Method and system for monitoring a nuclear power plant

Abstract

This invention discloses a method and system for monitoring nuclear power equipment. The method includes the following steps: extracting key parameters and their time-series data from the operating data of the nuclear power equipment based on its operating characteristics, and storing them; dynamically setting alarm thresholds for key parameters based on the operating specifications and historical key parameters of the nuclear power equipment; when key parameters of the nuclear power equipment are collected in real time, judging the alarm threshold for a single key parameter, and / or performing comprehensive analysis and anomaly detection on multiple key parameters to trigger a real-time alarm; calling historical key parameters and their time-series data of the nuclear power equipment, and calculating and predicting the future dynamic change rate of key parameters by continuously tracking the time-series data of key parameters, thereby triggering preventive maintenance alarms in advance. This invention can analyze the operating status of nuclear power equipment in real time, generate accurate alarm information, predict the future status of nuclear power equipment, and detect potential fault risks in advance.

Description

Technical Field

[0001] This invention relates to the field of equipment condition monitoring and fault early warning technology, and in particular to a method and system for monitoring nuclear power equipment. Background Technology

[0002] In recent years, with the rapid development of nuclear power technology and the continuous growth of nuclear power installed capacity, the safety and reliability of nuclear power equipment have become core issues in the operation of nuclear power plants. Nuclear power equipment typically operates in extreme environments such as high temperature, high pressure, and radiation. Changes in the condition of nuclear power equipment are often difficult to detect, and failures in nuclear power equipment can pose a serious threat to the safe operation of nuclear power plants.

[0003] Traditional nuclear power equipment monitoring methods primarily rely on alarm systems with fixed thresholds and single indicators. This static monitoring approach has several shortcomings when faced with dynamic changes in equipment operating parameters. For example, single-indicator alarm methods are susceptible to noise interference, leading to false alarms or missed alarms; fixed thresholds cannot adapt to changes in the operating status of nuclear power equipment and are difficult to accurately reflect the health status of the equipment. Furthermore, existing monitoring systems have limited capabilities in data processing and analysis, making it difficult to achieve real-time assessment and long-term prediction of the nuclear power equipment's status. This, to some extent, limits the maintenance efficiency and operational safety of nuclear power equipment. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a nuclear power equipment monitoring method and system, which addresses at least one defect of the related technologies mentioned in the background: the existing technologies have problems such as poor real-time performance, high false alarm / false alarm rates and lack of predictive ability.

[0005] The technical solution adopted by this invention to solve its technical problem is: to construct a nuclear power equipment monitoring method, comprising the following steps: Based on the operating characteristics of nuclear power equipment, key parameters and their time-series data are extracted from the operating data of the nuclear power equipment and stored. Based on the operating specifications and historical key parameters of the nuclear power equipment, the alarm thresholds of the key parameters are dynamically set. When key parameters of the nuclear power equipment are collected in real time, an alarm threshold is determined for a single key parameter, and / or a comprehensive analysis and anomaly detection are performed on multiple key parameters to trigger a real-time alarm; By calling upon the historical key parameters and their time-series data of the nuclear power equipment, and continuously tracking the time-series data of the key parameters, the future dynamic change rate of the key parameters is calculated and predicted, triggering preventive maintenance alarms in advance.

[0006] In the aforementioned nuclear power equipment monitoring method, preferably, the step of extracting key parameters and their time-series data from the operating data of the nuclear power equipment based on its operating characteristics and storing them includes: The raw operating data of the nuclear power equipment is collected in real time using sensors and data acquisition devices; The raw operating data of the nuclear power equipment collected is cleaned; Based on the operating characteristics of the nuclear power equipment, the key parameters and their time-series data are extracted from the cleaned operating data. The identification, key parameters, and time-series data of the nuclear power equipment are stored in a database, and a data index is established.

[0007] In the aforementioned nuclear power equipment monitoring method, preferably, the step of cleaning the collected raw operating data of the nuclear power equipment includes: The raw operating data of the nuclear power equipment is preprocessed, including handling missing values ​​and unifying timestamps; Detect and identify outliers using at least one of the statistical, machine learning, and rule-based methods. The cause of the anomaly was confirmed through cross-validation and manual review, and the missing values ​​were filled in by interpolation, or the outlier values ​​were replaced or deleted.

[0008] In the aforementioned nuclear power equipment monitoring method, preferably, the step of storing the nuclear power equipment's identifier, key parameters, and their time-series data in a database and establishing a data index includes: Choose the database type and table structure; The identification, key parameters, and time-series data of the nuclear power equipment are stored in the corresponding data tables; Create an index that is associated with a preset column in the data table.

[0009] In the aforementioned nuclear power equipment monitoring method, preferably, the step of dynamically setting the alarm threshold of the key parameters based on the operating specifications and historical key parameters of the nuclear power equipment includes: Based on the historical normal range of the key parameters under the aforementioned operating specifications, the alarm thresholds of the key parameters are dynamically relaxed.

