Method and system for predicting long-term life and evaluating health of a DCS system card of a nuclear power plant

Abstract

The application belongs to the technical field of long-term health evaluation of DCS system, aims to solve the problem of insufficient long-period prediction accuracy in traditional health evaluation, and discloses a nuclear power plant DCS system card device long-term life prediction and health evaluation method and system. The method determines the health degree reference path of each type of device, calculates the health degree path slope of each device at the current time, calculates the health degree slope similarity of the target device and the same type device at the same relative operation time, and then calculates the long-term health degree of the device to realize long-term health evaluation of the device. The system comprises a data acquisition module, a reference path generation module, a real-time health degree data acquisition module, a health degree path slope calculation module, a health degree slope similarity weight module, a long-term health prediction module and a long-term health evaluation module. The application better realizes the prediction of long-term health of the device, and provides dynamic and traceable decision basis for preventive maintenance and life management of the nuclear power DCS system.

Description

Technical Field

[0001] The present application belongs to the technical field of long-term health evaluation of DCS systems, and in particular relates to a method and system for long-term life prediction and health evaluation of DCS system components in nuclear power plants. Background Art

[0002] With the development of digital technology, all newly built power plants and most operating nuclear power plants currently use DCS (digital instrumentation and control system) technology, and some old power plants that use analog instruments have also gradually undergone digital control transformation.

[0003] As the nervous system and brain of a nuclear power plant, DCS has a global and systemic impact on the units. If a deviation occurs in any link during the DCS operation, it may cause the degradation of related nuclear power plant equipment. In severe cases, it may even lead to consequences such as reactor shutdown, turbine shutdown, and unit power reduction, posing a huge challenge to the safe, reliable, and economical operation of the nuclear power plant. Therefore, conducting research on DCS health management technology and gradually improving the reliability and intelligent operation and maintenance level of DCS play a vital role in the safe and stable operation of the units.

[0004] Existing nuclear power plant health status assessment methods have the following shortcomings:

[0005] Regarding evaluation indicators, shortcomings include incompleteness and irrational weighting. Regarding incompleteness, some methods lack comprehensive evaluation metrics, focusing solely on the operational status of hardware devices while lacking effective assessment of software system performance, functional implementation, and interaction with external systems. Regarding irrational weighting, different indicators have varying degrees of impact on DCS system health, and some methods lack a scientific basis for determining indicator weights, potentially leading to misjudgments of system health.

[0006] Second, in terms of data collection and processing, there are shortcomings such as incomplete data collection and insufficient data processing capabilities. Regarding incomplete data collection, existing methods fail to collect some health check data from active units, resulting in incomplete data and an inability to fully reflect the operating status of the DCS system. Regarding insufficient data processing capabilities, faced with massive amounts of DCS operating data, some traditional data processing methods struggle to effectively extract valuable information and fail to promptly identify potential problems.

[0007] Third, in terms of fault diagnosis and prediction, there are shortcomings such as limited fault diagnosis capabilities and low fault prediction accuracy. Regarding limited fault diagnosis capabilities, existing methods have certain limitations in fault diagnosis, making it difficult to accurately locate and diagnose some complex or hidden faults. Regarding low fault prediction accuracy, some methods lack effective algorithms and models for fault prediction, resulting in low prediction accuracy and an inability to provide an accurate basis for preventive maintenance.

[0008] Fourth, in terms of maintenance and management, there are shortcomings such as incomplete maintenance strategies and low intelligence levels. Regarding incomplete maintenance strategies, maintenance strategies based on existing evaluation methods are inadequate, lacking specificity and effectiveness, which can lead to over- or under-maintenance. Regarding low intelligence levels, some evaluation methods lack intelligent means and tools, making it impossible to achieve automatic monitoring, diagnosis, and early warning of DCS systems, increasing labor costs and management difficulties. Summary of the Invention

[0009] The main purpose of this application is to provide a long-term life prediction and health assessment method and system for nuclear power plant DCS system cards to solve the problem of lack of long-term prediction capabilities. Existing long-term health assessments are mostly based on short-term trend analysis and cannot effectively predict the health evolution of the equipment throughout its life cycle. This application can better predict the long-term health of the equipment by establishing a historical full life cycle decline curve for similar equipment.

