Method and device for analyzing operation events, electronic device and storage medium

Abstract

The present disclosure discloses an operation event analysis method and device, an electronic device and a storage medium. According to the present application, the event knowledge base is constructed based on historical event data and the causal analysis model is trained, the sufficient extraction and fusion analysis of event characteristics are realized, the direct cause and the systematic cause can be accurately identified without relying on manual annotation, the knowledge base and the model are dynamically updated through new event data to improve the generalization ability, the feedback document is automatically associated with the historical experience, the technical effects of improving the operation event analysis efficiency of the nuclear power plant, enhancing the accuracy and comprehensiveness of event cause identification, reducing the occurrence of repetitive failures, strengthening the experience inheritance and risk prevention and control capability, and further ensuring the operation safety and stability of the nuclear power plant are achieved.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, and more particularly to a method and apparatus for analyzing operational events, electronic devices, and storage media. Background Technology

[0002] Nuclear power plant operation event analysis and experience feedback generation, as a core component of the nuclear energy safety management system, are widely used for risk prevention and experience transfer throughout the entire life cycle of nuclear power plants. With the development of artificial intelligence technology, the nuclear power field has gradually introduced machine learning algorithms for event classification and importance grading. Among related technologies, the collaborative operation of algorithms such as support vector machines and decision trees with database systems has constructed an intelligent processing framework from data acquisition to document generation.

[0003] Existing event analysis methods rely on manually labeled classification models, but these suffer from shortcomings such as insufficient feature extraction and inadequate model generalization ability. Regarding event processing efficiency, current technologies take hours to days to analyze a single event, and the lack of historical data correlation analysis for new event feedback leads to frequent recurring failures. Summary of the Invention

[0004] This disclosure provides a method, apparatus, electronic device, and storage medium for analyzing operational events.

[0005] According to a first aspect of this disclosure, an analysis method for operational events is provided, comprising: Acquire historical event data and new event data to be analyzed during the operation of the nuclear power plant, and perform data standardization processing; Based on standardized historical event data, an event knowledge base is constructed that reflects the relationship between event types, fault manifestations, and causal factors. The new event data is subjected to feature extraction and fusion analysis based on a causal analysis model to identify the direct and systemic causes of the event and output the causal analysis results; wherein the causal analysis model is trained based on the historical event data. Based on the causal analysis results and the relationships in the event knowledge base, an experience feedback document containing causal analysis and improvement suggestions is generated. Based on the newly generated experience feedback documents and their corresponding event data, the event knowledge base and the causal analysis model are dynamically updated.

[0006] Optionally, the acquisition of historical event data and new event data to be analyzed during the operation of the nuclear power plant, and the data standardization processing, includes: The acquired event data is cleaned to remove invalid data and fill in missing items; The time information, numerical information, and text descriptions in the cleaned data are converted into a predefined unified format and standard units.

[0007] Optionally, the construction of an event knowledge base reflecting the correlation between event type, fault manifestation, and causal factors based on normalized historical event data includes: Establish a mapping relationship between event types and failure modes as the first level of the knowledge base; Based on the results of historical causal analysis, a causal chain between failure modes and specific causal elements is established as the second level of the knowledge base.

[0008] Optionally, the step of performing feature extraction and fusion analysis on the new event data based on the causal analysis model to identify the direct and systemic causes of the event, and outputting the causal analysis results, includes: The first neural network model is used to extract structured features from event data; The semantic features of the event description text are extracted using a second neural network model; The structured features are fused with semantic features, and an attention mechanism is used to determine the contribution of each feature to causal inference in order to generate the causal analysis results.

[0009] Optionally, the step of generating an experience feedback document containing cause analysis and improvement suggestions based on the causal analysis results and the relationships in the event knowledge base includes: Based on the preset document template, the basic information of the event, the results of causal analysis, and the related information matched from the event knowledge base are filled in to form the initial document; Based on the severity of the event, the scope of its impact, and the strength of the correlation between the causal factors, the improvement suggestions are prioritized and their content optimized.

[0010] Optionally, the method further includes: Based on the temporal characteristics of historical event data, a predictive model is used to predict the trend of event occurrence. Based on the prediction results, identify high-risk event types or equipment areas, and generate early warning information and preventive maintenance recommendations.

