WO2014173257A1 - 操作员状态评估的可靠性分析方法及装置 - Google Patents

操作员状态评估的可靠性分析方法及装置 Download PDF

Info

Publication number
WO2014173257A1
WO2014173257A1 PCT/CN2014/075729 CN2014075729W WO2014173257A1 WO 2014173257 A1 WO2014173257 A1 WO 2014173257A1 CN 2014075729 W CN2014075729 W CN 2014075729W WO 2014173257 A1 WO2014173257 A1 WO 2014173257A1
Authority
WO
WIPO (PCT)
Prior art keywords
psf
reliability
psfs
state
operator
Prior art date
Application number
PCT/CN2014/075729
Other languages
English (en)
French (fr)
Inventor
张力
李鹏程
戴立操
胡鸿
邹衍华
蒋建军
黄卫刚
戴忠华
王春辉
苏德颂
李晓蔚
Original Assignee
湖南工学院
南华大学
中广核核电运营有限公司
大亚湾核电运营管理有限责任公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 湖南工学院, 南华大学, 中广核核电运营有限公司, 大亚湾核电运营管理有限责任公司 filed Critical 湖南工学院
Publication of WO2014173257A1 publication Critical patent/WO2014173257A1/zh

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063Operations research, analysis or management
    • G06Q10/0639Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063Operations research, analysis or management
    • G06Q10/0639Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06398Performance of employee with respect to a job function

