Logging curve interpolation method, device, equipment, medium and program product

Abstract

Embodiments of the present application disclose a well logging curve interpolation method, device, equipment, medium and program product. The method comprises: obtaining original seismic data and original well logging curve data, and preprocessing the original seismic data and the original well logging curve data to determine a to-be-processed seismic data body, a wellside seismic trace data matrix and a well logging curve data matrix between target strata; for each to-be-processed seismic trace data in the to-be-processed seismic data body, the to-be-processed seismic trace data, the wellside seismic trace data matrix and the well logging curve data matrix are respectively solved as a query vector, a key vector and a value vector in a self-attention mechanism to determine a target interpolation well logging curve corresponding to the to-be-processed seismic trace data; and the target interpolation well logging curves are subjected to data homing processing according to a data processing mode of the preprocessing to determine a well logging curve interpolation result. The well logging curve obtained after the interpolation can better reflect the stratum change law in the work area, and the accuracy of the interpolation result is improved.

Description

Technical Field

[0001] This invention relates to the field of geophysical exploration technology, and in particular to a well logging curve interpolation method, apparatus, equipment, medium, and program product. Background Technology

[0002] Well logging technology is an important means of obtaining information about underground strata in geological exploration. Well logging curves can reveal crucial information such as the physical, chemical, and mineralogical properties of the formation. However, since well logging can only be performed within the wellbore, for areas without wells, it is necessary to use methods such as interpolation to obtain pseudo-logging curves or even 3D logging curve data volumes for geological analysis or as initial models for further work.

[0003] The interpolation methods currently used for well logging curves are often based on the well logging curves themselves, such as inverse distance weighting (IDW) interpolation and kriging interpolation.

[0004] However, while the above interpolation methods can yield results for well logging curves, they also have some significant limitations. For example, inverse distance-weighted interpolation primarily assigns weights based on the distance between data points and the points to be interpolated, ignoring the spatial correlation and geological characteristics of the formation data. This results in interpolation results with a pronounced bullseye effect, inconsistent with geological conditions. While Kriging interpolation considers the spatial correlation of the data, its statistically based method requires prior estimation of the data's variogram, which is often difficult to obtain accurately in practical applications, and similarly fails to accurately reflect formation variation patterns. Summary of the Invention

[0005] This invention provides a well logging curve interpolation method, apparatus, equipment, medium, and program product. Based on the self-attention mechanism, the correlation between seismic traces and each well bypass trace is taken into account in the well logging curve interpolation process, highlighting the phase control factor and avoiding the bullseye phenomenon that may occur when using well logging curve interpolation alone. This allows the interpolated well logging curve to better reflect the formation change pattern in the work area and improves the accuracy of the interpolation results.

[0006] In a first aspect, embodiments of the present invention provide a well logging curve interpolation method, comprising:

[0007] Acquire raw seismic data and raw well logging data, and preprocess the raw seismic data and raw well logging data to determine the seismic data volume to be processed between target formations, the well-side seismic trace data matrix, and the well logging data matrix;

[0008] For each seismic trace data in the seismic data volume to be processed, the seismic trace data to be processed, the well-side seismic trace data matrix and the well logging curve data matrix are respectively used as the query vector, key vector and value vector in the self-attention mechanism to solve the target interpolated well logging curve corresponding to the seismic trace data to be processed.

[0009] Based on the preprocessing data processing method, the data of each target interpolation logging curve is reallocated to determine the logging curve interpolation result.

[0010] Secondly, embodiments of the present invention also provide a well logging curve interpolation device, comprising:

[0011] The data acquisition module is used to acquire raw seismic data and raw well logging curve data, and to preprocess the raw seismic data and raw well logging curve data to determine the seismic data volume to be processed between the target formations, the well-side seismic trace data matrix, and the well logging curve data matrix.

[0012] The target curve determination module is used to solve the target interpolated well logging curve corresponding to each seismic trace data in the seismic data volume to be processed by using the seismic trace data to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix as query vector, key vector, and value vector in the self-attention mechanism, respectively.

[0013] The interpolation result determination module is used to perform data realignment processing on each target interpolated logging curve according to the preprocessed data processing method, and determine the logging curve interpolation result.

[0014] Thirdly, embodiments of the present invention also provide a logging curve interpolation device, which includes:

[0015] At least one processor; and a memory communicatively connected to the at least one processor;

[0016] The memory stores a programmable program that can be executed by at least one processor, which is executed by at least one processor to enable the at least one processor to implement the logging curve interpolation method of any embodiment of the present invention.

[0017] Fourthly, embodiments of the present invention also provide a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform the logging curve interpolation method of any embodiment of the present invention.

[0018] Fifthly, embodiments of the present invention also provide a computer program product, including a computer program, which, when executed by a processor, is used to perform the logging curve interpolation method of any embodiment of the present invention.

[0019] This invention provides a well logging curve interpolation method, apparatus, equipment, medium, and program product. It acquires raw seismic data and raw well logging curve data, preprocesses them to determine the target formation's unprocessed seismic data volume, well-side seismic trace data matrix, and well logging curve data matrix. For each unprocessed seismic trace in the unprocessed seismic data volume, the unprocessed seismic trace data, well-side seismic trace data matrix, and well logging curve data matrix are respectively used as the query vector, key vector, and value vector in a self-attention mechanism for solving, determining the target interpolated well logging curve corresponding to the unprocessed seismic trace data. Based on the preprocessed data processing method, each target interpolated well logging curve undergoes data realignment processing to determine the well logging curve interpolation result. By adopting the above technical solution, the principle of self-attention mechanism is applied to well logging curve interpolation. Unlike the conventional self-attention mechanism where the query vector, key vector, and value vector all come from the same data source, this method uses the seismic data volume to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix determined based on the original seismic data and original well logging data within the work area as the query vector, key vector, and value vector in the self-attention mechanism for solution. This completes the well logging curve interpolation for each seismic trace in the original seismic data. This allows the correlation between the seismic trace and each well-side trace to be fully considered during the interpolation process, highlighting phase control factors and avoiding the bullseye phenomenon that may occur when using only well logging curve interpolation. As a result, the well logging curves obtained after interpolation can better reflect the formation variation patterns within the work area, improving the accuracy of the interpolation results.

