Fault diagnosis method, device, medium and equipment for nuclear power control system

Abstract

The application discloses a fault diagnosis method, device, medium and equipment of a nuclear power control system, the method comprising: performing fault mode, influence and harmfulness analysis on a plurality of products in the nuclear power control system to obtain analysis results and constructing a fault-test correlation matrix; based on the analysis results, constructing an analytic hierarchy process model, and calculating the comprehensive weight corresponding to each fault mode in the scheme layer of the analytic hierarchy process model; obtaining the current state information of each test point in the nuclear power control system, matching the current state information of each test point with the fault-test correlation matrix, and obtaining a fuzzy fault group; based on the analytic hierarchy process model, obtaining the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault group, sorting the candidate fault modes according to the comprehensive weight from high to low, and generating a fault diagnosis result of the nuclear power control system. The application greatly improves the diagnosis accuracy under the condition of multiple fault concurrency, shortens the fault diagnosis time, and improves the maintenance efficiency.

Description

Technical Field

[0001] This invention relates to the field of nuclear power control technology, and in particular to a fault diagnosis method, apparatus, medium and equipment for a nuclear power control system. Background Technology

[0002] With the continued growth of global energy demand and increasing environmental awareness, nuclear power, as a highly efficient and clean energy source, is becoming increasingly important. The safe operation of nuclear power plants is the core of nuclear power development, and the reliability and stability of nuclear power control systems are directly related to the safety of nuclear power operation. In the operation of nuclear power control systems, fault diagnosis technology plays a crucial role in the timely detection and handling of system anomalies.

[0003] Existing fault diagnosis fusion technologies for nuclear power control systems typically rely on multi-source information association and integrated processing based on BIT (Build-In-Test) data, system self-diagnostic information, and manual inspection information to achieve optimal test point selection and fault location. However, BIT data is susceptible to the accuracy of test circuits and environmental interference (such as high temperature, radiation, electromagnetic interference, and other special operating conditions in nuclear power plants), leading to false alarms, missed alarms, or information distortion. System self-diagnostic information is mostly generated based on preset fault models, making it difficult to cover complex latent faults, gradual faults, and atypical fault modes in nuclear power control systems. Manual inspection information is also affected by the experience and operational standards of the inspectors, resulting in subjective bias.

[0004] Therefore, there is a need for a fault diagnosis method for nuclear power control systems with high fault diagnosis accuracy. Summary of the Invention

[0005] In view of this, the present invention provides a fault diagnosis method, apparatus, medium and equipment for nuclear power control systems, the main purpose of which is to solve the problem of low accuracy of existing fault diagnosis methods for nuclear power control systems.

[0006] According to one aspect of this application, a fault diagnosis method for a nuclear power plant control system is provided, the method comprising: Failure mode, impact and hazard analysis is performed on multiple products in the nuclear power control system to obtain the analysis results. Based on the analysis results, the system structure, function and preset diagnostic information of the nuclear power control system, a fault-test correlation matrix is ​​constructed, wherein the behavioral fault modes in the fault-test correlation matrix are listed as test points. Based on the analysis results, a hierarchical analysis model is constructed, and the comprehensive weight corresponding to each failure mode in the solution layer of the hierarchical analysis model is calculated. The hierarchical analysis model includes an objective layer, a criterion layer, and a solution layer. Obtain the current status information of each test point in the nuclear power control system, and match the current status information of each test point with the fault-test correlation matrix to obtain a fuzzy fault group; Based on the hierarchical analysis model, the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault group is obtained. According to the analysis results, the fault cause and design improvement scheme corresponding to each candidate fault mode are obtained. The candidate fault modes are sorted from high to low according to their comprehensive weights. Based on the comprehensive weight, fault cause and design improvement scheme corresponding to each candidate fault mode, the fault diagnosis result of the nuclear power control system is generated.

[0007] Optionally, the criterion layer of the hierarchical analysis model includes a severity criterion layer, an occurrence probability criterion layer, and a detectability criterion layer. The step of calculating the comprehensive weight corresponding to each failure mode in the solution layer of the hierarchical analysis model based on the analysis results includes: Based on the analysis results, the occurrence probability, severity, and detectability of each failure mode are obtained. The severity and detectability of each failure mode are converted into risk indicators to obtain the risk severity and risk undetectability of each failure mode. Based on the risk severity, risk undetectability and occurrence probability of each failure mode, the severity criterion layer weight, occurrence probability criterion layer weight and detectability criterion layer weight are calculated. Based on the detectability of each failure mode, calculate the first weight of each failure mode under the detectability criterion layer; based on the severity of each failure mode, calculate the second weight of each failure mode under the severity criterion layer; and based on the occurrence probability of each failure mode, calculate the third weight of each failure mode under the occurrence probability criterion layer. Based on the severity criterion layer weights, the occurrence probability criterion layer weights, the detectability criterion layer weights, and the first, second, and third weights corresponding to each failure mode, the comprehensive weight corresponding to each failure mode is calculated.

[0008] Optionally, the calculation of the severity criterion layer weights, the probability of occurrence criterion layer weights, and the detectability criterion layer weights based on the risk severity, risk undetectability, and occurrence probability corresponding to each failure mode includes: Based on the risk severity corresponding to each failure mode, the average risk severity is calculated; based on the risk undetectability corresponding to each failure mode, the average risk undetectability is calculated; based on the occurrence probability corresponding to each failure mode, the average risk occurrence probability is calculated. A judgment matrix is ​​constructed based on the average risk severity, the average risk undetectability, and the average risk occurrence probability. The eigenvalues ​​of the judgment matrix are calculated to obtain multiple eigenvalues ​​of the judgment matrix. The multiple eigenvalues ​​are then normalized to obtain normalized eigenvalues. The normalized feature values ​​are used as the weights for the severity criterion layer, the probability of occurrence criterion layer, and the detectability criterion layer, respectively.

[0009] Optionally, the step of constructing a judgment matrix based on the average risk severity, the average risk undetectability, and the average risk occurrence probability includes: Calculate the ratio of any two values ​​among the average risk severity, the average risk undetectability, and the average risk occurrence probability; Each ratio is rounded according to the analytic hierarchy process standard to obtain the rounded number of each ratio. Construct a judgment matrix based on the rounded value of each ratio.

[0010] Optionally, the step of using the normalized feature values ​​as weights for the severity criterion layer, the occurrence probability criterion layer, and the detectability criterion layer includes: The maximum value among the normalized feature values ​​is used as the weight of the detectability criterion layer, the median value among the normalized feature values ​​is used as the weight of the severity criterion layer, and the minimum value among the normalized feature values ​​is used as the weight of the occurrence probability criterion layer.

