Drum-shaped filter screen state monitoring system, method and equipment for nuclear power station

Abstract

The invention provides a drum-shaped filter screen state monitoring system, method and device for a nuclear power plant, and the system comprises a sensor module which is used for monitoring the operation state of a drum-shaped filter screen in real time; the data acquisition module is connected with the sensor module and is used for receiving real-time data from the sensor module; the data analysis module is connected with the data acquisition module and is used for receiving the preprocessed data and carrying out normalization processing and feature extraction on the preprocessed data so as to form to-be-detected data; the fault detection module is used for calculating an abnormal score of the to-be-detected data according to an isolated forest algorithm, comparing the abnormal score with a preset threshold value and generating abnormal data when the abnormal score is higher than the preset threshold value, and the abnormal data comprises the abnormal score and a fault type; and the alarm module is used for sending out an alarm signal when the abnormal data is received, so that accurate detection and prediction of the state of the filter screen can be realized, an early warning signal is sent out before a fault occurs, and the safety and reliability of operation of nuclear power station equipment are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of monitoring technology for cold source equipment in nuclear power plants, and in particular to a system, method and equipment for monitoring the condition of drum filters used in nuclear power plants. Background Technology

[0002] Drum filters play a crucial role in the cooling system of nuclear power plants. The cooling system provides cooling water to the nuclear reactor, ensuring its operation within a safe temperature range and preventing risks such as overheating. Cooling water typically contains a large amount of impurities, including plant debris, silt, and biological matter such as fish and shrimp. Especially in environments with complex water sources or external pollution, these impurities can affect the normal operation of the cooling system. Drum filters efficiently filter these impurities, ensuring the purity of the cooling water and thus guaranteeing the safety and stability of the nuclear power plant's cooling system. However, over time, drum filters are affected by various factors such as physical wear, corrosion, and scaling, leading to a gradual decline in their performance and even malfunction. Especially in complex operating environments, filter failures are often difficult to predict and detect, significantly impacting the safety of the nuclear power plant.

[0003] Currently, the drum filter monitoring methods in related technologies mainly rely on manual inspection and periodic maintenance, failing to achieve real-time monitoring and fault prediction. Although obvious faults (such as overload, abnormal vibration, etc.) can be detected manually, potential faults are difficult to detect in a timely manner because they depend on human judgment. Summary of the Invention

[0004] In view of this, it is necessary to provide a drum filter condition monitoring system, method and equipment for nuclear power plants that can overcome at least one of the above defects.

[0005] In a first aspect, embodiments of this application provide a drum filter condition monitoring system for nuclear power plants, used to monitor the condition of drum filters, the system comprising:

[0006] The sensor module is used to monitor the operating status of the drum-shaped filter screen in real time. The sensor module includes: an axial laser sensor for monitoring the axial displacement and deformation of the drum-shaped filter screen; a radial laser sensor for monitoring the radial displacement and gear runout of the drum-shaped filter screen; an axial eddy current sensor and a radial eddy current sensor for monitoring the bearing operating status and displacement changes of the drum-shaped filter screen; and a left bearing level sensor and a right bearing level sensor for monitoring changes in the lubricating oil level in the spindle oil tank of the drum-shaped filter screen.

[0007] The data acquisition module is used to receive real-time data from the sensor module and to sample and preprocess the real-time data to form preprocessed data.

[0008] The data analysis module is used to receive the preprocessed data, and to normalize and extract features from the preprocessed data to form the data to be detected.

[0009] The fault detection module is used to calculate the anomaly score of the data to be detected according to the isolated forest algorithm, compare the anomaly score with a preset threshold, and generate abnormal data when the anomaly score is higher than the preset threshold. The abnormal data includes the anomaly score and the fault type.

[0010] An alarm module is used to issue an alarm signal when the abnormal data is received;

[0011] The fault prediction module is used to acquire the time sequence information in the data to be detected, and process the data to be detected according to the time sequence information to form a feature sequence, wherein the data to be detected includes displacement information, vibration information and liquid level information;

[0012] Construct a long short-term memory network model;

[0013] The feature sequence is input into the long short-term memory network model to obtain the displacement prediction value, vibration prediction value and liquid level prediction value of the drum filter screen;

[0014] When the predicted displacement value, the predicted vibration value, or the predicted liquid level value exceeds the corresponding preset range, predicted abnormal data is generated.

[0015] The alarm module is also used to issue an alarm signal when it receives the predicted abnormal data.

[0016] In one embodiment, the fault prediction module is further configured to:

[0017] A hybrid prediction network is constructed based on time-series features at multiple scales to output a prediction sequence, which includes the displacement prediction value, the vibration prediction value, or the liquid level prediction value and the corresponding confidence level.

[0018] Based on the predicted sequence and historical fault data of the nuclear power plant, the expected fault occurrence time and corresponding confidence interval are output through the reliability model, wherein the confidence interval is dynamically adjusted according to the real-time operating parameters of the nuclear power plant.

[0019] In one embodiment, the dynamic adjustment of the confidence interval includes:

[0020] The real-time operating parameters of the nuclear power plant are obtained, including load, cooling water inlet temperature, and flow rate.

[0021] The real-time operating parameters are normalized and operating risk indicators are calculated according to preset weights;

[0022] The adjustment coefficient of the confidence interval is determined by nonlinear mapping based on the operational risk indicators.

[0023] The baseline confidence interval is adjusted according to the adjustment factor to obtain the confidence interval.

[0024] In one embodiment, the fault detection module constructs an isolated forest based on multiple isolated trees, and each isolated tree is generated through the following steps:

[0025] Randomly select a data feature;

[0026] Randomly select a segmentation point within the selected feature range;

[0027] Divide the data into two parts according to the split point, and recursively repeat until the stopping condition is met.

[0028] In one embodiment, the anomaly score is calculated from the average path length of all the isolated trees, where the path length of each isolated tree is the number of steps required to get from the root node to the leaf node of the data point.

[0029] In one embodiment, the fault detection module calculates the anomaly score according to an anomaly scoring formula, the anomaly scoring formula being:

[0030]

[0031] Where s(x) is the anomaly score, E(h(x)) is the average path length of point x, and c(n) is the normalization factor.