[0010] In the aforementioned nuclear power equipment monitoring method, preferably, the step of determining the alarm threshold for a single key parameter includes: Access the historical key parameters of the nuclear power equipment; Establish a correlation model between the historical key parameters of the nuclear power equipment and environmental data; Real-time acquisition of the current key parameters of the nuclear power equipment; Real-time acquisition of environmental data from the nuclear power equipment; The deviation between the current key parameters and the alarm threshold is analyzed, and dynamic calibration is performed using the correlation model in conjunction with the environmental data.

[0011] In the aforementioned nuclear power equipment monitoring method, preferably, the comprehensive analysis and anomaly detection of multiple key parameters includes: An anomaly detection algorithm based on machine learning or deep learning can further detect whether the correlation between the key parameter and other key parameters deviates abnormally when one of the key parameters has not yet exceeded the alarm threshold.

[0012] In the aforementioned nuclear power equipment monitoring method, preferably, the step of calling up the historical key parameters and their time-series data of the nuclear power equipment, and calculating and predicting the future dynamic change rate of the key parameters by continuously tracking the time-series data of the key parameters, thereby triggering a preventive maintenance alarm in advance, includes: By calling up the historical key parameters and their time-series data of the nuclear power equipment, continuously tracking the time-series data of the key parameters, and combining it with a fault case library of similar nuclear power equipment, a time series prediction model is used to calculate and predict the future dynamic change rate of the key parameters, thereby triggering preventive maintenance alarms in advance.

[0013] Preferably, in the nuclear power equipment monitoring method described above, the nuclear power equipment monitoring method further includes: The error of the time series prediction model is evaluated by comparing the predicted results with the actual measured values. Based on the error assessment results, the parameters of the time series prediction model are dynamically adjusted. The time series prediction model is updated regularly.

[0014] The present invention also constructs a nuclear power equipment monitoring system, comprising: The database module is used to extract key parameters and their time-series data from the operating data of the nuclear power equipment based on the operating characteristics of the nuclear power equipment, and to store them. The threshold setting module is used to dynamically set the alarm threshold of the key parameters based on the operating specifications and historical key parameters of the nuclear power equipment. The real-time alarm module is used to determine an alarm threshold for a single key parameter and / or to perform comprehensive analysis and anomaly detection on multiple key parameters when the key parameters of the nuclear power equipment are collected in real time, thereby triggering a real-time alarm; and, The predictive alarm module is used to call up the historical key parameters and their time-series data of the nuclear power equipment. By continuously tracking the time-series data of the key parameters, it calculates and predicts the future dynamic change rate of the key parameters and triggers preventive maintenance alarms in advance.

[0015] By implementing this invention, the following beneficial effects are achieved: Existing traditional monitoring methods typically rely on offline data analysis, which lacks real-time capability and makes it difficult to capture dynamic changes in the operation of nuclear power equipment in a timely manner. However, this invention, through the real-time collection, judgment, and analysis of key parameters, can analyze the operating status of nuclear power equipment in real time, promptly detect subtle dynamic changes, and ensure the real-time and dynamic nature of monitoring.

[0016] Existing alarm systems are typically based on fixed thresholds and single indicators, which are prone to false alarms or missed alarms. In contrast, this invention dynamically sets alarm thresholds for key parameters based on the operating specifications and historical key parameters of nuclear power equipment. When key parameters of nuclear power equipment are collected in real time, alarm threshold judgment is performed on individual key parameters, and / or comprehensive analysis and anomaly detection are performed on multiple key parameters. This generates accurate alarm information, significantly improving the accuracy and reliability of alarms, reducing the probability of false alarms and missed alarms, effectively improving the operational safety of nuclear power equipment, and optimizing maintenance strategies.

[0017] Existing technologies typically focus only on the current state of nuclear power equipment, lacking the ability to predict its future state and making it difficult to detect potential failure risks in advance. However, this invention, by calling upon historical key parameters and their time-series data of nuclear power equipment and continuously tracking the time-series data of key parameters, calculates and predicts the future dynamic change rate of key parameters. This enables the prediction of the future state of nuclear power equipment, the early detection of potential failure risks, and thus provides a scientific basis for nuclear power equipment maintenance, sensitively reflecting the dynamic characteristics of the equipment's state.

[0018] Meanwhile, this invention improves the overall level of automation, automating data collection, analysis, and alarm generation, reducing the need for manual intervention, and improving monitoring efficiency and response speed. Attached Figure Description

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 A flowchart of the nuclear power equipment monitoring method of the present invention is shown; Figure 2 A logical structure diagram of the nuclear power equipment monitoring system of the present invention is shown. Detailed Implementation

[0020] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0021] It should be noted that the flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0022] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0023] The following can be one, two, three or any number.

