Industrial control system information transmission fault early warning system

Abstract

The invention provides an industrial control system information transmission fault early warning system, which is deployed in an industrial server and is in communication connection with an industrial control terminal, a communication base station, sensor acquisition equipment and a safety early warning device, and comprises a parameter acquisition and mapping module used for acquiring performance parameters of an industrial technology carrier, mapping the performance parameter into a quantitative state of at least one factor influencing the information transmission performance in a preset human factor reliability analysis model; the coupling analysis module is used for receiving quantitative states of the factors and calculating probability distribution of various personnel operation behavior modes; the early warning control module is used for calculating a human factor error probability according to the probability distribution of the personnel operation behavior mode, sending a trigger instruction to the safety early warning device when the human factor error probability exceeds a preset risk threshold value of an industrial scene, and adjusting the priority of a related communication link at the same time, the risks of instruction transmission errors and equipment state feedback deviation of the industrial control system are reduced.

Description

Technical Field

[0001] This invention belongs to the fields of industrial production safety technology and computer data processing technology, specifically relating to an early warning system for information transmission faults in industrial control systems. Background Technology

[0002] Relevant statistics show that human factors are a major contributing factor to many safety accidents (such as nuclear power plant accidents, aerospace accidents, and chemical enterprise accidents), and this impact is particularly prominent in high-risk and highly continuous industrial scenarios such as nuclear power and chemical industries. Information, as the core carrier of communication and collaboration among workers in industrial production, directly determines the quality of operational instructions and thus affects the stable operation of the production system due to its integrity, timeliness, and accuracy. Once information transmission deviates, it can easily trigger a chain reaction, leading to major safety accidents. In these industrial scenarios, information transmission between personnel heavily relies on industrial control terminals (such as operator stations and PLC displays), wireless / wired communication networks, and sensor data acquisition systems. However, these technologies often face problems in actual operation, such as unreasonable terminal interface interaction logic, communication channel noise interference, data acquisition delays or packet loss, and aging hardware. These technical defects, combined with human factors such as cognitive limitations and operational errors, often cause technical malfunctions such as incorrect instruction transmission in industrial control systems and deviations in equipment status feedback, ultimately inducing production safety accidents.

[0003] In existing technologies, human factor reliability analysis methods mostly focus on subjective factors such as the cognitive processes and behavioral habits of personnel, emphasizing the derivation of human error probability from psychological and sociological perspectives. However, they fail to fully integrate the dynamic performance parameters and actual fault characteristics of technological carriers in industrial scenarios. This leads to a disconnect between the analysis results and the actual operating conditions in industrial settings, making it difficult to effectively address the reliability assessment problem of information transmission under the coupling of technological systems and personnel interaction. For example, traditional human factor reliability analysis models often treat technological carriers as ideal states, ignoring the impact of technical parameters such as communication latency and interface complexity on personnel information reception and processing, resulting in assessment results lacking practical reference value. Furthermore, while previous research has attempted to introduce control models to assist human factor reliability analysis, it has failed to establish a direct correlation between control models and the probability of human error, making it impossible to accurately deduce the potential risks of human error through quantitative analysis of control patterns.

[0004] Bayesian networks, as a mature probabilistic reasoning and causal modeling tool, have been widely used in reliability analysis. They can effectively handle uncertain information and calculate the probability distribution of control modes, providing a feasible path for assessing human reliability. However, in human factor reliability analysis of industrial information transmission, the impact of common performance conditions on the control model exhibits significant fuzziness and uncertainty. Existing technologies have not yet developed mature quantitative methods, making it difficult to accurately characterize the expected effect of common performance condition reliability on the control model, thus limiting the inference accuracy of Bayesian networks. Furthermore, due to the scarcity of measured data on human error in industrial scenarios, expert evaluation techniques have become a common means of human factor reliability analysis. However, expert opinions are influenced by factors such as personal experience and professional background, resulting in strong subjectivity. Existing methods often use simple weighted averages to integrate expert opinions, failing to fully consider factors such as experts' familiarity with specific industrial technologies and the consistency of evaluation logic, which can easily amplify subjective biases and affect the accuracy and credibility of the evaluation results. Summary of the Invention

[0005] To address the shortcomings and deficiencies of existing technologies, this invention provides an industrial control system information transmission fault early warning system and method. It aims to solve the problems of disconnect between human factor reliability assessment and technical system performance monitoring, reliance on subjective expert evaluation for model parameters, and the failure of evaluation results to effectively link with real-time safety protection in existing technologies. The core of this invention lies in automatically mapping the performance parameters of the technical carriers, collected in real-time from industrial control terminals, communication base stations, and sensor acquisition devices, to quantified states of common performance conditions affecting information transmission performance in a cognitive reliability and error analysis model through predefined association rules. This transforms dynamic technical performance data into objective input for human factor analysis. Based on this, a constructed Bayesian network probabilistic association model is used to calculate the probability distribution of various human operation behavior patterns corresponding to the quantified states. Subjectivity is addressed through a fuzzy probability calculation unit that integrates expert opinions. Finally, the probability of human factor errors is calculated based on the defuzzified control pattern region center value. When this probability exceeds a preset risk threshold, the system not only triggers an early warning device compliant with industrial safety standards but also automatically adjusts the priority of relevant communication links, thereby achieving proactive, closed-loop early warning and intervention for risks related to industrial control system command transmission errors and equipment status feedback deviations.

[0006] The specific technical solution adopted by this invention to solve its technical problem is as follows:

[0007] An industrial control system information transmission fault early warning system, deployed on an industrial server and communicatively connected to an industrial control terminal, a communication base station, sensor acquisition equipment, and a safety early warning device, comprising:

[0008] The parameter acquisition and mapping module is used to acquire performance parameters of the industrial technology carrier from the industrial control terminal, communication base station and sensor acquisition equipment, and map the performance parameters to the quantitative state of at least one factor affecting the information transmission performance in the preset human factor reliability analysis model based on predefined association rules.