[0032] In one embodiment, the fault prediction module is further configured to:

[0033] Obtain the changing trends of the displacement prediction value, the vibration prediction value, and the liquid level prediction value, and calculate the probability of failure based on the changing trends;

[0034] When the probability of the fault occurrence exceeds a preset probability threshold, a fault warning message is generated, which includes the fault type and the expected occurrence time.

[0035] In one embodiment, the fault prediction module is further configured to:

[0036] Obtain the current operating status parameters of the nuclear power plant;

[0037] Based on the current operating status parameters, a dynamic probability threshold is calculated using a S-shaped function;

[0038] The fault warning information is generated based on the dynamic probability threshold.

[0039] Secondly, embodiments of this application provide a method for monitoring the condition of a drum filter used in a nuclear power plant, applied to the drum filter condition monitoring system for nuclear power plants as described in the first aspect, the method comprising:

[0040] Real-time monitoring of the operating status of the drum filter;

[0041] Receive real-time data from the sensor module, and sample and preprocess the real-time data to form preprocessed data;

[0042] The preprocessed data is received, and the preprocessed data is normalized and features are extracted to form the data to be detected.

[0043] Anomaly scores are calculated for the data to be detected using the isolated forest algorithm. The anomaly scores are compared with a preset threshold. When the anomaly scores are higher than the preset threshold, anomaly data is generated. The anomaly data includes the anomaly scores and the fault type.

[0044] An alarm signal is triggered upon receiving the abnormal data.

[0045] Thirdly, embodiments of this application provide an electronic device, including:

[0046] processor;

[0047] Memory used to store processor-executable instructions;

[0048] The processor is configured to implement the method for monitoring the status of a drum filter for a nuclear power plant as described in the second aspect when executing the instructions.

[0049] The drum filter status monitoring system, method, and equipment for nuclear power plants provided in this application can monitor the key operating status of the drum filter in real time through sensor modules. By combining the isolated forest algorithm and the long short-term memory network model, it can achieve accurate detection and prediction of the filter status and issue early warning signals before a fault occurs, significantly improving the safety and reliability of nuclear power plant equipment operation. Attached Figure Description

[0050] Figure 1 This is a schematic diagram of a drum filter status monitoring system module for nuclear power plants provided in one embodiment of this application.

[0051] Figure 2 This is a schematic diagram of the installation of a sensor module provided in one embodiment of this application.

[0052] Figure 3 This is a schematic flowchart of a method for monitoring the condition of a drum filter for a nuclear power plant, provided in one embodiment of this application.

[0053] Figure 4A schematic diagram of an electronic device provided in an embodiment of this application.

[0054] Explanation of main component symbols

[0055] Nuclear Power Plant Drum Filter Condition Monitoring System 1

[0056] Sensor Module 10

[0057] Axial laser sensor 100

[0058] Radial laser sensor 200

[0059] Radial eddy current sensor 300

[0060] Axial eddy current sensor 400

[0061] Right bearing level sensor 500

[0062] Left bearing level sensor 600

[0063] Data acquisition module 20

[0064] Data Analysis Module 30

[0065] Fault detection module 40

[0066] Alarm module 50

[0067] Fault prediction module 60

[0068] Electronic devices 2

[0069] Processor 21

[0070] Memory 22

[0071] Method steps S100-S500 Detailed Implementation

[0072] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0073] It should be noted that, in the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application.

[0074] It should be noted that in the embodiments of this application, the terms "first," "second," etc., are used only for descriptive purposes and should not be construed as indicating or implying relative importance, nor as indicating or implying order. Features specified as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0075] Based on the embodiments described in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0076] Drum filters play a crucial role in the cooling system of nuclear power plants. The cooling system provides cooling water to the nuclear reactor, ensuring its operation within a safe temperature range and preventing risks such as overheating. Cooling water typically contains a large amount of impurities, including plant debris, silt, and biological matter such as fish and shrimp. Especially in environments with complex water sources or external pollution, these impurities can affect the normal operation of the cooling system. Drum filters efficiently filter these impurities, ensuring the purity of the cooling water and thus guaranteeing the safety and stability of the nuclear power plant's cooling system. However, over time, drum filters are affected by various factors such as physical wear, corrosion, and scaling, leading to a gradual decline in their performance and even malfunction. Especially in complex operating environments, filter failures are often difficult to predict and detect, significantly impacting the safety of the nuclear power plant.

[0077] Nuclear power plant cooling systems are typically exposed to high humidity environments. Impurities and corrosive substances in the water (such as salt and minerals) accelerate the corrosion process of drum filters, especially the metal parts. Regular inspections often fail to detect problems when small corrosion spots have formed on the filter surface.

[0078] During operation, the filter screen is affected by factors such as changes in water flow and velocity, resulting in a certain degree of vibration and wear. Over a long period of operation, the structure of the drum filter screen may undergo minor deformation or wear, which may not be detected during routine inspections.

[0079] Currently, the drum filter monitoring methods in related technologies mainly rely on manual inspection and periodic maintenance, failing to achieve real-time monitoring and fault prediction. Although obvious faults (such as overload, abnormal vibration, etc.) can be detected manually, potential faults are difficult to detect in a timely manner because they depend on human judgment.

[0080] In view of this, an embodiment of the present application provides a drum filter condition monitoring system, method and equipment for nuclear power plants. Through sensor modules, the key operating conditions of the drum filter (such as axial displacement, radial displacement, gear ring runout and lubricating oil level, etc.) are monitored in real time. Combined with the isolated forest algorithm and the long short-term memory network (LSTM) model, the system can accurately detect and predict the filter condition, issue early warning signals before a fault occurs, and significantly improve the safety and reliability of nuclear power plant equipment operation.

[0081] Figure 1 This application provides an embodiment of a drum filter condition monitoring system for nuclear power plants. For example... Figure 1 The nuclear power plant drum filter status monitoring system 1 shown includes: sensor module 10, data acquisition module 20, data analysis module 30, fault detection module 40, alarm module 50, and fault prediction module 60.

[0082] In this embodiment, the sensor module 10 is used to monitor the operating status of the drum filter in real time to provide high-precision operating parameters and provide reliable data support for subsequent data analysis and fault detection.

[0083] Understandably, the sensor module 10 is rationally arranged according to the actual operating conditions of the drum filter so as to fully cover key monitoring parameters and ensure the system's monitoring effect and fault identification capability.

