Well cementation quality control method and device based on graph structure

Abstract

The invention provides a well cementation quality control method and device based on a graph structure. Well cementation parameters of the target well are obtained, and a well cementation quality evaluation result of the target well is determined according to the well cementation parameters of the target well and a preset evaluation system; according to the well cementation quality evaluation result of the target well, whether the well cementation quality of the target well meets a preset quality requirement or not is judged; under the condition that the well cementation quality of the target well does not meet the preset quality requirement, target map structure data corresponding to the target well are obtained; determining a target well cementation parameter in the well cementation parameters of the target well according to a well cementation quality evaluation result of the target well, and determining a second well cementation parameter having a preset association relationship with the target well cementation parameter according to the target map structure data; and according to the target well cementation parameters and the second well cementation parameters, well cementation operation parameters of the target well are adjusted. Therefore, the defects of insufficient data utilization, inaccurate quality evaluation and lack of parameter adjustment strategies in the traditional cementing operation quality control are overcome.

Description

Technical Field

[0001] This specification belongs to the field of oil and gas development technology, and in particular relates to a graph-based cementing quality control method and device. Background Art

[0002] Currently, existing cementing operation control methods mainly evaluate cementing quality based on a single data source, and mainly set some simple quality indicators based on experience. This leads to problems such as insufficient utilization of multi-source data, inaccurate cementing quality assessment, and lack of effective optimization strategies for the operation process.

[0003] To address the above issues, no effective solutions have been proposed so far. Summary of the Invention

[0004] This specification provides a graph-based cementing quality control method and device. First, a preliminary assessment of cementing quality is performed using the target well's cementing parameters and a preset evaluation system. If the target well's cementing quality fails to meet preset requirements, a preset generation model is invoked, using graph data constructed from multiple first wells as a reference. By matching the target well's cementing parameters, the target cementing parameters and secondary cementing parameters with pre-set correlations are extracted, allowing targeted optimization and adjustment of cementing operation parameters. Compared to existing methods, this method achieves deep data fusion and correlation analysis, overcoming the shortcomings of traditional cementing operation control, which often suffer from insufficient data utilization, inaccurate quality assessments, and a lack of parameter adjustment strategies.

[0005] This specification provides a graph-based cementing quality control method, including:

[0006] Acquiring cementing parameters of a target well, and determining a cementing quality evaluation result of the target well based on the cementing parameters of the target well and a preset evaluation system;

[0007] Determining whether the cementing quality of the target well meets preset quality requirements based on the cementing quality assessment result of the target well;

[0008] When the cementing quality of the target well does not meet a preset quality requirement, target graph structure data corresponding to the target well is obtained; wherein the target graph structure data is constructed based on first cementing parameters of a plurality of first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node association relationships determined based on association relationships between the first cementing parameters;

[0009] determining a target cementing parameter among the cementing parameters of the target well according to a cementing quality assessment result of the target well, and determining a second cementing parameter having a preset correlation relationship with the target cementing parameter according to the target graph structure data;

[0010] According to the target cementing parameters and the second cementing parameters, the cementing operation parameters of the target well are adjusted. In one embodiment, before obtaining the target graph structure data corresponding to the target well, the method further includes:

[0011] Obtaining first cementing parameters and first cementing quality assessment results of a plurality of the first wells;

[0012] Using the first module of the preset generation model, based on the first cementing parameter, performing semantic extraction processing on the first cementing parameter to obtain a first feature vector; and using the second module of the preset generation model, based on the first cementing parameter, performing knowledge extraction processing on the first cementing parameter to obtain a second feature vector;

[0013] The target graph structure data is determined by utilizing the third module of the preset generation model according to the first eigenvector, the second eigenvector, and the first cementing quality evaluation result.

[0014] In one embodiment, the formation characteristics of the multiple first wells are different, and the first module of the preset generation model is used to perform semantic extraction processing on the first cementing parameter based on the first cementing parameter to obtain a first feature vector, and the second module of the preset generation model is used to perform knowledge extraction processing on the first cementing parameter based on the first cementing parameter to obtain a second feature vector, including:

[0015] Mapping the first cementing parameter to a preset space using a fourth module of the preset generation model to obtain a fourth eigenvector;

[0016] The first module of the preset generation model is used to perform semantic extraction processing on the fourth feature vector to obtain the first feature vector, and the second module of the preset generation model is used to perform knowledge extraction processing on the fourth feature vector to obtain the second feature vector.

[0017] In one embodiment, adjusting the cementing operation parameters of the target well according to the target cementing parameters and the second cementing parameters includes:

[0018] Performing multi-physics field coupling simulation processing according to the cementing parameters of the target well, and detecting whether there are abnormal working conditions during the simulation processing;

[0019] When an abnormal working condition is detected in the simulation process, performing matching verification processing on the abnormal working condition, the target cementing parameter and the second cementing parameter;

[0020] When the abnormal working condition, the target cementing parameters and the second cementing parameters successfully match, the cementing operation parameters of the target well are adjusted according to the cementing parameters among the target cementing parameters and the second cementing parameters that match the abnormal working condition.

[0021] In one embodiment, the method further comprises:

[0022] receiving an update request for the target graph structure data;

[0023] In response to the update request, determining a second cementing parameter and a second cementing quality evaluation result of a second well corresponding to the update request;

[0024] determining a similarity between the second cementing parameter and the first cementing parameter, and determining a candidate cementing parameter in the second cementing parameter based on the similarity;

[0025] The target graph structure data is updated according to the candidate cementing parameters and the second cementing quality evaluation result.

[0026] In one embodiment, the method further comprises:

[0027] Screening the cementing parameters of the target well to obtain a third cementing parameter, and generating verifiable information corresponding to the third cementing parameter;

[0028] The cementing quality evaluation result of the target well and the verifiable information are stored in a preset blockchain, and the cementing parameters of the target well except the third cementing parameter are stored in a preset storage device, so that when a traceability request for the cementing quality evaluation process of the target well is received, backtracking processing is performed according to the preset blockchain and the data stored in the preset storage device.

[0029] In one embodiment, the preset evaluation system includes scoring criteria corresponding to the cementing parameters and preset weights corresponding to each cementing parameter. The preset weights corresponding to the cementing parameters are determined based on historical cementing parameters of historical wells and historical cementing quality evaluation results.

[0030] This specification provides a cementing quality control device based on a graph structure, including:

[0031] An evaluation result determination module is used to obtain cementing parameters of a target well and determine a cementing quality evaluation result of the target well based on the cementing parameters of the target well and a preset evaluation system;

[0032] A demand judgment module is used to judge whether the cementing quality of the target well meets the preset quality requirements based on the cementing quality evaluation result of the target well;

[0033] a graph structure generation module, configured to obtain target graph structure data corresponding to the target well if the cementing quality of the target well does not meet preset quality requirements; wherein the target graph structure data is constructed based on first cementing parameters of a plurality of first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node association relationships determined based on association relationships between the first cementing parameters;

[0034] a parameter determination module, configured to determine a target cementing parameter among the cementing parameters of the target well according to a cementing quality assessment result of the target well, and to determine a second cementing parameter having a preset correlation relationship with the target cementing parameter according to the target graph structure data;

[0035] The parameter adjustment module is used to adjust the cementing operation parameters of the target well according to the target cementing parameters and the second cementing parameters.

[0036] This specification also provides an electronic device, including a processor and a memory for storing processor-executable instructions, wherein the processor implements a graph-structure-based cementing quality control method when executing the instructions.

[0037] This specification also provides a computer-readable storage medium having computer instructions stored thereon, which implement a graph-based cementing quality control method when the instructions are executed.

