclination, orientation, and filling status. Fractures are grouped based on this analysis and, combined with well trajectory data from conventional logging data, fracture densities in different orientations are calculated to form a single-well fracture density curve. By statistically analyzing fracture development patterns in core and imaging logging curves, the primary fracture development pattern in the study area is determined. Fracture locations are identified by combining the development characteristics of fractures of different occurrences. Imaging logging technology can utilize FMI imaging logging.

[0078] S3. According to the structural style and fracture analysis of the study area, the model forward modeling is performed based on the seismic data of the study area to obtain the model forward modeling results; the seismic data includes pre-stack seismic data and post-stack seismic data.

[0079] See Figure 3 , according to the structural style and fracture analysis, the geological models with different fault throws are established respectively ( Figure 3 The first figure in the figure), using different main frequencies (such as Figure 3 The seismic wavelets with main frequencies of 30HZ, 40HZ, 50HZ, 60HZ, and 70HZ are convolved with the formation reflection coefficient to obtain the model forward modeling results of the theoretical seismic waveform.

[0080] S4. Based on the forward modeling results, the multi-scale fracture classification standard is determined, and the fractures are classified into large-scale and medium-scale fractures, small-scale fractures, and micro-scale fractures.

[0081] See Figure 3Since different frequency data have different response characteristics to different fault throws, the higher the seismic frequency, the stronger the identification energy of small faults and cracks. According to the main frequency of actual seismic data, cracks are classified as follows: (1) Large and medium-scale cracks: the seismic phase axis is obviously dislocated or bent, and the cracks can be identified by the naked eye on the seismic profile; (2) Small-scale cracks: the seismic phase axis is slightly folded or there is a slight change in lateral energy, and there is no effective fault throw on the seismic profile; (3) Micro-scale cracks: natural crack zones that cannot be identified by the naked eye on the seismic profile. Specifically, when the main frequencies are 30HZ, 40HZ, and 50HZ, it can be seen that the seismic phase axis can still be found to be obviously dislocated at 10m, while there is no obvious dislocation on the seismic profile below 10m. Therefore, under these three main frequencies, cracks of 10m and above are large and medium-scale; similarly, when the main frequencies are 60HZ and 70HZ, cracks of 5m and above are large and medium-scale.

[0082] S5. Based on the post-stack seismic data of the study area and the seismic fault interpretation results, large and medium-scale fractures are obtained according to seismic attributes.

[0083] Conventional geometric seismic attributes such as curvature, coherence, dip, and ant body can be used on post-stack seismic data to characterize the development characteristics of large and medium-scale fractures, and verified by the earthquake fault interpretation results in the study area.

[0084] Curvature is a parameter that reflects the degree of bending of a curve or surface. The greater the structural principal curvature of a surface, the more curved it is and the easier it is to produce cracks. Therefore, the structural principal curvature reflects the distribution of cracks to a certain extent. When the curvature increases to its elastic limit, cracks will appear in areas with larger curvatures. Therefore, some high-curvature areas such as both sides of the fold axis, structural turning points, and both sides of the fault surface are often areas where cracks develop.

[0085] Because seismic waves are distorted at faults and fissures, faults and fractures can be effectively predicted by measuring the similarity and discontinuity between adjacent seismic data traces. Three coherence algorithms can be used, depending on the signal-to-noise ratio of the seismic data and the stability of the algorithm: cross-correlation-based, trace-similarity-based, and eigenvalue-based. Cross-correlation-based algorithms offer fast computational speed and minimal computational effort, but suffer from poor noise immunity, poor stability, and relatively low resolution. Algorithms based on trace similarity offer the advantages of strong anti-interference capabilities and higher resolution, but suffer from heavy computational effort, high computer memory requirements, and low lateral resolution. Eigenvalue-based algorithms offer higher resolution but also require more computation.

[0086] By detecting dip angles, faults can be identified by exploiting the difference in dip angle between the horizon and the fault plane. Where small faults exist, the dip angle will undergo a sudden, instantaneous change, resulting in a regular, sudden change band on the attribute slice or profile. When this change appears linear, it is considered a manifestation of a small fault. When interpreting seismic data at a horizon, if the continuity of the reflection symmetric axes is good, interpreters generally assume that the stratum contains no faults or fractures. However, due to the inherent resolution of the seismic data, small fractures may often exist in areas where the symmetric axes are distorted. These small fractures can be effectively identified using dip angle detection technology.

