A kinetically simulated engineered structured intelligent archival review method and system

Abstract

The present application relates to the technical field of data management and intelligent analysis of kinetic simulation engineering, and particularly relates to a structured intelligent archiving review method and system for kinetic simulation engineering. The input of the method is a target kinetic simulation engineering catalog and its configuration file, running artifacts, result artifacts and document basis; the processing process includes engineering scanning, case identification, specification input selection, configuration rule review, running health assessment, structured evidence organization, evidence constraint question and answer generation and report verification; the output is structured evidence, question and answer result, research report and audit record. Through the above technical chain, the present application realizes closed-loop processing from engineering data identification, evidence organization, question and answer generation to report verification, improves the kinetic simulation engineering data management efficiency, conclusion reliability and result traceability, forms a research report with complete evidence traceability, and comprehensively improves the simulation engineering data management efficiency, conclusion reliability and audit reliability.

Description

Technical Field

[0001] This invention relates to the field of data governance and intelligent analysis technology in dynamic simulation engineering, and in particular to a structured intelligent archiving and review method and system for dynamic simulation engineering. Background Technology

[0002] In current research and engineering practice, the simulation data mentioned above is mostly stored in a scattered manner in the form of directories and files. Researchers usually need to manually determine which is the standard input and which is the generated copy, and use different scripts to view the configuration parameters, running status and result data. For multiple parameter scan cases of the same model, it is also necessary to compare the configuration differences, running differences and result differences one by one, which is labor-intensive and prone to omissions.

[0003] Meanwhile, operational health assessments, results optimization, and research report compilation often lack a unified data organization method. Configuration issues, operational risks, results summaries, and documentation are usually scattered across different documents and toolchains, making it difficult to establish stable evidence citation relationships. When researchers further introduce AI-powered question answering or report generation tools, the lack of evidentiary boundary constraints can easily lead to answers that deviate from engineering facts and report conclusions that are difficult to trace.

[0004] Therefore, there is an urgent need to develop a comprehensive methodology that can be used for dynamic simulation engineering to uniformly complete engineering scanning, case identification, configuration review, operation evaluation, result processing, evidence archiving, question and answer generation, and report verification, so as to improve data management efficiency, the credibility of research conclusions, and subsequent review capabilities. Summary of the Invention

[0005] This invention discloses a structured intelligent archiving and review method and system for dynamic simulation engineering projects. The method's input includes the target dynamic simulation engineering project directory and its contained configuration files, operation logs, statistical files, result artifacts, and supporting documentation. The processing includes project scanning, case identification, specification input selection, configuration review, operation evaluation, structured evidence organization, evidence constraint question-and-answer generation, and report verification. The output includes structured evidence, question-and-answer results, research reports, and audit logs. Compared to existing technical approaches that only allow viewing configurations, logs, or results separately, this invention forms a closed-loop technical chain consisting of engineering data identification, evidence organization, question-and-answer generation, and report verification.

[0006] To achieve the above objectives, this invention provides a structured intelligent archiving and review method for dynamic simulation engineering, comprising the following steps: S1. Receive the target simulation project path, standardize the path and build a file scanning tree, automatically identify one or more case root directories based on the file scanning tree, and perform content-aware recognition on candidate configuration files to distinguish parameter files, XML configuration files and generated copies. S2. For each case, select the normative source configuration from the candidate configuration files, construct a case list containing the normative source configuration, generated copies, runtime artifacts and result artifacts, perform deterministic rule review based on the normative source configuration, and generate configuration discovery with evidence line numbers and evidence text. S3. Read the operation log, statistics file and / or residual file, extract the operation indicators and generate the operation health assessment results; S4. Write cases, artifacts, findings, indicators, and task events into the structured evidence storage layer, construct evidence bundles based on the structured evidence storage layer, and generate constrained narration, question-and-answer results, or research reports using the evidence bundles as input. S5. Perform validation on key assertions in the research report and output audit logs.

[0007] In some embodiments of this application, in S1, content-aware recognition includes: When a candidate configuration file contains parameter key-value syntax and a hierarchical structure, it is identified as a first-class parameter configuration file. When a candidate configuration file is in XML format and its root node meets the preset simulation configuration identifier, it is identified as a second type of structured configuration file.