Definitions

  • the present invention relates to the field of nuclear power plant detection and human factor reliability analysis, and more particularly to a reliability analysis method and apparatus for operator status assessment.
  • BACKGROUND OF THE INVENTION When an abnormal state occurs in a nuclear power plant, the operator will construct a reasonable and logical interpretation based on the state parameters of the nuclear power plant to evaluate the state of the power plant as a basis for subsequent response planning and response execution decisions. This system is called the Situation Assessment or Situation Awareness (SA).
  • SA Situation Awareness
  • the operator's correct state assessment of anomalous events is critical to the correct response of the operator's behavior.
  • the model consists of three components: the object (the actual current environment), the schema (the schema of the current environment), and the exploration (the search behavior in the environment).
  • Knowledge in the form of a schema or a mental model allows the operator to have a psychological expectation of information in the environment.
  • An active schema guides the operator's search and interpretation of specific information.
  • information obtained from the environment is schemated. Absorbing and using to revise and update the schema, again guiding the search for information to achieve contextual awareness, is a continuous process, as shown in Figure 1.
  • Embodiments of the present invention are directed to a reliability analysis method and apparatus for operator status evaluation, so as to solve the related art without considering the influence of the situational environment factors of the operator itself and their influence causality, thereby possibly To repeatedly calculate the likelihood of its effects, the probability of a state assessment error may cause an erroneous estimate.
  • a reliability analysis method for operator state evaluation including: determining a plurality of performance forming factors (Performance Shaping Factors, PSFs or PSFs) to be used, where The PSF is used for state evaluation; determining an association relationship between each PSF of the plurality of PSFs and other PSFs other than itself, and an association relationship between each PSF and a state evaluation reliability node; according to the plurality of PSFs and the plurality of The association relationship between the PSFs in the PSFs and the association relationship between the respective PSFs and the state evaluation reliability nodes establish a reliability analysis model based on the PSF causality operator state assessment to analyze the reliability of the operator Sex.
  • PSFs or PSFs Performance Shaping Factors
  • the reliability analysis model of the operator state assessment is a Bayesian network model.
  • establishing an operator status assessment based on the PSF causal relationship according to the relationship between the plurality of PSFs and the PSFs of the plurality of PSFs, and the association relationship between the respective PSFs and the state evaluation reliability nodes includes: determining a prior probability distribution of each root node PSF in the Bayesian network model in a different state, wherein the root node PSF is not Determining the PSF nodes pointed by the other nodes; determining the association relationship between the PSFs of the respective root nodes and other PSFs other than itself, and the association relationship between the PSFs and the state evaluation reliability nodes a conditional probability distribution in which the reliability node is in a different state from the state, wherein the child node PSF is a PSF node pointed to by other nodes; and the conditional probability distribution and the prior probability distribution are performed on a Bayesian network Caus
  • the value of the probability distribution of the Bayesian network model is calculated by a fuzzy method.
  • calculating, by the fuzzy method, the value of the probability distribution of the Bayesian network model comprises: calculating a prior probability distribution of each root node PSF in different states by using a triangular fuzzy number; Calculating a conditional probability distribution when the respective child node PSF and the state evaluation reliability node are in different states; determining a value of the prior probability and a value of the conditional probability by using a method of defocusing a triangle center of gravity, and A causal analysis of the Bayesian network is performed on the value of the prior probability and the value of the conditional probability to obtain the reliability of the state evaluation.
  • the method further includes: according to the determination of the operator state evaluation error, according to the relationship between the PSF of each root node and other PSFs other than itself, And determining, by the diagnostic analysis of the Bayesian network, a posteriori probability distribution of each root node PSF; and setting the root node PSF to a preset plurality of states The posterior probability distribution of the worst state in the middle is compared with the prior probability distribution of the worst state of the root node PSF in the preset plurality of states to obtain a key element affecting the reliability of the state assessment to determine the prevention state assessment Countermeasures for mistakes.
  • a reliability analysis apparatus for an operator status evaluation including: a first determining module, configured to determine a plurality of behavior forming factors PSF to be used, where the PSF is used a second determining module, configured to determine an association relationship between each PSF of the plurality of PSFs and other PSFs other than itself, and an association relationship between each PSF and a state evaluation reliability node; Establishing reliability of operator state evaluation based on PSF causality based on the plurality of PSFs, the association relationship between each of the plurality of PSFs, and the association relationship between the respective PSFs and the state evaluation reliability node The model is analyzed to analyze the reliability of the operator.
  • the reliability analysis model of the operator state assessment established by the establishing module is a Bayesian network model.
  • the establishing module includes: a first determining unit, configured to determine a prior probability distribution of each root node PSF in the Bayesian network model in a different state, where the root node PSF is not used by other nodes a PSF node that is directed to, a second determining unit, configured to determine, according to an association relationship between each of the root node PSFs and other PSFs other than itself, and an association relationship between the respective PSFs and the state evaluation reliability node a conditional probability distribution in which each of the child nodes PSF and the state evaluation reliability node are in different states, wherein the child node PSF is a PSF node pointed to by other nodes; and an analysis unit is configured to use the conditional probability distribution
  • the prior probability distribution is used to perform causal analysis of the Bayesian network to obtain the reliability of state evaluation under specific situation conditions.
  • the device further includes: a third determining module, configured to: according to the relationship between the respective root node PSF and other PSFs other than itself, and the The correlation between each PSF and the state evaluation reliability node is determined by a diagnostic analysis of the Bayesian network, and the posterior probability distribution of the PSF of each root node is determined; the comparison module is configured to place the root node PSF in a preset plurality of states The posterior probability of the worst state in the middle is compared with the prior probability of the worst state of the root node PSF in a preset plurality of states to obtain key factors affecting the reliability of the state assessment, to determine countermeasures for preventing the state assessment error.
  • a third determining module configured to: according to the relationship between the respective root node PSF and other PSFs other than itself, and the The correlation between each PSF and the state evaluation reliability node is determined by a diagnostic analysis of the Bayesian network, and the posterior probability distribution of the PSF of each root node is determined
  • the comparison module is configured to place the
  • the reliability relationship of the reliability nodes is evaluated, and a reliability analysis model of the operator state evaluation based on the PSF causality is established according to the above factors to analyze the reliability of the operator.
  • the related art does not consider the influence of the situational environmental factors on the operator itself and their causal relationship, which may bring the possibility of repeatedly calculating the influence, and may cause errors in the state estimation error probability.
  • FIG. 1 is a schematic diagram showing a state evaluation model of the related art
  • FIG. 2 is a flowchart showing a reliability analysis method of operator state evaluation according to an embodiment of the present invention
  • FIG. 3 shows an implementation of the present invention.
  • 4 is a schematic structural diagram of a reliability analysis device establishing module of an operator state evaluation according to an embodiment of the present invention
  • FIG. 1 is a schematic diagram showing a state evaluation model of the related art
  • FIG. 2 is a flowchart showing a reliability analysis method of operator state evaluation according to an embodiment of the present invention
  • FIG. 3 shows an implementation of the present invention.
  • 4 is a schematic structural diagram of a reliability analysis device establishing module of an operator state evaluation according to an embodiment of the present invention
  • FIG. 5 is a second structural diagram of a reliability analysis device for operator state evaluation according to an embodiment of the present invention
  • Figure 6 shows a schematic diagram of a Bayesian network model for operator state evaluation in a preferred embodiment of the present invention.
  • DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Based on the related technology, the influence of the situational environmental factors affected by the operators themselves and their influence causality are not considered, which may bring about the possibility of repeatedly calculating the influence thereof, and may cause a wrong estimation of the state estimation error probability.
  • the example provides a reliability analysis method for operator status evaluation, and the process of the method is as shown in FIG.
  • Step S202 determining a plurality of PSFs to be used, wherein the PSF is used for state evaluation.
  • Step S204 determining an association relationship between each PSF in the plurality of PSFs and other PSFs other than itself, and an association relationship between each PSF and the state evaluation reliability node;
  • Step S206 according to each of the plurality of PSFs and the plurality of PSFs Relationship between PSFs, and each
  • the relationship between the PSF and the state assessment reliability node establishes a reliability analysis model based on the PSF causality operator state assessment to analyze the reliability of the operator.
  • a reliability analysis model based on the PSF causality operator state assessment to analyze the reliability of the operator.
  • the related art does not consider the influence of the situational environmental factors on the operator itself and their causal relationship, which may bring the possibility of repeatedly calculating the influence, and may cause errors in the state estimation error probability.
  • the estimated problem provides qualitative and quantitative methods for the operator's state assessment reliability analysis, and provides countermeasures for the power plant to reduce the operator's state assessment failure probability.
  • the reliability analysis model of the operator state assessment may be a Bayesian network model.
  • the posterior probability distribution of the root node in the Bayesian network model of state evaluation in different states, the posterior probability of the root node PSF in the bad state (that is, the worst state in the preset multiple states) and the root node PSF are in poor condition
  • the prior probabilities of the states are compared to obtain the key elements (or key PSFs) that affect the reliability of the state assessment.
  • the reliability calculation of state evaluation is carried out, and according to the diagnostic analysis of Bayesian theory, the posterior probability distribution of the root node in the Bayesian network model is identified and the root is identified.
  • the prior probability value of the node PSF in a bad state is compared with the posterior probability value of the root node PSF in a bad state to obtain a key element that affects the reliability of the state evaluation.
  • the value of the probability distribution of the Bayesian network model can be further calculated by the fuzzy method.
  • each PSF can be evaluated by the triangular fuzzy number, and the above process of analyzing the reliability of the operator may be first calculated by the triangular fuzzy number.
  • the prior probability distribution of the root node PSF in different states then calculate the conditional probability distribution of each child node PSF and the state evaluation reliability node in different states by the triangular fuzzy number; then determine the prior probability by the method of triangle center of gravity defuzzification The value of the value and the conditional probability.
  • the conditional probability of the child node PSF and the state evaluation reliability node and the prior probability of the root node PSF are subjected to a causal analysis of the Bayesian network to obtain the reliability of the state evaluation.
  • the reliability analysis of the operator state assessment takes into account the causal relationship of the PSF, and the reliability estimation can be performed according to different states of different PSFs, which improves the accuracy of the reliability analysis.
  • the embodiment of the present invention further provides a reliability analysis device for assessing an operator state.
  • the structure of the device may be as shown in FIG. 3, including: a first determining module 10, configured to determine a plurality of PSFs to be used, where The PSF is used for state evaluation.
  • the second determining module 20 is coupled to the first determining module 10, and is configured to determine an association relationship between each PSF in the plurality of PSFs and other PSFs other than the PSF, and reliability of each PSF and state evaluation.
  • the association relationship between the nodes is coupled to the first determining module 10 and the second determining module 20, and is used to determine the reliability relationship between each PSF and each PSF according to multiple PSFs and multiple PSFs.