[0087] Ant-volume characterization of large and medium-scale fractures uses an ant colony algorithm to automatically track and identify fracture systems in seismic data. The principle is as follows: First, a large number of "ants" are pre-assigned to each segment of the seismic data according to a certain probability. The pre-assigned probability is related to the value of the data volume to be tracked. During the tracking process, the ants randomly select their tracking direction. The information they can identify for each sample point includes pheromone concentration and heuristic information. Once a fault meeting pre-assigned criteria is discovered, they automatically track the fracture system and simultaneously leave pheromones to update their information. Just like ants foraging for food, the pheromones left by the ants during the tracking process attract more ants. The search terminates when the pheromone concentration exceeds a pre-assigned threshold or when an initial termination condition is met, until the entire fracture system is identified. Compared to effective seismic information, noise in seismic profiles is always irregular, has lower energy, and is shorter in extent. The pheromones left by the ants during their tracking are often minimal, making the noise difficult to identify. Therefore, the ant colony algorithm ultimately produces an ant-volume with a high signal-to-noise ratio and a clear fracture system.

[0088] S6. Obtain small-scale fractures based on the post-stack seismic data and the preset fracture prediction model.

[0089] For small-scale cracks that cannot be effectively identified on seismic profiles, micro-fracture detection technology based on deep learning can be used for detection. Figure 4This technology uses a convolutional neural network algorithm based on U-net to segment images, achieving fully automatic supervised neural network learning based on big data. Specifically, first, to improve the imaging accuracy of fractures and overcome interference from non-geological factors such as background noise and acquisition errors, fault enhancement processing is performed on post-stack seismic data based on anisotropic diffusion theory. The above-mentioned large and medium-scale fracture prediction data volume is added as input on the basis of structural guidance filtering. When encountering larger faults or fractures, smoothing or filtering is only performed along the stratigraphic direction, and as little as possible is not performed perpendicular to the stratigraphic direction. This increases the imaging accuracy of the fracture and makes the cross-section clearer and crisper, effectively overcoming the damage to fractures caused by conventional filtering methods. Secondly, based on a preset fracture label database, the convolutional neural network algorithm of U-net is used for training to obtain a preset fracture prediction model. Finally, based on this preset fracture prediction model, the filtered seismic profile is segmented to achieve automatic prediction of small fractures. U-Net is a commonly used image segmentation network. The network structure is "U"-shaped and symmetrical and uses skip connections. The left half is the encoder for feature extraction, and the right half is decoded through upsampling to achieve an end-to-end effect. This connection method enables the decoding process to splice the feature vectors of the symmetric layers of the encoding process, fusing effective features at all scales, that is, combining features in low-level and high-level feature maps, avoiding computational errors and improving segmentation accuracy. Ultimately, it realizes fully automatic neural network learning based on big data and effectively identifies small-scale cracks.

[0090] S7. Based on the pre-stack seismic data of the study area and the fracture density curves of each single well, micro-scale fractures are obtained according to the pre-stack azimuthal anisotropic fracture prediction technology.

[0091] Because pre-stack seismic data contains more information than post-stack seismic data, pre-stack seismic azimuthal anisotropy technology can effectively predict micro-scale fractures that cannot be identified on seismic profiles. The pre-stack azimuthal fracture detection method is a seismic detection method based on longitudinal waves. When seismic P waves are reflected by fractured strata, the reflections produced are different due to the different azimuths between the P waves and the fractures. By utilizing the wide azimuth characteristics of 3D seismic data, the seismic P wave response characteristics at different azimuths can be extracted to detect the relative degree of fracture development. This method is particularly effective for high-angle fractures. Figure 5The specific approach is as follows: first, pre-stack CMP gather data analysis and azimuth division are carried out. Due to the almost non-existence of azimuthal anisotropy characteristics at zero offset or small offset, the coverage times of each azimuth at large offset are unevenly distributed and the data signal-to-noise ratio is low. Therefore, seismic traces with small offset and large offset are removed to ensure that the coverage times of each azimuth gather are relatively uniform and consistent. Based on the development characteristics of fracture strike and dip obtained from geological core and logging fracture analysis, the azimuth division scheme is determined while ensuring uniform distribution of coverage times and offset distance. Secondly, since the direction and density of fractures will have a great influence on seismic waves, the propagation speed of seismic waves in different azimuths is different. For example, the speed of seismic waves propagating along the direction of the fractures is the fastest, and the speed of seismic waves propagating perpendicular to the fractures is the slowest. Then, three-dimensional seismic attributes such as amplitude, frequency, and attenuation are extracted, and the optimal attributes are selected. The characteristics of three-dimensional seismic attributes changing with azimuth are used to fit the azimuthal ellipse, and the ellipse flattening and direction are calculated. The ellipse is then fitted with the fracture density results of well logging analysis, and the degree and direction of fracture development in the study area are finally obtained to provide a comprehensive description of the fracture density and direction.

