Position encoding determination method and well logging evaluation large model-based geological feature prediction method

By determining the spatial extent type and calculating the longitude and latitude weight coefficients for location encoding in the large oil and gas industry model, the problem that general large models cannot effectively handle the spatial location correlation in the oil and gas industry is solved, thus improving the prediction accuracy of the large oil and gas model.

WO2026237375A1PCT designated stage Publication Date: 2026-11-19RICHFIT INFORMATION TECH +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2025/144747
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-16
Filing Date
2025-12-23
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

The location encoding algorithm of general large models cannot effectively solve the problem of spatial location correlation in the oil and gas industry, resulting in insufficient accuracy of large models in the oil and gas industry in complex geological comprehensive evaluation.

Method used

By determining the type of the spatial range to be encoded, calculating the longitude and latitude weight coefficients based on the maximum and minimum longitude and latitude, isotropic processing is performed to generate location codes, and pre-training is conducted in conjunction with the spatial correlation of geological information.

Benefits of technology

It improves the spatial correlation considerations during the pre-training of large oil and gas models, thereby enhancing the accuracy of oil and gas reservoir spatial distribution prediction and exploration target selection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2025144747_19112026_PF_FP_ABST
    Figure CN2025144747_19112026_PF_FP_ABST
Patent Text Reader

Abstract

The present disclosure relates to the field of oil and gas field exploration and development, and relates to a position encoding determination method and a well logging evaluation large model-based geological feature prediction method. The position encoding determination method comprises: according to a preset spatial classification rule, identifying a geographic region to be encoded, and determining a spatial range type to which the geographic region belongs; if the spatial range type corresponds to a first spatial range, on the basis of the maximum longitude and latitude and the minimum longitude and latitude within the first spatial range, determining a longitude weight coefficient and a latitude weight coefficient within the first spatial range, wherein the first spatial range is a defined spatial range; and on the basis of the longitude weight coefficient and the latitude weight coefficient, calculating position encodings of points within the first spatial range. In the present solution, elements of geostatistics are comprehensively considered, and spatial position information is position-encoded, so that a large model based on the oil and gas industry can better consider spatial correlation of geological information during pre-training, thereby effectively improving the accuracy of the oil and gas large model.
Need to check novelty before this filing date? Find Prior Art

Description

Location coding determination and geological feature prediction method based on well logging evaluation model

[0001] Related applications

[0002] This application claims priority to Chinese Patent Application No. 202510637351.1, filed on May 16, 2025, and incorporates the entire contents of the aforementioned patent application as part of this application. Technical Field

[0003] This disclosure relates to the field of oil and gas field exploration and development, and in particular to a method for determining location codes and predicting geological features based on a large-scale well logging evaluation model. Background Technology

[0004] General-purpose large models, based on deep learning, especially the Transformer architecture, have proven successful in natural language processing (NLP), computer vision (CV), and other machine learning tasks. Their scale and complexity enable them to capture deep patterns in data, resulting in significant performance across various applications. The pre-training process of general-purpose large models must consider not only the semantic relevance of the sequence data itself but also the semantic relevance of the sequence data's positions; this is one of the core issues concerning the model's generalization ability.

[0005] The input tokens for large-scale models in the oil and gas industry are not language-based, but rather multi-dimensional information about a specific spatial location (such as well logging data), primarily describing the geological information at that location. Large-scale models in the oil and gas industry must consider not only the inherent relevance of geological information at a particular spatial location, but also the impact of different spatial locations on this relevance. However, the positional encoding of general-purpose large-scale models primarily considers the influence of word order, which is inconsistent with the spatial location impact of geological information. Therefore, current positional encoding algorithms for general-purpose large-scale language models cannot effectively address the spatial relevance of large-scale oil and gas models. General-purpose large-scale models face challenges in the oil and gas industry due to a lack of relevant professional knowledge in complex geological comprehensive evaluations.

[0006] The ultimate goal of large-scale models is to predict or generate geological information for a specific spatial location. Geological regional variables inherently possess both structure and randomness. Therefore, geological information at different spatial locations exhibits a certain correlation based on their proximity. Two locations that are closer together have a higher correlation, while two locations that are farther apart will inevitably have a lower correlation. This is similar to large language models, but there are also differences. The location in large language models is more local, even when considering long dependencies. In contrast, the spatial correlation of geological information exhibits a more global nature, meaning that absolute positional relationships exist throughout the entire considered space. The positional encoding in general-purpose large language models does not adequately address these issues, thus reducing the accuracy of industry-specific large-scale models. Summary of the Invention

[0007] To address the problems in the prior art, this disclosure provides a method for determining location coding and predicting geological features based on a large-scale well logging evaluation model. The method includes:

[0008] According to one aspect of the present disclosure, the spatial range type to which the spatial range to be encoded belongs is determined; if the spatial range type is a first spatial range, the longitude weight coefficient and latitude weight coefficient within the first spatial range are determined based on the maximum and minimum longitude and latitude within the first spatial range, wherein the first spatial range is a defined spatial range; and the location code of each point within the geographic area to be encoded is calculated based on the longitude weight coefficient and latitude weight coefficient.