[0010] Another object of the present application is to provide a method and system for long-term life prediction and health assessment of nuclear power plant DCS system components to address the problem of experience dependence. Existing health assessments often rely on expert experience threshold settings and are subject to subjective bias. The present application constructs a health prediction model based on PACE, which self-learns the performance degradation laws of equipment through massive historical data to achieve objective and quantitative health prediction.

[0011] Another purpose of the present application is to provide a method and system for long-term life prediction and health assessment of nuclear power plant DCS system components to solve the problem of delayed dynamic model updates. Conventional static models are difficult to absorb newly generated operation and maintenance data in a timely manner. The model-specific continuous learning mechanism provided by the present application can dynamically integrate newly added historical data, so that the prediction accuracy shows a convergence and improvement trend over time.

[0012] Another purpose of this application is to provide a method and system for long-term life prediction and health assessment of nuclear power plant DCS system components to solve the problem of inaccurate system-level assessment. The existing "short board effect" assessment method tends to exaggerate the impact of a single device failure. The hierarchical aggregation algorithm (equipment → cabinet → system) based on health weights in this application more scientifically reflects the true reliability status of complex systems.

[0013] Another purpose of this application is to provide a method and system for long-term life prediction and health assessment of nuclear power plant DCS system components to solve the problem of small sample prediction difficulties. The industry problem of scarce nuclear power equipment failure data is broken through the "similar equipment data sharing" mechanism, and the group degradation characteristics of the equipment group are used to make up for the lack of individual data. This application can better capture the commonalities and differences of equipment of the same type, and significantly improve the prediction credibility in low-data-volume scenarios.

[0014] In order to achieve the above objectives, this application provides the following technical solutions:

[0015] In a first aspect, the present application provides a method for long-term life prediction and health assessment of DCS system components in a nuclear power plant, comprising:

[0016] Step 1: Determine the health reference path for each type of equipment based on historical equipment health data;

[0017] Step 2: Based on the health data of each device and the health reference path obtained from the DCS real-time health evaluation, calculate the health path slope of each device at the current moment;

[0018] Step 3: Calculate the similarity of the health slopes of the target device and the same type of devices at the same relative operating time by comparing the health path slopes with the health reference path set of each type of device.

[0019] Step 4: Use the health slope similarity as a weight and combine it with the health reference path of the target device of the same type to calculate the long-term health of the device to achieve long-term health evaluation of the device;

[0020] Step 5: Based on the weight parameters used in the DCS real-time health evaluation and the long-term health of the equipment, the long-term health of the cabinet and the DCS system is calculated to achieve long-term health evaluation of the cabinet and the DCS system.

[0021] In some embodiments, offline data is used to construct a health reference path, and the offline data includes DCS equipment information, DCS system configuration, DCS system expert knowledge, maintenance records, alarm logs, reliability data, and spare parts information.

[0022] In some embodiments, in step 2, the health of device i obtained by DCS real-time health evaluation is assumed to be , then the health path slope of device i at the current moment can be obtained through the health sequence pass Calculation, k is the current moment, s i is the slope of the health path of device i at time k.

[0023] In some embodiments, a health reference path for each type of device is obtained based on historical online monitoring data and log data.

[0024] In some embodiments, a kernel function is used to calculate the health slope similarity between each reference path and the target path, and the slope difference of each point between each reference path and the target path is calculated respectively, and the calculated distance is converted into the health slope similarity with the help of the kernel function.

[0025] In some embodiments, the slope difference is calculated using the Euclidean distance formula as follows:

[0026]

[0027] Where x i is the slope data on each reference path, and x is the slope data on the target path.

[0028] In some embodiments, a Gaussian radial basis kernel function is used, and the calculation formula for the health slope similarity is as follows:

[0029]

[0030] Where, is the bandwidth of the kernel function, x i is the slope data on each reference path, and x is the slope data on the target path.