[0038] A graph-structured cementing quality control method provided herein obtains cementing parameters of a target well and determines a cementing quality assessment result of the target well based on the cementing parameters of the target well and a preset assessment system. Based on the cementing quality assessment result of the target well, it is determined whether the cementing quality of the target well meets preset quality requirements. If the cementing quality of the target well does not meet the preset quality requirements, target graph structure data corresponding to the target well is obtained. The target graph structure data is constructed based on first cementing parameters of a plurality of first wells using a preset generation model, and includes nodes corresponding to the first cementing parameters and node associations determined based on associations between the first cementing parameters. Based on the cementing quality assessment result of the target well, a target cementing parameter is determined among the cementing parameters of the target well, and a second cementing parameter having a preset association with the target cementing parameter is determined based on the target graph structure data. Cementing operation parameters of the target well are adjusted based on the target cementing parameter and the second cementing parameter. This approach first uses the target well's cementing parameters and a pre-set evaluation system to conduct a preliminary assessment of cementing quality. If the target well's cementing quality falls short of pre-set requirements, the pre-set generation model is further invoked, using the graph structure data constructed from multiple first wells as a reference. By matching the target well's cementing parameters, the target cementing parameters and the secondary cementing parameters with which they have a pre-set correlation are extracted, allowing for targeted optimization and adjustment of cementing operation parameters. Compared to existing methods, this approach achieves deep data fusion and correlation analysis, overcoming the shortcomings of traditional cementing operation control, such as insufficient data utilization, inaccurate quality assessments, and a lack of parameter adjustment strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of this specification, the following is a brief introduction to the drawings required for use in the embodiments. The drawings described below are only some of the embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0040] Figure 1 This is a flow chart of a cementing quality control method based on a graph structure provided by an embodiment of this specification;

[0041] Figure 2 This is a schematic diagram of the structure of an electronic device provided by an embodiment of this specification;

[0042] Figure 3 This is a schematic diagram of the structural composition of a cementing quality control device based on a graph structure provided by an embodiment of this specification;

[0043] Figure 4This is a flow chart of another graph-based cementing quality control method provided by an embodiment of this specification;

[0044] Figure 5 This is a schematic diagram of constructing a cementing knowledge graph provided by an embodiment of this specification;

[0045] Figure 6 This is a schematic diagram of a cementing operation optimization control system provided by an embodiment of this specification. DETAILED DESCRIPTION

[0046] To help those skilled in the art better understand the technical solutions in this specification, the following will provide a clear and complete description of the technical solutions in the embodiments of this specification, in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. All other embodiments derived by those skilled in the art based on the embodiments in this specification without creative effort shall fall within the scope of protection of this specification.

[0047] Cementing is a critical step in the oil drilling process. Its quality is directly related to the long-term stable production of oil and gas wells and the safety and economy of subsequent extraction operations. With the continuous development of the oil industry, the requirements for cementing quality are becoming increasingly higher.

[0048] Cementing technology has made significant progress over the past few decades. In the early days, cementing operations relied primarily on manual experience and simple monitoring methods, resulting in a relatively crude assessment of cementing quality. With the advancement of information technology, automated monitoring equipment has been introduced, enabling real-time monitoring of some cementing process parameters, such as pressure and displacement. In recent years, the rise of big data and artificial intelligence technologies has provided new ideas and methods for intelligent control of cementing operations.

[0049] However, cementing operation control currently faces numerous challenges. Firstly, the cementing process involves a variety of complex factors, including drilling parameters, formation characteristics, and cement slurry properties. These factors are interrelated and have complex influences, making it difficult for existing technologies to comprehensively and accurately analyze their combined impact on cementing quality. Secondly, there is a lack of scientific and systematic methods for evaluating cementing quality. Traditional evaluation indicators are limited and cannot fully reflect the overall quality of cementing operations, resulting in the inability to promptly identify potential problems and effectively control them.

[0050] Some existing methods rely primarily on a single data source to assess cementing quality. For example, they utilize only real-time pressure data from the cementing process, using a preset pressure threshold to determine whether the cementing process is normal. When the pressure exceeds the threshold, a cementing quality issue is considered. Simultaneously, simple quality indicators, such as whether the cement slurry return height meets design requirements, are established based on experience. Traditional statistical analysis methods are often used to analyze data, which is relatively simplistic and difficult to understand the complex underlying relationships. Furthermore, relying on a single data source fails to fully consider the numerous factors that influence cementing quality, such as formation characteristics and other drilling parameters. This results in biased assessments and can easily overlook potential quality issues. Furthermore, traditional statistical analysis methods are unable to effectively handle complex nonlinear relationships and accurately map the inherent connections between different factors and cementing quality. This results in inaccurate cementing quality assessments and hinders effective control of cementing operations. While existing technologies incorporate data analysis, these methods are relatively simple and can only handle linear relationships between parameters, making it difficult to accurately model the complex nonlinear relationships involved in cementing operations. This results in an inaccurate model, leading to significant errors in the assessment and prediction of cementing quality. Furthermore, while this technical solution utilizes some data to build the model, the data source remains relatively limited, failing to fully integrate other important data such as drilling and cementing. This limits the model's accuracy and reliability, making it unable to fully reflect the true state of cementing quality. Furthermore, the pre-set automatic control rules are relatively fixed, lacking flexibility and adaptability, making them difficult to cope with the complex and changing downhole conditions and preventing precise control of cementing operations.

[0051] To address the root causes of these issues, this manual first conducts a preliminary assessment of cementing quality using the target well's cementing parameters and a pre-set evaluation system. If the target well's cementing quality falls short of pre-set requirements, the system then calls upon a pre-set generation model, using graph data from multiple first wells as a reference. By matching the target well's cementing parameters, the system extracts the target cementing parameters and the secondary cementing parameters that have a pre-set correlation with them, allowing for targeted optimization and adjustment of cementing operation parameters. Compared to existing methods, this method achieves deep data fusion and correlation analysis, overcoming the shortcomings of traditional cementing operation control, which often include insufficient data utilization, inaccurate quality assessments, and a lack of parameter adjustment strategies.

[0052] See Figure 1 As shown, the embodiment of this specification provides a graph-based cementing quality control method, wherein the method is specifically applied to the server side. In specific implementation, the method may include the following contents:

[0053] S101: Acquire cementing parameters of a target well, and determine a cementing quality evaluation result of the target well based on the cementing parameters of the target well and a preset evaluation system;

[0054] S102: judging whether the cementing quality of the target well meets a preset quality requirement based on the cementing quality assessment result of the target well;

[0055] S103: If the cementing quality of the target well does not meet a preset quality requirement, obtaining target graph structure data corresponding to the target well; wherein the target graph structure data is constructed based on first cementing parameters of a plurality of first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node association relationships determined based on association relationships between the first cementing parameters;

[0056] S104: determining a target cementing parameter among the cementing parameters of the target well according to the cementing quality assessment result of the target well, and determining a second cementing parameter having a preset correlation relationship with the target cementing parameter according to the target graph structure data;

[0057] S105: Adjusting cementing operation parameters of the target well according to the target cementing parameters and the second cementing parameters.

[0058] The cementing parameters mentioned above can be obtained from the drilling logs of the oil well, detailed formation property data extracted from logging data, and information such as cement slurry formula and construction time collected from cementing operation records. Specifically, this data can include three categories: drilling, logging, and cementing. These include wellbore parameters during drilling (such as wellbore dimensions and well inclination), formation property data obtained from logging (such as resistivity and acoustic wave transit time), and construction parameters during cementing operations (such as cement slurry formula and displacement volume).

[0059] The above-mentioned preset evaluation system can be a cementing quality evaluation framework constructed based on multi-source data fusion and historical operation experience. The system sets clear quality standards and thresholds through a comprehensive evaluation of quantitative indicators of drilling, logging and cementing parameters (such as compressive strength, sealing, durability, etc.), and can objectively judge whether the cementing operation meets the preset requirements.

[0060] Specifically, it may include primary indicators, secondary indicators, evaluation criteria, data sources, weights corresponding to primary indicators, and weights corresponding to secondary indicators. Table 1 shows the preset evaluation system for cementing operations.

[0061] Table 1

[0062]

[0063]

[0064] The above-mentioned cementing quality assessment results can be presented in various forms through a preset assessment system. For example, a numerical comprehensive quality score (such as 0 to 100 points) is obtained by weighted calculation of the cementing parameters of the target well, which intuitively reflects the overall effect of the cementing operation and the bond strength. At the same time, the score can also be subdivided into grades such as excellent, good, medium, and poor, or qualitative labels such as "normal" and "abnormal" can be used to quickly determine whether the cementing quality meets the preset requirements. In addition, multiple key indicators such as cementing integrity index, bonding strength coefficient and risk level can also be output to provide a comprehensive and quantitative decision-making basis for subsequent cementing operation parameter adjustments and risk warnings.