[0009] According to one aspect of the present disclosure, determining the longitude weight coefficient and latitude weight coefficient within the first spatial range based on the maximum and minimum longitude and latitude within the first spatial range includes: performing isotropic processing on the longitude direction and latitude direction of the first spatial range based on the maximum and minimum longitude and latitude within the first spatial range to obtain the longitude weight coefficient and latitude weight coefficient.

[0010] According to one aspect of the present disclosure, isotropic processing is applied to the longitude and latitude directions of a first spatial range to obtain longitude weight coefficients and latitude weight coefficients, including: determining the longitude weight coefficients using a first formula based on the maximum and minimum longitude and latitude within the first spatial range. Where α represents the longitude weight coefficient, lon_max represents the maximum longitude, lon_min represents the minimum longitude, lat_max represents the maximum latitude, and lat_min represents the minimum latitude; based on the maximum and minimum longitude and latitude within the first spatial range, the latitude weight coefficient is determined using the second formula: Where β represents the latitude weight coefficient, lon_max represents the maximum longitude, lon_min represents the minimum longitude, lat_max represents the maximum latitude, and lat_min represents the minimum latitude.

[0011] According to one aspect of the present disclosure, calculating the location code of each point within a geographic area to be encoded based on longitude weight coefficients and latitude weight coefficients includes: determining the location code of each point within the geographic area to be encoded using the following formula based on the longitude weight coefficients, latitude weight coefficients, and the standardized longitude and latitude coordinates of each point: Where i represents the index of the position vector of a certain spatial location within the first spatial range, PE pos,2i PE represents the location code of the first element of the location vector that makes up each point within the geographic area to be encoded. pos,2i+1This represents the location code of the second element of the location vector that makes up each point within the geographic region to be encoded. The location codes of the first and second elements constitute the location codes of the points within the geographic region to be encoded. `norm_lon` represents the standardized longitude, `norm_lat` represents the standardized latitude, `Scale` represents the dimensionless location coefficient (default 100), `d_model` represents the dimension of the location vector (default 512), and `PE`... pos =PE(lon,lat) represents the location code of a point in space. In this embodiment, the first element of the location vector of each point in the geographic area to be encoded specifically represents the even number of elements in the location vector, and the odd number of elements represents the odd number of elements in the location vector. The location codes corresponding to the even number of elements and the odd number of elements are combined to obtain the location code of a point in the geographic area to be encoded. In this embodiment, the location vector of a point in the geographic area to be encoded consists of 512 elements, and PE is calculated according to the formula. pos,2i and PE pos,2i+1 This allows us to determine the location code corresponding to the location vector of a point within the geographic area to be encoded. pos,2i and PE pos,2i+1 =PE(lon,lat) both represent the position code of a point in space.

[0012] According to one aspect of the present disclosure, when considering the spatial location elevation of a geographic area to be encoded, the method further includes: determining an elevation weight coefficient for the geographic area to be encoded based on the Earth's surface area and reference elevation within the geographic area to be encoded; homogenizing the longitude and latitude coordinates within the geographic area to be encoded based on the elevation weight coefficients, and updating the longitude weight coefficients and latitude weight coefficients; and determining the location code of each point within the geographic area to be encoded based on the updated longitude weight coefficients, latitude weight coefficients, and elevation weight coefficients.

[0013] According to one aspect of the present disclosure, the method further includes: generating a location encoding vector for all points within the geographic area to be encoded based on the location encoding of each point within the geographic area to be encoded, wherein the location encoding vector is the location encoding vector corresponding to the geographic area to be encoded.

[0014] According to one aspect of the present disclosure, the method further includes: if the spatial range type is a second spatial range, standardizing the longitude and latitude coordinates within the second spatial range, wherein the second spatial range is an undefined spatial range; and calculating the location code of each point within the geographic area to be encoded based on the standardized longitude and latitude coordinates.

[0015] This disclosure provides a geological feature prediction method based on a well logging evaluation large model. The method includes: determining the geological information vector of the area to be encoded based on well logging data; determining the location encoding vector corresponding to the geographical area to be encoded according to any one of claims 1-7; and inputting the geological information vector and the location encoding vector into a pre-trained well logging evaluation large model to obtain the geological feature prediction result of the area to be encoded.

[0016] This disclosure provides a location coding determination device, comprising: a spatial range type determination unit, configured to identify a geographic region to be coded according to a preset spatial classification rule and determine the spatial range type to which the geographic region to be coded belongs; a coefficient determination unit, configured to determine the longitude weight coefficient and latitude weight coefficient within the first spatial range based on the maximum and minimum longitude and latitude within the first spatial range if the spatial range type is a first spatial range, wherein the first spatial range is a defined spatial range; and a location coding determination unit, configured to calculate the location code of each point within the geographic region to be coded based on the longitude weight coefficient and latitude weight coefficient.

[0017] This disclosure also provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a method for determining location encoding and predicting geological features based on a large well logging evaluation model.

[0018] This disclosure also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a method for determining location encoding and predicting geological features based on a large well logging evaluation model.

