An oilfield operation simulation scene variable mapping and parameter completion method and system

Abstract

The application relates to the field of oil and gas engineering simulation and digital twinning, and particularly discloses an oilfield operation simulation scene variable mapping and parameter completion method and system, which comprises the following steps: obtaining a multi-source heterogeneous engineering scheme document, extracting explicit structured parameters and implicit inferred parameters and labeling confidence, fusing the parameters into high-confidence parameters and low-confidence parameters, generating a missing identifier and extracting a working condition vector and a time sequence identifier; screening an effective mapping rule to perform differential mapping, performing multi-strategy completion of missing parameters in time sequence coupling, and obtaining physically self-consistent simulation scene parameter configuration through multi-scale physical constraint network rollback verification; inputting the parameter configuration into an inverse kinematics solver to generate a trajectory key frame, adaptively adjusting and rendering to generate a simulation scene, and performing multiple verifications and feedback corrections; and the system comprises functionally matched modules. The application realizes end-to-end automatic generation of unstructured engineering schemes into executable simulation scenes.

Description

Technical Field

[0001] This invention relates to the field of oil and gas engineering simulation and digital twin technology, and in particular to a method and system for variable mapping and parameter completion in oilfield operation simulation scenarios. Background Technology

[0002] Oilfield operations are a core component of oil and gas field development, characterized by the inability to visualize underground conditions, complex operational processes, and high operational risks. In recent years, digital twin and simulation technologies have been widely applied in oilfield operations, enabling purposes such as pre-simulation of construction plans, personnel training, and risk warning by recreating operational processes in a virtual environment. However, current technologies for generating simulation scenarios from engineering plans still suffer from the following significant shortcomings.

[0003] Engineering solutions typically exist in the form of multi-source heterogeneous sub-documents. Existing technologies only extract explicit parameters from single documents, failing to identify implicit parameters indirectly characterized by equipment models and process standards. Furthermore, they lack automatic verification mechanisms for version conflicts and parameter drift between multi-source documents, resulting in insufficient completeness of structured parameter extraction and unquantifiable confidence levels. Existing technologies generally use static mapping tables to convert engineering parameters to simulation parameters, which cannot handle nonlinear distortions of the same parameter under different work types and operating conditions. In particular, parameter jumps caused by hard switching of mapping rules are prone to occur in the transition interval between adjacent operating condition categories. At the same time, missing parameters are often filled independently or with default values, lacking physical consistency verification. Completion errors are easily amplified along the parameter coupling chain, causing physical and logical conflicts in the simulation scene. The lack of end-to-end automated compilation capabilities from unstructured engineering solutions to 3D executable simulation scenarios means that existing processes heavily rely on manual intervention to configure parameters and edit animation curves, resulting in semantic gaps between the engineering parameter space, the simulation scenario parameter space, and the 3D model joint space. Furthermore, existing technologies do not consider the constraints of obstacles in the work space and the resolution of mechanical redundancy degrees of freedom, leading to the risk of unexecutable trajectories. In addition, there is a lack of mechanisms to adaptively adjust the scene information density based on the characteristics of the work stage and the complexity of parameters, which can easily cause cognitive overload for trainees.

[0004] In summary, there is an urgent need for a method to generate oilfield operation simulation scenarios that can achieve collaborative parsing of multi-source documents and inference of implicit parameters, adaptive dynamic mapping of working conditions, completion of physically consistent missing parameters, and end-to-end automated generation. Summary of the Invention

[0005] The purpose of this invention is to address the problems in the existing technology, such as high reliance on manual intervention when generating unstructured oilfield engineering schemes into executable simulation scenarios, inability of static mapping to adapt across operating conditions, physical inaccuracies in missing parameter completion, and insufficient end-to-end automated compilation capabilities. The invention proposes a method and system for variable mapping and parameter completion in oilfield operation simulation scenarios.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] On the one hand, a method for variable mapping and parameter completion in oilfield operation simulation scenarios is provided, including the following specific steps:

[0008] S1. Obtain a set of oilfield operation engineering plan documents to be simulated. The set of engineering plan documents includes multi-source heterogeneous sub-documents of the same operation project. Based on the oilfield operation domain ontology dictionary and parameter pattern library, perform structured parsing and explicit parameter extraction on each sub-document to obtain explicit structured parameters, and label the extraction confidence level for each explicit structured parameter. Identify implicit parameters in each sub-document that are indirectly represented by equipment models, process standards, or operation procedures, and infer and complete the implicit parameters based on the domain knowledge base to obtain implicit inferred parameters. For parameters from different sub-documents that point to... Cross-document consistency checks are performed on parameters with the same engineering meaning to identify version conflict parameters and generate version consistency markers; the explicit structured parameters that pass the check are merged with implicit inferred parameters to form a structured engineering parameter set, and the structured engineering parameter set is divided into a high-confidence parameter subset and a low-confidence parameter subset according to the extraction confidence, the inferred confidence and the version consistency markers; the integrity of the structured engineering parameter set is checked to identify missing parameters and generate parameter missing identifiers, and at the same time, the working condition vector and the time sequence identifier of the work stage are extracted;

[0009] S2. Construct a dynamic mapping dictionary with a two-level index, comprising a first-level index based on job type and a second-level index based on working condition. A working condition transition smoothing mechanism is introduced in the second-level index. When the current working condition is in the transition interval between adjacent working condition categories, the dynamic activation factors of multiple mapping rules pointing to the same target simulation parameter are continuously interpolated. The effective weight of each mapping rule is determined based on the interpolated and fused activation factors. Based on the working condition vector and the job type, effective mapping rules whose dynamic activation factors meet threshold conditions are selected from the dynamic mapping dictionary. Combined with the version consistency marker, a precise mapping transformation is performed on the high-confidence parameter subset, and a conservative mapping transformation with a tolerable deviation range is performed on the low-confidence parameter subset, generating high-confidence simulation parameters and low-confidence simulation parameters respectively. A parameter confidence identifier is labeled for each parameter based on the confidence decay during the mapping transformation process. The high-confidence simulation parameters and the low-confidence simulation parameters are merged into a preliminary mapped simulation scenario parameter set.

[0010] For the missing parameters corresponding to the missing parameter identifiers, based on the position of the operation stage time sequence identifiers in the operation time sequence chain, a time-coupled completion strategy with physical transfer relationships to the parameters of upstream completed operation stages is invoked. Multi-strategy fusion completion is performed in the following priority order: physical constraint relationships first, similar working condition cases second, regression prediction third, and generative models as a fallback, to obtain the completed engineering parameters. The completed engineering parameters are then mapped to the simulation parameter space to generate completed simulation parameters. Using the high-confidence simulation parameters as anchor points, the parameters are mapped along the physical deterministic relationships in the oilfield operation parameter physical correlation database to the low-confidence simulation parameters. The true parameters and the completed simulation parameters are constrained and transmitted to calculate the constraint correction values ​​of the low-confidence simulation parameters and the completed simulation parameters. The simulation scene parameter set after the initial mapping is fused with the completed simulation parameters after constraint transmission, and a multi-scale physical constraint network is constructed based on the microscopic contact mechanics scale, the mesoscopic fluid dynamics scale, and the macroscopic structural mechanics scale. Through multi-scale physical consistency rollback verification and cross-scale conflict resolution iterative fine-tuning, the high-confidence simulation parameters remain locked during the iteration process, and only the parameters with confidence levels below a preset threshold are adjusted to obtain a physically self-consistent complete simulation scene parameter configuration.

[0011] S3. Based on the operation type and equipment specification parameters in the physically self-consistent complete simulation scene parameter configuration, adaptively select and instantiate a 3D model from the 3D model library; input the kinematic parameters in the physically self-consistent complete simulation scene parameter configuration into the inverse kinematics solver, introduce the operation space obstacle constraint and mechanical redundancy degree of freedom resolution strategy during the solution process, automatically generate the joint space motion trajectory keyframe sequence that meets the obstacle avoidance requirements, and obtain the animation curve data that can be executed by the 3D engine after interpolation and smoothing; link the animation curve data with the 3D model and perform scene rendering based on the environmental parameters and stage features in the physically self-consistent complete simulation scene parameter configuration to generate an oilfield operation simulation scene;

[0012] The simulation scenario of the oilfield operation is used to verify the integrity of the execution parameters, perform multi-scale physical consistency calculations and trajectory executability verification, and based on the verification results, the reputation of the mapping rules of the dynamic mapping dictionary, the model parameters of the multi-strategy fusion completion and the implicit parameter inference rules are updated.

[0013] On the other hand, an oilfield operation simulation scenario variable mapping and parameter completion system is provided to implement an oilfield operation simulation scenario variable mapping and parameter completion method, including:

[0014] The multi-source document collaborative parsing module is configured to acquire a set of engineering scheme documents, extract explicit structured parameters and implicit inferred parameters and label their confidence levels, perform cross-document consistency verification, identify version conflict parameters and generate version consistency markers, merge and divide them into high-confidence and low-confidence parameter subsets, generate missing markers, and extract working condition vectors and time series markers.

[0015] The dynamic mapping and physical self-consistent completion module is configured to combine version consistency markers with working condition vectors to filter effective mapping rules, map two types of parameters as high-confidence simulation parameters and low-confidence simulation parameters and mark the parameter confidence identifiers, complete missing parameters according to time sequence identifiers to obtain complete engineering parameters and map them into complete simulation parameters, and use high-confidence simulation parameters as anchor points to propagate to low-confidence simulation parameters and complete simulation parameters, and obtain physically self-consistent simulation scene parameter configurations through multi-scale physical constraint network rollback verification;

[0016] The scene linkage compilation and closed-loop verification module is configured to input the simulation scene parameter configuration into the inverse kinematics solver, introduce obstacle constraints and redundant degrees of freedom to generate trajectory keyframes and interpolate to obtain animation curve data, assemble the model and perform scene information density adaptive adjustment and rendering based on time sequence identifiers and parameter complexity to generate the simulation scene, perform multiple verifications and calculate the scene generation quality score, and provide feedback correction on the verification results.