[0065] For example, suppose the cementing quality assessment results of a well are as follows:

[0066] Cement bond quality: bond strength (excellent, 15 points), uniformity (good, 9 points), integrity (excellent, 6 points) → total score 30 / 50;

[0067] Annulus seal quality: air seal (excellent, 8 points), liquid seal (excellent, 6 points), micro-annulus (good, 4 points) → total score 18 / 30;

[0068] Construction process control: cement slurry performance (good, 3.5 points), displacement efficiency (excellent, 4 points), downhole condition control (good, 2.5 points) → total score 10 / 15;

[0069] Long-term stability: corrosion resistance (excellent, 4 points), formation stress resistance (good, 3.5 points), durability test (good, 2.5 points) → total score 10 / 10;

[0070] Construction efficiency: cementing operation time (good, 6 points), equipment utilization (excellent, 4.5 points), smooth process connection (good, 3 points) → total score 13.5 / 15;

[0071] Cost control: Material cost (excellent, 4 points), Labor cost (good, 3 points), Equipment cost (good, 3 points) → Total score 10 / 10;

[0072] Total score = ∑ (indicator score × total weight), which is ultimately used for horizontal comparison or decision-making improvement.

[0073] The target graph structure data described above can be an expression that abstracts the first cementing (or target well) parameters into nodes in a graph and establishes edge connections based on the mutual influence relationship between the parameters. This structure not only records the individual information of each cementing parameter, but also quantitatively describes the dependencies and interaction mechanisms between the parameters. For example, if the first cementing parameters include cement slurry density, injection rate, and well depth, then in the target graph structure, each parameter corresponds to a node; if data analysis shows that changes in injection rate have a significant impact on cement slurry density, and well depth also indirectly affects injection rate, then an edge is established between the cement slurry density node and the injection rate node, and another edge is established between the injection rate node and the well depth node, and appropriate weights are assigned to reflect the intensity of the influence.

[0074] The target cementing parameters mentioned above may be key parameters identified during the cementing quality assessment process as directly affecting the cementing effect. For example, if the assessment results show that cement slurry density has a significant impact on the overall cementing strength, then cement slurry density may be considered as the target cementing parameter.

[0075] The aforementioned second cementing parameters can be other related parameters determined based on the target graph structure data and having a preset correlation with the target cementing parameters. While these parameters are not the primary indicators that directly determine cementing quality, they have a certain degree of dependence or interaction with the target cementing parameters and can indirectly affect cementing quality. For example, if cement slurry density is the target cementing parameter, injection rate or displacement efficiency may be considered as related secondary cementing parameters. By adjusting these parameters, the cementing operation can be further optimized.

[0076] In some embodiments, adjusting the cementing operation parameters of the target well according to the target cementing parameters and the second cementing parameters may include:

[0077] Target cementing parameters represent key indicators that directly impact cementing quality (such as cement slurry density), while secondary cementing parameters are closely related auxiliary parameters (such as grouting rate or displacement efficiency) determined through graph-structured data. The target well cementing quality is first assessed. If the results are not up to standard, the cementing operation parameters are adjusted to improve the target cementing parameters based on the preset relationship between the two. For example, if the assessment shows that the cement slurry density is below the ideal range and analysis shows a positive correlation between the grouting rate and cement slurry density, the system may recommend reducing the grouting rate to allow the cement slurry more time to fully solidify and bond to the wellbore wall, improving the overall cementing quality.

[0078] Based on the above embodiment, a preliminary assessment of cementing quality is first performed using the target well's cementing parameters and a preset evaluation system. If the target well's cementing quality does not meet preset requirements, the preset generation model is further invoked, using the graph structure data constructed from multiple first wells as a reference. By matching the target well's cementing parameters, the target cementing parameters and the second cementing parameters with which they have a preset correlation are extracted, and targeted optimization and adjustment of the cementing operation parameters are then performed. Compared to existing methods, this method achieves deep data fusion and correlation analysis, overcoming the shortcomings of traditional cementing operation control, such as insufficient data utilization, inaccurate quality assessment, and lack of parameter adjustment strategies.

[0079] In some embodiments, before acquiring the target graph structure data corresponding to the target well, the method may further include the following steps when implemented:

[0080] S1: Acquire a plurality of first cementing parameters and first cementing quality evaluation results of the first wells;

[0081] S2: using the first module of the preset generation model, based on the first cementing parameter, performing semantic extraction processing on the first cementing parameter to obtain a first feature vector, and using the second module of the preset generation model, based on the first cementing parameter, performing knowledge extraction processing on the first cementing parameter to obtain a second feature vector;

[0082] S3: Utilizing the third module of the preset generation model, the target graph structure data is determined according to the first eigenvector, the second eigenvector, and the first cementing quality evaluation result.

[0083] In some embodiments, the first module of the preset generation model is used to perform semantic extraction processing on the first cementing parameter based on the first cementing parameter to obtain a first feature vector, and the second module of the preset generation model is used to perform knowledge extraction processing on the first cementing parameter based on the first cementing parameter to obtain a second feature vector. Specifically, the following steps may be included:

[0084] The first module of the pre-set generative model performs semantic extraction on the first cementing parameters of multiple first wells, extracting a first feature vector that reflects the overall semantics of the cementing operation (such as consistency and stability). Simultaneously, the second module of the pre-set generative model performs knowledge extraction on the same input, capturing specialized information related to cementing technology (such as bond strength and rheological properties) to form a second feature vector. For example, if the cementing parameters of a target well include cement slurry density, grouting rate, and well depth, semantic extraction may generate a vector describing overall stability, while knowledge extraction may highlight a vector reflecting cement slurry performance.

[0085] In some embodiments, the third module using the preset generation model to determine the target graph structure data according to the first eigenvector, the second eigenvector, and the first cementing quality assessment result may include:

[0086] First, the first feature vector obtained by the first module (reflecting the overall semantic information of cementing parameters) is fused with the second feature vector obtained by the second module (capturing cementing technical details and professional knowledge) to form a joint feature representation. Second, the joint feature representation is combined with the first cementing quality assessment result. The preset matching algorithm and rules are used to analyze the correlation between cementing parameters and determine the nodes of each key parameter in the graph structure. Finally, the target graph structure data is constructed based on the matching and association analysis results. The target graph structure data is presented as nodes (representing cementing parameters) and edges (representing preset correlations between parameters), which intuitively reveals the factors affecting the cementing quality of the target well.

[0087] Based on the above embodiment, the semantic and knowledge dual abstraction of cementing parameters is achieved. By constructing intuitive graph structure data, the factors affecting cementing quality are effectively revealed, providing a quantitative basis for subsequent parameter adjustment and operation optimization.

[0088] In some embodiments, the formation characteristics of the multiple first wells are different. The first module of the preset generation model is used to perform semantic extraction processing on the first cementing parameter based on the first cementing parameter to obtain a first feature vector. The second module of the preset generation model is used to perform knowledge extraction processing on the first cementing parameter based on the first cementing parameter to obtain a second feature vector. When the method is specifically implemented, it may further include the following contents:

[0089] S1: using the fourth module of the preset generation model, mapping the first cementing parameter to a preset space to obtain a fourth eigenvector;

[0090] S2: Using the first module of the preset generation model, perform semantic extraction processing on the fourth feature vector to obtain the first feature vector, and using the second module of the preset generation model, perform knowledge extraction processing on the fourth feature vector to obtain the second feature vector.

[0091] In some embodiments, the fourth module of the preset generation model is used to map the first cementing parameter to a preset space to obtain a fourth eigenvector. Specifically, the fourth module may include:

[0092] Since the formation characteristics of multiple first wells vary, in order to overcome these heterogeneities, the fourth module of the preset generative model uses the dimensionality reduction method in transfer learning to map the original first cementing parameters to a preset low-dimensional unified space to generate a fourth eigenvector. Specifically, this module learns the common features in historical data and performs nonlinear dimensionality reduction and feature alignment on cementing parameters (such as cement slurry density, grouting rate, well depth, etc.) under different formation conditions, so that the cementing parameters of each well can be compared and analyzed in the same space after dimensionality reduction. For example, suppose two first wells are located in sandstone and shale layers respectively, and their cementing parameters are quite different in the original space; after the mapping processing of the fourth module, these parameters are projected into a unified low-dimensional space. The generated fourth eigenvector not only retains the key information reflecting the cementing quality, but also eliminates the interference caused by formation differences.

