CN113963443B - Human error pattern recognition method, device, equipment and medium in nuclear power plant - Google Patents

Abstract

The invention discloses a method and a device for identifying a human error mode in a nuclear power plant, electronic equipment and a storage medium, wherein the method comprises the following steps: respectively carrying out hierarchical analysis and decomposition on a system function structure and an operation task in the system operation interaction process, and constructing an operation information base; according to the control information base, the target or the event/state is used as a guide to realize the operation navigation of the serialization action under the given task target; according to the control information base and the operation navigation, designing a personnel operation action mode recognition algorithm based on mode recognition; according to a personnel operation action mode recognition algorithm, the personnel error mode recognition in the nuclear power device is realized by combining real-time state monitoring data of the system. The invention realizes the identification of the typical human error modes in different operation scenes by detecting the consistency of the operation action modes and the equipment state modes in the control information base, thereby realizing early warning, preventing and reducing human error of operators and ensuring the operation safety of the marine nuclear power device.

Description

Human error pattern recognition method, device, equipment and medium in nuclear power plant

technical field

The invention belongs to the field of human factor reliability and nuclear safety analysis, in particular to a human error pattern recognition method, device, electronic equipment and storage medium in a nuclear power plant.

Background technique

Human factors play an important role in human-computer interaction systems, especially for complex systems such as nuclear power plants that rely heavily on human intervention by the operator, there is a complex dynamic interaction process between the automatic device and the operator's behavior, However, the unknown human-computer interaction process, especially human misoperation, will have an important impact on system failure. According to statistics, more than 70% of nuclear power accidents come from human error. Human error is used to describe the output or consequences of human actions, and refers to a series of human actions that exceed certain acceptability limits. It can be the actual operation action itself, or it can refer to human deliberate violations, which usually constitute accidents. main inducing reasons. Human error is often induced by a poor human-computer interaction design environment and manifested in the execution of system tasks.

The purpose of human reliability analysis is to identify potential human errors and evaluate the impact of human failure events on system risk. Human error identification is a basic component of human reliability analysis, and the rational division and definition of human actions and human error types is the basis. At present, there are many different methods to classify and model human error behavior patterns, such as THERP human error pattern classification based on the operator behavior model, Rasmussen human error pattern based on the rule/skill/knowledge (SRK) ladder model Pattern classification, CREAM human error pattern classification based on cause and effect, Reason human error pattern classification based on operational intent, SHERPA human error pattern classification based on cognitive process, and human failure event classification for PSA analysis references, etc.

In recent years, with the innovation and development of digital technology, the main control room of nuclear power plant is also fully digitalized. Different from the traditional human-computer interaction environment in the main control room, the operating mode, operating procedures, human-computer interaction behavior, and human error mode characteristics of nuclear power plants in the new digital control environment have undergone great changes. Because the analysis method fails to fully cover the consideration of the human error model under the transformation of the new digital environment, there is a potential new type of additional human error risk. In addition, although the automation, informatization, digitalization and intelligence of nuclear power plant systems have been significantly improved, the operating environment of marine nuclear power plant is extremely complex, the working conditions are changeable, and the structure, status and tasks of the system under different working conditions The purpose shows a strong stage and time-varying nature, making it still largely dependent on the operator's human intervention, coupled with the sharp deterioration of the control environment (high temperature and high humidity) and the long-term High-intensity operations cause a lot of cognitive and operational load pressure on the operator, which can easily lead to human error and endanger the safety of system operation. At the same time, the special combat mission of the marine nuclear power plant requires it to maintain its mobility in the face of extremely strange and emergency operating conditions. When the nuclear power plant has an unknown system abnormality or an accident beyond the design basis, the operator There are often no established operating procedures to follow. At this time, how to help operators respond to emergency operations in an environment without regulatory guidance is a major challenge for the intelligent operation and maintenance support technology of nuclear power plants at this stage.

Contents of the invention

In view of this, the present invention revolves around the nuclear power plant under the trend of new digital transformation with regulations (normal operating conditions, design basis accident conditions) and no regulations or regulations missing or incomplete (beyond design basis accident conditions or design extension conditions). conditions, extreme operating scenarios) scenarios, providing a human action pattern recognition method, device, electronic equipment, and storage medium in a nuclear power plant, aiming to achieve Typical human error pattern recognition and personnel operation action supervision under different operating scenarios (with or without regulations) to achieve early warning, prevent and reduce operator human error, and ensure the operation safety of marine nuclear power plants.

The first object of the present invention is to provide a method for recognizing a human action pattern in a nuclear power plant.

The second object of the present invention is to provide a personnel action pattern recognition device in a nuclear power plant.

The third object of the present invention is to provide an electronic device.

A fourth object of the present invention is to provide a storage medium.

First purpose of the present invention can be achieved by taking the following technical solutions:

A method for recognizing a human action pattern in a nuclear power plant, the method comprising:

Separately analyze and decompose the system function structure and operation tasks in the process of system operation interaction, and construct the operation information base;

According to the manipulation information base, guided by the target or event/state, realize the operation navigation of the sequenced action under the given task target;

According to the operation information base and the operation navigation, design a pattern recognition-based personnel operation action pattern recognition algorithm; wherein, the personnel operation action pattern recognition algorithm includes various personnel operation action patterns in the digital control environment of nuclear power plants The basis and method of discrimination;

According to the human operation action pattern recognition algorithm, combined with the real-time state monitoring data of the system, the human error pattern recognition in the nuclear power plant is realized.