[0011] According to a second aspect of this disclosure, an operational event analysis apparatus is provided, comprising: The acquisition unit is also used to acquire historical event data and new event data to be analyzed during the operation of the nuclear power plant; The processing unit is also used to perform normalization processing on the acquired data; The building unit is also used to construct an event knowledge base that reflects the relationship between event types, fault manifestations and causal elements based on normalized historical event data; The analysis unit is also used to perform feature extraction and fusion analysis on the new event data based on a causal analysis model, so as to identify the direct cause and systemic cause of the event and output the causal analysis results; wherein the causal analysis model is trained based on the historical event data; The generation unit is also used to generate an experience feedback document containing cause analysis and improvement suggestions based on the causal analysis results and the correlation in the event knowledge base; The update unit is also used to dynamically update the event knowledge base and the causal analysis model based on newly generated experience feedback documents and their corresponding event data.

[0012] Optionally, the processing unit is further configured to: The acquired event data is cleaned to remove invalid data and fill in missing items; The time information, numerical information, and text descriptions in the cleaned data are converted into a predefined unified format and standard units.

[0013] Optionally, the building unit is further configured to: Establish a mapping relationship between event types and failure modes as the first level of the knowledge base; Based on the results of historical causal analysis, a causal chain between failure modes and specific causal elements is established as the second level of the knowledge base.

[0014] Optionally, the analysis unit is further configured to: The first neural network model is used to extract structured features from event data; The semantic features of the event description text are extracted using a second neural network model; The structured features are fused with semantic features, and an attention mechanism is used to determine the contribution of each feature to causal inference in order to generate the causal analysis results.

[0015] Optionally, the generation unit is further configured to: Based on the preset document template, the basic information of the event, the results of causal analysis, and the related information matched from the event knowledge base are filled in to form the initial document; Based on the severity of the event, the scope of its impact, and the strength of the correlation between the causal factors, the improvement suggestions are prioritized and their content optimized.

[0016] Optional, also includes: The prediction unit is also used to predict the trend of event occurrence based on the temporal characteristics of historical event data and using a prediction model. The identification unit is further configured to identify high-risk event types or equipment areas based on the prediction results, and generate early warning information and preventative maintenance recommendations. According to a third aspect of this disclosure, an electronic device is provided, comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect above.

[0017] According to a fourth aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are configured to cause the computer to perform the method described in the first aspect above.

[0018] According to a fifth aspect of this disclosure, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method described in the first aspect above.

[0019] The operational event analysis method, apparatus, electronic device, and storage medium disclosed herein, through which an event knowledge base is constructed based on historical event data and a causal analysis model is trained, achieves full extraction and fusion analysis of event features. It can accurately identify direct and systemic causes of events without relying on manual annotation. Simultaneously, it dynamically updates the knowledge base and model with new event data to improve generalization ability, and automatically generates feedback documents by associating with historical experience. Therefore, it can solve the technical problems in existing event analysis methods, such as insufficient feature extraction and inadequate model generalization ability due to reliance on manual annotation, long analysis time, and frequent recurring failures due to the lack of historical data association analysis for new events. This achieves the technical effects of improving the efficiency of nuclear power plant operational event analysis, enhancing the accuracy and comprehensiveness of event cause identification, reducing the occurrence of recurring failures, strengthening experience transfer and risk prevention capabilities, and ultimately ensuring the operational safety and stability of nuclear power plants.

[0020] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0021] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein: Figure 1 A flowchart illustrating an operational event analysis method provided in this embodiment of the disclosure; Figure 2A schematic diagram of the structure of an operational event analysis device provided in an embodiment of this disclosure; Figure 3 A schematic diagram of the structure of an operational event analysis device provided in an embodiment of this disclosure; Figure 4 A schematic block diagram of an example electronic device provided for embodiments of this disclosure. Detailed Implementation

[0022] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0023] The following description, with reference to the accompanying drawings, outlines a method, apparatus, electronic device, and storage medium for analyzing operational events according to embodiments of this disclosure.

[0024] Figure 1 This is a flowchart illustrating an operational event analysis method provided in an embodiment of the present disclosure.