[0017] The beneficial effects of this invention are as follows:

[0018] 1. By combining multi-source heterogeneous document collaborative parsing and implicit parameter inference with cross-document consistency verification and confidence leveling, the problem of multi-source document version conflict, implicit parameter missingness and low efficiency of manual extraction is effectively solved. It realizes the automatic conversion from unstructured scheme to structured and reliable parameter set, and significantly improves the completeness and reliability of parameter extraction.

[0019] 2. By using a two-layer context-aware dynamic mapping and a smoothing mechanism for the transition zone of operating conditions, combined with temporal coupling completion, anchoring constraint propagation, and multi-scale physical self-consistency verification, the problems of static mapping jumping across operating conditions, unreliable physical completion of missing parameters, and error propagation are effectively solved. The mapping rule scene adaptation and physical self-consistency of completion results are achieved, which significantly improves the accuracy of simulation parameter configuration.

[0020] 3. By using parameter-driven inverse kinematics linkage compilation and redundant degree of freedom obstacle avoidance resolution, combined with scene information density adaptive adjustment and multiple verification closed-loop feedback, the problems of low efficiency of manual animation editing, training scene cognitive overload and scene non-executability are effectively solved, realizing end-to-end automatic generation from engineering parameters to 3D executable simulation scene and system self-learning optimization. Attached Figure Description

[0021] Figure 1 This is an overall flowchart of the method described in this invention;

[0022] Figure 2 This is an architecture diagram of a multi-scale physical constraint network.

[0023] Figure 3 Architecture diagram for the model that completes missing parameters in physical-data fusion;

[0024] Figure 4 This is a framework diagram of the system described in this invention. Detailed Implementation

[0025] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0026] Example 1

[0027] Please see the appendix Figure 1-3 A method for variable mapping and parameter completion in oilfield operation simulation scenarios includes the following specific steps:

[0028] S1. Obtain the engineering solution document set, extract explicit structured parameters and implicit inferred parameters, and label their confidence levels; perform cross-document consistency verification, identify version conflict parameters and generate version consistency markers, merge and divide them into high-confidence and low-confidence parameter subsets, generate missing parameter markers, and extract working condition vectors and time series markers; specifically including:

[0029] S11. Multi-source heterogeneous document normalization and collaborative extraction of explicit and implicit parameters, specifically including:

[0030] S111. Obtain the set of oilfield operation engineering plan documents to be simulated through a file interface or network interface. This set includes multiple heterogeneous sub-documents from various sources, such as the original engineering design, construction plan, equipment list, and process specification reference documents for the same operation project. The original formats of the sub-documents include PDF, Word, scanned image, and structured text formats.

[0031] Each sub-document undergoes format normalization. A document parsing engine converts PDF and Word files into plain text streams, and optical character recognition technology is used to convert scanned images into editable text. During normalization, the original document's paragraph structure, table layout, and heading hierarchy information are preserved.

[0032] Version number information is extracted from the header, footer, or cover area of ​​each sub-document. Version number formats include alphanumeric combinations such as V1.0 or 2024-A, and date codes such as 20240815. When extracting the compilation date, common date formats such as YYYY-MM-DD or YYYY year MM month DD day are recognized. When extracting applicable well numbers, well location identifiers are matched according to the oilfield well number coding rules. Well numbers typically consist of block codes, well type codes, and sequence numbers. Finally, a document metadata record containing the version number, compilation date, and applicable well number is generated.

[0033] S112. Based on a pre-built oilfield operation domain ontology dictionary and parameter pattern library, structured parsing and explicit parameter extraction are performed on each sub-document. The domain ontology dictionary covers professional terms and their synonyms and English abbreviations for drilling, fracturing, workover, and completion operations. Among them, the term WOB corresponds to drilling pressure, and the term SPM corresponds to pump stroke count. The parameter pattern library defines the name regular expression patterns, value format rules, and context trigger words for various parameters.

[0034] During the parsing process, the normalized text is first segmented into paragraphs and sentences, and then parameters are extracted using a rule-based method combined with a lightweight named entity recognition model. For parameters directly matched from the document text, the first confidence level is labeled based on the completeness of the match and the consistency of the units. The value of the first confidence level ranges from 0.85 to 1.00, with an optimal value of 0.95.

[0035] For parameters obtained through contextual inference—that is, parameters whose names or values ​​are not explicitly stated but inferred from the semantics of adjacent text—they are labeled with a second confidence level. The second confidence level ranges from 0.50 to 0.84, with a preferred value of 0.70. Each explicit structured parameter is uniformly represented as a tuple: In the formula: For the first One explicit structured parameter; For parameter names; To assign values ​​to the parameters; Units of measurement; For data types; Source line number; To extract confidence levels.

[0036] S113. Identify implicit parameters in each sub-document that are indirectly represented by equipment model, process standard, or operating procedure. Implicit parameters refer to engineering parameters that are not directly written in numerical form in the document but can be indirectly derived through equipment model, the process standard followed, or the operating procedure. The query and completion are based on an oilfield operations domain knowledge base, which includes an equipment parameter database, a process standard database, and an operating procedure database.

[0037] When the equipment model is identified, the system queries the equipment parameter database to obtain the corresponding rated operating parameters and factory performance parameters. Rated parameters include rated power, rated displacement, and rated pressure, while performance parameters include maximum hook load, maximum torque, and maximum speed. When the system identifies a process standard number or operating procedure name, the system queries the corresponding database to obtain recommended parameter ranges, such as drilling fluid density range, cement slurry mix ratio range, or fracturing fluid viscosity range.

[0038] When performing inference and completion for implicit parameters, if the knowledge base provides a fixed value, it is used directly; if a range is provided, the midpoint of the range is taken as the completion value. The inference confidence level is determined based on the authority of the knowledge base source. The inference confidence level from national or industry standard libraries ranges from 0.80 to 0.95, with a preferred value of 0.90; the range from equipment manufacturer manuals ranges from 0.75 to 0.90, with a preferred value of 0.85; and the range from internal company operating procedures ranges from 0.70 to 0.85, with a preferred value of 0.80.

[0039] S12. Cross-document consistency verification and confidence level identifier generation, specifically including:

[0040] S121. Integrate explicit structured parameters and implicit inferred parameters to establish a unified parameter pool, and perform deduplication and alignment based on parameter name and engineering meaning. During cross-document consistency verification, compare the values ​​of parameters pointing to the same engineering meaning in different sub-documents, i.e.: In the formula: For relative differences, For sub-documents The value of this parameter is... For sub-documents The value of this parameter is... To prevent extremely small positive numbers from being divided by zero, the range of values ​​is 10. -6 Up to 10 -3 Preferred value 10 -4 If the relative difference exceeds the preset tolerance threshold, it is determined to be a version conflict. The tolerance threshold for length parameters ranges from 1% to 5%, with a preferred value of 2%; the tolerance threshold for pressure parameters ranges from 2% to 8%, with a preferred value of 5%; and the tolerance threshold for density parameters ranges from 0.5% to 3%, with a preferred value of 1%.

[0041] When version conflicts exist, a baseline value is selected based on document source priority rules. The priority order is: design documents with higher approval levels take precedence over documents with newer compilation dates, and documents with newer compilation dates take precedence over reference standards with higher document type authority. After verification, a version consistency flag is generated for each parameter. No conflict is flagged as consistent, and conflict is flagged as conflict, and the baseline value source document number is recorded.

[0042] S122. During the partitioning, a comprehensive confidence score is calculated for each parameter. The comprehensive confidence score is obtained by weighting the extracted confidence score or inferred confidence score with the version consistency marker. If the version consistency marker is conflicting, a penalty score is deducted from the corresponding confidence score, i.e.: In the formula: For the first The overall confidence score of each parameter; For the original confidence level, take the parameter when explicit. When implicit parameters are taken ; This is a quantification value for version consistency; it is set to 1 when there is no conflict and 0 when there is a conflict. This is a conflict indicator; it is set to 1 if a version conflict exists, and 0 otherwise. The original confidence level weight ranges from 0.4 to 0.6, with a preferred value of 0.5. The version consistency weight ranges from 0.4 to 0.6, with a preferred value of 0.5. The conflict penalty score ranges from 0.1 to 0.3, with a preferred value of 0.2.

[0043] A confidence threshold is set, ranging from 0.75 to 0.85, with 0.80 being the preferred value. Parameters with a comprehensive confidence score greater than or equal to the threshold are classified into the high-confidence parameter subset, while parameters with a comprehensive confidence score less than the threshold are classified into the low-confidence parameter subset. The high-confidence parameter subset serves as the primary driving data for subsequent precise mapping, while the low-confidence parameter subset serves as auxiliary reference data and triggers a conservative mapping strategy.

[0044] S123. During integrity verification, a pre-stored list of key parameters for each job type is invoked. This list defines the required parameter sets for each of the drilling, fracturing, workover, and completion job types. Each parameter entry in the structured engineering parameter set is checked. If a parameter in the key parameter list is not present in the set or appears but has an empty value, it is determined to be a missing parameter. A binary parameter missing identifier is generated for each missing parameter, with a value of 1 indicating missing and a value of 0 indicating that it has been retrieved.