[0008] In some embodiments of this application, in S2, selecting the specification source configuration includes: Candidate configuration files are scored based on whether the file is located in the output directory, whether the file name matches the source configuration naming pattern, and the complexity of parameter recording, and the highest-scoring file is determined as the standard source configuration.

[0009] In some embodiments of this application, in S2, the deterministic rule review includes at least five of the following categories: absolute path detection, duplicate parameter detection, unknown parameter detection, unique label duplicate detection, event count consistency detection, time window continuity detection, filename and time window matching detection, and path file existence detection.

[0010] In some embodiments of this application, in S3, the operational health assessment is calculated comprehensively based on at least three of the following dimensions: time step stability, solver iterative behavior, residual monotonicity, late deceleration characteristics, and abnormal termination markers, and outputs a numerical score and a status label.

[0011] In some embodiments of this application, in S4, the structured evidence storage layer includes at least the entities Project, Case, Artifact, Finding, Metric, TaskRun, TaskEvent, and Report, and each discovery is associated with at least evidence text or evidence line number.

[0012] In some embodiments of this application, in S4, the constrained narration, question-and-answer results, or research report generation process is limited to referencing only findings, metrics, artifacts, task events, or document fragments in the evidence bundle, and includes a cue echo filtering and rollback text mechanism.

[0013] In some embodiments of this application, in step S5, the key assertion verification includes: Extract key conclusions from the research report, retrieve evidence records corresponding to each key conclusion, generate verification results based on the degree of evidence matching, and output audit documents for traceability.

[0014] In some embodiments of this application, a structured intelligent archiving and review system for dynamic simulation engineering is also disclosed, comprising: The project scanning module is used to receive the target simulated project path, standardize the path, and build a file scanning tree; The case identification module is used to automatically identify one or more case root directories based on the file scanning tree, and to perform content-aware identification on candidate configuration files to distinguish between parameter files, XML configuration files and generated copies; The inventory building module is used to select the specification source configuration from the candidate configuration files for each case and build an inventory of cases containing the specification source configuration, generated copies, runtime artifacts and result artifacts. The configuration review module is used to perform deterministic rule reviews based on the source configuration of the specification and generate configuration findings with evidence line numbers and evidence text. The operation review module is used to read operation logs, statistics files and / or residual files, extract operation indicators and generate operation health assessment results; Structured evidence storage is used to write cases, artifacts, findings, metrics, and task events into the structured evidence storage layer; The evidence-constrained AI interaction module is used to construct evidence bundles based on the structured evidence storage layer, and generate constrained narration, question-and-answer results or research reports with the evidence bundles as input. The assertion verification and auditing module is used to verify key assertions in research reports and output audit logs.

[0015] The advantages and beneficial effects of this invention compared to the prior art are: 1. This invention replaces traditional manual judgment, segment-by-segment browsing, and scattered viewing methods with automated engineering scanning, case identification, configuration review, and operational evaluation. As shown in the embodiments, the system can simultaneously complete automatic selection of standardized input, generation of duplicate differentiation, rapid location of configuration anomalies, and quantitative assessment of operational health within a single case, avoiding the tedious work of researchers repeatedly searching through XML configurations, log files, and result artifacts.

[0016] 2. By establishing a structured evidence storage layer that includes entities such as Project, Case, Artifact, Finding, and Metric, heterogeneous information originally scattered across directories, logs, diagrams, and documents is compressed into a unified set of structured facts. All findings are associated with evidence line numbers and evidence text, and the final report has evidence backtracking capabilities, providing clear clues for manual review and completing a closed-loop process from engineering data input to audit result output.

[0017] 3. During the question-and-answer generation and research report generation stages, the system constructs a constrained AI interaction chain based on structured evidence layers and document knowledge, limiting the output to citing only registered evidence. This mechanism effectively prevents the "illusion" problem caused by the lack of evidence boundaries in traditional AI tools, ensuring that all question-and-answer results and report conclusions are supported by engineering facts.

[0018] 4. This invention enables the entire process, from input scanning and deterministic rule review to operational health assessment and final audit output, to be completed within a single system. Compared to existing fragmented technical approaches that only allow viewing configurations, logs, or results separately, this invention achieves unified governance and closed-loop management of heterogeneous simulation data.