  • the association of nodes establishes a reliability analysis model of operator state assessment based on PSF causality to analyze the reliability of the operator.
  • the reliability analysis model of the operator state assessment established by the establishment module is a Bayesian network. Model.
  • the device can also identify key elements affecting the reliability of the state assessment through diagnostic analysis of the Bayesian network.
  • 4 is a schematic structural diagram of the establishing module 30, wherein the first determining unit 302 is configured to determine a prior probability distribution of PSFs of different root nodes in different states; and the second determining unit 304 is coupled to the first determining unit 302.
  • the analyzing unit 306 And coupling with the first determining unit 302 and the second determining unit 304, for performing the conditional probability of the child node PSF and the state evaluation reliability node and the prior probability of the root node PSF in a given context condition, Bayesian Causal analysis of the network to obtain the reliability of the state assessment.
  • the first determining unit 302 is further configured to calculate, when the PSF of each root node is in different states, by using the triangular fuzzy number and the triangular gravity center to solve the fuzzy method.
  • the second determining unit 304 is further configured to calculate a conditional probability distribution of each sub-node PSF and a state evaluation reliability node by a triangular fuzzy number and a triangular center of gravity de-fuzzification method; the analyzing unit 306, at a given Under the context condition, the conditional probability of each sub-node PSF and the state evaluation reliability node and the prior probability of the root node are subjected to a causal analysis of the Bayesian network to obtain the reliability of the state evaluation.
  • the apparatus may further include: a third determining module 40 coupled to the second determining module 20, for determining an operator status evaluation error, According to the relationship between the PSF of each root node and other PSFs other than itself, and the relationship between each PSF and the state evaluation reliability node, the PSF of the root node is determined to be in different states by the diagnostic analysis of the Bayesian network.
  • each module performs a corresponding function, wherein each module can be set in a server of the system, and when the operator analyzes according to different states, the operator state evaluation in the server
  • the reliability analysis device analyzes the reliability of the operator analysis process.
  • each module can also be set in the computer, and when the reliability analysis is required, it is controlled by the CPU.
  • the reliability analysis device of the above operator state evaluation By using the reliability analysis device of the above operator state evaluation, the accuracy of different operator state assessments can be further analyzed, which has certain practical significance.
  • Preferred embodiment Related technologies generally have the following disadvantages: (1) The current technology does not fully consider the characteristics of the digital human-machine system to construct the impact model of the state assessment; (2) The prior art does not fully consider the causal relationship of the PSF, so that the accuracy of the evaluation results Need to be improved; (3) The prior art lacks digital simulator data to support the quantification of state assessment.
  • a preferred embodiment of the present invention provides a reliability analysis method for operator status evaluation, which is described as follows: (1) Based on the situation environment analysis of the digital control room operator, Identify the PSF factors affecting the reliability of the operator's state assessment and its causal relationship and the causal relationship between the PSF and the state assessment reliability nodes. Establish a Bayesian network model for qualitative analysis of state assessments, and quantify the reliability of the operator's state assessment. Lay the foundation; (2) Bayesian network model based on established state assessment, through the simulator experiment to collect the prior probability and conditional probability of the network node, for difficult to collect data, event report analysis can be used to obtain the data.
  • the mental model is built through formal education, specific training, and operator experience, and is stored in the brain.
  • the state assessment process primarily develops a state model to describe the current plant state. If an event (such as an alarm) is very simple, the operator does not need any reasoning for the identification of the state of the plant, and is considered a skill-based state assessment. If an anomalous event belongs to a so-called "problem", the operator is required to explain the cause and effect of the problem to construct a state model, and the constructed state model is matched with the operator's mental model (ie, similarity matching), This process is called a regular state assessment.
  • the state model is the operator's understanding of the specific state of the system or component, and the state model is updated frequently as new information is collected.
  • the mental model is built through formal education, specific training, and operator experience, and is stored in the brain.
  • the state assessment process primarily develops a state model to assess the current state of the plant. If the operator is to evaluate the current state of the real power plant well, the operator needs to use his own mental model to identify the current state of the plant. This process is subject to the recognizability of the plant state, the operator's mental level / Mental models and the effects of psychological stress.
  • the mental level/mind model is derived from the knowledge and experience of the operator. The knowledge and experience are mainly influenced by the training of the organization and the communication and cooperation of the team.
  • the recognizability of the state presented by the power plant is mainly influenced by the automation level of the digital human-machine interface and system. If the digital human-machine interface is designed well, the information is conspicuous, and it is easy to collect information and identify The state of the system, if the system automation level is high, the operator does not participate in the specific task, it is easy to lose the understanding of the system state related to the task. In addition, the stress level has a great influence on the operator's matching between the state model and the mental model. The stress level is mainly affected by the severity of the event, the complexity of the task and the available time.
  • the complexity of the same task is mainly digitized.
  • the complexity of the tasks in the procedure is complicated by the tasks that the operator needs to complete.
  • the procedures or procedures are good for guiding the operator to respond to the plan.
  • the human-machine interface is not good. (such as many interface management tasks) it is difficult for operators to obtain useful information that is conducive to task completion.
  • the more serious the event the greater the psychological pressure of the operator, and the shorter the time available to complete the task, the greater the psychological pressure of the operator.
  • the status assessment is influenced by factors such as the level of communication and cooperation of the team, the level of training, digital procedures, digital human-machine interface, the severity of the incident, the available time of the incident, and the level of automation of the system.
  • the relationship diagram of the state assessment (or the Bayesian network model for state evaluation) is shown in Figure 6, where the lowest level of state assessment reliability is a state assessment reliability node.
  • the fuzzy Bayesian method for quantitative calculation of data acquisition and state evaluation is introduced.
  • (1) Data collection includes the following process.
  • the digital human-machine interface involved in the key tasks of the SGTR accident is evaluated according to the evaluation criteria of the human-machine interface design (from the ease of information collection, diagnosis and execution), and each screen is evaluated to obtain a probability distribution (assuming a total of Involving 100 screens, through the evaluation of the expert group, 90 pictures are good, 8 pictures are normal, 2 pictures are poor, then the prior probability distribution of the human-machine interface is: 0.9, 0.08, 0.02, the same can be The prior probability distribution of other influencing factors). Determine the conditional probability distribution of the network node.
  • the operator's knowledge and experience level or mental model level
  • stress level or easy identification of system state presentation
  • the experimental results of a group of operators which require the operator to assess the knowledge and experience acquired), the level of training and the level of communication in general (requires the operator to assess the knowledge and experience gained) As well as the training level difference and the general experimental results (the operator needs to assess the knowledge and experience acquired), so as to obtain the conditional probability distribution of the knowledge and experience part, the same reason, control the exchange of different levels of experimental variables, Get a conditional probability distribution of all knowledge and experience.
  • the conditional probability distribution of the node variables such as the complexity of the task can be obtained. If the variables are difficult to experiment (or difficult to control), expert judgment methods, event report statistical analysis or Modeling estimates such as regression techniques (see fuzzification described below).
  • the conditional probability distribution of the state evaluation reliability can be obtained.
  • the device is based on the data acquisition method judged by experts in the evaluation process, and the fuzzy method is adopted, and the evaluation procedure is as follows.
  • the evaluation procedure is as follows.
  • TM experts have different evaluation results of organizational factors due to different knowledge backgrounds and experiences, which affect the decision-making results. Therefore, it is necessary to form an expert group to eliminate this influence, and each expert assigns different weights.
  • the weight given by the z- th expert is Cl , ⁇ . e[0,l],
  • the probability of each factor being usually in the state through expert discussion can be evaluated by the triangular fuzzy number (ie, the most likely value; the best value; the worst value), such as (0.1, 0.3, 0.6). ), etc., or expressed in descriptive language, such as high, medium, low, etc.
  • a fuzzy membership function can be introduced to determine the probability value that a factor is in a certain state.
  • Discrete random variables V ⁇ X1, X2, ..., Xn ⁇
  • the corresponding nodes represent variables with finite state, and the nodes can be any abstract problems, such as device component status, test values, organizational factors, human diagnosis results, etc. .
  • the directed edge E represents the probability causal relationship between nodes.
  • the starting node of the directed edge is the parent node of the terminal node, called the child node, and the node without the parent node has only the child node is called the root node.
  • DAG implies a conditional independence hypothesis: Given its set of parent nodes, each variable is independent of its non-children. P is the quantitative part, which is the probability distribution on V. For the discrete case, it can be represented by the Conditional Probability Table (CPT), which is used to quantify the parent node to the child node. Impact.
  • CPT Conditional Probability Table
  • the probability distribution function of the root node is the edge probability distribution function. Since the probability of this kind of node is not conditional on other nodes, its probability is a prior probability, and other nodes are conditional probability distribution functions.
  • the chain rule shows that a BN is a description of the joint distribution of all variables in the DAG, and the edge probability and conditional probability of each node in the network can be calculated.
  • Bayesian network The analysis principle of Bayesian network is based on Bayes probability theory, and the analysis process is essentially the probability calculation process.
  • the analysis is mainly based on the following three formulas (ie, formulas 1-1, 1-2, and 1-3). Joint probability corpse ⁇ , ⁇ :
  • Bayesian network is as an analysis machine (also called inference engine) for calculating event beliefs. Its task is to calculate The probability of occurrence of certain events under the conditions of the evidence (or observation data). Assuming the evidence e is known, there are: In Bayesian network analysis, there are two main processes, namely causal analysis and diagnostic analysis. This will be explained separately below. Causal analysis is a top-down reasoning that is inferred from the cause. The Bayesian network analysis calculation is used to determine the probability of occurrence of the result for a given cause or evidence. Under normal circumstances, that is, each variable obeys the expert group judgment and the initial probability distribution based on the experimental data of the simulator.
  • the prior probability of the root node in the network can be expressed by Table 1 (the fuzzy prior probability of the root node). It is assumed that the fuzzy prior priorities of the exchange and cooperation levels of the obtained team are in different states (0.09, 0.10, 0.11), (0.29, 0.30, 0.31), (0.59, 0.60, 0.61), respectively, and the team is exchanged after the fuzzy solution. And the level of cooperation is in an inadequate state, acceptable state and sufficient state The rate distribution is (0.1, 0.3, 0.6). Similarly, the probability distribution of other node variables can be obtained through expert judgment or simulator experiment. Table 1
  • the conditional probability of the intermediate variable can be represented by Table 2.
  • the assumed data is shown in Table 2, that is, the conditional probability of the intermediate variable "Mental Model Level M M " (Knowledge and Experience Group I exchange level of cooperation, Training level).
  • Table 2 is the conditional probability of the intermediate variable "Mental Model Level M M " (Knowledge and Experience Group I exchange level of cooperation, Training level).
  • the Bayesian rule can be used to calculate the corresponding posterior probability. For example, to calculate the probability that the "team exchange and cooperation level" is in an "insufficient” state, according to formula (1-3):
  • the diagnostic analysis is to calculate the posterior probability of the root node PSF in a bad state, and compare it with the prior probability that they are in a bad state, and obtain the percentage change of each variable. Identify the factors most likely to trigger a state assessment error and provide decision support for the prevention of mistakes.

