Intelligent diagnosis and compliance automatic verification system and method for nuclear power plant maintenance procedures

Abstract

The application belongs to the technical field of digitalization and intelligent application in nuclear power industry, and provides an intelligent diagnosis and compliance automatic checking system and method for maintenance regulations of a nuclear power plant, the system comprising a rule base management module, an intelligent analysis module, a rule engine module and a user interaction module, the method creating and maintaining a machine-readable rule base for checking maintenance regulations of a nuclear power plant, acquiring and analyzing a maintenance regulation file to be checked, generating a structured regulation object model, matching and evaluating the regulation object model with rules in the rule base, generating a diagnosis result set, displaying the diagnosis result set to a user, and providing interactive modification guidance.The application performs automatic and real-time compliance checking and logic diagnosis on maintenance regulations of a nuclear power plant, so as to replace most repetitive and rule-based manual review work with machine automatic checking.

Description

Technical Field

[0001] This application belongs to the field of digital and intelligent application technology in the nuclear power industry, and in particular relates to an intelligent diagnosis and automatic compliance verification system and method for nuclear power plant maintenance procedures. Background Technology

[0002] Currently, the drafting, review, and revision processes for nuclear power plant maintenance procedures (including mechanical, instrumentation, and electrical procedures) primarily rely on the individual knowledge and manual inspection of procedure engineers, professional supervisors, and review experts. However, this highly human-dependent model has the following inherent flaws: (1) Efficiency bottleneck and high cost: Manual review of technical procedures that are often hundreds of pages long requires a lot of time and manpower, which is the main bottleneck in the life cycle management of procedures.

[0003] (2) The coverage and consistency of the inspection are difficult to guarantee: manual review is susceptible to fatigue, negligence and subjective judgment, making it difficult to ensure that every rule and every step is checked without omission. Different reviewers have different understandings of the regulations and standard provisions, resulting in inconsistent quality of procedures.

[0004] (3) Difficulty in knowledge transfer and retention: The review experience of senior experts is valuable tacit knowledge that is difficult to quantify, retain and replicate effectively. With the turnover of personnel, this experience may be lost, leading to a decline in overall review capabilities.

[0005] (4) Complex logical errors are difficult to detect: Complex logical errors that may exist in the procedures, such as infinite loops in jump steps, logical contradictions between parent and child steps, and lack of termination conditions in loop execution, are difficult to detect by manual static review, which poses a serious safety hazard to on-site execution.

[0006] (5) Delayed response to regulatory changes: When national nuclear safety regulations, industry standards or power plant technical specifications are updated, it is difficult for manual personnel to quickly locate all affected existing procedures and make batch modifications, which poses a compliance risk.

[0007] While some document proofreading software or format checking tools exist in the existing technology, they generally lack expertise in the nuclear power field and cannot understand the specific semantics, context, and stringent safety rules of maintenance procedures. Therefore, there is an urgent need in this field for an intelligent auxiliary system that deeply integrates nuclear power expertise, possesses reasoning capabilities, and can be seamlessly integrated into the procedure preparation workflow to resolve the aforementioned core contradictions of "quality, efficiency, consistency, and safety." Summary of the Invention

[0008] The purpose of this application is to overcome the shortcomings of the prior art and provide an intelligent diagnostic and automatic compliance verification system and method for nuclear power plant maintenance procedures. This system automates and performs real-time compliance verification and logical diagnosis of nuclear power plant maintenance procedures, replacing most of the repetitive and rule-based manual review work with automatic machine verification.

[0009] To achieve the above objectives, this application provides the following technical solution: Firstly, this application provides an intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures, comprising: The rule base management module is used to store and manage machine-readable rules for verifying nuclear power plant maintenance procedures; The intelligent parsing module is used to receive the maintenance procedures to be verified and parse them into a structured procedure object model. The rules engine module is used to match and evaluate the rules in the rule base management module with the procedure object model and generate diagnostic results. The user interaction module is used to display diagnostic results to users and provide interactive modification guidance.

[0010] As an implementable approach, the rule base management module is also used to add, delete, modify, and query validation rules through a graphical interface, as well as to manage rule versions and control their enabled / disabled status.