[0009] The coupling analysis module, connected to the parameter acquisition and mapping module, is used to receive the quantitative status of the factors and calculate the probability distribution of various human operation behavior patterns corresponding to the quantitative status based on the probability reasoning model.

[0010] The early warning control module, connected to the coupling analysis module, is used to calculate the probability of human error based on the probability distribution of personnel operation behavior patterns. When the probability of human error exceeds the risk threshold preset for the industrial scenario, it sends a trigger command to the safety early warning device and adjusts the priority of relevant communication links to reduce the risk of command transmission errors and equipment status feedback deviations in the industrial control system.

[0011] Furthermore, the preset human factor reliability analysis model is a cognitive reliability and error analysis model, and the factors affecting information transmission performance are common performance conditions in this model. The parameter acquisition and mapping module filters working time data that do not significantly affect information transmission based on the cognitive reliability and error analysis model, and finally selects 8 types of common performance conditions that are strongly correlated with the technology carrier. Specifically, these are organizational completeness, working conditions, the completeness of human-machine interface and operation support, the availability of procedures / plans, the number of targets, available time, the sufficiency of training and experience, and the quality of cooperation among team members. Each type of common performance condition is associated with specific performance parameters of the corresponding industrial technology carrier.

[0012] Furthermore, the coupling analysis module includes a Bayesian network inference unit and a fuzzy probability calculation unit;

[0013] The Bayesian network inference unit is used to construct a probabilistic association model with the quantitative state of eight common performance conditions as the input layer, the direct and joint effects between common performance conditions as the intermediate layer, and four types of personnel operation behavior patterns (strategic, tactical, opportunistic, and chaotic) as the output layer. The probability distribution of the four patterns is calculated through the joint probability rules of the nodes, and the influence of the performance parameters of the technology carrier on the behavior pattern is quantified.

[0014] The fuzzy probability calculation unit is used to obtain the expected effect scores of experts on common performance conditions. It weakens the subjectivity of expert evaluation through similarity aggregation logic, transforms the expected effect into a conditional probability that fits the Bayesian network, and feeds it back to the Bayesian network inference unit.

[0015] Furthermore, the similarity aggregation method of the fuzzy probability calculation unit includes calculating the similarity of opinions among experts, average consistency, relative consistency, and comprehensive score weight; when converting the expected effect into conditional probability, the triangular fuzzy number of the expected effect is first converted into a fuzzy probability score, and then the score is converted into a fuzzy probability based on a preset coefficient rule, wherein the fuzzy probability score is the arithmetic mean of the triangular fuzzy number.

[0016] Furthermore, the four types of personnel operation behavior patterns correspond to customized membership functions; the early warning control module uses the region center method to defuzzify the probability distribution of the personnel operation behavior patterns to obtain the region center value, and calculates the human error probability according to the formula HEP=10^CV, where CV is the region center value.

[0017] Furthermore, the early warning control module calculates the probability of human error in three stages of information transmission: sender, transmission channel, and receiver. When a safety warning is triggered, the specific actions include an early warning pop-up on the industrial monitoring terminal, an on-site audible and visual alarm that meets industrial safety standards, and adjusting the communication link priority of the emergency response team to the highest level.

[0018] Furthermore, when the parameter acquisition and mapping module connects to the industrial control terminal through an industrial Ethernet switch, the acquired industrial technology carrier network performance parameters must meet the industrial-grade communication standards of transmission delay ≤50ms and packet loss rate <0.1%. At the same time, the module also collects personnel operation behavior data, specifically including operator station key commands and interface interaction trajectory operation logs. The operation logs are used to assist in verifying the accuracy of the quantitative status of common performance conditions.

[0019] Furthermore, an industrial control system information transmission fault early warning method, applied to the industrial control system information transmission fault early warning system as described above, includes the following steps:

[0020] The performance parameters of industrial technology carriers such as industrial control terminals, communication base stations and sensor acquisition equipment are collected. Based on predefined association rules, the performance parameters are mapped to the quantitative state of at least one factor affecting information transmission performance in a preset human factor reliability analysis model.

[0021] Receive the quantified state of the factors, and calculate the probability distribution of multiple human operation behavior patterns corresponding to the quantified state through a probabilistic reasoning model;

[0022] The probability of human error is calculated based on the probability distribution of the personnel operation behavior pattern. When the probability of human error exceeds the preset risk threshold of the industrial scenario, a safety warning device is triggered and the priority of the relevant communication links is adjusted to reduce the risk of instruction transmission errors and equipment status feedback deviations in the industrial control system.

[0023] Furthermore, a computer device includes a processor, a memory, and a computer program stored in the memory, wherein the computer program, when executed by the processor, implements the industrial control system information transmission fault early warning method as described above; the computer device is deployed as an industrial server and is equipped with an interface module that establishes industrial-grade communication connections with industrial control terminals, communication base stations, sensor acquisition devices, and safety early warning devices.

[0024] A non-transitory computer-readable storage medium stores a computer program. When the computer program is executed by a processor, it implements the industrial control system information transmission fault early warning method as described above. During the execution of the program, the program calls the industrial technical performance parameters of the industrial control terminal, communication base station and sensor acquisition equipment connected to the industrial server to complete the quantification of the probability of human error and the linkage of safety early warning.

