A multi-channel adaptive subtraction method and system for protecting target layers

Abstract

This invention relates to the field of geophysical exploration technology, specifically to a multi-channel adaptive subtraction method and system for protecting target strata. The method includes optimized layer protection multi-channel adaptive subtraction processing after combining multi-channel layer protection models (masks). This method solves the problem that since multiple waves and primary waves in the original data do not completely satisfy the orthogonality condition, the effective signal is usually damaged during the adaptive subtraction process. It has the technical effect of effectively protecting the primary wave of the target strata and achieving a significant optimization and improvement in the matching subtraction results.

Description

Technical Field

[0001] This invention relates to the field of geophysical exploration technology, specifically to a multi-channel adaptive subtraction method and system for protecting target strata. Background Technology

[0002] In seismic exploration, multiples are one of the main factors affecting the signal-to-noise ratio and resolution of marine seismic data. They are usually regarded as noise and removed. Multiple suppression is one of the important steps in determining the success of seismic data processing and interpretation. The prediction-subtraction method based on the wave equation includes two processes: prediction and subtraction, and has a wider range of applicability. Since there are differences between the predicted multiples and the actual multiples in the original seismic data, an adaptive subtraction method is needed to match and subtract the predicted multiples from the original seismic data to achieve a good multiple suppression effect. In practical applications, the multiples and primary waves in the original data do not completely satisfy the orthogonality condition (i.e., they have a strict separability condition), and the effective signal is usually damaged during the adaptive subtraction process. Summary of the Invention

[0003] To address the technical problem of damage to effective signals during the matching subtraction process in multichannel adaptive subtraction methods, and specifically for scenarios requiring suppression of multiple waves through conventional wave equation-based predictive subtraction, this invention provides a multichannel adaptive subtraction method and system for protecting target layers.

[0004] In a first aspect, embodiments of the present invention provide a multi-channel adaptive subtraction method for protecting target layers, comprising:

[0005] Multiple wave model prediction: Based on the stratigraphic data of the target layer, a predicted multiple wave model is obtained through a multiple wave prediction method;

[0006] Extracting the primary wave model: Based on the stratigraphic data of the target layer, the primary wave model corresponding to the target layer is extracted using the phase axis matching method to obtain the extracted original primary wave model;

[0007] Determine the protection range: Based on the original primary wave model, through exponential inverse processing after steepness adaptation, obtain the layer protection model Mask of the target layer to determine the protection range of the original primary wave model;

[0008] Protecting effective signals: Based on the original seismic data and the predicted multiple wave model, and by introducing the layer protection model Mask, the optimized layer protection multi-channel adaptive subtraction processing after combining the multi-channel layer protection model Mask is used to obtain seismic data with layer protection after suppressing multiple waves, so as to protect the effective signals in the original seismic data.

[0009] Preferably, the steepness adaptation treatment includes:

[0010] Based on the original first-wave model, the original first-wave model is subjected to steepness optimization by configuring the scaling factor of the layer protection model function to obtain an amplified first-wave model.

[0011] Preferably, the steepness optimization process includes:

[0012] Initialize the scale factor configuration, where the initial scale factor configuration should be no less than 100, with a typical value of 100;

[0013] Based on the original primary wave model and the scale factor of the initial configuration, a visualized primary wave model is generated through the initial amplification process of the layer protection model function.

[0014] Based on the visualized, initially magnified primary wave model, determine its steepness suitability:

[0015] If its steepness is not suitable, the scale factor is adjusted to adapt its steepness to the needs of steepness optimization.

[0016] If the steepness is appropriate, then the amplified primary wave model is determined.

[0017] Preferably, the exponential inversion process includes:

[0018] Based on the amplified primary wave model, the amplified primary wave model is exponentially reversed by running the natural exponential function of the layer protection model function to form the layer protection model Mask, thereby determining the protection range of the original primary wave model.

[0019] Preferably, the layer protection model function is:

[0020] Mask = exp(-scale × Primarys) 2 )

[0021] Wherein, exp() is the natural exponential function; scale is a scale factor that adjusts the steepness of the Mask boundary of the layer protection model; Primaries is the original first-wave model; and Mask is the Mask of the layer protection model.

[0022] Preferably, the multi-layer protection model Mask combination includes:

[0023] Based on the original seismic data and the predicted multiple wave model, the combined seismic data and the multiple wave combined model are determined by combining the multi-channel layer protection through the multi-channel adaptive function of the layer protection. This allows the layer protection model Mask to be introduced into the original seismic data and the multiple wave model.

