An incident handling method and related device

Abstract

The application provides an accident handling method and related equipment, the method comprising: in the case that an accident occurs in a nuclear power system, predicting an accident type corresponding to the accident of the nuclear power system according to current operation parameters of the nuclear power system; predicting development trend information of the accident corresponding to the accident type in a future time period according to the current operation parameters and historical operation parameters of the nuclear power system; determining a handling strategy of the accident according to the development trend information, and sending the handling strategy to a target terminal, so that a user associated with the target terminal handles the accident based on the handling strategy. In the application, the accident handling efficiency of the nuclear power system is improved.

Description

Technical Field

[0001] This application relates to the field of nuclear power technology, and in particular to an accident handling method and related equipment. Background Technology

[0002] Handling operational accidents in nuclear power systems is an integral part of ensuring the safe operation of nuclear power systems.

[0003] In an exemplary technology, when an accident occurs in a nuclear power system, maintenance personnel can find a handling strategy in a standardized accident handling manual and then handle the accident according to the handling strategy.

[0004] However, handling the above-mentioned accidents requires manual search for handling strategies, and maintenance personnel need to have certain accident handling experience in order to find the handling strategies, which leads to a long handling time for accidents, that is, there is a problem of low accident handling efficiency in nuclear power systems. Summary of the Invention

[0005] Based on the above-mentioned technological status, this application provides an accident handling method and related equipment to solve the problem of low accident handling efficiency in nuclear power systems.

[0006] To achieve the above-mentioned technical objectives, this application proposes the following technical solution: Firstly, this application provides an accident handling method, including: In the event of an accident in a nuclear power system, the type of accident corresponding to the accident is predicted based on the current operating parameters of the nuclear power system. Based on the current operating parameters and the historical operating parameters of the nuclear power system, predict the development trend information of the accident corresponding to the accident type in the future time period; Based on the development trend information, a handling strategy for the incident is determined, and the handling strategy is sent to the target terminal so that the user associated with the target terminal can handle the incident based on the handling strategy.

[0007] In some implementations, predicting the development trend information of the accident corresponding to the accident type in the future time period based on the current operating parameters and the historical operating parameters of the nuclear power system includes: Determine the initial conditions and boundary conditions of the nuclear power system; Based on the initial conditions, the boundary conditions, and the current operating parameters, the key parameters of the nuclear power system are determined; Based on the key parameters and historical operating parameters, predict the development trend of the accident corresponding to the accident type in the future time period.

[0008] In some implementations, determining the initial conditions and boundary conditions of the nuclear power system includes: A spatial model of the nuclear power system is constructed, and the current state of the nuclear power system is configured on the simulation model to obtain an intermediate model; Based on the current operating parameters, the boundary conditions of the intermediate model are set to obtain the simulation model, wherein the initial conditions and the boundary conditions are determined based on the simulation model.

[0009] In some implementations, determining the key parameters of the nuclear power system based on the initial conditions, the boundary conditions, and the current operating parameters includes: Determine multiple mapping relationships corresponding to the simulation model; Based on the mapping relationships and the current operating parameters, the key parameters of the nuclear power system are determined.

[0010] In some implementations, predicting the development trend information of the accident corresponding to the accident type in the future time period based on the key parameters and the historical operating parameters includes: The key parameters, the accident type, and the historical operating parameters are input into the prediction model to obtain the prediction range of the key parameters in the future time period. The development trend information is determined based on the predicted values ​​within the predicted range.

[0011] In some implementations, determining the development trend information based on the predicted values ​​within the prediction interval includes: Determine the absolute value of the difference between the upper and lower limits in the prediction interval; If the absolute value of the difference is less than a preset threshold, the predicted value in the prediction interval is determined as the development trend information.

[0012] In some implementations, after determining the absolute value of the difference between the upper and lower limits in the prediction interval, the method further includes: If the absolute value of the difference is greater than or equal to a preset threshold, the development trend information is determined based on the simulation model corresponding to the nuclear power system and the current operating parameters.

[0013] In some implementations, before inputting the key parameters, the accident type, and the historical operating parameters into the prediction model, the method further includes: Multiple training samples are obtained, including the initial state of the nuclear power system, the future state of the nuclear power system, and the operation parameters corresponding to the operation of the nuclear power system during the process of the evolution from the initial state to the future state. The preset model is trained based on each of the training samples to obtain the prediction model.

[0014] In some implementations, acquiring multiple training samples includes: Acquire data corresponding to various accident scenarios of the nuclear power system; The data corresponding to each accident scenario is processed to obtain the initial state, future state, and operating parameters of the nuclear power system in each accident scenario; Based on the initial state, the future state, and the operation parameters corresponding to the accident scenario, a training sample corresponding to the accident scenario is constructed.

[0015] In some implementations, determining the incident handling strategy based on the development trend information includes: In the knowledge graph database, a strategy matching the development trend information is determined; The handling strategy for the incident is determined based on the matching strategy.

[0016] In some implementations, determining the incident handling strategy based on the matched strategy includes: Using the simulation model corresponding to the nuclear power system, the matched strategy is used to simulate accident handling and obtain effect parameters; If the effect parameter indicates that the matched strategy can handle the incident, the matched strategy is determined as the handling strategy.

[0017] In some implementations, sending the processing strategy to the target terminal includes: The prediction curve corresponding to the handling of the accident is determined according to the processing strategy. The processing strategy and the prediction curve are sent to the target terminal.

[0018] Secondly, this application provides an accident handling device, comprising: The acquisition module is used to predict the type of accident that occurs in a nuclear power system based on the current operating parameters of the nuclear power system in the event of an accident. The prediction module is used to predict the development trend information of the accident corresponding to the accident type in the future time period based on the current operating parameters and the historical operating parameters of the nuclear power system. The determination module is used to determine the handling strategy for the incident based on the development trend information, and send the handling strategy to the target terminal so that the user associated with the target terminal can handle the incident based on the handling strategy.