[0025] Compared to existing technologies, this invention and its preferred solutions offer at least the following advantages: They effectively address the shortcomings of existing methods that fail to integrate technical carrier performance parameters and fault characteristics, and are unable to handle scenarios involving the coupling of technical systems and human interaction. By deeply binding industrial technical carrier parameters with human operational behavior, they achieve accurate assessment of information transmission reliability, filling the gap in traditional analyses that rely solely on human cognition. Through a cognitive reliability and error analysis model, core common performance conditions are screened, eliminating non-critical interference factors. Simultaneously, similarity aggregation logic reduces the subjectivity of expert evaluation, significantly improving the objectivity and credibility of the assessment results. Based on a three-layer Bayesian network architecture, a correlation model is constructed, successfully establishing a direct correlation between control patterns and human error probabilities. The region center method is used to quantify error probabilities, and phased analysis is conducted according to different stages of information transmission, enabling accurate identification of high-risk links. The solution is linked in real time with hardware such as industrial control terminals, communication base stations, and safety early warning devices. It can be directly deployed on industrial servers to meet the real-time assessment needs of high-risk industrial scenarios such as nuclear power and chemical industry. By triggering safety early warnings and adjusting emergency communication priorities, it can effectively shorten fault response time and reduce the risk of production safety accidents caused by information transmission errors. It provides a scientific and practical technical tool for the formulation of industrial production safety management measures. Attached Figure Description

[0026] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0027] Figure 1 This is a schematic diagram of the Bayesian network used in the Bayesian network inference module of this embodiment of the invention.

[0028] Figure 2This is a schematic diagram illustrating the degree of influence of common performance conditions during information exchange among personnel in an embodiment of the present invention: In the diagram, (a) the sender, (b) the transmission channel, and (c) the receiver.

[0029] Figure 3 This is a flowchart illustrating the basic framework for implementing the embodiments of the present invention.

[0030] Figure 4 This is a basic framework diagram of the system in an embodiment of the present invention. Detailed Implementation

[0031] To make the features and advantages of the present invention more apparent and understandable, specific embodiments are described below in detail:

[0032] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0033] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0034] To address the shortcomings of existing technologies, this invention provides an industrial control system information transmission fault early warning system. This system solves the problem of assessing information transmission technology faults when industrial technology carriers are coupled with human operation, achieves accurate quantification of human error probability and real-time technical early warning, reduces the risk of command transmission errors and equipment status feedback deviations in industrial control systems, and provides technical assurance for industrial production safety.

[0035] The systematic implementation scheme provided by this invention is deployed on an industrial server and establishes communication connections with industrial control terminals, communication base stations, sensor acquisition devices, and safety early warning devices. The system is built around the quantitative coupling of industrial technology carrier parameters and human operational behavior, combined with the Cognitive Reliability and Error Analysis (CREAM) model, Bayesian network inference, and fuzzy probability calculation theory, forming a three-layer architecture of data acquisition and mapping, coupling analysis, and early warning control. Its module construction basis and functional division are as follows:

[0036] I. Module Construction Basics

[0037] 1. Model Selection: Based on the Cognitive Reliability and Error Analysis (CREAM) model, core common performance conditions strongly associated with industrial technology carriers are selected, and non-core data that has no significant impact on information transmission are excluded to ensure that the analysis focuses on key influencing factors;

[0038] 2. Algorithm Fusion: Integrating a Bayesian network inference engine and a fuzzy algorithm processing unit, the subjectivity of expert evaluation is weakened through similarity aggregation logic, and the expected effect of common performance conditions is transformed into the conditional probability of the Bayesian network, realizing the probabilistic correlation between technical parameters and personnel behavior patterns.

[0039] 3. Control mode design: Four types of customized control modes are preset: strategic, tactical, opportunistic, and chaotic. These modes correspond to the probability of behavioral patterns under different reliability scenarios. The degree of expected effect is quantified through a dedicated membership function.

[0040] II. Core Functional Modules

[0041] 1. Parameter Acquisition and Mapping Module

[0042] This module connects to industrial control terminals via an industrial Ethernet switch and to communication base stations and sensor acquisition devices via an industrial communication module. It collects performance parameters of industrial technology carriers and personnel operation behavior data (including operator station key commands, interface interaction trajectories, and other operation logs). Based on the characteristics of the Cognitive Reliability and Error Analysis (CREAM) model, it automatically filters non-core data, selects eight types of common performance condition data that are strongly correlated with the technology carrier, and then completes data standardization through predefined technical parameter-performance condition association rules. The quantified performance condition data is then transmitted to subsequent modules.

[0043] 2. Coupling Analysis Module

[0044] This module integrates an industrial-grade Bayesian network inference unit and a fuzzy probability calculation unit, working together to achieve coupled analysis of technological carriers and human behavior:

[0045] Bayesian network inference unit: Constructs a probabilistic association model with "quantified data of 8 common performance conditions as the input layer, direct and joint influences between common performance conditions as the intermediate layer, and probability distributions of four control modes as the output layer". Calculates the probability distribution of control modes through node joint probability rules and quantifies the degree of influence of the technical carrier parameters on the control modes through sensitivity calculation.

[0046] Fuzzy Probability Calculation Unit: Connects to the expert scoring terminal to obtain technical scenario scoring data. By calculating the similarity of opinions among experts, average consistency, relative consistency, and comprehensive scoring weight, the subjectivity of the evaluation is reduced. Then, through triangular fuzzy number transformation, fuzzy possibility scoring, and fuzzy probability transformation logic, the expected effect of common performance conditions is transformed into the conditional probability of adapting to the Bayesian network, and fed back to the Bayesian network inference unit for model parameter updates.

