A device and method for real-time processing of electrical imaging data while drilling

Abstract

The application provides a device and method for real-time processing of electrical imaging data while drilling, which comprises: a downhole imaging module for collecting resistivity imaging data of a required formation by using a downhole resistivity imaging instrument; an intelligent identification module for analyzing and identifying the collected resistivity imaging data by using a chip structure embedded with a set intelligent feature identification model, determining corresponding formation characteristics of well logging, and associating and fusing the corresponding formation characteristics with matching formation parameters into a feature identification result; and a real-time communication module for transmitting the fused feature identification result to a ground control system. The application establishes a formation feature identification model of resistivity imaging data while drilling based on an intelligent algorithm to perform real-time feature identification, overcomes the time limitation of ground data acquisition in the prior art, improves a data transmission mechanism, helps a ground engineer to obtain effective formation information in time, and thus makes a quick on-site decision, effectively reduces drilling risks, and optimizes the quality of drilling operations.

Description

Technical Field

[0001] This invention relates to the field of oil and gas exploration and development, specifically to a downhole real-time processing device and method for electrical imaging data while drilling. This method allows for real-time processing of electrical imaging data directly downhole during resistivity measurement while drilling, extracting the required feature information and transmitting it back to the surface in real time, thereby improving the accuracy and controllability of decision-making during the drilling process. Background Technology

[0002] In recent years, to adapt to the increasingly harsh logging environments of high-angle wells and horizontal wells, and to guide drilling operations with more comprehensive and reliable data support, logging-while-drilling (LOD) technology has become an important means of timely and accurate acquisition of drilling and geological data. Among them, LOD resistivity imaging logging plays a key role in real-time well site data acquisition, interpretation, on-site decision-making, and guiding geological steering drilling.

[0003] As the development of drilling instruments continues to advance, the collection and storage of downhole data has been gradually resolved. However, due to limitations in the data transmission speed during drilling, the current data processing strategy employed by drilling instruments involves batch-storing downhole imaging data locally. This data is then exported after drilling is completed, and surface engineers retrieve and analyze it to identify formation features such as fractures, caverns, and interlayer interfaces, thereby conducting a detailed formation evaluation. Clearly, this processing method suffers from poor data timeliness, insufficient flexibility, and high storage requirements for downhole instruments.

[0004] The information disclosed in the background section of this invention is intended only to enhance the understanding of the general background of this invention, and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides a downhole real-time processing device and method for electrical resistivity imaging data during drilling. The main objective of this invention is to overcome the current limitations of data transmission speed in electrical resistivity imaging instruments by performing real-time intelligent identification and feature extraction of data from the downhole environment, transmitting only a small amount of identified requirement information. This helps surface engineers obtain more formation information in real time, make rapid on-site decisions, reduce drilling risks, and improve logging performance and drilling efficiency. In one embodiment, the device includes:

[0006] The downhole imaging module is configured to acquire resistivity imaging data of the required formation in the downhole environment using a drilling resistivity imaging instrument.

[0007] The intelligent recognition module is connected to the downhole imaging module. It analyzes and identifies the collected resistivity imaging data through a chip structure with an embedded intelligent feature recognition model, determines the corresponding well logging formation features, and associates and fuses them with the matching formation parameters to form the feature recognition result.

[0008] The real-time communication module is configured to transmit the fused feature recognition results to the ground control system based on a set transmission protocol when such results are available.

[0009] The intelligent feature recognition model is pre-built by the model building module based on a set image data processing strategy and machine learning algorithm.

[0010] Preferably, in one embodiment, the resistivity imaging data acquired by the downhole imaging module includes a two-dimensional array of resistivity data of the formation around the well at the required depth, corresponding to a 360° radius around the well.

[0011] Furthermore, in one embodiment, the intelligent identification module is a chip structure that has passed high temperature and high pressure testing. Before the high temperature and high pressure test, the algorithm program for setting the intelligent feature identification module is imported into the chip structure by labeling the well area type using FPGA connectors and ARM connectors.

[0012] In an optional embodiment, the model building module constructs the intelligent feature recognition model through the following operations:

[0013] Step A1: Obtain sample data matching each well area type according to the set logic;

[0014] Step A2: Divide the training sample library and the test sample library based on the sample data;

[0015] Step A3: Based on the training samples, use the least squares regression algorithm based on neural networks to train and iteratively update the training, optimize and determine the key parameters corresponding to the model, and use the test samples to verify the accuracy of the model using the cross-validation method until the set training conditions are met.