[0019] This application provides an accident handling method and related equipment. In the event of an accident in a nuclear power system, the method predicts the accident type based on the current operating parameters of the nuclear power system. By using the current and historical operating parameters, it predicts the future development trend of the accident type over a future time period. Based on this trend, it determines an accident handling strategy and sends the strategy to a target terminal, enabling users associated with that terminal to handle the accident accordingly. This application determines the accident development trend by predicting the accident type, the current and historical operating parameters of the nuclear power system. This allows for the automatic and accurate determination of the accident handling strategy based on the trend information, eliminating the need for maintenance personnel to search for the strategy, reducing accident handling time, and improving the efficiency of nuclear power system accident handling. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0021] Figure 1 A flowchart of an accident handling method provided in this application embodiment Figure 1 .

[0022] Figure 2 The present application provides a flowchart of an accident handling method. Figure 2 .

[0023] Figure 3 The present application provides a flowchart of an accident handling method. Figure 3 .

[0024] Figure 4 The present application provides a flowchart of an accident handling method. Figure 4 .

[0025] Figure 5 This application provides a schematic diagram of the functional modules of an accident handling device.

[0026] Figure 6 This application provides a schematic diagram of the structure of an electronic device. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] First, let's explain the technical terms used in this application: VR, Virtual Reality; AR, Augmented Reality; AI, Artificial Intelligence; LSTM, Long Short-Term Memory; SGTR, Steam Generator Tube Rupture; DOP, Digital Operating Procedure; EOPs, Emergency Operating Procedures.

[0029] Accident Operation Procedures: Procedures that provide necessary guidance to operators to mitigate the consequences of transients and accidents that cause nuclear power plant parameters to exceed the limits of reactor protection systems or dedicated safety facilities.

[0030] Data-driven: A decision-making methodology that replaces intuition and experience with data analysis. It forms a closed-loop optimization mechanism by building automated decision-making models, and uses data as a core production factor to support efficient decision-making and action.

[0031] Digital twin: A technology that uses physical models, sensor data, and multidisciplinary simulation techniques to create a digital mapping in virtual space that is highly consistent with the physical entity, in order to achieve real-time monitoring, predictive analysis, and optimization.

[0032] Artificial intelligence (AI) refers to the theories, methods, and technologies that simulate human intelligence. It uses algorithms to enable computers to possess human-like intelligence, including technological approaches such as machine learning and deep learning.

[0033] Virtual and real-world interaction refers to the technology of integrating virtual information with the real environment through computer technology to achieve human-computer interaction, mainly including two forms: VR and AR.

[0034] Handling operational accidents in nuclear power systems is an integral part of ensuring the safe operation of nuclear power systems.

[0035] In one exemplary technology, when an accident occurs in a nuclear power system, maintenance personnel look up handling strategies in a standardized accident resolution manual and then handle the accident according to those strategies. However, this accident handling requires manual searching for handling strategies, and maintenance personnel need to have certain accident handling experience to find the correct strategies, resulting in a long handling time and low efficiency in nuclear power system accident handling. Furthermore, in emergency situations, manually searching and verifying handling steps is time-consuming and error-prone; the accident resolution manual is a static procedure and cannot be dynamically adjusted or optimized based on the real-time status of the power plant, making it difficult to cope with complex and ever-changing accident sequences; additionally, accident handling heavily relies on the personal experience and judgment of maintenance personnel and lacks the ability to utilize the massive amounts of power plant data for intelligent diagnosis and forward-looking prediction.

[0036] In another exemplary technique, the operational process of an accident is reproduced using a digital twin system operation tracking method, thereby replicating the state of the unit in the nuclear power system at the corresponding moment. This method generates an operation sequence based on generated samples and simulates the operation of all equipment in the initialized digital twin system according to the operation sequence.

[0037] However, the main focus of this method is to map real data from nuclear power systems to simulators for teaching and training using twin technology. It can help with tracing the causes of nuclear power system accidents, reviewing operations, and training operators, but it is only helpful after an accident and cannot handle accidents in a timely manner.

[0038] In another exemplary technology, based on the unit emergency accident procedures of a nuclear power system, three software tools were developed: data acquisition and image recognition, logic configuration, and screen configuration. These tools integrate neural network-based text information recognition technology to quickly realize the logic model and screen construction of the accident procedures, thus constructing an intelligent emergency accident procedure system. This system tracks and executes the emergency accident procedures, enabling rapid and accurate location of accidents and diagnosis of their causes.

[0039] The aforementioned method constructs an intelligent emergency accident procedure system. When an accident occurs in a nuclear power system, by collecting unit data, it can intuitively display the judgment criteria for the accident procedure, monitor equipment operating status and key parameters, achieve data-driven procedure tracking, and provide corresponding handling procedure guidance and suggestions. However, this method is more like an efficient "instruction manual" execution aid, mainly focusing on procedure tracking and guidance. Its capabilities for high-level automated decision support, especially at the levels of predictive diagnosis and autonomous decision-making, are relatively weak.

[0040] In another exemplary technology, a "Key Technology and Application of High-Reliability Operator Assistance System for Third-Generation Nuclear Power Systems" is disclosed. This method mainly innovates in three aspects: accident operation procedure system design, automated monitoring and diagnosis, and system reliability assurance. Regarding the strategic accident operation procedure system: it revolutionizes the traditional "event-oriented" accident procedures by introducing the concept of "state function," focusing more on the maintenance status of nuclear power plant safety functions. Regarding automated monitoring and diagnosis: the core adopts a cyclical, modular architecture of "diagnosis-processing-re-diagnosis." This means that accident handling is no longer a single linear process, but a dynamic, continuous assessment and response closed-loop process, thus simplifying the accident operation document system. Regarding system reliability assurance: to ensure the system's own reliability and avoid misleading operators, it introduces methods such as robustness analysis, power outage consequence analysis, and fire consequence analysis to design the reliability of sensors and information flows involved in accident handling. In addition, the system establishes a system verification method based on accident scenario simulation and human factor reliability evaluation. Through tools such as the HCR model and NASA-TLX scale, statistical analysis is conducted on the operator's diagnostic error probability and workload to verify the system's effectiveness.