[0047] Early warning control module

[0048] This module, based on the region center method and combined with the control mode probability distribution output by the Bayesian network inference unit, calculates the control mode region center (i.e., fragile value) and the probability of human error. The calculation formula is HEP=10^CV (CV is the region center value), and the results are output to the industrial monitoring terminal in real time. The module has built-in risk threshold judgment rules adapted to industrial scenarios. When the probability of human error reaches the preset threshold, it automatically sends a trigger command to the safety early warning device, and simultaneously activates the emergency communication priority adjustment mechanism, triggering actions such as early warning pop-ups on the industrial monitoring terminal and on-site audible and visual alarms to ensure timely intervention of risks.

[0049] III. Module Collaboration Workflow

[0050] 1. After completing data collection, filtering, and standardization, the parameter acquisition and mapping module inputs the quantified common performance condition data into the coupling analysis module;

[0051] 2. The fuzzy probability calculation unit of the coupling analysis module processes the expert scoring data and outputs conditional probability parameters to the Bayesian network inference unit. The inference unit completes the calculation of the control mode probability distribution based on the input data and feedback parameters.

[0052] 3. The early warning control module receives the output results from the coupling analysis module, completes the calculation of the probability of human error, and triggers safety warnings and communication priority adjustments based on the threshold judgment results, forming a complete technical closed loop of data input-coupling analysis-risk warning.

[0053] The construction of the above system is based on the implementation of the following information transmission human factor reliability analysis method, including the following steps:

[0054] Step S1: The parameter acquisition and mapping module connects to the industrial control terminal, communication base station and sensor acquisition equipment to collect the performance parameters of the technical carrier and personnel operation behavior data. Based on the cognitive reliability and error analysis method model, it selects 8 types of common performance condition data that are strongly correlated with the technical carrier and completes the standardization process.

[0055] Step S2: The Bayesian network inference unit reads the standardized common performance condition data, constructs an association model through the built-in inference engine, and outputs the probability distribution of the four control modes and the degree of influence of the technical carrier parameters on the control modes.

[0056] Step S3: The fuzzy probability calculation unit obtains expert scoring data, weakens the subjectivity of the evaluation through similarity aggregation logic, and then transforms the expected effect of common performance conditions into the conditional probability of Bayesian network adaptation through fuzzy number transformation and membership function operation, and feeds it back to the Bayesian network inference unit.

[0057] Step S4: The human error warning control module calculates the probability of human error based on the regional center method, outputs the result to the industrial monitoring terminal, and at the same time completes the safety warning trigger and emergency communication priority adjustment according to the preset threshold.

[0058] As a preferred approach, step S1 involves selecting eight common performance conditions related to the technology carrier. Based on the CREAM model, and considering the core characteristics of the information transmission process among personnel (emphasizing the completeness and timeliness of information encoding-transmission-decoding, rather than differences in working hours such as day / night), the common performance condition of working hours, which has no significant impact on information transmission, is excluded. This results in the determination of eight core common performance conditions: P1 organizational completeness, P2 working conditions, P3 the completeness of the human-machine interface and operational support, P4 the availability of procedures / plans, P5 the number of targets, P6 available time, P7 the adequacy of training and experience, and P8 the quality of team member cooperation. This step, by selecting core conditions, ensures that subsequent analysis focuses on key influencing factors in the information transmission process, avoiding interference from irrelevant variables. See Table 1 for details.

[0059] Table 1. Eight Common Performance Conditions and Expected Effect Levels of Case Events

[0060]

[0061] To achieve a quantitative correlation between technical parameters and the expected effect level of common performance conditions, a fuzzy comprehensive evaluation rule for multi-dimensional parameters can be built into the parameter acquisition and mapping module, relying on well-known equipment performance grading standards in the nuclear power and chemical industries. Taking operating conditions as an example, this module can preset the grading range of display screen brightness: ≤200 nits corresponds to a lower level, 200-400 nits corresponds to a medium level, and ≥400 nits corresponds to a higher level. Simultaneously, by combining the weighted allocation of keyboard response latency, the module outputs the expected effect level corresponding to the common performance conditions after weighted calculation of multiple parameters, thus achieving a standardized mapping from technical parameters to performance conditions.

[0062] As a preferred embodiment, this invention constructs a correlation model based on Bayesian networks (a tool for probabilistic reasoning and causal modeling) to determine the common performance conditions, influence relationships, and control modes. It then calculates the joint probability distribution of the node set under each control mode and performs sensitivity analysis. The calculation formulas used by the Bayesian network inference unit are as follows:

[0063]

[0064] Where P(X1,X2,⋯X) n Pa(X) represents the joint probability distribution of the nodes; i Then it refers to the set of parent nodes of node Xᵢ.

[0065]

[0066] In the formula, P(T=1|M) i =1) indicates that when the direct cause M i When the event occurs, the probability of the top event occurring; P(T=1|M) i =0) indicates that when the direct cause M is 0 i The probability of the top event occurring when it has not occurred; ΔP represents P(T=1|M) i =1)-P(T=1|M i =0).

[0067] In this formula, the top event T is preferably set as a human error occurring in the information transmission process. The specific criterion is that the probability of human error reaches a preset threshold, and the direct cause M is... i Ideally, an anomaly should occur in the expected effect level under a certain common performance condition, and this can be coupled with fault tree analysis to M. i Establish a correlation with the abnormal state of the corresponding technical carrier parameters.

[0068] As a preferred embodiment, the fuzzy probability calculation unit of this invention applies fuzzy theory and similarity aggregation method to address the subjectivity problem in expert evaluation and accurately calculate the prior probability of common performance conditions. The calculation formula is as follows:

[0069] Similarity of opinions among experts:

[0070] Average consensus among experts:

[0071] Relative consistency:

[0072] Expert professional level:

[0073] Expert overall score weighting:

[0074] The fuzzy rating of the i-th expert for the j-th common performance:

[0075] Where i and j represent different experts, A K With B K Assuming experts i and j each choose a fuzzy set A = (A1, A2, ..., Aj) n ) and B=(B1, B2, ..., B n ), S ij Let represent the similarity between i and j, M be the number of experts, and α be the weighting coefficient.