Further, the manipulation information library includes a procedure action library and a device status information library, wherein:

The construction of the procedure action library takes the standard operating procedures as input, and uses the hierarchical task analysis method to analyze and decompose the procedure tasks, and obtains the hierarchical sequence structure of "manipulation task objectives-subtasks-basic manipulation actions"; by analyzing the manipulation actions Digital coding processing of sequence information to generate a routine action library;

Or the procedure action library is obtained by importing digital operation procedures;

The equipment state information base uses system design data as input, and uses the abstract thinking of "goal-means" and "part-whole" to analyze and distribute the system function structure, and obtains "system function goal-system function-means structure-basic control Components" hierarchical correlation function structure allocation, through the interactive coupling of system functions and operation and manipulation tasks, an interactive process relationship model of "system function goals-manipulation task goals-subtasks-basic manipulation actions-manipulation action objects" is formed to generate equipment State information base table.

Further, the construction process of the procedure action library specifically includes:

Taking standard operating procedures as input, and based on the procedure response analysis, the access conditions of procedures under different operating conditions and scenarios are extracted; among them, the procedures under different operating conditions and scenarios include diagnostic procedures, response procedures and system procedures;

Taking the standard operating procedure as input, based on hierarchical task analysis, extracting the manipulation task target, decomposing the manipulation action sequence according to the requirements of the manipulation task target, and obtaining the sequence structure of "manipulation task target-subtask-basic operation steps";

abstracting the basic operation steps in the serialized structure as device actions for device state changes, and establishing a coupling association between the operation actions and the manipulated objects;

According to the coupling association between the operation action and the manipulated object, the attribute characteristics of the operation action are extracted, including the manipulated object, the start/end time of the operation action, and the task execution time, and are coded according to the order of action execution. The data format of "manipulating task target-belonging subtask-operating action-task time" is stored in the database table to form a procedure action library;

The construction process of the device state information base specifically includes:

Based on the input of system design data, determine the system design purpose, function and structural composition, and generate a list of equipment that can control the system;

According to the discrete and continuous state attribute characteristics of the equipment, the equipment is divided into discrete controllable parts and continuous controllable parts, so as to facilitate the identification of human error modes of incomplete operation and operation delay;

Coding the equipment name for the equipment, the equipment name encoding can intuitively reflect the equipment association system and tasks/subtasks;

The attribute characteristics of the equipment are extracted, and entered into the equipment state information database table through data abstraction and digital coding in the data structure format of "equipment-attribute-attribute value", thereby generating the equipment state information database table, the "equipment-attribute-attribute value" in The device includes the device type and device name; the operation action and the operation object are associated through the device action. The device action refers to the operation action performed during the modification of the device attribute, and the attribute value refers to the current value of the device attribute in the actual operating environment. .

Further, according to the manipulation information base, guided by the target or event/state, the operation navigation of the sequenced action under the given task target is realized, specifically including:

For maneuvering task scenarios guided by procedures, depending on the type of procedure, choose event or state-oriented, and automatically extract the maneuvering action sequence oriented to abnormal events or abnormal state responses based on the self-identification of working conditions and the access conditions of the procedures, and use them for online Operation and navigation prompts; for non-online applications, the operator can set the operating conditions, system function goals and manipulation task goals according to the needs, and perform operation and navigation prompts for specific task goals;

For special operating scenarios and environments without regulations, guided by system function goals, through successful path planning or expected manipulation response configuration settings, constructive operation navigation guidance suggestions are provided; wherein, the successful path planning includes two methods:

Based on the functional model of the multi-layer flow system, combined with the causal relationship mapping and the influence propagation rules, the successful path set leading to the functional goal of the system is obtained through the reasoning search of the anti-goal "flow" structure;

Based on the GO-FLOW reliability model of the system, through the analysis of the minimum path set, according to the signal "flow" direction and logical structure relationship of physical existence, and combined with the engineering experience feedback, the system success path set oriented by the mission goal is obtained;

After determining the manipulation task target under the specific operating conditions and task requirements, the manipulation action sequence is automatically associated and presented according to the tree structure of "manipulation task target-sub-task target-basic operation steps", and the manipulation is prompted according to the feedback of the operation supervision results The current task progress of the staff and the subsequent operations that need to be performed.

Further, the pattern recognition starts from the needs of the system to run tasks, and is judged by the access conditions of the procedures;

According to the personnel operation action pattern recognition algorithm, combined with the real-time state monitoring data of the system, the human error pattern recognition in the nuclear power plant is realized, specifically including:

Judging whether the access conditions of the regulations are met, if not, continue to monitor system operation; if so, then:

Extract the corresponding manipulation action sequence and project it to the operation navigation screen;

According to the operation navigation guide, record the execution time of the current action step, which is used for subsequent judgment whether the operation action is completed within the specified task time;

According to the system operating status parameters provided by real-time status monitoring and the actual operation instructions input by the operator, identify the mode of human error;

Display the identification results of the human error mode on the operation supervision interface;

Checks if the current action is the last step in the sequence of actions for the given task manipulation, and if so:

Check whether there are unexecuted actions in the task manipulation action sequence. If so, it is "operation omission" human error mode recognition, and the current recognition result of human error mode will be displayed on the operation supervision interface; if not, according to Navigation information to enter the next task;

If not, jump to judging whether the access conditions of the procedure are met, and continue to perform subsequent operations.

Further, the identification of the human error mode according to the system operating state parameters provided by the real-time state monitoring and the actual operation instructions input by the operator specifically includes:

The system operating state parameters and actual operation instructions are read into the equipment state information database through the external data interface, and the read data is compared with the default parameter settings in the equipment state information database. The operator performs an operation action on the current device, records the device whose parameters have changed, the parameter attribute and the parameter change value and integrates it into an operation action information, triggers the operation action event, and judges whether the operation object is based on the current monitoring operation event status mode matches, and if so:

Determine whether the equipment control mode and/or operation action match, if not, it is "wrong operation acts on the correct object" human error mode recognition; if so, jump to S1;

If not, then:

Judging whether the equipment control mode and/or action mode match, if so, it is "target selection error" human error mode recognition; if not, traverse the action library of the procedure to determine whether the operation action is covered in the action library of the procedure, if so, skip Go to S1; if not, it is "operation without procedures" human error pattern recognition;

S1:

According to the standard operation action sequence defined in the standard procedure action library, judge whether the execution sequence of the operation action matches, if so, then:

Combined with the execution time record of the current action step, judge whether the operation action is completed within the specified task time, if so, it is "correct operation" human error pattern recognition; if not, then: judge whether the manipulated object is a discrete variable, if so, If it is "delayed operation" human error pattern recognition; if not, it is "incomplete operation" human error pattern recognition;

If not, then:

Judging whether the operation action is in advance, if so, it is "operation ahead" human error pattern recognition; if not, it is "operation lag" human error pattern recognition.