[0041] However, the aforementioned approaches differ across generations in their core concepts, technological depth, and system capabilities. For example, regarding situational awareness, there is a lack of foresight; it remains essentially a "post-event response" model. While it excels at diagnosing and operating according to procedures based on current data, it cannot predict the evolution of an incident. This could cause operators to miss the optimal intervention window when dealing with complex and rapidly evolving incidents. Furthermore, the level of automation and adaptability in decision-making is relatively low. Additionally, the aforementioned automated diagnosis and execution still strictly rely on pre-set, static procedural logic and thresholds. It lacks the ability to autonomously optimize and judge outside of procedures, and it cannot adaptively adjust incident handling strategies and paths based on the real-time dynamic conditions of the power plant. Finally, regarding the new challenges of digital transformation, although existing technologies are themselves digital achievements, deeper digital transformation (such as the introduction of complex AI and digital twins) will bring new challenges such as cybersecurity, system integration, and new human-cause risks (such as over-reliance on automation). These are key issues that existing systems need to address as they evolve further.

[0042] In another exemplary technology, an intelligent accident handling procedure operation method for nuclear power systems is disclosed. By intelligently calculating the underlying information related to the operating status of the nuclear power system according to its function or purpose, multiple decision-related information related to the SOP procedure operation sheet and execution path are obtained. That is, the frequent call to the display screen is reduced, thus improving execution efficiency. Dynamic lines and intelligent diagnostic information can prevent operator misjudgment and improve execution accuracy. In addition, the automatic sequence execution function realizes automatic diagnosis and execution of the procedure, reducing the operator's workload.

[0043] While the aforementioned methods provide operators with decision-making information and support automated sequence execution through dynamic information integration and logical operations, most existing systems respond based on predefined logic and thresholds. For example, when a parameter exceeds a certain limit, a corresponding procedure is triggered. This is a "post-event response" model, which cannot predict the evolution of an accident and may miss the best handling opportunity. In other words, the "perception-cognition" is limited: it focuses on logical judgment and automated execution, and is insufficient in deep cognition (such as accident prediction and multi-strategy simulation).

[0044] To address the aforementioned technical problems, this application proposes an accident handling method that enables proactive prediction, intelligent diagnosis, dynamic execution, and collaborative management of accident handling, thereby significantly improving the safety and emergency response efficiency of nuclear power systems. The following detailed description of the accident handling method through various embodiments further illustrates this application.

[0045] Reference Figure 1 , Figure 1 A flowchart of an accident handling method provided in this application embodiment Figure 1 .like Figure 1 As shown, the accident handling method provided in this embodiment includes: Step S101: In the event of an accident in the nuclear power system, predict the type of accident corresponding to the accident based on the current operating parameters of the nuclear power system.

[0046] In this embodiment, the executing entity is an accident handling device. The accident handling device can be a terminal device in a nuclear power system used to handle accidents, or it can be the nuclear power system itself, or any terminal device with accident handling function. For ease of description, the term "device" will be used to refer to the accident handling device below.

[0047] The nuclear power system includes a data acquisition and sensing layer, comprised of a sensor network and actuator status devices. This layer is responsible for collecting real-time operating parameters from the nuclear power system and storing them in a real-time database. The system also includes an accident detection unit. When this unit detects an accident, it retrieves the system's real-time operating parameters at the time of the accident from the real-time database as the current operating parameters. These current operating parameters include data from various sensors within the nuclear power system, such as pressure, temperature, water level, and function of various devices. After obtaining the current operating parameters of the nuclear power system, the accident type corresponding to an accident occurring in the nuclear power system is predicted based on these parameters. In one example, abnormal operating parameters can be extracted from the current operating parameters and compared with pre-stored parameters associated with each type. If the parameters associated with a certain type contain all the abnormal operating parameters, then that type is taken as the accident type. In another example, the device is equipped with a prediction model. By inputting the current operating parameters into the prediction model, the model outputs the accident type. The prediction model is trained using training data, which includes the operating parameters of the nuclear power system at the time of an accident and the labeled accident type.

[0048] Step S102: Based on the current operating parameters and the historical operating parameters of the nuclear power system, predict the development trend information of the accident corresponding to the accident type in the future time period.

[0049] The device is equipped with a digital twin engine, which includes a dynamic accident process prediction model. The dynamic accident process prediction model can be constructed from time-series prediction networks such as long short-term memory networks or deep learning networks, and can make predictions using real-time data and historical data.

[0050] In response, the device acquires historical operating parameters of the nuclear power system. These historical operating parameters are, for example, the operating parameters of the nuclear power system within a historical time period, and the end point of the historical time period can be the time when an accident occurs in the nuclear power system.

[0051] After obtaining historical operating parameters, the device transmits the current operating parameters, historical operating parameters, and accident type to the dynamic accident progression prediction model. The model then predicts the development trend of the accident corresponding to the accident type over a future time period. This future time period can be several minutes, such as the next 5 minutes, and the development trend information can be the predicted values ​​of key parameters of the nuclear power system over this period. For example, if the accident type is accident A, then the key parameter of the nuclear power system is parameter B, and therefore it is necessary to predict the value of parameter B over the future time period.

[0052] Step S103: Determine the incident handling strategy based on the development trend information, and send the handling strategy to the target terminal so that the users associated with the target terminal can handle the incident based on the handling strategy.

[0053] After obtaining the trend information, the incident handling strategy can be determined based on this information. In one example, the trend information can be a predicted value; the numerical range within which the predicted value falls is then determined, and the strategy matching this numerical range is the handling strategy.