[0024] Preferably, the optimization layer protects the multi-channel adaptive subtraction process, which includes:

[0025] Based on combined seismic data and a multichannel multiple wave combined model, multichannel layer protection optimization subtraction is performed through a multichannel layer protection adaptive function to obtain seismic data with layer protection after suppression of multiple waves. This is to protect the primary wave data of the target protected layer in the process of removing multiple wave data from the original seismic data.

[0026] Preferably, the layer protection multichannel adaptive function is:

[0027]

[0028] Wherein, d is the original seismic data vector; s i M represents the i-th adaptive matched filter; i The i-th channel is the predicted multiple wave model, which is represented by a multiple wave convolution matrix; ⊙ is the Hadamard product; and Mask is the layer protection model Mask.

[0029] Preferably, the multiple wave model is:

[0030]

[0031] Wherein, m(t) is the predicted multiple of a single channel; N is the length of the single channel data; and n is the length of the adaptive matched filter.

[0032] Preferably, the original seismic data is:

[0033] d(t)={d(0), d(1), d(2)…d(N)}

[0034] Where t is a natural number 1…N.

[0035] Preferably, the multiple prediction method includes the extended interlayer multiple prediction method XIMP and the inverse scattering method.

[0036] Preferably, the in-phase shaft matching method includes:

[0037] Based on the stratigraphic data of the target stratigraphic level, pulse data constrained by stratigraphic time is generated. Based on the adaptive matching of the pulse data and the original seismic data, the original first wave model of the target stratigraphic level is extracted.

[0038] Preferably, the method further includes the following before the predicted multiple wave model:

[0039] The raw seismic data is acquired and preprocessed to obtain a post-stack dataset. Based on the post-stack dataset, the stratigraphic data of the target layer is extracted. The preprocessing of the raw seismic data includes the data preprocessing process of observation system definition, short-path recovery, regularization, removal of direct waves, noise suppression, dynamic correction, static correction, stacking, and imaging.

[0040] On the other hand, the present invention also provides a multi-channel adaptive subtraction system for protecting target layers, comprising:

[0041] The system is used to implement the above-mentioned multichannel adaptive subtraction method, and includes:

[0042] Multiple wave model prediction module: used to obtain the predicted multiple wave model based on the stratigraphic data of the target stratigraphic level through the multiple wave prediction method;

[0043] Primary wave model extraction module: Based on the stratigraphic data of the target stratigraphic level, the module extracts the primary wave model corresponding to the target stratigraphic level using the phase axis matching method to obtain the extracted original primary wave model;

[0044] The protection range determination module is used to obtain the layer protection model Mask of the target layer based on the original primary wave model through exponential inverse processing after steepness adaptation, so as to determine the protection range of the original primary wave model.

[0045] The effective signal protection module is used to obtain suppressed multiple seismic data with layer protection based on the original seismic data and the predicted multiple wave model. It introduces a layer protection model Mask and then performs optimized layer protection multi-channel adaptive subtraction processing after combining multiple layer protection model Masks to obtain seismic data with layer protection, thereby protecting the effective signals in the original seismic data. Attached Figure Description

[0046] Figure 1 This is a general flowchart of an embodiment of the present invention;

[0047] Figure 2 This is a detailed flowchart of an embodiment of the present invention;

[0048] Figure 3 This is an exploded flowchart of an embodiment of the present invention;

[0049] Figure 4 This is a flowchart illustrating the process of an embodiment of the present invention;

[0050] Figure 5 This is a flowchart of layer-protected multichannel adaptive subtraction according to an embodiment of the present invention;

[0051] Figure 6 is a set of comparison diagrams of actual seismic data tests according to an embodiment of the present invention;

[0052] Figure 6a It is a map of the raw seismic data containing multiple waves;

[0053] Figure 6b It is a predicted multiple wave model;

[0054] Figure 6c It is the layer protection model Mask of the target layer;

[0055] Figure 7 is a comparison diagram of CMP gathers in an embodiment of the present invention;

[0056] Figure 7a It is a multi-channel adaptive subtraction CMP gather map of the unprotected target layer;

[0057] Figure 7b It is a multi-channel adaptive subtraction CMP gather map with protected target layers;

[0058] Figure 8 is a set of superimposed cross-sectional comparison diagrams of an embodiment of the present invention;

[0059] Figure 8a It is a multichannel adaptive subtraction stacked profile of the unprotected target layer;

[0060] Figure 8b It is a multi-channel adaptive subtraction superimposed profile of the protected target layer; Detailed Implementation

[0061] The present invention will be further described below with reference to specific embodiments, but the scope of protection of the present invention is not limited thereto.