[0011] Optionally, the comprehensive weight for each failure mode is calculated based on the severity criterion layer weight, the occurrence probability criterion layer weight, the detectability criterion layer weight, and the first, second, and third weights corresponding to each failure mode, including: For each fault mode, calculate the first product of the detectability criterion layer weight and the first weight corresponding to the fault mode, calculate the second product of the severity criterion layer weight and the second weight corresponding to the fault mode, calculate the third product of the occurrence probability criterion layer weight and the third weight corresponding to the fault mode, and use the sum of the first product, the second product and the third product as the comprehensive weight corresponding to the fault mode.

[0012] Optionally, the current state information of each test point is matched with the fault-test correlation matrix to obtain a fuzzy fault group, including: The current state information of each test point is compared with a preset threshold, and the actual state value corresponding to the test point is obtained based on the comparison result. If the actual state value corresponding to the test point with an element of 1 in any fault mode in the fault-test correlation matrix is ​​also 1, then any fault mode is a candidate fault mode, and multiple candidate fault modes form a fault fuzzy group.

[0013] According to another aspect of this application, a fault diagnosis device for a nuclear power control system is provided, comprising: The correlation matrix construction module is used to perform fault mode, impact and hazard analysis on multiple products in the nuclear power control system, obtain analysis results, and construct a fault-test correlation matrix based on the analysis results, the system structure and function of the nuclear power control system and the preset diagnostic information. The behavioral fault modes in the fault-test correlation matrix are listed as test points. The comprehensive weight calculation module is used to construct a hierarchical analysis model based on the analysis results and calculate the comprehensive weight corresponding to each failure mode in the solution layer of the hierarchical analysis model. The hierarchical analysis model includes a target layer, a criterion layer, and a solution layer. The fuzzy fault group acquisition module is used to acquire the current status information of each test point in the nuclear power control system, and match the current status information of each test point with the fault-test correlation matrix to obtain the fuzzy fault group. The fault diagnosis result generation module is used to obtain the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault group based on the hierarchical analysis model, obtain the fault cause and design improvement scheme corresponding to each candidate fault mode according to the analysis results, sort the candidate fault modes in descending order of comprehensive weight, and generate the fault diagnosis result of the nuclear power control system based on the comprehensive weight, fault cause and design improvement scheme corresponding to each candidate fault mode.

[0014] Optionally, the comprehensive weight calculation module is further used for: Based on the analysis results, the occurrence probability, severity, and detectability of each failure mode are obtained. The severity and detectability of each failure mode are converted into risk indicators to obtain the risk severity and risk undetectability of each failure mode. Based on the risk severity, risk undetectability and occurrence probability of each failure mode, the severity criterion layer weight, occurrence probability criterion layer weight and detectability criterion layer weight are calculated. Based on the detectability of each failure mode, calculate the first weight of each failure mode under the detectability criterion layer; based on the severity of each failure mode, calculate the second weight of each failure mode under the severity criterion layer; and based on the occurrence probability of each failure mode, calculate the third weight of each failure mode under the occurrence probability criterion layer. Based on the severity criterion layer weights, the occurrence probability criterion layer weights, the detectability criterion layer weights, and the first, second, and third weights corresponding to each failure mode, the comprehensive weight corresponding to each failure mode is calculated.

[0015] Optionally, the comprehensive weight calculation module is further used for: Based on the risk severity corresponding to each failure mode, the average risk severity is calculated; based on the risk undetectability corresponding to each failure mode, the average risk undetectability is calculated; based on the occurrence probability corresponding to each failure mode, the average risk occurrence probability is calculated. A judgment matrix is ​​constructed based on the average risk severity, the average risk undetectability, and the average risk occurrence probability. The eigenvalues ​​of the judgment matrix are calculated to obtain multiple eigenvalues ​​of the judgment matrix. The multiple eigenvalues ​​are then normalized to obtain normalized eigenvalues. The normalized feature values ​​are used as the weights for the severity criterion layer, the probability of occurrence criterion layer, and the detectability criterion layer, respectively.

[0016] Optionally, the comprehensive weight calculation module is further used for: Calculate the ratio of any two values ​​among the average risk severity, the average risk undetectability, and the average risk occurrence probability; Each ratio is rounded according to the analytic hierarchy process standard to obtain the rounded number of each ratio. Construct a judgment matrix based on the rounded value of each ratio.

[0017] Optionally, the comprehensive weight calculation module is further used for: The maximum value among the normalized feature values ​​is used as the weight of the detectability criterion layer, the median value among the normalized feature values ​​is used as the weight of the severity criterion layer, and the minimum value among the normalized feature values ​​is used as the weight of the occurrence probability criterion layer.

[0018] Optionally, the comprehensive weight calculation module is further used for: For each fault mode, calculate the first product of the detectability criterion layer weight and the first weight corresponding to the fault mode, calculate the second product of the severity criterion layer weight and the second weight corresponding to the fault mode, calculate the third product of the occurrence probability criterion layer weight and the third weight corresponding to the fault mode, and use the sum of the first product, the second product and the third product as the comprehensive weight corresponding to the fault mode.

[0019] Optionally, the fuzzy fault group acquisition module is further configured to: The current state information of each test point is compared with a preset threshold, and the actual state value corresponding to the test point is obtained based on the comparison result. If the actual state value corresponding to the test point with an element of 1 in any fault mode in the fault-test correlation matrix is ​​also 1, then any fault mode is a candidate fault mode, and multiple candidate fault modes form a fault fuzzy group.

[0020] According to another aspect of this application, a storage medium is provided that stores at least one executable instruction, which causes a processor to perform an operation corresponding to the fault diagnosis method of the nuclear power control system described above.

[0021] According to another aspect of this application, a computer device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus; The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the fault diagnosis method of the nuclear power control system described above.