[0024] like Figure 1 As shown, some embodiments of the present invention disclose a method for monitoring nuclear power equipment, including the following steps: Based on the operating characteristics of nuclear power equipment, key parameters and their time-series data are extracted from the operating data of nuclear power equipment and stored. Based on the operating specifications and historical key parameters of nuclear power equipment, alarm thresholds for key parameters are dynamically set. When key parameters of nuclear power equipment are collected in real time, alarm thresholds are determined for individual key parameters, and / or multiple key parameters are comprehensively analyzed and anomalies are detected to trigger real-time alarms; By accessing historical key parameters and their time-series data of nuclear power equipment and continuously tracking the time-series data of key parameters, the future dynamic change rate of key parameters can be calculated and predicted, triggering preventive maintenance alarms in advance.

[0025] Existing traditional monitoring methods typically rely on offline data analysis, which lacks real-time capability and makes it difficult to capture dynamic changes in the operation of nuclear power equipment in a timely manner. However, this invention, through the real-time collection, judgment, and analysis of key parameters, can analyze the operating status of nuclear power equipment in real time, promptly detect subtle dynamic changes, and ensure the real-time and dynamic nature of monitoring.

[0026] Existing alarm systems are typically based on fixed thresholds and single indicators, which are prone to false alarms or missed alarms. In contrast, this invention dynamically sets alarm thresholds for key parameters based on the operating specifications and historical key parameters of nuclear power equipment. When key parameters of nuclear power equipment are collected in real time, alarm threshold judgment is performed on individual key parameters, and / or comprehensive analysis and anomaly detection are performed on multiple key parameters. This generates accurate alarm information, significantly improving the accuracy and reliability of alarms, reducing the probability of false alarms and missed alarms, effectively improving the operational safety of nuclear power equipment, and optimizing maintenance strategies.

[0027] Existing technologies typically focus only on the current state of nuclear power equipment, lacking the ability to predict its future state and making it difficult to detect potential failure risks in advance. However, this invention, by calling upon historical key parameters and their time-series data of nuclear power equipment and continuously tracking the time-series data of key parameters, calculates and predicts the future dynamic change rate of key parameters. This enables the prediction of the future state of nuclear power equipment, the early detection of potential failure risks, and thus provides a scientific basis for nuclear power equipment maintenance, sensitively reflecting the dynamic characteristics of the equipment's state.

[0028] Meanwhile, this invention improves the overall level of automation, automating data collection, analysis, and alarm generation, reducing the need for manual intervention, and improving monitoring efficiency and response speed.

[0029] In some embodiments, based on the operating characteristics of the nuclear power equipment, key parameters and their time-series data are extracted from the operating data of the nuclear power equipment and stored, specifically including: Real-time acquisition of raw operating data from nuclear power equipment is achieved through sensors and data acquisition devices. The raw operating data of the collected nuclear power equipment is cleaned; Based on the operating characteristics of nuclear power equipment, key parameters and their time-series data are extracted from the cleaned operating data. Nuclear power equipment identification, key parameters and their time-series data are stored in a database and an index is created to facilitate subsequent retrieval, analysis and querying.

[0030] Key parameters include, but are not limited to, at least one of temperature, pressure, vibration, rotational speed, and power. These key parameters can be determined based on the operating characteristics of different nuclear power equipment.

[0031] For example, a nuclear reactor is a nuclear island device and the core part of a nuclear power plant. It is responsible for generating nuclear fission reactions and releasing energy. Its operating characteristics include precise control of power output and temperature. Key parameters include temperature, pressure, power output, and neutron flux.

[0032] The steam generator is an interface device between the nuclear island and the conventional island. It is responsible for transferring the heat generated by the nuclear reactor to water to generate high-temperature and high-pressure steam. Its operating characteristics include ensuring efficient heat transfer and stable steam output. Key parameters include steam pressure, steam temperature, heat transfer efficiency, and water level.

[0033] The circulating water pump is a conventional island device responsible for transporting circulating cooling water and maintaining the normal operation of the nuclear power plant's cooling system. Its operating characteristics include ensuring stable flow and pressure, and key parameters include flow rate, head, efficiency, and motor current.

[0034] These devices play a crucial role in the operation of a nuclear power plant. The nuclear reactor generates nuclear energy, the steam generator converts thermal energy into steam, and the circulating water pumps maintain the stable operation of the cooling system, collectively ensuring the safety and efficiency of the nuclear power plant. By analyzing the operating characteristics of these nuclear power devices, we can better understand the operating mechanisms of nuclear power plants and provide support for the optimization and maintenance of these devices.

[0035] The time series data of key parameters include, but are not limited to, at least one of the following: mean, standard deviation, kurtosis, skewness, periodicity, and trend.

[0036] In some embodiments, key parameters of the nuclear power equipment are collected in real time, specifically including: Real-time acquisition of raw operating data from nuclear power equipment is achieved through sensors and data acquisition devices. The raw operating data of the collected nuclear power equipment is cleaned; Based on the operating characteristics of nuclear power equipment, key parameters are extracted from the cleaned operating data.

[0037] This invention significantly improves data processing efficiency and meets the needs of real-time monitoring through automated data acquisition and preprocessing.