[0020] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating a well logging curve interpolation method provided in Embodiment 1 of the present invention;

[0023] Figure 2 This is a flowchart illustrating a well logging curve interpolation method provided in Embodiment 2 of the present invention;

[0024] Figure 3 This is a flowchart illustrating a well logging curve interpolation method provided in Embodiment 3 of the present invention;

[0025] Figure 4 This is a three-dimensional representation of a channel sand deposition system provided in Embodiment 3 of the present invention;

[0026] Figure 5 This is an example image of a 102ms seismic amplitude slice of model data provided in Embodiment 3 of the present invention;

[0027] Figure 6 This is an example diagram of a model data well-crossing seismic profile provided in Embodiment 3 of the present invention;

[0028] Figure 7 This is an example diagram of a 102ms wave impedance slice of model data provided in Embodiment 3 of the present invention;

[0029] Figure 8 This is an example diagram of a wellbore impedance profile provided in Embodiment 3 of the present invention;

[0030] Figure 9 This is a curve example diagram of a well-side seismic trace data matrix provided in Embodiment 3 of the present invention;

[0031] Figure 10 This is an example curve diagram of a wave impedance curve matrix provided in Embodiment 3 of the present invention;

[0032] Figure 11 This is an example diagram of impedance slices for a 102ms phased interpolation wave model provided in Embodiment 3 of the present invention;

[0033] Figure 12 This is an example diagram of a well profile obtained by phased interpolation wave impedance of model data according to Embodiment 3 of the present invention;

[0034] Figure 13 This is an example diagram of a 102ms inverse distance weighted interpolation wave impedance slice provided in Embodiment 3 of the present invention;

[0035] Figure 14 This is an example diagram of a well-wave impedance profile obtained by inverse distance weighted interpolation of model data, provided in Embodiment 3 of the present invention.

[0036] Figure 15 This is a schematic diagram of the structure of a well logging curve interpolation device provided in Embodiment 4 of the present invention;

[0037] Figure 16 This is a schematic diagram of the structure of a well logging curve interpolation device provided in Embodiment 5 of the present invention. Detailed Implementation

[0038] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0039] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0040] Example 1

[0041] Figure 1 This is a flowchart illustrating a well logging curve interpolation method according to Embodiment 1 of the present invention. This embodiment is applicable to situations where, with the assistance of original seismic data, well logging curve interpolation is performed on seismic traces that do not contain the original well logging curve data, based on the original well logging curve data. This method can be executed by a well logging curve interpolation device, which can be implemented by software and / or hardware, and can be configured within a well logging curve interpolation equipment. Optionally, the well logging curve interpolation device can be an electronic device, such as a laptop, desktop computer, or smart tablet, etc. This embodiment of the present invention does not impose any limitations on this.

[0042] like Figure 1 As shown in the figure, the well logging curve interpolation method provided by the embodiment of the present invention specifically includes the following steps:

[0043] S101. Acquire raw seismic data and raw well logging curve data, and preprocess the raw seismic data and raw well logging curve data to determine the seismic data volume to be processed between the target formations, the well-side seismic trace data matrix, and the well logging curve data matrix.

[0044] In this embodiment, the raw seismic data can be specifically understood as the seismic data directly collected from each seismic trace during manual seismic surveys within the work area. The raw well logging curve data can be specifically understood as the curve data obtained by continuously recording various parameters varying with depth along the wellbore using various logging instruments inserted into each well within the work area. It can be understood that each type of parameter in each well within the work area can generate a corresponding well logging curve. Preprocessing can be specifically understood as preprocessing the raw seismic data and raw well logging curve data before well logging curve interpolation, preserving valid information and standardizing their format. The target formation can be specifically understood as the formation requiring geophysical exploration; it can be understood that target formations often contain targets of geophysical exploration interest, such as oil, gas, coal, and metals. The seismic data volume to be processed can be specifically understood as the collection of seismic data from each seismic trace at depths between the target formations, obtained after preprocessing the raw seismic data. The well-side seismic trace data matrix can be understood as a two-dimensional matrix composed of seismic data within the seismic data volume to be processed that correspond to the wells corresponding to the original well logging curve data. Specifically, the seismic trace closest to the well corresponding to the original well logging curve data is identified as the well-related well-side seismic trace. The well-side seismic trace data matrix is ​​obtained by sorting and combining the seismic data corresponding to each well-side seismic trace within the seismic data volume to be processed according to the well location. The well logging curve data matrix can be understood as a two-dimensional matrix obtained by sorting and combining the pre-processed well logging curves according to the well location.

[0045] Specifically, when well logging curve interpolation is required for the work area, the original seismic data and original well logging curve data obtained from artificial seismic exploration and well logging exploration within the work area will be acquired first. The original well logging curve data and original seismic data will then be standardized in terms of format, sampling rate, and length. Data located between the target strata will be extracted for preprocessing, resulting in a seismic data volume to be processed corresponding to the original seismic data, a well logging curve data matrix corresponding to the original well logging curve data, and a well-side seismic trace data matrix associated with both the original seismic data and the original well logging curve data.

[0046] S102. For each seismic trace data to be processed in the seismic data volume to be processed, the seismic trace data to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix are respectively used as the query vector, key vector, and value vector in the self-attention mechanism to solve the problem and determine the target interpolated well logging curve corresponding to the seismic trace data to be processed.

[0047] In this embodiment, the seismic trace data to be processed can be specifically understood as the seismic trace data corresponding to each seismic trace in the seismic data body to be processed. The self-attention mechanism can be specifically understood as a mechanism used to dynamically capture the dependencies between elements at different positions in the sequence when processing sequence data, and to generate new sequence representations based on these dependencies. The query vector can be specifically understood as the element currently being focused on in the self-attention mechanism or the target information to be obtained. The key vector can be specifically understood as a vector used to compare with the query vector to determine which elements in the sequence are most relevant to the current query. The value vector can be specifically understood as containing the actual information content, and is weighted and summed according to the degree of matching between the query vector and the key vector to finally form the output vector. The target interpolated logging curve can be specifically understood as the logging curve data related to the seismic trace corresponding to the seismic trace data to be processed, obtained after interpolating the logging curve data matrix.

[0048] Specifically, self-attention mechanisms are commonly used to extract dependencies between data sequences in an input data volume and generate new data sequences as output. In this embodiment of the invention, only the operational relationship between query vectors, key vectors, and value vectors in the self-attention mechanism is introduced. Seismic trace data to be processed, well-side seismic trace data matrices, and well logging curve data matrices from different data sources are used as query vectors, key vectors, and value vectors respectively for computation. Since the well-side seismic trace data and well logging curve data are fixed, but the seismic trace data to be processed is different each time, the well logging curve data matrix can be weighted according to the attention relationship between the input seismic trace data to be processed and the well-side seismic trace data matrix each time, obtaining the well logging curve corresponding to the seismic trace data to be processed. This obtained well logging curve is then determined as the target interpolated well logging curve corresponding to the seismic trace data to be processed.

[0049] In this embodiment of the invention, the principle of self-attention mechanism is applied to well logging curve interpolation. The input is adjusted to be seismic trace data to be processed, well-side seismic trace data matrix, and well logging curve data matrix from different data sources. This means that when interpolating well logging curves, the distance between the well logging curves themselves is no longer considered. Instead, attention relationship is solved between the seismic trace data to be interpolated and the well-side seismic trace data with the highest correlation to the well logging curve data. This allows the target interpolated well logging curve obtained after weighting the well logging curve data matrix according to the attention relationship to fully consider the correlation between the seismic traces and each well-side trace. At the same time, the calculation process of the target interpolated well logging curve relies solely on pure data-driven calculation, without the need for variogram prediction, correlation, or seismic phase calculation. It is not limited by distance and direction, and can better adapt to the well logging curve data interpolation requirements under various geological conditions, thereby improving the interpolation accuracy.