[0045] Simultaneously, working condition vectors are extracted from the engineering plan document collection. These vectors include formation pressure coefficient, geothermal gradient, rock drillability grade, and fluid viscosity. Each component is obtained through regular expression matching or numerical extraction. When extracting the time sequence identifiers for operation stages, the order of operation stages is identified based on the stage titles or timestamps in the documents, generating a time sequence identifier sequence containing stage names and stage numbers. Missing parameter identifier vectors are represented as follows: In the formula: This is a vector indicating missing parameters. For the first The missing parameter is indicated by a value of 1, which means it is missing and a value of 0 means it has been obtained. This represents the total number of parameters in the key parameter list. It also involves calculating the number of activated devices, moving entities, and physical calculation dimensions in the structured engineering parameter set, generating parameter complexity identifiers. These identifiers are used for subsequent information density grading adjustments in scene rendering.

[0046] S2. Combining version consistency markers and based on the working condition vector selection and effective mapping rules, map two types of parameters as high-confidence simulation parameters and low-confidence simulation parameters, and label the parameter confidence level; complete missing parameters based on time sequence identifiers to obtain complete engineering parameters, and map the complete engineering parameters to complete simulation parameters; using high-confidence simulation parameters as anchor points to propagate to low-confidence simulation parameters and complete simulation parameters, and after rollback verification through a multi-scale physical constraint network, obtain physically self-consistent simulation scene parameter configuration; specifically including:

[0047] S21. Two-layer dynamic mapping and credibility level conversion specifically include:

[0048] S211. Construct a two-layer dynamic mapping dictionary. The first-level index is the job type, including drilling, fracturing, workover, and completion categories. The second-level index is the working condition vector, including formation pressure coefficient, geothermal gradient, rock drillability grade, and fluid viscosity dimension. Each mapping rule includes source engineering parameters, target simulation parameters, mapping transformation function, applicable job type label set, and activation condition parameter set.

[0049] A smoothing mechanism for the working condition transition zone is introduced into the secondary index. This mechanism calculates the weighted Euclidean distance between the current working condition vector and the center vectors of adjacent working condition categories, i.e.: In the formula: This is the weighted Euclidean distance between the current working condition vector and the center vectors of adjacent working condition categories; This is the current operating condition vector; The center vector of adjacent working condition categories; The first of the current working condition vectors The components include formation pressure coefficient, geothermal gradient, rock drillability grade, and fluid viscosity. The first of the center vectors of adjacent working condition categories One component; For the first The weights of each working condition component satisfy... The weight of each component ranges from 0.05 to 0.40, and the preferred value is determined based on the degree of influence of the component on operational safety. Here, represents the dimension of the working condition vector, which is set to 4. Each working condition component has been normalized before calculating the weighted Euclidean distance, making... The distance is dimensionless. When the distance is less than the transition bandwidth, it is determined to be in the transition interval. The transition bandwidth ranges from 0.15 to 0.40, with a preferred value of 0.25.

[0050] Within the transition interval, for multiple mapping rules pointing to the same target simulation parameter, the Gaussian kernel function is used to calculate the degree of matching between the current operating condition and the rule activation condition, i.e.: In the formula: For the current working condition With rules Activation conditions The degree of matching between them The Gaussian kernel bandwidth parameter ranges from 0.1 to 2.0, with a preferred value of 0.5. This is then fused with the historical reputation score of the rule to obtain the dynamic activation factor, i.e.: In the formula: For mapping rules Under operating conditions The dynamic activator under, For rules Historical credibility For matching degree weight, The credibility weight ranges from 0.3 to 0.7, with a preferred value of 0.5. This is the preset threshold for the dynamic activation factor. The value range is from 0.30 to 0.60, with a preferred value of 0.45. Values ​​exceeding this threshold are filtered out. The rules constitute the effective mapping rule set .

[0051] S212. Based on the version consistency flag, apply the effective mapping rule set to each parameter in the high-confidence parameter subset. Perform exact mapping transformations. Exact mapping transformations include three types: direct mapping, computational mapping, and transformation mapping. Direct mapping only performs unit conversions; computational mapping performs explicit function calculations based on physical or empirical formulas; transformation mapping uses table lookup interpolation or nonlinear transformations.

[0052] For high-confidence parameters without version conflicts, the single rule with the highest dynamic activation factor is selected for transformation, and a unique value is output as the high-confidence simulation parameter. For parameters with version conflicts, a soft-voting fusion strategy is adopted, taking the weighted average of the mapping results of multiple rules, with the weights determined by the normalized value of the dynamic activation factor. After the accurate mapping transformation is completed, a mapping confidence score is generated based on the rule matching degree and historical reputation. The mapping confidence score ranges from 0.80 to 1.00, with an optimal value of 0.90.

[0053] S213. Apply the effective mapping rule set to each parameter in the low-confidence parameter subset. Perform a conservative mapping transformation with a tolerable deviation range. The conservative mapping transformation calculates the mapping result for each effective rule and outputs a tolerable deviation range flag. The tolerable deviation range is determined based on the parameter type: for length parameters, it is ±3% to ±8% of the mapping center value, with a preferred value of ±5%; for pressure parameters, it is ±5% to ±12% of the mapping center value, with a preferred value of ±8%; and for density parameters, it is ±2% to ±6% of the mapping center value, with a preferred value of ±4%.

[0054] The final conservative mapping interval is determined by performing interval union operation on the results of multiple rules, and the median of the interval is taken as the center value of the low-confidence simulation parameter. The mapping confidence value of the conservative mapping transformation ranges from 0.40 to 0.75, with an optimal value of 0.60. Each low-confidence simulation parameter is marked with a tolerable deviation range.

[0055] S214. Based on the original extraction confidence or inferred confidence and confidence attenuation of each parameter during the mapping transformation process, calculate the comprehensive confidence of each simulation parameter and label the parameter confidence level. The confidence attenuation is determined according to the mapping transformation function type. The attenuation value for direct mapping ranges from 0.00 to 0.05, with a preferred value of 0.02; the attenuation value for calculated mapping ranges from 0.03 to 0.10, with a preferred value of 0.06; and the attenuation value for conversion mapping ranges from 0.05 to 0.15, with a preferred value of 0.10. If unit conversion is involved, the basic unit conversion adds an attenuation of 0.01 to 0.03, with a preferred value of 0.02; and the composite unit conversion adds an attenuation of 0.03 to 0.05, with a preferred value of 0.04. If interpolation is involved, linear interpolation is subject to an additional attenuation of 0.02 to 0.05, with a preferred value of 0.03; nonlinear interpolation is subject to an additional attenuation of 0.05 to 0.10, with a preferred value of 0.08.

[0056] The overall confidence level equals the original confidence level minus the total attenuation, i.e.: In the formula: For the first The overall reliability of each simulation parameter For the first The original confidence levels of each parameter For the first The total confidence decay of each parameter In the formula: The decay amount is the type of the mapping transformation function. Convert the attenuation amount to units. This represents the attenuation amount in the interpolation calculation. The high-confidence simulation parameters and low-confidence simulation parameters are merged to form a preliminary mapped set of simulation scene parameters.

[0057] S22. Missing parameter completion for multi-strategy time-coupled operations, specifically including:

[0058] S221. Read the parameter missing identifier vector and filter for missing parameter indices with a value of 1. For each missing parameter, read the operation stage time sequence identifier sequence and determine the operation stage name and stage number to which the missing parameter belongs. According to the standard oilfield operation procedure, identify the set of upstream completed stages for the current stage. The upstream stages are all completed stages in the time sequence identifier sequence whose numbers are less than the current stage number.

[0059] Parameters from upstream operational phases are retrieved from the structured engineering parameter set, and parameters with physical transfer relationships to the currently missing parameters are selected as known condition parameters. These physical transfer relationships are determined based on an oilfield operational parameter physical correlation database, including material flow transfer relationships, energy flow transfer relationships, and state inheritance relationships. This database is stored using a graph database, employing a knowledge graph approach. In the formula: For the set of parameter nodes, This is a set of edges representing the relationships between parameters. Material flow transfer relationships refer to the material parameters output from the upstream stage serving as input parameters for the downstream stage. Energy flow transfer relationships refer to the transmission of pressure, temperature, or power parameters from the upstream stage to the downstream stage. State inheritance relationships refer to the maintenance and continuation of equipment state or wellbore state parameters between stages.

[0060] S222. Using the selected upstream completed stage parameters as known conditions, the first priority completion strategy is invoked. In the physical relationship database, starting from the known parameter node and ending at the missing parameter node, a breadth-first path search is performed to find a physically deterministic relationship path from the upstream completed stage parameter to the currently missing parameter. The physically deterministic relationship path consists of several physically constrained edges connected sequentially, with each edge corresponding to a physical formula or deterministic equation.

[0061] If one or more valid paths are found, and all intermediate parameters along the path have been assigned values ​​in the structured engineering parameter set or can be calculated using other known parameters, then the path is considered complete. The physical formulas along the path are then called, and the known parameter values ​​are substituted into the right side of the formula or directly rearranged to calculate the missing parameter values. Specifically: , , In the formula: Bottom hole pressure, in MPa; The pressure at the wellhead is expressed in MPa. Liquid column density, unit ; For gravitational acceleration, in units ; Vertical depth, in meters (m); This refers to the mechanical drilling rate, expressed in m / h. This is an empirical coefficient for drilling speed; Drilling pressure, unit: kN; The uniaxial compressive strength of rock is expressed in MPa. Rotational speed, unit ; It is a dimensionless empirical index; Displacement, unit: L / s; The diameter of the wellbore is in mm. The outer diameter of the drill string is in mm. The annular return velocity is expressed in m / s. The completion confidence of this completion strategy is directly set to 1.0, and it is marked as physical constraint completion.