[0062] Figure 1 , Figure 2 The following are the overall flowchart and detailed flowchart of this multichannel adaptive subtraction method. Figure 1 , 2 As shown, an embodiment of the present invention includes a multi-channel adaptive subtraction method for protecting target layers, comprising:

[0063] Multiple wave model prediction S101: Based on the stratigraphic data of the target layer, the predicted multiple wave model S201 is obtained through the multiple wave prediction method.

[0064] Extracting the primary wave model S102: Based on the stratigraphic data of the target layer, the primary wave model corresponding to the target layer is extracted using the phase axis matching method to obtain the extracted original primary wave model S202;

[0065] Determine the protection range S103: Based on the original primary wave model, through the exponential inverse processing after steepness adaptation, obtain the layer protection model Mask of the target layer to determine the protection range S203 of the original primary wave model.

[0066] Protecting the effective signal S104: Based on the original seismic data and the predicted multiple wave model, and by introducing the layer protection model Mask, the optimized layer protection multi-channel adaptive subtraction processing after combining the multi-channel layer protection model Mask is used to obtain the seismic data with layer protection after suppressing multiple waves, so as to protect the effective signal S204 in the original seismic data.

[0067] According to one embodiment of the present invention, such as Figure 1 , 2 As shown, the steepness adaptation process is as follows:

[0068] Based on the original first-wave model, the original first-wave model is subjected to steepness optimization by configuring the scaling factor of the layer protection model function to obtain an amplified first-wave model.

[0069] According to one embodiment of the present invention, such as Figure 3 , 4 As shown, the steepness optimization process includes:

[0070] Initialize the scale factor configuration, where the initial scale factor configuration should be no less than 100, with a typical value of 100;

[0071] Based on the original primary wave model and the scale factor of the initial configuration, a visualized primary wave model is generated through the initial amplification process of the layer protection model function.

[0072] Based on the visualized, initially magnified primary wave model, determine its steepness suitability:

[0073] If its steepness is not suitable, the scale factor is adjusted to adapt its steepness to the needs of steepness optimization.

[0074] If the steepness is appropriate, then the amplified primary wave model is determined.

[0075] According to one embodiment of the present invention, such as Figure 3 , 4 As shown, the exponential inverse processing includes:

[0076] Based on the amplified primary wave model, the amplified primary wave model is exponentially reversed by running the natural exponential function of the layer protection model function to form the layer protection model Mask, thereby determining the protection range of the original primary wave model.

[0077] According to one embodiment of the present invention, such as Figure 3 , 4 As shown, the layer protection model function is:

[0078] Mask = exp(-scale × Primarys) 2 )

[0079] Wherein, exp() is the natural exponential function; scale is a scale factor that adjusts the steepness of the Mask boundary of the layer protection model; Primaries is the original first-wave model; and Mask is the Mask of the layer protection model.

[0080] According to one embodiment of the present invention, such as Figure 3 , 4 As shown in Figure 5, the multi-layer protection model Mask combination includes:

[0081] Based on the original seismic data and the predicted multiple wave model, the combined seismic data and the multiple wave combined model are determined by combining the multi-channel layer protection through the multi-channel adaptive function of the layer protection. This allows the layer protection model Mask to be introduced into the original seismic data and the multiple wave model.

[0082] According to one embodiment of the present invention, such as Figure 3 , 4 As shown in Figure 5, the optimization layer protection multi-channel adaptive subtraction processing includes:

[0083] Based on combined seismic data and a multichannel multiple wave combined model, multichannel layer protection optimization subtraction is performed through a multichannel layer protection adaptive function to obtain seismic data with layer protection after suppression of multiple waves. This is to protect the primary wave data of the target protected layer in the process of removing multiple wave data from the original seismic data.

[0084] According to one embodiment of the present invention, such as Figure 3 , 4 As shown in Figure 5, the layer protection multichannel adaptive function is:

[0085]

[0086] Wherein, d is the original seismic data vector; s i M represents the i-th adaptive matched filter; i The i-th channel is the predicted multiple wave model, which is represented by a multiple wave convolution matrix; ⊙ is the Hadamard product; and Mask is the layer protection model Mask.

[0087] According to one embodiment of the present invention, such as Figure 3 , 4 As shown in Figure 5, the multiple wave model is as follows:

[0088]

[0089] Wherein, m(t) is the predicted multiple of a single channel; N is the length of the single channel data; and n is the length of the adaptive matched filter.

[0090] According to one embodiment of the present invention, such as Figure 3 , 4 As shown in Figure 5, the original seismic data is as follows:

[0091] d(t)={d(0), d(1), d(2)…d(N)}

[0092] Where t is a natural number 1…N.

[0093] According to one embodiment of the present invention, such as Figure 1 , 2 As shown, the multiple prediction method includes the extended interlayer multiple prediction method XIMP and the inverse scattering method.