[0054] In another example, the device uses a knowledge graph database to determine the strategy that matches the development trend information and then determines the accident handling strategy based on the matched strategy. The knowledge graph database stores a knowledge graph composed of prior knowledge about accident handling in nuclear power systems. This knowledge graph can be a state-action mapping engine. By determining the accident state through the development trend information, the action mapped to that state is determined in the knowledge graph; the mapped action is the handling strategy. Furthermore, the digital twin engine also sets up a simulation model, such as a high-fidelity physical model. Based on the actual parameter settings of the nuclear power system, the simulation model can simulate the thermal-hydraulic, seed physics, and other processes of the nuclear power system under normal and accident states. After determining the strategy that matches the development trend information, the device uses the corresponding simulation model of the nuclear power system to simulate the accident handling using the matched strategy to obtain effect parameters. These effect parameters can be the completion rate of the accident handling. When the effect parameters indicate that the matched strategy can handle the accident, the matched strategy is determined as the handling strategy. For example, when the completion rate reaches 95% or higher, it can be determined that the matched strategy can handle the accident. By using simulation models to accurately and completely deduce the matching strategies, the accuracy of the given processing strategies can be verified, thereby ensuring that accidents occurring in nuclear power systems can be effectively handled.

[0055] After determining the handling strategy, the strategy is sent to the target terminal, enabling the users associated with the target terminal to handle the incident in a timely manner based on the strategy.

[0056] The target terminal can be a device with a DOP (Dynamic Operation Point) human-machine interface, a device hosting a VR or AR training and drill platform, or a mobile emergency terminal. The DOP human-machine interface provides a graphical interface for the main control room operator, highlighting the current step, key parameters, system status, and AI operation suggestions, requiring operator confirmation for execution. VR / AR training and drill platform: Shares the same model with the digital twin engine, provides immersive accident handling training, and can record the operation process for review; Mobile emergency terminal: Provides a simplified interface and commands for on-site emergency personnel.

[0057] In addition, the device is equipped with a network security and access control layer, which can perform strict identity verification, authorization control, and operational auditing to ensure the network security and data integrity of the nuclear power system. Understandably, when a user associated with the target terminal performs accident handling based on the processing policy, they need to authenticate and authorize the device. If the authentication and authorization are successful, the device will respond to the user's operational instructions to the nuclear power system based on the target terminal.

[0058] Furthermore, the device determines the predicted curve after accident handling based on the processing strategy, and then sends the processing strategy and the predicted curve to the target terminal, enabling the user to handle the accident based on the predicted curve and the processing strategy. The predicted curve can show the numerical trend of key parameters of the nuclear power system after adopting the processing strategy, and can serve as the basis for decision-making on the processing strategy. The following example illustrates the accident handling process of a nuclear power system: 1. The nuclear power system detected a continuous drop in pressurizer pressure and an abnormally high level of radiation from the steam generator through sensors.

[0059] 2. The current operating parameters of the nuclear power system are transmitted to the digital twin engine of the device in real time. The intelligent diagnostic module in the digital twin system calculates quickly based on Bayesian network and outputs "SGTR accident" as the highest probability event, that is, determines the data type corresponding to the accident.

[0060] 3. The dynamic prediction module (data-driven) in the device initiates the SGTR special simulation to predict the loss trend of primary coolant charge in the future and provide early warning of the risk of core exposure.

[0061] 4. The state-action mapping engine is activated and automatically enters the SGTR processing procedure based on the prediction results, and dynamically decides to prioritize the "pressure reduction to safe injection setpoint" step, that is, to determine the processing strategy.

[0062] 5. Specific operation instructions and AI decision-making basis (such as prediction curves) are pushed to the DOP interface (target terminal) in the main control room and highlighted.

[0063] 6. After the operator reviews and confirms the instruction, the corresponding valves are automatically operated through the power plant control system, the nuclear power system enters the next state, and a closed-loop feedback is formed.

[0064] In this embodiment, in the event of an accident in a nuclear power system, the accident type is predicted based on the current operating parameters of the nuclear power system. Using both current and historical operating parameters, the future development trend of the accident type is predicted. Based on this trend, a handling strategy is determined and sent to the target terminal, allowing the associated user to process the accident accordingly. This embodiment determines the accident's development trend by predicting the accident type, current operating parameters, and historical operating parameters. This automatic and accurate determination of the handling strategy eliminates the need for maintenance personnel to search for the strategy, reducing processing time and improving the efficiency of nuclear power system accident handling.

[0065] Reference Figure 2 , Figure 2A flowchart of an accident handling method provided in this application embodiment Figure 2 ,based on Figure 1 In the embodiment shown, step S102 includes: Step S201: Determine the initial conditions and boundary conditions of the nuclear power system.

[0066] In this embodiment, the device needs to determine initial and boundary conditions based on real-time data. These initial and boundary conditions are not simply data imports, but a complex procedure called "data synchronization" to address issues of incomplete, inconsistent, and delayed data, thereby constructing a "virtual initial state" that most closely reflects the current real state of the nuclear power system. Initial conditions can be state variables of the nuclear power system at the time of an accident, such as the pressure, temperature, flow rate, density, and power of various devices within the system. Boundary conditions represent the interaction between the nuclear power system and the external environment, including static fixed boundary conditions and dynamic interactive boundary conditions. Static fixed boundary conditions include, but are not limited to, environmental conditions, equipment status within the nuclear power system, and valve openings. Environmental conditions include, for example, setting the current temperature of the final heat sink (seawater) based on meteorological data; equipment status, such as setting a pump to "operational" state based on control system signals and setting its speed to the measured value; and valve openings: setting the opening percentage of key valves as real-time feedback values.

[0067] Dynamic interactive boundary conditions are, for example, the model's execution of operational commands. When an operational command is issued (such as "adjust valve V01 opening to 50% in 10 seconds"), this future operation is pre-translated into changes in the model's boundary conditions at the corresponding time point. This allows the model to perform predictive simulations of "what will happen if this operation is performed." The model refers to the simulation model, and the boundary conditions are the conditions that need to be satisfied at each boundary of the simulation model.