[0062] S223. If the first priority strategy does not return a valid result, the second and third priority completion strategies are triggered sequentially: the second priority strategy is to complete the case search for similar working conditions, and the calculation process is as follows: retrieve the case search results with the highest similarity to the current working condition condition vector from the historical work database. In one case, the imputed value was calculated using locally weighted regression: , , In the formula: Provide completion values ​​for case retrieval; For the first The values ​​of the corresponding parameters in a number of similar cases; For the first The weight of each case; For the first The distance between each case and the current operating condition vector. , The first of the current working condition vectors One portion, For the first The working condition vector of the historical case. One portion, The number of dimensions for the working condition features; This is a bandwidth parameter, with a value ranging from 0.1 to 2.0, and a preferred value of 0.5. The number of cases to be retrieved ranges from 3 to 10, with a preferred value of 5. for The coefficient of variation for each case value; if Then let .

[0063] The third priority is regression prediction completion, which calls the pre-trained regression model to output predicted values ​​and prediction intervals, with the completion confidence level as follows: In the formula: To complete the confidence level for regression prediction; To predict the interval width; This is the width of the parameter value range. After completing the calculation of each priority, arbitration is performed. The preset difference threshold range is 0.05 to 0.20, with a preferred value of 0.10. If the confidence of the high-priority strategy is higher than that of the low-priority strategy, or the difference between the two does not exceed the preset difference threshold, the result of the high-priority strategy is adopted; otherwise, the completion value with the highest confidence among all strategies is adopted.

[0064] S224. If the confidence level of the completed value obtained after the above priority arbitration is still lower than the preset threshold, and the proportion of the number of missing parameters to the total number of parameters in the key parameter list exceeds the preset proportion, then the fourth priority generative model joint completion is initiated; where: In the formula: This represents the proportion of missing parameters. This represents the number of missing parameters. This represents the total number of parameters in the key parameter list. The preset threshold range is 0.50 to 0.70, with an optimal value of 0.60. The preset percentage range is 15% to 25%, with an optimal value of 20%.

[0065] The fourth-priority generative model joint completion uses a variational autoencoder as the generative model, and performs [the process] in the latent space. The mean of the second sample is used as the padding value: , In the formula: To complete the values ​​for the generative model; For the first Parameter values ​​generated by sub-latent space sampling; The number of samples is 50 to 200, with 100 being the preferred value. for The standard deviation of the sampling results; for The mean of the sampling results; if Then let After determining the final completion value, the completion engineering parameter is mapped and converted into a completion simulation parameter in the simulation parameter space. The mapping conversion is performed using the same mapping rules as S212, and the completion simulation parameter and its mapping confidence are output.

[0066] S23. Anchoring constraint propagation and multi-scale physical self-consistency verification, specifically including:

[0067] S231. Using high-confidence simulation parameters as anchor points, unidirectional constraint propagation is performed from the anchor points to low-confidence simulation parameters and supplementary simulation parameters along the physical deterministic relationships in the oilfield operation parameter physical correlation database. For each low-confidence simulation parameter or supplementary simulation parameter to be corrected, the physical deterministic relationship path between it and the anchor parameters is retrieved, and the theoretical constraint value is calculated based on the physical formulas on the path.

[0068] The deviation is obtained by comparing the current parameter value with the theoretical constraint value. The formula for calculating the constraint correction value is as follows: In the formula: For the first Constraint correction values ​​for low-confidence or completed simulation parameters; For the first The current value of each parameter after mapping or completion; This is a correction factor, with a value ranging from 0.6 to 1.0, and a preferred value of 0.8. For the first The deviation of each parameter In the formula: For the first Each parameter is a theoretical constraint value calculated based on the anchoring parameter through a physical deterministic relationship.

[0069] If multiple constraint paths exist, the weighted average of the correction values ​​calculated for each path is taken, i.e.: In the formula: For the first Correction values ​​for constrained path calculations. For the first The length of the path, The longest path length. To effectively constrain the total number of paths; For the first The weight of each path is determined based on its length; the shorter the path, the higher the weight. The weight of the shortest path ranges from 0.5 to 0.8, with a preferred value of 0.6. The weight of the second shortest path ranges from 0.2 to 0.4, with a preferred value of 0.3. Specifically: .

[0070] S232. The high-confidence simulation parameters in the initially mapped simulation scenario parameter set are merged with the low-confidence simulation parameters corrected by constraint propagation and the completed simulation parameters to form a complete simulation scenario parameter configuration. Based on this configuration, a multi-scale physical constraint network is constructed, including a microscopic contact mechanics sub-network, a mesoscopic fluid dynamics sub-network, and a macroscopic structural mechanics sub-network:

[0071] The microscopic contact mechanics subnetwork uses drill bit-rock contact parameters and casing-wellwall friction parameters as nodes, and Hertz's contact equation, Coulomb's friction law, and interfacial bond strength criterion as constraints, to verify interfacial stress and deformation compatibility. The mesoscopic fluid dynamics subnetwork uses drilling fluid performance parameters, annular return velocity parameters, and fracturing fluid loss parameters as nodes, and Navier-Stokes equations, mass conservation equations, and energy conservation equations as constraints, to verify flow field distribution and heat and mass transfer. The macroscopic structural mechanics subnetwork uses derrick load parameters, drill string axial force parameters, and casing buckling parameters as nodes, and Euler-Bernoulli beam equations, Mohr's strength criterion, and buckling equilibrium equations as constraints, to verify overall structural strength and stability. These three subnetworks achieve cross-scale coupling through bidirectional transmission of wall shear stress to the fluid boundary, fluid pressure to structural load, and structural deformation to the contact gap.

[0072] The cross-scale boundary condition transfer uses a weighted average mapping, such as the transfer of wall shear stress from the micro to the meso scale: In the formula: The number of microscopic contact interfaces; For the first Wall shear stress at a microscopic contact interface; As the interface area weight, satisfying .

[0073] S233. Substitute the fused complete simulation scene parameter configuration into the multi-scale physical constraint network and perform multi-scale physical consistency rollback verification. In the microscopic contact mechanics subnetwork, substitute the contact parameters into Hertz's contact equation and Coulomb's friction law to calculate the interface stress residual, i.e.: In the formula: For the residuals inside the micro-contact mechanics subnetwork, For the residuals of the Hertzian contact equation, For the residual of Coulomb's law of friction, Configure parameters for the complete simulation scene.

[0074] In the meso-level fluid dynamics subnetwork, fluid parameters are substituted into the Navier-Stokes equations and the mass conservation equations to calculate the velocity and pressure residuals of the flow field, i.e.: In the formula: For the residuals within the meso-level fluid dynamics subnetwork, For the Navier-Stokes equation residuals, This represents the residual of the mass conservation equation.

[0075] In the macroscopic structural mechanics subnetwork, the structural parameters are substituted into the Euler-Bernoulli beam equation and the Mohr strength criterion to calculate the stress residual and deformation residual, i.e.: In the formula: For the residuals within the macroscopic structural mechanics subnetwork. For the Euler-Bernoulli beam equation residuals, The residuals are for the Mohr strength criterion.

[0076] At the cross-scale boundary, the wall shear stress output by the micro-subnetwork is compared with the fluid boundary conditions of the meso-scale subnetwork to calculate the cross-scale coupling residual, i.e.: In the formula: For micro- to meso-scale coupled residuals, The wall shear stress output by the micro-subnetwork. The fluid boundary wall shear stress is input to the mesoscopic subnetwork.

[0077] By comparing the fluid pressure output by the mesoscale subnetwork with the structural load of the macroscale subnetwork, the cross-scale coupling residual is calculated, i.e.: In the formula: For meso-to-macro scale coupled residuals, The fluid pressure output by the mesoscopic subnetwork, The structural load pressure is input to the macroscopic subnetwork.

[0078] By comparing the structural deformation output by the macroscopic subnetwork with the contact gap of the microscopic subnetwork, the cross-scale coupling residual is calculated, i.e.: In the formula: For macroscopic to microscopic cross-scale coupled residuals, This represents the structural deformation output by the macroscopic subnetwork. The contact gap quantity is the input to the micro subnetwork.

[0079] The formula for calculating multi-scale physical consistency loss is: In the formula: This represents a multi-scale physical consistency loss. It is the set of all physical constraint edges in a multi-scale physical constraint network; For the first The weight of each physical constraint edge; For the first The residual function of the physical constraints, For dimensionless normalized residuals, This is the reference value for the physical quantity corresponding to the physical constraint. Identify the parameters whose intra-scale residuals or cross-scale coupling residuals exceed the preset tolerance threshold ε and their associated physical constraint relationships. The preset tolerance threshold ε ranges from 0.01 to 0.05, with a preferred value of 0.02.

[0080] S234. Using the physical residuals detected by multi-scale physical consistency rollback verification as the optimization target, perform cross-scale conflict resolution iterative fine-tuning on parameters with confidence levels below a preset threshold. The preset threshold ranges from 0.50 to 0.70, with a preferred value of 0.60. During the iteration process, high-confidence simulation parameters remain locked, and only low-confidence simulation parameters and completed simulation parameters are adjusted. The gradient descent method is used for iteration, i.e.: In the formula: For the first The subset of simulation scenario parameters to be adjusted in the next iteration; The learning rate ranges from 0.001 to 0.01, with a preferred value of 0.005. The physical consistency loss function is the gradient of the parameter to be adjusted.

[0081] After each iteration, the residuals within each scale and the cross-scale coupling residuals are recalculated. Iteration stops when all residuals fall within the preset tolerance threshold ε. The maximum number of iterations ranges from 50 to 200, with 100 being the preferred value. If convergence is not achieved after reaching the maximum number of iterations, the set of parameters with the smallest residuals during the iteration process is taken as the output to obtain the physically self-consistent complete simulation scene parameter configuration.