[0094] According to one embodiment of the present invention, such as Figure 1 , 2 As shown, the in-phase shaft matching method includes:

[0095] Based on the stratigraphic data of the target stratigraphic level, pulse data constrained by stratigraphic time is generated. Based on the adaptive matching of the pulse data and the original seismic data, the original first wave model of the target stratigraphic level is extracted.

[0096] According to one embodiment of the present invention, such as Figure 1 , 2 As shown, the predicted multiple wave model also includes:

[0097] The raw seismic data is acquired and preprocessed to obtain a post-stack dataset. Based on the post-stack dataset, the stratigraphic data of the target layer is extracted. The preprocessing of the raw seismic data includes the data preprocessing process of observation system definition, short-path recovery, regularization, removal of direct waves, noise suppression, dynamic correction, static correction, stacking, and imaging.

[0098] Meanwhile, the present invention also provides a multi-channel adaptive subtraction system for protecting target layers, comprising:

[0099] The system is used to implement the above-mentioned multichannel adaptive subtraction method, and includes:

[0100] Multiple wave model prediction module A101: used to obtain the predicted multiple wave model S201 based on the stratigraphic data of the target stratigraphic level through the multiple wave prediction method;

[0101] The primary wave model extraction module A102 is used to extract the primary wave model corresponding to the target layer based on the layer data of the target layer by using the phase axis matching method, so as to obtain the extracted original primary wave model S202.

[0102] The protection range determination module A103 is used to obtain the layer protection model Mask of the target layer based on the original primary wave model through exponential inverse processing after steepness adaptation, so as to determine the protection range S203 of the original primary wave model.

[0103] The effective signal protection module A104 is used to obtain the suppressed multiple wave seismic data with layer protection based on the original seismic data and the predicted multiple wave model. It is based on the introduction of the layer protection model Mask and the optimized layer protection multi-channel adaptive subtraction processing after combining the multi-channel layer protection model Mask, so as to protect the effective signal S204 in the original seismic data.

[0104] The distinguishing technical features of this case are:

[0105] Compared with existing and similar technologies, the adaptive subtraction method used in this case is different, that is, layer protection function is added to each adaptive subtraction method to form a corresponding layer protection adaptive subtraction method;

[0106] Specifically, this case adopts a multi-channel protection model Mask combined with optimized multi-channel adaptive subtraction processing for multi-channel protection. In addition, the layer protection model function and the multi-channel adaptive function for multi-channel protection are used in the specific function model.

[0107] Among them, the layer protection model function is:

[0108] Mask = exp(-scale × Primarys) 2 )

[0109] Among them, the layer-protected multichannel adaptive function is:

[0110]

[0111] The following is combined Figure 1 , 2 The working principle of this multichannel adaptive subtraction method shown in the embodiments will be explained.

[0112] like Figure 1 , 2As shown, this multi-channel adaptive subtraction method includes functional modules such as multiple wave model prediction, extraction of the primary wave model, determination of the protection range, and protection of the effective signal. Key processes include determining the protection range and protecting the effective signal. The "determining the protection range" process includes steepness adaptation processing and exponential inversion processing, while the "protecting the effective signal" process includes combining multi-channel layer protection model masks and optimizing the multi-channel adaptive subtraction processing of layer protection. In fact, determining the protection range involves obtaining the layer protection model mask of the target layer based on the original primary wave model through exponential inversion processing after steepness adaptation processing, thereby determining the protection range of the original primary wave model. Protecting the effective signal involves introducing layer protection based on the original seismic data and the predicted multiple wave model. Based on the model Mask, optimized layer-protected adaptive subtraction processing is performed by combining the multi-channel layer-protected model Mask to obtain seismic data with layer protection after suppression of multiple waves, thus protecting the effective signals in the original seismic data. Through this processing, this method addresses the problem of damage to effective signals during the matching subtraction process in conventional multi-channel adaptive subtraction methods. It improves upon conventional methods by converting the target layer primary wave into a layer-protected Mask and incorporating the Mask into the multi-channel adaptive subtraction method, achieving the goal of protecting the target layer primary wave data during the matching subtraction process. In summary, actual data testing verifies that this method can effectively protect the target layer primary wave, and the matching subtraction results are significantly optimized and improved.

[0113] The following is combined Figure 3 , 4 The working principle of this multichannel adaptive subtraction method shown in the embodiments will be explained.