[0011] As an implementable approach, the intelligent parsing module is also used to receive XML or JSON-based structured procedure data from the structured procedure editor as input and parse it into a procedure object model containing procedure cover objects, chapter title objects, and operation step objects.

[0012] As an implementable approach, the rule engine module is also used to execute a "match-evaluate-execute" inference loop, performing conditional truth value judgments on successfully matched rules and triggering diagnostic result generation actions for rules that meet the conditions.

[0013] As an implementable approach, the user interaction module includes a real-time diagnostic mode and a mandatory verification mode. The real-time diagnostic mode triggers partial verification during the user's editing process and presents the results. The mandatory verification mode triggers full verification when the user submits or publishes the procedure, and intercepts operations for errors and violations, while generating and exporting a detailed diagnostic report.

[0014] Secondly, this application provides a method for intelligent diagnosis and automatic compliance verification of nuclear power plant maintenance procedures, including: S1. Create and maintain a machine-readable rule base for verifying nuclear power plant maintenance procedures; S2. Obtain and parse the maintenance procedure document to be verified, and generate a structured procedure object model; S3. Match and evaluate the procedure object model with the rules in the rule base to generate a set of diagnostic results; S4. Display the diagnostic results to the user and provide interactive modification guidance.

[0015] As an feasible approach, in S1, the verification rules are written using a rule description language, which supports version iteration management and on-demand enable / disable control.

[0016] As an implementable approach, in S2, the structured elements of the procedure object model include operation step objects, which contain step IDs, description text, type identifiers, associated input areas, and jump target instructions.

[0017] As an implementable approach, S3 verifies the existence of the target step ID of the jump instruction, checks the consistency of parent-child step state inheritance, and determines whether the loop execution block has an explicitly defined termination condition.

[0018] As an feasible approach, in S4, in real-time diagnostic mode, verification is triggered by user editing operations, and feedback includes highlighting by issue level, displaying violation basis and repair suggestions on hover, and providing one-click correction for formatting issues; in forced verification mode, verification covers all content of the procedure, and if there are errors of the required level, the publishing process is blocked, and a categorized diagnostic report that can be exported is generated.

[0019] Compared with existing technologies, the intelligent diagnostic and automatic compliance verification system and method for nuclear power plant maintenance procedures provided in this application have the following advantages: (1) Leap in quality and safety: Through 100% automated inspection, defects in procedures caused by human negligence are fundamentally eliminated, significantly improving the technical accuracy and compliance of procedures and providing strong technical support for nuclear safety.

[0020] (2) Revolutionary improvement in efficiency: It will free engineers from tedious review work and is expected to improve the overall efficiency of procedure preparation and approval by 50% to 70%, and significantly shorten the procedure publication cycle.

[0021] (3) Standardization and consistency: By verifying all procedures in the plant through a unified rule base, the standardization of maintenance procedures has been effectively promoted, ensuring a high degree of consistency in the output quality of different procedures and different compilers.

[0022] (4) Knowledge assetization and inheritance: Valuable expert experience and regulatory standards are solidified into the system, forming digital assets for sustainable operation and iteration of enterprises, effectively solving the problem of knowledge inheritance.

[0023] (5) User experience optimization: The real-time diagnosis and intelligent prompt function is like an "AI assistant", providing immersive auxiliary writing experience for writers and reducing the work threshold and pressure. Attached Figure Description

[0024] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the technical description will be briefly introduced below.

[0025] Figure 1 A schematic diagram of the intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures provided in this application; Figure 2 A flowchart of the intelligent diagnosis and automatic compliance verification method for nuclear power plant maintenance procedures provided in this application. Detailed Implementation

[0026] The following detailed description provides further details on specific implementation methods.

[0027] like Figure 1 As shown, this application provides an intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures, comprising: The rule base management module is used to store and manage machine-readable rules for verifying nuclear power plant maintenance procedures; The intelligent parsing module is used to receive the maintenance procedures to be verified and parse them into a structured procedure object model. The rules engine module is used to match and evaluate the rules in the rule base management module with the procedure object model and generate diagnostic results. The user interaction module is used to display diagnostic results to users and provide interactive modification guidance.