[0050] S103. Based on the preprocessing data processing method, perform data realignment processing on each target interpolation logging curve to determine the logging curve interpolation result.

[0051] In this embodiment, data repositioning can be understood as the opposite of preprocessing, which enables the processed data to be repositioned to the original data format collected within the work area.

[0052] Specifically, for each target interpolated logging curve, the target interpolated logging curve is reverse-processed according to the preprocessed data processing method. That is, the sampling rate, length and vertex time value of the target interpolated logging curve are restored to the actual number of sampling points and the actual time value of the seismic trace corresponding to the target interpolated logging curve in the target formation. This realizes the vertical positioning of the target interpolated logging curve on the corresponding seismic trace. After the vertical positioning of all target interpolated logging curves is completed, the logging curve interpolation results in the work area are obtained.

[0053] It is understandable that, when performing well logging curve interpolation, well logging curve interpolation can be performed only on the seismic traces in the non-well-side seismic trace data matrix of the seismic data body to be processed, and the well logging curve data in the well logging curve data matrix can be used as the well logging curves on each seismic trace in the well-side seismic trace data matrix; alternatively, well logging curve interpolation can be performed on all seismic traces in the seismic data body to be processed, and the well logging curve interpolation results of all seismic traces in the work area can be obtained. Both of the above methods can be adopted, and the embodiments of the present invention do not limit this.

[0054] The technical solution of this embodiment acquires raw seismic data and raw well logging curve data, and preprocesses the raw seismic data and raw well logging curve data to determine the seismic data volume to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix between target formations. For each seismic trace data to be processed in the seismic data volume, the seismic trace data to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix are respectively used as the query vector, key vector, and value vector in the self-attention mechanism for solving to determine the target interpolated well logging curve corresponding to the seismic trace data to be processed. According to the preprocessed data processing method, the data of each target interpolated well logging curve is reset to determine the well logging curve interpolation result. By adopting the above technical solution, the principle of self-attention mechanism is applied to well logging curve interpolation. Unlike the conventional self-attention mechanism where the query vector, key vector, and value vector all come from the same data source, this method uses the seismic data volume to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix determined based on the original seismic data and original well logging data within the work area as the query vector, key vector, and value vector in the self-attention mechanism for solution. This completes the well logging curve interpolation for each seismic trace in the original seismic data. This allows the correlation between the seismic trace and each well-side trace to be fully considered during the interpolation process, highlighting phase control factors and avoiding the bullseye phenomenon that may occur when interpolating only well logging curves. As a result, the well logging curves obtained after interpolation can better reflect the formation variation patterns within the work area, improving the accuracy of the interpolation results.

[0055] Example 2

[0056] Figure 2This is a flowchart illustrating a well logging curve interpolation method provided in Embodiment 2 of the present invention. Based on the above optional implementation schemes, the present invention further optimizes the method. After obtaining the original seismic data and the original well logging curve data, the original seismic data and the original well logging curve data are first unified and aligned in form through time-depth conversion and calibration processing to obtain the corresponding intermediate well logging curve data. Then, through resampling, destination interlayer data extraction and data length unification processing, the original seismic data and the intermediate well logging curve data are converted into a seismic data volume to be processed, a well-side seismic trace data matrix and a well logging curve data matrix that can be used for self-attention mechanism operation. For each seismic trace to be processed, its information related to the well-side seismic trace data matrix is ​​substituted into the pre-configured attention vector determination formula to determine its attention vector related to each well-side seismic trace. Since there is a correlation and correspondence between each well-side seismic trace and the logging curve, the attention vector between the seismic trace to be processed and the well-side seismic trace data matrix can also be used to indicate the attention relationship between the seismic trace corresponding to the seismic trace to be processed and the logging curve. Therefore, multiplying the attention vector with the logging curve data yields the target interpolated logging curve that fully considers the correlation between the seismic traces in the work area and the well-side seismic traces, highlighting the phase control factors, so that the logging curve obtained after interpolation can better reflect the formation variation law in the work area and improve the accuracy of the interpolation results.

[0057] like Figure 2 As shown in the figure, the well logging curve interpolation method provided by the embodiment of the present invention specifically includes the following steps:

[0058] S201. Obtain raw seismic data and raw well logging curve data.

[0059] S202. Perform time-depth conversion and calibration processing on the original well logging curve data to determine the intermediate well logging curve data that is aligned with the original seismic data.

[0060] Specifically, since the original well logging curve data is depth curve data and the original seismic data is time-dimensional data, in order to determine the correlation between the two so that the well logging curve can be interpolated based on the seismic data, the original well logging curve data can first be converted from depth dimension to time dimension. Then, after the time-depth conversion is completed, the converted original well logging curve data and the original seismic data are calibrated according to the same time point, so that the intermediate well logging curve data obtained after calibration can be aligned with the seismic data in terms of geological characteristics.

[0061] S203. Resample the original seismic data and intermediate logging curve data, extract the destination interlayer data, and unify the data length to determine multiple seismic trace data sequences corresponding to the original seismic data, as well as multiple logging curve data sequences corresponding to the intermediate logging curve data.

[0062] Specifically, the original seismic data and intermediate well logging data are resampled to ensure that the resampled seismic data and well logging data contain the same number of sampling points within the same time period. This guarantees the analytical correspondence between the sampling points of corresponding seismic data and well logging curves. Furthermore, since geophysical exploration focuses only on information about the inter-strata regions potentially containing oil and gas deposits, it can be assumed that only well logging interpolation between the target strata is needed. Therefore, data extraction is performed from the resampled original seismic data and intermediate well logging data based on the time information corresponding to the target strata. Also, because the intermediate well logging data may not be continuous, meaning the extracted well logging data may not fully cover the time between the target strata, zero-padding or interpolation can be performed on the seismic data of each extracted seismic trace and the well logging data corresponding to each well to achieve uniform data length in the vertical direction. After the above processing steps, the data corresponding to each seismic trace in the original seismic data is processed into a seismic trace data sequence, and the data corresponding to each well in the intermediate well logging curve data is processed into a well logging curve data sequence.

[0063] For example, each processed seismic trace data sequence can be denoted as S_well, and each well logging curve data sequence can be denoted as C_well. The data length of each S_well and each C_well should be the same, and can be denoted as d_model. The d_model should be greater than or equal to the length of the longest seismic trace data after resampling and extraction of inter-layer data at the destination, and preferably an integer multiple of 8.

[0064] S204. Arrange the seismic trace data sequences according to the seismic trace order in the original seismic data to determine the seismic data volume to be processed.