[0022] By employing the above-described technical solutions, the technical solutions provided by the embodiments of the present invention have at least the following advantages: This application provides a fault diagnosis method, apparatus, medium, and equipment for nuclear power control systems. It performs fault mode, impact, and hazard analysis on multiple products within the nuclear power control system, obtains the analysis results, constructs a fault-test correlation matrix, and builds a hierarchical analysis model (AHP) based on the analysis results. It calculates the comprehensive weight corresponding to each fault mode in the scheme layer of the AHP model, obtains the current state information of each test point in the nuclear power control system, matches the current state information of each test point with the fault-test correlation matrix to obtain fuzzy fault groups, obtains the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault groups based on the comprehensive weight corresponding to each fault mode in the scheme layer of the AHP model, obtains the fault causes and design improvement schemes corresponding to the candidate fault modes based on the analysis results, and generates fault diagnosis results based on the comprehensive weights, fault causes, and design improvement schemes corresponding to the candidate fault modes. By matching the current state information of the test points with the fault-test correlation matrix, a large number of irrelevant faults are quickly screened out, resulting in fuzzy fault groups. Then, the AHP model incorporates multi-dimensional prior knowledge to accurately rank the candidate fault modes in the fuzzy fault groups, significantly improving the diagnostic accuracy under multiple concurrent fault conditions, shortening fault diagnosis time, and improving maintenance efficiency.

[0023] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0024] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals are used throughout the drawings to denote the same parts. In the drawings: Figure 1 A flowchart of a fault diagnosis method for a nuclear power control system provided in an embodiment of this application is shown; Figure 2 This paper illustrates a structural example of a hierarchical analysis model for a fault diagnosis method for a nuclear power control system provided in an embodiment of this application. Figure 3 This paper presents another flowchart of a fault diagnosis method for a nuclear power control system provided in an embodiment of this application. Figure 4 This paper presents another flowchart of a fault diagnosis method for a nuclear power control system provided in an embodiment of this application. Figure 5 Another flowchart of a fault diagnosis method for a nuclear power control system provided in an embodiment of this application is shown; Figure 6 This paper shows a block diagram of a fault diagnosis device for a nuclear power control system provided in an embodiment of this application.

[0025] in, Figure 6 In the Chinese section: 602-Correlation matrix construction module; 604-Comprehensive weight calculation module; 606-Fuzzy fault group acquisition module; 608-Fault diagnosis result generation module. Detailed Implementation

[0026] The present invention will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present invention can be combined with each other.

[0027] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the specific embodiments, structures, features, and effects according to the present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "an embodiment" or "an embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0028] To address the issue of low accuracy in existing fault diagnosis methods for nuclear power control systems, this application provides a fault diagnosis method for nuclear power control systems, such as... Figure 1 As shown, the method includes: 102: Perform failure mode, effects, and hazard analysis on multiple products in the nuclear power control system, obtain the analysis results, and construct a fault-test correlation matrix based on the analysis results, the system structure and functions of the nuclear power control system, and the preset diagnostic information. 104: Based on the analysis results, construct a hierarchical analysis model and calculate the comprehensive weight of each failure mode in the scheme layer of the hierarchical analysis model. The hierarchical analysis model includes the target layer, the criterion layer and the scheme layer. 106: Obtain the current status information of each test point in the nuclear power control system, match the current status information of each test point with the fault-test correlation matrix to obtain fuzzy fault groups; 108: Based on the hierarchical analysis model, obtain the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault group. According to the analysis results, obtain the fault cause and design improvement scheme corresponding to each candidate fault mode. Sort the candidate fault modes in descending order of comprehensive weight. Based on the comprehensive weight, fault cause and design improvement scheme corresponding to each candidate fault mode, generate the fault diagnosis result of the nuclear power control system.

[0029] Specifically, a comprehensive Failure Mode, Effects, and Criticality Analysis (FMECA) is conducted on multiple products (such as components, boards, equipment, and software) in the nuclear power control system. This process systematically identifies all potential failure modes and analyzes the cause, impact, probability of occurrence (O), severity (S), and detectability (D) of each failure mode. The specific process is as follows: Clearly define the analysis hierarchy (components, boards, equipment, systems, etc.), boundaries, and objectives (such as improving reliability and optimizing fault diagnosis); analyze the function, structure, and operating environment of the object; decompose the analysis object from top to bottom or from bottom to top, clarifying the functions, interfaces, and dependencies of each level, forming a decomposition list. For each decomposition unit, comprehensively list all possible failure modes (such as component short circuits, board signal distortion, and software logic errors), supplemented by historical failure data, expert experience, similar product cases, and failure physics analysis. Identify the direct / indirect causes of the failure mode, analyze the impact of the failure on the unit itself, the next higher level component, the system as a whole, and the end user / safety step by step, and then assess the probability of failure (O), severity (S), and detectability (D).

[0030] Based on the system structure, functional principles, preset diagnostic information, and FMECA analysis results, a binary fault-test correlation matrix (a D×n dimensional matrix) is established for the tested nuclear power control system. The rows represent D identified fault modes (F1, F2, ..., Fd), and the columns represent n system test points or sensors (T1, T2, ..., Tn). The value of the matrix element d_ij is: 1: This indicates that the occurrence of fault mode Fi will cause the output of test point Tj to exceed the normal range (i.e., it can be detected).

[0031] 0: Indicates that the occurrence of fault mode Fi is unrelated to the state of test point Tj or cannot be detected by it.

[0032] The fault-test correlation matrix describes the correlation between system faults and tests under specific conditions. For a given fault, all elements in the fault-test correlation matrix corresponding to the test points reachable by the information flow are "1". However, in engineering practice, signal attenuation can occur during transmission. In such cases, using only a single-state fault-test correlation matrix to diagnose the correlation of the entire system is limited and cannot fully reflect the characteristics of the entire system. This application uses the analytic hierarchy process (AHP) to calculate the relationship between faults and test points and assign weight values, using these weight values ​​to comprehensively reflect fault information.

[0033] Based on the analysis structure of FMECA, a hierarchical analysis model is constructed, such as... Figure 2 As shown, this hierarchical analysis model has three layers, including the target layer, the criterion layer, and the scheme layer.

[0034] Target layer: Fault priority ranking; The criterion layer selects three core criteria affecting fault priority: probability of occurrence (O), severity (S), and detectability (D). Severity (S): Measures the likelihood of a fault occurring, ranging from 0 (very severe) to 10 (very minor). A higher value indicates a less severe consequence and a smaller "risk contribution." Probability of occurrence (O): Measures the severity of the consequences of a fault on system function and safety, ranging from 0 (never will occur) to 1 (will definitely occur). A higher value indicates a higher probability and a greater risk. Detectability (D): Measures the ease with which a fault can be detected promptly and accurately using existing testing methods, ranging from 0 (undiagnosable) to 1 (diagnosable). A higher value indicates easier detection and a smaller "risk contribution." The hierarchical analysis model can handle testing uncertainties (such as unknown or distorted states at certain test points) and, through the trade-offs between criteria such as "detectability," can still provide reasonable fault priorities.