[0084] Specifically, please refer to the following: Figure 2 , Figure 2 This is an installation diagram of the sensor module 10 provided in one embodiment of this application. Figure 2 As shown, the sensor module 10 includes: an axial laser sensor 100, a radial laser sensor 200, a radial eddy current sensor 300, an axial eddy current sensor 400, a right bearing level sensor 500, and a left bearing level sensor 600.

[0085] Specifically, the axial laser sensor 100 is used to monitor the axial displacement and deformation of the drum-shaped filter screen; the radial laser sensor 200 is used to monitor the radial displacement and gear runout of the drum-shaped filter screen; the axial eddy current sensor 300 and the radial eddy current sensor 400 are used to monitor the bearing operating status and displacement changes of the drum-shaped filter screen; and the left bearing level sensor 500 and the right bearing level sensor 600 are used to monitor the changes in the lubricating oil level in the spindle oil tank of the drum-shaped filter screen.

[0086] Understandably, the axial laser sensor 100 is positioned along the axial direction of the drum-shaped filter screen, adjacent to the axial movement area of ​​the screen, ensuring real-time capture of axial changes. High-precision laser technology is used to acquire minute displacements and deformations in the axial direction, providing crucial data for filter screen stability analysis. The radial laser sensor 200 is positioned in the radial direction of the filter screen, close to the movement trajectory of the gear ring. It detects runout and radial displacement changes during gear ring operation, analyzing for abnormal vibrations or gear ring wear. The axial eddy current sensor 300 is positioned close to the axial movement direction of the bearing to ensure the capture of minute bearing displacement signals. Real-time feedback of the eddy current signal monitors the stability of the bearing's operating state, providing data support for determining bearing performance. The radial eddy current sensor 400 is located in the radial direction of the bearing to detect radial displacement and potential abnormal signals. It is used to promptly detect abnormal phenomena in bearing operation (such as wear or misalignment), providing early warning for preventing bearing failure. The right bearing level sensor 500 is installed on the right side of the spindle oil tank, close to the lubrication oil circuit. Used to monitor real-time changes in lubricating oil level, ensuring sufficient and stable lubricating oil supply and preventing bearing damage due to insufficient oil. The left bearing level sensor 600 is installed on the left side of the spindle oil tank, opposite to the right level sensor.

[0087] Together with the liquid level sensor on the right, it can dynamically monitor the lubricating oil level on both sides of the spindle oil tank to determine whether the lubrication system is operating normally.

[0088] Understandably, the layout and functional design of the sensor module 10 have been fully optimized to balance multi-dimensional monitoring needs and efficient data acquisition performance.

[0089] Furthermore, all sensor components employ high-precision measurement technology, possessing excellent anti-interference capabilities and adapting to the complex operating environment of nuclear power plant drum filters. Simultaneously, the sensor module 10 is designed with good expandability, allowing for the addition of other types of sensors, such as temperature or pressure sensors, to further enhance the system's monitoring capabilities and adaptability in the future.

[0090] The multi-source data acquired by sensor module 10 can provide comprehensive support for the subsequent data analysis module of the system. Combined with the isolated forest algorithm and long short-term memory network model, it can realize accurate detection and fault prediction of the operating status of the drum filter, and improve the safety and stability of nuclear power plant operation.

[0091] In this embodiment of the application, the data acquisition module 20 is connected to the sensor module 10 and is used to receive real-time data from the sensor module, and to sample and preprocess the real-time data to form preprocessed data.

[0092] Specifically, the data acquisition module 20 uses a high-precision sampling mechanism to uniformly process the multi-source data (such as axial displacement, radial displacement, vibration information, and liquid level information) transmitted by the sensor module 10. The preprocessing steps include data cleaning, noise reduction, and format standardization to ensure that the input data quality meets the requirements of subsequent data analysis.

[0093] Understandably, the data acquisition module 20 achieves efficient data transmission and processing through optimized hardware interfaces and communication protocols, enabling it to process concurrent data from multiple sensors in real time and possessing a certain redundancy mechanism to cope with data loss or abnormal situations.

[0094] In this embodiment of the application, the data analysis module 30 is connected to the data acquisition module 20 and is used to receive preprocessed data, normalize the preprocessed data and extract features to form data to be detected.

[0095] Specifically, the data analysis module 30 normalizes various sensor data to a unified numerical range, eliminating the impact of different data scales. During feature extraction, the module extracts key features based on the statistical characteristics of the signal, such as the signal mean, standard deviation, and spectral characteristics, to reflect the operating status of the drum filter.

[0096] Understandably, the data analysis module 30 employs efficient algorithms and optimized hardware resource management to ensure the real-time performance and accuracy of data processing, providing stable input for subsequent fault detection.

[0097] In this embodiment, the fault detection module 40 is connected to the data analysis module 30 and is used to calculate the anomaly score of the data to be detected according to the isolated forest algorithm, compare the anomaly score with a preset threshold, and generate abnormal data when the anomaly score is higher than the preset threshold. The abnormal data includes the anomaly score and the fault type.

[0098] Specifically, the fault detection module 40 performs anomaly analysis on the data to be detected by constructing an isolated forest model. The module determines the degree of anomaly based on preset thresholds and generates anomaly data including anomaly scores and specific fault types. For example, anomaly scores above the threshold may indicate specific problems such as excessive axial displacement of the filter screen or insufficient lubricating oil level.

[0099] Understandably, the fault detection module 40 uses a distributed algorithm to optimize the operating efficiency of the isolated forest model, enabling it to quickly and accurately identify abnormal situations in the complex operating environment of a nuclear power plant.

[0100] In this embodiment, the fault detection module 40 calculates an anomaly score according to the anomaly scoring formula, which is:

[0101]

[0102] Where s(x) is the anomaly score, E(h(x)) is the average path length of point x, and c(n) is the normalization factor.

[0103] Specifically, the Isolation Forest model calculates the path length of each isolated tree, comprehensively assesses the degree of difference between the data points and the normal data distribution, and finally derives an anomaly score. The higher the score, the more likely the data point is to be an anomaly.

[0104] Understandably, this formula, combined with the characteristics of the isolated forest model, can effectively adapt to multidimensional data distributions, and is especially suitable for anomaly detection needs in complex nuclear power plant systems.