[0065] Specifically, since each seismic trace data sequence corresponds to the data of each seismic trace in the original seismic data, and each seismic trace has its own unique line and trace number, the seismic trace data sequences can be arranged in order according to the line and trace numbers of each seismic trace in the original seismic data, and the resulting three-dimensional data tensor is determined as the seismic data volume to be processed.

[0066] Following the example above, since each seismic trace data sequence is a data sequence of the same length d_model obtained after zero padding or interpolation, arranging them according to the line number of the seismic trace will yield a seismic data body S_cube of equal length to be processed. Its dimensions can be represented as (n_line, n_cmp, d_model), where n_line and n_cmp are the number of lines and traces in the seismic exploration within the work area, respectively.

[0067] S205. Arrange the well logging curve data sequences according to the well location to determine the well logging curve data matrix.

[0068] Specifically, since each logging curve data sequence corresponds to a well in the work area, and each well has its own unique well location information, the well location information can be used to arrange the logging curve data sequences, and the resulting two-dimensional matrix can be determined as the logging curve data matrix.

[0069] For example, well location information may include coordinates, well name, and well identification number, etc., to identify the well. This embodiment of the invention does not impose any limitations on this. It is understood that when arranging each logging curve data sequence using well location information, the arrangement method of each logging curve data sequence is not limited. It is only necessary to arrange them based on parameters of the same type in the well location information during one arrangement process.

[0070] Following the example above, assuming n_well is the number of well locations, the well logging data matrix C_well_2D obtained by arranging the well logging data sequences according to the well locations can be represented by the dimension (n_well, d_model).

[0071] S206. The seismic trace data sequence corresponding to each well logging curve data sequence is determined as the well-side seismic trace data sequence, and each well-side seismic trace data sequence is arranged according to the well location to form a well-side seismic trace data matrix.

[0072] Specifically, since the well-side seismic trace data sequence can be understood as the seismic data sequence located geographically near the well corresponding to the well logging curve data sequence, that is, the seismic trace data sequence with the same geological characteristics as the well logging curve data sequence, the seismic trace data sequence that is geographically closest to each well logging curve data sequence can be selected as the well-side seismic trace data sequence. A two-dimensional matrix is ​​then constructed using the well location arrangement corresponding to the well logging curve data sequence, and this two-dimensional matrix is ​​determined as the well-side seismic trace data matrix.

[0073] Following the example above, the well-side seismic trace data matrix S_well_2D obtained by arranging the S_wells corresponding to each well point according to the well location can be represented by the dimension (n_well, d_model).

[0074] S207. For each seismic trace in the seismic data body to be processed, determine the attention vector corresponding to the seismic trace based on the seismic trace data to be processed, the well-side seismic trace data matrix, and the pre-configured attention vector determination formula.

[0075] The formula for determining the attention vector is as follows:

[0076]

[0077] Where Q represents the seismic trace data to be processed, K represents the well-side seismic trace data matrix, and d k represents the width of the well-side seismic trace data matrix, sqrt() is the square root function, and softmax() is the normalization function.

[0078] Specifically, seismic trace data to be processed is extracted sequentially from the seismic data volume according to the trace number, and used as the query vector in the self-attention mechanism. The well-side seismic trace data matrix is ​​used as the key vector in the self-attention mechanism. The transpose and matrix width of the key vector are substituted into the attention vector determination formula. The attention vector determination formula is solved to obtain the attention vector corresponding to the seismic trace data to be processed, which can indicate the correlation between the seismic trace data to be processed and the well-side seismic trace data matrix.

[0079] S208. The product of the attention vector and the logging curve data matrix is ​​determined as the target interpolated logging curve corresponding to the seismic trace data to be processed.

[0080] Specifically, since there is a correlation in geological characteristics between the well logging curve data matrix and the well-side seismic trace data matrix, the attention vector can also reflect the correlation between the location of the seismic trace data to be interpolated and the well logging curve data matrix. At this time, the attention vector is multiplied by the well logging curve data matrix, which is the median vector of the self-attention mechanism, and the output result is determined as the target interpolated well logging curve corresponding to the seismic trace data to be interpolated.

[0081] S209. Based on the preprocessing data processing method, perform data realignment processing on each target interpolation logging curve to determine the logging curve interpolation result.

[0082] Specifically, the data processing operation is reversed in S203. The obtained target interpolated logging curves are zero-padding or de-interpolation processed to restore the true number of sampling points of the target interpolated logging curves between the target layers. Then, the top surface time value of the target layer is added to the sampling point sequence number to realize the vertical positioning of the target interpolated logging curves. After processing all target interpolated logging curves, they are placed at the corresponding seismic traces in the original seismic data to obtain the logging curve interpolation results in the work area.

[0083] The technical solution of this embodiment, after acquiring the original seismic data and the original well logging curve data, first completes the form unification and alignment of the original seismic data and the original well logging curve data through time-depth conversion and calibration processing to obtain the corresponding intermediate well logging curve data. Then, through resampling, destination interlayer data extraction and data length unification processing, the original seismic data and the intermediate well logging curve data are converted into a seismic data volume to be processed, a well-side seismic trace data matrix and a well logging curve data matrix that can be used for self-attention mechanism operation. For each seismic trace to be processed, its information related to the well-side seismic trace data matrix is ​​substituted into the pre-configured attention vector determination formula to determine its attention vector related to each well-side seismic trace. Since there is a correlation and correspondence between each well-side seismic trace and the logging curve, the attention vector between the seismic trace to be processed and the well-side seismic trace data matrix can also be used to indicate the attention relationship between the seismic trace corresponding to the seismic trace to be processed and the logging curve. Therefore, multiplying the attention vector with the logging curve data yields the target interpolated logging curve that fully considers the correlation between the seismic traces in the work area and the well-side seismic traces, highlighting the phase control factors, so that the logging curve obtained after interpolation can better reflect the formation variation law in the work area and improve the accuracy of the interpolation results.

[0084] Example 3

[0085] Figure 3 This is a flowchart illustrating a well logging curve interpolation method provided in Embodiment 3 of the present invention. Based on the aforementioned optional technical solutions, this embodiment further optimizes the well logging curve interpolation process by introducing a well logging curve interpolation model that simultaneously includes an attention interpolation module and an optimization interpolation module. The aforementioned well logging curve interpolation processing method is integrated into the attention interpolation module, and an optimization interpolation module is subsequently added to perform a second optimization of the output well logging interpolation curve. This approach retains the aforementioned emphasis on phase control factors and avoids the bullseye phenomenon that may occur when relying solely on well logging curve interpolation. This allows the interpolated well logging curve to better reflect the formation variation patterns within the work area, thereby improving the accuracy of the interpolation results. Furthermore, the model's computational efficiency is higher, increasing the computational speed of well logging curve interpolation and better meeting the data processing needs of geophysical exploration.