[0025] like Figure 1 As shown, the method includes the following steps: Step 101: Obtain historical event data and new event data to be analyzed during the operation of the nuclear power plant, and perform data standardization processing; Providing a high-quality data foundation for nuclear power plant operation event analysis and experience feedback generation hinges on comprehensively acquiring effective data and ensuring its usability through standardized processing. Historical event data encompasses various relevant information accumulated during nuclear power plant operation, including but not limited to daily operation logs, detailed event reports, and equipment failure records. This data comes from a wide range of sources, including manually entered information by staff, real-time data captured by automated recording systems, and equipment operation data collected by various sensors. This historical data forms a crucial foundation for subsequent analysis and experience feedback. New event data to be analyzed refers to data related to various events that newly occur during nuclear power plant operation and require causal analysis and experience feedback generation. Timely acquisition of this type of data ensures the timeliness and relevance of the analysis.

[0026] Data normalization is a crucial step in this process, aiming to eliminate invalid information and inconsistencies in the data, thus clearing obstacles for subsequent analysis. The process begins with a comprehensive cleaning of both historical and new event data, removing noise, erroneous records, and blank information to ensure accuracy. Next, data from different sources and in different formats is standardized, such as unifying different date formats to a standard format and converting time units to a consistent unit. Missing values ​​are also appropriately filled to ensure data integrity. Through this series of normalization processes, historical event data and the new event data to be analyzed are combined into a structurally consistent, accurate, and complete dataset, providing reliable data support for subsequent event classification and causal analysis, effectively preventing data quality issues from affecting the accuracy and reliability of subsequent analysis results.

[0027] Step 102: Based on the normalized historical event data, construct an event knowledge base that reflects the relationship between event type, fault manifestation and causal factors; Standardized historical event data is the core foundation for building this knowledge base. This data, which has been cleaned, formatted, and completed, contains a large amount of real and valid event-related information, ensuring the accuracy and reliability of the knowledge base content.

[0028] The core of the event knowledge base lies in clearly presenting the inherent relationships between event types, failure manifestations, and causal elements. Event types encompass various event categories that may occur during nuclear power plant operation, such as mechanical failures, electrical failures, operational errors, and cooling system malfunctions; this categorizes the essential attributes of the events. Failure manifestations refer to the specific phenomena and characteristics exhibited when an event occurs, including observable and recordable manifestations such as abnormal equipment operation, shutdowns, and signs of component damage. Causal elements not only include the direct causes of the event but also cover deeper root causes and potential systemic problems, providing a comprehensive overview of the root causes of the event. During the construction process, in-depth mining and systematic analysis of standardized historical event data are conducted to extract the event type, failure manifestation, and causal element information for each historical event, thereby establishing a precise mapping relationship between the three. For example, this clarifies the specific failure manifestations often accompanied by a certain type of mechanical failure, as well as the corresponding direct and root causes, forming a structured knowledge system.

[0029] The establishment of these connections is not a simple accumulation of information, but rather based on statistical analysis and logical organization of a large amount of historical data, ensuring that each connection is supported by real-world cases and is reasonable. The resulting event knowledge base can systematically integrate key knowledge from historical events, enabling knowledge accumulation and reuse. This provides rapid and accurate knowledge references for the classification and causal analysis of subsequent new events and the generation of experience feedback, significantly improving the scientific rigor and efficiency of the entire analysis process.

[0030] Step 103: Based on the causal analysis model, perform feature extraction and fusion analysis on the new event data to identify the direct and systemic causes of the event, and output the causal analysis results; wherein, the causal analysis model is trained based on the historical event data; The effectiveness of the causal analysis model is based on sufficient training. Its training data comes from standardized historical event data, which covers complete information on various events during the operation of nuclear power plants, including event characteristics, causes, and correlations. Through systematic learning of this data, the model can grasp the inherent logical relationship between event characteristics and different types of causes, forming a reliable analytical capability.

[0031] When processing new event data, the first step is feature extraction. This involves comprehensively capturing various key features related to causal analysis from the new event data. These features cover multiple dimensions, including the specific manifestations of the new event, its scope of impact, changes in the operating status of equipment systems, and operational procedures, ensuring that no important information that might point to the cause is overlooked. Subsequently, feature fusion analysis is performed. The model integrates and deeply mines the extracted multi-dimensional features, combining them with historical causal patterns learned during training to analyze the causes of the new event layer by layer.