[0028] The rules stored in the rule base management module include at least: compliance rule base, anti-human factor rule base, logic rule base, and integrity rule base; the rule base management module provides a graphical interface for authorized administrators to add, delete, modify, query, and enable / disable rules.

[0029] The intelligent parsing module receives XML or JSON-based structured procedure data from the structured procedure editor as input and parses it into a procedure object model containing a procedure cover object, chapter title object, and operation step object. The operation step object contains step ID, description text, type, associated input area, and jump instruction information.

[0030] The working mechanism of the rule engine module includes pattern matching, condition evaluation, and action execution. Pattern matching is used to find rules related to the procedure object model, condition evaluation is used to determine whether the rule conditions are true, and action execution is used to generate a diagnostic result object containing the problem level, problem description, problem location, and remediation suggestions when the conditions are true.

[0031] The user interaction module includes a real-time diagnostic mode and a forced verification mode. The real-time diagnostic mode triggers partial verification during the user's editing process and presents the results through highlighted prompts, hovering windows, and one-click repair. The forced verification mode triggers full verification when the user submits or publishes a procedure, and intercepts operations for errors of a certain level, while generating and exporting a detailed diagnostic report.

[0032] Furthermore, this application also provides a method for intelligent diagnosis and automatic compliance verification of nuclear power plant maintenance procedures, such as... Figure 2 As shown, it includes the following steps: S1. Rule Pre-setting Steps: Create and maintain a machine-readable rule base for verifying nuclear power plant maintenance procedures; S2. Procedure parsing steps: Obtain and parse the maintenance procedure document to be verified, and generate a structured procedure object model; S3, Rule Reasoning Steps: Match and evaluate the procedure object model with the rules in the rule base to generate a set of diagnostic results; S4. Result Feedback Steps: Display the diagnostic results set to the user and provide interactive modification guidance.

[0033] In the rule pre-setting step, the rule base includes at least compliance rules based on the conversion of nuclear safety regulations, operating technical specifications and industry standards, human error prevention rules based on the application of human error prevention tools, logical rules for constraining the logical relationships between procedure steps, and integrity rules for verifying the essential elements of procedure documents.

[0034] The rule reasoning steps specifically include: using a rule engine to perform pattern matching to find relevant rules, evaluating whether the rule conditions are true, and executing defined actions when the conditions are true to generate diagnostic results containing the problem level, problem location, and remediation suggestions.

[0035] The results feedback steps include real-time diagnostic feedback: during the user's editing process, a local verification of the currently edited content is automatically triggered, and the diagnostic results are displayed in real time in the editing interface in the form of highlights and hovering windows.

[0036] The feedback process includes mandatory verification: when a user submits an application for approval or publishes it for effect, a full verification of the entire procedure is automatically triggered. If an error-level diagnostic result is found, the current operation is intercepted and a detailed diagnostic report is displayed. The user must resolve all errors before continuing.

[0037] Example This embodiment provides an intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures. It replaces most of the repetitive and rule-based manual review work with automated machine verification. Its specific objective is: Real-time intervention: Provide real-time diagnosis during the procedure drafting stage, allowing for "writing and checking" to identify problems in advance.

[0038] Mandatory compliance: Conduct mandatory verification before the procedures are published to prevent procedures with serious defects from proceeding to the next stage.

[0039] Knowledge solidification: Transforming regulations, standards, and expert experience into executable machine rules, thereby achieving the accumulation and reuse of knowledge.

[0040] Logical safeguards: Ensure that the logical relationships between procedural steps are rigorous and correct, and eliminate structural errors.

[0041] This system is integrated into the digital maintenance procedure management platform of nuclear power plants, and its core architecture consists of the following four modules working together: (1) Rule base management module It serves as the "brain" and knowledge center of the system, responsible for the storage, management, and maintenance of all verification rules.

[0042] Rule categories: Compliance rule base: Stores machine-readable rules converted from the Nuclear Safety Regulations, Operating Technical Specifications, and industry standards (such as NB / T standards). For example: "Steps involving radioactive operations must include the keyword 'dose monitoring' in the step description."