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Operations Research (AREA)
  • Marketing (AREA)
  • Game Theory and Decision Science (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Testing And Monitoring For Control Systems (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Complex Calculations (AREA)

Abstract

本发明公开了一种操作员状态评估的可靠性分析方法及装置,其中,该方法包括:确定待使用的多个PSF,其中,PSF用于状态评估;确定多个PSF中各个PSF与除自身外的其他PSF之间的关联关系以及各个PSF与状态评估可靠性节点的关联关系;根据上述各关联关系建立基于PSF因果关系的操作员状态评估的可靠性分析模型以分析操作员的可靠性。通过运用本发明,解决了相关技术没有考虑操纵员本身所受的情境环境因子的影响以及它们的影响因果关系,从而可能带来重复计算其影响的可能,对状态评估失误概率可能造成错误的估计的问题,进而为操纵员状态评估可靠性分析提供定性与定量方法,为电厂降低操纵员状态评估失效概率提供对策。

Description

操作员状态评估的可靠性分析方法及装置
技术领域 本发明涉及核电厂检测及人因可靠性分析领域, 更具体地, 涉及一种操作员状态 评估的可靠性分析方法及装置。 背景技术 当核电厂发生异常状态时, 操纵员将根据核电厂的状态参数情况构建一个合理的 和合乎逻辑的解释, 来评估电厂所处的状态, 作为后续的响应计划和响应执行决策的 依据, 这一系歹 !j行为过程称为状态评估 ( Situation Assessment or Situation Awareness, 简称为 SA)。操纵员对异常事件的正确状态评估对于操纵员行为的正确响应至关重要。 迄今为止, Endsley于 1995年发表在 "Human factors"期刊上名为" Toward a theory of situation awareness in dynamic" 中建立的状态评估模型中将其分为 3个层次:对当前环 境中关键元素的认知 (Perception) , 对当前状态的理解 (Comprehension) , 以及对未 来状态的预计 (Prqjection)。 Endsley 分析了影响状态评估的因素有个体因素和系统 / 任务因素。 Adams和 Tenney于 1995 年发表在 "Human factors"期刊上名为" Situation awareness and the cognitive management of complex-systems"中支持由 Neisser提出的知 觉环(Perceptual cycle)模型来描述状态评估,知觉环模型强调人与环境交互的动态性, 模型中包含三个组分: 对象 (实际的当前环境)、 图式 (当前环境的图式)和探索 (在 环境中的搜索行为)。以图式形式或心智模型构成的知识使操作者对环境中的信息产生 心理预期, 活跃的图式会指导操作者对特定信息的搜索和解释行为, 同时, 从环境中 获取信息会被图式吸收并用于修订和更新图式, 再次指导信息的搜索以达到对情境的 知觉, 是一个不断循环的过程, 如图 1 所示。 Bedney 和 Meister 于 1999 年发表在 "International Journal of Cognitive Ergonomics"期干1 J上名为 "Theory of activity and situation awareness"基于行为理论提出了状态评估的子系统交互模型, 共包括八个功能 模块, 即输入信息的含义、 印象-目标、 主观上认为相关的任务条件、 动机和重要性、 定向和探索行为、 评估的标准、 经验、 概念模型, 它们之间存在交互影响作用, 每个 功能模块对状态评估的形成有不同的功能。 这些模型基本上描述了操纵员处理信息和环境进行交互以获得状态感知的基本原 理和一般的特征, 在阐明状态评估的认知机理及影响状态评估的主要因素方面做出了 贡献, 但没有考虑数字化控制系统中的状态评估特征, 并且它们只是定性的分析, 而 没有对状态评估的可靠性进行量化。 在状态评估的定量建模方法, Miao A.X., Zachanas GL.和 Kao S.P.,于 1997年发表 的" A computational situation assessment model for nuclear power plant Operations" ^■文中 采用贝叶斯方法对操纵员的状态评估进行了定量计算, Kim M.C.和 Seong P.H于 2006 年在" An analytic model for situation assessment of nuclear power plant operators based on Bayesian inference"—文中基于贝叶斯分析建立了状态评估的定量分析模型, Kim 和 Seong于 2009年在 "A computational model for evaluating the effects of attention, memory, and mental models on situation assessment of nuclear power plant operators" ^ ^文也基于贝 叶斯网络建立了考虑人因影响的状态评估定量计算模型, 但是上述方法只是描述操纵 员状态评估的历程, 数据也只是基于假设, 并且没有考虑操纵员本身所受的情境环境 因子的影响以及它们的影响因果关系, 从而可能带来重复计算其影响的可能, 对状态 评估失误概率可能造成错误的估计。 发明内容 本发明实施例旨在提供一种操作员状态评估的可靠性分析方法及装置, 以至解决 相关技术没有考虑操纵员本身所受的情境环境因子的影响以及它们的影响因果关系, 从而可能带来重复计算其影响的可能, 对状态评估失误概率可能造成错误的估计的问 题。 根据本发明实施例的一个方面, 提供了一种操作员状态评估的可靠性分析方法, 包括: 确定待使用的多个行为形成因子(Performance Shaping Factors, 简称为 PSFs或 PSF), 其中, 所述 PSF用于状态评估; 确定所述多个 PSF中各个 PSF与除自身外的 其他 PSF之间的关联关系以及各个 PSF与状态评估可靠性节点的关联关系;根据所述 多个 PSF以及所述多个 PSF中各个 PSF之间的关联关系、 以及所述各个 PSF与所述 状态评估可靠性节点的关联关系建立基于 PSF因果关系的操作员状态评估的可靠性分 析模型以分析所述操作员的可靠性。 优选地, 所述操作员状态评估的可靠性分析模型为贝叶斯网络模型。 优选地,根据所述多个 PSF以及所述多个 PSF中各个 PSF之间的关联关系、 以及 所述各个 PSF与所述状态评估可靠性节点的关联关系建立基于 PSF因果关系的操作员 状态评估的可靠性分析模型以分析所述操作员的可靠性包括: 确定所述贝叶斯网络模 型中各个根节点 PSF处于不同状态的先验概率分布, 其中, 所述根节点 PSF为没有被 其他节点指向的 PSF节点;根据所述各个根节点 PSF与除自身外的其他 PSF之间的关 联关系、 以及所述各个 PSF与所述状态评估可靠性节点的关联关系确定所述各个子节 点 PSF和所述状态评估可靠性节点处于不同状态的条件概率分布, 其中, 所述子节点 PSF为被其他节点指向的 PSF节点; 将所述条件概率分布与所述先验概率分布进行贝 叶斯网络的因果分析以得到特定情境条件下的状态评估可靠性。 优选地, 通过模糊方法对所述贝叶斯网络模型的概率分布的值进行计算。 优选地, 通过模糊方法对所述贝叶斯网络模型的概率分布的值进行计算包括: 通 过三角模糊数计算所述各个根节点 PSF处于不同状态时的先验概率分布; 通过所述三 角模糊数计算所述各个子节点 PSF与所述状态评估可靠性节点处于不同状态时的条件 概率分布;通过三角形重心解模糊的方法确定所述先验概率的值和所述条件概率的值, 并将所述先验概率的值与所述条件概率的值进行贝叶斯网络的因果分析以得到所述状 态评估的可靠性。 优选地,根据所述多个 PSF以及所述多个 PSF中各个 PSF之间的关联关系、 以及 所述各个 PSF与所述状态评估可靠性节点的关联关系建立基于 PSF因果关系的操作员 状态评估的可靠性分析模型以分析所述操作员的可靠性之后, 还包括: 在确定操纵员 状态评估失误的情况下,根据所述各个根节点 PSF与除自身外的其他 PSF之间的关联 关系, 以及所述各个 PSF与所述状态评估可靠性节点的关联关系, 通过贝叶斯网络的 诊断分析确定所述各个根节点 PSF的后验概率分布;将所述根节点 PSF处于预设多个 状态中最差状态的后验概率分布与所述根节点 PSF处于所述预设多个状态中最差状态 的先验概率分布进行比较以得到影响状态评估可靠性的关键要素, 以确定预防状态评 估失误的对策。 根据本发明实施例的另一个方面,提供了一种操作员状态评估的可靠性分析装置, 包括: 第一确定模块, 用于确定待使用的多个行为形成因子 PSF, 其中, 所述 PSF用 于状态评估; 第二确定模块, 用于确定所述多个 PSF 中各个 PSF 与除自身外的其他 PSF之间的关联关系以及各个 PSF与状态评估可靠性节点的关联关系; 建立模块, 用 于根据所述多个 PSF、 所述多个 PSF中各个 PSF之间的关联关系、 以及所述各个 PSF 与所述状态评估可靠性节点的关联关系建立基于 PSF因果关系的操作员状态评估的可 靠性分析模型以分析所述操作员的可靠性。 优选地, 所述建立模块建立的所述操作员状态评估的可靠性分析模型为贝叶斯网 络模型。 优选地, 所述建立模块包括: 第一确定单元, 用于确定所述贝叶斯网络模型中各 个根节点 PSF处于不同状态的先验概率分布, 其中, 所述根节点 PSF为没有被其他节 点指向的 PSF节点; 第二确定单元, 用于根据所述各个根节点 PSF与除自身外的其他 PSF之间的关联关系、 以及所述各个 PSF与所述状态评估可靠性节点的关联关系确定 所述各个子节点 PSF和所述状态评估可靠性节点处于不同状态的条件概率分布,其中, 所述子节点 PSF为被其他节点指向的 PSF节点; 分析单元, 用于将所述条件概率分布 与所述先验概率分布进行贝叶斯网络的因果分析以得到特定情境条件下状态评估可靠 性。 优选地, 所述装置还包括: 第三确定模块, 用于在确定操纵员状态评估失误的情 况下, 根据所述各个根节点 PSF与除自身外的其他 PSF之间的关联关系, 以及所述各 个 PSF与状态评估可靠性节点的关联关系, 通过贝叶斯网络的诊断分析确定所述各个 根节点 PSF的后验概率分布; 比较模块, 用于将所述根节点 PSF处于预设多个状态中 最差状态的后验概率与根节点 PSF处于预设多个状态中最差状态的先验概率进行比较 以得到影响状态评估可靠性的关键要素, 以确定预防状态评估失误的对策。 本发明实施例采用在对操作员状态评估的可靠性进行分析时, 综合考虑了操作员 在状态评估的过程中的多个 PSF及各个 PSF与其他 PSF之间的关联关系, 以及各个 PSF与状态评估可靠性节点的关联关系, 并根据上述各因素建立基于 PSF因果关系的 操作员状态评估的可靠性分析模型以分析所述操作员的可靠性。 通过运用本发明实施 例, 解决了相关技术没有考虑操纵员本身所受的情境环境因子的影响以及它们的影响 因果关系, 从而可能带来重复计算其影响的可能, 对状态评估失误概率可能造成错误 的估计的问题, 进而为操纵员状态评估可靠性分析提供定性与定量方法, 为电厂降低 操纵员状态评估失效概率提供对策。 附图说明 附图用来提供对本发明的进一步理解, 构成本申请的一部分, 本发明的示意性实 施例及其说明用于解释本发明, 并不构成对本发明的不当限定。 在附图中: 图 1示出了相关技术的状态评估模型的示意图; 图 2示出了本发明实施例的操作员状态评估的可靠性分析方法的流程图; 图 3示出了本发明实施例的操作员状态评估的可靠性分析装置的结构示意图一; 图 4示出了本发明实施例的操作员状态评估的可靠性分析装置建立模块的结构示 意图; 图 5示出了本发明实施例的操作员状态评估的可靠性分析装置的结构示意图二; 以及 图 6示出了本发明优选实施例的操作员状态评估的贝叶斯网络模型的示意图。 具体实施方式 下面将参考附图并结合实施例, 来详细说明本发明实施例。 基于相关技术没有考虑操纵员本身所受的情境环境因子的影响以及它们的影响因 果关系, 从而可能带来重复计算其影响的可能, 对状态评估失误概率可能造成错误的 估计的问题, 本发明实施例提供了一种操作员状态评估的可靠性分析方法, 该方法的 流程如图 2所示, 包括步骤 S202至步骤 S206: 步骤 S202, 确定待使用的多个 PSF, 其中, PSF用于状态评估; 步骤 S204, 确定多个 PSF中各个 PSF与除自身外的其他 PSF之间的关联关系, 以及各个 PSF与状态评估可靠性节点的关联关系; 步骤 S206, 根据多个 PSF以及多个 PSF中各个 PSF之间的关联关系, 以及各个