[0105] In this embodiment, the alarm module 50 is connected to the fault detection module 40 and is used to issue an alarm signal when abnormal data is received.

[0106] Specifically, the alarm module 50 generates corresponding alarm signals based on abnormal data, including audible and visual alarms, system log recordings, and remote notifications, so that maintenance personnel can take timely countermeasures.

[0107] Understandably, the alarm module 50 has multi-level alarm capabilities and can flexibly adjust the alarm level according to the severity of the anomaly score, minimizing the interference of false alarms.

[0108] In this embodiment, the nuclear power plant drum filter condition monitoring system 1 further includes: a fault prediction module 60, connected to a data analysis module 30 and an alarm module 50, used to acquire time-series information in the data to be detected, process the data to be detected according to the time-series information to form a feature sequence, wherein the data to be detected includes displacement information, vibration information and liquid level information; construct a long short-term memory network model; input the feature sequence into the long short-term memory network model to obtain the displacement prediction value, vibration prediction value and liquid level prediction value of the drum filter; generate predicted abnormal data when the displacement prediction value, vibration prediction value or liquid level prediction value exceeds the corresponding preset range; the alarm module is also used to issue an alarm signal when receiving the predicted abnormal data.

[0109] Specifically, the fault prediction module 60 extracts the time-series features of the data to be detected, constructs a feature sequence, uses an LSTM model to perform time-series prediction, generates future operating state parameters, and compares the prediction results with a preset threshold to determine whether there is a potential fault risk.

[0110] Understandably, the fault prediction module 60 can identify potential fault trends in the drum filter in advance, providing a scientific basis for preventive maintenance and improving the operational reliability and service life of the equipment.

[0111] In this embodiment, the output formula of the fault prediction module 60 is:

[0112]

[0113] in, It is the predicted value at time k in the future, and f(X) is the long short-term memory network model.

[0114] Specifically, the LSTM model generates high-precision predictions of the future operating state of the drum filter by memorizing and processing the long-term and short-term dependencies of the data.

[0115] Understandably, the LSTM model can effectively capture the time-series characteristics of drum filter operation data and is suitable for prediction needs in complex dynamic environments.

[0116] In this embodiment, the fault prediction module 60 optimizes the long short-term memory network model using a mean squared error loss function. The formula for the loss function is:

[0117]

[0118] Where K is the value of the loss function. Let this be the value of the i-th sample at time k in the future, as predicted by the model. Let N be the value of the i-th sample at time k in the future, and N be the total number of samples.

[0119] Specifically, by minimizing the loss function value, the fault prediction module 60 ensures the accuracy and reliability of the LSTM model's prediction results.

[0120] Understandably, the mean squared error loss function can effectively measure the deviation between the predicted value and the true value, providing a clear target direction for model optimization.

[0121] In this embodiment of the application, the fault prediction module 60 is further configured to: construct a hybrid prediction network based on time-series features of multiple scales and output a prediction sequence, the prediction sequence including displacement prediction value, vibration prediction value or liquid level prediction value and corresponding confidence level; and output the expected fault occurrence time and corresponding confidence interval through a reliability model based on the prediction sequence and historical fault data of the nuclear power plant, wherein the confidence interval is dynamically adjusted according to the real-time operating parameters of the nuclear power plant.

[0122] Specifically, the fault prediction module 60 is also used to construct a hybrid prediction network based on time-series features at multiple scales. Transient, periodic, and trend features are extracted through short, medium, and long time windows, respectively, and the time-domain, frequency-domain, and time-frequency-domain features of each sensor are fused to form a unified multi-scale time-series representation. This multi-scale time-series representation is input into the hybrid prediction network, which may include prediction sub-models based on recurrent neural networks, temporal convolutional networks, or attention mechanisms, and may employ model ensemble strategies to output multi-step prediction sequences. Each predicted value is accompanied by a corresponding confidence assessment, which is obtained through uncertainty measures of model ensemble, regression intervals, or Monte Carlo sampling, and its interpretability is improved through calibration. Based on the prediction sequence and historical fault data of the nuclear power plant, the reliability model maps the prediction results to the expected fault occurrence time and a baseline confidence interval, providing a quantitative reference for subsequent early warning. The aforementioned hybrid prediction network and reliability model can run on edge devices or in the cloud, supporting online inference with sliding windows and periodic or triggered batch retraining. It also saves each prediction result, confidence level, and expected failure time as a historical record for subsequent model evaluation and online learning.

[0123] It is understandable that the implementation of the hybrid prediction network, confidence level, and predicted failure time can take many forms. For example, it can directly learn representations using end-to-end temporal convolutions or transformers, or it can employ stepwise feature extraction and fusion. Confidence estimation can be based on Bayesian approximation or quantile regression, or it can use calibration techniques such as temperature scaling. The reliability model can be based on empirical lifetime distributions, or it can be implemented through survival analysis or risk regressors in machine learning. For scenarios with limited computing power, a lightweight predictor with model pruning, quantization, or distillation can be used, and the complex training process can be moved to the cloud. For scenarios with high security requirements, dual-mode operation can be set up (a local lightweight predictor and a high-precision predictor in the cloud run in parallel), with inconsistencies between the two results serving as the trigger condition for manual review. The above parameters, such as window length, overlap rate, prediction step size, and initial width of the confidence interval, can be dynamically adjusted based on field data or learned online to continuously optimize performance.

[0124] In this embodiment, the dynamic adjustment of the confidence interval includes: obtaining the current real-time operating parameters of the nuclear power plant, including load, cooling water inlet temperature and flow rate; normalizing the real-time operating parameters and calculating the operating risk index according to a preset weight; determining the adjustment coefficient of the confidence interval through nonlinear mapping based on the operating risk index; and adjusting the baseline confidence interval according to the adjustment coefficient to obtain the confidence interval.

