Nuclear power plant health problem identification method and system
By integrating problem assessment models and expert identification results, the problems of low efficiency and insufficient coverage in the identification of health issues in nuclear power plants have been solved, achieving accurate and efficient identification and management, and improving the foresight and effectiveness of nuclear power safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-08
- Publication Date
- 2026-03-31
AI Technical Summary
Existing technologies are inefficient and subjective in identifying health problems in nuclear power plants, making it difficult to achieve early detection and warning. Furthermore, they cannot identify data that the problem assessment model cannot cover, resulting in incomplete identification.
A method for identifying health problems in nuclear power plants is constructed. By combining the automatic identification results of the problem assessment model with the human identification results of technical experts, the final identification results are generated through data preprocessing, screening and integration, including attributes such as the code, category and risk level of the health problem.
It enables accurate and efficient identification of health problems in nuclear power plants, improving the foresight, accuracy, and effectiveness of nuclear power safety management.
Smart Images
Figure CN121766754A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of nuclear safety management, and in particular to a method and system for identifying health problems in nuclear power plants. Background Technology
[0002] To enhance the foresight, accuracy, and effectiveness of nuclear power safety management, it is necessary to comprehensively monitor the health status of nuclear power plants and accurately and efficiently identify health issues. Currently, the identification of nuclear power plant health issues relies either on post-event reports and human experience for judgment, such as qualitative analysis of past events, or on anomaly importance assessment (SDP) and other problem assessment models. However, manual identification suffers from several drawbacks due to the wide range and large volume of data on nuclear power plant health status: low efficiency, high subjectivity, and difficulty in early detection and warning of potential risks or "precursor" anomalies. Automated identification methods, limited by the constraints of the problem assessment models used, cannot identify health issues for data not covered by these models or for events of high public concern requiring special analysis. Summary of the Invention
[0003] The technical health problem to be solved by this application is to provide a method and system for identifying health problems in nuclear power plants, addressing the aforementioned technical deficiencies in existing technologies.
[0004] The technical solution adopted in this application to solve its technical health problem is: constructing a method for identifying health problems in nuclear power plants, including: Step S10: Collect health status data of the nuclear power plant; Step S20: Preprocess the health status data and send the preprocessed health status data into a pre-established problem assessment model to obtain the first identification result of the current health problem of the nuclear power plant from the problem assessment model; Step S30: Select data that meets the set conditions from the collected health status data, and receive the second identification result of the nuclear power plant’s current health problems determined by experts based on the selected data; Step S40: Based on the first identification result and the second identification result, determine the final identification result of the current health problems of the nuclear power plant.
[0005] Optionally, step S40 includes: Step S41: According to the preset problem list template, the first identification result is converted into a first problem list, and the second identification result is converted into a second problem list. The problem list template includes multiple problem attributes of health problems, and the problem attributes include: code, category, and risk level. Step S42: Determine whether there are health problems with the same code in the first problem list and the second problem list. If so, delete the record of the health problem with the same code in the first problem list and update the first problem list. Step S43: Merge the updated first problem list and the second problem list to obtain a third problem list, wherein the third problem list includes the codes, categories, risk levels, and sources of multiple health problems, and the sources include automatic identification and expert judgment. Step S44: Based on the third problem list, determine the final identification result of the current health problems of the nuclear power plant.
[0006] Optionally, the first identification result is structured data; the second identification result is semi-structured data or unstructured data. Step S41 includes: According to the preset problem list template, the content fields of multiple problem attributes of each health problem are extracted from the first identification result and filled into the problem list template to generate the first problem list; The data content of the second identification result is identified to obtain structured data. According to the preset problem list template, the content fields of multiple problem attributes of each health problem are extracted from the structured data and filled into the problem list template to generate the second problem list.
[0007] Optionally, the problem attributes may also include the relevant system / device identifier and description of the health problem; Step S44 includes: Step S441: Analyze the relevant system / device identifiers and descriptions of each health problem in the third problem list to determine whether there are at least two health problems with common causes or commonalities in the third problem list; Step S442: For at least two health problems that have a common cause or commonality, the at least two health problems are merged into a new health problem, and the code, category, related system / device identifier, and description of the new health problem are determined. The higher risk level among the risk levels corresponding to the at least two health problems is taken as the risk level of the new health problem. Step S443: Add the code, category, risk level, related system / device identifier, and description of the new health problem to the third problem list; Step S444: Based on the third problem list, determine the final identification result of the current health problems of the nuclear power plant.
[0008] Optionally, after step S443, the method further includes: The new health problem is marked as the parent problem, and the at least two health problems are marked as child problems of the parent problem.