[0038] In some embodiments, the raw operating data of the collected nuclear power equipment is cleaned, specifically including: Preprocess the raw operating data of nuclear power equipment, including handling missing values ​​and unifying timestamps, to ensure the integrity and continuity of the data; Detect and identify outliers using at least one of the following methods: statistical methods, machine learning methods, and rule-based methods (such as thresholding and rate of change methods). The causes of anomalies were confirmed through cross-validation and manual review, and missing values ​​were filled in by interpolation, or outliers were replaced or deleted to ensure data quality.

[0039] In some embodiments, the identification, key parameters, and time-series data of the nuclear power equipment are stored in a database, and a data index is established, specifically including: Choose the database type (such as relational or time-series database) and table structure to ensure the scalability and efficiency of data storage; The identification, key parameters and their time-series data of nuclear power equipment are stored in the corresponding data tables; Create an index that is associated with a preset column in the data table. For example, the preset column is a frequently queried column, which is defined as a column whose number of queries within a certain period of time exceeds a preset value.

[0040] Creating indexes accelerates data retrieval and improves query efficiency. Indexes are associated with predefined columns to help quickly locate data. When optimizing indexes, select frequently queried columns and avoid excessive indexing to reduce storage space and impact on write operations. Through proper index design and optimization, ensure efficient data retrieval, providing reliable support for subsequent analysis and queries.

[0041] In some embodiments, historical key parameters of the nuclear power equipment are retrieved from a database. Based on the operating specifications and historical key parameters of the nuclear power equipment, alarm thresholds for key parameters are dynamically set, specifically including: Based on the historical normal range of key parameters under the operating specifications, the alarm thresholds of key parameters are dynamically relaxed.

[0042] For example, during the process of increasing the power output of a unit, the normal rise of certain key parameters (such as temperature and pressure) is expected. Based on the historical normal range of key parameters under the power level, the alarm threshold of key parameters will be dynamically relaxed to avoid false alarms caused by expected normal fluctuations, optimize performance, and reduce the possibility of false alarms and missed alarms.

[0043] In some embodiments, alarm threshold determination is performed on a single key parameter, specifically including: Accessing historical key parameters of nuclear power equipment; Establish a correlation model between historical key parameters of nuclear power equipment and environmental data; Real-time acquisition of current key parameters of nuclear power equipment; Real-time collection of environmental data (such as temperature and humidity) from nuclear power equipment; Analyze the deviation between current key parameters and alarm thresholds, and use an association model to perform dynamic calibration based on environmental data.

[0044] For example, increased ambient temperature may lead to decreased cooling efficiency, resulting in a slight increase in the equipment's operating temperature compared to normal. Humid or dusty environments may affect heat dissipation or cause changes in insulation performance. Existing technologies often fail to fully utilize historical operating data of the equipment, resulting in limited accuracy of monitoring results. This invention, however, combines historical key parameters of nuclear power equipment to construct a correlation model, placing the nuclear power equipment in its actual operating environment for comprehensive analysis, thereby improving the accuracy and effectiveness of alarms and ensuring the safe and stable operation of the equipment.

[0045] In some embodiments, comprehensive analysis and anomaly detection are performed on multiple key parameters, specifically including: Anomaly detection algorithms based on machine learning or deep learning can further detect whether there are abnormal deviations in the correlation between this key parameter and other key parameters when one of the key parameters has not yet exceeded the alarm threshold.

[0046] In other words, the judgment process not only involves threshold judgment for individual key parameters, but also comprehensive analysis and anomaly detection through multiple key parameters. When one key parameter has not yet exceeded the alarm threshold, but the correlation between this key parameter and other key parameters deviates abnormally (e.g., the drive current increases while the outlet pressure does not increase synchronously), anomaly detection algorithms based on machine learning or deep learning (such as isolated forests and autoencoders) can keenly capture this characteristic of equipment performance degradation. This can effectively identify systemic problems that cannot be detected by the alarm threshold of a single key parameter and identify potential abnormal patterns.

[0047] In some embodiments, historical key parameters and their time-series data of nuclear power equipment are invoked. By continuously tracking the time-series data of key parameters, the future dynamic change rate of key parameters is calculated and predicted, triggering preventive maintenance alarms in advance. Specifically, this includes: The system retrieves historical key parameters and their time-series data of nuclear power equipment from the database. By continuously tracking the time-series data of key parameters and combining it with a database of failure cases of similar nuclear power equipment, it uses time series prediction models (such as ARIMA, LSTM, etc.) to calculate and predict the future dynamic change rate of key parameters, thereby triggering preventive maintenance alarms in advance.

[0048] The dynamic rate of change calculation utilizes exponential smoothing to analyze the rate of change and trends of equipment parameters. Exponential smoothing employs a weighted averaging method, assigning higher weights to data from the most recent period to smooth the data and capture global trends. The process includes initializing smoothed values, calculating smoothed values, determining smoothing coefficients, recursively calculating smoothed values, calculating the rate of change and trend, and performing multi-step predictions. This effectively identifies the dynamic behavior of nuclear power equipment, providing support for fault early warning and maintenance decisions.