[0032] In this process, the model can clearly distinguish and identify the direct causes and systemic causes of an event. Direct causes refer to the immediate factors that directly lead to the occurrence of a new event; they are the surface-level triggering conditions. Systemic causes, on the other hand, are the deeper factors hidden behind the event, potentially involving issues affecting the stability of the nuclear power plant's operational system, such as equipment design defects, maintenance process loopholes, inadequate management mechanisms, and unsound technical standards. Through such feature extraction and fusion analysis, the model ultimately outputs structured causal analysis results, clearly presenting the direct and systemic causes of a new event. This ensures the comprehensiveness and accuracy of the analysis results, providing scientific and reliable support for subsequent experience feedback, improvement measure formulation, and other work based on these results. This helps operators fully understand the root causes of events, fundamentally solve problems, and prevent similar events from recurring.

[0033] Step 104: Based on the causal analysis results and the relationships in the event knowledge base, generate an experience feedback document containing causal analysis and improvement suggestions; The causal analysis results, serving as the core basis for document generation, clearly present the direct and systemic causes of new events, identifying the key issues that led to the events and forming the foundation for subsequent causal analysis and improvement recommendations. The correlations between event types, fault manifestations, and causal elements stored in the event knowledge base provide rich historical references for the documents. By accessing relevant information about similar events in the knowledge base, the analysis of current events can be made more in-depth and comprehensive.

[0034] When generating experience feedback documents, the cause analysis section focuses on the results of causal analysis, combining relevant data from the event knowledge base to elaborate on the direct causes and deep-seated systemic issues of the event. It also supplements the relevant causal characteristics of similar historical events, making the cause analysis more comprehensive and thorough, helping operations personnel clearly grasp the essence of the event. The improvement suggestion section, based on the conclusions of the cause analysis and referencing effective handling measures and experience summaries for similar events in the event knowledge base, proposes targeted and actionable improvement plans tailored to the specific circumstances of the current event. These suggestions draw on mature practices from historical experience while fitting the actual scenario of the current event, effectively guiding operations personnel to take precise measures to solve problems and prevent similar events from recurring.

[0035] The final generated experience feedback document organically combines cause analysis and improvement suggestions. It has a complete structure and clear logic, which can not only record the experience of handling the current event, but also provide a reliable reference for dealing with similar events in the future. It significantly improves the quality and practical application value of experience feedback on nuclear power plant operation events, and helps to continuously optimize the level of operation and management.

[0036] Step 105: Based on the newly generated experience feedback documents and their corresponding event data, dynamically update the event knowledge base and the causal analysis model.

[0037] The newly generated experience feedback documents encapsulate the results of in-depth analysis of new events, containing entirely new causal analysis conclusions, improvement suggestions, and corresponding event characteristic information. The associated event data covers the complete attribute details of the new events, and all of this constitutes the core basis for updates. When updating the event knowledge base, new correlations between the event type, fault manifestation, and causal elements corresponding to new events, as well as supplementary and improved information on existing correlations, are comprehensively integrated into the knowledge base. This fills potential knowledge gaps, enriches the correlation dimensions, and ensures that the knowledge base content always remains comprehensive and timely, covering a wider range of event scenarios.

[0038] For updates to the causal analysis model, new event data and their corresponding causal analysis results are used as new training data and input into the original model for iterative training. By continuously optimizing the model's parameter configuration, the model can accurately learn the characteristic patterns and causal logic of new events, continuously improving the efficiency of feature extraction and the accuracy of causal analysis for various new types of events. Through this dynamic update mechanism, the event knowledge base can continuously accumulate new operational experience, and the causal analysis model can continuously adapt to new event scenarios. The two form a virtuous cycle, ensuring that subsequent new event analysis and experience feedback document generation based on it always have reliable support, thereby guaranteeing the continuous optimization and efficient operation of the entire nuclear power plant's operational event analysis and experience feedback system.

[0039] In some embodiments, acquiring historical event data and new event data to be analyzed during the operation of a nuclear power plant, and performing data normalization processing includes: The acquired event data is cleaned to remove invalid data and fill in missing items; The time information, numerical information, and text descriptions in the cleaned data are converted into a predefined unified format and standard units.

[0040] The data cleaning process focuses on improving data quality. First, it involves a comprehensive screening of historical event data and new event data to be analyzed, accurately identifying and removing invalid data. This invalid data includes, but is not limited to, erroneous information caused by recording errors, noise data caused by system interference, and blank records without practical significance. At the same time, for any missing items in the data, they are reasonably filled in according to the inherent logic of the data and the characteristic patterns of similar data, ensuring that each event data has completeness and accuracy, providing a reliable foundation for subsequent processing.