[0031] In some embodiments, the weight value corresponding to each failed reference device is determined based on the similarity between each failed reference path and the target path, the remaining life of the target device is calculated based on the remaining life of each failed reference device and the corresponding weight value, and the health of the target device is obtained based on the health and correlation of each failed reference device.

[0032] In some embodiments, the target device health is calculated by taking the similarity between the target path and each reference path as a weighted average of the health, and the calculation formula is as follows:

[0033]

[0034]

[0035] Where, is the health of the target device, is the health of the i-th failed reference device, n is the number of failed reference devices, is the weight value corresponding to the i-th failed reference device.

[0036] In some embodiments, the PACE-based health prediction model is:

[0037] The improved model uses slope to calculate similarity instead of traditional numerical similarity calculation, which improves the model's dynamics and trend capture capabilities. Using numerical similarity calculations cannot capture trend characteristics, while using slope similarity calculations increases trend sensitivity due to the introduction of differences.

[0038] The kernel function is used to calculate the similarity between each reference path and the target path. First, the slope difference of each point between each reference path and the target path is calculated using the Euclidean distance method:

[0039]

[0040] Where x i is the slope data on each reference path, and x is the slope data on the target path;

[0041] The distance calculated above is converted into similarity with the help of kernel function. Specifically, Gaussian radial basis kernel function is used. Using this kernel function, the similarity value is smaller for larger distances and larger for smaller distances. The similarity value is calculated as follows:

[0042]

[0043] Where, Refers to the bandwidth of the kernel function, usually taken as , Indicates a small bandwidth situation. When the distance is close to 0, the weight generated will be very large. , a larger weight will be generated in a wider distance range.

[0044] In a second aspect, the present application provides a long-term life prediction and health assessment system for DCS system components in a nuclear power plant, comprising:

[0045] Data acquisition module, used to obtain offline data and online data;

[0046] A reference path generation module is used to store and generate a set of health reference paths for each type of equipment based on historical equipment health data;

[0047] Real-time health data acquisition module, used to obtain the health data of each device obtained from the DCS real-time health evaluation in real time;

[0048] The health path slope calculation module is used to calculate the health path slope of each device at the current moment based on real-time health data and health reference path;

[0049] The health slope similarity weight module is used to calculate the health slope similarity as the weight by comparing the health slope of the target device with that of the same type of devices;

[0050] Long-term health prediction module, used to predict the long-term health of target devices based on weights and reference paths of similar devices;

[0051] The long-term health evaluation module is used to integrate the long-term health of the target equipment with the DCS real-time weight parameters to calculate and output the long-term health of the cabinet and DCS system.

[0052] Compared with the existing technology, the long-term life prediction and health assessment method and system for DCS system cards in nuclear power plants provided by this application have the following beneficial effects:

[0053] This application has long-term prediction capabilities. By establishing a historical full life cycle decline curve for similar equipment, this application can better predict the long-term health of the equipment.

[0054] This application can better capture the commonalities and differences of devices of the same type, significantly improving the prediction credibility in low-data-volume scenarios.

[0055] Furthermore, the hierarchical aggregation algorithm (device → cabinet → system) based on health weights in this application more scientifically reflects the true reliability status of complex systems.

[0056] Furthermore, this application can significantly enhance long-term predictive maintenance capabilities. By constructing a PACE-based health prediction model for equipment degradation trajectories, it can achieve visual prediction of the long-term health of equipment → cabinet → system in the nuclear power DCS field.

[0057] Furthermore, the present application constructs a health prediction model based on PACE, which self-learns the performance degradation laws of equipment through massive historical data to achieve objective and quantitative health prediction.

[0058] Furthermore, the model provided in this application has a unique continuous learning mechanism that can dynamically integrate newly added historical data, so that the prediction accuracy shows a convergence and improvement trend over time.

[0059] Furthermore, this application features dynamic learning and evolution capabilities. The PACE-based health prediction model features online incremental learning, accelerating the convergence of prediction errors with each new piece of equipment health data. This self-optimization mechanism ensures the system continuously adapts to time-varying factors such as equipment aging and environmental changes.