[0093] Based on the above embodiment, the dimensionality reduction technology of transfer learning is used to map the high-dimensional, heterogeneous cementing parameters of multiple first wells, which are derived from different formation characteristics, into a preset unified low-dimensional space. This not only significantly reduces data dimensionality and noise interference, but also achieves effective alignment of data across different formations, laying a solid foundation for subsequent semantic and knowledge extraction. The processed fourth eigenvector can more accurately reflect key information about the cementing operation, improve the accuracy and robustness of cementing quality assessment, and provide more robust and transferable feature support for optimizing cementing operation parameters.

[0094] In some embodiments, the method of adjusting the cementing operation parameters of the target well according to the target cementing parameters and the second cementing parameters may further include the following when implemented:

[0095] S1: performing multi-physics field coupling simulation processing according to the cementing parameters of the target well, and detecting whether there are abnormal working conditions during the simulation processing;

[0096] S2: When an abnormal working condition is detected in the simulation process, performing matching verification processing on the abnormal working condition, the target cementing parameter, and the second cementing parameter;

[0097] S3: When the abnormal working condition, the target cementing parameters, and the second cementing parameters successfully match, adjusting the cementing operation parameters of the target well according to the cementing parameters among the target cementing parameters and the second cementing parameters that match the abnormal working condition.

[0098] In some embodiments, performing multi-physics field coupling simulation processing according to the cementing parameters of the target well and detecting whether there are abnormal working conditions during the simulation processing may include:

[0099] A comprehensive simulation of the target well's cementing parameters is conducted through multi-physics coupled simulation, encompassing key physical fields such as temperature, pressure, fluid dynamics, and formation stress. This simulation facilitates in-depth analysis of dynamic changes during the cementing process, such as cement slurry flow, interfacial bond strength, and temperature distribution. If abnormal operating conditions are detected during the simulation (such as sudden wellbore temperature changes, trapped annular air bubbles, or abnormal cement slurry thickening), further analysis will be conducted to determine their causes.

[0100] In some embodiments, when an abnormal operating condition is detected in the simulation processing, the abnormal operating condition, the target cementing parameter and the second cementing parameter are matched and verified. In specific implementation, for example, if an abnormal increase in the temperature of a certain section of the wellbore is found in the simulation, the phenomenon is matched with the cement slurry viscosity, the circulating pressure drop parameter and the formation permeability parameter to verify whether these parameters may cause the abnormality.

[0101] In some embodiments, when the abnormal operating condition, the target cementing parameters, and the second cementing parameters successfully match, the cementing operation parameters of the target well are adjusted based on the target cementing parameters and the cementing parameters of the second cementing parameters that match the abnormal operating condition. In specific implementation, for example, if the abnormal temperature increase is caused by high cement slurry viscosity, the system may recommend adjusting the cement slurry formula to reduce its viscosity, or optimizing the pumping displacement to reduce heat accumulation.

[0102] Specifically, the digital twin system collects dynamic parameters such as downhole pressure, temperature, and cement slurry return rate in real time, combines geological models with historical data to build a virtual twin, and simultaneously simulates cement slurry flow, cementation process, and annular sealing status (such as predicting cement slurry leakage path or gas channeling risk); when abnormal working conditions are monitored (such as sudden drop in pump pressure, abnormal return volume), the system automatically triggers graded intervention based on the preset multi-level response strategy - mild deviations (such as a 5% decrease in displacement efficiency) start a yellow warning and fine-tune the pump speed parameters, moderate risks (such as leakage rate of 2m 3 / h) switches to orange mode and injects plugging fiber slurry. In the event of a serious fault (such as wellhead pressure exceeding the threshold), the red shutdown command is executed and the emergency casing seal is activated. At the same time, the effectiveness of the optimization measures is verified in real time through the twin, forming a "perception-simulation-decision-verification" closed loop, ultimately achieving a transition from passive response to active control in cementing operations.

[0103] Based on the above examples, by combining multi-physics simulation with abnormal condition detection, potential risks can be identified early and targeted measures can be taken to prevent the spread or worsening of abnormalities. Furthermore, the matching verification process ensures the accuracy of parameter adjustments, thereby reducing resource waste caused by ineffective adjustments and improving the safety and quality stability of cementing operations.

[0104] In some embodiments, the method may further include the following when implemented:

[0105] S1: receiving an update request for the target graph structure data;

[0106] S2: In response to the update request, determining a second cementing parameter and a second cementing quality evaluation result of a second well corresponding to the update request;

[0107] S3: determining a similarity between the second cementing parameter and the first cementing parameter, and determining a candidate cementing parameter in the second cementing parameter based on the similarity;

[0108] S4: updating the target graph structure data according to the candidate cementing parameters and the second cementing quality evaluation result.

[0109] In some embodiments, determining the similarity between the second cementing parameter and the first cementing parameter, and determining candidate cementing parameters in the second cementing parameter based on the similarity, may include:

[0110] The similarity can be determined based on various metrics, such as numerical proximity, parameter trend matching, and parameter response behavior under specific operating conditions. For example, if the cement slurry viscosity in the first cementing parameter is 1.5 Pa·s, and a parameter in the second cementing parameter is 1.6 Pa·s, and the two values ​​are close and both exhibit stable flow characteristics under similar formation conditions, then this parameter can be selected as a candidate cementing parameter.

[0111] Based on the above examples, similarity analysis enables precise parameter selection and avoids blind parameter adjustments. Particularly in environments with complex formation conditions and strong parameter correlations, similarity-based candidate parameter screening helps quickly identify the parameters that best match the target operating conditions, improving the accuracy of cementing parameter adjustments. Furthermore, this method can leverage successful cases from historical data and draw on effective parameter configurations under similar geological conditions, further improving the reliability and success rate of cementing operations.

[0112] In some embodiments, the method may further include the following when implemented:

[0113] S1: screening the cementing parameters of the target well to obtain a third cementing parameter, and generating verifiable information corresponding to the third cementing parameter;

[0114] S2: The cementing quality evaluation result of the target well and the verifiable information are stored in a preset blockchain, and the cementing parameters of the target well except the third cementing parameter are stored in a preset storage device, so that when a traceability request for the cementing quality evaluation process of the target well is received, backtracking processing can be performed according to the preset blockchain and the data stored in the preset storage device.

[0115] Specifically, the cementing quality evaluation results and verifiable information of the target well are stored in a preset blockchain, which provides guarantees for the security and integrity of the cementing data. The blockchain has the characteristics of being tamper-proof and fully traceable, which can effectively prevent data from being maliciously tampered with. For example, the cementing quality evaluation results of the target well include key indicators such as cement slurry viscosity and well wall adhesion strength. At the same time, in order to enhance the integrity of the data, the system also uploads on-site measurement records, test reports and other information to the blockchain. At the same time, to improve data access efficiency, the system stores the remaining cementing parameters of the target well (such as density, displacement, annular pressure, etc.) in a preset storage device. When there is a need to trace the cementing quality evaluation process of the target well in the future, the key result data recorded in the blockchain and the detailed parameter data in the storage device can be combined to achieve a comprehensive retrospective analysis.

[0116] Furthermore, the introduction of smart contracts can further enhance the automation and security of the backtracking process. Smart contracts can be automatically triggered when specific conditions are met. For example, upon receiving a cementing quality inquiry, the smart contract automatically verifies the cementing quality evaluation results stored in the blockchain and cross-references the detailed parameter data stored in the storage device to verify that the cementing parameters meet the preset operating conditions. The introduction of smart contracts makes the backtracking process more transparent, objective, and efficient.

[0117] The above examples fully leverage the tamper-proof and traceable nature of blockchain to ensure the security and reliability of cementing quality evaluation data. Data integrity is protected even during data transmission and multi-party collaboration. Furthermore, the introduction of smart contracts automates and intelligently processes data traceability, reducing manual intervention and improving traceability efficiency. This provides strong technical support for rapid response and scientific analysis of cementing quality disputes.

[0118] In some embodiments, the method may further include the following when implemented:

[0119] The preset evaluation system includes scoring criteria corresponding to the cementing parameters and preset weights corresponding to each cementing parameter. The preset weights corresponding to the cementing parameters are determined based on historical cementing parameters of historical wells and historical cementing quality evaluation results.

