An unconventional situation online deduction device and method

Abstract

The application provides an unconventional situation online deduction device and method, comprising: a cognitive reasoning module for calling historical safety event data, analyzing the historical safety event data, and generating corresponding visualized review scenes; a dynamics modeling module for collecting equipment state parameters and environmental monitoring parameters in a hydropower station, fusing and analyzing the visualized review scenes, the equipment state parameters and the environmental monitoring parameters, and constructing a dynamics model of state evolution of the hydropower station; an online deduction module for determining risk bifurcation nodes and shock nodes of the hydropower station, performing online reasoning analysis on the risk bifurcation nodes and the shock nodes existing in the hydropower station, and obtaining a situation future evolution graph; and a risk identification module for determining risk step points existing in the hydropower station based on the situation future evolution graph, and synchronously visualizing the situation future evolution graph and the risk step points. The application is favorable for improving the predictability and response efficiency of industrial safety management.

Description

Technical Field

[0001] This invention relates to the field of industrial safety monitoring technology, and in particular to an unconventional situation online simulation device and method. Background Technology

[0002] Currently, unconventional situations refer to the dynamic evolution process in hydropower stations where a series of hidden and orderly abnormal events drive the system state to deviate from the normal operating range under the nonlinear coupling of multiple system parameters, and may ultimately lead to a step-like risk change through risk bifurcation, state oscillation, and other means. In the field of hydropower station safety production, perceiving, assessing and predicting the future evolution of the operational status is key to achieving early warning of risks and improving emergency response capabilities. Most existing hydropower station safety monitoring systems rely on threshold alarms based on real-time sensor data, that is, when a single parameter (such as temperature, pressure, unit vibration amplitude and dam displacement) exceeds a preset safety range, an alarm is triggered. Although this method is direct, it has obvious limitations. This approach is passive and cannot describe the dynamic evolution of the system state under the coupled effects of multiple parameters. At the same time, it is difficult to deal with unconventional situations caused by a series of hidden and orderly abnormal event chains. However, in existing technologies, although historical data structure models are used to assist decision-making, they often have the following shortcomings. First, the analysis of historical events is mostly isolated and static statistics, failing to effectively explore the temporal correlations and causal logic between events, and failing to form a deep understanding of the complete evolution chain of events. Second, in risk identification, it relies more on the judgment of the absolute value of state variables, while ignoring the step-like changes that may occur in the risk during the evolution process. Such abrupt changes are precisely the critical points for major accidents such as dam deformation intensification and unit flow interruption. Missing such warnings will lead to catastrophic consequences. Therefore, in order to overcome the above-mentioned technical problems, the present invention provides an unconventional situation online simulation device and method. Summary of the Invention

[0003] This invention provides an unconventional situation online simulation device and method, which integrates historical safety event data and real-time industrial parameters to achieve intelligent analysis and online simulation of the hydropower station's situation. Based on lessons learned from historical data, it generates a visualized recap scenario and constructs a dynamic model by combining real-time equipment status and environmental monitoring parameters to simulate the state evolution of the hydropower station. It can dynamically identify risk bifurcation nodes and oscillation nodes, predict the future evolution spectrum of the situation through online reasoning, and finally determine the risk step point and visualize it in real time, thereby helping operators to provide early warnings and make decisions. It is beneficial to improve the predictability and response efficiency of industrial safety management, reduce the probability of unexpected events, enhance risk identification capabilities through data-driven analysis, optimize emergency response processes, and support continuous improvement of safety strategies.

[0004] This invention provides an online simulation device for unconventional situations, comprising: The cognitive reasoning module is used to retrieve historical safety event data of hydropower stations from the built-in historical situation model library, and to perform intelligent analysis on the historical safety event data to generate a visual review scenario of the historical safety events. The dynamic modeling module is used to collect equipment status parameters and environmental monitoring parameters in hydropower stations, and to integrate and analyze the visualized recap scenarios with the equipment status parameters and environmental monitoring parameters to construct a dynamic model of the state evolution of hydropower stations. The online simulation module is used to determine the risk bifurcation nodes and oscillation nodes of the hydropower station based on the dynamic model, and to perform online reasoning analysis on the risk bifurcation nodes and oscillation nodes of the hydropower station to obtain the future evolution map of the situation. The risk identification module is used to determine the risk step points in the hydropower station based on the future evolution map of the situation, and to simultaneously visualize the future evolution map of the situation and the risk step points.