[0035] Based on the analysis results, the occurrence probability, severity, and detectability of each failure mode are obtained. Based on the occurrence probability, severity, and detectability of each failure mode, the weights of each criterion layer are calculated, namely the weights of the severity criterion layer, the occurrence probability criterion layer, and the detectability criterion layer. Based on the occurrence probability, severity, and detectability of each failure mode, the weights of each failure mode under each criterion layer are calculated. Then, based on the weights of the criterion layers and the weights of each failure mode under each criterion layer, the comprehensive weight of each failure mode in the analytic hierarchy process (AHP) model is calculated.

[0036] Sensors and data acquisition units are deployed on the equipment of the nuclear power plant control system. These sensors and data acquisition units serve as test points, acquiring the current status information of each test point (T1, T2, ..., Tn) in real time. After threshold judgment or data processing, the current status information of each test point is converted into a binary real-time test status vector T (an n-dimensional vector), where 1 indicates that an anomaly has been detected at the test point, and 0 indicates that the test point is in a normal state.

[0037] The real-time test state vector T is logically compared with each row of the fault-test correlation matrix (i.e., the correlation vector corresponding to each fault mode). The matching rule is: if the position in the real-time test state vector T corresponding to all elements with a value of 1 in a certain row vector (representing a certain fault Fi) of matrix D also has a value of 1, it means that the currently observed abnormal test mode covers all the anomalies that the fault mode can cause, and therefore the fault mode Fi is included in the candidate fault set—the fault fuzzy group F.

[0038] For each candidate fault mode in the selected fault fuzzy group F, a pre-calculated comprehensive weight is obtained from the hierarchical analysis model. Then, all fault modes in F are sorted from largest to smallest according to this comprehensive weight value. The higher the weight value, the higher the comprehensive probability of the fault mode occurring (taking into account probability, severity, and detectability).

[0039] The sorted fault list is output to maintenance personnel through a human-machine interface. The list not only displays the fault name, but also its priority weight, possible causes, and maintenance suggestions (from the FMECA database), providing the most direct and scientific support for maintenance decisions.

[0040] This application combines correlation and hierarchical analysis models, enabling rapid and accurate fault diagnosis. It is particularly suitable for industrial scenarios with high reliability and safety requirements, such as nuclear power DCS, steam turbine control, and safety instrumentation.

[0041] Compared with existing technologies, this application provides a fault diagnosis method for nuclear power control systems. It performs fault mode, impact, and hazard analysis on multiple products within the nuclear power control system, obtains the analysis results, constructs a fault-test correlation matrix, and builds a hierarchical analysis model (AHP) based on the analysis results. It calculates the comprehensive weight corresponding to each fault mode in the scheme layer of the AHP model, obtains the current state information of each test point in the nuclear power control system, matches the current state information of each test point with the fault-test correlation matrix to obtain fuzzy fault groups, obtains the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault groups based on the comprehensive weight corresponding to each fault mode in the scheme layer of the AHP model, obtains the fault causes and design improvement schemes corresponding to the candidate fault modes based on the analysis results, and generates fault diagnosis results based on the comprehensive weights, fault causes, and design improvement schemes corresponding to the candidate fault modes. By matching the current state information of the test points with the fault-test correlation matrix, a large number of irrelevant faults are quickly screened out, resulting in fuzzy fault groups. Then, the AHP model incorporates multi-dimensional prior knowledge to accurately rank the candidate fault modes in the fuzzy fault groups, significantly improving the diagnostic accuracy under multiple concurrent fault conditions, shortening fault diagnosis time, and improving maintenance efficiency.

[0042] In one embodiment of the present invention, such as Figure 3 As shown, the criterion layer of the hierarchical analysis model includes a severity criterion layer, an occurrence probability criterion layer, and a detectability criterion layer. Based on the analysis results, the comprehensive weight corresponding to each failure mode in the solution layer of the hierarchical analysis model is calculated, including: 302: Based on the analysis results, obtain the occurrence probability, severity, and detectability of each failure mode; 304: The severity and detectability of each failure mode are converted into risk indicators to obtain the risk severity and risk undetectability of each failure mode. Based on the risk severity, risk undetectability and occurrence probability of each failure mode, the severity criterion layer weight, occurrence probability criterion layer weight and detectability criterion layer weight are calculated. 306: Based on the detectability of each failure mode, calculate the first weight of each failure mode under the detectability criterion layer; based on the severity of each failure mode, calculate the second weight of each failure mode under the severity criterion layer; based on the occurrence probability of each failure mode, calculate the third weight of each failure mode under the occurrence probability criterion layer. 308: Based on the severity criterion layer weight, the occurrence probability criterion layer weight, the detectability criterion layer weight, and the first, second, and third weights corresponding to each failure mode, the comprehensive weight corresponding to each failure mode is calculated.

[0043] In this embodiment, based on the analysis results, the occurrence probability, severity, and detectability corresponding to each failure mode are obtained. The occurrence probability is the likelihood of the failure occurring, the severity is the level of seriousness of the consequences after the failure occurs, and the detectability is the ease with which the failure is detected. The raw values ​​of severity and detectability are converted into indicators that reflect risk. The occurrence probability is the risk indicator and does not need to be converted. The global importance ratio of severity, occurrence probability, and detectability at each criterion layer is determined through the risk indicators. For each failure mode, its individual score weights at the three criterion layer dimensions are calculated. Based on the global importance ratio and individual score weights, the comprehensive weight of each failure mode is calculated.

[0044] For example, for each failure mode, its score under the three criteria is calculated (i.e., single-criterion weights).

[0045] Calculate the weight of failure mode F_i under the severity (S_risk) criterion W_s_i = S_risk_i / (S_risk_1 + S_risk_2 + ... + S_risk_d) (i.e., the risk severity value of the failure / the sum of the risk severity values ​​of all failures); Calculate the weight of failure mode F_i under the probability (O_risk) criterion W_o_i=O_risk_i / (O_risk_1+O_risk_2+...+O_risk_d); Calculate the weight W_d_i of failure mode F_i under the undetectability (D_risk) criterion: W_d_i = D_risk_i / (D_risk_1 + D_risk_2 + ... + D_risk_d).

[0046] Through the above calculations, three vectors are obtained: W_s: Weight vector of all failure modes under the severity criterion.

[0047] W_o: The weight vector of all failure modes under the probability criterion.