Further, the display of the recognition result of the human error mode on the operation supervision interface specifically includes:

Feedback the executed operation action information and the recognition result of the corresponding human error mode to the operation navigation window, so that the operator can understand the current operation task process more conveniently;

The display of the operation supervision interface mainly includes the backtracking of the sequence of historical manipulation actions and the recognition results of the current manipulation actions. According to the recognition results of the current manipulation actions, combined with the warning information pop-up window, the potential human errors that may occur during the task execution process of the operator can be detected. It should be reminded that the traceback of the historical manipulation sequence includes the recognition results of the manipulation actions actually input by the operator in the past and the human error modes corresponding to the manipulation actions.

The second purpose of the present invention can be achieved by taking the following technical solutions:

A personnel action pattern recognition device in a nuclear power plant, said device comprising:

The control information base construction module is used to analyze and decompose the system function structure and the control tasks in the process of system operation interaction in a hierarchical manner, and construct the control information base;

An operation navigation implementation module, configured to realize the operation navigation of the serialized actions under the given task target according to the operation information base, guided by the target or event/state;

The personnel operation action pattern recognition algorithm design module is used to design a personnel operation action pattern recognition algorithm based on pattern recognition according to the operation information database and the operation navigation; wherein, the personnel operation action pattern recognition algorithm includes a variety of kernels Discrimination basis and method of personnel operation mode in the environment of digital control of power plant;

The human error pattern recognition module is used to realize the human error pattern recognition in the nuclear power plant according to the personnel operation action pattern recognition algorithm combined with the real-time status monitoring data of the system.

The third purpose of the present invention can be achieved by taking the following technical solutions:

An electronic device includes a processor and a memory for storing executable programs of the processor. When the processor executes the program stored in the memory, the above-mentioned human error pattern recognition method in a nuclear power plant is realized.

The fourth purpose of the present invention can be achieved by taking the following technical solutions:

A storage medium stores a program, and when the program is executed by a processor, the above-mentioned human error pattern recognition method in a nuclear power plant is realized.

Compared with the prior art, the present invention has the following beneficial effects:

1. The method provided by the present invention considers the new human-computer interaction behavior mode and human error mode under the digital control environment of nuclear power plants, and can realize Detection of human error patterns in various digital manipulation environments.

2. The method provided by the present invention is task-oriented and realizes the strong correlation coupling between the system function target and the operation and operation process. Tracing historical operation actions, monitoring the current task operation process, and guiding the expected implementation of the operation plan, combined with the evaluation of operational applicability and the identification of personnel operation actions, real-time operation warning can be achieved, thereby escorting the safe operation of nuclear power plants.

3. The successful path planning method provided by the present invention is guided by the success of the task, and can break through the safety guidelines for the unregulated operation of nuclear power plants in extremely unfamiliar and complex operating environments.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained according to the structures shown in these drawings without creative effort.

Fig. 1 is a flow chart of a human error pattern recognition method in a nuclear power plant according to Embodiment 1 of the present invention.

Fig. 2 is a schematic diagram of the principle of a human error pattern recognition method in a nuclear power plant according to Embodiment 1 of the present invention.

Fig. 3 is a flowchart of the reactor boron and water replenishment system in Embodiment 1 of the present invention.

FIG. 4 is a virtual simulation platform of Embodiment 1 of the present invention.

Fig. 5 is a flow chart of implementing human error pattern recognition in a nuclear power plant according to Embodiment 1 of the present invention.

Fig. 6 is a structural block diagram of a human error pattern recognition device in a nuclear power plant according to Embodiment 2 of the present invention.

FIG. 7 is a structural block diagram of an electronic device according to Embodiment 3 of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work all belong to the protection scope of the present invention . It should be understood that the specific embodiments described are only used to explain the present application and are not intended to limit the present application.

The present invention combines the manipulative behavior characteristics of personnel tasks in the digital control environment, and considers that in the system running without rules, the operator's operation behavior may not be determined solely based on whether the operator's actions follow the rules like the rules guide. Clearly distinguish whether it is a misoperation, and there is no established standard for judging the correctness of personnel actions in an environment without regulations. It is more based on the operator's behavior purpose and consequences, and makes applicability and safety evaluations, that is, from The attainability of system safety function goals and manipulation task goals is evaluated and divided into operator behavior types in many ways.

The present invention considers the following modes of action, specifically including:

Correct operation: It means that the operator successfully completes the prescribed operation action without error under the prescribed task time and under the prescribed conditions according to the designated operating procedures or successful path guidance.

Non-procedural operation: It has nothing to do with the established prescribed actions, and generally refers to the operator's actions beyond the scope of the prescribed action library.

Misoperation: generally refers to a human error in a certain behavior, or refers to the operator's action obviously deviates from the predetermined, required or expected standard and fails to achieve the expected effect.

For misoperation, the present invention comprehensively refers to the personnel action classification system proposed by Reason, Swain and Guttman et al., and divides personnel unsafe actions (Unsafe Actions) into the following two categories:

Error of Omission (EOO): An action is required but not implemented. It can also refer to human error caused by failure to meet system requirements or lack of time. Operation omission can be deliberate or unintentional.

Error of Commission (EOC): Refers to an action that is not related to the requirement but has been executed, so that the expected required state cannot be achieved.