[0043] Human error prevention rule base: Stores application rules for tools to prevent human error. For example: "After the execution of critical equipment operation procedures, an 'independent verification' step must be inserted"; "If the procedure includes numerical settings, 'three-stage communication' must be used for restatement and confirmation".

[0044] Logical rule base: Stores the logical constraint rules of the procedure structure. For example: "The target step ID of the 'jump to' instruction must actually exist"; "If the state of a step marked as a 'parent step' is 'N / A (not applicable)', then all its 'child steps' should automatically inherit the 'N / A' state"; "Loop execution blocks must explicitly define the loop termination condition".

[0045] Integrity rule base: Stores rules for essential elements of procedure documents. For example: "All procedures must include a 'Risk Analysis' section"; "Procedures involving special operations (such as high-altitude work, hot work) must list the relevant permits in the prerequisites"; "All tools and spare parts referenced in the procedures must be clearly listed in the corresponding sections".

[0046] Rule Expression and Storage: Rules can be written using industry-standard rule description languages ​​and stored as database records, supporting version management. The system provides a graphical interface that allows authorized administrators to add, delete, modify, and query rules, and supports enabling / disabling rules.

[0047] (2) Intelligent parsing module It is responsible for understanding and structuring the content of the procedure to be verified.

[0048] Input: Receives standard structured data (such as XML- or JSON-based procedure object models) from the upstream “Structured Procedure Editor”.

[0049] Parsing process: This module parses the procedure file into a series of objects that can be processed by the program, such as: 'procedure cover object', 'chapter title object', 'operation step object' (including step ID, description text, type, associated input area, jump instructions, etc.), 'table object', 'warning paragraph object', etc.

[0050] Output: A complete, structured procedure object model that clearly expresses the hierarchical structure of the procedure and the relationships between its elements, providing input for the rule engine.

[0051] (3) Rule Engine Module It is responsible for performing reasoning and judgment.

[0052] Working mechanism: It adopts a reasoning loop of "matching-evaluation-execution".

[0053] 1) Pattern matching: The engine matches the procedure object model output by the intelligent parsing module with the rules in the rule base management module to find all possible rules that may be triggered.

[0054] 2) Condition Evaluation: For a matched rule, the engine evaluates whether its condition part is true. For example, if the rule condition is "the step description contains 'closed' and there is no 'high voltage warning' before it", the engine will check the current step text and the paragraph preceding it.

[0055] 3) Execute the action: For rules whose conditions are true, execute the actions defined therein. A typical action is to generate a "diagnostic result" object, which includes: problem level (error, warning, tip), problem description, problem location (such as step ID), and repair suggestions.

[0056] (4) User interaction module Responsible for presenting diagnostic results to users in a user-friendly and efficient manner.

[0057] 1) Real-time diagnostic mode: Triggering timing: During the user's editing process, such as when entering text, inserting elements, or saving, the rule engine is automatically triggered to perform lightweight, localized validation.

[0058] Presentation format: Highlighting: In the editing interface, problematic elements (such as a step or a paragraph) will be highlighted with a specific background color (such as error - red, warning - yellow, hint - blue).

[0059] Hover window: When the user hovers the mouse over the highlighted area, a floating window is displayed, which details the cause of the problem, the violated rule, and specific modification suggestions.

[0060] One-click repair: For common issues such as formatting errors and numbering errors, a "one-click repair" button is provided. After the user confirms, the system will automatically perform the correction.

[0061] 2) Forced verification mode: Triggering timing: When a user performs critical operations such as "submit for approval" or "release and take effect".

[0062] Presentation format: Full scan: Perform a comprehensive rule check on the entire procedure.

[0063] Interception and Reporting: If an "Error" level violation is detected, the system will block the operation and display a detailed diagnostic report. The report is categorized by problem type and severity level, and can directly pinpoint the problem location. The user must resolve all errors before proceeding.

[0064] Report Export: Supports exporting diagnostic reports as PDF or Word documents for review records or communication.