[0009] Optionally, step S444 includes: The health issues in the third problem list are sorted to obtain a final problem list, which is then used as the final identification result of the current health issues of the nuclear power plant.
[0010] Optionally, the problem attribute further includes the discovery time; sorting the health problems in the third problem list includes: First, arrange the health issues in the third list of issues in descending order of risk level; For multiple health issues with the same risk level, they are arranged according to a preset category priority order or the order of discovery time from most recent to oldest. The categories include: safety category and management category, and the safety category has a higher priority than the management category.
[0011] Optionally, after step S40, the method further includes: Based on the pre-stored response settings information, for each health problem in the final problem list, the corresponding response mechanism information is determined. The response settings information includes multiple risk levels and the response mechanism information corresponding to each risk level. Output response mechanism information for each health issue so that the relevant personnel can respond. Update the treatment status, mitigation measures, and final results of each health problem to conduct closed-loop tracking of each health problem.
[0012] Optionally, after step S40, the method further includes: Based on the final problem list, the problem evaluation model is optimized.
[0013] The present invention also constructs a nuclear power plant health problem identification system, including a processor and a memory storing a computer program, wherein the processor implements the steps of the nuclear power plant health problem identification method described above when executing the computer program.
[0014] In the technical solution of this application, the automatic identification results of the problem assessment model and the human identification results of technical experts are integrated to determine the final identification results of the current health problems of the nuclear power plant. Therefore, the accurate and efficient identification of the health problems of the nuclear power plant is achieved, thereby improving the foresight, accuracy and effectiveness of nuclear power safety management. Attached Figure Description
[0015] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart of a method for identifying health problems in nuclear power plants according to an embodiment of the present invention. Detailed Implementation
[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0017] Figure 1 This is a flowchart of a method for identifying health problems in nuclear power plants according to an embodiment of the present invention. The method for identifying health problems in nuclear power plants according to this embodiment includes: Step S10: Collect health status data of the nuclear power plant; In this step, a unified data interface layer can be established to continuously or trigger-based collect health status data. This health status data includes baseline profile data, production data, and anomaly record data. Baseline profile data represents the nuclear power plant's actual condition and its compliance with regulations, standards, and licensing requirements. Production data comprises key data generated during the nuclear power unit's lifespan, such as safety production indicators, safety performance indicators, in-service inspection records, equipment reliability data, and radiation monitoring data. Anomaly record data records deviations from normal conditions, such as operational events, construction events, non-conformities, corrective action requirements (CARs), and status reports. Furthermore, the nuclear power plant's health status data can also include output data from a vulnerability identification module and an equipment status identification module. The vulnerability identification module's output data includes identified systemic and common problems; the equipment status identification module's output data includes potential or actual functional failures identified for specific equipment / components.
[0018] Step S20: Preprocess the health status data and send the preprocessed health status data into a pre-established problem assessment model to obtain the first identification result of the current health problem of the nuclear power plant from the problem assessment model; In this step, the problem assessment model may include, for example, an anomaly importance assessment (SDP) model, a precursor event analysis model, and a performance index analysis model. Since the collected health status data is structured data, after preprocessing, it can be directly input into the problem assessment model, which can output automatic identification results, such as the risk level and category of each identified health problem.
[0019] Regarding risk levels, they can be divided into four categories: A (major health problems), B (relatively serious health problems), C (general health problems), and D (other abnormalities). Category A further includes A1 (urgent response type) and A2 (high concern type). For A1 level health problems, the problem assessment model outputs a "red" abnormality, indicating an actual or potential high-risk consequence requiring immediate response action. For example, safety performance indicators are red, and SDP assessment results are red; for A2 level health issues, the problem assessment model outputs a specific "white" anomaly, but the recurrence rate is very high (e.g., more than 10 times in a group of plants within 5 years), and its trend changes need to be closely monitored; for B level health issues, the problem assessment model outputs a "yellow" anomaly, with high safety consequences and other factors, requiring the development of a plan and time-limited handling; for C level health issues, the problem assessment model outputs a "white" anomaly, with certain safety consequences or attention value, and is included in routine tracking management; for D level health issues, the problem assessment model outputs a "green" anomaly, currently with no significant safety impact, requiring no special response, but basic corrective actions must still be implemented according to relevant management requirements to prevent the problem from accumulating or escalating to a higher level.
[0020] Step S30: Select data that meets the set conditions from the collected health status data, and receive the second identification result of the nuclear power plant’s current health problems determined by experts based on the selected data; In this step, for data related to the following situations, an expert human judgment process can be initiated: data that cannot be covered by the problem assessment model (such as some experience feedback indicators and safety culture indicators); problems discovered through the vulnerability identification module and equipment status identification module that have not yet been modeled by the problem assessment model; events that require special judgment due to high social attention; and situations where there is controversy over the judgment results of the problem assessment model or a comprehensive judgment is required.