[0019] This scheme emphasizes the spatial correlation of regional variables, comprehensively considers geostatistical elements, and encodes spatial location information. This allows large-scale oil and gas industry models to better consider the spatial correlation of geological information during pre-training and adhere to linear correlation. Pre-training of industry-specific large-scale models can be carried out without modifying the architecture of general large-scale models, thereby effectively improving the accuracy of oil and gas large-scale models. For example, when pre-training a large-scale model for basin-level reservoir parameter prediction, historical well logging data, seismic attribute data, and reservoir parameter data within the basin have already been collected. Using the location encoding provided in the embodiments of this disclosure, the spatial distribution patterns of well logging data, seismic data, and reservoir parameters can be effectively obtained. Furthermore, this can optimize the prediction of oil and gas reservoir spatial distribution and the selection of exploration targets. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 is a flowchart of a method for determining the location code for well logging evaluation according to an embodiment of the present disclosure;

[0022] Figure 2 is a flowchart of a method for cross-validating and completing design parameters according to an embodiment of this disclosure.

[0023] Figure 3 shows a flowchart of a method for determining longitude weight coefficients and latitude weight coefficients within a first spatial range according to an embodiment of this disclosure;

[0024] Figure 4 illustrates a location coding determination method that considers the spatial location elevation of a geographic area to be coded according to an embodiment of this disclosure.

[0025] Figure 5 shows a flowchart of a prediction model training method based on an oil and gas field according to an embodiment of this disclosure;

[0026] Figure 6 is an architecture diagram of a Transformer model according to an embodiment of this disclosure;

[0027] Figure 7 is a schematic diagram of a location coding determination device for well logging evaluation according to an embodiment of the present disclosure;

[0028] Figure 8 shows a schematic diagram of the structure of a computer device according to an embodiment of the present disclosure.

[0029] Explanation of symbols in the attached drawings: 701, Spatial range type determination unit; 702, Coefficient determination unit; 703, Position encoding determination unit; 802, Computer equipment; 804, Processor; 806, Memory; 808, Drive mechanism; 810, Input / output module; 812, Input device; 814, Output device; 816, Presentation device; 818, Graphical user interface; 820, Network interface; 822, Communication link; 824, Communication bus. Detailed Implementation

[0030] To enable those skilled in the art to better understand the technical solutions in this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0031] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure 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 this disclosure 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 non-exclusive inclusion; for example, a process, method, apparatus, product, or device 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 devices.

[0032] This disclosure provides method operation steps as shown in the embodiments or flowcharts, but based on conventional or non-inventive labor, more or fewer operation steps may be included. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only execution order. In actual system or device products, the methods shown in the embodiments or drawings can be executed sequentially or in parallel.

[0033] It should be noted that the location coding determination method and the geological feature prediction method based on the well logging evaluation model disclosed herein can be used in the field of oil and gas field exploration and development. This disclosure does not limit the application field of the location coding determination method and the geological feature prediction method based on the well logging evaluation model.

[0034] General large language models have extremely long contextual understanding, but due to the limitation of GPU memory resources, these large models do not necessarily need to be trained on very long texts during actual training. How to ensure that the model can support a length far exceeding that during pre-training without causing a decrease in the model's generalization ability is one of the core problems of large models. This problem is also referred to as the extrapolation of large models.

[0035] Pre-training of large-scale well logging evaluation models in the oil and gas industry is inseparable from location encoding. Vectors corresponding to multiple adjacent depth points in the same geographic area to be encoded have high correlation. However, vectors corresponding to depth points of adjacent wells in the same formation, although not in the same context and are far apart, may be more correlated. But this correlation cannot be captured by large-scale language models in conventional domains. Therefore, this application needs to determine the location encoding of the input vectors so that the large model can better consider the spatial correlation of geological information during pre-training.

[0036] Figure 1 shows a schematic diagram of a location coding determination system for well logging evaluation according to an embodiment of the present invention. The system may include a terminal 101 and a server 102. The terminal 101 and server 102 communicate via a network, which may include a Local Area Network (LAN), a Wide Area Network (WAN), the Internet, or a combination thereof, and is connected to a website, user equipment (e.g., a computing device), and a backend system. Staff can send a request to the server 102 to determine the spatial range type via the terminal 101. Upon receiving the request, the server 102 identifies the geographic area to be coded according to preset spatial classification rules, determining the spatial range type to which the geographic area to be coded belongs. If the spatial range type is a first spatial range, the server determines the longitude weight coefficient and latitude weight coefficient within the first spatial range based on the maximum and minimum longitude and latitude within the first spatial range. Based on the longitude and latitude weight coefficients, the server calculates the location code for each point within the first spatial range and sends the location code to the terminal 101.

[0037] In this embodiment of the disclosure, server 102 may be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0038] In an optional embodiment, terminal 101 may be an electronic device, including but not limited to self-service terminal equipment, desktop computer, tablet computer, laptop computer, smart wearable device, etc. Optionally, the operating system running on the electronic device may include but not limited to Android, iOS, Linux, Windows, etc. Of course, terminal 101 is not limited to the above-mentioned physical electronic devices; it may also be software running on the above-mentioned electronic devices.