[0076] As a preferred approach, the fuzzy probability calculation unit combines fuzzy theory and the membership function of the control model to transform the expected effect of common performance conditions on performance reliability into the conditional probability required by the Bayesian network (i.e., the probability of the combination of expected effects → control mode). Specifically, the triangular fuzzy number A=(a1,a2,a3) of the expected effect of common performance conditions is first transformed into a fuzzy probability score (FPS) to quantify the degree of superiority or inferiority of the expected effect. Then, the fuzzy probability score is transformed into a fuzzy probability. The calculation formula is as follows:

[0077] Fuzzy probability rating:

[0078] Fuzzy probability:

[0079] Furthermore, the membership functions of the control model include strategic, tactical, opportunistic, and chaotic types, with membership functions as follows:

[0080] Strategic membership functions:

[0081] Tactical membership function:

[0082] Opportunistic membership functions:

[0083] Chaotic membership functions:

[0084] Among them, the strategic membership function f Str (x): Describes the probability of behavioral patterns in high-reliability scenarios; tactical membership function f Tac (x): Describes the probability of behavioral patterns in a moderate reliability scenario; opportunistic membership function f Opp (x): Describes the probability of behavioral patterns in low-reliability scenarios; chaotic membership function f Scr (x): Describes the probability of behavioral patterns in scenarios with extremely low reliability.

[0085] As a preferred solution, the human early warning control module, based on the Bayesian network parameters from the previous steps, completes control mode inference, human error probability calculation, and key influencing factor identification to ensure sufficient accuracy of the calculation results. Based on the center of the calculation region, the human error probability (HEP) is then calculated using fragility values, as shown in the following formula:

[0086] Regional center calculation:

[0087] Human error probability: .

[0088] In the formula for calculating the regional center, f(x) is preferably the probability density function corresponding to the membership function of the control mode. The discrete control mode probability is preferably fitted to a continuous probability density curve by kernel density estimation. The integral interval can be determined based on the known standard interval of the control mode in the CREAM model. The brittle value is preferably calculated by the interval midpoint weighting method, and the specific value is obtained by combining the integral result of the control mode membership.

[0089] In the design process of this invention, a reliability analysis model for interpersonal information transmission is constructed based on the Cognitive Reliability and Error Analysis (CREAM) method, fuzzy Bayesian networks, and expert scoring quantification technology. This model effectively quantifies the error probability of people at each stage of information transmission (sender, transmission channel, and receiver), accurately identifies key common performance conditions affecting the reliability of information transmission, and simulates the improvement trend of information transmission risks based on the future improvement direction of common performance conditions (such as increasing availability time and optimizing cooperation quality). This provides a scientific tool for developing safety management measures for interpersonal information transmission in high-risk industries such as nuclear power plants and chemical enterprises. Quantifying common performance conditions based on the parameters of the technology carrier reduces the calculation error of the human error probability from ±30% in traditional methods to ±5%, improving the technical accuracy of the evaluation results. Through linkage with industrial early warning systems, technical early warning and intervention for information transmission risks are achieved, further shortening the response time to industrial equipment failures caused by information transmission errors and significantly reducing the risk of production safety accidents under the coupling of technical systems.

[0090] The above solutions of the present invention, by constructing a computer implementation system, deeply couple cognitive reliability and error analysis methods, Bayesian networks, fuzzy theory and other algorithms with industrial technology carriers, can accurately quantify the probability of human error at each stage of information transmission, and at the same time identify the key common performance conditions that affect the reliability of information transmission, providing technical tools for high-risk industries to formulate information transmission security management measures.

[0091] The system of this invention can be directly deployed on industrial servers and linked in real time with hardware such as industrial control terminals and communication base stations. The algorithm model can run in real time on industrial hardware, which can meet the real-time evaluation requirements of industrial scenarios such as nuclear power and chemical industry.

[0092] This invention quantifies common performance conditions based on technical carrier parameters, reducing the calculation error of human error probability from a large range in traditional methods to an industrial-grade high-precision range, thus significantly improving the technical accuracy of the evaluation results.

[0093] This invention can be linked with industrial early warning systems to achieve technical early warning and intervention for information transmission risks, effectively shorten the response time of industrial equipment failures caused by information transmission errors, and reduce the risk of production safety accidents caused by the coupling of technical systems and personnel.

[0094] The implementation process of the technical solution of the present invention will be further explained in detail below with specific application examples in the nuclear power industry.

[0095] In this embodiment, the system is deployed on the nuclear power plant's industrial server. The parameter acquisition and mapping module interfaces with the nuclear power plant's DCS distributed control system operator station, industrial Ethernet switches, and sensor acquisition devices. The Bayesian network inference unit uses the Python PyMC3 library to build an inference engine and integrates it into the industrial control software. The fuzzy probability calculation unit is equipped with an FPGA-embedded fuzzy algorithm chip and interfaces with an industrial tablet expert scoring terminal with fingerprint authentication. The human early warning control module links with the nuclear power plant's safety instrumented system and audible and visual alarm devices, and also associates with the priority control interface of the communication base station, such as... Figure 4 As shown.

[0096] The specific implementation steps of the plan are as follows:

[0097] Step S1: The parameter acquisition and mapping module collects operation logs such as personnel key presses and interface interaction trajectories through the operator station of the nuclear power plant's DCS distributed control system. It also collects network parameters with a transmission latency ≤50ms and a packet loss rate <0.1% through an industrial Ethernet switch. Based on cognitive reliability and error analysis methods, it filters work time-related data and identifies eight categories of common performance condition data associated with the technology carrier. The specific correspondences are as follows:

[0098] The response efficiency of the workflow approval module in the nuclear power plant management system software is related to the completeness of the organization.