[0049] In some embodiments, the method further includes: The error of the time series prediction model is evaluated by comparing the predicted results with the actual measured values. Based on the error assessment results, the parameters of the time series prediction model are dynamically adjusted. Regularly update the time series forecasting model to ensure it adapts to changes in equipment characteristics and operating environment.

[0050] In some embodiments, the parameters of the time series prediction model are dynamically adjusted based on the error evaluation results, specifically including: Based on the error assessment results, the error type (e.g., persistent overestimation or underestimation) or operating condition (e.g., equipment start-up and shutdown) is identified. Fixed parameter correction amounts are preset based on experience to achieve rapid fine-tuning. Simultaneously, adaptive optimization algorithms (e.g., online variants of gradient descent) are used to automatically calculate the gradient of the error relative to the model parameters and dynamically update (in small increments) the parameters along the direction of error reduction. This allows the model to gradually adapt to slow data changes, minimizing error and improving prediction accuracy. The adjusted model is tested with real-time data to ensure performance improvement. Finally, the entire adjustment process is logged and model performance is continuously tracked to ensure real-time performance and stability.

[0051] This invention provides an intelligent optimization mechanism that dynamically optimizes model parameters, thereby improving the system's operating efficiency and reliability in complex environments.

[0052] In some embodiments, the method further includes: After an alarm is triggered, it is prioritized and categorized according to its severity and urgency. Management personnel are notified through multiple methods, and detailed alarm reasons and scientific maintenance or operation suggestions are provided to support decision-making for equipment maintenance. Record the trigger time, parameter values, handling measures, and results of each alarm, and analyze the data regularly to optimize the judgment logic and ensure the safe and stable operation of nuclear power equipment.

[0053] like Figure 2 As shown, some embodiments of the present invention also disclose a nuclear power equipment monitoring system, comprising: The database module is used to extract key parameters and their time-series data from the operating data of nuclear power equipment based on the operating characteristics of the equipment, and then store them. The threshold setting module is used to dynamically set alarm thresholds for key parameters based on the operating specifications and historical key parameters of nuclear power equipment. The real-time alarm module is used to determine alarm thresholds for individual key parameters and / or to perform comprehensive analysis and anomaly detection for multiple key parameters when real-time data of critical parameters of nuclear power equipment is collected, thereby triggering a real-time alarm; and, The predictive alarm module is used to call up the historical key parameters and their time-series data of nuclear power equipment. By continuously tracking the time-series data of key parameters, it calculates and predicts the future dynamic change rate of key parameters and triggers preventive maintenance alarms in advance.

[0054] Existing technologies are difficult to adapt to the specific requirements of different equipment, and the system has poor scalability and maintainability. In contrast, the present invention adopts a modular design, with each functional module being independent and expandable, possessing good scalability and compatibility, and can flexibly adapt to the monitoring needs of different nuclear power equipment.

[0055] Existing traditional monitoring methods typically rely on offline data analysis, which lacks real-time capability and makes it difficult to capture dynamic changes in the operation of nuclear power equipment in a timely manner. However, this invention, through the real-time collection, judgment, and analysis of key parameters, can analyze the operating status of nuclear power equipment in real time, promptly detect subtle dynamic changes, and ensure the real-time and dynamic nature of monitoring.

[0056] Existing alarm systems are typically based on fixed thresholds and single indicators, which are prone to false alarms or missed alarms. In contrast, this invention dynamically sets alarm thresholds for key parameters based on the operating specifications and historical key parameters of nuclear power equipment. When key parameters of nuclear power equipment are collected in real time, alarm threshold judgment is performed on individual key parameters, and / or comprehensive analysis and anomaly detection are performed on multiple key parameters. This generates accurate alarm information, significantly improving the accuracy and reliability of alarms, reducing the probability of false alarms and missed alarms, effectively improving the operational safety of nuclear power equipment, and optimizing maintenance strategies.

[0057] Existing technologies typically focus only on the current state of nuclear power equipment, lacking the ability to predict its future state and making it difficult to detect potential failure risks in advance. However, this invention, by calling upon historical key parameters and their time-series data of nuclear power equipment and continuously tracking the time-series data of key parameters, calculates and predicts the future dynamic change rate of key parameters. This enables the prediction of the future state of nuclear power equipment, the early detection of potential failure risks, and thus provides a scientific basis for nuclear power equipment maintenance, sensitively reflecting the dynamic characteristics of the equipment's state.

[0058] Meanwhile, this invention improves the overall level of automation, automating data collection, analysis, and alarm generation, reducing the need for manual intervention, and improving monitoring efficiency and response speed.