Specific human error modes considered include:

Operation omission (E1): The operator omits a necessary operation intentionally or unintentionally during the execution of the operation task;

Target selection error (E2): The operator selects the wrong object target during the execution of the operation action;

Operation Execution Error (E3): The operator performed an incorrect operation on the correct target object;

Incomplete operation (E4): The operator's operation is not sufficient, for example, the valve opening is not opened or closed to the specified position;

Operational delay (E5): The operator fails to complete the operation within the specified time for the action task, for example, the valve is not opened or closed within the specified time;

Operation in advance: (E6): The operation action is completed earlier than the prescribed steps;

Operation lag (E7): The operation action is completed later than the prescribed steps.

The above types of human error are only used as case illustrations, and the modes and types of human error in actual scenarios may far exceed the scope of the definitions given above.

Example 1:

As shown in Figure 1 and Figure 2, a human error pattern recognition method in a nuclear power plant specifically includes the following steps:

S101. Construct a manipulation information base, where the manipulation information base includes a device status information base and a procedure action base.

Through the hierarchical analysis and decomposition of the system's functional structure and procedural tasks, a system equipment state information base and a procedural action base are constructed.

As shown in Figure 3, this embodiment takes the operation of the boron and water supply system of the reactor as an example, and describes the construction process of the control information database in combination with a hypothetical case scenario. The specific construction process is as follows:

S1011. Construct a device status information database.

The equipment state information base uses system design data as input, and uses the abstract thinking of "goal-means" and "part-whole" to analyze and distribute the system function structure, and obtains "system function goal-system function-means structure-basic control Components" are assigned hierarchically related functional structures, and through the interactive coupling of system functions and operation and manipulation tasks, an interactive process relationship model of "system function goals-manipulation task goals-subtasks-basic manipulation actions-manipulation action objects" is formed.

Through the hierarchical analysis and decomposition of the system function structure, the system equipment status information database is constructed. The specific steps include:

(1) Based on the input system design data, realize the system design purpose and function decomposition.

The reactor boron and water supply system is an important auxiliary support system for the primary circuit, which mainly realizes the following functions:

Reactivity control (FG1): Provide boric acid solution and desalted and deoxygenated water to the primary loop system to assist in the control of boric acid concentration;

Volume control (FG2): Provide desalted and deoxygenated water to the primary circuit system to assist in realizing the volume fluctuation of the reactor coolant in the primary circuit;

Chemical control (FG3): Configure and inject chemicals such as hydrazine and lithium hydroxide through the chemical addition box to assist in the control of oxygen content and PH value in the primary loop reactor coolant.

(2) According to the functional structure diagram of the system, decompose the functional structure of the system.

Specifically include:

Boric acid injection pipeline: It consists of two redundant boric acid tanks, two boric acid transport pumps and valves on the branch pipeline. During normal operation, concentrated boric acid solution is delivered by one of the boric acid tanks and one boric acid delivery pump, and the boric acid tank and boric acid delivery pump on the other branch pipeline are in standby state;

Desalination and deoxygenation water injection pipeline: It consists of two desalination and deoxygenation water tanks with redundant structure, two desalination and deoxygenation water pumps and their branch pipelines and valves. During normal operation, water is supplied by one of the desalination and deoxygenation water tanks and one desalination and deoxygenation water pump, and the desalination and deoxygenation water tank and desalination and deoxygenation water pump on the other branch pipeline are in standby state;

Chemical addition pipeline: consists of a chemical addition tank and valves on a single branch pipeline. During normal operation, hydrazine and lithium hydroxide solutions are added through the chemical addition tank, and the chemical solvent is sent to the primary circuit of the reactor through the topping pipeline.

(3) According to the functional structure of the system, list the equipment composition of the system and classify the equipment.

The system specifically includes the following components:

(3-1) Controllable components: refer to equipment that can be operated and controlled, and its control variables involve discrete switch states, continuous switch states, and continuous process variables.

(3-1-1) Discrete switchgear, specifically including:

Manual control valve: REA001HOV, REA002HOV, REA003HOV, REA004HOV, REA005HOV, REA006HOV, REA007HOV, REA008HOV, REA009HOV, REA010HOV, REA011HO;

Pneumatic control valve: REA001PCV, REA002PCV, electric control valve REA002MOV;

Desalination and deoxygenation water pump: REA001PO, REA002PO;

Boric acid transfer pump: REA003PO, REA004PO.

(3-1-2) Continuous switching state equipment, specifically including:

Pneumatic isolation valves: REA001PIV, REA002PIV.

(3-1-3) Continuous process variable control equipment, specifically including:

Desalination and deoxygenation water tank: REA001BA, REA002BA;

Boric acid tank: REA003BA, REA004BA;

Chemical addition box: REA005BA;

Ingredients box: REA006BA;

Volume control box: RCV001BA.

(3-2) Uncontrollable components: Refers to equipment that cannot be operated and controlled, and its equipment status is not affected by external operations.

Includes check valve REA00CV.

(3-3) According to the equipment type, extract the attribute characteristics of the equipment, specifically including:

Control mode: automatic control and manual control;

Action mode: open action and close action;

State mode: open state, closed state, valve opening value and process variable value;

Action execution time: time requirement.

(4) Based on the above system function structure decomposition and distribution results, the basic attribute characteristics of the obtained system equipment are extracted, sorted by category, and the equipment status information database table is generated. As shown in Table 1, the equipment status information database table fields include equipment class, Subclass, device name, device number, device action, device attribute (control mode, action mode, state mode), device control variable attribute value, etc.

Table 1 Device status information database table

S1012, constructing a procedure action library.

The construction of the procedure action library takes the standard operating procedure as input, analyzes and decomposes the procedure task by applying the hierarchical task analysis method, and obtains the hierarchical sequence structure of "manipulation task target-subtask-basic operation action"; through the sequence information of the operation action Digital code processing to generate a routine action library. The procedure action library can also be obtained by importing digital operation procedures.