[0114] like Figure 3 , 4 As shown, the "determining the protection range" process includes steepness adaptation processing and exponential inversion processing. "Stiffness adaptation processing" involves optimizing the original primary wave model by configuring the scale factor of the layer protection model function to obtain an amplified primary wave model; that is, adjusting the steepness of the Mask boundary by configuring the scale factor. Simultaneously, "exponential inversion processing" involves exponentially inverting the amplified primary wave model by running the natural exponential function of the layer protection model function to form a layer protection model Mask, thereby determining the protection range of the original primary wave model. Through these processes, the target stratum primary wave is transformed into a layer protection Mask; that is, the target stratum primary wave is extracted from the target stratum information and transformed into a layer protection Mask using the following formula:

[0115] Mask = exp(-scale × Primarys)2 )

[0116] Where exp represents the natural exponential function; the scale factor is used to adjust the steepness of the mask boundary, the larger the value, the steeper the mask boundary; and Primarys are the primary waves of the extracted target layers.

[0117] Mask = exp(-scale × Primarys) 2 )

[0118] like Figure 3 As shown, firstly, the seismic data is preliminarily processed; secondly, the horizon file is imported into the seismic data, and the primary wave of the corresponding horizon is extracted using the horizon prediction multiple wave model data; thirdly, the primary wave is converted into the corresponding horizon protection mask using the above formula; finally, the multiple wave is subtracted using the improved horizon protection multichannel adaptive subtraction method; the verification with actual seismic data proves that this method can effectively protect the primary wave of the target horizon, and the matching subtraction results are significantly optimized and improved.

[0119] The layer protection model function involved in this multichannel adaptive subtraction method is:

[0120] Mask = exp(-scale × Primarys) 2 )

[0121] To prepare for the next stage of processing multi-channel protection model Mask combination and optimization layer protection multi-channel adaptive subtraction processing, the protection range of the original primary wave model was clarified.

[0122] Meanwhile, the "determining the protection range" process also includes a steepness optimization process. Simply put, the scale factor can generally be a positive real number greater than 100. The larger the value, the larger the protection range, but the steeper the boundary of the range. The choice of Mask must ensure that the protection range is as large as possible, but at the same time, the boundary of the range must have a certain smooth slope. Therefore, the setting of the scale value must balance these two effects, and the default value is 100. The process includes: first, initializing the first scale factor, then generating an enlarged primary wave model, observing the enlarged primary wave model, and judging whether the steepness angle and smoothness meet the needs of "steepness adaptation processing" through engineering experience. If the requirements are met, the scale factor is determined, and the enlarged primary wave model based on the scale factor is determined. If the requirements are not met, within the compensation range, the scale factor is configured in an adjustment manner to adapt its steepness to the needs of steepness optimization processing, thereby finally forming a suitable protection range for the original primary wave model.

[0123] The following is combined Figure 3 , 4The working principle of this multichannel adaptive subtraction method shown in the embodiments will be explained.

[0124] like Figure 3 , 4 As shown, the "multi-channel layer protection model mask combination" process includes multi-channel layer protection model mask combination and optimized layer protection multi-channel adaptive subtraction processing. The "multi-channel layer protection model mask combination" involves combining the original seismic data and the predicted multiple wave model using a layer protection multi-channel adaptive function to determine the combined seismic data and the multi-channel multiple wave model, thus introducing the layer protection model mask into the original seismic data and the multiple wave model. Simultaneously, the "optimized layer protection multi-channel adaptive subtraction processing" involves optimizing the multi-channel layer protection subtraction based on the combined seismic data and the multi-channel multiple wave model using a layer protection multi-channel adaptive function to obtain seismic data with layer protection after suppressing multiple waves, thereby removing the primary wave data of the protected layer from the original seismic data during the multiple wave model process. To achieve layer protection multi-channel adaptive subtraction, the layer protection mask is introduced into the conventional multi-channel adaptive subtraction method, which can be derived. During the calculation process, the layer protection mask needs to be applied to both the original seismic data and the predicted multiple wave data. The conventional multi-channel adaptive subtraction method can be expressed as:

[0125]

[0126] Where d represents the original seismic record and s represents the adaptive matched filter.

[0127] The predicted multiple convolution matrix M is:

[0128]

[0129] Where m(t) is the predicted multiple wave and N is the length of the adaptive filter.

[0130] The original seismic data is as follows:

[0131] d(t)={d(0), d(1), d(2)…d(N)}

[0132] Where t is a natural number 1…N;

[0133] After introducing the layer protection mask, the improved layer protection multichannel adaptive subtraction formula can be obtained:

[0134]

[0135] Where E is the least squares error function, Mask is the layer protection matrix, ⊙ is the Hadamard product, and the meanings of the other symbols are the same as in the above formula.

[0136] Through algorithm improvements, the layer protection multi-channel adaptive subtraction method can effectively protect the primary wave of the target layer and optimize the matching subtraction results.