[0005] Preferably, an unconventional situation online deduction device, including a cognitive reasoning module, comprises: The data extraction unit is used to extract historical safety event data of hydropower stations from the historical situation model library; The mining unit is used to perform event chain mining on historical safety event data of hydropower stations to identify multiple ordered sub-events corresponding to a safety event. The link construction unit is used to construct a complete evolution link of historical security events based on the event logic relationship of multiple ordered sub-events; The visualization retrospective scenario generation unit is used to generate visualization retrospective scenarios of historical security events based on the complete evolution chain.

[0006] Preferably, an unconventional situation online simulation device includes a mining unit, comprising: The attribute extraction subunit is used to extract the event type attribute and timestamp attribute of each event record in the historical security event data to generate a basic event set; The initial event sequence generation sub-unit is used to arrange the event records of the basic event set according to the timestamp attribute in a preset order to generate the initial event sequence; The event flow analysis subunit is used for: Perform event flow analysis on the initial event sequence, the event flow analysis including: calculating the time interval and state transition probability between adjacent events; The time interval difference is compared with a preset time threshold, and the state transition probability is compared with a preset probability threshold. When the time interval difference is greater than the preset time threshold, or when the state transition probability is less than the preset probability threshold, it is determined that there is a state transition point between the current adjacent events. The sub-unit is used to divide the initial event sequence into multiple consecutive sub-event sequence segments based on the state transition points. The type determination subunit is used to identify the event type in each sub-event sequence segment based on the event type attribute, and to determine the type of each sub-event in each sub-event sequence segment; Event relationship diagram construction sub-units are used for: The sub-event types are used as nodes, and temporal association edges are established between adjacent nodes according to the time order of the sub-event sequence segments. Traverse each sub-event sequence segment to determine the temporal order and co-occurrence frequency of event types in each sub-event sequence segment, and determine the causal logical relationship between sub-event types based on the temporal order and co-occurrence frequency; Based on the causal logical relationship between sub-event types, construct the causal logical edge of the corresponding node; An event relationship graph is constructed based on the nodes, temporal correlation edges, and causal logic edges of each sub-event sequence segment. The event chain corresponding to the starting point to the target endpoint is determined based on the event relationship graph, and multiple ordered sub-events are determined based on the event chain.

[0007] Preferably, an unconventional situation online simulation device includes a dynamic modeling module, comprising: Device configuration unit, used for: The system obtains the collection requirements for different types of parameters based on the management terminal, analyzes the collection requirements, and determines the collection cycle and single collection quantity for different types of parameters. Based on the acquisition cycle and the amount of data acquired in a single acquisition, the parameters of each sensor node pre-deployed in the hydropower station are adapted. The data acquisition unit is used for: Based on the parameter adaptation results, each sensor node is controlled in parallel to collect equipment status parameters and environmental monitoring parameters in the hydropower station, and the collected equipment status parameters and environmental monitoring parameters are marked with source tags based on the location source of each sensor node. Upload the device status parameters and environmental monitoring parameters marked with the source to the data center.