[0086] like Figure 3 As shown in the figure, the well logging curve interpolation method provided by the embodiment of the present invention specifically includes the following steps:

[0087] S301. Acquire raw seismic data and raw well logging curve data, and preprocess the raw seismic data and raw well logging curve data to determine the seismic data volume to be processed between the target formations, the well-side seismic trace data matrix, and the well logging curve data matrix.

[0088] S302. Based on the seismic trace data matrix and logging curve data matrix near the well, a logging curve interpolation model is trained.

[0089] The well logging curve interpolation model includes an attention interpolation module and an optimization interpolation module.

[0090] The data processing method of the attention interpolation module is consistent with the processing method of using the seismic trace data to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix as query vector, key vector, and value vector in the self-attention mechanism, respectively.

[0091] In this embodiment, the well logging curve interpolation model can be specifically understood as a neural network model used to perform interpolation processing on the input data based on a self-attention mechanism, and to extract features and optimize predictions on the interpolated data. The optimized interpolation module can be specifically understood as a neural network module used to optimize the input well logging curve interpolation data based on trained and fixed parameters, and to output data in a form that is closer to the actual well logging curve data.

[0092] Specifically, before performing well logging curve interpolation, different work areas have different geological characteristics. If well logging curve interpolation is required for a specific work area, a well logging curve interpolation model suitable for the geological conditions of that work area needs to be trained. At this time, a training sample set can be constructed based on the well-side seismic trace data matrix and well logging curve data matrix that meet the requirements after processing and have clear geological characteristics. A well logging curve interpolation model containing an attention interpolation module and an optimization interpolation module can be obtained by training the constructed training sample set.

[0093] In this embodiment of the invention, based on the processed well-side seismic trace data matrix and well logging curve data matrix, a well logging curve interpolation model suitable for the work area is specifically constructed. This ensures the adaptability of subsequent well logging curve interpolation processing to the work area data and improves interpolation accuracy. Simultaneously, since the trained well logging curve interpolation model already includes an attention interpolation module capable of implementing the well logging curve interpolation data processing methods provided in the above embodiments, the effects of the well logging curve interpolation processing in the above embodiments are preserved. Furthermore, adding an optimized interpolation module to the well logging curve interpolation model achieves a second optimization of the well logging curve interpolation data obtained from simple data computation, further improving the accuracy of the interpolation results.

[0094] S303. For each seismic trace in the seismic data body to be processed, input the seismic trace data to be processed into the attention interpolation module in the well logging curve interpolation model to determine the intermediate interpolation well logging curve.

[0095] Specifically, for each seismic trace in the seismic data body to be processed, the seismic trace to be processed is used as the input of the attention interpolation module in the well logging curve interpolation model. After processing it using the well logging curve interpolation method provided in the above embodiments, the output result of the attention interpolation module is determined as the intermediate interpolated well logging curve.

[0096] S304. Input the intermediate interpolated logging curve into the optimized interpolation module in the logging curve interpolation model to determine the target interpolated logging curve corresponding to the seismic trace data to be processed.

[0097] Specifically, the intermediate interpolated logging curve is used as the input to the optimization interpolation module in the logging curve interpolation model. The optimization interpolation module optimizes and predicts the intermediate interpolated logging curve, and the output of the optimization interpolation module is determined as the target interpolated logging curve corresponding to the seismic trace data to be processed.

[0098] S305. Based on the preprocessing data processing method, perform data realignment processing on each target interpolation logging curve to determine the logging curve interpolation result.

[0099] Understandably, for the interpretation of seismic data, the wave impedance curve is the most important logging curve, and wave impedance data has a clear functional relationship with seismic amplitude. That is, the reflection coefficient R at the interface between the upper and lower strata... ef The earthquake reflection coefficient formula R can be used. ef = (Z2-Z1) / (Z2+Z1) can be obtained, where Z1 and Z2 are the wave impedances of the upper and lower strata above the reflection interface, respectively; and the seismic amplitude can be obtained by the convolution formula S = W*R ef The result is obtained, where S represents seismic data, W represents wavelet, and * represents the convolution operator. The process of calculating seismic reflection amplitude data from wave impedance data is called forward modeling, and the result is called the seismic synthetic record. Therefore, based on the characteristics of wave impedance data, this embodiment of the invention provides two different training methods for well logging curve interpolation models.

[0100] For example, when the logging curve type is a wave impedance curve, a logging curve interpolation model is trained based on the well-side seismic trace data matrix and the logging curve data matrix. This can be achieved in the following way:

[0101] 1) Substitute the well-side seismic trace data matrix and well logging curve data matrix into the attention interpolation module, and extract multiple seismic trace data to be processed from the seismic data volume to be processed as training seismic trace data.

[0102] 2) Construct a training sample set with each training seismic trace as a sample and the same training seismic trace as a label.

[0103] 3) Input each training sample in the training sample set into the initial well logging curve interpolation model in sequence, and perform forward modeling on the output of the initial well logging curve interpolation model to determine the intermediate seismic trace data.

[0104] In this embodiment, the initial logging curve interpolation model can be specifically understood as a logging curve interpolation model that has not been trained with parameter adjustments.

[0105] Specifically, based on the above embodiments, the attention interpolation module requires the use of the well-side seismic trace data matrix and the well logging curve data matrix. Therefore, the well-side seismic trace data matrix and the well logging curve data matrix can be directly substituted into the attention interpolation module to complete the configuration of the attention interpolation module. Simultaneously, since the well logging curve is a wave impedance curve, it can be back-modeled into seismic trace data based on the wave impedance curve output by the well logging curve interpolation model. Therefore, when training the wave impedance curve type well logging curve interpolation model, multiple seismic trace data to be processed can be directly extracted from the seismic data volume to be processed, and these can be used simultaneously as training seismic trace data and corresponding labels. Each set of training seismic trace data and its corresponding label is used as a training sample to construct a training sample set. Then, each training sample in the training sample set is sequentially input into the initial well logging curve interpolation model that has not been trained on the parameter tuning network in the optimized interpolation module. The output of the initial well logging curve interpolation model is then back-modeled to obtain intermediate seismic trace data. Since the intermediate seismic trace data is obtained by forward modeling the wave impedance curves interpolated from the initial well logging curve interpolation model, it theoretically needs to be as close as possible to the actual training seismic trace data to ensure the accuracy of the well logging curve interpolation data output by the well logging curve interpolation model. Therefore, a loss function can be constructed based on each intermediate seismic trace data and the training seismic trace data in each training sample, and the optimized interpolation module in the initial well logging curve interpolation model can be iteratively optimized based on the loss function. After training is completed, the well logging curve interpolation model can be obtained.