PSF与状态评估可靠性节点的关联关系建立基于 PSF因果关系的操作员状态评估的可 靠性分析模型以分析操作人员的可靠性。 本发明实施例采用在对操作员状态评估的可靠性进行分析时, 综合考虑了操作员 在状态评估的过程中的多个 PSF及各个 PSF与其他 PSF之间的关联关系,并根据上述 PSF及其之间的关联关系以及各个 PSF与状态评估可靠性节点的关联关系建立操作员 状态评估的可靠性分析模型以分析操作人员的可靠性。 通过运用本发明实施例, 解决 了相关技术没有考虑操纵员本身所受的情境环境因子的影响以及它们的影响因果关 系, 从而可能带来重复计算其影响的可能, 对状态评估失误概率可能造成错误的估计 的问题, 进而为操纵员状态评估可靠性分析提供定性与定量方法, 为电厂降低操纵员 状态评估失效概率提供对策。 在实施过程中, 操作员状态评估的可靠性分析模型可以为贝叶斯网络模型。 在根 据多个 PSF以及多个 PSF中各个 PSF之间的关联关系, 以及各个 PSF与状态评估可 靠性节点的关联关系建立操作员状态评估的可靠性分析模型以分析操作人员的可靠性 的过程中,需要先确定各个根节点 PSF处于不同状态的先验概率分布;再根据各个 PSF 与除自身外的其他 PSF之间的关联关系以及各个 PSF与状态评估可靠性节点的关联关 系确定各个 PSF与状态评估可靠性节点处于不同状态的条件概率分布; 通过贝叶斯网 络的因果分析 (或根据结果逐步推理) 计算操纵员状态评估的可靠性, 随后通过贝叶 斯网络的诊断分析, 得到状态评估的贝叶斯网络模型中的根节点处于不同状态的后验 概率分布, 将根节点 PSF处于不良状态 (即预设多个状态中最差状态) 的后验概率与 根节点 PSF处于不良状态的先验概率进行比较以得到影响状态评估可靠性的关键要素 (或称关键 PSF)。 实施过程中, 是根据贝叶斯理论的因果分析, 进行状态评估可靠性计算, 并且根 据贝叶斯理论的诊断分析, 识别状态评估贝叶斯网络模型中根节点的后验概率分布, 并将根节点 PSF处于不良状态的先验概率值与根节点 PSF处于不良状态的后验概率值 进行比较以得到影响状态评估可靠性的关键要素。 但由于各个 PSF的状态等级评定的复杂性和不确定性, 可以进一步通过模糊方法 对贝叶斯网络模型的概率分布的值进行计算。 在通过模糊方法对贝叶斯网络模型的概率分布的值进行计算的过程中, 可以利用 三角模糊数对各个 PSF进行评价, 则上述分析操作员可靠性的过程可以是先通过三角 模糊数计算各个根节点 PSF处于不同状态时的先验概率分布; 再通过三角模糊数计算 各个子节点 PSF与状态评估可靠性节点处于不同状态的条件概率分布; 再通过三角形 重心解模糊的方法确定先验概率的值和条件概率的值。 在给定的情境条件下, 将子节 点 PSF和状态评估可靠性节点的条件概率与根节点 PSF的先验概率进行贝叶斯网络的 因果分析以得到状态评估的可靠性。 通过上述方法进行操作员状态评估的可靠性分析过程中,考虑了 PSF的因果关系, 可根据不同 PSF的不同状态进行可靠性估计, 提升了可靠性分析的准确度。 本发明实施例还提供了一种操作员状态评估的可靠性分析装置, 该装置的结构示 意可以如图 3所示, 包括: 第一确定模块 10, 用于确定待使用的多个 PSF, 其中, PSF 用于状态评估; 第二确定模块 20, 与第一确定模块 10耦合, 用于确定多个 PSF中各 个 PSF与除自身外的其他 PSF之间的关联关系以及各个 PSF与状态评估可靠性节点的 关联关系; 建立模块 30, 与第一确定模块 10和第二确定模块 20耦合, 用于根据多个 PSF以及多个 PSF中各个 PSF之间的关联关系,以及各个 PSF与状态评估可靠性节点 的关联关系建立基于 PSF因果关系的操作员状态评估的可靠性分析模型以分析操作人 员的可靠性。 其中, 建立模块建立的操作员状态评估的可靠性分析模型为贝叶斯网络 模型。 优选地, 本装置还可以通过贝叶斯网络的诊断分析, 识别影响状态评估可靠性 的关键要素。 图 4示出了建立模块 30的结构示意图, 其中, 第一确定单元 302, 用于确定各个 根节点的 PSF处于不同状态的先验概率分布;第二确定单元 304,与第一确定单元 302 耦合,用于根据各个 PSF与除自身外的其他 PSF之间的关联关系, 以及各个 PSF与状 态评估可靠性节点的关联关系确定各个子节点 PSF与状态评估可靠性节点的条件概率 分布; 分析单元 306, 与第一确定单元 302和第二确定单元 304耦合, 在给定的情境 条件下,用于将子节点 PSF和状态评估可靠性节点的条件概率与根节点 PSF的先验概 率进行贝叶斯网络的因果分析以得到状态评估的可靠性。 在实施过程中, 如果选择三角模糊数的方式进行状态估计的可靠性分析, 则第一 确定单元 302,还用于通过三角模糊数和三角形重心解模糊的方法计算各个根节点 PSF 处于不同状态时的先验概率分布; 第二确定单元 304, 还用于通过三角模糊数和三角 形重心解模糊的方法计算各个子节点 PSF和状态评估可靠性节点的条件概率分布; 分 析单元 306, 在给定的情境条件下, 用于将各个子节点 PSF和状态评估可靠性节点的 条件概率与根节点的先验概率进行贝叶斯网络的因果分析以得到状态评估的可靠性。 图 5示出了上述装置的一种优选实施方式的结构示意图, 该装置还可以包括: 第 三确定模块 40, 与第二确定模块 20耦合, 用于在确定操纵员状态评估失误的情况下, 根据各个根节点的 PSF与除自身外的其他 PSF之间的关联关系,以及各 PSF与状态评 估可靠性节点的关联关系, 通过贝叶斯网络的诊断分析确定各个根节点的 PSF处于不 同状态的后验概率分布; 比较模块 50, 与第三确定模块 40和第一确定模块 10耦合, 用于将根节点 PSF处于不良状态的后验概率分布与根节点 PSF处于不良状态的先验概 率分布进行比较以得到影响状态评估可靠性的关键要素, 以确定预防状态评估失误的 对策。 在上述装置实施上述方法的过程中, 其各个模块执行着相应的功能, 其中, 各模 块都可以设置在系统的服务器中, 当操作员根据不同的状态进行分析时, 服务器中的 操作员状态评估的可靠性分析装置对操作员分析过程的可靠性进行分析。 当然, 各模 块也可以设置在计算机中, 当需要进行可靠性分析时, 通过 CPU进行控制。通过运用 上述操作员状态评估的可靠性分析装置, 可以进一步分析不同操作员状态评估的准确 性, 具有一定的实用意义。 优选实施例 相关技术通常存在以下的缺点: (1 )现在技术没有充分考虑数字化人 -机系统的特 征来构建状态评估的影响模型; (2) 现有技术没有充分考虑 PSF的因果关系, 使得评 估的结果精度有待提升; (3 ) 现有技术缺乏数字化模拟机数据来支持状态评估的定量 化。 基于上述待解决的问题, 本发明优选实施例提供了一种操作员状态评估的可靠性 分析方法, 该方法要保护技术方案阐述如下: (1 ) 基于数字化主控室操纵员的情境环 境分析,识别影响操纵员的状态评估可靠性的 PSF因子及其因果关系以及各 PSF与状 态评估可靠性节点的因果关系, 建立状态评估定性分析的贝叶斯网络模型, 为操纵员 状态评估可靠性的量化奠定基础; (2) 基于建立的状态评估的贝叶斯网络模型, 通过 模拟机实验来收集网络节点的先验概率和条件概率, 对于难以收集的数据, 可采用事 件报告分析来获取数据。 为确保数据和结果的有效性, 建立一种操纵员状态评估可靠 性评定的模糊贝叶斯方法, 提高分析的精度。 下面通过具体实施例来说明本优选实施例的详尽技术方案。 首先, 介绍数字化核电厂主控室操纵员的状态评估行为。 当核电厂发生异常状态时, 操纵员将根据核电厂的状态参数情况构建一个合理的 和合乎逻辑的解释, 来评估电厂所处的状态, 作为后续的响应计划和响应执行决策的 依据。 这一系列过程称为状态评估, 并涉及两个相关的模型, 即状态模型和心智模型。 状态模型就是操纵员对特定电厂系统所呈现出的状态的理解, 并且当收集到新信息的 时候, 状态模型会被经常更新。 心智模型是通过正式的教育、 具体的培训和操纵员经 验来构建的, 并且存储在大脑中。 状态评估过程主要就是发展一个状态模型来描述当 前的电厂状态。 如果一个事件 (如报警) 非常简单, 操纵员对电厂状态的辨识不需要任何推理, 则认为是技能型的状态评估。如果一个异常事件属于所谓的 "问题", 要求操纵员对该 问题产生的原因和影响进行说明来构建状态模型, 并且构建好的状态模型与操纵员的 心智模型进行匹配(即相似性匹配), 则这个过程称为规则型的状态评估。 同样, 对于 不熟悉的状态模式, 要求操纵员评估和预测可能的电厂状态, 然后分析问题空间的结 构和功能之间更加抽象的逻辑关系, 进行深层次的分析, 逐渐形成一个状态模型并进 行验证, 最后确定电厂状态, 这个过程被认为是知识型的状态评估。 其次, 介绍操纵员的状态评估的贝叶斯网络模型。 通过组建专家组 (包括核电厂操纵员班组以及人因专家) 识别出影响操纵员状态 评估可靠性的影响因素以及它们的因果关系, 一般来说, 当核电厂发生异常事件后, 操纵员的状态评估涉及两个相关的模型, 即状态模型和心智模型。 状态模型就是操纵 员对系统或组件的特定状态的理解, 并且当收集到新信息的时候, 状态模型会被经常 更新。 心智模型是通过正式的教育、 具体的培训和操纵员经验来构建的, 并且存储在 大脑中。 状态评估过程主要就是发展一个状态模型来评估当前的电厂状态。 如果操纵 员要很好地评估出真实的电厂当前状态, 则操纵员需要利用其自身的心智模型去辨识 出电厂当前的状态, 这个过程受电厂状态呈现的易识别性、 操纵员的心智水平 /心智模 型以及心理压力的影响。 心智水平 /心智模型来源于操纵员的知识和经验, 知识和经验 主要受组织培训的影响和班组的交流与合作的影响, 如果培训不够, 则操纵员的知识 和经验会受影响, 班组的交流与合作可以补充操纵员个体的知识和经验的不足。 电厂所呈现的状态的易识别性 (状态模型的另一种解释) 主要受数字化人机界面 和系统的自动化水平的影响, 如果数字化人机界面设计好, 则信息醒目, 容易搜集信 息和识别出系统所处的状态, 如果系统自动化水平高, 则操纵员没有参与到具体的任 务中, 则容易丧失与任务相关的系统状态的理解。 另外, 压力水平对操纵员在状态模 型和心智模型之间的匹配有很大的影响, 压力水平主要受事件的严重度、 任务的复杂 性及可用时间的影响, 同样任务的复杂性主要受数字化规程设计的好坏与数字化人机 界面设计的好坏的影响, 规程中的任务复杂则操纵员需要完成的任务复杂, 规程或程 序好有利于指导操纵员做出响应计划, 人机界面不好 (如诸多的界面管理任务) 则操 纵员难以获取有利于任务完成的有用信息。 再者, 事件越严重, 操纵员的心理压力越 大, 完成任务的可用时间越短, 则操纵员的心理压力越大。 通过上述分析, 状态评估 受班组的交流与合作水平、 培训水平、 数字化规程、 数字化人机界面、 事件的严重度、 事故处置的可用时间以及与系统的自动化水平等因素的影响, 这些 PSF因子与状态评 估的影响关系图 (或称状态评估的贝叶斯网络模型) 见图 6所示, 其中, 最下层的状 态评估可靠性就是一种状态评估可靠性节点。 最后, 介绍数据的获取和状态评估定量计算的模糊贝叶斯方法。