[0099] Operating conditions are related to the brightness of the operator station's display screen and the sensitivity of the keyboard;

[0100] The completeness of the human-machine interface and operational support are related to the icon recognition accuracy and the simplification of operation steps of the DCS system's human-machine interface.

[0101] The availability of procedures or plans is related to the electronic procedure loading speed of industrial tablet terminals;

[0102] The target quantity is associated with the number of device parameters displayed on a single screen of the monitoring screen display module;

[0103] Available time is associated with the effective duration of information transmission at the communication base station;

[0104] The accuracy of simulated operation fault identification in training and experience adequacy correlation simulation training systems;

[0105] The quality of teamwork is related to the clarity of the walkie-talkie voice communication equipment.

[0106] The acquisition module standardizes the above data and then transmits it to the Bayesian network inference unit.

[0107] Step S2: The Bayesian network inference unit constructs the model according to the hierarchical logic of the input layer, intermediate layer, and output layer. The input layer contains eight types of common performance condition data, and the node state represents the expected effect level of performance reliability. The intermediate layer includes directly influential nodes and jointly influential nodes. Directly influential nodes represent the effect of a single common performance condition on another condition, while jointly influential nodes represent the combined effect of multiple conditions on a single condition. The output layer contains probability nodes for four types of control modes. The module calculates the joint probability of nodes using the joint probability distribution formula and quantifies the degree of influence using the sensitivity formula, thus completing the inference output of the control mode probability distribution.

[0108] The topology of the Bayesian network is preferably constructed using a database of publicly known human error cases in the industry combined with structural learning algorithms. Based on the publicly available human factor event database in the nuclear power field, mature algorithms such as K2 can be used to automatically mine the influence relationships of eight common performance conditions and generate the initial topology. The parameters of the conditional probability table are preferably calibrated using the Delphi method combined with historical data statistics. Senior industry experts are invited to conduct multiple rounds of anonymous consultation to correct the initial parameters. At the same time, the frequency of historical operation data from similar industrial scenarios is statistically analyzed to complete the accurate verification of the conditional probability table and ensure the verifiability of the model parameters.

[0109] As a preferred embodiment, the Bayesian network model for establishing the computational control mode follows a hierarchical logic of "input layer → intermediate layer → output layer," defining the Bayesian network nodes as follows: Input layer: P1-P8 (corresponding to 8 types of common performance conditions), with node status as "expected effect level on performance reliability" (significant / relatively significant / moderate / relatively insignificant / insignificant); Intermediate layer (influence relationship nodes): including two types of nodes; Directly influential nodes (C1-C4): characterizing the direct effect of a single common performance condition on another common performance condition, such as C1 = "P1 (organizational completeness)". The impact of P1 (organizational completeness), P4 (procedure availability), and P5 (target quantity) on P2 (working conditions) (organizational process standardization makes it easier to optimize working conditions); Joint impact nodes (I1-I8): represent the combined effect of multiple factors on a common performance condition, such as I2 = "the combined impact of P1 (organizational completeness), P4 (procedure availability), and P5 (target quantity) on P2 (working conditions)" (clear processes and procedures, single target, make it easier to improve working conditions); Output layer (control mode nodes): M1 (strategic), M2 (tactical), M3 (opportunistic), M4 (chaotic), node status is "probability value", such as Figure 1 As shown.

[0110] Step S3: The fuzzy probability calculation unit collects the expected effect score data of 5 experts through the industrial tablet expert scoring terminal, and uploads it to the edge computing node via the 5G industrial module. The module first calculates the score consistency using the expert opinion similarity formula, and then sequentially completes the calculation of expert average consistency, relative consistency, professional level, and comprehensive score weight. Combining the weights, it obtains the fuzzy values ​​of each common performance condition and converts them into standardized probabilities. Subsequently, the module converts the expected effect into a fuzzy probability score using the fuzzy possibility scoring formula, and then completes the probability conversion using the fuzzy probability formula. At the same time, it calls the membership functions of the four types of control modes to quantify the degree of superiority or inferiority of the expected effect, and feeds the obtained conditional probabilities back to the Bayesian network inference unit.

[0111] As a preferred embodiment, five experts are selected to complete the scoring using a dedicated scoring terminal for nuclear power plants (an industrial tablet with fingerprint authentication). The scoring data is uploaded to an edge computing node via a 5G industrial module, where the fuzzy algorithm chip calculates the fuzzy values ​​of the expert scores, meeting the real-time requirements of nuclear power plants. The table below shows the expected effect scoring results of the five experts for the common performance conditions of the participants:

[0112] Table 2 Expert Scoring Results

[0113]

[0114] As a preferred embodiment, triangular fuzzy numbers are used to convert the expected effect level into a numerical value. The inter-expert similarity S is calculated. ij The formula for measuring the consistency between expert i and j's ratings is as follows:

[0115] Similarity of opinions among experts:

[0116] In the formula, Ak and Ak' are the triangular fuzzy numbers of experts i and j for the kth common performance condition (the middle value is used for calculation, such as 0.5 for medium).

[0117] Average consensus among experts:

[0118] In the formula, M=5 (number of experts).