[0059] In some embodiments, based on the operating characteristics of the nuclear power equipment, key parameters and their time-series data are extracted from the operating data of the nuclear power equipment and stored, specifically including: Real-time acquisition of raw operating data from nuclear power equipment is achieved through sensors and data acquisition devices. The raw operating data of the collected nuclear power equipment is cleaned; Based on the operating characteristics of nuclear power equipment, key parameters and their time-series data are extracted from the cleaned operating data. Nuclear power equipment identification, key parameters and their time-series data are stored in a database and an index is created to facilitate subsequent retrieval, analysis and querying.

[0060] Key parameters include, but are not limited to, at least one of temperature, pressure, vibration, rotational speed, and power. These key parameters can be determined based on the operating characteristics of different nuclear power equipment.

[0061] For example, a nuclear reactor is a nuclear island device and the core part of a nuclear power plant. It is responsible for generating nuclear fission reactions and releasing energy. Its operating characteristics include precise control of power output and temperature. Key parameters include temperature, pressure, power output, and neutron flux.

[0062] The steam generator is an interface device between the nuclear island and the conventional island. It is responsible for transferring the heat generated by the nuclear reactor to water to generate high-temperature and high-pressure steam. Its operating characteristics include ensuring efficient heat transfer and stable steam output. Key parameters include steam pressure, steam temperature, heat transfer efficiency, and water level.

[0063] The circulating water pump is a conventional island device responsible for transporting circulating cooling water and maintaining the normal operation of the nuclear power plant's cooling system. Its operating characteristics include ensuring stable flow and pressure, and key parameters include flow rate, head, efficiency, and motor current.

[0064] These devices play a crucial role in the operation of a nuclear power plant. The nuclear reactor generates nuclear energy, the steam generator converts thermal energy into steam, and the circulating water pumps maintain the stable operation of the cooling system, collectively ensuring the safety and efficiency of the nuclear power plant. By analyzing the operating characteristics of these nuclear power devices, we can better understand the operating mechanisms of nuclear power plants and provide support for the optimization and maintenance of these devices.

[0065] The time series data of key parameters include, but are not limited to, at least one of the following: mean, standard deviation, kurtosis, skewness, periodicity, and trend.

[0066] In some embodiments, key parameters of the nuclear power equipment are collected in real time, specifically including: Real-time acquisition of raw operating data from nuclear power equipment is achieved through sensors and data acquisition devices. The raw operating data of the collected nuclear power equipment is cleaned; Based on the operating characteristics of nuclear power equipment, key parameters are extracted from the cleaned operating data.

[0067] This invention significantly improves data processing efficiency and meets the needs of real-time monitoring through automated data acquisition and preprocessing.

[0068] In some embodiments, the raw operating data of the collected nuclear power equipment is cleaned, specifically including: Preprocess the raw operating data of nuclear power equipment, including handling missing values ​​and unifying timestamps, to ensure the integrity and continuity of the data; Detect and identify outliers using at least one of the following methods: statistical methods, machine learning methods, and rule-based methods (such as thresholding and rate of change methods). The causes of anomalies were confirmed through cross-validation and manual review, and missing values ​​were filled in by interpolation, or outliers were replaced or deleted to ensure data quality.

[0069] In some embodiments, the identification, key parameters, and time-series data of the nuclear power equipment are stored in a database, and a data index is established, specifically including: Choose the database type (such as relational or time-series database) and table structure to ensure the scalability and efficiency of data storage; The identification, key parameters and their time-series data of nuclear power equipment are stored in the corresponding data tables; Create an index that is associated with a preset column in the data table. For example, the preset column is a frequently queried column, which is defined as a column whose number of queries within a certain period of time exceeds a preset value.

[0070] Creating indexes accelerates data retrieval and improves query efficiency. Indexes are associated with predefined columns to help quickly locate data. When optimizing indexes, select frequently queried columns and avoid excessive indexing to reduce storage space and impact on write operations. Through proper index design and optimization, ensure efficient data retrieval, providing reliable support for subsequent analysis and queries.

[0071] In some embodiments, historical key parameters of the nuclear power equipment are retrieved from a database. Based on the operating specifications and historical key parameters of the nuclear power equipment, alarm thresholds for key parameters are dynamically set, specifically including: Based on the historical normal range of key parameters under the operating specifications, the alarm thresholds of key parameters are dynamically relaxed.

[0072] For example, during the process of increasing the power output of a unit, the normal rise of certain key parameters (such as temperature and pressure) is expected. Based on the historical normal range of key parameters under the power level, the alarm threshold of key parameters will be dynamically relaxed to avoid false alarms caused by expected normal fluctuations, optimize performance, and reduce the possibility of false alarms and missed alarms.

[0073] In some embodiments, alarm threshold determination is performed on a single key parameter, specifically including: Accessing historical key parameters of nuclear power equipment; Establish a correlation model between historical key parameters of nuclear power equipment and environmental data; Real-time acquisition of current key parameters of nuclear power equipment; Real-time collection of environmental data (such as temperature and humidity) from nuclear power equipment; Analyze the deviation between current key parameters and alarm thresholds, and use an association model to perform dynamic calibration based on environmental data.