[0060] Furthermore, this application realizes visualization of the entire life cycle and constructs a three-dimensional spatiotemporal heat map of equipment health (time × location × health value), which can trace the performance degradation path of any equipment in the past few years and provide data support for technological transformation and spare parts reserves.

[0061] Furthermore, the present application realizes the advancement of the safety barrier, because it has a longer-term health prediction capability, can identify hidden degradation early, and advance the discovery node of potential faults.

[0062] Furthermore, this application realizes knowledge accumulation and inheritance, stores the equipment degradation laws, and forms a database containing multiple types of typical degradation patterns, solving the pain point problem of difficulty in inheriting the experience of experts in the nuclear power industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solution of this application, the following is a brief introduction to the drawings required for the technical description.

[0064] Figure 1 A flowchart of a method for long-term life prediction and health assessment of DCS system components in a nuclear power plant provided in an embodiment of the present application;

[0065] Figure 2 A full flow chart of the long-term health evaluation of the DCS system provided in the embodiment of this application;

[0066] Figure 3 PACE flow chart provided for the embodiments of this application;

[0067] Figure 4 A schematic diagram of the target device fitting analysis provided in an embodiment of the present application;

[0068] Figure 5 This is a structural diagram of the long-term life prediction and health assessment system for DCS system cards in a nuclear power plant provided in an embodiment of the present application. DETAILED DESCRIPTION

[0069] The following is further explained in detail through specific implementation methods.

[0070] like Figures 1 to 4 As shown, the embodiment of the present application provides a method for long-term life prediction and health assessment of DCS system components in a nuclear power plant, including:

[0071] Step 1: Determine the health reference path for each type of equipment based on historical equipment health data;

[0072] Step 2: Based on the health data of each device obtained from the DCS real-time health evaluation and the health reference path in step 1, calculate the health path slope of each device at the current moment;

[0073] Step 3: Calculate the similarity of the health slopes of the target device and the same type of devices at the same relative operating time using the health path slope calculated in step 2 and the health reference path set of each type of device in step 1;

[0074] Step 4: Use the health slope similarity in step 3 as a weight, and combine it with the health reference path of the target device of the same type to calculate the long-term health of the device to achieve long-term health evaluation of the device;

[0075] Step 5: The weight parameters used in the DCS real-time health evaluation are combined with the equipment long-term health in step 4 to calculate the long-term health of the cabinet and DCS system, so as to achieve the long-term health evaluation of the cabinet and the DCS system.

[0076] In the embodiment of the present application, offline data is used to construct a health reference path, as shown in Table 1.

[0077] Table 1 Offline data of long-term health assessment

[0078]

[0079] The online data is processed by the DCS real-time health assessment method, including monitoring data and log data, as shown in Table 2.

[0080] Table 2 Long-term health evaluation online data

[0081]

[0082] In an embodiment of the present application, historical fault records and reliability data in offline data are used to calculate the health reference path of each device. The health reference path reflects the trend of health changes over time. Devices of the same type include reference paths calculated by multiple devices.

[0083] The health calculated from real-time health evaluation accumulates over time to form a target path. The health slopes of multiple reference paths and the target path at the same relative run time are calculated and then converted into a path similarity metric. The health of the target device, known as the target device health, is estimated by using the historical health of each device of the same type and the corresponding weights for similarity. This in turn leads to the calculation of the health evaluation of the DCS cabinet and system.

[0084] The Path Classification and Estimation (PACE) algorithm provided in this application includes two sub-steps: path acquisition and health prediction. Figure 3 Steps 1 to 4 include the implementation process of the PACE algorithm, namely, obtaining the health reference path and predicting the long-term health of the equipment.

[0085] Path acquisition: The historical health data of the target device and multiple failed reference devices are obtained respectively. The target path is obtained from the online data of the target device, and multiple reference paths are obtained from the historical health data of each failed reference device.