[0008] Preferably, an unconventional situation online simulation device includes a dynamic modeling module, comprising: The hydropower station analysis unit is used to perform a global scan of the hydropower station and construct a dynamic state-space model of the hydropower station. The state-space model includes state equations and observation equations. The state equations characterize the dynamic law of the evolution of the internal state vector of the hydropower station over time, and the observation equations are used to establish the mapping relationship between the internal state vector of the hydropower station and the observation vectors directly measured by the sensors. Filter configuration unit, used for: The system obtains visualized retrospective scenarios of historical security events, extracts structured event data corresponding to the visualized retrospective scenarios, and extracts initial estimates and error covariance matrices of key state variables before and after the occurrence of historical security events based on the structured event data corresponding to the visualized retrospective scenarios. The filter is initialized and configured based on the initial estimates of the key state variables and the error covariance matrix. Dynamic model building unit, used for: Based on the initial configuration results, a recursive filtering estimation loop is performed on the dynamic state-space model of the hydropower station, specifically as follows: In each cycle, the state of the hydropower station at the current moment is predicted based on the state equation, and the equipment state parameters and environmental monitoring parameters of the hydropower station are collected after the prediction. The prediction results are corrected based on equipment status parameters and environmental monitoring parameters, and the optimal state prediction value of the hydropower station at the current moment is obtained based on the corrected results. A dynamic model of the state evolution of a hydropower station is obtained based on the recursive filtering estimation results.

[0009] Preferably, an unconventional situation online simulation device includes an online simulation module comprising: The simulation unit is used to perform numerical simulation of the state evolution process of a hydropower station based on a dynamic model, and to obtain dynamic change data of state parameters. The risk bifurcation node identification unit is used to identify risk bifurcation nodes in the hydropower station based on the dynamic change data of the state parameters and according to a preset method. The oscillation node identification unit is used to monitor the oscillation mode of the state variable in real time based on the dynamic change data of the state parameters, and to identify the corresponding oscillation node when the oscillation amplitude and oscillation frequency exceed the preset stability domain threshold. The reasoning and analysis unit is used to take the risk bifurcation node and oscillation node as key state points of the evolution path, and analyze the causal relationship and temporal sequence relationship between the key state points based on the multivariate reasoning engine to determine the evolution path of the deduced system state. The map generation unit is used to construct and output a future evolution map of the situation based on the evolution path.

[0010] Preferably, in an unconventional situation online simulation device, the graph generation unit maps the evolution path into a directed graph structure, where nodes represent risk bifurcation nodes and oscillation nodes, edges represent causal and temporal relationships between nodes, and the directed graph structure is visualized and rendered using a graph rendering algorithm to generate the future evolution graph of the situation.

[0011] Preferably, an unconventional situation online simulation device includes a risk identification module, comprising: The sequence extraction unit is used to extract continuous path sequences representing the evolution of the system state over time in the future situation evolution map; The sampling unit is used to sample the system state variables in the continuous path sequence and determine the target state value of the system state variables at discrete time points. The risk step point judgment unit is used to analyze system variables based on target state values ​​and determine whether there is a risk step point. The synchronous display unit is used to highlight the risk step point when there is a risk step point, and overlay and render it onto the corresponding position of the future evolution map of the situation to generate a fused view. At the same time, the fused view is sent to the visualization terminal for synchronous display.

[0012] Preferably, an unconventional situation online simulation device, including a risk step point judgment unit, comprises: The analysis sub-unit is used to calculate the first derivative sequence of the system state variables using a numerical difference algorithm based on the target state value, and then calculate the second derivative sequence of the system state variables again using a numerical difference algorithm based on the first derivative sequence. The risk step point determination sub-unit is used to scan the second derivative sequence, identify local extreme points in the second derivative sequence where the absolute value of each value is greater than a preset mutation threshold, and use the local extreme points as risk step points.

[0013] This invention provides a method for online simulation of unconventional situations in hydropower stations, comprising: Step 1: Retrieve historical security event data from the built-in historical situation model library, and perform intelligent analysis on the historical security event data to generate a visual review scenario of the historical security events; Step 2: Collect equipment status parameters and environmental monitoring parameters in the hydropower station, and perform integrated analysis on the visualized review scenario and equipment status parameters and environmental monitoring parameters to construct a dynamic model of the hydropower station's state evolution; Step 3: Based on the dynamic model, determine the risk bifurcation nodes and oscillation nodes of the hydropower station, and perform online reasoning analysis on the risk bifurcation nodes and oscillation nodes of the hydropower station to obtain the future evolution map of the situation; Step 4: Based on the future evolution map of the situation, identify the risk step points in the hydropower station, and simultaneously visualize the future evolution map of the situation and the risk step points.