[0074] For example, increased ambient temperature may lead to decreased cooling efficiency, resulting in a slight increase in the equipment's operating temperature compared to normal. Humid or dusty environments may affect heat dissipation or cause changes in insulation performance. Existing technologies often fail to fully utilize historical operating data of the equipment, resulting in limited accuracy of monitoring results. This invention, however, combines historical key parameters of nuclear power equipment to construct a correlation model, placing the nuclear power equipment in its actual operating environment for comprehensive analysis, thereby improving the accuracy and effectiveness of alarms and ensuring the safe and stable operation of the equipment.

[0075] In some embodiments, comprehensive analysis and anomaly detection are performed on multiple key parameters, specifically including: Anomaly detection algorithms based on machine learning or deep learning can further detect whether there are abnormal deviations in the correlation between this key parameter and other key parameters when one of the key parameters has not yet exceeded the alarm threshold.

[0076] In other words, the judgment process not only involves threshold judgment for individual key parameters, but also comprehensive analysis and anomaly detection through multiple key parameters. When one key parameter has not yet exceeded the alarm threshold, but the correlation between this key parameter and other key parameters deviates abnormally (e.g., the drive current increases while the outlet pressure does not increase synchronously), anomaly detection algorithms based on machine learning or deep learning (such as isolated forests and autoencoders) can keenly capture this characteristic of equipment performance degradation. This can effectively identify systemic problems that cannot be detected by the alarm threshold of a single key parameter and identify potential abnormal patterns.

[0077] In some embodiments, historical key parameters and their time-series data of nuclear power equipment are invoked. By continuously tracking the time-series data of key parameters, the future dynamic change rate of key parameters is calculated and predicted, triggering preventive maintenance alarms in advance. Specifically, this includes: The system retrieves historical key parameters and their time-series data of nuclear power equipment from the database. By continuously tracking the time-series data of key parameters and combining it with a database of failure cases of similar nuclear power equipment, it uses time series prediction models (such as ARIMA, LSTM, etc.) to calculate and predict the future dynamic change rate of key parameters, thereby triggering preventive maintenance alarms in advance.

[0078] The dynamic rate of change calculation utilizes exponential smoothing to analyze the rate of change and trends of equipment parameters. Exponential smoothing employs a weighted averaging method, assigning higher weights to data from the most recent period to smooth the data and capture global trends. The process includes initializing smoothed values, calculating smoothed values, determining smoothing coefficients, recursively calculating smoothed values, calculating the rate of change and trend, and performing multi-step predictions. This effectively identifies the dynamic behavior of nuclear power equipment, providing support for fault early warning and maintenance decisions.

[0079] In some embodiments, the system further includes: The model adjustment module is used to evaluate the error of the time series prediction model by comparing the prediction results with the actual measured values, and dynamically adjust the parameters of the time series prediction model based on the error evaluation results.

[0080] The model update module is used to periodically update the time series forecasting model to ensure that it adapts to changes in equipment characteristics and operating environment.

[0081] In some embodiments, the parameters of the time series prediction model are dynamically adjusted based on the error evaluation results, specifically including: Based on the error assessment results, the error type (e.g., persistent overestimation or underestimation) or operating condition (e.g., equipment start-up and shutdown) is identified. Fixed parameter correction amounts are preset based on experience to achieve rapid fine-tuning. Simultaneously, adaptive optimization algorithms (e.g., online variants of gradient descent) are used to automatically calculate the gradient of the error relative to the model parameters and dynamically update (in small increments) the parameters along the direction of error reduction. This allows the model to gradually adapt to slow data changes, minimizing error and improving prediction accuracy. The adjusted model is tested with real-time data to ensure performance improvement. Finally, the entire adjustment process is logged and model performance is continuously tracked to ensure real-time performance and stability.

[0082] This invention provides an intelligent optimization mechanism that dynamically optimizes model parameters, thereby improving the system's operating efficiency and reliability in complex environments.

[0083] In some embodiments, the system further includes: The alarm module is used to prioritize and classify alarms according to their severity and urgency after they are triggered, notify management personnel through multiple methods, and provide detailed alarm reasons, scientific maintenance or operation suggestions to provide decision support for equipment maintenance. The recording module is used to record the trigger time, parameter values, handling measures and results of each alarm, and to analyze the data regularly to optimize the judgment logic and ensure the safe and stable operation of nuclear power equipment.

[0084] It is understood that the above embodiments only illustrate some implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can freely combine the above embodiments or technical features without departing from the concept of the present invention, and can also make several modifications and improvements, all of which fall within the protection scope of the present invention. That is, the embodiments described "in some embodiments" can be freely combined with any of the preceding and following embodiments. Therefore, all equivalent transformations and modifications made within the scope of the claims of the present invention should be covered by the claims of the present invention.