[0086] Health prediction: Calculate the similarity between each reference path and the target path respectively, and use the obtained similarity between each reference path and the target path and the health of each failed reference device of the same type to estimate the long-term health of the target device. The amount of data used to achieve prediction based on the health data of the target device and each failed reference device is large, and the health data of the target device, each failed reference device, and the correlation between the target device and each failed reference device can be fully utilized to achieve device health prediction, effectively improving the accuracy of product health prediction. Health prediction can be achieved based on the health data of the target device and the failed reference device, without the need to know the failure threshold of the device in advance. Compared with the traditional prediction method based on the trend of degradation data changes, dynamic device health prediction can be achieved.

[0087] In step 1, set the device type to , a total of Types, namely , for a certain device type , assuming there is The historical health data of the same type of devices, the recording time of each device is from the time the device is online to the time the device fails, such as ,So One of its kind The health reference path is its Health The health sequence .

[0088] In step 2, the method used in this application has the characteristics of strong adaptability. Any method to calculate the health data of each device obtained by DCS real-time health evaluation, that is, to calculate the real-time health, is applicable to this method. The health of (current health at that moment), then the device The slope of the health path at the current moment can be obtained through the health sequence Make the following calculation, that is , is the current moment, s i is the slope of the health path of device i at time k.

[0089] In step 3, the health path of the target device (referred to as the target path) is based on real-time online monitoring data and log data, and is obtained through the accumulation of real-time online data health evaluation over a period of time, such as the device current The target path at time .

[0090] In step 3, when the slopes of the target path and the reference path at the same relative time are similar, a greater similarity is assigned; when they are different, a smaller similarity value should be assigned. If the path similarity between the target device and a failed reference device is low, the data accuracy of the failed reference device is low, and the influence of the failed reference device data in the prediction should be reduced. Conversely, when the path similarity between the target device and a failed reference device is high, the data accuracy of the failed reference device is high, and the influence of the failed reference device data in the prediction should be increased.

[0091] In step 3, the kernel function is used to calculate the similarity between each reference path and the target path. First, the slope difference of each point between each reference path and the target path is calculated separately, using the Euclidean distance calculation method:

[0092]

[0093] Where x i is the slope data on each reference path, and x is the slope data on the target path;

[0094] The distance calculated above is converted into similarity with the help of kernel function. Specifically, Gaussian radial basis kernel function is used. Using this kernel function, the similarity value is smaller for larger distances and larger for smaller distances. The similarity value is calculated as follows:

[0095]

[0096] Where, Refers to the bandwidth of the kernel function, usually taken as , Indicates a small bandwidth situation. When the distance is close to 0, the weight generated will be very large. , a larger weight will be generated in a wider distance range.

[0097] In step 4, the weight value corresponding to each failed reference device is determined according to the similarity between each failed reference path and the target path, and the remaining life of the target device is calculated by the remaining life of each failed reference device and the corresponding weight value, that is, the similarity value obtained by the kernel function above To determine the weight of each failure reference device data, so as to increase the influence of failure reference device data with high similarity and reduce the influence of failure reference device data with low similarity. Ultimately, the health of each failure reference device and its correlation can be comprehensively considered to obtain the accurate target device health.

[0098] The health of the target device is calculated by taking the similarity between the target path and each reference path as the weighted average of the health. The diagram is as follows: Figure 4 As shown, Figure 4 middle, 、 、 、 They represent the health reference paths of four similar devices. The health calculation formula is as follows:

[0099]

[0100]

[0101] Where, is the health of the target device, For the The health of a failed reference device, is the number of failed reference devices, For the The weight value corresponding to each failed reference device.

[0102] Comprehensively evaluate the health of the cabinet based on the health of each target device in the cabinet.

[0103] Comprehensive evaluation of cabinet health relies on the weights of devices in different classifications, which can be determined based on the relative importance of devices in different classifications in Table 3. For example, if the importance of CC1 is 1, the importance of CC2, NC, and RTM devices is 1 / 5, 1 / 7, and 1 / 9, respectively. Their total importance is approximately 1.454968. Dividing the importance of each device classification by the total importance yields weights of 0.687, 0.137, 0.098, and 0.076, respectively. The weighted sum of these weights and the device health yields the cabinet health.