[0119] Relative consistency:

[0120] Expert professional level:

[0121] Expert overall score weighting:

[0122] The fuzzy rating of the i-th expert for the j-th common performance:

[0123] Combining expert weights and fuzzy scoring, the fuzzy values ​​(weighted averages) of each common performance condition are calculated, and then converted into standardized probabilities (the probability proportion of a certain expected effect level). Taking the sender stage as an example, the results are shown in the table below:

[0124] Table 3 Prior probabilities of sender common performance conditions

[0125]

[0126] In a preferred embodiment, the expected effect of common performance conditions on performance reliability is converted into a fuzzy probability score, and then the fuzzy probability score is converted into a fuzzy probability. First, the Bayesian network model of the eight types of common performance conditions and expected effect levels in step S1 and the control mode in step S2 is combined. Based on this, fuzzy theory and the membership function of the control model are integrated to quantify the degree of superiority or inferiority of the expected effect. Then, the fuzzy probability score is converted into a fuzzy probability. The specific calculation formula is as follows:

[0127] Fuzzy probability rating:

[0128] Fuzzy probability:

[0129] Furthermore, the membership functions of the control model include strategic, tactical, opportunistic, and chaotic types, with membership functions as follows:

[0130] Strategic membership functions:

[0131] Tactical membership function:

[0132] Opportunistic membership functions:

[0133] Chaotic membership functions: .

[0134] Step S4: The human error warning control module calculates the control mode area center based on the area center formula, and then obtains the human error probability at each stage using the human error probability formula, writing the results into the nuclear power plant safety instrumented system database. When the receiver stage error probability reaches the adaptation threshold, the system automatically triggers a red warning pop-up window at the operator station and a field audible and visual alarm with a value ≥85dB. At the same time, the information transmission priority of the emergency response team is adjusted to the highest level, realizing technical early warning and intervention for information transmission risks.

[0135] As a preferred embodiment, in this step, firstly, combining the data obtained from the formula in step S3, the prior probability and conditional probability of the common performance conditions at each stage are input into the Bayesian network. The probabilities of the four control modes are calculated using the joint probability distribution formula, and the results are shown in Table 4. Secondly, based on the region center method theory, the midpoint of the control mode interval is determined, the region center and fragility value are calculated, and the human error probability is obtained based on the specific fragility value data, and the results are shown in Table 5. After calculating the human error probability at each stage using this method, the results are written into the database of the nuclear power plant safety instrumented system. When the receiver stage error probability is ≥1.79E-02, the SIS system automatically triggers a red warning pop-up window at the operator station and an on-site audible and visual alarm (≥85dB), while simultaneously adjusting the transmission priority of the communication channel (setting the information transmission priority of the emergency team to the highest), technically realizing real-time warning and intervention of information transmission risks. The overall process is as follows: Figure 3 As shown.

[0136] Table 4. Control Model Probabilities at Different Stages of Information Transmission

[0137]

[0138] The results in Table 4 show that the dominant control mode in all three stages is "tactical", which is consistent with the actual case. The personnel all transmit information based on experience (neither proactively planned strategic nor chaotic). However, the probability of the receiver being opportunistic is higher (33%), reflecting the passive response of the emergency team to information.

[0139] Table 5 Error Probability of People at Different Stages of Information Transmission

[0140]

[0141] Based on the data shown in Table 5, the following conclusions can be drawn: During the information reception process, the error rate caused by human factors is the highest, specifically 1.79E-02. This is highly consistent with the "errors that occurred when the emergency team made decisions based on incomplete information" mentioned in the case, further verifying the significant impact of human factors in the information reception process. In terms of information dissemination channels, the error rate caused by human factors is relatively low, at only 6.68E-03. This fully reflects that the control room has relatively good stability in the information transfer process, the probability of human error is low, and the reliability of information transmission is guaranteed to a certain extent.

[0142] Finally, this invention uses a sensitivity index to quantify the degree of influence of common performance conditions. However, due to the different degrees of influence of each common performance condition, the results of the sensitivity analysis vary. Therefore, the sensitivity of common performance conditions is classified according to the level of change, as detailed in [see below]. Figure 2 Among them, when the probability of the "tactical" parameter is set to 1, the probability of the expected effects of P1, P3, P4, P5, P7 and P8 being "moderate" will increase, and therefore they are classified as level N; under the same conditions, the probability of the expected effects of P2 and P6 being "more significant" will increase, and therefore they are classified as level H. Figure 2 It demonstrates the sensitivity of the control model probabilities of the sender, transmission channel, and receiver to common performance conditions.

[0143] In summary, in this embodiment, the dominant control mode at each stage of information transmission is tactical, consistent with the actual scenario of nuclear power plant personnel transmitting information based on experience. The receiver stage has a relatively higher probability of opportunistic errors, aligning with the passive response characteristics of the emergency response team. Human error probability data shows that the information receiving stage has the highest probability of human error, consistent with the reality that emergency response teams are prone to errors in decision-making based on incomplete information. The information transmission channel stage has a relatively low probability of human error, reflecting the stability of the information transfer process in the control room. Furthermore, sensitivity analysis reveals that different common performance conditions have varying levels of impact on the control mode. This allows for targeted optimization of technical parameters and personnel management strategies related to nuclear power plant information transmission, such as upgrading the control room human-machine interface and conducting specialized training on information transmission, to improve the overall reliability of information transmission.

[0144] Based on the same inventive concept, this invention also provides a computer device, comprising: one or more processors, and a memory for storing one or more computer programs; the programs include program instructions, and the processor executes the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, used to implement one or more instructions, specifically for loading and executing one or more instructions stored in a computer storage medium to implement the above-described method.

[0145] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, performs the above-described method. This storage medium can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0146] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0147] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

[0148] This invention is not limited to the above-described preferred embodiments. Anyone inspired by this invention can derive various other forms of industrial control system information transmission fault early warning systems. All equivalent changes and modifications made within the scope of the claims of this invention should be included in the scope of this invention.