[0041] The standardization of formats and units aims to eliminate data heterogeneity. It converts time information in the cleaned data into a predefined, unified format, ensuring consistent time representation across different sources and facilitating subsequent time-based analysis and comparison. It also converts various numerical information into industry-standard units, such as converting different time lengths to minutes and standardizing parameter values ​​from different measurement methods to avoid analytical errors caused by unit differences. Simultaneously, it organizes text descriptions according to predefined standardized formats, ensuring consistent and clear presentation of event information. Through these operations, event data that might otherwise vary in format and unit becomes a unified, standardized dataset. This effectively guarantees that subsequent stages, such as building an event knowledge base and running causal analysis models, can proceed smoothly based on a consistent data foundation, preventing data format inconsistencies or unit discrepancies from affecting the reliability and accuracy of the analysis results.

[0042] In some embodiments, constructing an event knowledge base reflecting the relationship between event type, fault manifestation, and causal factors based on normalized historical event data includes: Establish a mapping relationship between event types and failure modes as the first level of the knowledge base; Based on the results of historical causal analysis, a causal chain between failure modes and specific causal elements is established as the second level of the knowledge base.

[0043] The process of building an event knowledge base achieves structure and efficiency by establishing hierarchical relationships, ensuring a clear and traceable connection between event types, fault manifestations, and causal factors. The core of the first level is establishing a mapping relationship between event types and fault modes. Event types cover various event categories that may occur in nuclear power plant operation, including mechanical faults, electrical faults, operational errors, and cooling system failures. Each event type corresponds to a specific fault mode, which is the specific fault form and manifestation characteristics presented when the event occurs. For example, fault modes corresponding to mechanical faults may include bearing wear, pump instability, and internal component damage; fault modes corresponding to electrical faults may include transformer overheating and cable overload. By sorting and standardizing historical event data, each event type is accurately matched with its corresponding fault modes, forming the top-level classification structure of the knowledge base, laying the foundation for subsequent deeper associations.

[0044] The second layer constructs causal chains between failure modes and specific causal elements based on historical causal analysis results. These causal elements encompass the direct and root causes of the failure mode, as well as potential systemic causes. By deeply mining causal relationship data from historical events, the specific causes behind each failure mode are clarified. For example, the direct cause of a pump's unstable operation might be abnormal operating parameters, the root cause might be bearing wear, and the systemic cause might be an unreasonable maintenance cycle. These causal relationships are solidified into the knowledge base in chain form, forming a hierarchical and logically coherent knowledge system. This layered construction approach makes the event knowledge base structure clearer and the relationships more precise, facilitating rapid subsequent queries and providing a clear knowledge traceability path for the analysis of new events. This ensures that the knowledge base-based analysis process is efficient and reliable, fully leveraging the knowledge value of historical event data.

[0045] In some embodiments, the feature extraction and fusion analysis of the new event data based on the causal analysis model to identify the direct and systemic causes of the event and output the causal analysis results includes: The first neural network model is used to extract structured features from event data; The semantic features of the event description text are extracted using a second neural network model; The structured features are fused with semantic features, and an attention mechanism is used to determine the contribution of each feature to causal inference in order to generate the causal analysis results.

[0046] The first neural network model is specifically designed to extract structured features from event data. These features are information with fixed formats and quantifiable attributes, including directly quantifiable or standardized attributes such as the time of occurrence, the time of impact, the type of equipment involved, and the category to which the event belongs. The model accurately captures these objective and explicit features by analyzing the standardized new event data layer by layer, providing a solid factual basis for causal analysis. The second neural network model focuses on extracting semantic features from event description text. Event description text contains rich unstructured information such as the process of fault occurrence and operational details. The model can delve into the deeper semantics behind the text, such as implicit information like environmental conditions at the time of the fault, equipment operating status, and abnormal links in the operational process. It transforms textual information into feature vectors that can be used for analysis, supplementing key content at the subjective descriptive level that structured features fail to cover.

[0047] The feature fusion stage organically integrates the extracted structured features and semantic features to form a comprehensive feature set that fully reflects the multi-dimensional attributes of the new event. Subsequently, an attention mechanism is introduced to assign weights to the fused features. This mechanism can automatically identify the importance of different features in causal inference. For example, semantic features directly related to the fault may contribute more to the identification of the direct cause, while structured features related to the long-term operating status of the equipment may be more valuable for the analysis of systemic causes. By quantifying the contribution of each feature, the core influencing factors are highlighted and the interference of secondary information is weakened. Finally, based on the weighted comprehensive features, the direct cause and systemic cause of the event are accurately identified, and the causal analysis results with rigorous logic and reliable results are output, ensuring that the cause identification is both comprehensive and in line with the essence of the event.