Claims

1. A nuclear power plant monitoring method, characterized by, Includes the following steps: Based on the operating characteristics of nuclear power equipment, key parameters and their time-series data are extracted from the operating data of the nuclear power equipment and stored. Based on the operating specifications and historical key parameters of the nuclear power equipment, the alarm thresholds of the key parameters are dynamically set. When key parameters of the nuclear power equipment are collected in real time, an alarm threshold is determined for a single key parameter, and / or a comprehensive analysis and anomaly detection are performed on multiple key parameters to trigger a real-time alarm; By calling upon the historical key parameters and their time-series data of the nuclear power equipment, and continuously tracking the time-series data of the key parameters, the future dynamic change rate of the key parameters is calculated and predicted, triggering preventive maintenance alarms in advance.

2. The nuclear power plant monitoring method according to claim 1, characterized in that, Based on the operating characteristics of the nuclear power equipment, the extraction and storage of key parameters and their time-series data from the operating data of the nuclear power equipment includes: The raw operating data of the nuclear power equipment is collected in real time using sensors and data acquisition devices; The raw operating data of the nuclear power equipment collected is cleaned; Based on the operating characteristics of the nuclear power equipment, the key parameters and their time-series data are extracted from the cleaned operating data. The identification, key parameters, and time-series data of the nuclear power equipment are stored in a database, and a data index is established.

3. The nuclear power plant monitoring method according to claim 2, characterized in that, The cleaning of the raw operating data of the nuclear power equipment includes: The raw operating data of the nuclear power equipment is preprocessed, including handling missing values ​​and unifying timestamps; Detect and identify outliers using at least one of the statistical, machine learning, and rule-based methods. The cause of the anomaly was confirmed through cross-validation and manual review, and the missing values ​​were filled in by interpolation, or the outlier values ​​were replaced or deleted.

4. The nuclear power plant monitoring method according to claim 2, characterized in that, The step of storing the identification, key parameters, and time-series data of the nuclear power equipment in a database and establishing a data index includes: Choose the database type and table structure; The identification, key parameters, and time-series data of the nuclear power equipment are stored in the corresponding data tables; Create an index that is associated with a preset column in the data table.

5. The nuclear power plant monitoring method according to claim 1, characterized in that, The method of dynamically setting alarm thresholds for key parameters based on the operating specifications and historical key parameters of the nuclear power equipment includes: Based on the historical normal range of the key parameters under the aforementioned operating specifications, the alarm thresholds of the key parameters are dynamically relaxed.

6. The nuclear power plant monitoring method according to claim 5, characterized in that, The determination of the alarm threshold for a single key parameter includes: Access the historical key parameters of the nuclear power equipment; Establish a correlation model between the historical key parameters of the nuclear power equipment and environmental data; Real-time acquisition of the current key parameters of the nuclear power equipment; Real-time acquisition of environmental data from the nuclear power equipment; The deviation between the current key parameters and the alarm threshold is analyzed, and dynamic calibration is performed using the correlation model in conjunction with the environmental data.

7. The nuclear power plant monitoring method according to claim 1, characterized by, The comprehensive analysis and anomaly detection of multiple key parameters includes: An anomaly detection algorithm based on machine learning or deep learning can further detect whether the correlation between the key parameter and other key parameters deviates abnormally when one of the key parameters has not yet exceeded the alarm threshold.

8. The nuclear power plant monitoring method according to claim 1, characterized by, The process of accessing historical key parameters and their time-series data of the nuclear power equipment, continuously tracking the time-series data of these key parameters, calculating and predicting the future dynamic change rate of these key parameters, and triggering preventative maintenance alarms in advance includes: By calling up the historical key parameters and their time-series data of the nuclear power equipment, continuously tracking the time-series data of the key parameters, and combining it with a fault case library of similar nuclear power equipment, a time series prediction model is used to calculate and predict the future dynamic change rate of the key parameters, thereby triggering preventive maintenance alarms in advance.

9. The nuclear power equipment monitoring method according to claim 8, characterized in that, The nuclear power equipment monitoring method also includes: The error of the time series prediction model is evaluated by comparing the predicted results with the actual measured values. Based on the error assessment results, the parameters of the time series prediction model are dynamically adjusted. The time series prediction model is updated regularly.

10. A nuclear power plant monitoring system characterized by, include: The database module is used to extract key parameters and their time-series data from the operating data of the nuclear power equipment based on the operating characteristics of the nuclear power equipment, and to store them. The threshold setting module is used to dynamically set the alarm threshold of the key parameters based on the operating specifications and historical key parameters of the nuclear power equipment. The real-time alarm module is used to determine the alarm threshold for a single key parameter and / or to perform comprehensive analysis and anomaly detection on multiple key parameters when the key parameters of the nuclear power equipment are collected in real time, thereby triggering a real-time alarm. as well as, The predictive alarm module is used to call up the historical key parameters and their time-series data of the nuclear power equipment. By continuously tracking the time-series data of the key parameters, it calculates and predicts the future dynamic change rate of the key parameters and triggers preventive maintenance alarms in advance.

Images