[0068] For example, a spatial model corresponding to the nuclear power system is constructed, and the current state of the nuclear power system is configured on the simulation model to obtain an intermediate model. The boundary conditions of the intermediate model are set based on the current operating parameters to obtain a simulation model. The initial conditions and boundary conditions are then extracted based on the simulation model.

[0069] Step S202: Determine the key parameters of the nuclear power system based on the initial conditions, boundary conditions, and current operating parameters.

[0070] After determining the initial and boundary conditions, the key parameters of the nuclear power system are determined using the initial and boundary conditions and the current operating parameters.

[0071] In one example, a nuclear power system has a set of partial differential equations describing the conservation of mass, energy, and momentum. By substituting the boundary conditions, initial conditions, and current operating parameters into these partial differential equations, the key parameters of the nuclear power system can be solved.

[0072] In another example, multiple mapping relationships corresponding to the simulation model are determined. Since the simulation system itself provides initial and boundary conditions, key parameters of the nuclear power system can be determined based on these mapping relationships and current operating parameters. Specifically, the core of the simulation model is a set of partial differential equations describing the conservation of mass, energy, and momentum in the nuclear power system. The device first discretizes the simulation model in space and time, dividing the continuous physical space of the nuclear power system's reactor, pipes, and equipment into millions or even hundreds of millions of tiny computational grids. On each grid, based on given initial conditions and satisfying boundary constraints, these partial differential equations are solved. The solution process requires simulating complex coupling phenomena such as flow, heat transfer, and neutron dynamics in the nuclear power system. During the solution process, time propagation is required; that is, starting from t=0 (the initial state), a new state with an extremely short time step (such as a millisecond-level time step) is calculated. This new state is then used as the starting point to continue calculating the next time step, and so on, thus achieving time-provisioned simulation of the physical processes. During and after the model is solved, the simulation model can output two types of parameters: virtual values ​​of measurable parameters and non-measurable but crucial internal states.

[0073] Virtual values ​​of measurable parameters: At the actual installation location of the sensor, the calculated values ​​of temperature, pressure, etc. at that point are output. These values ​​can be continuously compared with real-time measured values ​​to verify the accuracy of the model and diagnose sensor faults.

[0074] Unmeasurable yet crucial internal states: This is the greatest value of high-fidelity models (simulation models). They can output the state at any location. For example, local power peaks in the reactor core: real-time calculation of power distribution within the core, providing early warning of potential localized overheating; Bubble proportion distribution: Predicting the generation and distribution of steam bubbles in the primary loop is crucial for determining the boiling state of the coolant; Critical heat flux margin: Calculates how much safety margin is left before the current state deviates from the dangerous state of nucleation boiling. This is one of the highest priority parameters in accident management. Stress and fatigue: Output the thermal and mechanical stresses of critical equipment components (such as pressure vessels and main pipelines).

[0075] Step S203: Based on key parameters and historical operating parameters, predict the development trend information of accidents corresponding to the accident type in the future time period.

[0076] After determining the key parameters, the system predicts the development trend of accidents corresponding to different accident types over future time periods based on these parameters and historical operating parameters. For example, the key parameters, historical operating parameters, and accident type can be input into a dynamic accident progression prediction model to obtain the development trend information output by the model.

[0077] In this embodiment, the high-fidelity physical model achieves: Synchronization: By using data assimilation, the initial state of the virtual object is synchronized with that of the physical object; Prediction: Given boundary conditions (including future operations), extrapolate future states based on first principles. Insight: The critical internal states of a nuclear power system that cannot be directly measured but determine its safety.

[0078] Ultimately, these key output state parameters flow in real time to the dynamic accident process prediction model and autonomous decision engine (the module used to determine the handling strategy) in the device, becoming the ultimate basis for intelligent diagnosis and forward-looking decision-making, thus truly closing the intelligent loop of "perception-cognition-decision-execution".

[0079] Figure 3 A flowchart of an accident handling method provided in this application embodiment Figure 3 ,based on Figure 2 In the embodiment shown, step S203 includes: Step S301: Input the key parameters, accident type, and historical operating parameters into the prediction model to obtain the prediction range of the key parameters in the future time period.

[0080] In this embodiment, the device is equipped with a prediction model, such as the dynamic accident process prediction model described above. After obtaining the key parameters, the device inputs the key parameters, the accident type, and historical operating parameters into the prediction model to obtain the prediction range of the key parameters over a future time period. The prediction range is, for example, a numerical range consisting of the upper and lower limits of the key parameters over the future time period.

[0081] Step S302: Determine the development trend information based on the predicted values ​​of the prediction interval.

[0082] After determining the prediction interval, the development trend information can be determined based on the predicted values ​​within the prediction interval. For example, the device determines the absolute value of the difference between the upper and lower limits within the prediction interval and compares this absolute value with a preset threshold. If the absolute value of the difference is less than the preset threshold, the predicted value within the prediction interval is used as the development trend information; that is, a value can be randomly selected from the prediction interval as the development trend information. If the absolute value of the difference is greater than or equal to the preset threshold, the development trend information is determined based on the simulation system corresponding to the nuclear power system and the current operating parameters.

[0083] For example, if the prediction range is too wide (e.g., the upper and lower limits of the pressure prediction for the next 10 minutes differ by more than 2 MPa), it indicates that the current state may be outside the coverage of the prediction model's training data, making the prediction unreliable. In this case, the device automatically triggers a high-fidelity model calculation, using its more accurate result to replace or correct the prediction model's forecast, and adds this new data point to the database for subsequent updates to the prediction model.

[0084] In this embodiment, the prediction model predicts the forecast range of key parameters in the future time period, thereby accurately determining the development trend information based on the forecast range.

[0085] Figure 4 A flowchart of an accident handling method provided in this application embodiment Figure 4 .based on Figure 3 In the embodiment shown, before step S301, the method further includes: Step S401: Obtain multiple training samples, including the initial state of the nuclear power system, the future state, and the operation parameters corresponding to the operation of the nuclear power system during the process of evolving from the initial state to the future state.