[0106] It is understood that when extracting training seismic trace data from the seismic data body to be processed, all seismic trace data to be processed in the seismic data body to be processed can be extracted, or only a portion of the seismic trace data to be processed can be extracted. The more seismic trace data to be processed extracted, the larger the data volume of the training sample set is, and the more accurate the training of the well logging curve interpolation model will be, but the training speed will be slower. The number of training seismic trace data to be extracted can be determined according to actual needs, and the embodiments of the present invention do not impose any restrictions on this.

[0107] For example, when the logging curve type is not a wave impedance curve, a logging curve interpolation model is trained based on the well-side seismic trace data matrix and the logging curve data matrix. This can be achieved in the following way:

[0108] 1) Substitute the well-side seismic trace data matrix and well logging curve data matrix into the attention interpolation module.

[0109] 2) Construct a training sample set with each well-side seismic trace data sequence in the well-side seismic trace data matrix as a sample and the well logging curve data sequence corresponding to each well-side seismic trace data sequence as a label.

[0110] 3) The initial logging curve interpolation model is trained using the training sample set to obtain the logging curve interpolation model.

[0111] Specifically, based on the above embodiments, the attention interpolation module requires the use of the well-side seismic trace data matrix and the well logging curve data matrix. Therefore, the well-side seismic trace data matrix and the well logging curve data matrix can be directly substituted into the attention interpolation module to complete the configuration of the attention interpolation module. Simultaneously, for well logging curves of non-impedance curve types, each well-side seismic trace data sequence in the well-side seismic trace data matrix can be directly used as a sample, and the corresponding well logging curve data sequence can be used as a label to construct training samples. The initial well logging curve interpolation model is trained using a training sample set constructed from multiple training samples, resulting in the well logging curve interpolation model. The loss function required during training is based on the output of the initial well logging curve interpolation model and the well logging curve data sequence.

[0112] Optionally, to more clearly verify the interpolation effect of well logging curves, this embodiment of the invention also provides a set of three-dimensional model data as an example to further illustrate the implementation process of this embodiment. The three-dimensional model data uses a channel sand deposition system in a horizontally layered formation as an example. Figure 4 This is a three-dimensional representation of a channel sand sedimentary system provided in Embodiment 3 of the present invention. The data includes, for example: Figure 4 The model's impedance data volume includes the seismic data volume generated from the model's impedance through convolution forward modeling, and the impedance curves of eight wells extracted from the model's impedance data volume. The impedance data volume and the seismic data volume have values ​​of 200, 131, and 198 in the line, cmp, and time dimensions, respectively. To more clearly illustrate the model data, this embodiment also provides the following images, using a horizontal slice at 102ms and a longitudinal profile through the wells as examples to demonstrate the model data. Figure 5 This is an example image of a 102ms seismic amplitude slice of model data provided in Embodiment 3 of the present invention. Figure 6 This is an example diagram of a model data well-crossing seismic profile provided in Embodiment 3 of the present invention. Figure 7 This is an example diagram of a 102ms wave impedance slice of model data provided in Embodiment 3 of the present invention. Figure 8 This is an example diagram of a model data wellbore impedance profile provided in Embodiment 3 of the present invention.

[0113] In this embodiment, the wave impedance data, seismic data, and wave impedance curves of the aforementioned model are resampled at a sampling rate of 1ms. Since the above data are all time-domain data, no alignment operation is required. If it were actual data, time-depth calibration would be required through synthetic records to align the wave impedance curves with the seismic data. Because the formation is horizontally layered, this embodiment directly uses all vertical data for calculation, with a vertical range of 1–198ms and a sampling rate of 1ms, totaling 198 sample points. This is equivalent to 1ms for the top layer and 198ms for the bottom layer. To facilitate subsequent network operations, d_model = 200 is used; therefore, the seismic trace data and wave impedance curves at the well points need to be padded with two zeros. Since there are a total of 8 wells, the seismic trace data sequence and wave impedance curve sequence are arranged in ascending order of trace number, resulting in a two-dimensional matrix S_well_2D of the well-side seismic trace data and a two-dimensional matrix C_well_2D of the wave impedance curves, both with dimensions (8, 200). Figure 9 This is an example curve diagram of a well-side seismic trace data matrix provided in Embodiment 3 of the present invention. Figure 10 This is an example curve diagram of a wave impedance curve matrix provided in Embodiment 3 of the present invention. Seismic traces are then extracted from the original seismic record and zero-padding is performed to obtain S_cube, with dimensions (200, 131, 200). The S_cube data is processed using the well logging curve interpolation method and inverse distance weighted interpolation method provided in the embodiments of the present invention, taking 102ms horizontal slices and longitudinal profile data through the well as examples. Figure 11 This is an example diagram of impedance slices for a 102ms phased interpolation wave model provided in Embodiment 3 of the present invention. Figure 12 This is an example diagram of a well-pass profile using phased interpolation wave impedance based on model data, provided in Embodiment 3 of the present invention. Figure 13 This is an example diagram of a 102ms inverse distance weighted interpolation wave impedance slice provided in Embodiment 3 of the present invention. Figure 14 This is an example diagram of a well-wave impedance profile obtained by inverse distance weighted interpolation of model data according to Embodiment 3 of the present invention. It is clearly evident from this diagram that… Figure 11 and Figure 7 , Figure 12 and Figure 8 The displayed cross-sectional effect is consistent, and Figure 13 and Figure 14 Only the values ​​near the well point match the model data, while the rest exhibit a "bull's eye phenomenon" centered on the well point, which is obviously incorrect. This shows that the results of the embodiments of the present invention can better reflect the depositional phenomena of the reservoir and achieve the purpose of phase-controlled interpolation.

[0114] The technical solution of this embodiment introduces a well logging curve interpolation model that simultaneously includes an attention interpolation module and an optimization interpolation module during the well logging curve interpolation process. The aforementioned well logging curve interpolation processing method is integrated into the attention interpolation module, and an optimization interpolation module is added subsequently to perform a second optimization of the output well logging interpolation curve. This approach retains the factors highlighted by phasing and avoids the bullseye phenomenon that may occur with interpolation focused solely on well logging curves. This ensures that the interpolated well logging curve better reflects the formation variation patterns within the work area, thereby improving the accuracy of the interpolation results. Furthermore, the model's computational efficiency is higher, increasing the computational speed for well logging curve interpolation and better meeting the data processing needs of geophysical exploration.

[0115] Example 4

[0116] Figure 15 This is a schematic diagram of the structure of a well logging curve interpolation device provided in Embodiment 4 of the present invention, as shown below. Figure 15 As shown, the well logging curve interpolation device includes a data acquisition module 41, a target curve determination module 42, and an interpolation result determination module 43.

[0117] The data acquisition module 41 is used to acquire raw seismic data and raw well logging curve data, and preprocess the raw seismic data and raw well logging curve data to determine the seismic data volume to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix between the target formations. The target curve determination module 42 is used to solve the target interpolated well logging curve corresponding to each seismic trace data to be processed in the seismic data volume by using the seismic trace data to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix as query vector, key vector, and value vector in the self-attention mechanism, respectively. The interpolation result determination module 43 is used to perform data repositioning processing on each target interpolated well logging curve according to the preprocessed data processing method to determine the well logging curve interpolation result.