[0082] S3. Input the simulation scene parameter configuration into the inverse kinematics solver, introduce obstacle constraints and redundant degrees of freedom to generate trajectory keyframes, and interpolate to obtain animation curve data; assemble the model and perform adaptive adjustment and rendering of scene information density based on temporal identifiers and parameter complexity to generate the simulation scene; perform multiple verifications, calculate the scene generation quality score, and provide feedback corrections based on the verification results; specifically including:

[0083] S31. Parameter-driven linked compilation and trajectory keyframe generation, specifically including:

[0084] The S311 3D model adopts a layered structure, including a geometric model layer, a physical property layer, and a joint hierarchy. The geometric model layer stores triangular meshes, surface textures, and spatial topology information; the physical property layer stores mass, density, elastic modulus, and inertia tensor; the joint hierarchy uses a parent-child node chain connection to define the degrees of freedom of translational, rotational, and ball joints. Based on the job type and equipment specifications, a matching model is retrieved from the 3D model library. If a perfect match exists, it is directly called; otherwise, the model with the highest feature similarity is selected, triggering parameter-driven deformation. The equipment specifications are obtained by mapping and transforming the structured engineering parameter set extracted in S1 using S2. Parameter-driven deformation adjusts the vertex coordinates of the geometric model layer according to the length, diameter, and height dimensions in the equipment specifications, using bone binding and mesh deformation algorithms to ensure consistency with the specifications. Static attribute parameters, including material color, surface roughness, and mass distribution, are assigned to the physical property layer interface; initial state parameters, including initial position coordinates, orientation angle, and scaling ratio, are assigned to the root node transformation matrix, completing the parameterized instantiation of the 3D model.

[0085] The S312. Inverse kinematics solver adopts a modular structure, including a target pose input module, an obstacle distance potential field calculation module, a Jacobi iterative solution engine, and a trajectory output module. The target pose input module receives the target pose sequence from the end effector, including the target's 3D coordinates and quaternions. The obstacle distance potential field calculation module pre-defines the obstacle space envelope, approximating the obstacle geometry using a sphere or bounding box, and calculates the shortest distance between the joint trajectory and the obstacle envelope in real time. An obstacle distance potential field term is introduced into the solver's objective function; the potential field term is defined as:

[0086] ,

[0087] In the formula: For joint configuration The obstacle distance potential field term; This is the potential field gain coefficient, with a value range of 10. 3 Up to 10 5 Preferred value 10 4 ; The shortest distance between the joint trajectory and the obstacle envelope; The safe distance threshold ranges from 0.5 m to 2.0 m, with 1.0 m being the preferred value. The potential field term monotonically increases as the distance between the joint movement trajectory and the obstacle decreases, guiding the trajectory away from the obstacle.

[0088] The Jacobi iterative solution engine takes minimizing the target pose error as its core objective, adds a penalty term for the obstacle potential field, and iteratively updates the joint angles using the Jacobi matrix transpose method, with a single iteration step size of [missing information]. The value ranges from 0.01 to 0.10, with a preferred value of 0.05. The maximum number of iterations ranges from 50 to 200, with a preferred value of 100. The guide trajectory deviates from the obstacle.

[0089] S313. The redundant degrees of freedom optimization module is embedded in the inverse kinematics solver. For mechanical systems with redundant degrees of freedom, such as a seven-DOF robotic arm or a traveling block system, it performs joint optimization in the Jacobian null space. The first optimization objective is to minimize joint displacements, i.e., to minimize the Euclidean distance between the current joint angle and the initial joint angle; the second optimization objective is to smooth joint velocities, i.e., to minimize the sum of squared differences in the changes in joint angles between adjacent time steps. The two optimization objectives are weighted and combined as follows: In the formula: To jointly optimize the objective function for redundant degrees of freedom, For the first The coordinate values ​​of each joint. For the first The initial coordinate values ​​of each joint. For the first The first keyframe moment The coordinate values ​​of each joint. This represents the total number of joint degrees of freedom. This represents the total number of keyframes. The value is the joint displacement weight, ranging from 0.3 to 0.7, with a preferred value of 0.5; The joint velocity smoothing weight ranges from 0.3 to 0.7, with a preferred value of 0.5.

[0090] The constraints include end effector target pose constraints and obstacle distance potential field constraints. The pose position error tolerance ranges from 0.001m to 0.01m, with a preferred value of 0.005m. The attitude angle error tolerance ranges from 0.1° to 1.0°, with a preferred value of 0.5°. The solver searches for the optimal joint configuration in the redundant degrees of freedom space. The keyframe time interval ranges from 0.1s to 1.0s, with a preferred value of 0.5s, generating a sequence of keyframes for the joint spatial motion trajectory that meets obstacle avoidance requirements. Each keyframe records the angular displacement of each joint and the corresponding timestamp.

[0091] S314. A cubic spline interpolation algorithm is used to smooth the keyframe sequence of joint spatial motion trajectories. Using the time corresponding to each keyframe as a node and the joint angle as the interpolation variable, polynomial coefficients for each segment are calculated to ensure continuity of joint angle, joint angular velocity, and joint angular acceleration at the nodes. The interpolation time step ranges from 0.01s to 0.05s, with a preferred value of 0.02s. The output is animation curve data that the 3D engine can continuously execute, including the angular displacement, angular velocity, and angular acceleration of each joint at each time step.

[0092] S32. Adaptive adjustment of scene information density and rendering output, specifically including:

[0093] S321. Establish the binding relationship between animation curve data and the joint hierarchy of the 3D model, assigning the joint angular displacement, angular velocity, and angular acceleration of each time step in the animation curve data to the rotation or translation channel of the corresponding joint node. For rotary joints, map the angular displacement to the rotation angle around the joint axis; for translational joints, map the angular displacement to the linear displacement along the joint axis; for ball joints, decompose the angular displacement into rotational components along three orthogonal axes. After the linkage assembly is completed, the components of the 3D model will produce continuous motion as the time step progresses.

[0094] Simultaneously, the environmental parameters in the physically consistent complete simulation scene parameter configuration are mapped to the scene graph parameters of the rendering engine; the lighting direction parameters are mapped to the azimuth and elevation angles of directional light sources; the light intensity parameters are mapped to the radiant flux of the light source; the weather condition parameters are mapped to the atmospheric scattering coefficient and precipitation particle density; and the initial viewpoint position parameters are mapped to the initial world coordinates of the camera. The scene graph parameters are organized in a hierarchical node tree, with the root node being the virtual scene world coordinate system, and child nodes including light source nodes, camera nodes, environment nodes, and model instance nodes. The attributes of each node are updated in real time with the environmental parameters.

[0095] S322. Read the sequence of operation phase identifiers and identify the name of the currently active operation phase. Retrieve matching effect resources from the scene effects library based on the current phase name. Effect resources include particle systems, dynamic lighting, volumetric fog, and screen post-processing materials. Particle systems are used to simulate drilling fluid jets, fracturing fluid backflow, or rock cuttings splash; dynamic lighting is used to simulate well blowout flames or equipment warning lights; volumetric fog is used to simulate steam or dust environments; and screen post-processing materials are used to simulate underwater refraction or high-temperature heat wave distortion.

[0096] Simultaneously, the parameter complexity quantification value is read. This value is directly calculated from the parameter complexity identifier generated by S123, specifically based on the number of simultaneously activated devices, the number of moving entities, and the number of physical calculation dimensions in the simulation scene parameter configuration. In the formula: This is a quantified value for the parameter complexity, ranging from 5 to 50. This refers to the number of devices that can be activated simultaneously. The number of moving entities; To calculate the number of dimensions in physical calculations; These are weighting coefficients, with values ​​ranging from 0.8 to 1.5, and an optimal value of 1.0.

[0097] The density of scene information display is adjusted based on parameter complexity. When the complexity is below 15, all equipment details and parameter labels are displayed. When the complexity is between 15 and 30, the internal structure of secondary equipment and redundant parameter labels are hidden. When the complexity is above 30, only the outline of core equipment and key safety parameters are displayed. The viewing prompt hierarchy is adjusted synchronously, including the display priority of highlighted prompts, arrow guidance, and text annotations. Special effects resources are bound to the corresponding equipment nodes in the 3D model, so that the simulation scene presents corresponding stage-specific visual effects at different operation stages.

[0098] S323. Invoke the 3D rendering engine, load the assembled 3D model, animation curve data, and scene graph parameters, and execute frame-by-frame scene rendering. The rendering pipeline includes a geometry processing stage, a rasterization stage, and a pixel shading stage. The geometry processing stage updates the vertex positions of each model instance based on the animation curve data. The rasterization stage projects the triangular mesh onto screen space. The pixel shading stage calculates lighting, shadows, and environmental effects based on the scene graph parameters. The rendering output frame rate ranges from 30fps to 60fps, with 30fps being the preferred value. The resolution ranges from 1920×1080 to 3840×2160, with 1920×1080 being the preferred value.

[0099] It generates oilfield operation simulation scenarios that support free switching of perspectives. Users can adjust the camera's azimuth, pitch, and line-of-sight using a mouse or touch control to achieve 360-degree panoramic observation. Real-time adjustment of key parameters is supported, dynamically modifying equipment operating parameters during simulation and instantly reflecting them in the scene. Operation process playback is supported, allowing users to play generated animation curve data forward or backward using a timeline control. Progress display is also supported, overlaying the current operation stage name, completion percentage, and estimated remaining time onto the scene interface, forming a complete interactive simulation training environment.