[0092] S8. Take large-scale and medium-scale cracks, small-scale cracks and micro-scale cracks as input variables, and use fuzzy neural network to fuse them to obtain a multi-scale fusion crack prediction data body.

[0093] Based on the relationship between oil and gas reservoirs and fractures, as well as their geological characteristics, the obtained large-scale, medium-scale, small-scale, and microscale fractures were used as input for multi-scale information fusion using fuzzy neural network technology. Fuzzy neural networks are a machine learning algorithm that uses neural network approximation techniques to find the parameters of a fuzzy system (i.e., fuzzy sets and fuzzy rules). Since both fuzzy systems and neural networks are nonlinear input-output mappings, in principle, fuzzy systems can be equivalently represented by neural networks. The structure of the fuzzy system determines the equivalent neural network structure, where each layer and each node of the neural network corresponds to a part of the fuzzy system. Therefore, unlike conventional black-box neural networks, all parameters of this neural network have physical meaning. The neural network derived from the fuzzy system through equivalent transformation is called a fuzzy neural network (FNN). This FNN is completely equivalent to the fuzzy system in terms of input and output ports, while its internal weights and node parameters can be modified through neural network learning. This technology combines the powerful structural knowledge representation capabilities of fuzzy logic reasoning with the robust self-learning capabilities of neural networks, effectively leveraging their respective strengths while addressing their respective shortcomings.

[0094] Specifically, see Figure 6A fuzzy neural network consists of five layers. Designed based on the working process of a fuzzy system, it is a fuzzy inference system implemented using a neural network: The first layer is the input layer, with the number of nodes equal to the number of input variables. The second layer is the membership function layer for the input variables, which implements the fuzzification of the input variables. The third layer is the "AND" layer, with the number of nodes equal to the product of the number of fuzzy sets for each input variable (i.e., the number of rules). The fourth layer is the "OR" layer, with the number of nodes equal to the number of fuzzy partitions of the output variable. Each node represents a fuzzy partition (i.e., a fuzzy set) of the output variable, and its output is the membership function value of each fuzzy degree of the output variable. The third and fourth layers can also be collectively referred to as the fuzzy inference layer. The fifth layer is the defuzzification layer, or the "clarification layer," with the number of nodes equal to the number of output variables. This layer implements the "clarification" operation, which converts the membership values ​​of the fuzzy sets of the output variables obtained by fuzzy rule reasoning (i.e., the outputs of each node in the fourth layer) into precise numerical values ​​of the output variables. In this example, based on the relationship between oil and gas reservoirs and fractures and their geological characteristics, multi-scale prediction data representing fractures is quantified as input. Fuzzy transformation principles are then used to comprehensively evaluate and analyze various attributes, ultimately generating a fracture prediction data volume. Multi-information fusion is a theoretical approach for processing multi-source information to generate comprehensively usable information. The introduction of fuzzy neural networks can characterize both the macroscopic characteristics of the network and the analysis of network topology at different scales. This approach effectively characterizes the distribution patterns and development level of shale fracture networks, as well as the dynamic expansion of these networks.

[0095] Through the embodiments of the present application, multi-disciplinary integration of geological, seismic, and well logging information is achieved, and complex convolutional neural networks and multi-information fusion technology based on fuzzy optimization theory are used to achieve high-precision identification and multi-scale prediction of fractures. This effectively solves the cross-scale problem of seismic and geological data, effectively characterizes the distribution pattern and development degree of shale fracture networks, and provides a new method for realizing multi-scale characterization of shale fractures and studying the dynamic expansion mechanism of fractures, providing technical support for the large-scale exploration and development of shale oil and gas.

[0096] Example 2

[0097] Based on the same inventive concept, the present application provides a multi-scale information fusion underground crack prediction device, see Figure 7 , specifically including:

[0098] The first acquisition module 101 is used to obtain the structural style of the study area determined based on geological data;

[0099] The first analysis module 102 is used to obtain the fracture analysis results of the study area determined based on the well logging data, and obtain the fracture density curve of each single well according to the fracture analysis results;

[0100] The model forward modeling module 103 is used to perform model forward modeling based on the seismic data of the study area according to the structural pattern and fracture analysis of the study area, obtain model forward modeling results, and obtain multi-scale fracture classification standards; the seismic data includes pre-stack seismic data and post-stack seismic data;

[0101] The first fracture determination module 104 is used to obtain large and medium-scale fractures based on seismic attributes based on the post-stack seismic data of the study area and the seismic fault interpretation results;

[0102] The second fracture determination module 105 obtains small-scale fractures based on the post-stack seismic data and a preset fracture prediction model;

[0103] The third fracture determination module 106 is used to obtain micro-scale fractures based on the pre-stack seismic data of the study area and the fracture density curves of each single well according to the pre-stack azimuthal anisotropic fracture prediction technology;

[0104] The fusion module 107 is used to take large-scale and medium-scale cracks, small-scale cracks and micro-scale cracks as input variables and fuse them using a fuzzy neural network to obtain a multi-scale fused crack prediction data body.