[0125] Specifically, the dynamic adjustment of the confidence interval involves acquiring the current real-time operating parameters of the nuclear power plant. These parameters include, but are not limited to, unit load, cooling water inlet temperature, cooling water flow rate, and other indicators affecting equipment stress or aging. These operating parameters are then standardized and normalized, and combined into an operational risk index according to preset or online-learned weights. Based on this operational risk index, a confidence interval adjustment coefficient is obtained through a nonlinear mapping function (e.g., sigmoid or other monotonically increasing functions) to expand or contract the baseline confidence interval. Simultaneously, upper and lower limits are applied to the adjusted confidence interval to ensure safety. When real-time operating parameters experience sudden changes or exceed safety thresholds, the system immediately recalculates the confidence interval and generates a temporarily increased risk level, issuing a high-priority alarm if necessary. When real-time parameters are missing or deemed unreliable, the system employs data interpolation, estimation based on nearby time periods, or a more conservative default adjustment coefficient to ensure that the confidence interval is not unduly narrowed. The aforementioned dynamic adjustments can be executed collaboratively at the edge or in the cloud, and the input parameters, calculation process, adjustment coefficients, and results of each adjustment are logged for traceability and compliance review. Meanwhile, the confidence interval mapping strategy and various weights can be optimized online or periodically based on experience data, so that the adjustment strategy can be continuously improved as equipment ages and operating modes change.

[0126] Understandably, the specific implementation of dynamic adjustment of confidence intervals allows for various engineering variations. The weights of operational risk indicators can be initially set by engineering experts based on experience and then fine-tuned on field data with the goal of minimizing false alarms or missed alarms. The nonlinear mapping function can be set manually or learned using a small neural network or tree model. To meet safety redundancy requirements, a dual-track strategy can be set: a regular track for daily fine-tuning, and a safety track using a more conservative expansion coefficient under high-risk conditions. To facilitate operational decision-making, the system can output the confidence intervals before and after adjustment, the adjustment coefficient, and the dominant operational parameters causing the adjustment in the alarm information, allowing operational personnel to quickly determine and handle the situation. Furthermore, to prevent false alarms caused by frequent oscillations, a minimum time interval can be set for changes in the adjustment coefficient, or a smoothing filtering strategy can be used. These various variations and parameter settings facilitate the provision of sufficient implementation examples in the specification, thereby enhancing the enforceability and resistance to rejection of the claims.

[0127] In this embodiment, the fault prediction module 60 is further configured to: obtain the changing trends of displacement prediction value, vibration prediction value and liquid level prediction value, calculate the probability of fault occurrence based on the changing trends; and generate fault warning information when the probability of fault occurrence exceeds a preset probability threshold, the fault warning information including the fault type and the expected occurrence time.

[0128] Specifically, the fault prediction module 60 first processes the multi-step prediction output sequence from the long short-term memory network model. The prediction sequences are organized into three groups according to the variable categories (axial displacement d, radial vibration v, and bearing fluid level l). For each predicted sequence, the module computes a normalized excess measure. Then, the normalized excess measures are linearly weighted and mapped to a single fault occurrence probability using a sigmoid function. The fault type is determined by the variable that contributes the most, for example, when w... d z d =max{w d z d ,w v z v ,w l z l When the fault type is determined to be "abnormal displacement", the fault type is determined to be "abnormal vibration" when the vibration item is at its maximum; the fault type is determined to be "insufficient lubrication / oil leakage" when the liquid level item is at its maximum or the liquid level deviation direction is negative.

[0129] Regarding the calculation of the Estimated Time of Occurrence (ETA), the fault prediction module 60 uses the following method: for each prediction sequence, it finds the earliest moment on the time axis when the corresponding safety threshold is crossed, i.e., it calculates the minimum k. * Make The condition for exceeding the limit is met; if multiple variables exceed the limit simultaneously, the earliest k is selected. * As the overall estimated occurrence time; if the predicted sequence does not exceed the limit within a given window, the approximate time to reach the threshold can be estimated based on linear or quadratic fitting of the sequence. The module records the estimated occurrence time in the fault warning information in both absolute time (e.g., "2025-XX-XX 14:30") and relative time (e.g., "estimated in 12 minutes"). When the calculated P fault Exceeding the preset probability threshold P th When the time limit is reached (or the dynamic probability threshold is exceeded, see below), the system generates and issues a fault warning message, which includes: fault type, fault occurrence probability P. fault The expected occurrence time, main contributing characteristics, and confidence interval are provided to enable maintenance personnel to make quick decisions and take appropriate actions.

[0130] Understandably, this calculation process considers both the most severe deviation of each monitored variable within the prediction window and the fusion of multivariate information into a single probability index through weighted averages and S-shaped mapping, facilitating unified decision-making. By clearly defining the expected occurrence time and main contributing characteristics of the early warning, operations and maintenance personnel can prioritize and take targeted measures. Furthermore, this fault occurrence probability can be used in conjunction with the anomaly score output by fault detection modules (isolated forest, reconstruction error, etc.) to reduce the false alarm rate. For example, a fusion scoring strategy can be adopted, and the fused decision-making basis can be simultaneously provided in the early warning information, improving the interpretability and reliability of the alarm.

[0131] In this embodiment, the fault prediction module 60 is further configured to: obtain the current operating status parameters of the nuclear power plant; calculate the dynamic probability threshold using an S-shaped function based on the current operating status parameters; and generate fault warning information based on the dynamic probability threshold.

[0132] Specifically, the fault prediction module 60 obtains the set of operating status parameters x = [x1, x2, ..., x] of the nuclear power plant in real time from the monitoring system or data analysis module. m The parameters include, but are not limited to, cooling water flow rate, inlet water temperature, speed / load, ambient temperature, and power supply voltage. The module first standardizes each parameter:

[0133]

[0134] Where μ i ,σ i The parameters are x respectively i The historical mean and standard deviation (which can be updated periodically). Then calculate the running severity index (SeverityIndex) S:

[0135]

[0136] Where α i The weighting coefficients reflect the impact of various operating parameters on the risk of the drum filter. These weights can be determined through historical data regression or expert scoring and adjusted online. Based on this severity indicator S, the module uses a sigmoid function to generate a dynamic probability threshold P. th (S):

[0137]

[0138] Where P min With P max Here, S and S0 represent the lower and upper limits of the threshold, respectively, k is the curvature coefficient, and S0 is the center point (i.e., the threshold is at the midpoint when S = S0). In practical settings, when the severity of the operation increases (S increases), k can be set to < 0 or the weight direction can be adjusted to make P... th The threshold is lowered to improve the sensitivity of early warnings under high-risk conditions; conversely, the threshold is increased to reduce false alarms under low-risk conditions. The calculated P... th (S) and the probability of failure P fault Comparison: When P fault >P th When (S) is triggered, the system generates a fault warning message; otherwise, it does not trigger or only generates a low-level prompt.