[0100] S33. Multiple verification and closed-loop feedback correction, specifically including:

[0101] S331. Compare the simulation scenario parameter configuration with the pre-stored list of key parameters for each job type. The list of key parameters is maintained separately for each job type, covering the name, data type, and value range requirements of all required parameters for that job type. Iterate through each parameter entry in the list, searching for the same parameter in the simulation scenario parameter configuration. If the search fails or the parameter value is null or invalid, the parameter is considered missing. Generate a missing marker for each missing parameter and record its corresponding equipment system category and process stage category. If any key parameter is missing, trigger the return mechanism, passing the missing parameter list to the multi-strategy fusion completion step to re-execute the completion process. After re-completion, perform integrity verification again until all key parameters are complete or the maximum number of retries is reached. The maximum number of retries ranges from 2 to 5, with 3 being the preferred value.

[0102] S332. Substitute the complete configuration of the simulation scene parameters into the constructed multi-scale physical constraint network, and perform physical consistency verification at the microscopic contact mechanics scale, mesoscopic fluid dynamics scale, and macroscopic structural mechanics scale. At the microscopic scale, calculate the stress and friction residuals at the drill bit-rock interface; at the mesoscopic scale, calculate the velocity and pressure residuals of the annular flow field; and at the macroscopic scale, calculate the stress and deformation residuals of the derrick structure. Simultaneously, calculate the coupling residuals at the cross-scale boundaries, including the wall shear stress transfer residual from micro to mesoscopic, the fluid pressure load transfer residual from meso to macroscopic, and the structural deformation contact gap residual from macroscopic to microscopic. The preset tolerance threshold for physical residuals within each scale ranges from 0.01 to 0.05, with a preferred value of 0.02. The preset tolerance threshold for cross-scale coupling residuals ranges from 0.02 to 0.08, with a preferred value of 0.05. If any residual exceeds the corresponding tolerance threshold, the physical consistency verification is deemed unsuccessful.

[0103] S333. Perform forward kinematics back-substitution verification on the keyframe sequence of joint spatial motion trajectories. Substitute the joint coordinate values ​​corresponding to each keyframe into the forward kinematics equations to calculate the spatial pose of the end effector in the actual joint configuration frame by frame, including the three-dimensional position coordinates and the attitude rotation matrix. Compare the calculated actual pose with the preset target pose of the keyframe to calculate the position deviation and attitude angle deviation, i.e.: , In the formula: The positional deviation is preset to a precision range of 0.001m to 0.01m, with a preferred value of 0.005m. This is the actual position vector of the end effector; The target position vector of the end effector; The attitude angle deviation is preset to a precision range of 0.1° to 1.0°, with a preferred value of 0.5°. This is the actual attitude angle vector of the end effector; This represents the target attitude angle vector of the end effector. The above attitude angle deviations are... The minimum circular arc distance is used for calculation. If Euler angles are used, then the angular components are calculated as follows: Perform periodic adjustments.

[0104] Simultaneously, each joint coordinate value is checked frame by frame to ensure it does not exceed the mechanical limit range of that joint. The mechanical limits include the minimum angle limit, maximum angle limit, minimum displacement limit, and maximum displacement limit of the joint. If the position deviation or attitude deviation exceeds the accuracy range, or if any joint coordinate exceeds the mechanical limit, the trajectory executability verification is deemed to have failed.

[0105] S334. Based on the comprehensive parameter integrity verification results, multi-scale physical consistency verification results, and trajectory executability verification results, a comprehensive quality score is generated for the calculated scene. The quality score is obtained by weighted summation of the integrity score, physical consistency score, and trajectory executability score. The integrity score equals the ratio of the number of complete key parameters to the total number of key parameters; the physical consistency score equals the ratio of the number of constraint edges whose residuals at each scale are below the tolerance threshold to the total number of constraint edges; and the trajectory executability score equals the ratio of the number of keyframes that meet the requirements for pose deviation and joint constraints to the total number of keyframes. Specifically: In the formula: Generate a comprehensive quality score for the scene. The number of complete key parameters, This represents the total number of key parameter lists. This represents the number of constraint edges where the residual is below the tolerance threshold. This represents the total number of physical constraint edges. The number of keyframes required to meet both pose deviation and joint limitation requirements. This represents the total number of keyframes. These are weighting coefficients, with values ​​ranging from 0.25 to 0.40, and the preferred value being 1 / 3.

[0106] The preset pass threshold for the overall quality score ranges from 0.75 to 0.90, with a preferred value of 0.85. If the score is below the pass threshold, the failed item is located in the mapping, completion, or inference stage, and the verification failure information is categorized and fed back to the corresponding module for correction. The optimization of the document source priority rules includes adjusting the weight of the sub-document compilation date, the authority of the document type, and the approval level.

[0107] Example 2

[0108] See appendix Figure 4 A variable mapping and parameter completion system for oilfield operation simulation scenarios, used to implement a method for variable mapping and parameter completion in oilfield operation simulation scenarios, including:

[0109] The multi-source document collaborative parsing module is configured to acquire a set of engineering scheme documents, extract explicit structured parameters and implicit inferred parameters and label their confidence levels, perform cross-document consistency verification, identify version conflict parameters and generate version consistency markers, merge and divide them into high-confidence and low-confidence parameter subsets, generate missing markers, and extract working condition vectors and time series markers.

[0110] The dynamic mapping and physical self-consistent completion module is configured to combine version consistency markers with working condition vectors to filter effective mapping rules, map two types of parameters as high-confidence simulation parameters and low-confidence simulation parameters and mark the parameter confidence identifiers, complete missing parameters according to time sequence identifiers to obtain complete engineering parameters and map them into complete simulation parameters, and use high-confidence simulation parameters as anchor points to propagate to low-confidence simulation parameters and complete simulation parameters, and obtain physically self-consistent simulation scene parameter configurations through multi-scale physical constraint network rollback verification;

[0111] The scene linkage compilation and closed-loop verification module is configured to input the simulation scene parameter configuration into the inverse kinematics solver, introduce obstacle constraints and redundant degrees of freedom to generate trajectory keyframes and interpolate to obtain animation curve data, assemble the model and perform scene information density adaptive adjustment and rendering based on time sequence identifiers and parameter complexity to generate the simulation scene, perform multiple verifications and calculate the scene generation quality score, and provide feedback correction on the verification results.

[0112] Example 3

[0113] To verify the effectiveness of the variable mapping and parameter completion method for oilfield operation simulation scenarios described in this invention, a comparative experiment was conducted based on real engineering scheme data from a certain oilfield operation simulation training system.

[0114] Experimental dataset: A total of 80 real oilfield operation engineering plan documents were collected, including 40 drilling engineering plans, 25 fracturing construction plans, and 15 well workover operation plans. In each plan, 15%–25% of parameters were randomly selected to simulate common scenarios in actual engineering plans, such as parameter abbreviations, missing items, and inconsistencies between multiple source documents.

[0115] Comparison method:

[0116] Method A (Traditional Manual Method): Oilfield simulation engineers with more than 5 years of experience manually read the engineering plan documents, manually extract parameters and configure the simulation scene, and manually adjust the pose and animation trajectory of the 3D model.

[0117] Method B (Fixed Mapping Table + Default Value Method): A preset static parameter lookup table is used to perform a one-to-one mapping conversion from engineering parameters to simulation parameters. Missing parameters are directly filled with industry-standard default values, without physical consistency verification.

[0118] Method C (Fixed Mapping Table + Single Regression Completion Method): A static mapping table is used for parameter transformation. Missing parameters are completed using a single multivariate regression model trained on historical data. After completion, there is no iterative fine-tuning or physical verification.

[0119] The method of this invention: The method described in this application includes multi-source heterogeneous document collaborative parsing and implicit parameter inference, two-layer context-aware dynamic mapping, temporally coupled multi-strategy completion, anchored constraint transmission and multi-scale physical self-consistency verification, inverse kinematics linkage compilation under redundant degree of freedom obstacle avoidance constraints, and closed-loop verification feedback.

[0120] Evaluation indicators:

[0121] Scene generation success rate: The percentage of scenes that successfully pass parameter integrity verification, multi-scale physical consistency verification, and trajectory executability verification.

[0122] Average scene generation time: The total time from inputting a collection of engineering solution documents to outputting an interactive simulation scene;

[0123] Mean relative error of imputation: The mean relative error between the imputed value and the true value of the scheme (pre-recorded in the experimental database) for missing parameters only;

[0124] Physical consistency first pass rate: The proportion of parameter sets that pass the multi-scale physical consistency check on the first attempt after the mapping is completed (without triggering cross-scale conflict resolution iterative fine-tuning).

[0125] The experimental results are shown in Table 1 below:

[0126] Table 1: Comparison of Results

[0127]

[0128] Results analysis:

[0129] (1) In terms of scene generation success rate, the method of the present invention achieves 95.0%, which is 22.5 percentage points higher than method A, 27.5 percentage points higher than method C, and 51.2 percentage points higher than method B. Method B uses static mapping and default value filling, which causes the mapping rules to fail and the completed parameters to deviate significantly from the physical constraints when switching working conditions, resulting in a large number of scenes failing the basic verification. Although method C introduces regression completion, it lacks physical consistency closed-loop verification, and the completion error propagates along the parameter coupling chain, so the final scene pass rate is still limited. The present invention uses a two-layer context-aware dynamic mapping dictionary to adaptively match the working condition transition zone mapping rules, combined with multi-strategy fusion completion and multi-scale physical constraint network rollback verification, which fundamentally ensures the physical self-consistency of the parameter set and the executability of the scene.

[0130] (2) Regarding the average generation time, the method of this invention takes only 18 minutes, which is 90% shorter than manual method A, and 60% and 65% shorter than semi-automatic methods B and C, respectively. Method A requires engineers to read multi-source documents page by page, manually extract parameters, and manually configure simulation models and animation curves; although methods B / C eliminate some manual mapping steps, manual intervention is still required to adjust the mapping table in the face of changes in working conditions. This invention automatically generates joint trajectory keyframes through automated collaborative parsing of multi-source heterogeneous documents, implicit parameter inference, two-layer dynamic mapping routing, and inverse kinematics solver, realizing one-click automatic compilation from "engineering scheme text parameters" to "3D executable animation", compressing the overall scene generation efficiency from hours to minutes.