[0039] Furthermore, it should be noted that Figure 1 shows only one application environment provided by this disclosure. In actual applications, it may also include multiple terminals 101, and this disclosure does not impose any restrictions.

[0040] Figure 2 shows a flowchart of a location coding method for well logging evaluation according to an embodiment of the present disclosure, specifically including steps 201 to 203:

[0041] Step 201: Identify the geographic region to be coded according to the preset spatial classification rules and determine the spatial range type to which the geographic region to be coded belongs.

[0042] In this step, the geographic region to be encoded can be understood as the data used as input to the large language model in subsequent processing; it can be a three-dimensional spatial region. Within this geographic region, there are multiple points whose location encoding needs to be determined. Therefore, the method disclosed herein is used to perform location encoding on each point within the spatial range to be encoded, obtaining the location-encoded data for each point within the spatial range. Further, this step identifies the geographic region to be encoded according to preset spatial classification rules. Specifically, it identifies whether the spatial range type of the geographic region to be encoded belongs to a defined regional spatial range or an undefined regional spatial range.

[0043] In this disclosure, the pre-defined spatial classification rules include, but are not limited to, rules for dividing regions based on administrative divisions and geological structures. In this disclosure, the spatial scope of the oil and gas industry encompasses the physical and geographical boundaries of its entire lifecycle activities, including resource exploration, development, production, transportation, and waste disposal.

[0044] Step 202: If the spatial range type is the first spatial range, determine the longitude weight coefficient and latitude weight coefficient within the first spatial range based on the maximum and minimum longitude and latitude within the first spatial range, where the first spatial range is a predefined spatial range.

[0045] In this step, the first spatial scope is a predefined spatial scope, which specifically refers to a geographical or functional area pre-delineated and assigned specific attributes through clear rules, standards, or protocols. Such spatial scopes typically have clear boundaries, classification attributes, and application purposes, used to support scenarios such as resource management, asset tracking, data interoperability, or compliance control. In this embodiment, physical or logical boundaries can be determined by coordinates, geographical landmarks (such as rivers, roads), or protocols (such as contract blocks), for example: oilfield development contract blocks, pipeline safety buffer zones, refinery operation areas, etc. In this embodiment, types can be divided according to preset rules (such as administrative divisions, geological structures, functional uses, etc.) to facilitate unified management, for example: oil and gas basins, fault zone geological structural units. In this disclosure, the first spatial scope is dynamically expandable, and its scope can be adjusted or new categories added (such as adding exploration blocks or emergency response zones) according to needs.

[0046] In this step, after determining the first spatial range, the maximum longitude, maximum latitude, minimum longitude, and minimum latitude within the geographic area to be coded are determined based on the distribution range of the geographic area to be coded along latitude and longitude. Furthermore, the longitude weight coefficient and latitude weight coefficient are determined. These coefficients are used to process the weighted calculation of geospatial data to reflect the differences in the impact of different latitudes and longitudes within the first spatial range on the location coding analysis results.

[0047] In this disclosure, the latitude weight coefficient is a weight designed to correct for area deformation caused by the Earth's curvature. Since the Earth is a sphere, the actual distance corresponding to the same longitude interval differs at different latitudes (longest at the equator, decreasing towards the poles). Therefore, the longitude weight coefficient is typically used for data processing in non-uniform longitude grids. In an ideal uniform grid, the longitude weight coefficient may be 1, but if the grid resolution varies with longitude (e.g., denser observations in certain areas), the weight needs to be adjusted.

[0048] Step 203: Calculate the location code of each point within the geographic area to be coded based on the longitude weight coefficient and the latitude weight coefficient.

[0049] In this step, after calculating the longitude weight coefficient and latitude weight coefficient, and before calculating the location code of each point in the geographic area to be coded, it is necessary to standardize the longitude and latitude of each point. Finally, the standardized longitude and latitude, as well as the longitude weight coefficient and latitude weight coefficient, are used to calculate the location code of each point in the geographic area to be coded.

[0050] This application considers the spatial correlation of geological regional variables, encodes spatial information by location, and takes into account the spatial correlation of geological information, which greatly improves the rationality of pre-training of large industry models and effectively improves the prediction accuracy of large oil and gas models.

[0051] Figure 3 shows a flowchart of a method for determining longitude weight coefficients and latitude weight coefficients within a first spatial range according to an embodiment of this disclosure, specifically including steps 301 to 302:

[0052] Step 301: Determine the longitude weight coefficient using the first formula based on the maximum and minimum latitude and longitude within the first spatial range.

[0053] This step uses the following first formula to determine the longitude weighting coefficients based on the maximum latitude, minimum latitude, maximum longitude, and minimum longitude within the first spatial range:

[0054] Where α represents the longitude weight coefficient, lon_max represents the maximum longitude, lon_min represents the minimum longitude, lat_max represents the maximum latitude, and lat_min represents the minimum latitude.