[0016] Specifically, in one embodiment, the model building module performs the following operations to obtain sample data matching each well type:

[0017] a1. Considering the diversity of well area types, statistically analyze drilling resistivity imaging data of a set scale;

[0018] a2. Preprocess the drilling resistivity imaging data for each well area type;

[0019] a3. Identify and acquire formation feature data characterized by preprocessed drilling resistivity imaging data;

[0020] a4. Based on the acquired formation feature data, the associated drilling resistivity imaging data are labeled, and all associated labeled combination data are statistically analyzed as a sample dataset, which is then classified and stored by well area type.

[0021] Furthermore, in one embodiment, when the model building module preprocesses the drilling resistivity imaging data for each well type, it includes the following operations:

[0022] The collected electrical imaging data during drilling was segmented using the OTSU automatic thresholding method to obtain binary images. Then, the connected components were calculated using a neighborhood calculation method with set parameters to segment the regions.

[0023] In an optional embodiment, the downhole communication module is further provided with a transmission decision unit, which is configured to perform availability assessment on the existing feature recognition results. If it is determined that there are features required by the current depth of the formation in the feature recognition results, an upload command is generated.

[0024] Furthermore, in one embodiment, the device further includes a downhole data quality control module configured to perform real-time defect removal, equalization, gain adjustment, and threshold segmentation on the resistivity imaging data acquired downhole, in order to control the quality of the feature recognition input data.

[0025] In another embodiment, considering other aspects of the application, the apparatus further includes a construction assistance module configured to analyze the identified periodic geological features, make construction recommendation data, and transmit it to the surface control system.

[0026] Based on other aspects of the apparatus described in any one or more of the above embodiments, the present invention also provides a method for real-time downhole processing of electrical imaging data while drilling, the method comprising:

[0027] The downhole imaging procedure involves using a downhole, drilling-while-drilling resistivity imaging instrument to acquire resistivity imaging data of the required formation.

[0028] The downhole identification process involves calling the set intelligent feature recognition model embedded in the downhole chip structure based on the well area type of the current data, analyzing and identifying the resistivity imaging data, determining the corresponding well logging formation features, and associating and fusing them with the matched formation parameters to form the feature recognition result.

[0029] In the real-time communication process, when there are fused feature recognition results, they are transmitted to the ground control system based on the set transmission protocol.

[0030] The intelligent feature recognition model is pre-built by the model building module based on a set image data processing strategy and machine learning algorithm.

[0031] Compared with the closest prior art, the present invention also has the following beneficial effects:

[0032] This invention provides a downhole real-time processing device and method for logging-while-drilling (LWD) resistivity imaging data. The invention establishes a formation feature identification model based on an intelligent algorithm for LWD resistivity imaging data; embeds the established intelligent model algorithm into a chip module; acquires formation resistivity imaging data using a LWD resistivity imaging instrument; performs quality processing on the LWD resistivity imaging data; uses the chip module to perform real-time feature identification on the acquired LWD resistivity imaging data; and sends corresponding depth point and feature identification result instructions to the surface. This invention establishes a formation feature identification model based on an intelligent algorithm for LWD resistivity imaging data, performing real-time feature identification. This overcomes the time limitations of existing technologies for surface data acquisition, improves the data transmission mechanism, and helps surface engineers obtain timely and effective formation information. This provides data support for on-site decision-making and adjustment of drilling parameters, effectively reducing drilling risks, optimizing drilling operation quality, and improving the overall operability and practicality of logging-while-drilling technology.

[0033] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description, claims, and drawings. Attached Figure Description

[0034] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0035] Figure 1 This is a schematic diagram of the downhole real-time processing device for electrical imaging data while drilling provided in an embodiment of the present invention;

[0036] Figure 2 This is a comparison of image data before and after quality control processing by the downhole real-time processing device for downhole electrical imaging data provided in this embodiment of the invention;

[0037] Figure 3 This is a diagram illustrating the identification effect of the entry and exit interface using the downhole real-time processing device for drilling electrical imaging data provided in another embodiment of the present invention.

[0038] Figure 4 This is a flowchart illustrating a method for real-time downhole processing of electrical imaging data provided in another embodiment of the present invention. Detailed Implementation

[0039] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples. Those skilled in the art will then fully understand how the present invention uses technical means to solve technical problems and achieve technical effects, and will be able to implement the present invention specifically based on the above-described implementation process. It should be noted that, as long as there is no conflict, the various embodiments and features of the present invention can be combined with each other, and the resulting technical solutions are all within the protection scope of the present invention.