[0048] W_d: The weight vector of all failure modes under the detectability criterion.

[0049] The weight of each failure mode under each criterion is combined with the weight Wc of the criterion layer to obtain the final comprehensive weight of each failure mode relative to the overall objective (failure priority).

[0050] Formula: W=[W_o, W_s, W_d]×Wc; When expanded, it means: W_i=(W_o_i×w_O_risk)+(W_s_i×w_S_risk)+(W_d_i×w_D_risk), where: W_i: The overall weight of the failure mode F_i.

[0051] W_o_i, W_s_i, W_d_i: The weights of the fault mode F_i under the three single criteria.

[0052] w_O_risk, w_S_risk, w_D_risk: Weights of the probability of occurrence criterion layer, the severity criterion layer, and the detectability criterion layer.

[0053] In one embodiment of the present invention, such as Figure 4 As shown, based on the risk severity, risk undetectability, and occurrence probability corresponding to each failure mode, the weights of the severity criterion layer, the occurrence probability criterion layer, and the detectability criterion layer are calculated, including: 402: Based on the risk severity corresponding to each failure mode, the average risk severity is calculated; based on the risk undetectability corresponding to each failure mode, the average risk undetectability is calculated; based on the occurrence probability corresponding to each failure mode, the average risk occurrence probability is calculated. 404: Construct a judgment matrix based on the average risk severity, the average risk undetectability, and the average risk occurrence probability; 406: Calculate the eigenvalues ​​of the judgment matrix to obtain multiple eigenvalues ​​of the judgment matrix, and normalize the multiple eigenvalues ​​to obtain normalized eigenvalues. 408: The normalized feature values ​​are used as the weights of the severity criterion layer, the probability of occurrence criterion layer, and the detectability criterion layer, respectively.

[0054] Specifically, the ratio of each pair of the average risk severity, average risk undetectability, and average risk occurrence probability is calculated. Each ratio is rounded according to the analytic hierarchy process (AHP) standard to obtain a rounded value for each ratio. Based on the rounded values ​​of each ratio, a judgment matrix is ​​constructed, and the eigenvalues ​​of the judgment matrix are calculated. The maximum value among the normalized eigenvalues ​​is used as the weight of the detectability criterion layer, the median value among the normalized eigenvalues ​​is used as the weight of the severity criterion layer, and the minimum value among the normalized eigenvalues ​​is used as the weight of the occurrence probability criterion layer.

[0055] For example, the severity S and detectability D corresponding to each failure mode are converted into risk indicators: Risk severity (S_risk): S_risk = (10 - S_original) / 10, where S_original is the severity of each failure mode obtained from the analysis results, with 0 indicating the most severe. After conversion, S_risk = 1 indicates the most severe risk.

[0056] Probability of occurrence of risk (O_risk): O_risk = O_original, where O_original is the probability of occurrence of each failure mode obtained from the analysis results. The value of O itself represents the probability, and the larger the value, the higher the risk. No conversion is required.

[0057] Risk undetectability (D_risk): D_risk = 1 - D_original, where D_original is the detectability of each failure mode obtained from the analysis results. The original D_original is 1, which means easy to detect. After conversion, D_risk = 0, which means low risk (easy to detect). The original D_original is 0, which means undiagnosable. After conversion, D_risk = 1, which means high risk (difficult to detect).

[0058] Now, the larger the values ​​of the three new indicators S_risk, O_risk, and D_risk, the greater the risk contribution of the criterion to fault priority.

[0059] Calculate the average of all failure modes on each transformed risk criterion as a benchmark for comparison.

[0060] Calculate the average risk severity (S_risk_avg): S_risk_avg=(S_risk1+S_risk2+...+S_riskd) / d; Calculate the average probability of occurrence of the risk of neutrality (O_risk_avg): O_risk_avg=(O_risk1+O_risk2+...+O_riskd) / d; Calculate the average risk undetectability (D_risk_avg): D_risk_avg=(D_risk1+D_risk2+...+D_riskd) / d; The relative importance of the three risk criteria is determined by comparing their average values. Will The average values ​​of the three risk criteria are compared pairwise, and a judgment matrix is ​​generated based on the comparison values. The eigenvalues ​​[w_S, w_O, w_D] of the judgment matrix A are calculated, and the eigenvalues ​​are normalized to obtain the weights of the severity criterion layer, the probability of occurrence criterion layer, and the detectability criterion layer. w_S_risk = w_S / (w_S + w_O + w_D); w_O_risk = w_O / (w_S + w_O + w_D); w_D_risk = w_D / (w_S + w_O + w_D); The consistency ratio CR of the matrix was calculated to be less than 0.1, which passed the consistency test and proved that the weight allocation was reasonable.

[0061] In one embodiment, based on the severity criterion layer weights, the occurrence probability criterion layer weights, the detectability criterion layer weights, and the first, second, and third weights corresponding to each failure mode, a comprehensive weight corresponding to each failure mode is calculated, including: For each failure mode, calculate the first product of the detectability criterion layer weight and the first weight corresponding to the failure mode, calculate the second product of the severity criterion layer weight and the second weight corresponding to the failure mode, calculate the third product of the occurrence probability criterion layer weight and the third weight corresponding to the failure mode, and use the sum of the first, second and third products as the comprehensive weight corresponding to the failure mode.

[0062] Specifically, the criterion-level weights are the importance percentages assigned to the three dimensions (detectability, severity, and probability of occurrence), while the failure mode weights are the scores for each failure mode under a single criterion dimension. The formula for each dimension is calculated as "criterion-level weight × failure mode weight under that criterion dimension," and then the three products are summed to obtain the overall weight of the failure mode. The higher the overall weight, the more attention this failure requires.

[0063] In one embodiment, the current state information of each test point is matched with the fault-test correlation matrix to obtain a fuzzy fault group, including: comparing the current state information of each test point with a preset threshold, and obtaining the actual state value corresponding to the test point based on the comparison result; if the actual state value corresponding to the test point with an element of 1 in any fault mode in the fault-test correlation matrix is ​​also 1, then any fault mode is a candidate fault mode, and multiple candidate fault modes form a fuzzy fault group.

[0064] Specifically, a normal / abnormal threshold is first set for each test point (for example, a sensor value exceeding 50 is abnormal). Then, the real-time status information of the test point is compared with the threshold to determine whether each test point is actually "normal (0)" or "abnormal (1)". In the fault-test correlation matrix, the rows represent the fault mode and the columns represent the test points. An element of 1 in the i-th row and j-th column means "when the fault in the i-th row occurs, the corresponding test point in the j-th column is abnormal"; an element of 0 means it is irrelevant.