[0139] Understandably, employing a dynamic probability threshold based on a sigmoid function allows the system threshold to no longer be a fixed static value, but rather automatically adjusted according to the on-site operating conditions, thus balancing the needs of detection sensitivity and false alarm suppression. This mechanism allows for the automatic lowering of the trigger threshold to provide early warning during high-load, abnormal environments, or critical operating phases of the nuclear power plant, while raising the threshold during stable, low-load phases to avoid unnecessary interference. The weighting coefficients of the operational severity index S, the upper and lower limits of the threshold, and the sigmoid parameters (k, S0) can all be trained offline or optimized online using historical fault samples and operational experience; simultaneously, the system records the value of the dynamic threshold and the triggering decision for each instance, facilitating post-event auditing and model improvement. Furthermore, to ensure safety, the system supports forcing the dynamic threshold to a protection value and triggering corresponding protection or shutdown commands under emergency conditions, thereby linking predictive warnings with safety protection and improving the overall reliability of nuclear power plant operation.

[0140] In this embodiment, the fault detection module 40 constructs an isolated forest based on multiple isolated trees. Each isolated tree is generated through the following steps: randomly selecting a data feature; randomly selecting a split point within the selected feature range; dividing the data into two parts according to the split point, and recursively repeating until the stopping condition is met.

[0141] Specifically, isolated forests generate multiple isolated trees using the methods described above, and utilize the overall structure of the forest to identify outliers.

[0142] Understandably, this method can efficiently process multidimensional datasets and ensure anomaly detection accuracy in complex environments.

[0143] In this embodiment of the application, the anomaly score is calculated from the average path length of all isolated trees, where the path length of each isolated tree is the number of steps required to get from the root node to the leaf node of the data point.

[0144] Specifically, by calculating the path length, the isolated forest model quantifies the degree of anomaly of data points and generates anomaly scores with reference value.

[0145] Understandably, this scoring method can intuitively reflect the degree of deviation of data points from the normal distribution, which helps to quickly locate potential problems.

[0146] In an exemplary embodiment, the operation of a drum filter condition monitoring system for a nuclear power plant is as follows:

[0147] When the drum filter starts operating, the sensor module 10 monitors its operating status in real time. Specifically, the axial laser sensor 100 and radial laser sensor 200 collect data on the axial displacement, radial displacement, and gear runout of the filter; the axial eddy current sensor 300 and radial eddy current sensor 400 monitor the operating status and displacement changes of the bearings; and the left bearing level sensor 500 and right bearing level sensor 600 detect changes in the lubricating oil level in the spindle oil tank. This real-time data is transmitted to the data acquisition module 20 via a high-precision interface.

[0148] The data acquisition module 20 samples and preprocesses the received sensor data. The preprocessing includes signal denoising, data format standardization, and missing value handling, resulting in high-quality preprocessed data. Subsequently, the data analysis module 30 normalizes the preprocessed data, unifying the sensor data to the same scale range and extracting feature values ​​such as the mean displacement, the spectral characteristics of vibration, and the fluctuation trend of liquid level. These features are integrated into the data to be detected for subsequent analysis.

[0149] The fault detection module 40 receives the data to be detected and calculates anomaly scores for each data point based on the isolated forest algorithm. For example, if the system detects an axial displacement score of 0.85 at a certain moment, exceeding the system's preset threshold of 0.8, the data is determined to be abnormal. The fault detection module then generates an abnormal data record, including the abnormal score, the specific fault type (e.g., excessive axial displacement), and related information. Upon receiving the abnormal data, the alarm module 50 immediately issues an audible and visual alarm and sends a remote notification to the monitoring system, prompting maintenance personnel to inspect and handle the drum filter.

[0150] To further enhance the reliability of equipment operation, a fault prediction module 60 is also deployed in the system. This module utilizes the time-series feature data generated by the data analysis module 30 to construct and train a Long Short-Term Memory (LSTM) network model. The fault prediction module uses this model to predict the axial displacement trend of the drum filter screen within the next hour, generating a displacement prediction value. At a certain moment, if the prediction result indicates that the axial displacement will reach a dangerous threshold within the next 15 minutes, the module generates abnormal prediction data, and the alarm module 50 issues a warning signal, prompting maintenance personnel to take intervention measures in advance.

[0151] Throughout the process, the system combines real-time monitoring, anomaly detection, and trend prediction to comprehensively assess the operating status of the drum filter. By integrating the Isolation Forest algorithm and the LSTM model, the system can not only promptly detect operational anomalies but also effectively predict potential faults, ensuring the stable operation of the nuclear power plant.

[0152] In an exemplary embodiment, the fault prediction module 60 of the nuclear power plant drum filter condition monitoring system operates as follows:

[0153] During normal operation of the drum filter, the data analysis module 30 continuously receives and processes preprocessed data from the sensor module 10, extracting feature data such as displacement, vibration, and liquid level information. This data is organized into a time-series feature sequence in chronological order, forming the input dataset for prediction.

[0154] The fault prediction module 60 uses a constructed Long Short-Term Memory (LSTM) network model to train and predict feature sequences. For example, the data input to the LSTM model includes the axial displacement change, radial vibration amplitude, and dynamic change trend of the lubricating oil level of the drum filter over the past 10 minutes. The model outputs predictions of the axial displacement value, vibration amplitude, and oil level value for the next 15 minutes.

[0155] At a certain moment, the LSTM model calculates a predicted axial displacement of 0.45 mm for the next 15 minutes, while the system's preset threshold is 0.4 mm. Because the predicted value exceeds the threshold, the fault prediction module 60 generates a predicted anomaly data record, including the predicted displacement value, the time of the anomaly, and the predicted anomaly type (e.g., excessive axial displacement). This anomaly data is immediately transmitted to the alarm module 50.

[0156] Upon receiving the predicted abnormal data, the alarm module 50 issues a warning signal, including audible and visual alarm prompts and remote push notifications, to notify maintenance personnel of the potential risk of axial displacement exceeding limits. Simultaneously, the system interface displays detailed information about the abnormal data, such as a trend graph of the predicted displacement value and possible future time points for the anomaly.