[0040] Although the flowchart describes the operations as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. The order of the operations can be rearranged. A process can terminate when its operation is complete, but it may also have additional steps not included in the diagram. A process can correspond to a method, function, procedure, subroutine, subroutine, etc.

[0041] Computer equipment includes user equipment and network equipment. User equipment or clients include, but are not limited to, computers, smartphones, PDAs, etc.; network equipment includes, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers. Computer equipment can operate independently to implement this invention, or it can connect to a network and implement this invention through interaction with other computer equipment in the network. The network in which the computer equipment is located includes, but is not limited to, the Internet, wide area network, metropolitan area network, local area network, VPN network, etc.

[0042] The term “and / or” as used herein includes any and all combinations of one or more of the associated items listed. When a unit is referred to as “connected” or “coupled” to another unit, it may be directly connected to or coupled to said other unit, or there may be an intermediate unit present.

[0043] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments. Unless the context clearly indicates otherwise, the singular forms “a” and “an” as used herein are also intended to include the plural. It should also be understood that the terms “comprising” and / or “including” as used herein specify the presence of the stated features, integers, steps, operations, units, and / or components, without excluding the presence or addition of one or more other features, integers, steps, operations, units, components, and / or combinations thereof.

[0044] To adapt to increasingly harsh logging environments such as high-angle and horizontal wells, and to guide drilling operations with more comprehensive and reliable data support, logging-while-drilling (LOD) technology has become an important means of timely and accurate acquisition of drilling and geological data. Among these technologies, resistivity imaging logging while drilling plays a crucial role in real-time well site data acquisition, interpretation, on-site decision-making, and guiding geological-guided drilling.

[0045] As the development of drilling instruments continues to advance, the collection and storage of downhole data has been gradually resolved. However, due to limitations in the data transmission speed during drilling, the current data processing strategy employed by drilling instruments involves batch-storing downhole imaging data locally. This data is then exported after drilling is completed, and surface engineers retrieve and analyze it to identify formation features such as fractures, caverns, and interlayer interfaces, thereby conducting a detailed formation evaluation. Clearly, this processing method suffers from poor data timeliness, insufficient flexibility, and high storage requirements for downhole instruments.

[0046] To address the aforementioned issues, this invention provides a downhole real-time processing device and method for electrical imaging data during drilling. By processing and identifying imaging data in real time downhole, a large amount of imaging data is converted into a small amount of required information, which can help surface engineers quickly understand the downhole situation, make efficient decisions, effectively reduce drilling risks, and improve logging results and drilling efficiency.

[0047] The following describes the detailed flow of the method according to an embodiment of the present invention with reference to the accompanying drawings, the steps of which can be executed in a computer system containing, for example, a set of computer-executable instructions. Although the logical order of the steps is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0048] Example 1

[0049] Figure 1 This diagram illustrates the structure of the downhole real-time processing device for electrical imaging data provided in Embodiment 1 of the present invention. (Refer to...) Figure 1 It can be seen that the device includes:

[0050] The downhole imaging module is configured to acquire resistivity imaging data of the required formation in the downhole environment using a drilling resistivity imaging instrument.

[0051] The intelligent recognition module is connected to the downhole imaging module. It analyzes and identifies the collected resistivity imaging data through a chip structure with an embedded intelligent feature recognition model, determines the corresponding well logging formation features, and associates and fuses them with the matching formation parameters to form the feature recognition result.

[0052] The real-time communication module is configured to transmit the fused feature recognition results to the ground control system based on a set transmission protocol when such results are available.

[0053] The intelligent feature recognition model is pre-built by the model building module based on a set image data processing strategy and machine learning algorithm.

[0054] By using the transposition method described in the above embodiments to process downhole logging imaging data, key formation information reflected in the downhole electrical imaging data can be extracted in real time. This transforms a large amount of imaging data into a small number of prompts. Compared to the previous method of obtaining a large amount of downhole electrical imaging data stored in the instrument for formation analysis only after drilling is completed, this invention can significantly improve timeliness, helping surface engineers obtain more formation information in real time, thereby enabling rapid on-site decision-making. If the accuracy of the intelligent recognition model is high enough, it can effectively reduce drilling risks and improve logging results and drilling efficiency.