[0104] Table 3 Importance of expert experience knowledge equipment classification and deduction value when equipment fails

[0105]

[0106] In some embodiments, after performing a long-term health evaluation of the cabinet and DCS system, the present application performs auxiliary instrumentation and control operation and maintenance work, specifically including:

[0107] (1) Through the identified faulty or abnormal equipment and related fault information, the expert knowledge base can be called in the health assessment module to find the right equipment fault or maintenance plan to assist in equipment problem troubleshooting and maintenance work;

[0108] (2) If the expert knowledge base does not cover the fault or maintenance data of the relevant equipment, the user can call the fault simulation function in the health assessment module, rely on the DCS parallel system to simulate the operation event, and deduce the development trend of the equipment fault to study the impact of the fault and determine the potential defects of the system;

[0109] (3) For equipment failures that have occurred, users can automatically generate equipment work orders based on the expert knowledge base;

[0110] (4) Users can use the card diagnostic device provided with the platform to diagnose faulty or abnormal equipment, locate faulty components, and facilitate subsequent repairs of the card;

[0111] (5) Every six months or a year, the platform will provide health evaluation reports, including the health status, trends, and number of failures of the DCS system and each cabinet, and provide optimization suggestions for regular testing or preventive maintenance cycles based on this data; at the same time, the platform will provide corresponding optimization suggestions based on various health evaluation indicators (for example, working environment, spare parts status, equipment aging, etc.).

[0112] The long-term health assessment method for DCS systems in this application uses a large amount of device health data accumulated through real-time health assessments to construct a PACE model to predict future health status. The PACE model is a data-driven model that estimates the health of target devices based on the historical health trends of similar devices over time. As historical data continues to accumulate, the model will continue to gain more data support, and predictions of future health will become increasingly accurate. Once the device health is predicted, it will be used to weightedly calculate the cabinet health and system health.

[0113] In addition, if Figure 5 As shown, the embodiment of the present application also provides a long-term life prediction and health assessment system for DCS system cards in a nuclear power plant, including:

[0114] Data acquisition module, used to obtain offline data and online data;

[0115] A reference path generation module is used to store and generate a set of health reference paths for each type of equipment based on historical equipment health data;

[0116] Real-time health data acquisition module, used to obtain the health data of each device obtained from the DCS real-time health evaluation in real time;

[0117] The health path slope calculation module is used to calculate the health path slope of each device at the current moment based on real-time health data and health reference path;

[0118] The health slope similarity weight module is used to calculate the health slope similarity as the weight by comparing the health slope of the target device with that of the same type of devices;

[0119] A long-term health prediction module, configured to predict the long-term health of the target device based on the weight and reference paths of similar devices;

[0120] The long-term health evaluation module is used to integrate the long-term health of the target equipment with the DCS real-time weight parameters to calculate and output the long-term health of the cabinet and DCS system.

[0121] The long-term life prediction and health assessment system for DCS system cards in nuclear power plants provided in the above embodiment can implement the technical solution described in the above method embodiment. The specific implementation principles of the above modules can be found in the corresponding contents of the above method embodiment and will not be repeated here.

[0122] In addition, an embodiment of the present application also provides an electronic device including a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the above-mentioned method for long-term life prediction and health assessment of nuclear power plant DCS system card components is implemented.

[0123] In addition, an embodiment of the present application also provides a computer-readable storage medium, on which computer-readable instructions are stored. When the computer-readable instructions are executed, the above-mentioned nuclear power plant DCS system card component long-term life prediction and health assessment method is implemented.

[0124] Computer-readable media include permanent and non-permanent, removable and non-removable media that can store information using any method or technology. The information can be computer-readable instructions, data structures, program modules or other data.

[0125] The above description is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any changes or replacements that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed in this application should be covered by the scope of protection of the present application.