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall technical route of the present invention; Figure 2 This is a flowchart illustrating the engineering scanning and case list construction process in an embodiment of the present invention. Figure 3 This is a flowchart illustrating the structured evidence layer supporting question-and-answer and report generation in an embodiment of the present invention; Figure 4 This is a flowchart illustrating the research report verification and audit output process in an embodiment of the present invention. Figure 5 This is a schematic diagram of the real-time scanning workbench interface in an embodiment of the present invention; Figure 6 This is a schematic diagram of the AI ​​workbench interface in an embodiment of the present invention; Figure 7 This is a schematic diagram of the health assessment result interface in an embodiment of the present invention; Figure 8 This is a schematic diagram of the report generation result interface in an embodiment of the present invention. Detailed Implementation

[0021] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product is in use. They are used only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," and "connect" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0022] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0023] like Figure 1 As shown, this invention provides a structured intelligent archiving and review method for dynamic simulation engineering, comprising the following steps: S1. Receive the target simulation project path, standardize the path and build a file scanning tree, automatically identify one or more case root directories based on the file scanning tree, and perform content-aware recognition on candidate configuration files to distinguish parameter files, XML configuration files and generated copies. In one embodiment, a schematic diagram of the real-time scanning workbench interface is shown below. Figure 5 As shown.

[0024] S2. For each case, select the normative source configuration from the candidate configuration files, construct a case list containing the normative source configuration, generated copies, runtime artifacts and result artifacts, perform deterministic rule review based on the normative source configuration, and generate configuration discovery with evidence line numbers and evidence text. In one embodiment, the AI ​​workbench and evidence panel interface are as follows: Figure 6 As shown.

[0025] S3. Read the operation log, statistics file and / or residual file, extract the operation indicators and generate the operation health assessment results; In one embodiment, the health assessment results interface is as follows: Figure 7 As shown, the report generation results interface is as follows: Figure 8 As shown.

[0026] S4. Write cases, artifacts, findings, indicators, and task events into the structured evidence storage layer, construct evidence bundles based on the structured evidence storage layer, and generate constrained narration, question-and-answer results, or research reports using the evidence bundles as input. S5. Perform validation on key assertions in the research report and output audit logs.

[0027] In some embodiments of this application, in S1, content-aware recognition includes: When a candidate configuration file contains parameter key-value syntax and a hierarchical structure, it is identified as a first-class parameter configuration file. When a candidate configuration file is in XML format and its root node meets the preset simulation configuration identifier, it is identified as a second type of structured configuration file.

[0028] In some embodiments of this application, in S2, selecting the specification source configuration includes: Candidate configuration files are scored based on whether the file is located in the output directory, whether the file name matches the source configuration naming pattern, and the complexity of parameter recording, and the highest-scoring file is determined as the standard source configuration.

[0029] In some embodiments of this application, in S2, the deterministic rule review includes at least five of the following categories: absolute path detection, duplicate parameter detection, unknown parameter detection, unique label duplicate detection, event count consistency detection, time window continuity detection, filename and time window matching detection, and path file existence detection.

[0030] In some embodiments of this application, in S3, the operational health assessment is calculated comprehensively based on at least three of the following dimensions: time step stability, solver iterative behavior, residual monotonicity, late deceleration characteristics, and abnormal termination markers, and outputs a numerical score and a status label.

[0031] In some embodiments of this application, in S4, the structured evidence storage layer includes at least the entities Project, Case, Artifact, Finding, Metric, TaskRun, TaskEvent, and Report, and each discovery is associated with at least evidence text or evidence line number.

[0032] In some embodiments of this application, in S4, the constrained narration, question-and-answer results, or research report generation process is limited to referencing only findings, metrics, artifacts, task events, or document fragments in the evidence bundle, and includes a cue echo filtering and rollback text mechanism.

[0033] In some embodiments of this application, in step S5, the key assertion verification includes: Extract key conclusions from the research report, retrieve evidence records corresponding to each key conclusion, generate verification results based on the degree of evidence matching, and output audit documents for traceability.