[0055] Specifically, in order to ensure that the collected drilling imaging data can comprehensively characterize the features of the downhole formation, in one embodiment, the resistivity imaging data collected by the downhole imaging module includes a two-dimensional array of resistivity data of the formation around the well at the required depth corresponding to 360° of the well.

[0056] Considering that the acquired resistivity imaging data may contain erroneous data points or interference factors that are detrimental to subsequent feature recognition, in a preferred embodiment, the device of the present invention further includes a downhole data quality control module, which is configured to perform real-time bad point removal, equalization processing, gain adjustment, and threshold segmentation processing on the downhole acquired resistivity imaging data, so as to control the quality of the feature recognition input data and fundamentally avoid interference from inferior source data and other factors in the image data.

[0057] Furthermore, referring to the acquisition instruments and computing boards related to the resistivity imaging data used, a high-temperature resistant circuit board is designed. Based on the relevant circuits, a combined intelligent module is constructed. Based on this, the present invention uses relevant FPGA and ARM emulators to write the established feature recognition model algorithm program into the chip structure of the circuit board. Therefore, in a preferred embodiment, the intelligent recognition module is a chip structure that has passed high-temperature and high-pressure testing. Before the high-temperature and high-pressure test, the algorithm program for setting the intelligent feature recognition model is imported into the chip structure, categorized by well area type, using FPGA and ARM connectors.

[0058] In practical applications, during the pre-construction of the intelligent feature recognition model, the model building module constructs the intelligent feature recognition model through the following operations:

[0059] Step A1: Obtain sample data matching each well area type according to the set logic;

[0060] Step A2: Divide the training sample library and the test sample library based on the sample data;

[0061] Step A3: Based on the training samples, use the least squares regression algorithm based on neural networks to train and iteratively update the training, optimize and determine the key parameters corresponding to the model, and use the test samples to verify the accuracy of the model using the cross-validation method until the set training conditions are met.

[0062] Furthermore, in step A1, in order to ensure the comprehensiveness and reliability of the sample data from the source, the model building module performs the following operations to obtain sample data matching each well area type:

[0063] a1. Considering the diversity of well area types, statistically analyze drilling resistivity imaging data of a set scale;

[0064] a2. Preprocess the drilling resistivity imaging data for each well area type;

[0065] a3. Identify and acquire formation feature data characterized by preprocessed drilling resistivity imaging data;

[0066] a4. Based on the acquired formation feature data, the associated drilling resistivity imaging data are labeled, and all associated labeled combination data are statistically analyzed as a sample dataset, which is then classified and stored by well area type.

[0067] In practical applications, the following approach can be used to establish a formation feature identification model based on intelligent algorithms for drilling resistivity imaging data: a) Collect a large amount of drilling resistivity imaging data;

[0068] b. Preprocess the drilling resistivity imaging data;

[0069] Preprocessing mainly includes image thresholding and region segmentation to obtain useful feature data from the image;

[0070] c. Classify the feature data of electrical imaging while drilling and establish a sample library;

[0071] Features refer to the main formation features reflected on the drilling imaging data map, including but not limited to solution holes, caverns, fractures, laminae, gravels, and layer interfaces, which can be selected according to actual evaluation needs; classifying the drilling electrical imaging data refers to labeling the preprocessed feature data.

[0072] d. Establish a high-precision intelligent recognition model based on appropriate artificial intelligence algorithms;

[0073] A portion of the sample database is selected as the training set, and the other portion is selected as the test set. The model is trained using appropriate intelligent algorithms; these intelligent algorithms include, but are not limited to, machine learning and deep learning algorithms.

[0074] Additionally, it should be noted that when more drilling electrical imaging data is obtained, the model can be further trained and updated based on the new data to optimize the model parameters.

[0075] In practical applications, the chip structure with embedded intelligent feature recognition model algorithm can be integrated with the downhole imaging data acquisition instrument in the form of a downhole black box. This eliminates the need for additional large components, achieving efficient intelligent recognition of downhole data while minimizing the impact on the overall space occupancy of the downhole processing device.

[0076] Each set of sample data obtained based on the above strategy is a combination of resistivity imaging data with known required formation characteristics, which can serve as excellent input for intelligent learning algorithms. Furthermore, the sample data acquired in this invention covers different well types. By constructing matching intelligent feature recognition models based on well type, the influence of differences between different well types can be minimized, significantly improving the matching degree between the intelligent feature recognition model and the input data to be recognized, thus effectively enhancing the accuracy of the recognition results.