[0055] Step 302: Determine the latitude weighting coefficient using the second formula based on the maximum and minimum latitude and longitude within the first spatial range.

[0056] Where β represents the latitude weighting coefficient, lon_max represents the maximum longitude, lon_min represents the minimum longitude, lat_max represents the maximum latitude, and lat_min represents the minimum latitude. In this disclosure, α+β=1.

[0057] In some embodiments of this disclosure, before calculating the location code of each point within the geographic area to be coded based on the longitude weight coefficient and the latitude weight coefficient, the method further includes: standardizing the latitude and longitude coordinates of each point within the first spatial range based on the maximum and minimum latitude and longitude within the first spatial range, to obtain the standardized latitude and longitude coordinates of each point. Then, the location code is calculated based on the standardized latitude and longitude coordinates and the longitude weight coefficient and latitude weight coefficient.

[0058] In this step, the latitude and longitude coordinates of each point within the first spatial range are standardized using the following formula. Specifically, based on the maximum and minimum longitudes within the first spatial range, the longitude coordinates of each point within the first spatial range are standardized using the following formula:

[0059] Where norm_lon represents the standardized longitude;

[0060] Based on the maximum and minimum latitude within the first spatial range, the latitude coordinates of each point within the first spatial range are standardized using the following formula:

[0061] Here, norm_lat represents the standardized latitude.

[0062] Based on the longitude weight coefficient, latitude weight coefficient, and the standardized longitude and latitude coordinates of each point, the location code of each point within the geographic area to be coded is determined using the following formula:

[0063] Where i represents the index of the position vector of a certain spatial location within the first spatial range, norm_lon represents the standardized longitude, norm_lat represents the standardized latitude, Scale represents the dimensionless position coefficient, which defaults to 100, d_model represents the dimension of the position vector, which defaults to 512, and PE pos =PE(lon,lat) represents the position code of a point in space.

[0064] In this disclosure, a spatial location is encoded as a multidimensional location vector as follows: location vector [v1, v2, v3, ...], where i is the index position of the location vector.

[0065] In the formula for calculating the location code, the longitude weight coefficient and the latitude weight coefficient are used to perform isotropic processing on the longitude and latitude directions of the first spatial range, respectively. This achieves isotropic processing on the longitude and latitude directions of the first spatial range, simplifies the computational complexity, and improves the computational efficiency of the large oil and gas model.

[0066] Figure 4 illustrates a location coding determination method considering the spatial location elevation of a geographic area to be coded according to an embodiment of this disclosure, specifically including steps 401 to 403:

[0067] Step 401: Determine the elevation weight coefficient of the geographic region to be coded based on the Earth's surface area and reference elevation within the geographic region to be coded.

[0068] The geographic region to be coded within the latitude and longitude range can be triangulated into multiple triangles of latitude and longitude coordinates. The side length and area of ​​each triangle can be calculated, and the summation can be used to obtain the area within the region.

[0069] In this case, assume that the three sides of one of the partitioned triangles are s1, s2, and s3.

[0070] Where ss represents the average side length of the triangle;

[0071] A Δi Represents the area of ​​the triangle into which the parts are divided;

[0072] A=∑A Δi,i=1,2,...,N;A represents the Earth's surface area within the geographic region to be encoded, consisting of lon_min, lon_max, lat_min, and lat_max.

[0073] Furthermore, the radius of the equivalent circular area of ​​the geographic region to be encoded is calculated using the following formula:

[0074] Where A represents the Earth's surface area within the geographic region to be coded, and r represents the radius of the equivalent circular area of ​​the geographic region to be coded.

[0075] In this disclosure, the region to be encoded is a three-dimensional spatial range, and the geographic coordinates within the three-dimensional spatial range are three-dimensional information (X,Y,Z), where X and Y are planar positions and Z is a longitudinal position. All three positions exhibit severe anisotropy, and the influence of both planar and longitudinal positions on the correlation needs to be considered. However, the influencing factors (weights) they use to represent the correlation are different. Therefore, it is necessary to determine an influencing factor to balance the final impact of planar and longitudinal positions on the correlation of geological information.

[0076] This step determines the elevation weighting coefficient using the following formula:

[0077] Where ele_ref represents the maximum depth of the geographic area to be encoded. Considering that the current well depth range is basically within 10,000 meters, the default ele_ref = 10,000, and r is the radius of the equivalent circular area of ​​the geographic area to be encoded.

[0078] Here, norm_ele represents the standardized spatial location elevation.

[0079] Step 402: Based on the elevation weight coefficient, perform isotropic processing on the longitude and latitude coordinates within the geographic area to be coded to obtain the updated longitude weight coefficient and latitude weight coefficient.

[0080] In the field of geophysics, longitude converges at the poles, causing the actual distance corresponding to the same longitude and latitude interval to vary with latitude; the geomagnetic field and gravity field are affected by mantle convection and plate movement, resulting in directional differences; uneven distribution of satellite orbits or ground observation stations leads to differences in the direction of data coverage, all of which contribute to latitude and longitude anisotropy.