[0065] If the actual state value of all test points with elements equal to 1 corresponding to a certain fault mode is 1, then this fault mode is listed as a candidate fault mode; multiple such candidate fault modes are grouped together to form a fault fuzzy group.

[0066] In one embodiment, the process of the fusion fault diagnosis method based on correlation model and analytic hierarchy process is as follows: Figure 5 As shown, the detailed steps are as follows: Step one involves constructing the FMECA correlation matrix and AHP weights based on the FMECA analysis. FMECA systematically analyzes the potential failure modes, causes, and impacts of components, boards, equipment, and software, and quantitatively assesses the severity (S), probability of occurrence (O), and detectability (D) of each failure mode. The analysis results directly serve as the basis for constructing the FMECA correlation matrix and as input for calculating the AHP model weights.

[0067] Step two involves constructing a fault-test correlation matrix D based on historical system data, design documents, and the system's FMECA analysis results. For a system with 100 fault modes (m=100) and 50 test points (n=50), D is a 100×50 binary matrix. This matrix can be stored in a relational database or a file, and is maintained and version-controlled by the "Correlation Matrix Management Module".

[0068] Step 3: Using the AHP method, construct as follows Figure 2 The hierarchical model is shown. Based on the system FMECA analysis information from step one, the importance of each of the three criteria (failure probability, severity, and detectability) relative to the target is compared pairwise to form a judgment matrix. Finally, the weights of each criterion are combined with the scoring weights of each fault under different criteria to obtain a comprehensive weight vector W (a 100-dimensional vector). This process can be completed offline by the "AHP weight calculation module" or during system initialization. Assuming that after analyzing all failure modes, the average value of the transformed criteria is obtained: S_risk_avg (average risk severity) = 0.40; O_risk_avg (average probability of risk occurrence) = 0.05; D_risk_avg (average risk undetectability) = 0.65; Calculate the ratio between each pair of criteria: a_SO=S_risk_avg / O_risk_avg=0.40 / 0.05=8.0; a_SD=S_risk_avg / D_risk_avg=0.40 / 0.65≈0.6154; a_OD=O_risk_avg / D_risk_avg=0.05 / 0.65≈0.0769; Round the ratio to the AHP standard scale: a_SO=8.0 → Rounded to 8 (which is exactly the standard ratio); a_SD = 0.6154 → 1 / 0.6154 ≈ 1.625 → rounded to 2, therefore a_SD = 1 / 2; a_OD=0.0769→1 / 0.0769≈13→Since 13>9, we take the maximum value of 9, therefore a_OD=1 / 9; The resulting judgment matrix A is:

[0069] Calculate the eigenvalues ​​of the judgment matrix A, normalize the eigenvalues, and use the normalized eigenvalues ​​as the criterion layer weight vector: W i =[0.358,0.055,0.587]; Calculate the consistency index (CI) and the consistency ratio (CR): CI=(λ_max-n) / (n-1)=(3.038-3) / 2=0.019; RI (Randomization Index, 0.58 when n=3); CR=CI / RI=0.019 / 0.58≈0.0328<0.1.

[0070] The calculated consistency ratio CR of the matrix is ​​less than 0.1, passing the consistency test and proving that the weight allocation is reasonable. The ranking of the W_i index is the fault ranking.

[0071] This result reflects the following in the system: Undetectability (D_risk) has the greatest impact on fault priority, with a weight of 0.587; The impact of severity (S_risk) is second, with a weight of 0.358; The probability of occurrence (O_risk) has the least impact, with a weight of only 0.055.

[0072] This weighting distribution indicates that, in the system under test, the difficulty of detecting the fault is the most important consideration, followed by the severity of the fault, while the probability of the fault occurring is relatively less important.

[0073] Step four: The real-time test state vector T is a 50-dimensional binary vector. Real-time state information for each test point is obtained from the field control network. After preprocessing such as threshold judgment and filtering, the real-time state information is used to generate the test state vector. When comparing test state vectors, bitwise AND or equality operations can be used: that is, for the i-th row D_i of the test-correlation matrix D, check whether (T&D_i)==D_i (where & is the bitwise AND operation). If satisfied, then fault i is the fault fuzzy group F.

[0074] Step five involves extracting the weight value corresponding to each fault in the fault fuzzy group F from the pre-calculated comprehensive weight vector W. Then, the faults in the fault fuzzy group F are sorted from largest to smallest according to their weight values. A higher weight indicates a greater overall probability of the fault occurring (considering probability, severity, and detectability), and thus a higher ranking. This step is crucial for the fusion of AHP decision and correlation models, resolving the problem of prioritizing faults when multiple faults are matched simultaneously.

[0075] Step six: Output the sorted fault list obtained in step five as the final diagnostic result. This result can be displayed to maintenance personnel in a clear list format through the DCS fault diagnosis information interface, or it can be passed to other maintenance management systems via an interface. The output information should at least include the fault name (or ID), sorting position, and its weight value, thereby providing direct and efficient support for maintenance decisions and guiding maintenance personnel to prioritize troubleshooting the most likely faults.

[0076] This completes one diagnostic cycle. The system will continue to cycle through steps four through six to achieve real-time online fault diagnosis.

[0077] Furthermore, as a response to the above Figure 1 The implementation of the method shown in this invention provides a fault diagnosis device for a nuclear power control system, such as... Figure 6 As shown, the device includes: The correlation matrix construction module 602 is used to perform fault mode, impact and hazard analysis on multiple products in the nuclear power control system, obtain analysis results, and construct a fault-test correlation matrix based on the analysis results, the system structure and function of the nuclear power control system and the preset diagnostic information. The behavioral fault modes in the fault-test correlation matrix are listed as test points. The comprehensive weight calculation module 604 is used to construct a hierarchical analysis model based on the analysis results and calculate the comprehensive weight corresponding to each failure mode in the scheme layer of the hierarchical analysis model. The hierarchical analysis model includes the target layer, the criterion layer and the scheme layer. The fuzzy fault group acquisition module 606 is used to acquire the current status information of each test point in the nuclear power control system, and match the current status information of each test point with the fault-test correlation matrix to obtain the fuzzy fault group. The fault diagnosis result generation module 608 is used to obtain the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault group based on the hierarchical analysis model, obtain the fault cause and design improvement scheme corresponding to each candidate fault mode according to the analysis results, sort the candidate fault modes in descending order of comprehensive weight, and generate the fault diagnosis result of the nuclear power control system based on the comprehensive weight, fault cause and design improvement scheme corresponding to each candidate fault mode.