[0048] In some embodiments, generating an experience feedback document containing cause analysis and improvement suggestions based on the causal analysis results and the relationships in the event knowledge base includes: Based on the preset document template, the basic information of the event, the results of causal analysis, and the related information matched from the event knowledge base are filled in to form the initial document; Based on the severity of the event, the scope of its impact, and the strength of the correlation between the causal factors, the improvement suggestions are prioritized and their content optimized.

[0049] The pre-built document template is a standardized framework that clearly defines the areas for displaying basic event information, presenting causal analysis results, supplementing related information, and listing improvement suggestions, providing a unified structural guide for document generation. When filling in the initial document, basic event information is directly extracted from the standardized new event data, including key content such as the event occurrence time, involved equipment, event type, and impact time, ensuring the authenticity and accuracy of the information; the causal analysis results fully incorporate the identified direct and systemic causes, clearly presenting the core issues of the event; the related information matched from the event knowledge base covers historical handling experience, cause correlation patterns, and other content related to similar events, providing a strong historical reference for the analysis of the current event. This information is precisely filled in according to the pre-set logical order and position of the template, forming a comprehensive and well-structured initial document.

[0050] Based on this, improvement suggestions are prioritized and optimized according to the severity, scope of impact, and correlation strength of causal factors of the event. The severity of the event is determined by its potential threat to the operational safety of the nuclear power plant. The scope of impact includes dimensions such as the operation of equipment systems affected by the event. The correlation strength of causal factors refers to the degree of consistency between the current event cause and the causes of similar events in the knowledge base. Combining these three key dimensions, improvement suggestions targeting core causes with a wide scope of impact and high severity are prioritized. At the same time, the content of the suggestions is optimized. Priority suggestions are supplemented with detailed implementation steps, clear execution standards, and reasonable time nodes, while non-priority suggestions maintain concise and accurate expression to ensure that the suggestions are both in line with the actual scenario and highly operable. The final experience feedback document has a clear structure and coherent logic, which can provide comprehensive and powerful guidance for operations personnel to quickly grasp key information about the event, take effective countermeasures, and prevent similar events from happening again.

[0051] In some embodiments, the method further includes: Based on the temporal characteristics of historical event data, a predictive model is used to predict the trend of event occurrence. Based on the prediction results, identify high-risk event types or equipment areas, and generate early warning information and preventive maintenance recommendations.

[0052] This step aims to proactively mitigate operational risks at nuclear power plants through forward-looking analysis, providing predictive support for safety management. The temporal characteristics of historical event data encompass key information reflecting temporal relationships, such as the chronological order of events, frequency of occurrence in different time periods, failure intervals, and seasonal or time-period distribution patterns. These characteristics reveal underlying patterns in event occurrence and form the core basis for trend prediction. Based on these temporal characteristics, a suitable predictive model is used to scientifically extrapolate event trends. The model employs deep learning to capture the changing patterns in historical time-series data, accurately identifying trends in event frequency, periodic fluctuations, and changes in the probability of sudden occurrences, thereby outputting the possible scenarios for various events in the future.

[0053] Subsequently, based on the prediction results, high-risk identification was conducted, screening out event types with a significantly increasing probability of occurrence, as well as equipment areas with concentrated failure frequencies, a high number of predicted failures, or a continuously rising risk level. The specific scope of high-risk event types and high-risk equipment areas was clearly defined to ensure accurate and efficient risk positioning. Finally, based on the identification results, targeted early warning information and preventative maintenance recommendations were generated. The early warning information clearly listed high-risk objects, the potential risk level, and the possible timeframe, allowing operators to keep abreast of risk dynamics. The preventative maintenance recommendations, combining high-risk characteristics and historical event handling experience, proposed actionable preemptive measures, such as shortening inspection cycles and increasing inspection items for high-risk equipment areas, optimizing maintenance processes, improving protection mechanisms, and adjusting operating parameters for high-risk event types. By intervening in advance to eliminate potential hazards, the probability of events is effectively reduced, further enhancing the safety and stability of nuclear power plant operations and achieving a management upgrade from passive response to proactive prevention.