[0014] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a structural diagram of an unconventional situation online simulation device according to an embodiment of the present invention; Figure 2 This is a structural diagram of the cognitive reasoning module in an unconventional situation online deduction device according to an embodiment of the present invention; Figure 3 This is a flowchart of an online simulation method for unconventional situations in an embodiment of the present invention. Detailed Implementation

[0017] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0018] Example 1: This example provides an online simulation device for unconventional situations, such as... Figure 1 As shown, it includes: The cognitive reasoning module is used to retrieve historical safety event data of hydropower stations from the built-in historical situation model library, and to perform intelligent analysis on the historical safety event data to generate a visual review scenario of the historical safety events. First, historical security event data (e.g., records of past equipment cascading failures, chemical spill response reports, etc.) related to the current focus scenario are retrieved from the historical situation model library as needed. This data is then subjected to intelligent analysis (such as causal analysis and pattern recognition) to generate a visual debriefing scenario. This is typically a dynamic, graphical timeline or flowchart that clearly shows the cause, development, key decision points, and final outcome of the historical event.

[0019] Specifically, historical safety event data can include: abnormal dam displacement events, turbine unit over-vibration events, reservoir landslide displacement events, and power plant flooding accidents. The visualized review scenario can dynamically display a complete event chain, such as "continuous rise in reservoir water level, accelerated rate of dam displacement, and a step increase in seepage flow."

[0020] The dynamic modeling module is used to collect equipment status parameters and environmental monitoring parameters in hydropower stations, and to integrate and analyze the visualized recap scenarios with the equipment status parameters and environmental monitoring parameters to construct a dynamic model of the state evolution of hydropower stations. Secondly, real-time equipment status parameters (such as motor speed, valve opening, and bearing temperature) and environmental monitoring parameters (such as area temperature, combustible gas concentration, and wind speed) are collected from sensors and control systems at the industrial site. The evolutionary logic contained in the generated visual replay scenario is then integrated and analyzed with the collected real-time parameters. Through mathematical modeling methods, a dynamic model that describes how the overall state of the hydropower station changes over time is constructed. This model can simulate the interaction between parameters and the dynamic behavior of the system.

[0021] The online simulation module is used to determine the risk bifurcation nodes and oscillation nodes of the hydropower station based on the dynamic model, and to perform online reasoning analysis on the risk bifurcation nodes and oscillation nodes of the hydropower station to obtain the future evolution map of the situation. Simulations are performed using a constructed dynamic model. During the simulation, risk bifurcation nodes (key points where multiple possible future development paths emerge, such as forks in the road where different emergency plans are chosen) and oscillation nodes (links where the system state fluctuates continuously and is difficult to stabilize, such as periodic and drastic changes in pressure or temperature) are automatically identified in the model. Based on these key nodes, the module performs online reasoning analysis across multiple paths and scenarios (such as "if-then" analysis) to deduce various possible development situations, ultimately generating a future evolution map of the situation containing multiple possible paths and outcomes. This map is similar to a "map" that marks various safe or risky states that may be reached from the present. For example, a risk bifurcation node could be the decision point of "whether the reservoir water level should continue to rise or release water after exceeding the flood limit"; an oscillation node could be "the active power of the generating units fluctuating continuously within the range of 100MW to 120MW".

[0022] The risk identification module is used to determine the risk step points in the hydropower station based on the future evolution map of the situation, and to simultaneously visualize the future evolution map of the situation and the risk step points.