[0120] Specifically, within the aforementioned pre-defined evaluation system, the scoring criteria and the pre-defined weights for each cementing parameter together form the core basis for cementing quality assessment. The scoring criteria measure the ideal range or acceptance criteria for each cementing parameter, while the pre-defined weights determine the degree of impact each parameter has on overall cementing quality. These weights are determined based on an in-depth analysis of historical well data, combining a large amount of historical cementing parameters and cementing quality assessment results, and utilizing statistical analysis or machine learning models to determine the importance of each parameter.

[0121] For example, in the cementing quality assessment of an oil well, cement slurry density, displacement, and annular pressure are important cementing parameters. Analysis of historical well data reveals that cement slurry density has a greater impact on cementing quality, so it is assigned a higher weight (e.g., 0.35); whereas annular pressure has a relatively smaller impact on cementing quality, so it is assigned a lower weight (e.g., 0.15). Assuming the ideal cement slurry density range is 1.8 to 2.0 g / cm 3 , if the measured density is 1.85g / cm 3 , then the parameter can get a score close to full score; if the ideal value of the annular pressure is 8-12MPa, but the measured value is 7MPa, then the parameter may only get a partial score.

[0122] Finally, the comprehensive cementing quality score can be obtained based on the weighted calculation of the scores of each parameter and its weight, providing a scientific and quantitative reference basis for the overall quality assessment.

[0123] As can be seen from the above, the embodiments of this specification provide a graph-structured cementing quality control method, which obtains cementing parameters of a target well and determines a cementing quality assessment result of the target well based on the cementing parameters of the target well and a preset assessment system; determines whether the cementing quality of the target well meets preset quality requirements based on the cementing quality assessment result of the target well; and obtains target graph structure data corresponding to the target well if the cementing quality of the target well does not meet the preset quality requirements; wherein the target graph structure data is constructed based on first cementing parameters of multiple first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node associations determined based on associations between the first cementing parameters; determines a target cementing parameter among the cementing parameters of the target well based on the cementing quality assessment result of the target well, and determines a second cementing parameter having a preset association with the target cementing parameter based on the target graph structure data; and adjusts cementing operation parameters of the target well based on the target cementing parameter and the second cementing parameter. This approach first uses the target well's cementing parameters and a pre-set evaluation system to conduct a preliminary assessment of cementing quality. If the target well's cementing quality falls short of pre-set requirements, the pre-set generation model is further invoked, using the graph structure data constructed from multiple first wells as a reference. By matching the target well's cementing parameters, the target cementing parameters and the secondary cementing parameters with which they have a pre-set correlation are extracted, allowing for targeted optimization and adjustment of cementing operation parameters. Compared to existing methods, this approach achieves deep data fusion and correlation analysis, overcoming the shortcomings of traditional cementing operation control, such as insufficient data utilization, inaccurate quality assessments, and a lack of parameter adjustment strategies.

[0124] See Figure 2 As shown, an embodiment of this specification also provides a specific electronic device, wherein the electronic device includes a network communication port 201, a processor 202 and a memory 203, and the above structures are connected through internal cables so that each structure can perform specific data interaction.

[0125] The network communication port 201 can be used to obtain cementing parameters of a target well, and determine a cementing quality evaluation result of the target well according to the cementing parameters of the target well and a preset evaluation system.

[0126] The processor 202 can be specifically used to determine whether the cementing quality of the target well meets the preset quality requirements based on the cementing quality assessment result of the target well; if the cementing quality of the target well does not meet the preset quality requirements, obtain target graph structure data corresponding to the target well; wherein the target graph structure data is constructed based on the first cementing parameters of multiple first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node association relationships determined based on the association relationships between the first cementing parameters; based on the cementing quality assessment result of the target well, determine the target cementing parameter among the cementing parameters of the target well, and based on the target graph structure data, determine the second cementing parameter that has a preset association relationship with the target cementing parameter; and based on the target cementing parameter and the second cementing parameter, adjust the cementing operation parameters of the target well.

[0127] The memory 203 may be specifically used to store corresponding instruction programs.

[0128] Based on the above method, the relevant structural performance of electronic equipment can be effectively utilized, the data processing speed of electronic equipment can be improved, and the cementing quality control method based on the graph structure can be efficiently implemented.

[0129] In this embodiment, the network communication port 201 can be a virtual port that is bound to different communication protocols, thereby being capable of sending or receiving different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. Furthermore, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM or CDMA; it can also be a Wi-Fi chip; or it can be a Bluetooth chip.

[0130] In this embodiment, the processor 202 may be implemented in any suitable manner. For example, the processor may take the form of a microprocessor or a processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, an embedded microcontroller, etc. This specification is not intended to limit this.

[0131] In this embodiment, the memory 203 may include layers. In a digital system, anything that can store binary data can be a memory. In an integrated circuit, a circuit with a storage function that has no physical form is also called a memory, such as RAM, FIFO, etc. In a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0132] An embodiment of the present specification further provides a computer-readable storage medium based on the graph-structure-based cementing quality control method, which obtains cementing parameters of a target well and determines a cementing quality assessment result of the target well based on the cementing parameters of the target well and a preset assessment system; determines whether the cementing quality of the target well meets preset quality requirements based on the cementing quality assessment result of the target well; and obtains target graph structure data corresponding to the target well if the cementing quality of the target well does not meet the preset quality requirements; wherein the target graph structure data is constructed based on first cementing parameters of multiple first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node associations determined based on associations between the first cementing parameters; determines a target cementing parameter among the cementing parameters of the target well based on the cementing quality assessment result of the target well, and determines a second cementing parameter having a preset association with the target cementing parameter based on the target graph structure data; and adjusts cementing operation parameters of the target well based on the target cementing parameter and the second cementing parameter.

[0133] In this embodiment, the storage medium includes, but is not limited to, random access memory (RAM), read-only memory (ROM), cache, hard disk drive (HDD), or memory card. The memory can be used to store computer program instructions. The network communication unit can be an interface configured in accordance with the standards specified by the communication protocol for network connection communication.

[0134] In this embodiment, the functions and effects specifically implemented by the program instructions stored in the computer-readable storage medium can be explained in comparison with other implementations and will not be repeated here.

[0135] See Figure 3 At the software level, the embodiments of this specification also provide a cementing quality control device based on a graph structure, which may specifically include the following structural modules:

[0136] An evaluation result determination module 301 is used to obtain cementing parameters of a target well and determine a cementing quality evaluation result of the target well based on the cementing parameters of the target well and a preset evaluation system;

[0137] The demand judgment module 302 is used to judge whether the cementing quality of the target well meets the preset quality requirements based on the cementing quality evaluation result of the target well;

[0138] A graph structure generation module 303 is configured to obtain target graph structure data corresponding to the target well if the cementing quality of the target well does not meet preset quality requirements; wherein the target graph structure data is constructed based on first cementing parameters of a plurality of first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node association relationships determined based on association relationships between the first cementing parameters;

[0139] a parameter determination module 304 for determining a target cementing parameter among the cementing parameters of the target well based on the cementing quality assessment result of the target well, and determining a second cementing parameter having a preset correlation relationship with the target cementing parameter based on the target graph structure data;

[0140] The parameter adjustment module 305 is configured to adjust the cementing operation parameters of the target well according to the target cementing parameters and the second cementing parameters.

[0141] In some embodiments, before the above-mentioned graph structure generation module 303, during specific implementation, the first cementing parameters and the first cementing quality assessment results of the plurality of first wells are obtained; a feature vector determination module is used to use the first module of the preset generation model to perform semantic extraction processing on the first cementing parameters based on the first cementing parameters to obtain a first feature vector, and use the second module of the preset generation model to perform knowledge extraction processing on the first cementing parameters based on the first cementing parameters to obtain a second feature vector; and use the third module of the preset generation model to determine the target graph structure data according to the first feature vector, the second feature vector and the first cementing quality assessment result.

[0142] In some embodiments, the formation characteristics of the multiple first wells are different. When the above-mentioned feature vector determination module is implemented, the fourth module of the preset generation model is used to map the first cementing parameters to the preset space to obtain the fourth feature vector; the first module of the preset generation model is used to perform semantic extraction processing on the fourth feature vector to obtain the first feature vector; and the second module of the preset generation model is used to perform knowledge extraction processing on the fourth feature vector to obtain the second feature vector.