( 1 ) 数据采集包括以下过程。
( 1.1 ) 基于模拟机实验的数据获取。 确定网络节点的先验概率分布: 选择典型的事故场景(如 SGTR、 LOCA、 主蒸汽 管道破口、 全厂失电等) 进行实验, 对事故场景中关键点的任务所涉及的数字化人机 界面、 数字化规程、 任务的复杂性、 事故场场景下的时间窗口、 交流与合作水平、 培 训水平等影响因子进行评定, 识别主要影响因素的概率分布。 比如针对 SGTR事故的 关键任务所涉及的数字化人机界面按人机界面设计好坏的评定标准 (从信息搜集、 诊 断和执行的容易度方面) 进行每个画面进行评定, 得到概率分布 (假设共涉及 100副 画面, 通过专家组的评定, 得到 90幅画面是好的, 8幅画面一般, 2幅画面差, 则得 到人机界面的先验概率分布为: 0.9,0.08,0.02, 同理可得其他影响因素的先验概率分 布)。 确定网络节点的条件概率分布。 在实验过程中对操纵员的知识和经验水平 (或称 心智模型水平)、 压力水平、 状态模型水平(或称系统状态呈现的易识别性)等进行评 定。这需要被试针对每完成一个关键的任务进行评定(要求实事求是)。统计评定结果, 得到心智模型水平(假设有三个水平, 好、 中、 差)、 压力水平(假设有三个水平, 好、 中、 差)、 状态模型水平 (假设有三个水平, 好、 中、 差)等的条件概率分布。 同时通 过状态评估的实验结果的统计分布, 得到状态评估可靠性的条件概率分布。 比如培训和交流水平影响操纵员的知识和经验或心智模型水平, 则选择不同培训 水平的人员进行实验, 包括培训水平好、 中、 差以及交流水平一般的情况下分别进行 实验, 得到培训水平好的一组操纵员的实验结果 (需操纵员对其获取的知识和经验进 行评定)、培训水平中等与交流水平一般的情况下的实验结果(需操纵员对其获取的知 识和经验进行评定) 以及培训水平差和交流一般的实验结果 (需操纵员对其获取的知 识和经验进行评定), 从而得到知识和经验的一部分的条件概率分布, 同理, 控制好交 流不同水平的实验变量, 可得到所有知识和经验的条件概率分布。 同样, 控制好其他 可以控制的变量, 可得到任务的复杂性等节点变量的条件概率分布, 如果对于难以进 行实验的变量(或难以控制的), 可采用专家判断的方法、事件报告统计分析或回归技 术等进行建模估计(见下面介绍的模糊化处理)。最终根据状态评估可靠性的测量结果, 可得到状态评估可靠性的条件概率分布。
( 1.2 ) 基于专家判断的数据获取。 对于难以进行实验的变量(或难以控制的), 比如事件的严重度等节点变量等, 可 采用专家判断的方法来获取数据 (如果有足够多的事件样本, 也可采用事件统计的方 法来获取)。 由于因素状态等级评定的复杂性和不确定性, 以及专家知识、 能力、 经验 的有限性, 使某些专家难以确定因素状态等级的确切值, 因此, 导致专家可能用描述 性语言或用范围值来表达, 如"大约 7"、 " 很可能在 5-7这个范围"、 " ( 3, 5, 7 ) "等。 并且, 决策者认为, 模糊判断比确切值判断更可信, 更符合人们的真实思维, 因此, 我们提出通过模糊方法对 PSF因子处于不同状态的概率分布进行评价, 本优选实施例 的装置在评价过程中是基于专家判断的的数据获取方式, 采用模糊方法, 其评价程序 如下。 第一, 组建专家组。 不同的专家由于知识背景和经验不同对组织因素的评价结果 不同, 从而影响决策结果, 因此, 需组建专家组来消除这种影响, 并且每个专家分配 不同的权重。假设有™个专家组成的专家组,且第 z个专家赋予的权重为 Cl, ς. e[0,l],
∑ς=ι。
;=1 第二, 确定 PSF处于不同状态的概率。 通过专家讨论将每个因素通常处于何种状 态的概率可采用三角模糊数 (即最有可能的值; 最好的值; 最差的值) 对组织因素进 行评价, 如 (0.1, 0.3, 0.6) 等, 或用描述性语言来表示, 如高、 中、 低等。 对于描 述性语文可引入模糊隶属函数来确定因素处于某种状态的概率值。 第三, 计算各因素的综合概率分布值并解模糊。 依据专家权重和相应的因素状态概率,可计算各因素的状态概率分布, 公式如下: 二^^^+^2 ^? ' ^" ^ 其中, 是因素 的模糊综合得分, 它是一个三角 模糊数: (∑l^ck,±m^ck,±u^ck) 0 为了将综合的三角模糊数转化为确切值, 可通过三角形重心解模糊的方法求解, 其公式参考如下: F JUiH +l 其中, M '表示最大可能值, ^ ^表示最可能
3
值, 表示最小可能值。
(2) 贝叶斯网络的分析 贝叶斯网络(简称为 BN)是由节点和边组成的有向无环图(Directed Acyclic Graph, 简称为 DAG), 可以用 N=«V, E>, ?>来描述。 离散随机变量 V={X1, X2, ..., Xn} 对应的节点表示具有有限状态的变量,节点可以是任何抽象的问题,如设备部件状态、 测试值、 组织因素、 人的诊断结果等。 有向边 E表示节点间的概率因果关系, 有向边 的起始节点 是终节点 的父节点, 称为子节点, 没有父节点只有子节点的节点称为 根节点。 DAG蕴涵了一个条件独立假设: 给定其父节点集, 每一个变量独立于它的非 子孙节点。 P 为定量部分, 是 V 上的概率分布, 对于离散情况, 可用条件概率表 (Conditional Probability Table, 简称为 CPT) 来表示, 用于定量说明父节点对子节点 的影响。 根节点的概率分布函数为边缘概率分布函数, 由于该类节点的概率不以其他 节点为条件, 故其概率为先验概率, 其他节点为条件概率分布函数。 链式法则表明一个 BN就是在 DAG中所有变量的联合分布的一种描述,并且网络 中每个节点的边缘概率和条件概率都可计算。贝叶斯网络的分析原理是基于 Bayes 概 率理论, 分析过程实质上就是概率计算过程。主要根据下列三个公式(即公式 1-1、 1-2 和 1-3) 进行分析计算。 联合概率尸 ^···,^^:
P(U) = P(X1,X2,---,X„) = lPiX,
Figure imgf000014_0001
其中, τ,为 X,父节点的集合。 ,.的边缘概率为:
P(Xt)= ∑ P(U) (1-2) 贝叶斯网络的主要应用就是作为一个用于计算事件信念的分析机 (也可以称为推 理机), 其任务是计算"在给定的证据(或观察数据) 的条件下, 某些事件的发生概率。 假设已知证据 e, 则有:
Figure imgf000014_0002
在贝叶斯网络分析中, 主要包括两个过程, 即因果分析和诊断分析。 下面分别对 其进行说明。 因果分析由原因推知结论, 是一种自顶向下的推理。 在给定的原因或证据的条件 下, 使用贝叶斯网络分析计算, 求出结果发生的概率。 在正常情况下, 即各变量服从 专家组判断和基于模拟机实验数据得到的初始概率分布, 比如网络中的根节点的先验 概率可用表 1 (根节点的模糊先验概率) 来表示。 假设得到的班组的交流与合作水平 的处于不同状态的模糊先验概率为(0.09,0.10,0.11)、 (0.29,0.30,0.31)、 (0.59,0.60,0.61), 分别解模糊后得到班组交流与合作水平处于不充分状态、 可接受状态和充分状态的概 率分布为 (0.1 ,0.3,0.6), 同理可通过专家判断或模拟机实验得到其他节点变量的概率 分布。 表 1
Figure imgf000015_0001
同理可得中间变量的条件概率, 可用表 2来表示, 假设得到的数据见表 2所示, 即中间变量"心智模型水平 MM"的条件概率 ^ (知识和经验 I班组交流合作水平,培训水平)。 表 2
Figure imgf000015_0002
Figure imgf000016_0001
则班组的交流与合作水平、 培训水平引起操纵员的知识和经验(或心智模型水平) 处 于"低"水平状态的概率 (用 Ρ(ΜΜ=ΜΜΟ表示)可根据公式(1-2)有: xP(
Figure imgf000016_0002
)x[P(TR= TR )xP(MM=MM, 11 C。=Co , TR = TR )+P(TR = ¾2) xP(MM=Mu 11 C。 =C02, TR = T^2
=c 3, =rR,3)] 将获得的数据代入公式可得 P(MM= 1), 同样可计算得 Ρ(ΜΜ= Μ,2), 和 ( M= M,3)O因此,可得到了中间变量"知识和经验"的概率处于不同状态的概率分布。 同理可计算得到其他节点变量的概率分布。 最终计算得到状态评估可靠性 P=R。 诊断分析是由结论推知原因, 是一种自底向上的推理过程。 目的是在已知结果时, 找出产生该结果的各种原因的可能性。 已知发生了某些结果, 根据贝叶斯网络计算, 得到造成该结果发生的原因和发生的概率。 在状态评估可靠性的模糊贝叶斯网络模型 中, 假设已发生状态评估失误, 则利用贝叶斯法则可计算出相应的后验概率。 比如要 计算"班组交流与合作水平 "处于 "不充分"状态的概率, 则根据公式 (1-3) 可得:
P(C0 = C0,1,RR =RR,l)
P(C0 = C0.\ \RR = RR.l) ,其中, RR =i?fl2表示发生状态评估
^ P(RR =RR.2) ^
失误 由公式 (1-1) 可算出^^^^^,^ 二^ , 由公式 (1-2) 可算出尸0¾ = ¾2), 从而可以算出所求的值。 诊断分析是将计算得到根节点 PSF处于不良状态的后验概率, 分别与它们处于不 良状态的先验概率进行比较, 可得到各变量变化的百分比。 识别最有可能引发状态评 估失误的影响因素, 为失误的预防提供决策支持。 通过运用本发明实施例, 可以为数字化主控室操纵员状态评估可靠性分析提供定 性与定量方法与工具, 为电厂降低操纵员状态评估失效概率提供对策; 为核电厂数字 化主控室操纵员人因可靠性分析 (HRA) 与概率安全评价 (PSA) 提供操纵员状态评 估可靠性接口数据与计算工具, 建立的状态评估可靠性计算的模糊贝叶斯方法, 提高
HRA和 PSA分析的精度; 为核电厂数字化主控室操纵员防人因失误培训与场景开发 提供支持; 为数字化工业系统主控室作业人员状态评估或者决策行为的可靠性分析与 安全风险评估提供技术支持与工具。 显然, 本领域的技术人员应该明白, 上述的本发明的各模块或各步骤可以用通用 的计算装置来实现, 它们可以集中在单个的计算装置上, 或者分布在多个计算装置所 组成的网络上, 可选地, 它们可以用计算装置可执行的程序代码来实现, 从而可以将 它们存储在存储装置中由计算装置来执行,或者将它们分别制作成各个集成电路模块, 或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。 这样, 本发明不限 制于任何特定的硬件和软件结合。 以上所述仅为本发明的优选实施例而已, 并不用于限制本发明, 对于本领域的技 术人员来说, 本发明可以有各种更改和变化。 凡在本发明的精神和原则之内, 所作的 任何修改、 等同替换、 改进等, 均应包含在本发明的保护范围之内。