[0105] The implementation principle and technical effects of the device for predicting underground cracks using multi-scale information fusion provided in an embodiment of the present invention are similar to those of any of the aforementioned method embodiments and will not be described in detail here.

[0106] Example 3

[0107] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the method for predicting underground cracks by multi-scale information fusion described in any of the aforementioned embodiments is implemented.

[0108] The computer-readable storage medium may be included in the device / apparatus described in the above embodiments, or may exist independently without being incorporated into the device / apparatus. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of the present invention.

[0109] According to an embodiment of the present invention, a computer-readable storage medium may be a non-volatile computer-readable storage medium, such as, but not limited to, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0110] Example 4

[0111] Based on the same inventive concept, the present application provides an electronic device, referring to Figure 8 , including a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112 and the memory 113 communicate with each other through the communication bus 114,

[0112] Memory 113, for storing computer programs;

[0113] The processor 111 is configured to implement the underground crack prediction method based on multi-scale information fusion described in any of the aforementioned embodiments when executing the program stored in the memory 113 .

[0114] The implementation principle and technical effects of the electronic device provided by the embodiment of the present invention are similar to those of any of the aforementioned method embodiments and will not be repeated here.

[0115] The memory 113 can be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read-Only Memory), an EPROM, a hard disk, or a ROM. The memory 113 has storage space for program code for executing any of the method steps described above. For example, the storage space for program code can include individual program codes for implementing each of the steps in the method described above. These program codes can be read from or written to one or more computer program products. These computer program products include program code carriers such as a hard disk, a compact disc (CD), a memory card, or a floppy disk. Such computer program products are typically portable or fixed storage units. The storage unit can have storage segments or storage space arranged similarly to the memory 113 in the electronic device described above. The program code can be compressed, for example, in a suitable form. Typically, the storage unit includes a program for executing the method steps according to an embodiment of the present invention, i.e., code that can be read by, for example, the processor 111, and when executed by the electronic device, causes the electronic device to execute the various steps in the method described above.

[0116] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for predicting underground cracks based on multi-scale information fusion, characterized in that: The following steps are involved: Obtain the structural style of the study area based on geological data; Obtaining fracture analysis results of a study area determined based on well logging data, and obtaining fracture density curves of each single well based on the fracture analysis results; According to the structural style and fracture analysis of the study area, a model forward modeling is performed based on the seismic data of the study area to obtain a model forward modeling result; the seismic data includes pre-stack seismic data and post-stack seismic data; Based on the forward modeling results of the model, a multi-scale fracture classification standard is determined, and the fractures are classified into large-scale fractures, small-scale fractures and micro-scale fractures; Based on the post-stack seismic data of the study area and taking the seismic fracture interpretation results as the standard, the large and medium-scale fractures are obtained according to seismic attributes; Obtaining small-scale fractures based on the post-stack seismic data and a preset fracture prediction model; Based on the pre-stack seismic data of the study area and the fracture density curves of each single well, the micro-scale fractures are obtained according to the pre-stack azimuthal anisotropic fracture prediction technology; The large and medium scale cracks, the small scale cracks and the micro scale cracks are used as input variables and fused using a fuzzy neural network to obtain a multi-scale fused crack prediction data volume.

2. The underground crack prediction method based on multi-scale information fusion according to claim 1, characterized in that: The method of determining a multi-scale fracture classification standard based on the forward modeling results of the model and classifying the fractures into large-scale and medium-scale fractures, small-scale fractures, and micro-scale fractures comprises the following steps: According to the structural style and the fracture analysis, geological models with different fault throws are established respectively, and theoretical seismic waveforms are obtained by convolving seismic wavelets with different main frequencies with the formation reflection coefficient; The cracks are classified according to the dominant frequency of the seismic data. If the image shows that the seismic phase axis is obviously dislocated or bent and can be identified with the naked eye on the seismic section, it is a large-scale or medium-scale crack; if the image shows that the seismic phase axis is slightly folded or there is a slight change in the lateral energy and there is no effective fault throw on the seismic section, it is a small-scale crack; if the image cannot be identified with the naked eye on the seismic section, it is a microscale crack.