[0081] In this step, the longitude coordinates within the geographic area to be coded are made isotropic according to the following formula:

[0082] Where γ represents the elevation weight coefficient, α represents the longitude weight coefficient before the update, β represents the latitude weight coefficient before the update, and α′ represents the longitude weight coefficient after the update.

[0083] Correspondingly, the latitudinal coordinates within the first spatial range are homogenized according to the following formula;

[0084] Where γ represents the elevation weighting coefficient, α represents the longitude weighting coefficient before the update, β represents the latitude weighting coefficient before the update, and β′ represents the latitude weighting coefficient after the update. And α′+β′+γ=1.

[0085] In the above formula, the longitude weight coefficient and latitude weight coefficient before the update are calculated according to the formula in Figure 3. In this disclosure, isotropic processing of longitude and latitude can eliminate the non-uniformity caused by directional differences in geospatial data, making it have consistent statistical characteristics or physical properties in all directions, eliminating directional deviations and geometric distortions, and achieving isotropy.

[0086] Step 403: Determine the location code of each point within the geographic area to be coded based on the updated longitude weight coefficient, latitude weight coefficient, and elevation weight coefficient.

[0087] Where γ represents the elevation weight coefficient, α′ represents the updated longitude weight coefficient, and β′ represents the updated latitude weight coefficient. norm_lon represents the standardized longitude, norm_lat represents the standardized latitude, Scale represents the dimensionless position coefficient (default 100), d_model represents the dimension of the position vector (default 512), and PE... pos,2i This represents the location code of a point within the first spatial range.

[0088] Based on the above formula, the location codes of each point within the geographic area to be coded are obtained.

[0089] According to one embodiment of this disclosure, the method further includes: generating a location encoding vector for all points within the geographic area to be encoded based on the location encoding of each point within the geographic area to be encoded, wherein the location encoding vector is the location encoding vector corresponding to the geographic area to be encoded.

[0090] According to one embodiment of this disclosure, if the spatial range type is a second spatial range, the longitude and latitude coordinates within the second spatial range are standardized. Specifically, the following formula is used to standardize the longitude and latitude coordinates within the second spatial range:

[0091] Where lon represents the longitude coordinates within the second spatial range; lat represents the latitude coordinates within the second spatial range; norm_lon represents the standardized longitude coordinates within the second spatial range, and norm_lat represents the standardized latitude coordinates within the second spatial range. In this step, the standardized longitude and latitude coordinates are substituted into the location coding calculation formula in Figure 3 to calculate the location code of each point within the second spatial range. The longitude weight coefficient and latitude weight coefficient of the second spatial range can be set to half.

[0092] Calculate the location code of each point within the undefined spatial range based on the standardized longitude and latitude.

[0093] This disclosure emphasizes the spatial correlation of regional variables, comprehensively considers geostatistical elements, and encodes spatial location information. This allows large-scale oil and gas industry models to better consider the spatial correlation of geological information during pre-training and to follow linear correlation. Pre-training of industry-specific large-scale models can be carried out without modifying the architecture of general large-scale models, thereby effectively improving the accuracy of oil and gas large-scale models.

[0094] Figure 5 shows a flowchart of a prediction model training method based on an oil and gas field according to an embodiment of this disclosure, specifically including steps 501 to 503:

[0095] Step 501: Determine the geological information vector of the area to be encoded based on the well logging data of the area to be encoded.

[0096] In the direction of predicting geological features of oil and gas fields, the data input into the prediction model is vector data corresponding to at least one location in the area to be encoded, and the prediction model is expected to predict geological information. Therefore, these vector data can be called geological information vectors. Geological information vectors are actually indirect responses to geological information. Each dimension of the geological information vector can reflect a small part of the geological information, including but not limited to: lithology, physical properties, oil-bearing properties, etc.

[0097] In this disclosure, the logging curve of a well in a study area within a first spatial range can be used as data to be input into the prediction model. Before becoming input data, the logging curve of this well is converted into a geological information vector, thereby obtaining the address information vectors corresponding to multiple locations of this well. For example, this well is called well A. There are logging curves for a certain depth range of well A (e.g., depth 2500 meters to depth 2520 meters). These logging curves correspond to 100 depth points (spatial points) within this depth range, and each depth point corresponds to at least one value in the logging curve. Through at least one value in the logging curve, the porosity, permeability, oil saturation, etc., of the 100 depth points within the depth range of 2500 meters to 2520 meters of well A can be calculated. These are all geological information. Converting the geological information of each depth point into vector form yields the geological information vectors of each depth point. These values ​​can form vectors, such as: [GR,CAL,SP,RLLM,RLLD,AC,DEN,CNL,...].

[0098] Step 502: Determine the location encoding vector corresponding to the geographic area to be encoded.

[0099] Following the example in step 501, this step calculates location coding vectors for 100 depth points within the range of 2500 meters to 2520 meters in well A, resulting in 100 location coding vectors corresponding to each of the 100 depth points. This yields 100 location coding vectors and their corresponding 100 geological information vectors.