[0054] Corresponding to the above-described method for analyzing operational events, this invention also proposes an apparatus for analyzing operational events. Since the apparatus embodiments of this invention correspond to the method embodiments described above, details not disclosed in the apparatus embodiments can be referred to in the method embodiments, and will not be repeated here.

[0055] Figure 2 A schematic diagram of the structure of an operational event analysis device provided in an embodiment of this disclosure, as shown below. Figure 2 As shown, it includes: Acquisition unit 21 is also used to acquire historical event data and new event data to be analyzed during the operation of the nuclear power plant; The processing unit 22 is also used to perform normalization processing on the acquired data; Building unit 23 is also used to build an event knowledge base that reflects the relationship between event type, fault manifestation and causal elements based on normalized historical event data; The analysis unit 24 is also used to perform feature extraction and fusion analysis on the new event data based on the causal analysis model, so as to identify the direct cause and systemic cause of the event and output the causal analysis results; wherein the causal analysis model is trained based on the historical event data; The generation unit 25 is also used to generate an experience feedback document containing cause analysis and improvement suggestions based on the causal analysis results and the correlation in the event knowledge base; The update unit 26 is also used to dynamically update the event knowledge base and the causal analysis model based on the newly generated experience feedback documents and their corresponding event data.

[0056] Furthermore, in one possible implementation of this disclosure, the processing unit 22 is further configured to: The acquired event data is cleaned to remove invalid data and fill in missing items; The time information, numerical information, and text descriptions in the cleaned data are converted into a predefined unified format and standard units.

[0057] Furthermore, in one possible implementation of this disclosure embodiment, the construction unit 23 is further configured to: Establish a mapping relationship between event types and failure modes as the first level of the knowledge base; Based on the results of historical causal analysis, a causal chain between failure modes and specific causal elements is established as the second level of the knowledge base.

[0058] Furthermore, in one possible implementation of this disclosure, the analysis unit 24 is further configured to: The first neural network model is used to extract structured features from event data; The semantic features of the event description text are extracted using a second neural network model; The structured features are fused with semantic features, and an attention mechanism is used to determine the contribution of each feature to causal inference in order to generate the causal analysis results.

[0059] Furthermore, in one possible implementation of this disclosure embodiment, the generation unit 25 is further configured to: Based on the preset document template, the basic information of the event, the results of causal analysis, and the related information matched from the event knowledge base are filled in to form the initial document; Based on the severity of the event, the scope of its impact, and the strength of the correlation between the causal factors, the improvement suggestions are prioritized and their content optimized.

[0060] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, it also includes: Prediction unit 27 is also used to predict the trend of event occurrence based on the temporal characteristics of historical event data using a prediction model; The identification unit 28 is also used to identify high-risk event types or equipment areas based on the prediction results, and generate early warning information and preventive maintenance suggestions.

[0061] It should be noted that the foregoing explanation of the method embodiments also applies to the apparatus of the embodiments of this disclosure, and the principle is the same. Therefore, the embodiments of this disclosure are not limited thereto.

[0062] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0063] Figure 4 A schematic block diagram of an example electronic device 400 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0064] like Figure 4 As shown, device 400 includes a computing unit 401, which can perform various appropriate actions and processes based on a computer program stored in ROM (Read-Only Memory) 402 or a computer program loaded from storage unit 408 into RAM (Random Access Memory) 403. RAM 403 may also store various programs and data required for the operation of device 400. The computing unit 401, ROM 402, and RAM 403 are interconnected via bus 404. I / O (Input / Output) interface 405 is also connected to bus 404.

[0065] Multiple components in device 400 are connected to I / O interface 405, including: input unit 406, such as keyboard, mouse, etc.; output unit 407, such as various types of monitors, speakers, etc.; storage unit 408, such as disk, optical disk, etc.; and communication unit 409, such as network card, modem, wireless transceiver, etc. Communication unit 409 allows device 400 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0066] The computing unit 401 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, CPUs (Central Processing Units), GPUs (Graphics Processing Units), various special-purpose AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, DSPs (Digital Signal Processors), and any suitable processor, controller, microcontroller, etc. The computing unit 401 performs the various methods and processes described above, such as operational event analysis methods. For example, in some embodiments, the operational event analysis methods may be implemented as computer software programs tangibly contained in a machine-readable medium, such as storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on device 400 via ROM 402 and / or communication unit 409. When the computer program is loaded into RAM 403 and executed by the computing unit 401, one or more steps of the methods described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured to perform the aforementioned operational event analysis method by any other suitable means (e.g., by means of firmware).