[0077] In one specific embodiment, the model building module preprocesses the drilling resistivity imaging data for each well type by including the following operations:

[0078] The collected electrical imaging data during drilling was segmented using the OTSU automatic thresholding method to obtain binary images. Then, the connected components were calculated using a neighborhood calculation method with set parameters to segment the regions.

[0079] Furthermore, in practical applications, the downhole communication module of the present invention can adopt a real-time identification and real-time transmission to the surface method, or it can perform periodic transmission according to a set strategy. Based on this, in one embodiment, the downhole communication module is also provided with a transmission decision unit, which is configured to perform availability identification on the existing feature identification results. If it is determined that there are features required by the current depth stratum in the feature identification results, an upload command is generated. For example, when using a chip module to perform intelligent layer interface identification on downhole electrical imaging data, the real-time data acquired at time i undergoes three preprocessing steps: equalization, gain adjustment, and threshold segmentation. Then, the intelligent algorithm in the chip identifies the layer interface features. If the features of the downhole electrical imaging data match the layer interface, YES is output, and the real-time communication module is controlled to send the depth position parameters and feature identification results to the surface control system. For example, the downhole data transmission device can be used to transmit the depth and feature identification results of formations, fractures, layer interfaces, caves, gravel, etc., to the surface in the form of instructions with a limited number of characters. If the identified feature is not a layer interface, NO is output. If the intelligent model outputs YES at time i (i.e., layer interface information is identified), the depth and layer interface instructions at time i are uploaded to the surface. If layer interface information is not identified (the program outputs NO), intelligent identification is performed at the next time step.

[0080] Furthermore, when employing the technical solution of this invention, surface personnel or the control center do not need to wait for the well to be pulled out and retrieve the logging-while-drilling (LWD) data from the downhole storage module for analysis. Instead, they can obtain significant and effective formation features during the logging-while-drilling process, thus providing decision support for subsequent construction control. However, in actual construction, only personnel with a high level of expertise can typically summarize reasonable construction control recommendations based on limited feature data, limiting the application scenarios. Therefore, in an optional embodiment, the device further includes a construction assistance module configured to analyze the identified periodic formation feature results, make construction recommendation data, and transmit it to the surface control system. This allows even less skilled personnel to promptly detect unsuitable downhole construction conditions and adjust to the optimal logging method.

[0081] In addition, it should be noted that the present invention aims to utilize the intelligent identification model algorithm of the downhole environment in real time to realize the effective characteristics of the downhole measured information. Before well tripping, the information is transmitted to the surface control system or personnel in a smaller data length. In practical applications, the technical concept and execution strategy of the present invention can also be applied to other downhole construction measurement data besides resistivity logging image data according to actual needs, so as to support the optimization processing of downhole data in oil and gas well exploration and development, and provide more comprehensive and flexible optimization support for downhole construction decisions.

[0082] Furthermore, some existing engineering teams employ methods that involve detailed processing directly downhole, such as analyzing well logging imaging data downhole and extracting fracture-related features for specific calculations to identify detailed fracture distribution and attribute information. This places a high burden on downhole equipment and requires a high level of data processing expertise from personnel, necessitating both drilling and logging expertise as well as data statistical analysis expertise. This approach has significant limitations and insufficient practicality. In contrast, this invention utilizes an integrated intelligent feature recognition model, ensuring reliability and recognition efficiency while effectively lowering the specific professional requirements for personnel, making it more suitable for widespread promotion and comprehensive application.

[0083] Application example illustration:

[0084] Taking the intelligent identification of the layer interface of well S in well zone of type X as an example:

[0085] The first step is to build a matching feature recognition model;

[0086] First, the collected electrical imaging data from drilling was segmented using the OTSU automatic thresholding method to obtain binary images. Then, connected components were calculated using an 8-neighborhood calculation method to further segment the regions, such as... Figure 2As shown, the layer interface features are labeled to establish a sample library;

[0087] 70% of the data was selected as the training set for model training, and the remaining 30% was used as the validation set. A least squares regression model based on a neural network was used for training. Key parameters in the model were determined based on the training set sample data, and cross-validation was used to verify the model's accuracy. The main parameters of the model built in this example are as follows: 14 latent variables, 4 neurons, and a test set compliance rate of 94%.