Introduce conditional analysis and hierarchical task decomposition through the procedure, extract the manipulation action sequence oriented by the manipulation task goal, build the procedure action library, and use it as a standard reference basis for the consistency detection and identification of personnel operation action patterns. The specific steps include:

(1) According to the operating characteristics of the reactor boron and water supply system, the system function task decomposition is realized, including:

Boron dilution (MG1): replace the reactor coolant in the primary circuit with an equal amount of desalted and deoxygenated water to reduce the concentration of boric acid in the primary circuit coolant, and stop dilution when the replenished desalted and deoxygenated water reaches the set value ; At least one desalination and deoxygenation water supply line should be put into operation to ensure the successful completion of the boron dilution task.

Boronation (MG2): replace the reactor coolant in the primary circuit with an equal amount of concentrated boric acid solution to increase the concentration of boric acid in the primary circuit coolant, and stop borylation when the supplied boric acid solution reaches the set value; at least A series of boric acid solution supply lines to ensure the successful completion of the boronation task.

Automatic boron replenishment (MG3): supply boron-containing water with the same concentration as the primary reactor coolant for volume control without changing the concentration of the primary reactor coolant; during the replenishment process, keep the flow of desalinated and oxygenated water constant, The flow rate of boric acid is given by calculation. Boric acid and desalted and deoxygenated water are mixed in the mixing channel and then injected into the charging pipeline to complete automatic boron replenishment; Successful completion of the boron resupply mission.

Manual boron replenishment (MG4): used for the initial water filling or replenishment of the refueling water tank and the water filling and exhausting of the container control box, replenishing boron-containing water with the same concentration as the coolant of the primary loop reactor, used for volume control, without changing the primary loop reactor The concentration of the coolant; during the replenishment process, the flow rate of desalination and deoxygenation water and boric acid is manually set by the operator, and the replenishment is stopped manually when the replenishment amount reaches the set value; at least one row of desalination and deoxygenation water injection pipelines and one row of boric acid should be put in Injection lines to ensure successful completion of automated boron replenishment missions.

Addition of chemicals (MG5): Inject desalination and oxygenation water into the pipeline of the chemical addition tank through a desalination and deoxygenation water pump, and add hydrazine or lithium hydroxide solution in the chemical addition tank to change the primary reactor coolant Content and PH value to achieve the purpose of chemical control; during the chemical addition process, at least one line of desalination and deoxygenation water supply pipelines and chemical addition tank pipelines should be invested to ensure the successful completion of the chemical addition task.

(2) Taking manual boron supply (MG4) as the target of the manipulation task, through a hypothetical scene without procedures, successful path planning, example hierarchical task decomposition and manipulation action sequence extraction process, the manipulation action sequence is shown in Table 2.

The manual boron replenishment task of the reactor boron and water replenishment system is divided into five subtasks, including:

Open the isolation valve: open the pneumatic isolation valve REA001PIV and REA002PIV (Step-1), which is a combination of disordered actions, which can be simply understood as an operating procedure at the functional level.

Switch the pump from automatic to manual control mode: switch the desalination and deoxygenation water pump REA002PO and boric acid delivery pump REA002PO from automatic to manual control mode (Step-2), which is also a combination of disordered actions.

Turn on the desalination and oxygen removal pump and the boric acid delivery pump: turn on the boric acid delivery pump REA004PO (Step-3) and the desalination and oxygen removal pump REA002PO (Step-4).

Start boron supply: adjust the opening of the pneumatic control valve REA001PCV to 80% (Step-5), adjust the opening of the pneumatic control valve REA002PCV to 60% (Step-6), and open the electric control valve REA002MOV (Step-7 ).

Stop boron supply: turn off the pneumatic control valve REA001PCV (Step-8), close the pneumatic control valve REA002PCV (Step-9), close the electric control valve REA002MOV (Step-10), close the boric acid delivery pump REA004PO (Step-11), close Desalination and deoxygenation water pump REA002PO (Step-12), close the pneumatic isolation valve REA001PIV (Step-13), and close the pneumatic isolation valve REA002PIV (Step-14).

Table 2 Manipulation Action Sequence List

(3) Based on the results of hierarchical task decomposition, the basic operation steps under the tree structure of "manipulation task objectives-subtasks-basic operation steps" are organized in the form of serialization or function combination to form a routine action library to realize various The automatic association between the level task goal-subtask goal-basic operation action.

The information list fields of the procedure action library include: working condition scenario, task target, subtask target, operation action step sequence code, operation action mode, manipulated object association code, control variable minimum boundary value, control variable maximum boundary value and action execution time etc.

S102. The designer operates an action pattern recognition algorithm.

Based on the state pattern recognition technology, according to the operation information base and operation navigation, design the personnel operation action pattern recognition algorithm; wherein, the personnel operation action pattern recognition algorithm includes the discrimination of personnel operation action patterns in the digital control environment of various nuclear power plants Basis and method for human error pattern recognition in digital control environment of nuclear power plant.

S103. The real-time status monitoring data of the system, including system running status implementation monitoring input and actual operation action instruction input.

Based on the QT virtual simulation platform, the generation, monitoring and reading of the process state parameters of the simulation system and the manual operation command signals of the operator are carried out.

As shown in Figure 4, through the QT Designer virtual simulation environment, the human-computer interaction and the generation, monitoring and reading of system operating state parameters and operator operation instructions during the manual boron supply process under refueling and shutdown conditions are simulated. Simulation data includes:

Real-time monitoring input of system operating status: including system equipment status, control mode, process variable monitoring value, etc.; for example, the boric acid delivery pump REA004PO is in manual control mode and is on, and the concentration of boric acid is 1300ppm.

Actual operation command input: manual open signal, manual close signal, etc.

S104. According to the human operation action pattern recognition algorithm, combined with the real-time status monitoring data of the system, the human error pattern recognition in the nuclear power plant is realized.