[0021] When filtering the collected health status data, the filtering can be based on the prompts output by the problem assessment model (such as prompts indicating that the model cannot cover certain data). Alternatively, relevant personnel can select the following data from the collected data: data that the problem assessment model has not yet modeled, data related to events of high social concern, and data for which there is controversy regarding the output results of the problem assessment model.
[0022] For the selected data, experts will consult and make judgments based on established criteria such as the "Classification Method for Abnormal Events and Unit Status in Nuclear Power Plants" to generate a second identification result for the current health problems of the nuclear power plant. This result may specifically include the risk level of the health problem (e.g., A1, A2, B, C, D), category labels (e.g., "equipment-related," "management-related," "common problems," etc.), detailed descriptions (which may include text descriptions, root cause analyses, etc.), relevant system / equipment identifiers (e.g., system / equipment functional location, unit number, etc.), problem discovery time, expert signatures, and other information. Furthermore, this second identification result can be entered into the system by filling out a standard form or by uploading a text file.
[0023] Step S40: Based on the first identification result and the second identification result, determine the final identification result of the current health problems of the nuclear power plant.
[0024] In this step, the identification results from the two methods are integrated to generate the final identification result of the current health problem. The final identification result includes the identified multiple health problems and the risk level and category of each health problem.
[0025] The technical solution of this embodiment provides a systematic, standardized, and intelligent method for identifying health problems in nuclear power plants. This method integrates the automatic identification results of the problem assessment model with the human identification results of technical experts to determine the final identification result of the current health problems in the nuclear power plant. Therefore, it achieves accurate and efficient identification of health problems in nuclear power plants, thereby improving the foresight, accuracy, and effectiveness of nuclear power safety management.
[0026] Further, in an optional embodiment, step S40 includes: Step S41: According to the preset problem list template, the first identification result is converted into a first problem list, and the second identification result is converted into a second problem list. The problem list template includes multiple problem attributes of health problems, and the problem attributes include: code, category, and risk level. In this step, the identification results from both methods are first standardized. The standardized first and second problem lists have unified problem attributes, specifically including: the health problem's coded ID (a unique identifier, which can be generated based on the health problem's description, related system / equipment identification, and discovery time according to preset coding rules), risk level (standardized to one of A1, A2, B, C, or D), category (standardized to a preset category label, such as equipment, management, or common problems), source (standardized to automatic identification or expert judgment), description (problem details), related system / equipment identification, discovery time, etc.
[0027] Step S42: Determine whether there are health problems with the same code in the first problem list and the second problem list. If so, delete the record of the health problem with the same code in the first problem list and update the first problem list. In this step, when the risk level or category of the same health problem (identified by a unique identifier) differs between two identification results, arbitration is conducted according to a pre-defined conflict arbitration mechanism (expert judgment takes precedence). This is because the problem assessment module's output may ignore relevant information, while experts make judgments based on a more in-depth risk assessment, which is more comprehensive and authoritative. Therefore, once experts participate in the judgment, their results override the automated identification results. For example, for a certain health problem, the risk level in the first identification result is A1 (red); the risk level in the second identification result is B, then the final risk level is B. As another example, for a certain health problem, the category in the first identification result is equipment-related, and the category in the second identification result is management-related, then the final category is management-related.
[0028] Step S43: Merge the updated first problem list and the second problem list to obtain a third problem list, wherein the third problem list includes the code, category, risk level, and source of the health problem, and the source includes automatic identification and expert judgment. Step S44: Based on the third problem list, determine the final identification result of the current health problems of the nuclear power plant.
[0029] In this embodiment, the identification results obtained from the two identification methods are first standardized to obtain a first problem list and a second problem list. Then, if the risk levels or categories of the same health problem differ between the two problem lists, the risk level or category of the health problem is arbitrated according to a preset conflict arbitration mechanism (expert judgment takes priority). Finally, the two arbitrated problem lists are merged.
[0030] Further, in an optional embodiment, the first identification result is structured data; the second identification result is semi-structured data or unstructured data, therefore, step S41 includes: According to the preset problem list template, the content fields of multiple problem attributes of each health problem are extracted from the first identification result and filled into the problem list template to generate the first problem list; The data content of the second identification result is identified to obtain structured data. According to the preset problem list template, the content fields of multiple problem attributes of each health problem are extracted from the structured data and filled into the problem list template to generate the second problem list.