[0086] Step S402: Train the preset model based on each training sample to obtain the prediction model.

[0087] In this embodiment, the device acquires multiple training samples and trains a preset model based on each training sample to obtain a prediction model. The training samples include the initial state of the nuclear power system, the future state, and the operating parameters corresponding to the operations performed by the nuclear power system during the evolution from the initial state to the future state.

[0088] In one example, training samples could be manually labeled samples of operating parameters and handling strategies for accidents occurring in nuclear power systems.

[0089] In another example, data corresponding to various accident scenarios in the nuclear power system are obtained, and the data corresponding to each accident scenario is processed to obtain the initial state, future state, and operating parameters of the nuclear power system in each accident scenario. Based on the initial state, future state, and operating parameters corresponding to the accident scenario, training samples corresponding to the accident scenario are constructed.

[0090] For example, using a high-fidelity physical model, simulations of massive accident scenarios (covering design baseline accidents, beyond design baseline accidents, and combinations of various initial causes and failures) are performed on a supercomputing platform, generating a large-scale, high-fidelity, labeled simulation dataset. Each training sample contains an initial state, an operation sequence (a sequence of operation parameters arranged according to operation time), and a future state.

[0091] Training with Physically Constrained Embeddings: The key to training a predictive model (such as a deep neural network) using this dataset lies not in ordinary training, but in incorporating physical knowledge as a constraint embedding loss function: Traditional Loss: This aims to make the predicted value as close as possible to the output of a high-fidelity model (supervised learning). Physical Information Loss: This incorporates residuals from physical conservation laws into the loss function. For example, the state parameters predicted by the neural network must also approximately satisfy the mass and energy conservation equations. This is the core idea of ​​physically information neural networks. This ensures that even in regions of sparse data, the neural network's predictions will not severely violate physical laws.

[0092] Output: A trained "prediction model" with embedded physical laws is obtained, which can respond in milliseconds.

[0093] Online extrapolation of the predictive model, i.e., dynamic fusion of the predictive model first and the high-fidelity model as a correction. When real-time data floods in, the two types of models are dynamically fused according to the following strategy: State synchronization and initialization: After data assimilation, real-time data is used to set consistent initial and boundary conditions for both the high-fidelity model and the prediction model.

[0094] Predictive model predictions: For most routine monitoring and prediction requests requiring millisecond-level response times, the prediction model is invoked first. The prediction model outputs the predicted trajectory of key parameters for a future time period (e.g., the next 30 minutes) within milliseconds.

[0095] The roles and triggering conditions of the high-fidelity model: Role 1: Perform "ultimate validation" of critical decisions. When the autonomous decision engine generates a high-risk operation suggestion (such as "recommendation to perform safety injection") based on the prediction of the surrogate model, before finally recommending it to the operator, the system will asynchronously start a high-fidelity model in the background to perform a precise and complete deduction of the operation strategy to verify the reliability of the prediction model.

[0096] Role Two: "Calibrate" when uncertainty is too high. Continuously monitor the confidence interval width of the prediction model. If the confidence interval is too wide (e.g., a difference of more than 2 MPa between the upper and lower bounds in a pressure prediction for the next 10 minutes), it indicates that the current state may exceed the coverage of the prediction model's training data, making the prediction unreliable. In this case, automatically trigger a high-fidelity model calculation, using its more accurate result to replace or correct the surrogate model's prediction, and add this new data point to the database for subsequent updates to the surrogate model.

[0097] This is a crucial closed loop. The real power plant data accumulated during online operation, along with the validation results of the high-fidelity model at key points, forms an incremental dataset. This new data is then used periodically or irregularly to fine-tune or retrain the fast proxy model, allowing it to continuously adapt to changes in the actual characteristics of the power plant (such as equipment aging), thus achieving lifelong learning for the model.

[0098] The aforementioned fusion computing process achieves three values ​​within the device: Balancing speed and accuracy: Using predictive models to meet the needs of real-time monitoring and rapid response, and using high-fidelity models to ensure the accuracy and reliability of final decisions.

[0099] Credibility of predictions: Through physical constraint training and uncertainty quantification, the predictions of the prediction model are no longer "black box mystery", but transparent and trustworthy results with physical basis and confidence assessment.

[0100] The predictive model's adaptive evolution: Through online calibration and continuous learning, the predictive capabilities of the entire digital twin can continuously improve and evolve over time and with the accumulation of data, becoming increasingly closer to the real nuclear power system. This entire mechanism is the underlying computational guarantee that enables the digital twin engine to perform "proactive intelligent decision-making."

[0101] Corresponding to the above-described accident handling method, this application also provides an accident handling device. Figure 5 This is a schematic diagram of a module of an accident handling device provided in an embodiment of this application. The accident handling device 500 provided in this embodiment includes: The acquisition module 510 is used to predict the type of accident that occurs in the nuclear power system based on the current operating parameters of the nuclear power system in the event of an accident. The prediction module 520 is used to predict the development trend information of accidents corresponding to accident types in the future time period based on current operating parameters and historical operating parameters of the nuclear power system. The determination module 530 is used to determine the handling strategy for the incident based on the development trend information, and send the handling strategy to the target terminal so that the users associated with the target terminal can handle the incident based on the handling strategy.

[0102] In some implementations, the accident handling device 500 is also used for: Determine the initial and boundary conditions of the nuclear power system; Based on the initial conditions, boundary conditions, and current operating parameters, determine the key parameters of the nuclear power system; Based on key parameters and historical operating parameters, predict the development trend of accidents corresponding to different accident types in the future.

[0103] In some implementations, the accident handling device 500 is also used for: A spatial model of the nuclear power system is constructed, and the current state of the nuclear power system is configured on the simulation model to obtain an intermediate model; Based on the current operating parameters, the boundary conditions of the intermediate model are set to obtain the simulation model. The initial conditions and boundary conditions are determined based on the simulation model.