As shown in Figure 5, based on the real-time simulation monitoring data generated by the QT virtual simulation platform, and according to the human operation action pattern recognition algorithm, the human error pattern recognition in the nuclear power plant is realized. The specific steps include:

Operation navigation guidance: Based on the protocol action library, through the selection of the manipulation task target, the manipulation action sequence associated with the "manual boron supply" task will be presented in a tree-structured navigation bar.

Simulation of system operation and manipulation process: Based on the QT virtual simulation platform, simulate and generate system equipment status control signals and system process status monitoring parameters, specifically including control mode and status mode digital signals of controllable components, volume control tank water level, reactor coolant Continuous process variable monitoring values such as oxygen content, boric acid concentration, and pH value.

Real-time monitoring data reading: through the external data interface, read the real-time monitoring signal generated by the QT virtual simulation platform or simulator into the control information database, update the device control variable attribute value in the device status information database, and Changes automatically trigger associated action events;

Consistency detection of operation action: compare and analyze the actual operation action event triggered by the state change with the expected operation action in the action library;

Personnel action pattern recognition: Through consistency detection, the recognition of personnel operation action patterns and potential human error patterns is realized, as shown in Table 3. The described embodiment takes into account the typical action patterns and human errors that may occur during the execution of tasks by the operator, specifically including correct operation, irregular operation, incomplete operation, operation delay, operation advance, operation lag, and target selection. Errors, wrong execution of operations, omissions of operations, etc.

Table 3. Recognition results of human operation action patterns and potential human error patterns

Operation monitoring interface display: According to the recognition results of personnel action patterns, combined with the warning information pop-up window, the operator will be prompted for potential human errors that may occur during the task execution process, and the information of the executed operation actions will be fed back to the operation navigation window. Operators can more easily understand the current operation process.

The human action pattern recognition algorithm and examples in the digital control environment of nuclear power plant provided by the present invention are not limited to the above typical human action type recognition. Improve the basis for the identification of personnel action types to achieve a wider range of human error pattern recognition and coverage.

Those skilled in the art can understand that all or part of the steps in the method for implementing the above embodiments can be completed by instructing related hardware through a program, and the corresponding program can be stored in a computer-readable storage medium.

It should be noted that although method operations of the above-described embodiments are described in a particular order in the drawings, this does not require or imply that the operations must be performed in that particular order, or that all illustrated operations must be performed to achieve the desired results. . Rather, the depicted steps may be performed in an altered order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step for execution, and/or one step may be decomposed into multiple steps for execution.

Example 2:

As shown in Figure 6, this embodiment provides a human error pattern recognition device in a power plant, which includes a control information database construction module 601, an operation navigation realization module 602, a personnel operation action pattern recognition algorithm design module 603 and a human error pattern recognition device. Due to error pattern recognition module 604, wherein:

The manipulation information base construction module 601 is used for hierarchically analyzing and decomposing the system function structure and manipulation tasks in the system operation interaction process respectively, and constructing the manipulation information base;

An operation navigation realization module 602, configured to realize the operation navigation of the sequenced actions under the given task target according to the manipulation information base and guided by the target or event/state;

The personnel operation action pattern recognition algorithm design module 603 is used to design a personnel operation action pattern recognition algorithm based on pattern recognition according to the manipulation information base and the operation navigation; wherein, the personnel operation action pattern recognition algorithm includes a variety of Discrimination basis and method for personnel operation mode in digital control environment of nuclear power plant;

The human error pattern recognition module 604 is used to realize the human error pattern recognition in the nuclear power plant according to the human error pattern recognition algorithm combined with the real-time status monitoring data of the system;

Among them, the human error pattern recognition module 604 includes an operation supervision interface display unit, which is used to display the interface display of the human error pattern recognition results, and timely remind the operator of potential or ongoing possible harmful actions through the design of early warning information pop-up windows , improve its safety and situation awareness, and feed back the information of the executed operation actions to the operation navigation window, so that the operator can more easily understand the progress of the current operation task.

For the specific implementation of each module in this embodiment, please refer to the above-mentioned embodiment 1, and will not go into details one by one here; In , the above-mentioned function allocation can be completed by different functional modules according to needs, that is, the internal structure is divided into different functional modules, so as to complete all or part of the functions described above.

Example 3:

This embodiment provides an electronic device, which can be a computer. As shown in FIG. To provide computing and control capabilities, the memory includes a non-volatile storage medium 706 and an internal memory 707, the non-volatile storage medium 706 stores an operating system, computer programs and databases, and the internal memory 707 is a non-volatile storage The operating system in the medium and the operation of the computer program provide an environment. When the processor 702 executes the computer program stored in the memory, the human error pattern recognition method in the nuclear power plant of the above-mentioned embodiment 1 is realized, as follows:

Separately analyze and decompose the system function structure and operation tasks in the process of system operation interaction, and construct the operation information base;

According to the manipulation information base, guided by the target or event/state, realize the operation navigation of the sequenced action under the given task target;

According to the operation information base and the operation navigation, design a pattern recognition-based personnel operation action pattern recognition algorithm; wherein, the personnel operation action pattern recognition algorithm includes various personnel operation action patterns in the digital control environment of nuclear power plants The basis and method of discrimination;

According to the human operation action pattern recognition algorithm, combined with the real-time state monitoring data of the system, the human error pattern recognition in the nuclear power plant is realized.

Example 4:

This embodiment provides a storage medium, which is a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the human error mode in the nuclear power plant of the above-mentioned embodiment 1 is realized. The identification method is as follows:

Separately analyze and decompose the system function structure and operation tasks in the process of system operation interaction, and construct the operation information base;

According to the manipulation information base, guided by the target or event/state, realize the operation navigation of the sequenced action under the given task target;

According to the operation information base and the operation navigation, design a pattern recognition-based personnel operation action pattern recognition algorithm; wherein, the personnel operation action pattern recognition algorithm includes various personnel operation action patterns in the digital control environment of nuclear power plants The basis and method of discrimination;

According to the human operation action pattern recognition algorithm, combined with the real-time state monitoring data of the system, the human error pattern recognition in the nuclear power plant is realized.