[0157] Furthermore, the fault prediction module 60 optimizes the LSTM model using the mean squared error loss function, making the prediction results more accurate. During long-term operation, the system continuously performs adaptive training on the LSTM model, gradually improving prediction performance and reducing the possibility of false positives and false negatives.

[0158] Through this predictive process, the system can issue an early warning 15 minutes in advance when the drum filter screen is likely to experience an axial displacement exceeding the limit, providing maintenance personnel with sufficient time to intervene, such as adjusting the filter screen operating parameters or performing preventive maintenance, thereby effectively preventing the occurrence of the fault.

[0159] Figure 3 This is a schematic flowchart of a method for monitoring the condition of a drum filter used in a nuclear power plant, provided in one embodiment of this application. Figure 3 As shown, the method for monitoring the condition of drum filters used in nuclear power plants includes at least the following steps:

[0160] S100: Monitor the operating status of the drum filter in real time; S200: Receive real-time data from the sensor module, sample and preprocess the real-time data to form preprocessed data; S300: Receive the preprocessed data, normalize the preprocessed data and extract features to form the data to be detected; S400: Calculate the anomaly score of the data to be detected according to the isolated forest algorithm, compare the anomaly score with a preset threshold, and generate abnormal data when the anomaly score is higher than the preset threshold. The abnormal data includes the anomaly score and the fault type; S500: Trigger an alarm signal when abnormal data is received.

[0161] S100: Real-time monitoring of the operating status of the drum filter.

[0162] In this embodiment of the application, the method for monitoring the status of a drum filter in a nuclear power plant includes, in step S100, real-time monitoring of the operating status of the drum filter. The specific monitoring method can be referred to... Figure 1 , Figure 2 The details and their corresponding descriptions will not be repeated here.

[0163] S200: Receives real-time data from the sensor module and samples and preprocesses the real-time data to form preprocessed data.

[0164] In this embodiment of the application, the method for monitoring the condition of a drum filter in a nuclear power plant includes, in step S200, receiving real-time data from a sensor module, and sampling and preprocessing the real-time data to form preprocessed data. The specific processing method can be referred to... Figure 1 , Figure 2 The details and their corresponding descriptions will not be repeated here.

[0165] S 300: Receives preprocessed data, normalizes the preprocessed data, and extracts features to form the data to be detected.

[0166] In this embodiment of the application, the method for monitoring the condition of drum filters used in nuclear power plants includes receiving preprocessed data, normalizing the preprocessed data, and extracting features to form data to be detected. Specific processing methods can be referred to... Figure 1 , Figure 2 The details and their corresponding descriptions will not be repeated here.

[0167] S 400: Calculate the anomaly score of the data to be detected based on the isolated forest algorithm, compare the anomaly score with a preset threshold, and generate anomaly data when the anomaly score is higher than the preset threshold. The anomaly data includes the anomaly score and the fault type.

[0168] In this embodiment of the application, the method for monitoring the condition of drum filters in nuclear power plants includes step S400, which involves calculating anomaly scores for the data to be detected using the isolated forest algorithm, comparing the anomaly scores with a preset threshold, and generating anomaly data when the anomaly score is higher than the preset threshold. The anomaly data includes anomaly scores and fault types. Specific processing methods can be found in [reference needed]. Figure 1 , Figure 2 The details and their corresponding descriptions will not be repeated here.

[0169] S 500: Triggers an alarm signal when abnormal data is received.

[0170] In this embodiment of the application, the method for monitoring the status of drum filters in nuclear power plants includes triggering an alarm signal upon receiving abnormal data in step S500. The specific processing method can be found in [reference needed]. Figure 1 , Figure 2 The details and their corresponding descriptions will not be repeated here.

[0171] Figure 4 This is an electronic device 2 provided in one embodiment of this application. For example... Figure 4 As shown, the electronic device 2 includes at least the following components: processor 21 and memory 22.

[0172] In this embodiment, the memory 22 is used to store executable instructions of the processor 21, which, when configured to execute instructions, implement... Figure 1 The method for monitoring the condition of drum filters used in nuclear power plants is shown.

[0173] In this application embodiment, a computer-readable storage medium includes instructions that instruct a device to perform a method for monitoring the condition of a drum filter for a nuclear power plant, as described in the second aspect. For example, the instructions instruct the device to perform... Figure 1 The method for monitoring the condition of drum filters used in nuclear power plants is shown in steps S100 to S500.

[0174] In one embodiment of this application, the program operating in the electronic device 2 may be a program that controls a central processing unit (CPU) or similar device to achieve the functions of the above-described embodiments of the present invention (a program that enables the computer to function). Then, the information processed by these devices is temporarily stored in random access memory (RAM) during processing, and subsequently stored in various ROMs such as read-only memory (Flash ROM) and hard disk drives (HDDs), and read, corrected, and written by the CPU as needed.

[0175] It should be noted that a portion of the electronic device 2 described above can also be implemented using a computer. In this case, the program for implementing the control function can be recorded on a computer-readable recording medium, and the program recorded on the recording medium can be read into the computer and executed.

[0176] It should be noted that the "computer" mentioned here refers to a computer built into electronic device 2, which employs hardware including an operating system and peripheral devices. Furthermore, "computer-readable recording media" refers to removable media such as floppy disks, magneto-optical disks, ROMs, and CD-ROMs, as well as storage devices such as hard drives built into the computer.

[0177] Furthermore, a "computer-readable recording medium" can include: a medium that dynamically stores a program for a short period of time, such as a communication line used when transmitting a program via a network such as the Internet or a communication line such as a telephone line; or a medium that stores a program for a fixed period of time, such as volatile memory inside a computer that serves as a server or client in this case. In addition, the aforementioned program can be a program used to implement the above-mentioned functions, or it can be a program that can implement the above-mentioned functions by combining with programs already recorded in the computer.

[0178] Furthermore, the electronic device 2 in the above embodiments can also be implemented as an assembly (device group) composed of multiple devices. Each device constituting the device group can possess some or all of the functions or functional blocks of the electronic device 2 in the above embodiments. As a device group, it is sufficient to have all the functions or functional blocks of the electronic device 2.