[0104] In some implementations, the accident handling device 500 is also used for: Determine multiple mapping relationships corresponding to the simulation model; Based on the various mapping relationships and current operating parameters, the key parameters of the nuclear power system are determined.

[0105] In some implementations, the accident handling device 500 is also used for: By inputting key parameters, accident types, and historical operating parameters into the prediction model, the prediction range of key parameters for future time periods can be obtained. Based on the predicted values ​​within the forecast range, information on development trends is determined.

[0106] In some implementations, the accident handling device 500 is also used for: Determine the absolute value of the difference between the upper and lower limits of the prediction interval; If the absolute value of the difference is less than a preset threshold, the predicted value in the prediction interval will be determined as the development trend information.

[0107] In some implementations, the accident handling device 500 is also used for: If the absolute value of the difference is greater than or equal to a preset threshold, the development trend information is determined based on the simulation model of the nuclear power system and the current operating parameters.

[0108] In some implementations, the accident handling device 500 is also used for: Multiple training samples were obtained, including the initial state of the nuclear power system, the future state of the nuclear power system, and the corresponding operating parameters of the nuclear power system as it evolves from the initial state to the future state. The preset model is trained based on each training sample to obtain the prediction model.

[0109] In some implementations, the accident handling device 500 is also used for: Acquire data corresponding to various accident scenarios in the nuclear power system; The data corresponding to each accident scenario is processed to obtain the initial state, future state, and operating parameters of the nuclear power system in each accident scenario; Based on the initial state, future state, and operational parameters corresponding to the accident scenario, training samples corresponding to the accident scenario are constructed.

[0110] In some implementations, the accident handling device 500 is also used for: In knowledge graph databases, strategies for matching development trend information are determined; The incident handling strategy is determined based on the matching strategy.

[0111] In some implementations, the accident handling device 500 is also used for: Using a simulation model corresponding to the nuclear power system, the matching strategy is simulated for accident handling to obtain effect parameters; If the effect parameters indicate that the matched strategy can handle the incident, the matched strategy is determined as the handling strategy.

[0112] In some implementations, the accident handling device 500 is also used for: Determine the predicted curve corresponding to the handling of the accident based on the handling strategy; The processing strategy and prediction curve are sent to the target terminal.

[0113] The accident handling apparatus and the accident handling method provided in the above embodiments of this application belong to the same application concept and can execute the accident handling method provided in any of the above embodiments of this application. They have the corresponding functional modules and beneficial effects for executing the accident handling method. Technical details not described in detail in this embodiment can be found in the specific processing content of the accident handling method provided in the above embodiments of this application, and will not be repeated here.

[0114] The functions implemented by each module in the accident handling device can be implemented by the same or different processors, and this application embodiment does not limit this.

[0115] It should be understood that the modules in the above-described accident handling device can be implemented by a processor calling firmware. For example, the system includes a processor connected to a memory containing instructions. The processor calls the instructions stored in the memory to implement any of the above methods or to implement the functions of each module of the device. The processor can be a general-purpose processor, such as a CPU or microprocessor, and the memory can be internal to the device or external to the system. Alternatively, the modules in the system can be implemented as hardware circuits. By designing the hardware circuits, some or all of the module functions can be implemented. The hardware circuits can be understood as one or more processors. For example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all of the above modules are implemented by designing the logical relationships between the components within the circuit. In another implementation, the hardware circuit can be implemented using a PLD, such as an FPGA, which can include a large number of logic gates. The connection relationships between the logic gates are configured through configuration files, thereby implementing the functions of some or all of the above modules. All modules of the above-described accident handling device can be implemented entirely by a processor calling firmware, entirely by hardware circuits, or partially by a processor calling firmware with the remaining parts implemented by hardware circuits.

[0116] In this application embodiment, the processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a CPU, microprocessor, GPU, or DSP. In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are fixed or reconfigurable. For example, the processor may be a hardware circuit implemented as an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above modules. Furthermore, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, or DPU.

[0117] As can be seen, each module in the above-mentioned accident handling device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor types.

[0118] Furthermore, the modules in the above-mentioned accident handling device can be integrated in whole or in part, or they can be implemented independently. In one implementation, these modules are integrated together and implemented in the form of a System-on-Chip (SoC). The SoC may include at least one processor for implementing any of the above methods or implementing the functions of the modules of the device. The at least one processor can be of different types, such as CPU and FPGA, CPU and artificial intelligence processor, CPU and GPU, etc.

[0119] This application provides a schematic diagram of the structure of an electronic device, see [link]. Figure 6 As shown, the electronic device includes a memory 600 and a processor 610; wherein the memory 600 is connected to the processor 610 and is used to store programs; the processor 610 is used to implement the accident handling method disclosed in any of the above embodiments by running the programs stored in the memory 600.

[0120] Specifically, the aforementioned electronic device may further include: a bus, a communication interface 620, an input device 630, and an output device 640. The electronic device may also include a data transceiver module, an image monitoring module, and a signal monitoring module.

[0121] The processor 610, memory 600, communication interface 620, input device 630, and output device 640 are interconnected via a bus. Among them: A bus can include a pathway for transmitting information between various components in an electronic device.

[0122] The processor 610 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0123] The processor 610 may include a main processor, as well as a baseband chip, modem, etc.

[0124] The memory 600 stores a program that executes the technical solution of this invention, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 600 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.

[0125] Input device 630 may include a device for receiving user input data and information, such as a keyboard, mouse, camera, scanner, light pen, voice input device, touch screen, pedometer, or gravity sensor.

[0126] Output device 640 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0127] The communication interface 620 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0128] The processor 610 executes the program stored in the memory 600 and calls other devices, which can be used to implement the various steps of any of the accident handling methods provided in the above embodiments of this application.