[0034] In some embodiments of this application, a structured intelligent archiving and review system for dynamic simulation engineering is also disclosed, comprising: The project scanning module is used to receive the target simulated project path, standardize the path, and build a file scanning tree; The case identification module is used to automatically identify one or more case root directories based on the file scanning tree, and to perform content-aware identification on candidate configuration files to distinguish between parameter files, XML configuration files and generated copies; The inventory building module is used to select the specification source configuration from the candidate configuration files for each case and build an inventory of cases containing the specification source configuration, generated copies, runtime artifacts and result artifacts. The configuration review module is used to perform deterministic rule reviews based on the source configuration of the specification and generate configuration findings with evidence line numbers and evidence text. The operation review module is used to read operation logs, statistics files and / or residual files, extract operation indicators and generate operation health assessment results; Structured evidence storage is used to write cases, artifacts, findings, metrics, and task events into the structured evidence storage layer; The evidence-constrained AI interaction module is used to construct evidence bundles based on the structured evidence storage layer, and generate constrained narration, question-and-answer results or research reports with the evidence bundles as input. The assertion verification and auditing module is used to verify key assertions in research reports and output audit logs.

[0035] This invention achieves the transformation of dynamic simulation engineering data from decentralized storage to unified governance by constructing a closed-loop technology chain covering engineering scanning, case identification, configuration review, operational evaluation, structured evidence organization, evidence-constrained question-and-answer generation, and report verification. On the one hand, automated processing significantly improves data management efficiency and reduces manual screening costs. On the other hand, by establishing a structured evidence storage layer that includes evidence line numbers, evidence text, and multi-dimensional entities, the traceability and verifiability of research conclusions are ensured. At the same time, using structured evidence as the boundary constraint for the AI ​​generation process effectively avoids the risk of answers deviating from engineering facts, ultimately forming a research report with complete evidence backtracking capabilities, comprehensively improving the data governance efficiency, conclusion credibility, and audit reliability of simulation engineering.

[0036] Specifically, the technical approach of this invention is as follows: 1. Project Directory and Data Collection: The process begins by receiving the root path of the target dynamics simulation project. The input for this step is the project directory and its hierarchical structure. The processing includes path normalization, scan depth control, and initial file type screening. The output is the project scan tree and a list of basic workpieces. The scanning phase focuses on identifying configuration files, log files, statistical files, result files, drawing files, and their directory relationships, providing a unified entry point for subsequent case identification and evidence organization.

[0037] 2. Standard Input Identification and Case List Construction: Based on the scan tree, one or more candidate case directories are discovered, and the configuration type is identified according to the file content. For configuration files using parameter key-value syntax, their scope structure and key parameters are extracted; for configuration files using XML structure, their key nodes and fields are identified. Then, standard input files are selected according to criteria such as whether they are located in the output directory, file naming pattern, and record complexity to construct the case list. The output of this step is a standard input list, a run artifact list, a result artifact list, and a case-level evidence index. The engineering scan and case list construction process is as follows: Figure 2 As shown.

[0038] 3. Configuration Review and Operational Assessment: A deterministic rule review is performed on the standard input files, focusing on absolute paths, duplicate definitions, unknown fields, consistency of event counts, time window matching relationships, and the existence of external files. For runtime artifacts, time steps, solver behavior, residual trends, and abnormal termination markers are extracted to form an operational health score and status label. The output of this step includes not only findings and indicators but also traceable evidence line numbers, evidence text, and their severity levels.

[0039] 4. Structured Evidence Layer Construction: Projects, cases, artifacts, findings, metrics, tasks, events, and reporting objects are uniformly registered in the structured evidence layer. This evidence layer serves as the sole source of facts for subsequent Q&A, reporting, and verification, maintaining a stable mapping between configuration reviews, operational assessments, results summaries, and documentation. This step compresses heterogeneous information previously scattered across directories, logs, graphs, and documents into a unified set of structured facts.

[0040] 5. Evidence-Constrained Question-and-Answer and Report Generation: A constrained question-and-answer and report generation chain is constructed, bounded by a structured evidence layer and document knowledge. This chain can output dialogical analysis responses for the current task, provide general questions and answers for the entire project, and offer specific questions and answers for different simulation types such as ASPECT and Badlands. In the report generation phase, the system organizes the research report content according to preset chapter templates. The key technical effect of this step is that all questions and answers and report content are bounded by registered evidence, thus avoiding output that deviates from engineering facts. The structured evidence layer supports the question-and-answer and report generation process as follows: Figure 3 As shown.