[0065] Based on the system in this embodiment, this implementation provides a method for intelligent diagnosis and automatic compliance verification of nuclear power plant maintenance procedures, including: S1. Rule Pre-setting Steps: Create and maintain a set of verification rules, including compliance rules, anti-human error rules, logical rules, and integrity rules, through the rule base management module; S2. Procedure parsing steps: Obtain the structured maintenance procedure file through the intelligent parsing module, and parse it to generate a procedure object model; S3. Rule reasoning steps: The rule engine module matches and evaluates the conditions of the procedure object model with the verification rule set to generate a set of diagnostic results. S4. Result Feedback Steps: The diagnostic results are fed back to the user through the user interaction module, and interactive modification guidance is provided.

[0066] The above description is only a specific embodiment of this application, but the protection scope of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application.

Claims

1. An intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures, characterized in that, include: The rule base management module is used to store and manage machine-readable rules for verifying nuclear power plant maintenance procedures; The intelligent parsing module is used to receive the maintenance procedures to be verified and parse them into a structured procedure object model. The rules engine module is used to match and evaluate the rules in the rule base management module with the procedure object model and generate diagnostic results. The user interaction module is used to display diagnostic results to users and provide interactive modification guidance.

2. The intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures according to claim 1, characterized in that, The rule base management module is also used to add, delete, modify, and query validation rules through a graphical interface, as well as to manage rule versions and control their enabled / disabled status.

3. The intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures according to claim 1, characterized in that, The intelligent parsing module is also used to receive XML or JSON-based structured procedure data from the structured procedure editor as input and parse it into a procedure object model containing procedure cover objects, chapter title objects, and operation step objects.

4. The intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures according to claim 1, characterized in that, The rules engine module is also used to execute the "match-evaluate-execute" inference loop, to perform conditional truth value judgments on successfully matched rules, and to trigger the diagnostic result generation action for rules that meet the conditions.

5. The intelligent diagnostic and automatic compliance verification system for nuclear power plant maintenance procedures according to claim 1, characterized in that, The user interaction module includes a real-time diagnostic mode and a mandatory verification mode. The real-time diagnostic mode triggers partial verification during the user's editing process and presents the results. The mandatory verification mode triggers full verification when the user submits or publishes a procedure, and intercepts operations for errors and violations, while generating and exporting a detailed diagnostic report.

6. A method for intelligent diagnosis and automatic compliance verification of nuclear power plant maintenance procedures, characterized in that, include: S1. Create and maintain a machine-readable rule base for verifying nuclear power plant maintenance procedures; S2. Obtain and parse the maintenance procedure document to be verified, and generate a structured procedure object model; S3. Match and evaluate the procedure object model with the rules in the rule base to generate a set of diagnostic results; S4. Display the diagnostic results to the user and provide interactive modification guidance.

7. The intelligent diagnosis and automatic compliance verification method for nuclear power plant maintenance procedures according to claim 6, characterized in that, In S1, the verification rules are written using a rule description language, which supports version iteration management and on-demand enable / disable control.

8. The intelligent diagnosis and automatic compliance verification method for nuclear power plant maintenance procedures according to claim 6, characterized in that, In S2, the structured elements of the procedure object model include operation step objects, which contain step ID, description text, type identifier, associated input area, and jump target instruction.

9. The intelligent diagnosis and automatic compliance verification method for nuclear power plant maintenance procedures according to claim 6, characterized in that, In S3, the existence of the target step ID of the jump instruction is verified, the consistency of the inheritance of the parent and child step states is checked, and it is determined whether the loop execution block has a clearly defined termination condition.

10. The intelligent diagnosis and automatic compliance verification method for nuclear power plant maintenance procedures according to claim 6, characterized in that, In S4, in real-time diagnostic mode, verification is triggered by user editing operations, and feedback forms include highlighting by problem level, displaying violation basis and repair suggestions on hover, and providing one-click correction for formatting issues; In the mandatory verification mode, the verification covers the entire content of the procedure. If an error level problem is found, the release process will be blocked, and a categorized diagnostic report that can be exported will be generated.

Images