[0118] The technical solution of this invention applies the principle of self-attention mechanism to well logging curve interpolation. Unlike conventional self-attention mechanisms where the query vector, key vector, and value vector all originate from the same data, this invention uses the seismic data volume to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix determined based on the original seismic data and original well logging data within the work area as the query vector, key vector, and value vector in the self-attention mechanism for solution. This completes the well logging curve interpolation for each seismic trace in the original seismic data. This allows the correlation between the seismic trace and each well-side trace to be fully considered during the interpolation process, highlighting phase control factors and avoiding the bullseye phenomenon that may occur when interpolating only well logging curves. As a result, the interpolated well logging curves can better reflect the formation variation patterns within the work area, improving the accuracy of the interpolation results.

[0119] Optionally, the data acquisition module 41 is specifically used for:

[0120] The original well logging curve data were converted to time and depth and calibrated to determine intermediate well logging curve data that were aligned with the original seismic data.

[0121] The original seismic data and intermediate logging curve data were resampled, the target interlayer data were extracted, and the data length was uniformly processed to determine multiple seismic trace data sequences corresponding to the original seismic data and multiple logging curve data sequences corresponding to the intermediate logging curve data.

[0122] The seismic trace data sequences are arranged in the order of the seismic traces in the original seismic data and determined as the seismic data volume to be processed.

[0123] The data sequences of each logging curve are arranged according to the well location to form a logging curve data matrix;

[0124] The seismic trace data sequence corresponding to each well logging curve data sequence is determined as the well-side seismic trace data sequence, and each well-side seismic trace data sequence is arranged according to the well location to form the well-side seismic trace data matrix.

[0125] Optionally, the target curve determination module 42 is specifically used for:

[0126] Based on the seismic trace data to be processed, the well-side seismic trace data matrix, and the pre-configured attention vector determination formula, the attention vector corresponding to the seismic trace data to be processed is determined.

[0127] The product of the attention vector and the logging curve data matrix is ​​used to determine the target interpolated logging curve corresponding to the seismic trace data to be processed;

[0128] The formula for determining the attention vector is as follows:

[0129]

[0130] Where Q represents the seismic trace data to be processed, K represents the well-side seismic trace data matrix, and d k represents the width of the well-side seismic trace data matrix, sqrt() is the square root function, and softmax() is the normalization function.

[0131] Optionally, the well logging curve interpolation device also includes:

[0132] After determining the seismic data volume to be processed between the target formations, the well-side seismic trace data matrix, and the well logging curve data matrix, a well logging curve interpolation model is trained based on the well-side seismic trace data matrix and the well logging curve data matrix; the well logging curve interpolation model includes an attention interpolation module and an optimization interpolation module;

[0133] For each seismic trace in the seismic data body to be processed, the seismic trace data to be processed is input into the attention interpolation module in the well logging curve interpolation model to determine the intermediate interpolation well logging curve.

[0134] The intermediate interpolated logging curves are input into the optimized interpolation module in the logging curve interpolation model to determine the target interpolated logging curve corresponding to the seismic trace data to be processed.

[0135] The data processing method of the attention interpolation module is consistent with the processing method of using the seismic trace data to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix as query vector, key vector, and value vector in the self-attention mechanism, respectively.

[0136] Optionally, when the logging curve type is a wave impedance curve, a logging curve interpolation model is trained based on the well-side seismic trace data matrix and the logging curve data matrix, including:

[0137] Substitute the well-side seismic trace data matrix and the well logging curve data matrix into the attention interpolation module, and extract multiple seismic trace data from the seismic data volume to be processed as training seismic trace data.

[0138] Construct a training sample set with each training seismic trace as a sample and the same training seismic trace as a label;

[0139] Each training sample in the training sample set is sequentially input into the initial well logging curve interpolation model, and the output of the initial well logging curve interpolation model is forward modeled to determine the intermediate seismic trace data.

[0140] A loss function is constructed based on the intermediate seismic trace data and the training seismic trace data in each training sample. The optimization interpolation module is then trained using the loss function to obtain the well logging curve interpolation model.

[0141] Optionally, when the logging curve type is not a wave impedance curve, a logging curve interpolation model is trained based on the well-side seismic trace data matrix and the logging curve data matrix, including:

[0142] Substitute the seismic trace data matrix and logging curve data matrix from the well into the attention interpolation module;

[0143] A training sample set is constructed, with each well-side seismic trace data sequence in the well-side seismic trace data matrix as a sample and the well logging curve data sequence corresponding to each well-side seismic trace data sequence as a label;

[0144] The initial logging curve interpolation model is trained using a training sample set to obtain the logging curve interpolation model.

[0145] The logging curve interpolation device provided in this embodiment of the invention can execute the logging curve interpolation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0146] Example 5

[0147] Figure 16 This is a schematic diagram of a well logging curve interpolation device according to Embodiment 5 of the present invention. The well logging curve interpolation device 50 can be a digital computer of various forms, such as a laptop computer, desktop computer, workbench, personal digital assistant, server, blade server, mainframe computer, and other suitable computers. The well logging curve interpolation device 50 can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0148] like Figure 16 As shown, the logging curve interpolation device 50 includes at least one processor 51 and a memory, such as a read-only memory (ROM) 52 and a random access memory (RAM) 53, communicatively connected to the at least one processor 51. The memory stores a programmable program executable by the at least one processor. The processor 51 can perform various appropriate actions and processes based on the programmable program stored in the ROM 52 or loaded from the storage unit 58 into the RAM 53. The RAM 53 can also store various programs and data required for the operation of the logging curve interpolation device 50. The processor 51, ROM 52, and RAM 53 are interconnected via a bus 54. An input / output (I / O) interface 55 is also connected to the bus 54. Optionally, the processor can be an FPGA.

[0149] Multiple components in the logging curve interpolation device 50 are connected to the I / O interface 55, including: an input unit 56, such as a keyboard, mouse, etc.; an output unit 57, such as various types of displays, speakers, etc.; a storage unit 58, such as a disk, optical disk, etc.; and a communication unit 59, such as a network card, modem, wireless transceiver, etc. The communication unit 59 allows the logging curve interpolation device 50 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0150] Processor 51 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 51 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 51 performs the various methods and processes described above, such as well logging curve interpolation methods.

[0151] In some embodiments, the logging curve interpolation method may be implemented as a programmable program tangibly contained in a computer-readable storage medium, such as storage unit 58. In some embodiments, part or all of the programmable program may be loaded and / or installed onto the logging curve interpolation device 50 via ROM 52 and / or communication unit 59. When the programmable program is loaded into RAM 53 and executed by processor 51, one or more steps of the logging curve interpolation method described above may be performed. Alternatively, in other embodiments, processor 51 may be configured to perform the logging curve interpolation method by any other suitable means (e.g., by means of firmware).