[0041] 6. Report Verification and Audit Output: Key conclusions in the research report are extracted and verified to check for direct evidence or documentary support, and corresponding audit records are output. This step ensures the final report has evidence traceability and provides clear clues for manual review, thus completing the closed-loop process from engineering data input to audit result output. The research report verification and audit output process is as follows: Figure 4 As shown.

[0042] The following verification is based on a specific embodiment.

[0043] Step 1, Case Selection and Significance: The software analysis object selected in this embodiment is the Badlands case test-100-1-1-1e-6-2-0.36-15000000-0.0. The reason for choosing this case is that it has standardized input files, runtime logs, generated copies, and a large number of result artifacts, which can fully cover the entire process of this invention from engineering scanning, configuration review, runtime evaluation to result archiving.

[0044] Specifically, in this case, the system identified one standard input file (input.xml), one generated copy (output / input.xml), one runtime artifact (running.log), and 80 result artifacts. This structure is highly suitable as a case study because manual review requires researchers to not only determine whether input.xml or output / input.xml should be the entry point for analysis, but also to sift through a large number of HDF5, XMF, and log files to identify truly representative sources of evidence—a process that is inefficient and prone to omissions.

[0045] Step 2, Configure the review results: In this embodiment, the system recognizes input.xml as a standardized input file and performs deterministic rule review on it, ultimately outputting four warnings. Unlike manually browsing XML segment by segment, the system directly binds the exception category, evidence line number, and evidence text, enabling the centralized display of exceptions that were originally scattered in different configuration locations.

[0046] The main configuration anomalies detected by the system include: 1. absolute_path: Absolute paths were detected near lines 52, 68, 76, and 84, indicating that this case depends on a specific cluster directory and has poor portability when migrated to other workspaces; 2. path_not_found: The absolute path mentioned above cannot be directly located in the current workspace context, which means that researchers may still be unable to reproduce the experiment if they only copy the case directory; 3. Duplicate_unique_tag: Duplicate wbase tags were identified in waveglobal near lines 178 and 191, indicating a risk of opaque parsing priority. 4. filename_time_mismatch: The tectonic event settings detect a mismatch between the time period implied by the filename and the declared time window, indicating a possible incorrect correspondence between the constructed displacement file and the event interval.

[0047] If the above checks are performed manually, researchers need to repeatedly check the scattered sections such as sea level curves, tectonic events, and wave parameters in the same input.xml file, and manually compare the relationship between file names and time windows, which is significantly less efficient in locating the data.

[0048] Step 3, Run the evaluation and archive the results: During the operational analysis phase, the system used running.log as the primary source of evidence, parsing 12,016 lines of log data to extract 6,000 step records and generating a health conclusion for the watch status, with a comprehensive score of 20. The core evidence was: a median early step time of 21.45 seconds, a median late step time of 30.65 seconds, and a slowdown_ratio of 1.4289, indicating a significant late-stage deceleration.

[0049] Compared to manual log reading, the system eliminates the need for researchers to manually retrieve time information and status changes segment by segment. Instead, it directly outputs operational status labels, key metrics, and evidence summaries. Furthermore, the system registers 80 outcome artifacts for this case, including flow.time. .hdf5, sed.time .hdf5, tin.time The .hdf5 file and the corresponding XMF / XDMF index file provide a unified result foundation for subsequent result extraction, question answering, and report generation.

[0050] Step 4, Comprehensive Analysis of Significance: As can be seen from the specific examples above, this invention is not merely a catalog display, but can simultaneously accomplish the following tasks within a single case: automatically selecting standardized inputs, differentiating generated copies, quickly locating configuration anomalies, quantifying operational risks, registering result artifacts, and generating a unified evidence set. For researchers, this means that they can obtain structured and traceable review results without having to manually read input.xml, manually verify absolute paths and time windows, and then browse through 12016 lines of running.log and 80 output artifacts line by line.

[0051] Therefore, this embodiment specifically demonstrates that the present invention can significantly improve the efficiency of anomaly identification, reduce the cost of manual screening in real Badlands cases, and organize configuration review, operation evaluation and result archiving into the same closed-loop technology chain.