[0131] (3) Regarding the average relative error of the completion, the method of this invention has an error of only 6.2%, which is 73.7% lower than that of method B and 51.6% lower than that of method C. Method B uses fixed default values ​​and does not consider the differences in specific working conditions, resulting in the largest error; Method C uses a single regression model, which cannot handle small samples and nonlinear working conditions. The four-level priority completion strategy chain established in this invention, namely "physical hard constraint priority → similar working condition case retrieval → regression prediction → generative model fallback", combines the temporal coupling completion mechanism to use the parameters of the upstream already operated stage to physically transfer constraints to the missing parameters of the downstream, and uses high-confidence simulation parameters as anchor points to correct low-confidence and completion parameters through anchor constraint transmission, which significantly improves the completion accuracy.

[0132] (4) Regarding the first-pass rate of physical consistency, the method of this invention achieved 91.7%, which is 13.4 percentage points higher than method A, 36.1 percentage points higher than method C, and 60.5 percentage points higher than method B. This indicator directly verifies the effectiveness of the original multi-scale physical constraint network rollback verification and cross-scale conflict resolution iterative fine-tuning mechanism of this invention. Method A relies on the personal experience of engineers to make physical rationality judgments, which is highly subjective and difficult to cover cross-scale coupling constraints; methods B / C completely lack automatic physical verification. This invention uses joint constraint verification of three scales: microscopic contact mechanics, mesoscopic fluid dynamics, and macroscopic structural mechanics. It locks the iteration space with high-confidence parameters as anchoring benchmarks and only performs cross-scale conflict resolution fine-tuning on parameters with confidence levels below the threshold. This ensures that the parameter set after completion and mapping meets the physical residual tolerance requirements within each scale and at the cross-scale boundary during the first verification, ensuring that the simulation scenario is not only numerically accurate but also physically reliable, and can be directly used for subsequent engineering simulation analysis and training simulations.

[0133] In summary, the experimental results show that the present invention significantly outperforms existing comparative methods in four dimensions: scene generation success rate, generation efficiency, completion accuracy, and physical consistency. This fully verifies the effectiveness and advancement of the method described in the present invention in the field of automatic generation of oilfield operation simulation scenes.

[0134] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention is limited to these examples; within the framework of the invention, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of the different aspects of the invention as described above, which are not provided in detail for the sake of brevity.

[0135] This invention is intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this specification. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for variable mapping and parameter completion in oilfield operation simulation scenarios, characterized in that, include: Obtain the engineering plan document set, extract explicit structured parameters and implicit inferred parameters, and label their confidence levels respectively; Perform cross-document consistency verification, identify version conflict parameters and generate version consistency markers, merge and divide them into high-confidence parameter subsets and low-confidence parameter subsets, generate missing markers, and extract working condition vectors and time series markers; Based on the version consistency marker and the working condition vector filtering effective mapping rules, the two types of parameters are mapped as high-confidence simulation parameters and low-confidence simulation parameters, and the parameter confidence level is marked. Missing parameters are filled in based on the time sequence identifier to obtain the completion engineering parameters, and the completion engineering parameters are mapped and converted into completion simulation parameters; The simulation parameters are propagated from high-confidence simulation parameters to low-confidence simulation parameters and complete simulation parameters. After rollback verification through a multi-scale physical constraint network, a physically self-consistent simulation scene parameter configuration is obtained. The simulation scene parameters are configured and input into the inverse kinematics solver. Obstacle constraints and redundant degrees of freedom are introduced to generate trajectory keyframes, and interpolation is used to obtain animation curve data. The model is assembled and, based on the timing identifier and parameter complexity, the scene information density is adaptively adjusted and rendered to generate a simulation scene. Perform multiple verifications, calculate the quality score for the scenario, and provide feedback and corrections based on the verification results.

2. The method for variable mapping and parameter completion in oilfield operation simulation scenarios according to claim 1, characterized in that, The process of obtaining the engineering solution document set, extracting explicit structured parameters and implicit inferred parameters, and labeling their confidence levels specifically includes: Obtain the set of oilfield operation engineering plan documents to be simulated, including the original engineering design, construction plan, equipment list and process specification reference documents for the same operation project. Normalize the format of each sub-document and extract the version number, compilation date and applicable well number of each sub-document. Based on the domain ontology dictionary and parameter pattern library, each sub-document is structured and explicit parameters are extracted to obtain explicit structured parameters. The parameters that are directly matched are labeled as the first confidence level, and the parameters obtained by context reasoning are labeled as the second confidence level. The first confidence level and the second confidence level are collectively referred to as the extraction confidence level. Identify implicit parameters in each sub-document that are indirectly represented by equipment model, process standard or operating procedure, query the corresponding rated parameters, factory performance parameters or recommended parameter range based on the domain knowledge base, infer and complete the implicit parameters and mark the inference confidence level.

3. The method for variable mapping and parameter completion in oilfield operation simulation scenarios according to claim 2, characterized in that, The process of performing cross-document consistency verification, identifying version conflict parameters and generating version consistency markers, fusing and dividing them into high-confidence parameter subsets and low-confidence parameter subsets, generating missing identifiers, and extracting operating condition vectors and time series identifiers, specifically includes: Explicit structured parameters and implicit inferred parameters are merged to form a set of structured engineering parameters. Parameters pointing to the same engineering meaning undergo cross-document consistency checks. If relative differences exist in values, a baseline value is selected based on document source priority rules, and a version consistency flag is generated. The relative differences... for: In the formula: For sub-documents The value of this parameter is... For sub-documents The value of this parameter is... To prevent extremely small positive numbers from being divided by zero; Based on the extraction confidence, inference confidence, and version consistency marker, a comprehensive confidence score is calculated. ,Right now: In the formula: For the original confidence level, take the parameter when explicit. When implicit parameters are taken ; This is a quantification value for version consistency; it is set to 1 when there is no conflict and 0 when there is a conflict. This is a conflict indicator; it is set to 1 if a version conflict exists, and 0 otherwise. The original confidence weights, For version consistency weights, The conflict penalty score is used to divide the fused parameter set into a high-confidence parameter subset and a low-confidence parameter subset. The structured engineering parameter set is checked for completeness against the list of key parameters required for the pre-stored job type. Missing parameters are identified and parameter missing identifiers are generated. At the same time, the working condition vector and job phase sequence identifier are extracted from the engineering plan document set.

4. The method for variable mapping and parameter completion in oilfield operation simulation scenarios according to claim 1, characterized in that, The process, combining the version consistency marker with the effective mapping rules for filtering working condition vectors, maps two types of parameters as high-confidence simulation parameters and low-confidence simulation parameters, and marks the parameter confidence level, specifically including: A two-level dynamic mapping dictionary is constructed, with job type as the first-level index and working condition as the second-level index. A working condition transition smoothing mechanism is introduced to continuously interpolate the dynamic activation factors of multiple mapping rules in the transition interval. The current working condition vector and job type are obtained, and the dynamic activation factors of each candidate rule are calculated. ,Right now: In the formula: For rules Historical credibility For matching degree weight, As a weight for reputation, For the current working condition With rules Activation conditions The degree of matching between them In the formula: For Gaussian kernel bandwidth parameters, The weighted Euclidean distance between the current working condition vector and the center vectors of adjacent working condition categories is used to filter rules that exceed a preset threshold to form an effective mapping rule set. Based on the version consistency flag, for each parameter in the high confidence parameter subset, the effective mapping rule set is invoked to perform precise mapping transformation, generating high confidence simulation parameters and their mapping confidence; For each parameter in the low-confidence parameter subset, a conservative mapping transformation is performed by calling the effective mapping rule set to generate low-confidence simulation parameters and their mapping confidence. The conservative mapping transformation outputs parameter values ​​with tolerable deviation range labels. The overall reliability of each simulation parameter is calculated based on the original extraction confidence level and the confidence decay. And label the parameter confidence level, where: In the formula: For the first The original confidence levels of each parameter For the first The total confidence decay of each parameter In the formula: The decay amount is the type of the mapping transformation function. Convert the attenuation amount to units. This is the attenuation amount for interpolation calculation; the high-confidence simulation parameters and low-confidence simulation parameters are merged into a set of simulation scene parameters after preliminary mapping.

5. The method for variable mapping and parameter completion in oilfield operation simulation scenarios according to claim 4, characterized in that, The process of completing missing parameters based on time sequence identifiers to obtain completion engineering parameters, and mapping and converting the completion engineering parameters into completion simulation parameters, specifically includes: Based on the time sequence identifier of the operation stage, determine the stage position of each missing parameter in the operation time sequence chain, and identify the upstream completed stage parameters that have a physical transmission relationship with the current missing parameter. Given the parameters of the upstream completed stages, the system searches the physical relational database for the physical deterministic relational path to the currently missing parameter. If the path exists and the intermediate parameters are complete, the system directly calculates the completed value using the physical formula on the path. Specifically: , , In the formula: For the bottom hole pressure, For wellhead pressure, For the density of the liquid column, It is the acceleration due to gravity. For vertical depth, For mechanical drilling speed, This is an empirical coefficient for drilling speed. For drilling pressure, For the uniaxial compressive strength of rock, For rotational speed, It is a dimensionless empirical index. For displacement, The diameter of the wellbore. The outer diameter of the drill string. For circular return speed; If the first priority strategy is not applicable or the confidence level is lower than the preset threshold, the second priority case retrieval completion and the third priority regression prediction completion will be triggered in sequence. If the confidence level of the high priority strategy is higher than that of the low priority strategy, or the difference between the two does not exceed the preset threshold, the result of the high priority strategy will be adopted; otherwise, the completion value with the highest confidence level among the strategies will be adopted. If the confidence level of the completed value is still lower than the preset threshold after arbitration by each priority strategy, and the scale of missing parameters exceeds the preset proportion of the total parameters, then the fourth priority generative model joint completion is initiated to determine the final completed value and map the completed engineering parameters to the completed simulation parameters, wherein: , In the formula: To complete the values ​​for the generative model, For the first Parameter values ​​generated by sub-latent space sampling For the number of samples, for The standard deviation of the sampling results for The mean of the sampling results.