3. The underground crack prediction method based on multi-scale information fusion according to claim 1, characterized in that: Based on the post-stack seismic data of the study area, large and medium-scale fractures are predicted according to seismic attributes, including the following steps: Based on the post-stack seismic data of the study area, the large and medium-scale fractures are characterized using geometric seismic attributes, including curvature, coherence, dip, and ant body, and verified using the fault interpretation results of the study area, wherein the fault interpretation results of the study area are determined based on the geological data of the study area.

4. The underground crack prediction method based on multi-scale information fusion according to claim 1, characterized in that: The method of obtaining small-scale fractures based on the post-stack seismic data and a preset fracture prediction model comprises the following steps: Based on the anisotropic diffusion theory, the post-stack seismic data of the study area are subjected to fault enhancement processing to obtain a seismic profile; Based on a preset fracture label database, a U-net convolutional neural network algorithm is used for training to obtain the preset crack prediction model; The seismic profile is segmented based on the preset crack prediction model to obtain small-scale cracks.

5. The underground crack prediction method based on multi-scale information fusion according to claim 1, characterized in that: The method of obtaining microscale fractures based on the pre-stack seismic data of the study area and the fracture density curves of each single well according to the pre-stack anisotropic fracture prediction technology includes the following steps: Performing azimuth division based on the pre-stack seismic data of the study area to form a plurality of azimuth gathers; Extracting frequency, amplitude, and attenuation attributes from each of the azimuth gathers, and performing optimal attribute selection; Performing azimuthal ellipse fitting, wherein the major axis direction of the ellipse is the direction of the microscale crack; The ellipse flattening is fitted with the single well fracture density curve to obtain the density of the microscale fractures.

6. The underground crack prediction method based on multi-scale information fusion according to claim 1, characterized in that: The method of obtaining the structural pattern of the research area determined based on geological data comprises the following steps: Conduct geological simulation based on geological profile and obtain dynamic data; By analyzing outcrops and cores, the overall structural pattern, lithology, and stress state of the study area were established; The fracture development state is obtained by combining the dynamic data, the fracture parameters are counted, the distribution of the fractures in the formation corresponding to the core is predicted and calibrated, and the structural style is established, which includes normal faults, reverse faults and strike-slip faults.

7. The underground crack prediction method based on multi-scale information fusion according to claim 1, characterized in that: The method of obtaining the fracture analysis results of the study area determined based on the well logging data and obtaining the fracture density curve of each single well according to the fracture analysis results includes the following steps: Obtaining the logging data based on imaging logging technology and conventional logging data; Performing the fracture analysis based on the well logging data, wherein the fracture analysis includes the inclination, tendency, and filling condition of the fracture; Fracture grouping is performed based on the fracture analysis, and fracture densities in different directions are calculated in combination with well trajectory data in the conventional logging data to form the single well fracture density curve.

8. A device for predicting underground cracks using multi-scale information fusion, characterized in that: include: The first acquisition module is used to obtain the structural style of the study area determined based on geological data; The first analysis module is used to obtain the fracture analysis results of the study area determined based on the well logging data, and obtain the fracture density curve of each single well according to the fracture analysis results; A model forward modeling module is used to perform model forward modeling based on the structural style and fracture analysis of the study area and the seismic data of the study area to obtain model forward modeling results and obtain multi-scale fracture classification standards; the seismic data includes pre-stack seismic data and post-stack seismic data; The first fracture determination module is used to obtain large and medium-scale fractures based on the post-stack seismic data of the study area and the seismic fault interpretation results according to the seismic attributes; The second fracture determination module is used to obtain small-scale fractures based on the post-stack seismic data and a preset fracture prediction model; A third fracture determination module is configured to obtain microscale fractures based on the pre-stack seismic data of the study area and the fracture density curves of each single well according to the pre-stack azimuthal anisotropic fracture prediction technology; The fusion module is used to take the large and medium scale cracks, the small scale cracks and the micro scale cracks as input variables, and fuse them using a fuzzy neural network to obtain a multi-scale fused crack prediction data body.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the underground crack prediction method based on multi-scale information fusion as described in any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: The processor, the communication interface, the memory and the communication bus are connected to each other via the communication bus. Memory for storing computer programs; The processor is configured to implement the underground crack prediction method based on multi-scale information fusion as described in any one of claims 1 to 7 when executing the program stored in the memory.

Images