[0137] like Figure 3 As shown, the entire process includes, firstly, the data preparation stage:

[0138] (1) Including but not limited to the following processing steps: observation system definition, short-path recovery, regularization, removal of direct waves, noise suppression, dynamic correction, static correction, stacking, and imaging.

[0139] (2) Extract the layers that need to be protected from the post-stack data, import the layer files into the seismic data, use the layer to predict the multiple wave model data M, and extract the primary waves of the corresponding layers.

[0140] Secondly, there is the multi-channel adaptive subtraction stage with layer protection:

[0141] (1) The primary wave is converted into the corresponding layer protection mask using the method (layer protection model function) proposed in this patent;

[0142] (2) The improved layer protection multichannel adaptive subtraction method (layer protection multichannel adaptive function) is used to subtract multiple waves.

[0143] Finally, there is the result output stage: outputting the calculation results, including the first wave after suppressing multiple waves, the subtracted multiple waves, and the adaptive filter.

[0144] The layer-protected multichannel adaptive function involved in this multichannel adaptive subtraction method is:

[0145]

[0146] To obtain seismic data with layered protection after suppression of multiples, and to protect the primary wave data of the target protected layer during the process of removing multiples from the original seismic data, the aim is to protect the primary wave data of the target protected layer during the matching subtraction process based on the seismic data with layered protection after suppression of multiples.

[0147] At the same time, this case once again combines Figure 5 As shown, the multi-channel adaptive subtraction process for the entire layer protection is further described:

[0148] The specific implementation steps of layer-protected multi-channel adaptive subtraction are as follows:

[0149] Step 1: Preliminary processing of seismic data to obtain stacked or migrated data volumes;

[0150] Step 2: Extract the target layer from the stacked or migrated data volume and import it into the seismic data;

[0151] Step 3: Predict the multiple wave model of the target layer, and extract the primary wave of the target layer at the same time;

[0152] Step 4: Use the layer protection model function to convert the primary wave of the target layer into a layer protection mask;

[0153] Step 5: Load the seismic data, multiple wave model, and Mask into the layer protection multichannel adaptive subtraction algorithm (layer protection multichannel adaptive function) to obtain the final layer protection multiple wave suppression result.

[0154] The following is combined Figure 1 , 2 The working principle of this multichannel adaptive subtraction method shown in the embodiments will be explained.

[0155] like Figure 1 , 2 As shown, this multichannel adaptive subtraction method also includes a data preparation stage, which includes multiple wave model prediction and primary wave model extraction. The purpose of these two processes is to obtain the predicted multiple wave model and the extracted primary wave model. The "multiple wave model prediction" involves obtaining the predicted multiple wave model based on the stratigraphic data of the target layer using the multiple wave prediction method. Meanwhile, the primary wave model extraction involves obtaining the primary wave model corresponding to the target layer based on the stratigraphic data of the target layer using the phase axis matching method to obtain the extracted primary wave model. The above methods complete the data preparation work of the data preparation stage, providing a data foundation for the next stage of steepness adaptive processing and exponential inverse processing.

[0156] like Figure 1 , 2As shown, this multichannel adaptive subtraction method also includes a data preprocessing stage, which includes: acquiring raw seismic data and preprocessing the raw seismic data to obtain a post-stack dataset; and extracting the layer data of the target layer based on the post-stack dataset. The preprocessing of the raw seismic data includes observation system definition, near-path recovery, regularization, removal of direct waves, noise suppression, dynamic correction, static correction, stacking, and imaging. The true multiples in the raw data refer to the original data containing both primary waves (effective signals) and multiples (noise). The so-called "true multiples" refers to the multiples contained in the raw data. The predicted multiples refer to the multiples calculated using a prediction method, such as model-driven or data-driven methods, which differ somewhat from the true multiples in the raw seismic data. Furthermore, the original data (containing primary waves + true multiples) and the predicted multiples are adaptively subtracted to obtain the primary waves. The data preprocessing stage includes, but is not limited to, observation system definition, short-path recovery, regularization, removal of direct waves, noise suppression, dynamic correction, static correction, stacking, and imaging. Then, the layers to be protected are extracted from the post-stack data, and finally, the layer files are imported into the seismic data.

[0157] The working principle of this multichannel adaptive subtraction method will be explained below with reference to the embodiments shown in Figures 6, 7, and 8.