[0129] This application also proposes a chip, which includes a processor and a data interface. The processor reads and runs a program stored in the memory through the data interface to execute the accident handling method described in any of the above embodiments. For the specific processing procedure and its beneficial effects, please refer to the above-described embodiments of the accident handling method.

[0130] In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the accident handling methods according to various embodiments of this application as described in any of the above embodiments of this specification.

[0131] Computer program products can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the power device, as a standalone firmware package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0132] Furthermore, embodiments of this application may also be storage media storing a computer program, which is executed by a processor to perform the steps of the accident handling method according to various embodiments of this application described in any of the above embodiments of this specification, specifically implementing the steps of the above accident handling method.

[0133] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0134] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0135] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.

[0136] The units of the apparatus in the various embodiments of this application can be merged, divided, and deleted according to actual needs.

[0137] It should be understood, in the several embodiments provided in this application, that the disclosed terminals, devices, and methods can be implemented in other ways. For example, the terminal embodiments described above are merely illustrative; for instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0138] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.

[0139] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or as firmware functional modules or sub-modules.

[0140] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer firmware, or a combination of both. To clearly illustrate the interchangeability of hardware and firmware, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or firmware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0141] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly using hardware, firmware units executed by a processor, or a combination of both. The firmware units can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the art.

[0142] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0143] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An accident handling method, characterized in that, include: In the event of an accident in a nuclear power system, the type of accident corresponding to the accident is predicted based on the current operating parameters of the nuclear power system. Based on the current operating parameters and the historical operating parameters of the nuclear power system, predict the development trend information of the accident corresponding to the accident type in the future time period; Based on the development trend information, a handling strategy for the incident is determined, and the handling strategy is sent to the target terminal so that the user associated with the target terminal can handle the incident based on the handling strategy.

2. The accident handling method according to claim 1, characterized in that, The prediction of the development trend information of the accident corresponding to the accident type in the future time period based on the current operating parameters and the historical operating parameters of the nuclear power system includes: Determine the initial conditions and boundary conditions of the nuclear power system; Based on the initial conditions, the boundary conditions, and the current operating parameters, the key parameters of the nuclear power system are determined; Based on the key parameters and historical operating parameters, predict the development trend of the accident corresponding to the accident type in the future time period.

3. The accident handling method according to claim 2, characterized in that, Determining the initial conditions and boundary conditions of the nuclear power system includes: A spatial model of the nuclear power system is constructed, and the current state of the nuclear power system is configured on the simulation model to obtain an intermediate model; Based on the current operating parameters, the boundary conditions of the intermediate model are set to obtain the simulation model, wherein the initial conditions and the boundary conditions are determined based on the simulation model.

4. The accident handling method according to claim 3, characterized in that, The process of determining the key parameters of the nuclear power system based on the initial conditions, the boundary conditions, and the current operating parameters includes: Determine multiple mapping relationships corresponding to the simulation model; Based on the mapping relationships and the current operating parameters, the key parameters of the nuclear power system are determined.

5. The accident handling method according to claim 2, characterized in that, The step of predicting the development trend information of the accident corresponding to the accident type in the future time period based on the key parameters and the historical operating parameters includes: The key parameters, the accident type, and the historical operating parameters are input into the prediction model to obtain the prediction range of the key parameters in the future time period. The development trend information is determined based on the predicted values ​​within the predicted range.

6. The accident handling method according to claim 5, characterized in that, Determining the development trend information based on the predicted values ​​within the predicted range includes: Determine the absolute value of the difference between the upper and lower limits in the prediction interval; If the absolute value of the difference is less than a preset threshold, the predicted value in the prediction interval is determined as the development trend information.

7. The accident handling method according to claim 6, characterized in that, After determining the absolute value of the difference between the upper and lower limits in the prediction interval, the method further includes: If the absolute value of the difference is greater than or equal to a preset threshold, the development trend information is determined based on the simulation model corresponding to the nuclear power system and the current operating parameters.

8. The accident handling method according to claim 5, characterized in that, Before inputting the key parameters, the accident type, and the historical operating parameters into the prediction model, the process also includes: Multiple training samples are obtained, including the initial state of the nuclear power system, the future state of the nuclear power system, and the operation parameters corresponding to the operation of the nuclear power system during the process of the evolution from the initial state to the future state. The preset model is trained based on each of the training samples to obtain the prediction model.

9. The accident handling method according to claim 8, characterized in that, The acquisition of multiple training samples includes: Acquire data corresponding to various accident scenarios of the nuclear power system; The data corresponding to each accident scenario is processed to obtain the initial state, future state, and operating parameters of the nuclear power system in each accident scenario; Based on the initial state, the future state, and the operation parameters corresponding to the accident scenario, a training sample corresponding to the accident scenario is constructed.

10. The accident handling method according to claim 1, characterized in that, The step of determining the handling strategy for the accident based on the development trend information includes: In the knowledge graph database, a strategy matching the development trend information is determined; The handling strategy for the incident is determined based on the matching strategy.

11. The accident handling method according to claim 10, characterized in that, Determining the handling strategy for the incident based on the matching strategy includes: Using the simulation model corresponding to the nuclear power system, the matched strategy is used to simulate accident handling and obtain effect parameters; If the effect parameter indicates that the matched strategy can handle the incident, the matched strategy is determined as the handling strategy.

12. The accident handling method according to any one of claims 1-11, characterized in that, Sending the processing strategy to the target terminal includes: The prediction curve corresponding to the handling of the accident is determined according to the processing strategy. The processing strategy and the prediction curve are sent to the target terminal.

13. An accident handling device, characterized in that, include: The acquisition module is used to predict the type of accident that occurs in a nuclear power system based on the current operating parameters of the nuclear power system in the event of an accident. The prediction module is used to predict the development trend information of the accident corresponding to the accident type in the future time period based on the current operating parameters and the historical operating parameters of the nuclear power system. The determination module is used to determine the handling strategy for the incident based on the development trend information, and send the handling strategy to the target terminal so that the user associated with the target terminal can handle the incident based on the handling strategy.

Images