[0100] In this step, following the method for determining the location codes of each point in the geographic area to be coded as shown in Figures 2 to 4 above, a location coding vector corresponding to the geographic area to be coded is generated based on the spatial relationship of each point. Alternatively, in this disclosure, a location coding matrix corresponding to the geographic area to be coded can also be generated based on the spatial relationship of each point.

[0101] Step 503: Input the geological information vector and the location encoding vector into the pre-trained well logging evaluation model to obtain the geological feature prediction results of the area to be encoded.

[0102] In this step, the geological information vectors and location coding vectors corresponding to each point in the geographic area to be coded are used as fusion vectors and input into the well logging evaluation large model. This disclosure does not limit the fusion calculation method for the geological information vectors and location coding vectors. The well logging evaluation large model in this disclosure is constructed based on the Transformer model.

[0103] Figure 6 shows an architecture diagram of a Transformer model according to an embodiment of this disclosure.

[0104] In Transformer models, special positional encoding methods are used, primarily combinations of sine and cosine functions and RoPE (Rotated Position Vector) encoding. In Transformer models, positional encoding enables the model to distinguish inputs at different locations, assigning a fixed vector to each position in the input sequence. The positional encoding vector is added to the word vectors in the input sequence to fuse positional information. This application integrates geostatistical factors, considering the correlation of spatial positional information at a specific location and different locations within the geological space of the study area, and encodes this spatial information. This allows large models to better consider the spatial correlation of geological information during pre-training, improving the rationality of pre-training large-scale oil and gas models and ultimately enhancing their prediction accuracy.

[0105] Figure 7 is a schematic diagram of a position encoding determination device according to an embodiment of this disclosure. The figure illustrates the basic structure of the position encoding determination device. The functional units and modules can be implemented in software, or the position encoding can be determined using a general-purpose chip or a specific chip. Specifically, the device includes:

[0106] The spatial range type determination unit 701 is used to identify the geographic area to be coded according to the preset spatial classification rules and determine the spatial range type to which the geographic area to be coded belongs.

[0107] The coefficient determination unit 702 is used to determine the longitude weight coefficient and latitude weight coefficient within the first spatial range based on the maximum and minimum longitude and latitude within the first spatial range if the spatial range type is a first spatial range, wherein the first spatial range is a defined spatial range;

[0108] The location coding determination unit 703 is used to calculate the location code of each point in the geographic area to be coded based on the longitude weight coefficient and the latitude weight coefficient.

[0109] Figure 8 shows a schematic diagram of a computer device provided in an embodiment of this disclosure. The location coding determination and geological feature prediction method based on a large well logging evaluation model of this application can be applied to a computer device. The computer device 802 may include one or more processors 804, such as one or more central processing units (CPUs), each of which can implement one or more hardware threads. The computer device 802 may also include any memory 806 for storing any kind of information such as code, settings, data, etc. Non-limitingly, for example, the memory 806 may include any type of RAM, any type of ROM, flash memory device, hard disk, optical disk, etc. More generally, any memory can use any technology to store information. Further, any memory can provide volatile or non-volatile retention of information. Further, any memory can represent a fixed or removable component of the computer device 802. In one case, when the processor 804 executes associated instructions stored in any memory or combination of memories, the computer device 802 can perform any operation of the associated instructions. The computer device 802 also includes one or more drive mechanisms 808 for interacting with any memory, such as a hard disk drive mechanism, an optical disk drive mechanism, etc.

[0110] Computer device 802 may also include an input / output module 810 (I / O) for receiving various inputs (via input device 812) and providing various outputs (via output device 814). A specific output mechanism may include a presentation device 816 and an associated graphical user interface (GUI) 818. In other embodiments, the input / output module 810 (I / O), input device 812, and output device 814 may be omitted, and the device may function solely as a computer device within a network. Computer device 802 may also include one or more network interfaces 820 for exchanging data with other devices via one or more communication links 822. One or more communication buses 824 couple the components described above together.

[0111] Communication link 822 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. Communication link 822 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0112] Corresponding to the methods in Figures 2 to 5, this disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the above-described methods.

[0113] This disclosure also provides a computer-readable instruction, wherein when a processor executes the instruction, the program therein causes the processor to perform the method shown in Figures 2 to 5.

[0114] It should be understood that in the various embodiments of this disclosure, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this disclosure.

[0115] It should also be understood that, in the embodiments of this disclosure, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, the character " / " in this disclosure generally indicates that the preceding and following related objects have an "or" relationship.

[0116] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this disclosure can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0117] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0118] In the embodiments provided in this disclosure, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, apparatuses, or units, or they may be electrical, mechanical, or other forms of connection.

[0119] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this disclosure, depending on actual needs.

[0120] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0121] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0122] This disclosure uses specific embodiments to illustrate the principles and implementation methods of this disclosure. The description of the above embodiments is only for the purpose of helping to understand the methods and core ideas of this disclosure. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this disclosure. Therefore, the content of this disclosure should not be construed as a limitation of this disclosure.