[0158] In existing technologies, based on different assumptions and prior conditions, multiple suppression methods can be broadly classified into three categories: filtering methods, wave equation prediction subtraction methods, and machine learning methods. Direct filtering methods based on signal analysis mainly utilize digital signal filtering techniques to suppress multiples based on the characteristic differences between primary and multiple waves in a specific domain, such as predictive deconvolution, FK filtering, Radon transform, and cluster filtering. Wave equation prediction subtraction methods involve both prediction and subtraction processes and have wider applicability. Multiple prediction can be performed in two ways: model-driven and data-driven. Model-driven methods require prior knowledge of the subsurface model's parameters and are often used for processing seabed-related multiples in marine seismic data. Data-driven methods, such as surface correlated multiple removal (SRME) and backscattering surface multiple removal, do not require prior information from subsurface models and directly utilize raw seismic data to suppress multiples. Machine learning methods focus on extracting features from seismic data through trained neural networks; once the neural network is successfully trained, it can quickly and effectively remove multiples. Since the predicted multiples differ from the actual multiples in the raw seismic data, adaptive subtraction methods are needed to match and subtract the predicted multiples from the raw seismic data to achieve good multiple suppression. Commonly used adaptive subtraction methods for multiples include independent component analysis, pattern matched filtering, non-steady-state linear regression, higher-order sparse Radon transform, and least squares adaptive matched filtering. Since the multiple waves and the primary waves in the original data cannot fully satisfy the orthogonality condition (i.e., have a strict separability condition), the effective signal is usually damaged during the adaptive subtraction process. This invention improves the conventional multi-channel adaptive subtraction method by converting the primary wave of the target layer into a layer protection mask and introducing the mask into the multi-channel adaptive subtraction method, thereby achieving the purpose of protecting the primary wave data of the target protection layer during the matching subtraction process.

[0159] As mentioned above, this case provides a summary of the existing technology. Next, we will further describe the technical effects obtained by this multichannel adaptive subtraction method based on the experimental results in conjunction with Figures 6, 7, and 8.

[0160] As shown in Figures 6, 7, and 8, in this data test, Figure 6 shows the actual seismic data test, in which... Figure 6a For raw seismic data containing multiple waves, Figure 6b For the predicted multiple wave model, Figure 6cTo extract the primary wave based on the layer and transform it using the layer protection model function, the target layer mask is obtained. The original seismic data, multiple wave model, and mask shown in Figure 6 are loaded into the layer protection multi-channel adaptive subtraction algorithm (layer protection multi-channel adaptive function) to obtain the final layer protection multi-channel suppression results, as shown in Figures 7 and 8. Figure 7 shows the layer protection multi-channel adaptive subtraction results on the CMP gather, where... Figure 7a , Figure 7b The figures show the multichannel adaptive subtraction results with and without layer protection, respectively. It can be seen that after applying layer protection, the energy and continuity of the phase axis of the target layer shown in the box are significantly enhanced. To further illustrate the advantages of the layer protection method, the matching subtraction results are superimposed and plotted as Figure 8, where... Figure 8a , Figure 8b These are the multichannel adaptive subtraction results for unprotected and protected layers, respectively. From the stacked profiles, it can be seen that the stacked profile with protected layers (…) Figure 8b The energy and continuity of the phase axis of the target layer shown in the box are significantly better than those of the stacked profile without layer protection. Figure 8a In summary, through verification with actual seismic data, this method can effectively protect the primary wave of the target layer, and the matching subtraction results are significantly optimized and improved.

[0161] The various embodiments in this invention are described in a progressive manner. For the same or similar parts between the various embodiments, please refer to each other. Each embodiment focuses on describing the differences from other embodiments.

[0162] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more features.

[0163] The scope of protection of this invention is not limited to the embodiments described above. Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its scope and spirit. If these modifications and variations fall within the scope of the claims of this invention and their equivalents, then the intent of this invention also includes these modifications and variations.

Claims

1. A multi-channel adaptive subtraction method for protecting target layers, characterized in that, include: Multiple wave model prediction: Based on the stratigraphic data of the target layer, a predicted multiple wave model is obtained through a multiple wave prediction method; Extracting the primary wave model: Based on the stratigraphic data of the target layer, the primary wave model corresponding to the target layer is extracted using the phase axis matching method to obtain the extracted original primary wave model; Determine the protection range: Based on the original primary wave model, through exponential inverse processing after steepness adaptation, obtain the layer protection model Mask of the target layer to determine the protection range of the original primary wave model; Protecting effective signals: Based on the original seismic data and the predicted multiple wave model, and by introducing the layer protection model Mask, the optimized layer protection multi-channel adaptive subtraction processing after combining the multi-channel layer protection model Mask is used to obtain seismic data with layer protection after suppressing multiple waves, so as to protect the effective signals in the original seismic data.