[0088] The second step is to embed the developed intelligent algorithm into the chip;

[0089] Based on the acquisition and processing board related to resistivity imaging data, a high-temperature resistant circuit board was designed. The related circuits form an intelligent module, and the program is written into the chip of the circuit board through relevant FPGA emulators and ARM emulators.

[0090] The third step is to use the chip module to perform intelligent layer interface recognition on the downhole electrical imaging data while drilling.

[0091] The real-time data acquired at time i from downhole is subjected to three preprocessing processes: equalization, gain adjustment, and threshold segmentation. Then, the intelligent algorithm in the chip is used to identify the layer interface features. If the features of the drilling electrical imaging data match the layer interface, YES is output; if the identified features are not layer interfaces, NO is output.

[0092] The fourth step is to determine whether an upload command is required.

[0093] If the intelligent model outputs YES at time i (i.e., the layer interface information has been identified), then the depth and layer interface command at time i are uploaded to the ground; if the layer interface information is not identified (the program outputs NO), then intelligent identification is performed at the next time step; Figure 3 As shown, in this example, layer interface commands were reported at four depths: 6021, 6028, 6032, and 6037, with a high recognition accuracy.

[0094] The scheme described in the above embodiments of the present invention involves pre-establishing a formation feature identification model based on intelligent algorithms for drilling resistivity imaging data; embedding the established intelligent model algorithm into a chip module; acquiring formation resistivity imaging data using a drilling resistivity imaging instrument; performing quality processing on the drilling resistivity imaging data; using the chip module to perform real-time feature identification on the acquired drilling resistivity imaging data; and then sending the corresponding depth point and feature identification result instructions to the surface. This significantly improves timeliness, helps surface engineers obtain more effective formation information in real time, thereby enabling rapid on-site decision-making, effectively reducing drilling risks, and improving logging results and drilling efficiency.

[0095] In the downhole real-time processing device for electrical imaging data provided in this embodiment of the invention, each module or unit structure can operate independently or in combination according to actual calculation and recognition needs to achieve the corresponding technical effects.

[0096] Example 2

[0097] The apparatus has been described in detail in the embodiments disclosed above. Based on other aspects of the apparatus described in any one or more of the above embodiments, the present invention also provides a method for real-time downhole processing of electrical imaging data while drilling. This method is applied to the real-time downhole processing apparatus for electrical imaging data while drilling described in any one or more of the above embodiments. Specific embodiments are given below for detailed description.

[0098] Specifically, Figure 4 The diagram shows a flowchart of the downhole real-time processing method for electrical imaging data provided in an embodiment of the present invention, as shown below. Figure 4 As shown, the method includes:

[0099] The downhole imaging procedure involves using a downhole, drilling-while-drilling resistivity imaging instrument to acquire resistivity imaging data of the required formation.

[0100] The downhole identification process involves calling the set intelligent feature recognition model embedded in the downhole chip structure based on the well area type of the current data, analyzing and identifying the resistivity imaging data, determining the corresponding well logging formation features, and associating and fusing them with the matched formation parameters to form the feature recognition result.

[0101] In the real-time communication process, when there are fused feature recognition results, they are transmitted to the ground control system based on the set transmission protocol.

[0102] The intelligent feature recognition model is pre-built by the model building module based on a set image data processing strategy and machine learning algorithm.

[0103] Specifically, in one embodiment, during the downhole imaging step, the resistivity imaging data acquired by the downhole imaging module includes a two-dimensional array of resistivity data of the formation around the well at the required depth, corresponding to 360° of the formation.

[0104] Furthermore, in one embodiment, the method further includes a downhole data quality control step, which involves real-time removal of bad pixels, equalization processing, gain adjustment, and threshold segmentation of the resistivity imaging data acquired downhole, in order to control the quality of the feature recognition input data.

[0105] In practical applications, a high-temperature resistant circuit board is designed based on the acquisition instruments and computing boards used for resistivity imaging data. A combined intelligent module is constructed based on this circuitry. Therefore, this invention uses FPGA and ARM emulators to write the program into the chip structure of the circuit board. Thus, in one embodiment, the chip structure used in the downhole identification step is a chip structure that has passed high-temperature and high-pressure testing. Before the high-temperature and high-pressure test, the algorithm program for setting up the intelligent feature identification module is imported into the chip structure, categorized by well area type, using FPGA and ARM connectors.