6. The method for variable mapping and parameter completion in oilfield operation simulation scenarios according to claim 5, characterized in that, The process of propagating simulation parameters from high-confidence parameters to low-confidence and completed simulation parameters, and then performing rollback verification through a multi-scale physical constraint network, yields a physically consistent simulation scene parameter configuration, specifically including: Using high-confidence simulation parameters as anchor points, unidirectional constraint propagation is performed along the physical deterministic relationships in the physical correlation database to low-confidence simulation parameters and completed simulation parameters, and the constraint correction values ​​of each parameter relative to the anchor parameters are calculated. ,Right now: In the formula: For the first The current value of each parameter after mapping or completion. For correction factor, For the first The deviation of each parameter In the formula: For the first Each parameter is a theoretical constraint value calculated based on the anchoring parameter through a physical deterministic relationship; The high-confidence simulation parameters are fused with the low-confidence simulation parameters and the completed simulation parameters after constraint transmission correction. A multi-scale physical constraint network is constructed based on the micro-contact mechanics, meso-fluid dynamics and macro-structural mechanics scales. The physical equilibrium equations and constitutive relations of each scale are used as constraint edges, and each parameter is used as a node. Cross-scale coupling is achieved through boundary condition transmission. The fused simulation scene parameter configuration is substituted into the multi-scale physical constraint network, and a multi-scale physical consistency rollback check is performed. Physical residuals are calculated within each scale and at cross-scale boundaries to identify parameters where cross-scale physical conflicts occur and their associated physical constraint relationships. The multi-scale physical consistency loss is then calculated. The calculation formula is: In the formula: Let be the set of all physical constraint edges in a multi-scale physical constraint network. For the first The weight of each physical constraint edge, For the first The residual function of the physical constraints, For dimensionless normalized residuals, This is the reference value for the physical quantity corresponding to the physical constraint; Using the physical residuals detected by multi-scale physical consistency rollback verification as the optimization target, cross-scale conflict resolution iterative fine-tuning is performed on parameters with confidence levels below a preset threshold. High-confidence simulation parameters remain locked during the iteration process until the physical residuals at each scale and the cross-scale coupling residuals are reduced to within the preset tolerance range, thus obtaining a physically self-consistent simulation scenario parameter configuration.

7. The method for variable mapping and parameter completion in oilfield operation simulation scenarios according to claim 1, characterized in that, The process involves inputting the simulation scene parameters into the inverse kinematics solver, introducing obstacle constraints and redundant degrees of freedom to generate trajectory keyframes, and interpolating to obtain animation curve data. Specifically, this includes: Based on the operation type and equipment specification parameters in the physically self-consistent simulation scenario parameter configuration, a matching 3D model is retrieved from the 3D model library, static attribute parameters are assigned to the attribute interface of the 3D model, and initial state parameters are assigned to the transformation matrix of the 3D model to complete the parameterized instantiation of the 3D model. The kinematic parameters in the simulation scene parameter configuration are input into the inverse kinematics solver. A potential field term representing the distance to obstacles in the workspace is introduced into the objective function. This potential field term is defined as follows: In the formula: For joint configuration The obstacle distance potential field term, The potential field gain coefficient is... The shortest distance between the joint trajectory and the obstacle envelope. This is the safe distance threshold; For a mechanical system with redundant degrees of freedom, the joint optimization objective is to minimize joint displacements and smooth joint velocities, i.e.: In the formula: To jointly optimize the objective function for redundant degrees of freedom, For the first The coordinate values ​​of each joint. For the first The initial coordinate values ​​of each joint. For the first The first keyframe moment The coordinate values ​​of each joint. This represents the total number of joint degrees of freedom. This represents the total number of keyframes. For joint displacement weights, We assign smoothing weights to joint velocities; and, under the premise of satisfying the target pose of the end effector and the potential field constraints of the obstacle distance, we solve for the optimal joint configuration in the redundant degrees of freedom space to generate a keyframe sequence of joint spatial motion trajectories that meets obstacle avoidance requirements. The keyframe sequence is smoothed using a spline interpolation algorithm to generate animation curve data that can be continuously executed by the 3D engine.

8. The method for variable mapping and parameter completion in oilfield operation simulation scenarios according to claim 7, characterized in that, The assembly model then performs adaptive adjustment and rendering of scene information density based on timing identifiers and parameter complexity to generate a simulation scene, specifically including: The animation curve data is linked and assembled with the 3D model, and the environmental parameters in the simulation scene parameter configuration are mapped to the scene graph parameters of the rendering engine. The environmental parameters include the lighting direction, lighting intensity, weather conditions and the initial position of the viewpoint. Based on the timing identifiers and parameter complexity of the work phases, special effects resources matching the current work phase are retrieved from the scene effects library. The density of scene information display and the level of view prompts are adjusted, and the special effects resources are bound to the corresponding device nodes in the 3D model. This allows the simulation scene to present corresponding stage-specific visual effects at different work phases, including parameter complexity. The calculation formula is: In the formula: For parameter complexity, quantification value The number of devices that can be activated simultaneously. For the number of moving entities, To calculate the physical dimension, These are weighting coefficients; The system invokes a 3D rendering engine to perform scene rendering based on the 3D model, animation curve data, and scene diagram parameters after linkage assembly. This generates an oilfield operation simulation scene that supports free switching of perspectives, real-time adjustment of key parameters, playback of the operation process, and display of work progress.

9. The method for variable mapping and parameter completion in oilfield operation simulation scenarios according to claim 8, characterized in that, The process of performing multiple verifications, calculating a quality score for the scenario, and providing feedback corrections to the verification results specifically includes: The simulation scenario parameter configuration is checked against the list of key parameters required for this job type, and it is checked whether there are any missing or unassigned key parameters. If there are any missing parameters, the missing parameters are marked and the multi-strategy fusion completion step is returned to re-trigger the completion. The simulation scene parameter configuration is substituted into the multi-scale physical constraint network, and the physical residuals of each scale constraint edge and the coupling residuals at the cross-scale boundary are calculated respectively. If the physical residual at any point exceeds the preset tolerance threshold, the physical consistency verification is deemed to have failed. The keyframe sequence of joint space motion trajectory is validated by back-substituting the forward kinematics solution. The joint coordinates corresponding to each keyframe are substituted into the forward kinematics equation to calculate the actual pose of the end effector. Specifically: , In the formula: This is the positional deviation. This is the actual position vector of the end effector. The target position vector of the end effector. This refers to the attitude angle deviation. This is the actual attitude angle vector of the end effector. The target attitude angle vector of the end effector is used to determine whether the deviation between the actual pose and the target pose is within the preset accuracy range, and at the same time to verify whether the coordinates of each joint exceed the mechanical limit. Based on the parameter integrity verification results, multi-scale physical consistency verification results, and trajectory executability verification results, a comprehensive quality score is generated for the calculated scenario. ,Right now: In the formula: The number of complete key parameters, This represents the total number of key parameter lists. This represents the number of constraint edges where the residual is below the tolerance threshold. This represents the total number of physical constraint edges. The number of keyframes required to meet both pose deviation and joint limitation requirements. This represents the total number of keyframes. The weighting coefficient is used; if it is lower than the preset passing threshold, the mapping step, completion step or inference step corresponding to the failed item is located, and the verification failure information is classified and fed back, which is used for updating the rule reputation of the dynamic mapping dictionary, optimizing the model parameters of multi-strategy fusion completion, correcting the implicit parameter inference rules and optimizing the document source priority rules, respectively.

10. A system for variable mapping and parameter completion in oilfield operation simulation scenarios, characterized in that, A method for implementing variable mapping and parameter completion in an oilfield operation simulation scenario as described in any one of claims 1-9 includes: The multi-source document collaborative parsing module is configured to acquire a set of engineering scheme documents, extract explicit structured parameters and implicit inferred parameters and label their confidence levels, perform cross-document consistency verification, identify version conflict parameters and generate version consistency markers, merge and divide them into high-confidence and low-confidence parameter subsets, generate missing markers, and extract working condition vectors and time series markers. The dynamic mapping and physical self-consistent completion module is configured to combine version consistency markers with working condition vectors to filter effective mapping rules, map two types of parameters as high-confidence simulation parameters and low-confidence simulation parameters and mark the parameter confidence identifiers, complete missing parameters according to time sequence identifiers to obtain complete engineering parameters and map them into complete simulation parameters, and use high-confidence simulation parameters as anchor points to propagate to low-confidence simulation parameters and complete simulation parameters, and obtain physically self-consistent simulation scene parameter configurations through multi-scale physical constraint network rollback verification; The scene linkage compilation and closed-loop verification module is configured to input the simulation scene parameter configuration into the inverse kinematics solver, introduce obstacle constraints and redundant degrees of freedom to generate trajectory keyframes and interpolate to obtain animation curve data, assemble the model and perform scene information density adaptive adjustment and rendering based on time sequence identifiers and parameter complexity to generate the simulation scene, perform multiple verifications and calculate the scene generation quality score, and provide feedback correction on the verification results.

Images