[0078] This application provides a fault diagnosis device for nuclear power control systems. Compared with existing technologies, it performs fault mode, impact, and hazard analysis on multiple products in the nuclear power control system, obtains analysis results, constructs a fault-test correlation matrix, and builds a hierarchical analysis model based on the analysis results. It calculates the comprehensive weight corresponding to each fault mode in the scheme layer of the hierarchical analysis model, obtains the current state information of each test point in the nuclear power control system, matches the current state information of each test point with the fault-test correlation matrix to obtain fuzzy fault groups, obtains the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault group based on the comprehensive weight corresponding to each fault mode in the scheme layer of the hierarchical analysis model, obtains the fault causes and design improvement schemes corresponding to the candidate fault modes based on the analysis results, and generates fault diagnosis results based on the comprehensive weights, fault causes, and design improvement schemes corresponding to the candidate fault modes. By matching the current state information of the test points with the fault-test correlation matrix, a large number of irrelevant faults are quickly screened out to obtain fuzzy fault groups. Then, the hierarchical analysis model is used to incorporate multi-dimensional prior knowledge to accurately sort the candidate fault modes in the fuzzy fault groups, significantly improving the diagnostic accuracy under multiple concurrent fault conditions, shortening the fault diagnosis time, and improving maintenance efficiency.

[0079] In one embodiment, the comprehensive weight calculation module is further used for: Based on the analysis results, obtain the occurrence probability, severity, and detectability of each failure mode; The severity and detectability of each failure mode are converted into risk indicators to obtain the risk severity and risk undetectability of each failure mode. Based on the risk severity, risk undetectability and occurrence probability of each failure mode, the severity criterion layer weight, occurrence probability criterion layer weight and detectability criterion layer weight are calculated. Based on the detectability of each failure mode, calculate the first weight of each failure mode under the detectability criterion layer; based on the severity of each failure mode, calculate the second weight of each failure mode under the severity criterion layer; and based on the occurrence probability of each failure mode, calculate the third weight of each failure mode under the occurrence probability criterion layer. Based on the severity criterion layer weights, the occurrence probability criterion layer weights, the detectability criterion layer weights, and the first, second, and third weights corresponding to each failure mode, the comprehensive weight corresponding to each failure mode is calculated.

[0080] In one embodiment, the comprehensive weight calculation module is further used for: Based on the risk severity corresponding to each failure mode, the average risk severity is calculated; based on the risk undetectability corresponding to each failure mode, the average risk undetectability is calculated; based on the occurrence probability corresponding to each failure mode, the average risk occurrence probability is calculated. A judgment matrix is ​​constructed based on the average risk severity, the average risk undetectability, and the average risk occurrence probability. The eigenvalues ​​of the judgment matrix are calculated to obtain multiple eigenvalues ​​of the judgment matrix. The multiple eigenvalues ​​are then normalized to obtain normalized eigenvalues. The normalized feature values ​​are used as the weights for the severity criterion layer, the probability of occurrence criterion layer, and the detectability criterion layer, respectively.

[0081] In one embodiment, the comprehensive weight calculation module is further used for: Calculate the ratio of each pair of the average risk severity, average risk undetectability, and average risk occurrence probability; Each ratio is rounded according to the analytic hierarchy process standard to obtain the rounded number of each ratio. Construct a judgment matrix based on the rounded value of each ratio.

[0082] In one embodiment, the comprehensive weight calculation module is further used for: The maximum value among the normalized feature values ​​is used as the weight of the detectability criterion layer, the median value among the normalized feature values ​​is used as the weight of the severity criterion layer, and the minimum value among the normalized feature values ​​is used as the weight of the occurrence probability criterion layer.

[0083] In one embodiment, the comprehensive weight calculation module is further used for: For each failure mode, calculate the first product of the detectability criterion layer weight and the first weight corresponding to the failure mode, calculate the second product of the severity criterion layer weight and the second weight corresponding to the failure mode, calculate the third product of the occurrence probability criterion layer weight and the third weight corresponding to the failure mode, and use the sum of the first, second and third products as the comprehensive weight corresponding to the failure mode.

[0084] In one embodiment, the fuzzy fault group acquisition module is further configured to: The current state information of each test point is compared with a preset threshold, and the actual state value corresponding to the test point is obtained based on the comparison result. If the actual state value corresponding to the test point with an element of 1 in any fault mode in the fault-test correlation matrix is ​​also 1, then any fault mode is a candidate fault mode, and multiple candidate fault modes form a fault fuzzy group.

[0085] According to one embodiment of the present invention, a storage medium is provided, the storage medium storing at least one executable instruction, which can execute the fault diagnosis method of the nuclear power control system in any of the above method embodiments.

[0086] The present invention provides a computer device according to one embodiment, and the specific implementation of the computer device is not limited by the specific embodiments of the present invention.

[0087] The computer device may include a processor, a communications interface, memory, and a communications bus.

[0088] The processor, communication interface, and memory communicate with each other via a communication bus.

[0089] A communication interface is used to communicate with other network elements, such as clients or other servers.

[0090] The processor is used to execute programs, specifically the relevant steps in the above-described embodiments of the fault diagnosis method for nuclear power control systems.

[0091] Specifically, the program may include program code, which includes computer operation instructions.

[0092] The processor may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The computer device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0093] Memory is used to store programs. Memory may include high-speed RAM, and may also include non-volatile memory, such as at least one disk drive.

[0094] Specifically, the program can be used to cause the processor to execute the fault diagnosis method of the nuclear power control system in any of the above method embodiments.

[0095] It will be apparent to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. In one embodiment, they can be implemented using device-executable program code, thereby allowing them to be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular hardware and software combination.

[0096] The above embodiments are merely exemplary embodiments of this application and are not intended to limit this application. The scope of protection of this application is defined by the claims. Those skilled in the art can make various modifications or equivalent substitutions to this application within its substance and scope of protection, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of this application.