[0179] It is understood that the nuclear power plant drum filter condition monitoring system 1, method, and equipment 2 provided in this application embodiment monitor the key operating states of the drum filter (such as axial displacement, radial displacement, gear ring runout, and lubricating oil level) in real time through the sensor module 10. Combined with the isolated forest algorithm and the Long Short-Term Memory (LSTM) network model, it achieves accurate detection and prediction of the filter state, issuing early warning signals before faults occur, significantly improving the safety and reliability of nuclear power plant equipment operation. The system introduces the isolated forest algorithm to calculate anomaly scores and combines it with the LSTM network to analyze and predict time-series data. This allows for accurate identification of potential fault types and locations of the drum filter, generating predicted anomaly data, supporting advance maintenance planning, and reducing downtime and repair costs caused by equipment failures. The alarm module can not only trigger alarms when the anomaly score exceeds a threshold but also generate multi-level alarm signals based on predicted anomaly data, guiding maintenance personnel to prioritize serious faults, further improving operation and maintenance efficiency and the safety factor of nuclear power plant operation.

[0180] Those skilled in the art should recognize that the above embodiments are only used to illustrate this application and are not intended to limit this application. Any appropriate changes and variations made to the above embodiments within the essential spirit and scope of this application fall within the scope of protection claimed in this application.

Claims

1. A condition monitoring system for drum filters used in nuclear power plants, characterized in that, The system includes: The sensor module is used to monitor the operating status of the drum-shaped filter screen in real time. The sensor module includes: an axial laser sensor for monitoring the axial displacement and deformation of the drum-shaped filter screen; a radial laser sensor for monitoring the radial displacement and gear runout of the drum-shaped filter screen; an axial eddy current sensor and a radial eddy current sensor for monitoring the bearing operating status and displacement changes of the drum-shaped filter screen; and a left bearing level sensor and a right bearing level sensor for monitoring changes in the lubricating oil level in the spindle oil tank of the drum-shaped filter screen. The data acquisition module is used to receive real-time data from the sensor module and to sample and preprocess the real-time data to form preprocessed data. The data analysis module is used to receive the preprocessed data, and to normalize and extract features from the preprocessed data to form the data to be detected. The fault detection module is used to calculate the anomaly score of the data to be detected according to the isolated forest algorithm, compare the anomaly score with a preset threshold, and generate abnormal data when the anomaly score is higher than the preset threshold. The abnormal data includes the anomaly score and the fault type. An alarm module is used to issue an alarm signal when the abnormal data is received; The fault prediction module is used to acquire the time-series information in the data to be detected, process the data to be detected according to the time-series information to form a feature sequence, wherein the data to be detected includes displacement information, vibration information and liquid level information; construct a long short-term memory network model, and input the feature sequence into the long short-term memory network model to obtain the displacement prediction value, vibration prediction value and liquid level prediction value of the drum filter. When the predicted displacement value, the predicted vibration value, or the predicted liquid level value exceeds the corresponding preset range, predicted abnormal data is generated. The alarm module is also used to issue an alarm signal when it receives the predicted abnormal data.

2. The nuclear power plant drum filter condition monitoring system according to claim 1, characterized in that, The fault prediction module is also used for: A hybrid prediction network is constructed based on time-series features at multiple scales to output a prediction sequence, which includes the displacement prediction value, the vibration prediction value, or the liquid level prediction value and the corresponding confidence level. Based on the predicted sequence and historical fault data of the nuclear power plant, the expected fault occurrence time and corresponding confidence interval are output through the reliability model, wherein the confidence interval is dynamically adjusted according to the real-time operating parameters of the nuclear power plant.

3. The nuclear power plant drum filter condition monitoring system according to claim 2, characterized in that, The dynamic adjustment of the confidence interval includes: The real-time operating parameters of the nuclear power plant are obtained, including load, cooling water inlet temperature, and flow rate. The real-time operating parameters are normalized and operating risk indicators are calculated according to preset weights; The adjustment coefficient of the confidence interval is determined by nonlinear mapping based on the operational risk indicators. The baseline confidence interval is adjusted according to the adjustment factor to obtain the confidence interval.

4. The nuclear power plant drum filter condition monitoring system according to claim 1, characterized in that, The fault detection module constructs an isolated forest based on multiple isolated trees, and each isolated tree is generated through the following steps: Randomly select a data feature; Randomly select a segmentation point within the selected feature range; Divide the data into two parts according to the split point, and recursively repeat until the stopping condition is met.

5. The condition monitoring system for drum-shaped filters in nuclear power plants according to claim 4, characterized in that, The anomaly score is calculated from the average path length of all the isolated trees, where the path length of each isolated tree is the number of steps required to get from the root node to the leaf node of the data point.

6. The condition monitoring system for drum-shaped filters in nuclear power plants according to claim 1, characterized in that, The fault detection module calculates the anomaly score according to the anomaly scoring formula, which is: Where s(x) is the anomaly score, E(h(x)) is the average path length of point x, and c(n) is the normalization factor.

7. The condition monitoring system for drum-shaped filters in nuclear power plants according to claim 1, characterized in that, The fault prediction module is also used for: Obtain the changing trends of the displacement prediction value, the vibration prediction value, and the liquid level prediction value, and calculate the probability of failure based on the changing trends; When the probability of the fault occurrence exceeds a preset probability threshold, a fault warning message is generated, which includes the fault type and the expected occurrence time.

8. The condition monitoring system for drum-shaped filters in nuclear power plants according to claim 7, characterized in that, The fault prediction module is also used for: Obtain the current operating status parameters of the nuclear power plant; Based on the current operating status parameters, a dynamic probability threshold is calculated using a S-shaped function; The fault warning information is generated based on the dynamic probability threshold.

9. A method for monitoring the condition of a drum filter used in a nuclear power plant, applied to the drum filter condition monitoring system for nuclear power plants as described in any one of claims 1 to 8, characterized in that, The method includes: Real-time monitoring of the operating status of the drum filter; Receive real-time data from the sensor module, and sample and preprocess the real-time data to form preprocessed data; The preprocessed data is received, and the preprocessed data is normalized and features are extracted to form the data to be detected. Anomaly scores are calculated for the data to be detected using the isolated forest algorithm. The anomaly scores are compared with a preset threshold. When the anomaly scores are higher than the preset threshold, anomaly data is generated. The anomaly data includes the anomaly scores and the fault type. An alarm signal is triggered upon receiving the abnormal data.

10. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to implement the method for monitoring the status of a drum filter for a nuclear power plant as described in claim 9 when executing the instructions.

Images