[0031] In this embodiment, since the first identification result is structured data (e.g., risk level, category, etc.), the content fields of the corresponding question attributes can be directly mapped to the corresponding positions in the question list template according to the question list template. The second identification result, however, is generally obtained through form entry or text upload, and is semi-structured or unstructured data. Therefore, it is necessary to first perform content recognition to extract key information (such as risk level, category), and then directly map the content fields of the corresponding question attributes to the corresponding positions in the question list template according to the question list template.
[0032] Furthermore, in an optional embodiment, the problem attributes in the problem list template also include the relevant system / device identifiers and descriptions of the health problems. Therefore, the generated first and second problem lists also include the relevant system / device identifiers and descriptions of each health problem. Moreover, step S44 specifically includes: Step S441: Analyze the relevant system / device identifiers and descriptions of each health problem in the third problem list to determine whether there are at least two health problems with common causes or commonalities in the third problem list. For example, feature matching based on category, relevant system / device, and semantic similarity of description can be used to determine whether there are at least two health problems with common causes or commonalities in the third problem list. Step S442: For at least two health problems that have a common cause or commonality, the at least two health problems are merged into a new health problem, and the code, category, related system / device identifier, and description of the new health problem are determined. The higher risk level among the risk levels corresponding to the at least two health problems is taken as the risk level of the new health problem. In this step, for example, for the following two health issues in the third problem list: leakage from pump A and leakage from pump B, both of which are rooted in a defective seal design, these two health issues can be merged into a new health issue. When determining the risk level of this new health issue, the one with the higher risk level between the two health issues, leakage from pump A and leakage from pump B, should be selected.
[0033] Step S443: Add the code, category, risk level, related system / device identifier, and description of the new health problem to the third problem list; Step S444: Based on the third problem list, determine the final identification result of the current health problems of the nuclear power plant.
[0034] In this embodiment, clustering of health problems of the same type (common cause or commonality) in the third problem list is achieved.
[0035] Furthermore, following step S443, the method further includes: marking the new health problem as a parent problem, and marking the at least two health problems as child problems of the parent problem. This creates a "parent-child" structure in the third problem list.
[0036] Further, in an optional embodiment, step S444 includes: sorting the health issues in the third problem list to obtain a final problem list, and using the final problem list as the final identification result of the current health issues of the nuclear power plant. In this embodiment, sorting the health issues in the third problem list displays the importance and urgency of each health issue.
[0037] Furthermore, the issue attribute also includes the discovery time, and when sorting the various health issues in the third issue list, the sorting can be performed as follows: First, arrange the health issues in the third list of issues in descending order of risk level; For multiple health issues with the same risk level, they are arranged according to a preset category priority order or the order of discovery time from most recent to oldest. The categories include: safety category and management category, and the safety category has a higher priority than the management category.
[0038] In this embodiment, for each health issue in the third issue list, they are first sorted by risk level, and then sorted by category priority or discovery time from recent to distant. This helps with resource allocation and ensures that high-risk issues are handled first.
[0039] Furthermore, in an optional embodiment, after step S40, the method further includes: Based on the pre-stored response settings information, for each health problem in the final problem list, the corresponding response mechanism information is determined. The response settings information includes multiple risk levels and the response mechanism information corresponding to each risk level. Output response mechanism information for each health issue so that the relevant personnel can respond. Update the treatment status, mitigation measures, and final results of each health problem to conduct closed-loop tracking of each health problem.
[0040] In this embodiment, the response settings can be as follows: For A1-level health issues, the corresponding response mechanism is to immediately activate emergency response or special handling procedures to ensure rapid intervention for high-risk matters; for A2 or B-level health issues, the corresponding response mechanism is to include them in key work plans, clarify the responsible parties, rectification measures, and completion deadlines, and implement time-limited closed-loop management; for C-level health issues, the corresponding response mechanism is to include them in daily monitoring mechanisms and track trends through regular reviews and reports; and for C-level health issues, the corresponding response mechanism is to conduct necessary experience feedback and basic rectification, and provide timely feedback on the processing results in the system. Once the final issue list is obtained, the corresponding response mechanism's prompts are automatically triggered based on the risk level of each health issue in the final issue list, allowing relevant personnel to respond promptly. Furthermore, the processing status, mitigation measures, and final results of all health issues are updated in real time, thereby achieving full-process closed-loop tracking.
[0041] Furthermore, in an optional embodiment, after step S40, the method further includes: optimizing the problem assessment model based on the final problem list. In this embodiment, after obtaining the final problem list, relevant data is also synchronously fed back to the platform for continuous optimization of the problem assessment model, thereby improving the overall intelligent judgment capability.