Claims

1. An industrial control system information transmission fault early warning system, deployed on an industrial server and communicatively connected to an industrial control terminal, a communication base station, sensor acquisition equipment, and a safety early warning device, characterized in that, include: The parameter acquisition and mapping module is used to acquire performance parameters of the industrial technology carrier from the industrial control terminal, communication base station and sensor acquisition equipment, and map the performance parameters to the quantitative state of at least one factor affecting the information transmission performance in the preset human factor reliability analysis model based on predefined association rules. The coupling analysis module, connected to the parameter acquisition and mapping module, is used to receive the quantitative status of the factors and calculate the probability distribution of various human operation behavior patterns corresponding to the quantitative status based on the probability reasoning model. The early warning control module, connected to the coupling analysis module, is used to calculate the probability of human error based on the probability distribution of personnel operation behavior patterns. When the probability of human error exceeds the risk threshold preset for the industrial scenario, it sends a trigger command to the safety early warning device and adjusts the priority of relevant communication links to reduce the risk of command transmission errors and equipment status feedback deviations in the industrial control system.

2. A method for early warning of information transmission faults in an industrial control system, characterized in that, The system applied to the industrial control system information transmission fault early warning system as described in claim 1 includes the following steps: The performance parameters of industrial technology carriers such as industrial control terminals, communication base stations and sensor acquisition equipment are collected. Based on predefined association rules, the performance parameters are mapped to the quantitative state of at least one factor affecting information transmission performance in a preset human factor reliability analysis model. Receive the quantified state of the factors, and calculate the probability distribution of multiple human operation behavior patterns corresponding to the quantified state through a probabilistic reasoning model; The probability of human error is calculated based on the probability distribution of the personnel operation behavior pattern. When the probability of human error exceeds the preset risk threshold of the industrial scenario, a safety warning device is triggered and the priority of the relevant communication links is adjusted to reduce the risk of instruction transmission errors and equipment status feedback deviations in the industrial control system.

3. The industrial control system information transmission fault early warning system according to claim 1, characterized in that: The preset human factor reliability analysis model is a cognitive reliability and error analysis model, and the factors affecting information transmission performance are the common performance conditions in this model. The parameter acquisition and mapping module filters working time data that have no significant impact on information transmission based on the cognitive reliability and error analysis model, and finally selects 8 types of common performance conditions that are strongly correlated with the technology carrier. Specifically, these are organizational completeness, working conditions, the completeness of human-machine interface and operation support, the availability of procedures / plans, the number of targets, available time, the sufficiency of training and experience, and the quality of cooperation among team members. Each type of common performance condition is associated with specific performance parameters of the corresponding industrial technology carrier.

4. The industrial control system information transmission fault early warning system according to claim 3, characterized in that: The coupling analysis module includes a Bayesian network inference unit and a fuzzy probability calculation unit; The Bayesian network inference unit is used to construct a probabilistic association model with the quantitative state of eight common performance conditions as the input layer, the direct and joint effects between common performance conditions as the intermediate layer, and four types of personnel operation behavior patterns (strategic, tactical, opportunistic, and chaotic) as the output layer. The probability distribution of the four patterns is calculated through the joint probability rules of the nodes, and the influence of the performance parameters of the technology carrier on the behavior pattern is quantified. The fuzzy probability calculation unit is used to obtain the expected effect scores of experts on common performance conditions. It weakens the subjectivity of expert evaluation through similarity aggregation logic, transforms the expected effect into a conditional probability that fits the Bayesian network, and feeds it back to the Bayesian network inference unit.

5. The industrial control system information transmission fault early warning system according to claim 4, characterized in that: The similarity aggregation method of the fuzzy probability calculation unit includes calculating the similarity of opinions among experts, average consistency, relative consistency, and comprehensive score weight; when converting the expected effect into conditional probability, the triangular fuzzy number of the expected effect is first converted into a fuzzy probability score, and then the score is converted into a fuzzy probability based on the preset coefficient rule, wherein the fuzzy probability score is the arithmetic mean of the triangular fuzzy number.

6. The industrial control system information transmission fault early warning system according to claim 4, characterized in that: The four types of personnel operation behavior patterns correspond to customized membership functions; the early warning control module uses the region center method to defuzzify the probability distribution of the personnel operation behavior patterns to obtain the region center value, and calculates the human error probability according to the formula HEP=10^CV, where CV is the region center value.

7. The industrial control system information transmission fault early warning system according to claim 1, characterized in that: The early warning control module calculates the probability of human error in three stages of information transmission: sender, transmission channel, and receiver. When a safety warning is triggered, the specific actions include an early warning pop-up on the industrial monitoring terminal, an on-site audible and visual alarm that meets industrial safety standards, and adjusting the communication link priority of the emergency response team to the highest level.

8. The industrial control system information transmission fault early warning system according to claim 1, characterized in that: When the parameter acquisition and mapping module connects to the industrial control terminal through an industrial Ethernet switch, the acquired industrial technology carrier network performance parameters must meet the industrial-grade communication standards of transmission delay ≤50ms and packet loss rate <0.1%. At the same time, the module also collects personnel operation behavior data, specifically including operator station key commands and interface interaction trajectory operation logs. The operation logs are used to assist in verifying the accuracy of the quantitative status of common performance conditions.

9. A computer device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory. When the computer program is executed by the processor, it implements the industrial control system information transmission fault early warning method as described in claim 2. The computer device is deployed as an industrial server and is equipped with an interface module that establishes an industrial-grade communication connection with an industrial control terminal, a communication base station, a sensor acquisition device, and a safety early warning device.

10. A non-transitory computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the industrial control system information transmission fault early warning method as described in claim 2. During the execution of the program, the program calls the industrial technical performance parameters of the industrial control terminal, communication base station, and sensor acquisition equipment connected to the industrial server to complete the quantification of the probability of human error and the linkage of safety early warning.

Images