[0152] Optionally, embodiments of the present invention also provide a computer program product, including a computer program that, when executed by a processor, implements the well logging curve interpolation method provided in any embodiment of the present invention.

[0153] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more programmable programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a memory system, at least one input device, and at least one output device, and transmitting data and instructions to the memory system, the at least one input device, and the at least one output device.

[0154] Programmable programs for implementing the methods of the present invention can be written in any combination of one or more programming languages. These programmable programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the programmable programs cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The programmable programs can be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0155] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a programmable program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0156] To provide user interaction, the systems and techniques described herein can be implemented on a logging curve interpolation device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the logging curve interpolation device. Other types of devices can also be used to provide user interaction; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0157] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0158] A computing system can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by programmable programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0159] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0160] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A well logging curve interpolation method, characterized in that, include: Acquire raw seismic data and raw well logging curve data, and preprocess the raw seismic data and raw well logging curve data to determine the seismic data volume to be processed between target formations, the well-side seismic trace data matrix, and the well logging curve data matrix; For each seismic trace data to be processed in the seismic data body to be processed, the seismic trace data to be processed, the well-side seismic trace data matrix and the logging curve data matrix are respectively used as the query vector, key vector and value vector in the self-attention mechanism to solve for the target interpolated logging curve corresponding to the seismic trace data to be processed. Based on the preprocessing data processing method, the data of each target interpolation logging curve is reallocated to determine the logging curve interpolation result.

2. The well logging curve interpolation method according to claim 1, characterized in that, The preprocessing of the raw seismic data and the raw well logging data to determine the seismic data volume to be processed between target formations, the well-side seismic trace data matrix, and the well logging data matrix includes: The original well logging curve data is subjected to time-depth conversion and calibration processing to determine intermediate well logging curve data aligned with the original seismic data; The original seismic data and the intermediate logging curve data are resampled, destination interlayer data are extracted, and data length is unified to determine multiple seismic trace data sequences corresponding to the original seismic data and multiple logging curve data sequences corresponding to the intermediate logging curve data. The seismic trace data sequences are arranged in the order of the seismic traces in the original seismic data to determine the seismic data volume to be processed. The well logging curve data sequences are arranged according to well location to form a well logging curve data matrix; The seismic trace data sequence corresponding to each of the well logging curve data sequences is determined as the well-side seismic trace data sequence, and the well-side seismic trace data sequences are arranged according to the well location to form the well-side seismic trace data matrix.

3. The well logging curve interpolation method according to claim 1, characterized in that, The step of using the seismic trace data to be processed, the well-side seismic trace data matrix, and the logging curve data matrix as the query vector, key vector, and value vector in the self-attention mechanism, respectively, to determine the target interpolated logging curve corresponding to the seismic trace data to be processed includes: Based on the seismic trace data to be processed, the well-side seismic trace data matrix, and the pre-configured attention vector determination formula, the attention vector corresponding to the seismic trace data to be processed is determined; The product of the attention vector and the logging curve data matrix is ​​determined as the target interpolated logging curve corresponding to the seismic trace data to be processed; The formula for determining the attention vector is as follows: Where Q is the seismic trace data to be processed, K is the well-side seismic trace data matrix, and d k The width of the well-side seismic trace data matrix is ​​given by sqrt(), where sqrt() is the square root function and softmax() is the normalization function.

4. The well logging curve interpolation method according to claim 1, characterized in that, After determining the seismic data volume to be processed between the target strata, the well-side seismic trace data matrix, and the well logging curve data matrix, the following is also included: Based on the well-side seismic trace data matrix and the well logging curve data matrix, a well logging curve interpolation model is trained; wherein, the well logging curve interpolation model includes an attention interpolation module and an optimization interpolation module; For each seismic trace in the seismic data body to be processed, the seismic trace is input into the attention interpolation module in the logging curve interpolation model to determine the intermediate interpolation logging curve. The intermediate interpolated logging curve is input into the optimized interpolation module in the logging curve interpolation model to determine the target interpolated logging curve corresponding to the seismic trace data to be processed. The data processing method of the attention interpolation module is consistent with the processing method of solving the seismic trace data to be processed, the well-side seismic trace data matrix, and the well logging curve data matrix as query vector, key vector, and value vector in the self-attention mechanism, respectively.

5. The well logging curve interpolation method according to claim 4, characterized in that, When the logging curve type is a wave impedance curve, the step of training a logging curve interpolation model based on the well-side seismic trace data matrix and the logging curve data matrix includes: The well-side seismic trace data matrix and the well logging curve data matrix are substituted into the attention interpolation module, and multiple seismic trace data to be processed are extracted from the seismic data body to be processed as training seismic trace data. Construct a training sample set using the training seismic trace data described above as samples and the same training seismic trace data as labels; Each training sample in the training sample set is sequentially input into the initial well logging curve interpolation model, and the output of the initial well logging curve interpolation model is forward modeled to determine the intermediate seismic trace data. A loss function is constructed based on the intermediate seismic trace data and the training seismic trace data in the training samples, and the optimized interpolation module is trained using the loss function to obtain the well logging curve interpolation model.

6. The well logging curve interpolation method according to claim 4, characterized in that, When the logging curve type is not a wave impedance curve, the step of training a logging curve interpolation model based on the well-side seismic trace data matrix and the logging curve data matrix includes: Substitute the well-side seismic trace data matrix and the well logging curve data matrix into the attention interpolation module; A training sample set is constructed, with each well-side seismic trace data sequence in the well-side seismic trace data matrix as a sample and the well logging curve data sequence corresponding to each well-side seismic trace data sequence as a label; The initial logging curve interpolation model is trained using the training sample set to obtain the logging curve interpolation model.

7. A well logging curve interpolation device, characterized in that, include: The data acquisition module is used to acquire raw seismic data and raw well logging curve data, and to preprocess the raw seismic data and raw well logging curve data to determine the seismic data volume to be processed between target formations, the well-side seismic trace data matrix, and the well logging curve data matrix. The target curve determination module is used to solve for each seismic trace data to be processed in the seismic data body to be processed by using the seismic trace data to be processed, the well-side seismic trace data matrix and the logging curve data matrix as query vector, key vector and value vector in the self-attention mechanism, respectively, to determine the target interpolated logging curve corresponding to the seismic trace data to be processed. The interpolation result determination module is used to perform data realignment processing on each of the target interpolated logging curves according to the preprocessed data processing method, and determine the logging curve interpolation result.

8. A well logging curve interpolation device, characterized in that, include: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a programmable program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the logging curve interpolation method according to any one of claims 1-6.

9. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the logging curve interpolation method as described in any one of claims 1-6.

10. A computer program product comprising a computer program, which, when executed by a processor, performs the well logging curve interpolation method as described in any one of claims 1-6.

Images