[0106] Furthermore, in one embodiment, the intelligent feature recognition model is constructed in the model building step by the following operations:

[0107] Step A1: Obtain sample data matching each well area type according to the set logic;

[0108] Step A2: Divide the training sample library and the test sample library based on the sample data;

[0109] Step A3: Based on the training samples, use the least squares regression algorithm based on neural networks to train and iteratively update the training, optimize and determine the key parameters corresponding to the model, and use the test samples to verify the accuracy of the model using the cross-validation method until the set training conditions are met.

[0110] In one embodiment, in step A1, the model building module performs the following operations to obtain sample data matching each well area type:

[0111] a1. Considering the diversity of well area types, statistically analyze drilling resistivity imaging data of a set scale;

[0112] a2. Preprocess the drilling resistivity imaging data for each well area type;

[0113] a3. Identify and acquire formation feature data characterized by preprocessed drilling resistivity imaging data;

[0114] a4. Based on the acquired formation feature data, the associated drilling resistivity imaging data are labeled, and all associated labeled combination data are statistically analyzed as a sample dataset, which is then classified and stored by well area type.

[0115] Furthermore, in a preferred embodiment, the process of preprocessing the drilling resistivity imaging data for each well type includes:

[0116] The collected electrical imaging data during drilling was segmented using the OTSU automatic thresholding method to obtain binary images. Then, the connected components were calculated using a neighborhood calculation method with set parameters to segment the regions.

[0117] In one optional embodiment, a transmission decision step is included before uploading the feature recognition results to the surface control system. This step assesses the availability of the existing feature recognition results. If it is determined that the feature recognition results contain features required for the current depth of the formation, an upload command is generated. For example, when using a chip module to perform intelligent layer interface recognition on downhole electrical imaging data, the real-time data acquired at time i undergoes three preprocessing steps: equalization, gain adjustment, and threshold segmentation. Then, the intelligent algorithm in the chip identifies the layer interface features. If the downhole electrical imaging data features match the layer interface, YES is output; if the identified features are not layer interfaces, NO is output. If the intelligent model outputs YES at time i (i.e., layer interface information is identified), the depth and layer interface command at time i are uploaded to the surface. If no layer interface information is identified (the program outputs NO), intelligent recognition is performed at the next time step.

[0118] Furthermore, in one embodiment, the method further includes a construction assistance step: analyzing the identified periodic stratigraphic characteristics, making construction recommendation data, and transmitting it to the surface control system.

[0119] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0120] It should be noted that, in other embodiments of the present invention, the method can also combine one or more of the above embodiments to obtain a new downhole real-time processing method for logging-while-drilling electrical imaging data, so as to optimize the logging-while-drilling technology.

[0121] It should be noted that, based on the methods in any one or more embodiments of the present invention described above, the present invention also provides a storage medium storing program code that can implement the methods described in any one or more embodiments. When the program code is executed by the operating system, it can implement the downhole real-time processing method for drilling electrical imaging data as described above.

[0122] It should be understood that the embodiments disclosed herein are not limited to the specific structures, processing steps, or materials disclosed herein, but should be extended to equivalent substitutions of these features as understood by those skilled in the art. It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting.

[0123] The phrase "an embodiment" in the specification means that a specific feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. Therefore, the phrase "an embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment.

[0124] While the embodiments disclosed in this invention are as described above, the content is merely for the purpose of facilitating understanding of the invention and is not intended to limit the invention. Any person skilled in the art to which this invention pertains may make any modifications and variations in form and detail of the implementation without departing from the spirit and scope disclosed herein; however, the scope of patent protection for this invention shall still be determined by the scope defined in the appended claims.

Claims

1. A downhole real-time processing device for electrical imaging data while drilling, characterized in that, The device includes: The downhole imaging module is configured to acquire resistivity imaging data of the required formation in the downhole environment using a drilling resistivity imaging instrument. The intelligent recognition module is connected to the downhole imaging module. It analyzes and identifies the collected resistivity imaging data through a chip structure with an embedded intelligent feature recognition model, determines the corresponding well logging formation features, and associates and fuses them with the matching formation parameters to form the feature recognition result. The real-time communication module is configured to transmit the fused feature recognition results to the ground control system based on a set transmission protocol when such results are available. The intelligent feature recognition model is pre-built by the model building module based on a set image data processing strategy and machine learning algorithm; Using FPGA and ARM connectors, the algorithm program for setting up the intelligent feature recognition model is imported into the chip structure by labeling it by well area type; the chip structure embedded with the intelligent feature recognition model algorithm is integrated with the downhole imaging module in the form of a downhole black box; The model building module performs the following operations to obtain sample data for building the intelligent feature recognition model: a1. Considering the diversity of well area types, statistically analyze drilling resistivity imaging data of a set scale; a2. Preprocess the drilling resistivity imaging data for each well area type; a3. Identify and acquire formation feature data characterized by preprocessed drilling resistivity imaging data; the formation feature data includes data on solution holes, caverns, fractures, laminae, gravel, and layer interfaces. a4. Based on the acquired formation feature data, the associated drilling resistivity imaging data are labeled, and all the combined data of the associated labels are counted as a sample dataset, which is then classified and stored by well area type. When the model building module preprocesses the drilling resistivity imaging data for each well type, it includes the following operations: The collected electrical imaging data during drilling was segmented using the OTSU automatic thresholding method to obtain binary images. Then, the connected components were calculated using a neighborhood calculation method with set parameters to segment the regions.