Claims

权 利 要 求 书
1. 一种操作员状态评估的可靠性分析方法, 包括:
确定待使用的多个行为形成因子 PSF, 其中, 所述 PSF用于状态评估; 确定所述多个 PSF中各个 PSF与除自身外的其他 PSF之间的关联关系以及 各个 PSF与状态评估可靠性节点的关联关系;
根据所述多个 PSF、 所述多个 PSF中各个 PSF之间的关联关系、 以及所述 各个 PSF与所述状态评估可靠性节点的关联关系建立基于 PSF因果关系的操作 员状态评估的可靠性分析模型以分析所述操作员的可靠性。
2. 根据权利要求 1所述的方法, 其中, 所述操作员状态评估的可靠性分析模型为 贝叶斯网络模型。
3. 根据权利要求 2所述的方法, 其中, 根据所述多个 PSF、 所述多个 PSF中各个 PSF之间的关联关系、 以及所述各个 PSF与所述状态评估可靠性节点的关联关 系建立基于 PSF因果关系的操作员状态评估的可靠性分析模型以分析所述操作 员的可靠性包括:
确定所述贝叶斯网络模型中各个根节点 PSF 处于不同状态的先验概率分 布, 其中, 所述根节点 PSF为没有被其他节点指向的 PSF节点;
根据所述各个根节点 PSF与除自身外的其他 PSF之间的关联关系、以及所 述各个 PSF与所述状态评估可靠性节点的关联关系确定各个子节点 PSF和所述 状态评估可靠性节点处于不同状态的条件概率分布, 其中, 所述子节点 PSF为 被其他节点指向的 PSF节点; 将所述条件概率分布与所述先验概率分布进行贝叶斯网络的因果分析以得 到特定情境条件下的状态评估可靠性。
4. 根据权利要求 2或 3所述的方法, 其中, 通过模糊方法对所述贝叶斯网络模型 的概率分布的值进行计算。
5. 根据权利要求 4所述的方法, 其中, 通过模糊方法对所述贝叶斯网络模型的概 率分布的值进行计算包括:
通过三角模糊数计算所述各个根节点 PSF 处于不同状态时的先验概率分 布; 通过所述三角模糊数计算所述各个子节点 PSF与所述状态评估可靠性节点 处于不同状态时的条件概率分布;
通过三角形重心解模糊的方法确定所述先验概率的值和所述条件概率的 值, 并将所述先验概率的值与所述条件概率的值进行贝叶斯网络的因果分析以 得到所述状态评估的可靠性。
6. 根据权利要求 3所述的方法, 其中, 根据所述多个 PSF、 所述多个 PSF中各个 PSF之间的关联关系、 以及所述各个 PSF与所述状态评估可靠性节点的关联关 系建立基于 PSF因果关系的操作员状态评估的可靠性分析模型以分析所述操作 员的可靠性之后, 还包括:
在确定操纵员状态评估失误的情况下, 根据所述各个根节点 PSF与除自身 外的其他 PSF之间的关联关系,以及所述各个 PSF与所述状态评估可靠性节点 的关联关系, 通过贝叶斯网络的诊断分析确定所述各个根节点 PSF的后验概率 分布;
将所述根节点 PSF处于预设多个状态中最差状态的后验概率分布与所述根 节点 PSF处于所述预设多个状态中最差状态的先验概率分布进行比较以得到影 响状态评估可靠性的关键要素, 以确定预防状态评估失误的对策。
7. 一种操作员状态评估的可靠性分析装置, 包括:
第一确定模块,用于确定待使用的多个行为形成因子 PSF,其中,所述 PSF 用于状态评估;
第二确定模块,用于确定所述多个 PSF中各个 PSF与除自身外的其他 PSF 之间的关联关系以及各个 PSF与状态评估可靠性节点的关联关系;
建立模块, 用于根据所述多个 PSF、 所述多个 PSF中各个 PSF之间的关联 关系、 以及所述各个 PSF与所述状态评估可靠性节点的关联关系建立基于 PSF 因果关系的操作员状态评估的可靠性分析模型以分析所述操作员的可靠性。
8. 根据权利要求 7所述的装置, 其中, 所述建立模块建立的所述操作员状态评估 的可靠性分析模型为贝叶斯网络模型。
9. 根据权利要求 8所述的装置, 其中, 所述建立模块包括:
第一确定单元, 用于确定所述贝叶斯网络模型中各个根节点 PSF处于不同 状态的先验概率分布, 其中, 所述根节点 PSF为没有被其他节点指向的 PSF节 点; 第二确定单元,用于根据所述各个根节点 PSF与除自身外的其他 PSF之间 的关联关系、 以及所述各个 PSF与所述状态评估可靠性节点的关联关系确定各 个子节点 PSF和所述状态评估可靠性节点处于不同状态的条件概率分布,其中, 所述子节点 PSF为被其他节点指向的 PSF节点;
分析单元, 用于将所述条件概率分布与所述先验概率分布进行贝叶斯网络 的因果分析以得到特定情境条件下状态评估可靠性。 根据权利要求 9所述的装置, 其中, 所述装置还包括: 第三确定模块, 用于在确定操纵员状态评估失误的情况下, 根据所述各个 根节点 PSF与除自身外的其他 PSF之间的关联关系,以及所述各个 PSF与状态 评估可靠性节点的关联关系, 通过贝叶斯网络的诊断分析确定所述各个根节点 PSF的后验概率分布;
比较模块, 用于将所述根节点 PSF处于预设多个状态中最差状态的后验概 率与所述根节点 PSF处于预设多个状态中最差状态的先验概率进行比较以得到 影响状态评估可靠性的关键要素, 以确定预防状态评估失误的对策。
PCT/CN2014/075729 2013-04-23 2014-04-18 操作员状态评估的可靠性分析方法及装置 WO2014173257A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201310142890.5 2013-04-23
CN201310142890.5A CN103218689B (zh) 2013-04-23 2013-04-23 操作员状态评估的可靠性分析方法及装置