Claims

1. A fault diagnosis method for a nuclear power plant control system, characterized in that, include: Failure mode, effects, and hazards analysis were performed on multiple products in the nuclear power control system. The analysis results were obtained, and a fault-test correlation matrix was constructed based on the analysis results, the system structure and functions of the nuclear power control system, and the preset diagnostic information. Based on the analysis results, a hierarchical analysis model is constructed, and the comprehensive weight corresponding to each failure mode in the solution layer of the hierarchical analysis model is calculated. The hierarchical analysis model includes an objective layer, a criterion layer, and a solution layer. Obtain the current status information of each test point in the nuclear power control system, and match the current status information of each test point with the fault-test correlation matrix to obtain a fuzzy fault group; Based on the hierarchical analysis model, the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault group is obtained. According to the analysis results, the fault cause and design improvement scheme corresponding to each candidate fault mode are obtained. The candidate fault modes are sorted from high to low according to their comprehensive weights. Based on the comprehensive weight, fault cause and design improvement scheme corresponding to each candidate fault mode, the fault diagnosis result of the nuclear power control system is generated. Based on the risk severity, risk undetectability, and occurrence probability corresponding to each failure mode, the weights of the severity criterion layer, the occurrence probability criterion layer, and the detectability criterion layer are calculated, including: Based on the risk severity corresponding to each failure mode, the average risk severity is calculated; based on the risk undetectability corresponding to each failure mode, the average risk undetectability is calculated; based on the occurrence probability corresponding to each failure mode, the average risk occurrence probability is calculated. Among these, the average risk undetectability is greater than the average risk severity, and the average risk severity is greater than the average risk occurrence probability. A judgment matrix is ​​constructed based on the average risk severity, the average risk undetectability, and the average risk occurrence probability. The eigenvalues ​​of the judgment matrix are calculated to obtain multiple eigenvalues ​​of the judgment matrix. The multiple eigenvalues ​​are then normalized to obtain normalized eigenvalues. The normalized feature values ​​are used as the weights of the severity criterion layer, the probability of occurrence criterion layer, and the detectability criterion layer, respectively. The judgment matrix is ​​constructed based on the average risk severity, the average risk undetectability, and the average risk occurrence probability, including: Calculate the ratio of any two values ​​among the average risk severity, the average risk undetectability, and the average risk occurrence probability; Each ratio is rounded according to the analytic hierarchy process standard to obtain the rounded number of each ratio. Construct a judgment matrix based on the rounded value of each ratio; The step of using the normalized feature values ​​as weights for the severity criterion layer, the probability of occurrence criterion layer, and the detectability criterion layer includes: The maximum value among the normalized feature values ​​is used as the weight of the detectability criterion layer, the median value among the normalized feature values ​​is used as the weight of the severity criterion layer, and the minimum value among the normalized feature values ​​is used as the weight of the occurrence probability criterion layer.

2. The fault diagnosis method for a nuclear power control system as described in claim 1, characterized in that, The criterion layer of the hierarchical analysis model includes a severity criterion layer, an occurrence probability criterion layer, and a detectability criterion layer. Based on the analysis results, the comprehensive weight corresponding to each failure mode in the solution layer of the hierarchical analysis model is calculated, including: Based on the analysis results, the occurrence probability, severity, and detectability of each failure mode are obtained. The severity and detectability of each failure mode are converted into risk indicators to obtain the risk severity and risk undetectability of each failure mode. Based on the risk severity, risk undetectability and occurrence probability of each failure mode, the severity criterion layer weight, occurrence probability criterion layer weight and detectability criterion layer weight are calculated. Based on the detectability of each failure mode, calculate the first weight of each failure mode under the detectability criterion layer; based on the severity of each failure mode, calculate the second weight of each failure mode under the severity criterion layer; and based on the occurrence probability of each failure mode, calculate the third weight of each failure mode under the occurrence probability criterion layer. Based on the severity criterion layer weights, the occurrence probability criterion layer weights, the detectability criterion layer weights, and the first, second, and third weights corresponding to each failure mode, the comprehensive weight corresponding to each failure mode is calculated.

3. The fault diagnosis method for a nuclear power control system as described in claim 2, characterized in that, The comprehensive weight for each failure mode is calculated based on the severity criterion layer weight, the occurrence probability criterion layer weight, the detectability criterion layer weight, and the first, second, and third weights corresponding to each failure mode, including: For each fault mode, calculate the first product of the detectability criterion layer weight and the first weight corresponding to the fault mode, calculate the second product of the severity criterion layer weight and the second weight corresponding to the fault mode, calculate the third product of the occurrence probability criterion layer weight and the third weight corresponding to the fault mode, and use the sum of the first product, the second product and the third product as the comprehensive weight corresponding to the fault mode.

4. The fault diagnosis method for a nuclear power plant control system as described in any one of claims 1-3, characterized in that, The current state information of each test point is matched with the fault-test correlation matrix to obtain a fuzzy fault group, including: The current state information of each test point is compared with a preset threshold, and the actual state value corresponding to the test point is obtained based on the comparison result. If the actual state value corresponding to the test point with an element of 1 in any fault mode in the fault-test correlation matrix is ​​also 1, then any fault mode is a candidate fault mode, and multiple candidate fault modes form a fault fuzzy group.

5. A fault diagnosis device for a nuclear power control system based on the fault diagnosis method of the nuclear power control system according to claim 1, characterized in that, include: The correlation matrix construction module is used to perform fault mode, impact and hazard analysis on multiple products in the nuclear power control system, obtain analysis results, and construct a fault-test correlation matrix based on the analysis results, the system structure and function of the nuclear power control system and the preset diagnostic information. The comprehensive weight calculation module is used to construct a hierarchical analysis model based on the analysis results and calculate the comprehensive weight corresponding to each failure mode in the solution layer of the hierarchical analysis model. The hierarchical analysis model includes a target layer, a criterion layer, and a solution layer. The fuzzy fault group acquisition module is used to acquire the current status information of each test point in the nuclear power control system, and match the current status information of each test point with the fault-test correlation matrix to obtain the fuzzy fault group. The fault diagnosis result generation module is used to obtain the comprehensive weight corresponding to each candidate fault mode in the fuzzy fault group based on the hierarchical analysis model, obtain the fault cause and design improvement scheme corresponding to each candidate fault mode according to the analysis results, sort the candidate fault modes in descending order of comprehensive weight, and generate the fault diagnosis result of the nuclear power control system based on the comprehensive weight, fault cause and design improvement scheme corresponding to each candidate fault mode.

6. A storage medium storing at least one executable instruction that causes a processor to perform the operation of the fault diagnosis method for a nuclear power control system as described in any one of claims 1-4.

7. A computer device, comprising: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the fault diagnosis method of the nuclear power control system as described in any one of claims 1-4.

Images