[0042] The present invention also constructs a nuclear power plant health problem identification system, which includes a processor and a memory storing a computer program, and the processor implements the steps of the nuclear power plant health problem identification method described above when executing the computer program.
[0043] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. A method for identifying a health problem of a nuclear power plant, characterized by, The method comprises the following steps: Step S10, collecting health state data of the nuclear power plant; Step S20, preprocessing the health state data, inputting the preprocessed health state data into a pre-established problem assessment model, and obtaining a first identification result of a current health problem of the nuclear power plant from the problem assessment model; Step S30, screening data meeting a set condition from the collected health state data, and receiving a second identification result of the current health problem of the nuclear power plant determined by an expert according to the screened data; Step S40, determining a final identification result of the current health problem of the nuclear power plant according to the first identification result and the second identification result.
2. The nuclear power plant health problem identification method according to claim 1, characterized by, The step S40 comprises the following steps: Step S41, converting the first identification result into a first problem list and converting the second identification result into a second problem list according to a pre-set problem list template, wherein the problem list template comprises a plurality of problem attributes of health problems, and the problem attributes comprise coding, category, risk level; Step S42, judging whether there is a health problem with the same coding in the first problem list and the second problem list, if there is, deleting the record of the health problem with the same coding in the first problem list, and updating the first problem list; Step S43, merging the updated first problem list and the second problem list to obtain a third problem list, wherein the third problem list comprises coding, category, risk level, and source of a plurality of health problems, and the source comprises automatic identification and expert determination; Step S44, determining the final identification result of the current health problem of the nuclear power plant according to the third problem list.
3. The nuclear power plant health problem identification method according to claim 2, characterized by, The first identification result is structured data; The second identification result is semi-structured data or unstructured data; The step S41 comprises the following steps: According to the pre-set problem list template, extracting content fields of a plurality of problem attributes of each health problem from the first identification result, and filling the content fields into the problem list template to generate the first problem list; Identifying data content of the second identification result to obtain structured data, and extracting content fields of a plurality of problem attributes of each health problem from the structured data according to the pre-set problem list template, and filling the content fields into the problem list template to generate the second problem list.
4. The nuclear power plant health problem identification method according to claim 2, characterized by, The problem attributes further comprise relevant system / device identification and description of the health problem; The step S44 comprises the following steps: Step S441, analyzing the relevant system / device identification and description of each health problem in the third problem list, and judging whether there are at least two health problems with common causes or commonness in the third problem list; Step S442, for at least two health problems with common causes or commonness, merging the at least two health problems into a new health problem, and determining coding, category, relevant system / device identification, and description of the new health problem, and taking a higher risk level in risk levels corresponding to the at least two health problems as a risk level of the new health problem. Step S443, adding the code, category, risk level, related system / device identification, and description of the new health problem to the third problem list; Step S444, determining the final identification result of the current health problems of the nuclear power plant according to the third problem list.
5. The nuclear power plant health problem identification method according to claim 4, characterized by, After the step S443, further comprising: Marking the new health problem as a parent problem and marking the at least two health problems as child problems of the parent problem.
6. The nuclear power plant health problem identification method according to Claim 4, characterized by, The step S444 comprises: Arranging the health problems in the third problem list to obtain a final problem list, and taking the final problem list as the final identification result of the current health problems of the nuclear power plant.
7. The nuclear power plant health problem identification method according to claim 6, characterized by, The problem attribute further comprises a discovery time; and the arranging the health problems in the third problem list comprises: Arranging the health problems in the third problem list in descending order of risk level first; For multiple health problems with the same risk level, arranging them in a preset category priority order or a discovery time order from recent to remote, wherein the category comprises a safety category and a management category, and the priority of the safety category is higher than that of the management category.
8. The nuclear power plant health problem identification method according to claim 6, characterized by, After the step S40, further comprising: According to the pre-stored response setting information, determining the response mechanism information corresponding to each health problem in the final problem list, wherein the response setting information comprises multiple risk levels and response mechanism information corresponding to each risk level; Outputting the response mechanism information of each health problem to enable corresponding personnel to respond; Updating the processing status, mitigation measures, and final effectiveness of each health problem to perform closed-loop tracking of each health problem.
9. The nuclear power plant health problem identification method according to Claim 6, characterized by, After the step S40, further comprising: Optimizing the problem evaluation model according to the final problem list.
10. A nuclear power plant health issue identification system comprising a processor and a memory having stored therein a computer program, the system being characterized by, The processor implements the steps of the nuclear power plant health problem identification method of any one of claims 1-9 when executing the computer program.