Publications (1)

Publication Number Publication Date
WO2014173257A1 true WO2014173257A1 (zh) 2014-10-30

Family

ID=48816453

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/075729 WO2014173257A1 (zh) 2013-04-23 2014-04-18 操作员状态评估的可靠性分析方法及装置

Country Status (2)

Country Link
CN (1) CN103218689B (zh)
WO (1) WO2014173257A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3460733A4 (en) * 2017-01-31 2020-01-08 Mitsubishi Heavy Industries, Ltd. DESIGN PROCESS SUPPORT SYSTEM, CONTROL PROCEDURE AND PROGRAM

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103198438A (zh) * 2013-04-23 2013-07-10 湖南工学院 响应计划的可靠性分析方法及装置
CN103678886B (zh) * 2013-11-26 2017-01-04 北京空间飞行器总体设计部 一种基于地面测试数据的卫星贝叶斯网络健康确定方法
CN103970018A (zh) * 2014-05-07 2014-08-06 湖南工学院 获取人机交互复杂度数据的方法和装置
CN104392752B (zh) * 2014-10-13 2016-11-30 中国科学院合肥物质科学研究院 一种实时在线的核反应堆故障诊断与监测系统
CN104966158B (zh) * 2015-06-12 2021-11-16 中国核电工程有限公司 影响操作员不干预时间敏感事故的筛选方法
CN106411854B (zh) * 2016-09-06 2019-01-29 中国电子技术标准化研究院 一种基于模糊贝叶斯的网络安全风险评估方法
CN109284925A (zh) * 2018-09-21 2019-01-29 南华大学 一种班组情景意识的测量方法、装置、设备及存储介质
CN109447495A (zh) * 2018-11-08 2019-03-08 南华大学 核电厂组织行为可靠性评估方法、装置及设备
CN109636164B (zh) * 2018-11-27 2022-07-19 南华大学 一种用于工业系统的人因安全评估方法、装置及存储介质
CN112486940A (zh) * 2019-09-12 2021-03-12 伊姆西Ip控股有限责任公司 用于事件排序的方法、设备和计算机程序产品
CN113240688A (zh) * 2021-06-01 2021-08-10 安徽建筑大学 一种一体化洪涝灾害精准监测预警方法
CN114740774B (zh) * 2022-04-07 2022-09-27 青岛沃柏斯智能实验科技有限公司 一种通风柜安全操作的行为分析控制系统
CN115018334A (zh) * 2022-06-15 2022-09-06 湖南科技大学 一种基于模糊贝叶斯网络的瓦斯爆炸风险评估方法

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2015181A1 (en) * 2007-06-15 2009-01-14 Atomic Energy Council - Institute of Nuclear Energy Research Software safety analysis and evaluation system
CN103198438A (zh) * 2013-04-23 2013-07-10 湖南工学院 响应计划的可靠性分析方法及装置

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2015181A1 (en) * 2007-06-15 2009-01-14 Atomic Energy Council - Institute of Nuclear Energy Research Software safety analysis and evaluation system
CN103198438A (zh) * 2013-04-23 2013-07-10 湖南工学院 响应计划的可靠性分析方法及装置

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
DONG, CHEN: "Application and Research of the Bayesian Network in Power System Fault Diagnosis", CHINESE MASTER'S DISSERTATION FULL-TEXT DATABASE, 25 May 2010 (2010-05-25) *
YIN: "Xiaowei ''Reliability Evaluation of Mechanical System Based on Bayesian Network", CHINESE DOCTORAL DISSERTATION FULL-TEXT DATABASE, 30 November 2010 (2010-11-30) *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3460733A4 (en) * 2017-01-31 2020-01-08 Mitsubishi Heavy Industries, Ltd. DESIGN PROCESS SUPPORT SYSTEM, CONTROL PROCEDURE AND PROGRAM

Also Published As

Publication number Publication date
CN103218689A (zh) 2013-07-24
CN103218689B (zh) 2017-09-22

Similar Documents

Publication Publication Date Title
WO2014173257A1 (zh) 操作员状态评估的可靠性分析方法及装置
KR102118670B1 (ko) Ict 인프라 관리 시스템 및 이를 이용한 ict 인프라 관리 방법
CN111124840B (zh) 业务运维中告警的预测方法、装置与电子设备
Groth et al. A hybrid algorithm for developing third generation HRA methods using simulator data, causal models, and cognitive science
Akash et al. Human trust-based feedback control: Dynamically varying automation transparency to optimize human-machine interactions
WO2014173258A1 (zh) 响应计划的可靠性分析方法及装置
Cailliau et al. Runtime monitoring and resolution of probabilistic obstacles to system goals
Kim et al. An analytic model for situation assessment of nuclear power plant operators based on Bayesian inference
CN112069726A (zh) 一种基于贝叶斯网络的风险分析评估方法及装置
US11593648B2 (en) Methods and systems for detection and isolation of bias in predictive models
CN112379325A (zh) 一种用于智能电表的故障诊断方法及系统
Duan et al. Fault diagnosis for complex systems based on reliability analysis and sensors data considering epistemic uncertainty
US10545465B2 (en) System and method for selecting controllable parameters for equipment operation safety
Shirley et al. Development of a quantitative Bayesian network mapping objective factors to subjective performance shaping factor evaluations: An example using student operators in a digital nuclear power plant simulator
Lee et al. A computational model for evaluating the effects of attention, memory, and mental models on situation assessment of nuclear power plant operators
Zhao et al. CMS-BN: A cognitive modeling and simulation environment for human performance assessment, part 2—Application
Jensen et al. Reasoning about system-level failure behavior from large sets of function-based simulations
Boring et al. A research roadmap for computation-based human reliability analysis
CN113377595B (zh) 故障诊断方法、装置、电子设备和存储介质
CN116049642A (zh) 一种故障诊断方法、系统、电子设备及计算机存储介质
Santoso et al. Nuclear plant fault diagnosis using probabilistic reasoning
Perreault et al. Deriving prognostic continuous time Bayesian networks from D-matrices
US20210110287A1 (en) Causal Reasoning and Counterfactual Probabilistic Programming Framework Using Approximate Inference
US20230161637A1 (en) Automated reasoning for event management in cloud platforms
Baras et al. Integrated, distributed fault management for communication networks

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14787674

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 14787674

Country of ref document: EP

Kind code of ref document: A1