[0052] In this application, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. In case of any inconsistency, the meaning set forth in this specification or derived from the content described herein shall prevail. Furthermore, the terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit the scope of this application.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A structured intelligent archiving and review method for dynamic simulation engineering, characterized in that, Includes the following steps: S1. Receive the target simulation project path, normalize the path and construct a file scanning tree, automatically identify one or more case root directories based on the file scanning tree, and perform content-aware recognition on candidate configuration files to distinguish parameter files, XML configuration files and generated copies. S2. For each case, select the normative source configuration from the candidate configuration files, construct a case list containing the normative source configuration, generated copies, runtime artifacts and result artifacts, perform deterministic rule review based on the normative source configuration, and generate configuration discovery with evidence line numbers and evidence text. S3. Read the operation log, statistics file and / or residual file, extract the operation indicators and generate the operation health assessment results; S4. Write cases, artifacts, findings, indicators, and task events into the structured evidence storage layer, construct evidence bundles based on the structured evidence storage layer, and generate constrained narration, question-and-answer results, or research reports using the evidence bundles as input. S5. Perform validation on key assertions in the research report and output audit logs.

2. The structured intelligent archiving and review method for dynamic simulation engineering according to claim 1, characterized in that, In S1, content-aware recognition includes: When a candidate configuration file contains parameter key-value syntax and a hierarchical structure, it is identified as a first-class parameter configuration file. When a candidate configuration file is in XML format and its root node meets the preset simulation configuration identifier, it is identified as a second type of structured configuration file.

3. The structured intelligent archiving and review method for dynamic simulation engineering according to claim 2, characterized in that, In step S2, the selection of the standard source configuration includes: Candidate configuration files are scored based on whether the file is located in the output directory, whether the file name matches the source configuration naming pattern, and the complexity of parameter recording, and the highest-scoring file is determined as the standard source configuration.

4. The structured intelligent archiving and review method for dynamic simulation engineering according to claim 3, characterized in that, In S2, the deterministic rule review includes at least five of the following categories: absolute path detection, duplicate parameter detection, unknown parameter detection, unique label duplicate detection, event count consistency detection, time window continuity detection, filename and time window matching detection, and path file existence detection.

5. The structured intelligent archiving and review method for dynamic simulation engineering according to claim 4, characterized in that, In S3, the operational health assessment is calculated comprehensively based on at least three of the following dimensions: time step stability, solver iterative behavior, residual monotonicity, late deceleration characteristics, and abnormal termination markers, and outputs a numerical score and status label.

6. The structured intelligent archiving and review method for dynamic simulation engineering according to claim 5, characterized in that, In S4, the structured evidence storage layer includes at least the entities Project, Case, Artifact, Finding, Metric, TaskRun, TaskEvent, and Report, and each discovery is associated with at least evidence text or evidence line number.

7. The structured intelligent archiving and review method for dynamic simulation engineering according to claim 6, characterized in that, In S4, the process of generating constrained narration, question-and-answer results, or research reports is limited to referencing only findings, indicators, artifacts, task events, or document fragments in the evidence bundle, and includes a cue echo filtering and rollback text mechanism.

8. The structured intelligent archiving and review method for dynamic simulation engineering according to claim 7, characterized in that, In S5, the key assertion verification includes: Extract key conclusions from the research report, retrieve evidence records corresponding to each key conclusion, generate verification results based on the degree of evidence matching, and output audit documents for traceability.

9. A structured intelligent archiving and review system for dynamic simulation engineering, characterized in that, include: The project scanning module is used to receive the target simulated project path, standardize the path, and construct a file scanning tree; The case identification module is used to automatically identify one or more case root directories based on the file scanning tree, and to perform content-aware identification on candidate configuration files to distinguish between parameter files, XML configuration files and generated copies; The inventory building module is used to select the specification source configuration from the candidate configuration files for each case and build an inventory of cases containing the specification source configuration, generated copies, runtime artifacts and result artifacts. The configuration review module is used to perform deterministic rule reviews based on the source configuration of the specification and generate configuration findings with evidence line numbers and evidence text. The operation review module is used to read operation logs, statistics files and / or residual files, extract operation indicators and generate operation health assessment results; Structured evidence storage is used to write cases, artifacts, findings, metrics, and task events into the structured evidence storage layer; The evidence-constrained AI interaction module is used to construct evidence bundles based on the structured evidence storage layer, and generate constrained narration, question-and-answer results or research reports with the evidence bundles as input. The assertion verification and auditing module is used to verify key assertions in research reports and output audit logs.

Images