2. The multichannel adaptive subtraction method according to claim 1, characterized in that, The steepness adaptation process: Based on the original first-wave model, the original first-wave model is subjected to steepness optimization by configuring the scaling factor of the layer protection model function to obtain an amplified first-wave model.

3. The multichannel adaptive subtraction method according to claim 2, characterized in that, The steepness optimization process includes: Initialize the scale factor configuration, where the initial scale factor configuration should be no less than 100, with a typical value of 100; Based on the original primary wave model and the scale factor of the initial configuration, a visualized primary wave model is generated through the initial amplification process of the layer protection model function. Based on the visualized, initially magnified primary wave model, determine its steepness suitability: If its steepness is not suitable, the scale factor is adjusted to adapt its steepness to the needs of steepness optimization. If the steepness is appropriate, then the amplified primary wave model is determined.

4. The multichannel adaptive subtraction method according to claim 1, characterized in that, The exponential inverse processing includes: Based on the amplified primary wave model, the amplified primary wave model is exponentially reversed by running the natural exponential function of the layer protection model function to form the layer protection model Mask, thereby determining the protection range of the original primary wave model.

5. The multichannel adaptive subtraction method according to claim 2 or 4, characterized in that, include: The layer protection model function is: Mask=exp(-scale×Primaries 2 ) Wherein, exp() is the natural exponential function; scale is a scale factor that adjusts the steepness of the Mask boundary of the layer protection model; Primaries is the original first-wave model; and Mask is the Mask of the layer protection model.

6. The multichannel adaptive subtraction method according to claim 1, characterized in that, The multi-layer protection model Mask combination includes: Based on the original seismic data and the predicted multiple wave model, the combined seismic data and the multiple wave combined model are determined by combining the multi-channel layer protection through the multi-channel adaptive function of the layer protection. This allows the layer protection model Mask to be introduced into the original seismic data and the multiple wave model.

7. The multichannel adaptive subtraction method according to claim 6, characterized in that, The optimized layer protects the multichannel adaptive subtraction process, which includes: Based on combined seismic data and a multichannel multiple wave combined model, multichannel layer protection optimization subtraction is performed through a multichannel layer protection adaptive function to obtain seismic data with layer protection after suppression of multiple waves. This is to protect the primary wave data of the target protected layer in the process of removing multiple wave data from the original seismic data.

8. The multichannel adaptive subtraction method according to claim 6 or 7, characterized in that, include: The layer protection multichannel adaptive function is: Wherein, d is the original seismic data vector; s i M represents the i-th adaptive matched filter; i The i-th channel is the predicted multiple wave model, which is represented by a multiple wave convolution matrix; ⊙ is the Hadamard product; and Mask is the layer protection model Mask.

9. The multichannel adaptive subtraction method according to claim 1, characterized in that, The multiple prediction methods include the extended interlayer multiple prediction method XIMP and the inverse scattering method.

10. The multichannel adaptive subtraction method according to claim 1, characterized in that, The in-phase shaft matching method includes: Based on the stratigraphic data of the target stratigraphic level, pulse data constrained by stratigraphic time is generated. Based on the adaptive matching of the pulse data and the original seismic data, the original first wave model of the target stratigraphic level is extracted.

11. The multichannel adaptive subtraction method according to claim 1, characterized in that, The predicted multiple wave model also includes: The raw seismic data is acquired and preprocessed to obtain a post-stack dataset. Based on the post-stack dataset, the stratigraphic data of the target layer is extracted. The preprocessing of the raw seismic data includes the data preprocessing process of observation system definition, short-path recovery, regularization, removal of direct waves, noise suppression, dynamic correction, static correction, stacking, and imaging.

12. A multi-channel adaptive subtraction system for protecting target layers, characterized in that, include: The system is used to implement the multichannel adaptive subtraction method as described in any one of claims 1-11, and includes: Multiple wave model prediction module: used to obtain the predicted multiple wave model based on the stratigraphic data of the target stratigraphic level through the multiple wave prediction method; Primary wave model extraction module: Based on the stratigraphic data of the target stratigraphic level, the module extracts the primary wave model corresponding to the target stratigraphic level using the phase axis matching method to obtain the extracted original primary wave model; The protection range determination module is used to obtain the layer protection model Mask of the target layer based on the original primary wave model through exponential inverse processing after steepness adaptation, so as to determine the protection range of the original primary wave model. The effective signal protection module is used to obtain suppressed multiple seismic data with layer protection based on the original seismic data and the predicted multiple wave model. It introduces a layer protection model Mask and then performs optimized layer protection multi-channel adaptive subtraction processing after combining multiple layer protection model Masks to obtain seismic data with layer protection, thereby protecting the effective signals in the original seismic data.

Images