[0067] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application-Specific Standard Products), SOCs (System-on-Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0068] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0069] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, RAM, ROM, EPROM (Electrically Programmable Read-Only Memory) or flash memory, optical fiber, CD-ROM (Compact Disc Read-Only Memory), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0070] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0071] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include LANs (Local Area Networks), WANs (Wide Area Networks), the Internet, and blockchain networks.

[0072] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. A server can be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service system that addresses the shortcomings of traditional physical hosts and VPS (Virtual Private Server) services, such as high management difficulty and weak business scalability. Servers can also be servers for distributed systems or servers incorporating blockchain technology.

[0073] It's important to note that artificial intelligence (AI) is the study of enabling computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, and planning). It encompasses both hardware and software technologies. AI hardware technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, and big data processing. AI software technologies primarily include computer vision, speech recognition, natural language processing, machine learning / deep learning, big data processing, and knowledge graph technologies.

[0074] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0075] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for analyzing operational events, characterized in that, include: Acquire historical event data and new event data to be analyzed during the operation of the nuclear power plant, and perform data standardization processing; Based on standardized historical event data, an event knowledge base is constructed that reflects the relationship between event types, fault manifestations, and causal factors. The new event data is subjected to feature extraction and fusion analysis based on a causal analysis model to identify the direct and systemic causes of the event and output the causal analysis results; wherein the causal analysis model is trained based on the historical event data. Based on the causal analysis results and the relationships in the event knowledge base, an experience feedback document containing causal analysis and improvement suggestions is generated. Based on the newly generated experience feedback documents and their corresponding event data, the event knowledge base and the causal analysis model are dynamically updated.

2. The method according to claim 1, characterized in that, The acquisition of historical event data and new event data to be analyzed during the operation of the nuclear power plant, and the subsequent data standardization processing, includes: The acquired event data is cleaned to remove invalid data and fill in missing items; The time information, numerical information, and text descriptions in the cleaned data are converted into a predefined unified format and standard units.

3. The method according to claim 1, characterized in that, The event knowledge base, constructed based on normalized historical event data and reflecting the correlation between event types, fault manifestations, and causal factors, includes: Establish a mapping relationship between event types and failure modes as the first level of the knowledge base; Based on the results of historical causal analysis, a causal chain between failure modes and specific causal elements is established as the second level of the knowledge base.

4. The method according to claim 1, characterized in that, The feature extraction and fusion analysis of the new event data based on the causal analysis model to identify the direct and systemic causes of the event, and the output of the causal analysis results include: The first neural network model is used to extract structured features from event data; The semantic features of the event description text are extracted using a second neural network model; The structured features are fused with semantic features, and an attention mechanism is used to determine the contribution of each feature to causal inference in order to generate the causal analysis results.

5. The method according to claim 1, characterized in that, The step of generating an experience feedback document containing cause analysis and improvement suggestions based on the causal analysis results and the relationships in the event knowledge base includes: Based on the preset document template, the basic information of the event, the results of causal analysis, and the related information matched from the event knowledge base are filled in to form the initial document; Based on the severity of the event, the scope of its impact, and the strength of the correlation between the causal factors, the improvement suggestions are prioritized and their content optimized.

6. The method according to claim 1, characterized in that, The method further includes: Based on the temporal characteristics of historical event data, a predictive model is used to predict the trend of event occurrence. Based on the prediction results, identify high-risk event types or equipment areas, and generate early warning information and preventive maintenance recommendations.

7. An operational event analysis device, characterized in that, include: The acquisition unit is also used to acquire historical event data and new event data to be analyzed during the operation of the nuclear power plant; The processing unit is also used to perform normalization processing on the acquired data; The building unit is also used to construct an event knowledge base that reflects the relationship between event types, fault manifestations and causal elements based on normalized historical event data; The analysis unit is also used to perform feature extraction and fusion analysis on the new event data based on a causal analysis model, so as to identify the direct cause and systemic cause of the event and output the causal analysis results; wherein the causal analysis model is trained based on the historical event data; The generation unit is also used to generate an experience feedback document containing cause analysis and improvement suggestions based on the causal analysis results and the correlation in the event knowledge base; The update unit is also used to dynamically update the event knowledge base and the causal analysis model based on newly generated experience feedback documents and their corresponding event data.

8. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method according to any one of claims 1-6.

Images