Claims

1. A method for long-term life prediction and health assessment of DCS system components in a nuclear power plant, characterized in that: include: Step 1: Determine the health reference path for each type of equipment based on historical equipment health data; Step 2: Based on the health data of each device and the health reference path obtained from the DCS real-time health evaluation, calculate the health path slope of each device at the current moment; Step 3: Calculate the similarity of the health slopes of the target device and the same type of devices at the same relative operating time by comparing the health path slopes with the health reference path set of each type of device. The kernel function is used to calculate the fitness slope similarity between each reference path and the target path. The slope difference of each point between each reference path and the target path is calculated respectively. The calculated distance is then converted into fitness slope similarity using the kernel function. Step 4: Use the health slope similarity as a weight and combine it with the health reference path of the target device of the same type to calculate the long-term health of the device to achieve long-term health evaluation of the device; Step 5: Based on the weight parameters used in the DCS real-time health evaluation and the long-term health of the equipment, the long-term health of the cabinet and the DCS system is calculated to achieve long-term health evaluation of the cabinet and the DCS system.

2. The method for long-term life prediction and health assessment of DCS system components in a nuclear power plant according to claim 1, characterized in that: Offline data is used to build a health reference path. Offline data includes DCS equipment information, DCS system configuration, DCS system expert knowledge, maintenance records, alarm logs, reliability data, and spare parts information.

3. The method for long-term life prediction and health assessment of DCS system components in a nuclear power plant according to claim 1, characterized in that: In step 2, let the health of device i obtained by DCS real-time health evaluation be hd i，k , then the health path slope of device i at the current moment can be obtained through the health sequence H i =(h i,1 ,h i,2 ,…,h i,L ) through s i =hd i,k -h i,k-1 Calculation, k is the current moment, s i is the slope of the health path of device i at time k.

4. The method for long-term life prediction and health assessment of DCS system components in a nuclear power plant according to claim 1, characterized in that: Based on historical online monitoring data and log data, obtain the health reference path for each type of device.

5. The method for long-term life prediction and health assessment of DCS system components in a nuclear power plant according to claim 1, characterized in that: The slope difference is calculated using the Euclidean distance formula as follows: Where x i is a data point on each reference path, and x is a data point on the target path.

6. The method for long-term life prediction and health assessment of DCS system components in a nuclear power plant according to claim 1, characterized in that: Using the Gaussian radial basis kernel function, the calculation formula for the health slope similarity is as follows: Where σ is the bandwidth of the kernel function, x i is a data point on each reference path, and x is a data point on the target path.

7. The method for long-term life prediction and health assessment of DCS system components in a nuclear power plant according to claim 1, characterized in that: The weight value corresponding to each failure reference device is determined based on the similarity between each failure reference path and the target path. The remaining life of the target device is calculated based on the remaining life of each failure reference device and the corresponding weight value. The health of the target device is obtained based on the health and correlation of each failure reference device.

8. The method for long-term life prediction and health assessment of DCS system components in a nuclear power plant according to claim 7, characterized in that: The target device health is calculated by taking the similarity between the target path and each reference path as the weighted average of the health. The calculation formula is as follows: Where RL0 is the health of the target device, RL i is the health of the i-th failed reference device, n is the number of failed reference devices, w i is the weight value corresponding to the i-th failed reference device.

9. A long-term life prediction and health assessment system for DCS system components in a nuclear power plant, characterized in that: The method for long-term life prediction and health assessment of a nuclear power plant DCS system card component according to any one of claims 1 to 8 comprises: Data acquisition module, used to obtain offline data and online data; A reference path generation module is used to store and generate a set of health reference paths for each type of equipment based on historical equipment health data; Real-time health data acquisition module, used to obtain the health data of each device obtained from the DCS real-time health evaluation in real time; The health path slope calculation module is used to calculate the health path slope of each device at the current moment based on real-time health data and health reference path; The health slope similarity weight module is used to calculate the health slope similarity as the weight by comparing the health slope of the target device with that of the same type of devices; Long-term health prediction module, used to predict the long-term health of target devices based on weights and reference paths of similar devices; The long-term health evaluation module is used to integrate the long-term health of the target equipment with the DCS real-time weight parameters to calculate and output the long-term health of the cabinet and DCS system.

Images