[0143] In some embodiments, the parameter adjustment module 305 is specifically implemented to perform multi-physical field coupling simulation processing according to the cementing parameters of the target well, and detect whether there are abnormal working conditions in the simulation processing; when it is detected that there are abnormal working conditions in the simulation processing, the abnormal working conditions, the target cementing parameters and the second cementing parameters are matched and verified; when the abnormal working conditions, the target cementing parameters and the second cementing parameters are successfully matched, the cementing operation parameters of the target well are adjusted according to the cementing parameters in the target cementing parameters and the second cementing parameters that match the abnormal working conditions.

[0144] In some embodiments, during specific implementation, an update request for the target graph structure data is received; in response to the update request, a second cementing parameter and a second cementing quality evaluation result of a second well corresponding to the update request are determined; the similarity between the second cementing parameter and the first cementing parameter is determined, and based on the similarity, candidate cementing parameters in the second cementing parameter are determined; and based on the candidate cementing parameters and the second cementing quality evaluation result, the target graph structure data is updated.

[0145] In some embodiments, during specific implementation, the cementing parameters of the target well are screened to obtain a third cementing parameter, and verifiable information corresponding to the third cementing parameter is generated; the cementing quality evaluation result of the target well and the verifiable information are stored in a preset blockchain, and the cementing parameters of the target well other than the third cementing parameter are stored in a preset storage device, so that when a traceability request for the cementing quality evaluation process of the target well is received, backtracking processing is performed according to the preset blockchain and the data stored in the preset storage device.

[0146] In some embodiments, during specific implementation, the preset evaluation system includes scoring criteria corresponding to the cementing parameters and preset weights corresponding to each of the cementing parameters. The preset weights corresponding to the cementing parameters are determined based on historical cementing parameters of historical wells and historical cementing quality evaluation results.

[0147] It should be noted that the units, devices or modules described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. For the convenience of description, the above devices are described in terms of functions and are divided into various modules and described separately. Of course, when implementing this specification, the functions of each module can be implemented in the same software and / or hardware, or the modules that implement the same function can be implemented by a combination of sub-modules or sub-units. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods, such as units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0148] As can be seen from the above, the graph-based cementing quality control device provided in the embodiments of this specification first performs a preliminary assessment of cementing quality using the cementing parameters of the target well and a preset evaluation system. If the target well's cementing quality fails to meet preset requirements, the device further invokes a preset generation model, using the graph structure data constructed from multiple first wells as a reference. By matching the target well's cementing parameters, the device extracts the target cementing parameters and the second cementing parameters that have a preset correlation with them, thereby performing targeted optimization and adjustment of the cementing operation parameters. Compared to existing methods, this method achieves deep data fusion and correlation analysis, overcoming the shortcomings of traditional cementing operation control, such as insufficient data utilization, inaccurate quality assessment, and the lack of parameter adjustment strategies.

[0149] In a specific scenario example, a graph-based cementing quality control method and device provided in this specification can be applied. First, the cementing quality is preliminarily evaluated using the cementing parameters of the target well and a preset evaluation system. When it is found that the cementing quality of the target well does not meet the preset requirements, the preset generation model is further called, and the graph structure data constructed from multiple first wells is used as a reference. By matching the cementing parameters of the target well, the target cementing parameters and the second cementing parameters that have a preset correlation relationship with them are extracted, and then the cementing operation parameters are optimized and adjusted in a targeted manner. Compared with existing methods, deep fusion and correlation analysis of data are achieved, overcoming the shortcomings of insufficient data utilization, inaccurate quality evaluation, and lack of parameter adjustment strategies in traditional cementing operation control. The specific implementation process may include the following. See. Figure 4 shown.

[0150] S1: Data acquisition and preprocessing (drilling-testing-cementing multi-source data acquisition and processing to form a high-quality cementing quality assessment data set).

[0151] Collect three major categories of data: drilling, logging, and cementing. This includes, but is not limited to, wellbore parameters (such as wellbore dimensions and inclination) acquired during drilling, formation characteristics (such as resistivity and acoustic transit time) acquired through logging, and cementing parameters (such as cement slurry formulation and displacement volume). The collected data is pre-processed through cleaning, denoising, and normalization to improve data quality.

[0152] Specifically, consider the cementing operation of an actual oil well in an oil field. Drilling parameters are obtained from the well's drilling logs, detailed formation characteristics are extracted from well logging data, and information such as cement slurry formulation and construction time are collected from cementing operation records. This data is cleaned to remove outliers and noise, and then normalized to a uniform dimension and range.

[0153] S2: Construct a cementing knowledge graph (extract the mapping relationship between each feature and cementing quality, and construct a cementing knowledge graph).

[0154] See Figure 5 As shown in the figure, based on intelligent algorithms and combined with domain knowledge, preprocessed data is analyzed and mined. Different types of data are associated to construct a cementing knowledge graph. In the knowledge graph, nodes represent different entities (such as formation characteristics and cementing parameters), and edges represent relationships between entities (such as the impact of certain formation characteristics on cementing quality). In this way, the impact of different characteristics on cementing quality is comprehensively mapped.

[0155] Specifically, deep learning algorithms were used to analyze preprocessed data. For example, a correlation was discovered between formation permeability and cement slurry water loss rate. "Formation permeability" and "cement slurry water loss rate" were treated as entities, and a relationship between them was established, which was then integrated into the cementing knowledge graph. Through a series of data mining and correlation analysis, a complete cementing knowledge graph was constructed, and a transfer learning update mechanism was introduced to adapt to the operational conditions of different blocks.

[0156] S3: Introduce a knowledge graph update mechanism based on transfer learning (introduce a knowledge graph update mechanism based on transfer learning to realize cross-block knowledge transfer).

[0157] A BERT-TransE hybrid model was used, along with a pre-trained language model to extract semantic features from multi-source data. Cross-block knowledge transfer was achieved through a relational projection matrix. A migration evaluation metric was established, with a cosine similarity threshold set above 0.85 to initiate knowledge fusion. A GNN (Graph Neural Network) was applied to update node embeddings, and an improved Node2Vec algorithm was used to calculate node similarity. A test set containing over 500 cementing scenarios was constructed, and the inference accuracy of the migrated graph was verified.

[0158] Specifically, BERT's pre-trained language model capabilities are used to extract semantic features from multi-source data, effectively capturing deep semantic information from complex text, parameter records, and engineering data. Combined with the TransE model, cross-block knowledge transfer is achieved through a relational projection matrix. Specifically, the BERT model first extracts features from the raw cementing data, such as multi-dimensional parameters such as well depth, formation pressure, temperature, and wellbore structure, to form high-dimensional semantic vectors. Subsequently, the TransE model constructs a cementing knowledge graph based on the inherent connections between these feature vectors, forming a graph structure in which nodes represent parameters and edges represent parameter relationships.

[0159] During the knowledge transfer process, the cosine similarity is used to assess the similarity of well parameters in the new block with those in the existing knowledge graph. If the similarity of parameters such as depth, formation pressure, and temperature of a well in the new block exceeds 0.85 (with an adjustable threshold) with those of a well in the existing graph, the knowledge fusion mechanism is triggered. At this point, cementing experience from the original graph (such as slurry ratio, displacement control, and thickening time) can be transferred and applied to wells in the new block, achieving experience reuse.

[0160] For example, suppose that in the knowledge graph of Block A of an oil field, a well with a depth of 2800m, a formation pressure of 30MPa, and a temperature of 90°C uses a density of 1.85g / cm 3 When a well in new Block B is discovered with parameters such as a depth of 2850m, a formation pressure of 29MPa, and a temperature of 92°C, and its similarity calculation result is 0.88 (exceeding the threshold of 0.85), the system can automatically recommend relevant cementing parameters from Block A and further optimize them for application in Block B.

[0161] In addition, we use GNN (Graph Neural Network) to embed and update nodes in the graph, ensuring that the features of parameter nodes remain effectively represented under dynamic changes. To improve the matching accuracy of node similarity, we introduce an improved Node2Vec algorithm, further optimizing the node wandering strategy and feature aggregation to achieve more accurate similarity calculation.

[0162] This method constructed a test set containing over 500 cementing scenarios. Experimental results showed that the graphs after knowledge transfer significantly improved inference accuracy, further validating the effectiveness of this method in knowledge reuse and optimization in cementing projects. This migration mechanism not only improves cementing efficiency in new wells but also reduces the trial-and-error cost of parameter tuning, providing a more intelligent, data-driven solution for cementing projects.