2. The apparatus according to claim 1, characterized in that, The resistivity imaging data acquired by the downhole imaging module includes a two-dimensional array of resistivity data of the formation around the well at the required depth, corresponding to a 360° radius around the well.

3. The apparatus according to claim 1, characterized in that, The intelligent recognition module is a chip structure that has passed high temperature and high pressure testing. Before the high temperature and high pressure test, the algorithm program for setting the intelligent feature recognition model is imported into the chip structure.

4. The apparatus according to claim 1, characterized in that, The model building module constructs the intelligent feature recognition model through the following operations: Step A1: Obtain sample data matching each well area type according to the set logic; Step A2: Divide the training sample library and the test sample library based on the sample data; Step A3: Based on the training samples, use the least squares regression algorithm based on neural networks to iteratively update and train the model, optimize and determine the key parameters corresponding to the model, and use the test samples to verify the accuracy of the model using the cross-validation method until the set training conditions are met.

5. The apparatus according to claim 1, characterized in that, The real-time communication module is also equipped with a transmission decision unit, which is configured to perform availability assessment on the existing feature recognition results. If it is determined that there are features required for the current depth strata in the feature recognition results, an upload command is generated.

6. The apparatus according to claim 1, characterized in that, The device also includes a downhole data quality control module, which is configured to perform real-time bad pixel removal, equalization processing, gain adjustment, and threshold segmentation processing on the resistivity imaging data acquired downhole, so as to control the quality of the feature recognition input data.

7. The apparatus according to claim 1, characterized in that, The device also includes a construction assistance module, which analyzes the identified periodic geological features, makes construction recommendations, and transmits the data to the ground control system.

8. A method for real-time downhole processing of electrical imaging data while drilling, wherein the method is applied to the apparatus described in any one of claims 1 to 7, characterized in that, The method includes: The downhole imaging process involves using a downhole, drilling-while-drilling resistivity imaging instrument to acquire resistivity imaging data of the required formation. The downhole identification process involves calling the set intelligent feature recognition model embedded in the downhole chip structure based on the well area type of the current data to analyze and identify the resistivity imaging data, determine the corresponding well logging formation features, and correlate and fuse them with the matching formation parameters to form the feature recognition result. Specifically, the algorithm program of the set intelligent feature recognition model is imported into the chip structure by labeling it with the well area type using FPGA connectors and ARM connectors. The chip structure with the embedded intelligent feature recognition model algorithm is then integrated with the downhole imaging module in the form of a downhole black box. In the real-time communication process, when there are fused feature recognition results, they are transmitted to the ground control system based on the set transmission protocol. The intelligent feature recognition model is pre-built by the model building module based on a set image data processing strategy and machine learning algorithm; the following operations are performed to obtain sample data for building the intelligent feature recognition model: a1. Considering the diversity of well area types, statistically analyze drilling resistivity imaging data of a set scale; a2. Preprocess the drilling resistivity imaging data for each well area type; a3. Identify and acquire formation feature data characterized by preprocessed drilling resistivity imaging data; the formation feature data includes data on solution holes, caverns, fractures, laminae, gravel, and layer interfaces. a4. Based on the acquired formation feature data, the associated drilling resistivity imaging data are labeled, and all the combined data of the associated labels are counted as a sample dataset, which is then classified and stored by well area type. When the model building module preprocesses the drilling resistivity imaging data of various well types, it includes the following operations: the collected drilling resistivity imaging data is segmented using the OTSU automatic threshold segmentation method to obtain binary images, and then connected components are calculated using a neighborhood calculation method with set parameters to segment the regions.

Images