[0023] Finally, a deep scan of the future evolution map of the situation is performed to identify risk jump points where risk values ​​suddenly and abruptly increase (e.g., the critical point where the risk jumps instantly from a "warning" state to a "high-risk" state). The system synchronously and dynamically visualizes the future evolution map of the situation and all identified risk jump points through a monitoring screen or terminal interface. Operators can intuitively see the future risk evolution path and the most critical danger points, thus providing decision support for taking measures in advance. For example, a risk jump point could be "when the second derivative of the dam displacement exceeds a preset threshold, the system determines that the dam is about to enter an accelerated deformation stage," and this point is highlighted in red on the evolution map.

[0024] In this embodiment, the historical situation model library refers to a database that stores data related to past industrial safety incidents, including event records, processing procedures, and result information, for subsequent analysis and reference.

[0025] In this embodiment, historical security incident data refers to specific information about past security incidents, such as incident type, time, scope of impact, and response measures, which serves as the basis for analysis.

[0026] In this embodiment, the visual debriefing scenario refers to the process of reproducing historical security events through graphics or animation, which facilitates an intuitive understanding of the event's evolution and key points.

[0027] In this embodiment, the equipment status parameters refer to the vibration parameters of the turbine generator set (such as swing and vortex shell vibration), the displacement and deformation parameters of the dam (such as horizontal displacement of the dam crest and vertical displacement of the dam foundation), and the reservoir water level and seepage parameters (such as upstream water level and dam foundation seepage pressure) during the operation of the hydropower station equipment.

[0028] In this embodiment, environmental monitoring parameters refer to external condition data in the hydropower station, such as air quality, humidity, and concentration of toxic gases, which are used to assess environmental risks.

[0029] In this embodiment, the dynamic model of the state evolution of the hydropower station refers to a model built on mathematical and physical principles, which describes the dynamic change law of the industrial system under the influence of internal and external factors.

[0030] In this embodiment, risk bifurcation nodes refer to key decision points or turning points in the process of industrial situation evolution that may trigger different risk paths.

[0031] In this embodiment, the oscillation node refers to a segment in the situation that fluctuates frequently or is unstable, which can easily lead to increased risk or loss of control.

[0032] In this embodiment, the future evolution map refers to a chart or map generated through simulation, which shows the possible development trend and risk distribution of the hydropower station.

[0033] In this embodiment, the risk step point refers to the critical moment or position in the evolution of the situation where the risk level suddenly jumps, which requires special attention and intervention.

[0034] The working principle and beneficial effects of the above technical solution are as follows: By integrating historical safety event data and real-time industrial parameters, it enables intelligent analysis and online simulation of the hydropower station's status. Based on lessons learned from historical data, it generates visualized recap scenarios and constructs a dynamic model by combining real-time equipment status and environmental monitoring parameters to simulate the state evolution of the hydropower station. It can dynamically identify risk bifurcation nodes and oscillation nodes, predict the future evolution of the status through online reasoning, and ultimately determine the risk step point and visualize it in real time, thereby helping operators to provide early warnings and make decisions. It is conducive to improving the predictability and response efficiency of industrial safety management, reducing the probability of unexpected events, enhancing risk identification capabilities through data-driven analysis, optimizing emergency response processes, and supporting continuous improvement of safety strategies.

[0035] Example 2: Based on Example 1, this example provides an online simulation device for unconventional situations, such as... Figure 2 As shown, the cognitive reasoning module includes: The data extraction unit is used to extract historical safety event data of hydropower stations from the historical situation model library; The mining unit is used to perform event chain mining on historical safety event data of hydropower stations to identify multiple ordered sub-events corresponding to a safety event. The link construction unit is used to construct a complete evolution link of historical security events based on the event logic relationship of multiple ordered sub-events; The visualization retrospective scenario generation unit is used to generate visualization retrospective scenarios of historical security events based on the complete evolution chain.

[0036] In this embodiment, the event chain refers to a sequence of multiple sub-events arranged in chronological order and causal logic, which describes the complete development process of a single security event from its initial cause to its final result.

[0037] In this embodiment, ordered sub-events refer to the independent stages or links that constitute a complete security event. These links have a chronological order in terms of occurrence time and a logical causal relationship.