Claims

1. A method of position encoding determination, characterized by, The method includes: Based on the preset spatial classification rules, the geographic region to be coded is identified, and the spatial range type to which the geographic region to be coded belongs is determined. If the spatial range type is a first spatial range, the longitude weight coefficient and latitude weight coefficient within the first spatial range are determined based on the maximum and minimum longitude and latitude within the first spatial range, wherein the first spatial range is a defined spatial range; Based on the longitude weight coefficient and latitude weight coefficient, calculate the location code of each point within the geographic area to be coded.

2. The method of claim 1, wherein, Based on the maximum and minimum latitude and longitude within the first spatial range, the longitude weight coefficient and latitude weight coefficient within the first spatial range are determined as follows: Based on the maximum and minimum latitude and longitude within the first spatial range, the longitude and latitude directions within the first spatial range are subjected to isotropic processing to obtain longitude weight coefficients and latitude weight coefficients.

3. The method of claim 2, wherein, The longitude and latitude directions of the first spatial range are respectively isotropically processed to obtain the longitude weight coefficients and latitude weight coefficients, including: According to the maximum latitude and longitude and the minimum latitude and longitude within the first spatial range, a longitude weight coefficient is determined by using a first formula: Where α represents the longitude weight coefficient, lon_max represents the maximum longitude, lon_min represents the minimum longitude, lat_max represents the maximum latitude, and lat_min represents the minimum latitude; According to the maximum latitude and the minimum latitude within the first spatial range, a latitude weight coefficient is determined by using a second formula: Where β represents the latitude weight coefficient, lon_max represents the maximum longitude, lon_min represents the minimum longitude, lat_max represents the maximum latitude, and lat_min represents the minimum latitude.

4. The method of claim 2, wherein, Calculating the location code of each point within the geographic area to be coded, based on the longitude and latitude weight coefficients, includes: According to the longitude weight coefficient, the latitude weight coefficient and the normalized longitude and latitude coordinates of each point, the position code of each point in the geographic region to be encoded is determined by using the following formula: Where i represents the index of the position vector of a certain spatial location within the first spatial range, PE pos,2i PE represents the location code of the first element of the location vector that makes up each point within the geographic area to be encoded. pos,2i+1 This represents the position code of the second element of the position vector that makes up each point within the geographic region to be encoded. The position codes of the first and second elements constitute the position codes of the points within the geographic region to be encoded. `norm_lon` represents the standardized longitude, `norm_lat` represents the standardized latitude, `Scale` represents the dimensionless position coefficient, `d_model` represents the dimension of the position vector, and `PE` represents the position vector. pos,2i and PE pos,2i+1 =PE(lon,lat) both represent the position code of a point in space.

5. The method of claim 1, wherein, When considering the spatial location and elevation of the geographic area to be encoded, the method further includes: The elevation weighting coefficient of the geographic region to be encoded is determined based on the Earth's surface area and reference elevation within the geographic region to be encoded. Using the elevation weight coefficients, the longitude and latitude coordinates within the geographic area to be coded are processed isotropically to obtain updated longitude weight coefficients and latitude weight coefficients. Based on the updated longitude weight coefficient, latitude weight coefficient, and elevation weight coefficient, the location code of each point within the geographic area to be coded is determined.

6. The method of claim 5, wherein, The updated longitude weighting coefficients and latitude weighting coefficients are determined using the following formulas: Where γ represents the elevation weight coefficient, α represents the longitude weight coefficient before the update, β represents the latitude weight coefficient before the update, and α′ represents the longitude weight coefficient after the update. Where γ represents the elevation weight coefficient, α represents the longitude weight coefficient before the update, β represents the latitude weight coefficient before the update, and β′ represents the latitude weight coefficient after the update.

7. The method of claim 1, wherein, The method further includes: Based on the location codes of each point within the geographic region to be encoded, a location code vector is generated corresponding to all points within the geographic region to be encoded, and the location code vector is the location code vector corresponding to the geographic region to be encoded.

8. The method of claim 1, wherein, The method further includes: If the spatial range type is a second spatial range, the longitude and latitude coordinates within the second spatial range are standardized, wherein the second spatial range is an undefined spatial range; Based on the standardized longitude and latitude coordinates, the location code of each point within the geographic area to be coded is calculated.

9. A method for predicting geological features of a large model based on well logging, characterized in that, The method includes: Based on the well logging data of the area to be coded, determine the geological information vector of the area to be coded; The method according to any one of claims 1-8 determines the location encoding vector corresponding to the geographic region to be encoded; The geological information vector and the location encoding vector are input together into the pre-trained well logging evaluation model to obtain the geological feature prediction results of the area to be encoded.

10. A position encoding determination apparatus characterized by comprising: The device includes: The spatial range type determination unit is used to identify the geographic region to be coded according to the preset spatial classification rules and determine the spatial range type to which the geographic region to be coded belongs. The coefficient determination unit is used to determine the longitude weight coefficient and latitude weight coefficient within the first spatial range based on the maximum and minimum longitude and latitude within the first spatial range if the spatial range type is a first spatial range, wherein the first spatial range is a defined spatial range; The location coding determination unit is used to calculate the location code of each point in the geographic area to be coded based on the longitude weight coefficient and the latitude weight coefficient.

11. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 9.