[0163] S4: Introducing a multi-condition adaptive KPI dynamic weighting algorithm (establishing a dynamically optimized cementing operation KPI system to quantitatively evaluate operation quality).

[0164] A dynamic weight calculation model was constructed (input layer: feature vectors of drilling, logging, and cementing data; processing layer: fuzzy rough set theory was used to simplify features and eliminate redundant parameter interference; output layer: an improved NSGA-II algorithm was applied for multi-objective optimization to generate the Pareto frontier solution set). An adaptive adjustment mechanism was introduced to establish a weight influence factor matrix. Details are shown in Table 2.

[0165] Table 2

[0166] Working condition type Quality Weight Efficiency weight Cost Weight Conventional vertical well 0.45 0.35 0.20 Highly deviated well 0.60 0.25 0.15 Ultra-deep wells 0.50 0.30 0.20

[0167] At the same time, it is equipped with an online learning module, which automatically updates the weight model after completing 10 well operations, and the loss function adopts the hybrid form of MAE+MSE.

[0168] Specifically, key indicators of cementing quality are determined based on the well's geological conditions and production requirements, such as ensuring the average compressive strength of the cement sheath is at least a certain value and that cement slurry displacement efficiency reaches a certain percentage. These indicators are incorporated into the KPI system, with corresponding weights assigned to each. Combined with a KPI dynamic weighting algorithm that adapts to multiple operating conditions, the optimal KPI evaluation criteria are tailored for each well.

[0169] S5: Introducing a cementing digital twin system (combining the cementing digital twin system with a graded response strategy for abnormal working conditions to monitor downhole conditions in real time).

[0170] A cementing digital twin was constructed through multi-physics coupled modeling, including a cement slurry rheology model (using the Herschel-Bulkley constitutive equation), a wellbore-formation coupled model (using the finite element method to solve the temperature-stress field distribution), and a cementing interface simulation (using molecular dynamics to simulate the interface bonding strength). This system is equipped with a real-time data-driven system, including data interfaces supporting industrial protocols such as Modbus and OPC UA. The synchronization mechanism includes a clock synchronization module with a 250ms refresh cycle and a verification indicator system. Details are shown in Table 3.

[0171] Table 3

[0172] Verification Dimension Detection method Eligibility criteria Mechanical properties Triaxial stress testing Compressive strength ≥14MPa Sealing Airtightness test Leakage rate <0.01mL / min Durability Temperature cycle test (-20~150℃) Performance degradation ≤5% / 100 times

[0173] Abnormal Condition Grading Response Strategy: A three-level classification model (Level 1 (routine abnormality): parameter deviation <10%, automatic adjustment; Level 2 (serious abnormality): deviation 10-30%, manual confirmation; Level 3 (accident level): deviation >30%, expert consultation) is established. This is coupled with an intelligent diagnosis engine (using an XGBoost + LightGBM hybrid model to rank feature importance) to control response timeliness. See Table 4 for details.

[0174] Table 4

[0175] level Response time limit Disposal completion rate Level 1 ≤30s ＞95％ Level 2 ≤5min ＞90％ Level 3 ≤30min ＞85％

[0176] Specifically, the digital twin system collects dynamic parameters such as downhole pressure, temperature, and cement slurry return rate in real time, combines geological models with historical data to build a virtual twin, and simultaneously simulates cement slurry flow, cementation process, and annular sealing status (such as predicting cement slurry leakage path or gas channeling risk); when abnormal working conditions are monitored (such as sudden drop in pump pressure, abnormal return volume), the system automatically triggers graded intervention based on the preset multi-level response strategy - mild deviations (such as a 5% decrease in displacement efficiency) start a yellow warning and fine-tune the pump speed parameters, moderate risks (such as leakage rate of 2m 3 / h) switches to orange mode and injects plugging fiber slurry. In the event of a serious fault (such as wellhead pressure exceeding the threshold), the red shutdown command is executed and the emergency casing seal is activated. At the same time, the effectiveness of the optimization measures is verified in real time through the twin, forming a "perception-simulation-decision-verification" closed loop, ultimately achieving a transition from passive response to active control in cementing operations.

[0177] S6: Introduce a blockchain-based quality traceability module (build a blockchain-based quality traceability module and propose targeted optimization strategies).

[0178] Based on the quality assessment results, the knowledge graph is used for reasoning. The causes of quality risks are analyzed, and combined with historical data and experience, targeted optimization strategies are proposed, such as adjusting the cement slurry formula and optimizing the displacement process, to achieve real-time optimization control of cementing operations.

[0179] The blockchain architecture utilizes Hyperledger Fabric consortium blockchain with four consensus nodes. The smart contract includes 12 key quality node recording rules. The data storage mechanism utilizes the SM3 national encryption algorithm and is connected to the National Time Service Center's NTP server.

[0180] Based on the above embodiments, full use is made of multi-source data from drilling, logging, and cementing to construct a cementing knowledge graph based on transfer learning and self-update, comprehensively mapping the influence of different features on cementing quality, and building a scientific KPI system through a multi-condition adaptive KPI dynamic weight algorithm to achieve accurate evaluation and effective control of cementing operation quality. In combination with a cementing quality digital twin verification system and an abnormal working condition graded response strategy, downhole status is monitored and perceived in real time. Finally, through a blockchain-based quality traceability module, combined with historical data and experience, targeted optimization strategies are proposed to achieve real-time optimization control of cementing operations, thereby overcoming the shortcomings of existing technologies.

[0181] In some embodiments, during cementing operations, data is collected in real time and compared with KPI indicators. For example, by monitoring the actual displacement volume and displacement time of the cement slurry, combined with the relationship between displacement efficiency and cementing quality in the knowledge graph, the current displacement efficiency is evaluated to see if it meets the requirements. If the displacement efficiency is found to be lower than the set value, the knowledge graph is used to analyze possible causes, such as poor mud performance or insufficient displacement equipment pressure. Based on the quality assessment results, if the low displacement efficiency is found to be due to mud performance issues, solutions to similar problems in historical data are obtained from the knowledge graph, such as adding a chemical additive to improve mud performance. The operator adjusts the operation parameters based on the optimization strategy, achieving real-time optimization of the cementing operation.

[0182] In some embodiments, see Figure 6 As shown in Figure 1, a multi-layered processor-based system architecture consists of four main layers: The data layer, which includes a data acquisition system and preprocessing modules, is responsible for collecting data from various sensors and devices and performing preliminary data processing to ensure the accuracy of subsequent model training and analysis. The algorithm and model layer includes a cementing knowledge graph and a KPI evaluation system. The former is used to build and store cementing-related expertise, while the latter is used to evaluate key performance indicators, helping to identify potential problems and optimize operations. The monitoring and control layer includes a digital twin system and a hierarchical response strategy. The digital twin system simulates the cementing process in real time using a virtual model, while the hierarchical response strategy takes appropriate control measures based on the severity of abnormal conditions. The reasoning and optimization layer includes a quality traceability module and optimization measure recommendation. The former is used to trace and analyze data and decisions during the cementing process, while the latter provides specific suggestions for operation optimization through reasoning models. The user interface, serving as the system's human-computer interaction interface, connects storage, input devices, displays, and network interfaces to ensure data visualization, interaction, and storage. The overall system implements a closed-loop process of data acquisition, analysis, monitoring and control, and optimization reasoning, improving the quality and safety of cementing operations.

[0183] Based on the above embodiments, the present invention aims to solve the problems of insufficient utilization of multi-source data, inaccurate cementing quality assessment, and lack of effective optimization strategies in the existing cementing operation control methods. By comprehensively utilizing three types of data, namely drilling, logging, and cementing, a knowledge graph for analyzing factors affecting cementing quality is constructed, and a knowledge graph update mechanism based on transfer learning is embedded to solve the problems that the traditional cementing knowledge graph relies on manually labeled data and is difficult to cope with complex and changeable downhole working conditions (such as high-temperature and high-pressure formations, salt-gypsum layers, and other special geological conditions), resulting in delayed knowledge updates and poor cross-scenario adaptability. A multi-condition adaptive KPI dynamic weight algorithm is proposed to make up for the shortcomings of the static weights determined by the traditional analytic hierarchy process (AHP) that cannot adapt to changes in working conditions such as different well depths and formation pressures, resulting in distorted KPI assessments. A cementing quality digital twin verification system is constructed to improve the situation that traditional cementing quality detection relies on post-logging and cannot achieve real-time quality prediction during the operation process. At the same time, a graded response strategy for abnormal working conditions is formulated to solve the problem that traditional abnormal processing uses a single threshold alarm and cannot distinguish the severity of the fault; a blockchain-based quality traceability module is built to improve the problems of traditional quality traceability such as easy data tampering and unclear responsibility definition, realize accurate assessment and real-time optimization control of cementing operation quality, and improve the stability and reliability of cementing quality.