It should be noted that the computer-readable storage medium in this embodiment may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the above two. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, electrical connections with one or more wires, portable computer diskettes, hard disks, 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 of the above.

In summary, the method for human error pattern recognition in nuclear power plants provided by the present invention revolves around system function goals and manipulation task goals, based on system function structure analysis and hierarchical task analysis, and by analyzing the status of controllable system equipment The extraction of attribute characteristics and manipulation action sequence, construction of manipulation information base, and design of personnel manipulation movement pattern recognition algorithm based on this, realizes effective recognition of human error patterns in the new digital manipulation environment, and provides operational navigation and supervision The method can realize navigation prompts and real-time control task process supervision for the control action sequence under the established system function target or control task target, timely warning of potential human errors, especially the successful path planning provided in extremely unfamiliar emergency operation scenarios and environments The method can effectively assist and guide the operator to successfully bring the nuclear power plant back to a safe state, thereby providing escort for the safe operation of the nuclear power plant.

The above is only a preferred embodiment of the patent of the present invention, but the scope of protection of the patent of the present invention is not limited thereto. Equivalent replacements or changes to the technical solutions and their inventive concepts all fall within the scope of protection of the invention patent.

Claims (7)

1. A human error pattern recognition method in a nuclear power plant, characterized in that the method comprises: Respectively analyze and decompose the system function structure and operation tasks in the process of system interaction in a hierarchical manner, and construct the operation information library; the operation information library includes two parts: the procedure action library and the equipment status information library, in which the procedure action library is imported through digital procedures Or based on the hierarchical task analysis of the standardized operating procedures, the equipment status information base is obtained through the decomposition and construction of the hierarchical functional structure; through the interactive coupling of system functions and operation and manipulation tasks, a "system function target-operation task target-subtask -Basic manipulation action-manipulation action object”interaction process relationship model; According to the manipulation information base, guided by the target or event/state, realize the operation navigation of the sequenced action under the given task target; According to the operation information base and the operation navigation, design a pattern recognition-based personnel operation action pattern recognition algorithm; wherein, the personnel operation action pattern recognition algorithm includes various personnel operation action patterns in the digital control environment of nuclear power plants The basis and method of discrimination; According to the human operation action pattern recognition algorithm, combined with the real-time status monitoring data of the system, if the access conditions of the regulations are met, the corresponding manipulation action sequence is extracted and projected to the operation navigation screen; according to the system operating status parameters provided by the real-time status monitoring Realize the recognition of human error patterns in nuclear power plants with the actual operation instructions input by the operator, and display the recognition results of human error patterns on the operation monitoring interface; According to the system operating state parameters provided by real-time state monitoring and the actual operation instructions input by the operator, the human error pattern recognition in the nuclear power plant is realized, specifically including: Read the system operating state parameters and actual operation instructions into the equipment state information database, compare and analyze the read data with the default parameter settings in the equipment state information database, and if it detects that the data changes, it is considered that the operator has changed the current The device performs an operation action, records the device whose parameters have changed, parameter attributes, and parameter change values and integrates it into an operation action information, triggers an operation action event, and judges whether the operation object matches based on the current monitoring operation event status mode. , then: determine whether the equipment control mode and/or operation action match, if not, it is "wrong operation acts on the correct object" human error mode recognition; if so, jump to S1; If not, then: judge whether the equipment control mode and/or operation action match, if yes, it is "target selection error" human error pattern recognition; if not: traverse the procedure action library, and judge whether the operation action is covered in the procedure action library , if yes, jump to S1; if not, it is "operation without procedures" human error pattern recognition; S1: According to the standard operation action sequence defined in the standard procedure action library, judge whether the execution sequence of the operation action matches. If so, then: combined with the execution time record of the current action step, determine whether the operation action is completed within the specified task time. If yes, then: "Correct operation" human error pattern recognition; if not, then: judge whether the manipulated object is a discrete variable, if so, then "operation delay" human error pattern recognition; if not, then "operation incomplete" human error pattern recognition; If not, then: judge whether the operation action is in advance, if so, it is "operation ahead" human error pattern recognition; if not, it is "operation lag" human error pattern recognition. 2. The human error pattern recognition method in the nuclear power plant according to claim 1, wherein the construction process of the procedure action library specifically includes: Taking standard operating procedures as input, and based on the procedure response analysis, the access conditions of procedures under different operating conditions and scenarios are extracted; among them, the procedures under different operating conditions and scenarios include diagnostic procedures, response procedures and system procedures; Taking the standard operating procedure as input, based on hierarchical task analysis, extracting the manipulation task target, decomposing the manipulation action sequence according to the requirements of the manipulation task target, and obtaining the sequence structure of "manipulation task target-subtask-basic operation steps"; abstracting the basic operation steps in the serialized structure as device actions for device state changes, and establishing a coupling association between the operation actions and the manipulated objects; According to the coupling association between the operation action and the manipulated object, the attribute characteristics of the operation action are extracted, including the manipulated object, the start/end time of the operation action, and the task execution time, and are coded according to the order of action execution. The data format of "manipulating task target-belonging subtask-operating action-task time" is stored in the database table to form a procedure action library; The construction process of the device state information base specifically includes: Based on the input of system design data, determine the system design purpose, function and structural composition, and generate a list of equipment that can control the system; According to the discrete and continuous state attribute characteristics of the equipment, the equipment is divided into discrete controllable parts and continuous controllable parts, so as to facilitate the identification of human error modes of incomplete operation and operation delay; Coding the equipment name for the equipment, the equipment name encoding can intuitively reflect the equipment association system and tasks/subtasks; The attribute characteristics of the equipment are extracted, and entered into the equipment state information database table through data abstraction and digital coding in the data structure format of "equipment-attribute-attribute value", thereby generating the equipment state information database table; wherein, the "equipment-attribute-attribute The device in Value" includes the device type and device name. 3. The human error pattern recognition method in a nuclear power plant according to claim 1, characterized in that, according to the control information base, the target or event/state is oriented to realize the given task target. Operation navigation for serialized actions, including: For maneuvering task scenarios guided by procedures, depending on the type of procedure, choose event or state-oriented, and automatically extract the maneuvering action sequence oriented to abnormal events or abnormal state responses based on the self-identification of working conditions and the access conditions of the procedures, and use them for online Operation and navigation prompts; for non-online applications, the operator can set the operating conditions, system function goals and manipulation task goals according to the needs, and perform operation and navigation prompts for specific task goals; For special operating scenarios and environments without regulations, guided by system function goals, through successful path planning or expected manipulation response configuration settings, constructive operation navigation guidance suggestions are provided; wherein, the successful path planning includes two methods: Based on the functional model of the multi-layer flow system, combined with the causal relationship mapping and the influence propagation rules, the successful path set leading to the functional goal of the system is obtained through the reasoning search of the anti-goal "flow" structure; Based on the GO-FLOW reliability model of the system, through the analysis of the minimum path set, according to the signal "flow" direction and logical structure relationship of physical existence, and combined with the engineering experience feedback, the system success path set oriented by the mission goal is obtained; After determining the manipulation task target under the specific operating conditions and task requirements, the manipulation action sequence is automatically associated and presented according to the tree structure of "manipulation task target-sub-task target-basic operation steps", and the manipulation is prompted according to the feedback of the operation supervision results The current task progress of the staff and the subsequent operations that need to be performed. 4. The human action pattern recognition method in nuclear power according to claim 1, wherein the recognition result of the human error pattern is displayed on the operation supervision interface, specifically comprising: Feedback the executed operation action information and the recognition result of the corresponding human error mode to the operation navigation window, so that the operator can understand the current operation task process more conveniently; The display of the operation supervision interface mainly includes the backtracking of the sequence of historical manipulation actions and the recognition results of the current manipulation actions. According to the recognition results of the current manipulation actions, combined with the warning information pop-up window, the potential human errors that may occur during the task execution process of the operator can be detected. It should be reminded that the traceback of the historical manipulation sequence includes the recognition results of the manipulation actions actually input by the operator in the past and the human error modes corresponding to the manipulation actions. 5. A human error pattern recognition device in a nuclear power plant, characterized in that the device comprises: The control information base building module is used to perform hierarchical analysis and decomposition on the system function structure and manipulation tasks during the system operation interaction process, and construct the control information base; the control information base includes two parts: the procedure action base and the equipment status information base, Among them, the procedure action library is obtained by importing digital procedures or based on hierarchical task analysis of standardized operating procedures, and the equipment status information library is obtained by decomposing and constructing hierarchical functional structures; through the interactive coupling of system functions and operation and manipulation tasks, a "system function Target-manipulation task target-subtask-basic manipulation action-manipulation action object” interaction process relationship model; An operation navigation implementation module, configured to realize the operation navigation of the serialized actions under the given task target according to the operation information base, guided by the target or event/state; The personnel operation action pattern recognition algorithm design module is used to design a personnel operation action pattern recognition algorithm based on pattern recognition according to the operation information database and the operation navigation; wherein, the personnel operation action pattern recognition algorithm includes a variety of kernels Discrimination basis and method of personnel operation mode in the environment of digital control of power plant; The human error pattern recognition module is used to extract the corresponding manipulation action sequence and project it to the operation navigation screen according to the personnel operation action pattern recognition algorithm, combined with the real-time status monitoring data of the system, if the access conditions of the regulations are met, The system operating state parameters provided by real-time state monitoring and the actual operation instructions input by the operator realize the recognition of human error patterns in nuclear power plants, and display the recognition results of human error patterns on the operation monitoring interface. According to the real-time state monitoring provided The operating state parameters of the system and the actual operation instructions input by the operator realize the recognition of human error patterns in nuclear power plants, including: Read the system operating state parameters and actual operation instructions into the equipment state information database, compare and analyze the read data with the default parameter settings in the equipment state information database, and if it detects that the data changes, it is considered that the operator has changed the current The device performs an operation action, records the device whose parameters have changed, parameter attributes, and parameter change values and integrates it into an operation action information, triggers an operation action event, and judges whether the operation object matches based on the current monitoring operation event status mode. , then: determine whether the equipment control mode and/or operation action match, if not, it is "wrong operation acts on the correct object" human error mode recognition; if so, jump to S1; If not, then: judge whether the equipment control mode and/or operation action match, if yes, it is "target selection error" human error pattern recognition; if not: traverse the procedure action library, and judge whether the operation action is covered in the procedure action library , if yes, jump to S1; if not, it is "operation without procedures" human error pattern recognition; S1: According to the standard operation action sequence defined in the standard procedure action library, judge whether the execution sequence of the operation action matches. If so, then: combined with the execution time record of the current action step, determine whether the operation action is completed within the specified task time. If yes, then: "Correct operation" human error pattern recognition; if not, then: judge whether the manipulated object is a discrete variable, if so, then "operation delay" human error pattern recognition; if not, then "operation incomplete" human error pattern recognition; If not, then: judge whether the operation action is in advance, if so, it is "operation ahead" human error pattern recognition; if not, it is "operation lag" human error pattern recognition. 6. An electronic device, comprising a processor and a memory for storing a program executable by the processor, wherein the processor implements the core described in any one of claims 1-4 when executing the program stored in the memory Human Error Pattern Recognition Method in Power Plant. 7. A storage medium, storing a program, characterized in that, when the program is executed by a processor, the human error pattern recognition method in a nuclear power plant according to any one of claims 1-4 is realized.

Images