[0184] Furthermore, it is possible to achieve: 1. Comprehensive and integrated analysis: By integrating the three major categories of data, namely drilling, logging, and cementing, it is possible to comprehensively consider various factors affecting cementing quality, overcome the limitations of existing technologies that rely solely on a single data source, and improve the accuracy and reliability of cementing quality assessment. 2. A scientific assessment system: Utilizing a multi-condition KPI dynamic weight algorithm to formulate a KPI system, the weights of each indicator are scientifically determined, making the assessment results more objective and accurate, and better reflecting the actual quality of cementing operations. 3. Visualization and interpretability: The construction of a cementing knowledge graph can intuitively display the relationship between different factors and cementing quality, making it easier for technical personnel to understand and analyze, and providing clear guidance for the optimization of cementing operations. 4. Intelligent optimization decision-making: Based on the reasoning mechanism of the knowledge graph and quality assessment results, it can automatically give targeted optimization suggestions, realize intelligent control and optimization of cementing operations, and improve the efficiency and quality of cementing operations.

[0185] Although this specification provides the method operation steps as described in the embodiments or flow charts, more or fewer operation steps may be included based on conventional or non-creative means. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only execution order. When the device or client product in practice is executed, it can be executed in sequence or in parallel according to the method shown in the embodiments or the drawings (for example, a parallel processor or a multi-threaded processing environment, or even a distributed data processing environment). The term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, product or device including a series of elements includes not only those elements, but also includes other elements that are not explicitly listed, or also includes elements inherent to such process, method, product or device. In the absence of more restrictions, it is not excluded that there are other identical or equivalent elements in the process, method, product or device including the elements. Words such as first and second are used to represent names and do not represent any particular order.

[0186] Those skilled in the art will also appreciate that, in addition to implementing the controller in pure computer-readable program code, it is entirely possible to implement the same functionality by logically programming the method steps in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, embedded microcontrollers, and the like. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be considered structures within the hardware component. Alternatively, the devices for implementing various functions can be considered both software modules implementing the method and structures within the hardware component.

[0187] Through the description of the above embodiments, it can be seen that those skilled in the art can clearly understand that this specification can be implemented by means of software plus the necessary general hardware platform. Based on this understanding, the technical solution of this specification can essentially be embodied in the form of a software product. This computer software product can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a mobile terminal, a server, or a network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of this specification.

[0188] Although the present specification has been described through embodiments, those skilled in the art will appreciate that there are many modifications and variations to the present specification without departing from the spirit of the present specification. It is intended that the appended claims include these modifications and variations without departing from the spirit of the present specification.

Claims

1. A cementing quality control method based on graph structure, characterized in that: include: Acquiring cementing parameters of a target well, and determining a cementing quality evaluation result of the target well based on the cementing parameters of the target well and a preset evaluation system; Determining whether the cementing quality of the target well meets preset quality requirements based on the cementing quality assessment result of the target well; When the cementing quality of the target well does not meet a preset quality requirement, target graph structure data corresponding to the target well is obtained; wherein the target graph structure data is constructed based on first cementing parameters of a plurality of first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node association relationships determined based on association relationships between the first cementing parameters; determining a target cementing parameter among the cementing parameters of the target well according to a cementing quality assessment result of the target well, and determining a second cementing parameter having a preset correlation relationship with the target cementing parameter according to the target graph structure data; The cementing operation parameters of the target well are adjusted according to the target cementing parameters and the second cementing parameters.

2. The method according to claim 1, characterized in that Before acquiring the target graph structure data corresponding to the target well, the method further includes: Obtaining first cementing parameters and first cementing quality assessment results of a plurality of the first wells; Using the first module of the preset generation model, based on the first cementing parameter, performing semantic extraction processing on the first cementing parameter to obtain a first feature vector; and using the second module of the preset generation model, based on the first cementing parameter, performing knowledge extraction processing on the first cementing parameter to obtain a second feature vector; The target graph structure data is determined by utilizing the third module of the preset generation model according to the first eigenvector, the second eigenvector, and the first cementing quality evaluation result.

3. The method according to claim 2, characterized in that The formation characteristics of the multiple first wells are different, the first module of the preset generation model is used to perform semantic extraction processing on the first cementing parameter based on the first cementing parameter to obtain a first feature vector, and the second module of the preset generation model is used to perform knowledge extraction processing on the first cementing parameter based on the first cementing parameter to obtain a second feature vector, including: Mapping the first cementing parameter to a preset space using a fourth module of the preset generation model to obtain a fourth eigenvector; The first module of the preset generation model is used to perform semantic extraction processing on the fourth feature vector to obtain the first feature vector, and the second module of the preset generation model is used to perform knowledge extraction processing on the fourth feature vector to obtain the second feature vector.

4. The method according to claim 1, wherein The adjusting the cementing operation parameters of the target well according to the target cementing parameters and the second cementing parameters includes: Performing multi-physics field coupling simulation processing according to the cementing parameters of the target well, and detecting whether there are abnormal working conditions during the simulation processing; When an abnormal working condition is detected in the simulation process, performing matching verification processing on the abnormal working condition, the target cementing parameter and the second cementing parameter; When the abnormal working condition, the target cementing parameters and the second cementing parameters successfully match, the cementing operation parameters of the target well are adjusted according to the cementing parameters among the target cementing parameters and the second cementing parameters that match the abnormal working condition.

5. The method according to claim 1, wherein The method further comprises: receiving an update request for the target graph structure data; In response to the update request, determining a second cementing parameter and a second cementing quality evaluation result of a second well corresponding to the update request; determining a similarity between the second cementing parameter and the first cementing parameter, and determining a candidate cementing parameter in the second cementing parameter based on the similarity; The target graph structure data is updated according to the candidate cementing parameters and the second cementing quality evaluation result.

6. The method according to claim 1, characterized in that The method further comprises: Screening the cementing parameters of the target well to obtain a third cementing parameter, and generating verifiable information corresponding to the third cementing parameter; The cementing quality evaluation result of the target well and the verifiable information are stored in a preset blockchain, and the cementing parameters of the target well except the third cementing parameter are stored in a preset storage device, so that when a traceability request for the cementing quality evaluation process of the target well is received, backtracking processing is performed according to the preset blockchain and the data stored in the preset storage device.

7. The method according to claim 1, characterized in that The preset evaluation system includes scoring criteria corresponding to the cementing parameters and preset weights corresponding to each cementing parameter. The preset weights corresponding to the cementing parameters are determined based on historical cementing parameters of historical wells and historical cementing quality evaluation results.

8. A cementing quality control device based on graph structure, characterized in that: include: An evaluation result determination module is used to obtain cementing parameters of a target well and determine a cementing quality evaluation result of the target well based on the cementing parameters of the target well and a preset evaluation system; A demand judgment module is used to judge whether the cementing quality of the target well meets the preset quality requirements based on the cementing quality evaluation result of the target well; a graph structure generation module, configured to obtain target graph structure data corresponding to the target well if the cementing quality of the target well does not meet preset quality requirements; wherein the target graph structure data is constructed based on first cementing parameters of a plurality of first wells using a preset generation model, and the target graph structure data includes nodes corresponding to the first cementing parameters and node association relationships determined based on association relationships between the first cementing parameters; a parameter determination module, configured to determine a target cementing parameter among the cementing parameters of the target well according to a cementing quality assessment result of the target well, and to determine a second cementing parameter having a preset correlation relationship with the target cementing parameter according to the target graph structure data; The parameter adjustment module is used to adjust the cementing operation parameters of the target well according to the target cementing parameters and the second cementing parameters.

9. An electronic device, characterized in that: The invention comprises a processor and a memory for storing processor-executable instructions, wherein when the processor executes the instructions, the steps of the graph-structure-based cementing quality control method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the cementing quality control method based on graph structure according to any one of claims 1 to 7 are implemented.

Images