[0038] In this embodiment, the complete evolutionary link refers to the complete path model of a certain historical security event from its origin, development to its outcome, formed after identifying all ordered sub-events and clarifying their interrelationships.

[0039] The working principle and beneficial effects of the above technical solution are as follows: by deeply analyzing historical safety events, it is possible to transform scattered accident information into a complete event development chain with a clear structure and definite cause and effect, and reproduce it in an intuitive and visual form, which effectively improves the efficiency and depth of learning from historical events and enhances the systematicness and foresight of industrial safety management.

[0040] Example 3: Based on Example 2, this example provides an unconventional situation online simulation device and a mining unit, including: The attribute extraction subunit is used to extract the event type attribute and timestamp attribute of each event record in the historical security event data to generate a basic event set; The initial event sequence generation sub-unit is used to arrange the event records of the basic event set according to the timestamp attribute in a preset order to generate the initial event sequence; The event flow analysis subunit is used for: Perform event flow analysis on the initial event sequence, the event flow analysis including: calculating the time interval and state transition probability between adjacent events; The time interval difference is compared with a preset time threshold, and the state transition probability is compared with a preset probability threshold. When the time interval difference is greater than the preset time threshold, or when the state transition probability is less than the preset probability threshold, it is determined that there is a state transition point between the current adjacent events. The sub-unit is used to divide the initial event sequence into multiple consecutive sub-event sequence segments based on the state transition points. The type determination subunit is used to identify the event type in each sub-event sequence segment based on the event type attribute, and to determine the type of each sub-event in each sub-event sequence segment; Event relationship diagram construction sub-units are used for: The sub-event types are used as nodes, and temporal association edges are established between adjacent nodes according to the time order of the sub-event sequence segments. Traverse each sub-event sequence segment to determine the temporal order and co-occurrence frequency of event types in each sub-event sequence segment, and determine the causal logical relationship between sub-event types based on the temporal order and co-occurrence frequency; Based on the causal logical relationship between sub-event types, construct the causal logical edge of the corresponding node; An event relationship graph is constructed based on the nodes, temporal correlation edges, and causal logic edges of each sub-event sequence segment. The event chain corresponding to the starting point to the target endpoint is determined based on the event relationship graph, and multiple ordered sub-events are determined based on the event chain.

[0041] In this embodiment, the basic event set refers to the set of original event records extracted from historical data, where each event contains basic attributes such as type and occurrence time.

[0042] In this embodiment, the initial event sequence refers to a complete timeline sequence formed by arranging all records in the basic event set strictly in chronological order of occurrence.

[0043] In this embodiment, calculating the time interval and state transition probability between adjacent events includes: The state transition probability is calculated as follows: For any two adjacent event types A and B in the initial event sequence, count the number of times event type A is followed immediately by event type B in the entire sequence, denoted as N(A→B); Let N(A) be the total number of times event type A occurs in the entire sequence. Then the state transition probability P(B|A) = N(A→B) / N(A); When the event record is unstructured text, the corresponding event type is first extracted from each log text using a preset event type dictionary or a rule-based keyword matching method. Then, the event type sequence is generated by sorting by timestamp, and the probability is calculated.

[0044] In this embodiment, a state transition point refers to a key position in the initial event sequence that marks the end of one event development stage and the beginning of another new stage. It is determined by excessively large time intervals or excessively low state transition probabilities between adjacent events.

[0045] In this embodiment, a sub-event sequence segment refers to a number of shorter, closely related consecutive event segments obtained by cutting the long initial event sequence using state transition points.

[0046] In this embodiment, the event relationship graph refers to a network graph that uses sub-event types as nodes and edges to represent the temporal order and causal logic relationships between them, and is used to comprehensively describe the association structure of event elements.

[0047] In this embodiment, the temporal association edge refers to a connecting line in the event relationship graph that is only used to indicate that two sub-event types have a sequential relationship in terms of occurrence time.

[0048] In this embodiment, a causal logic edge refers to a causal connection line in an event relationship graph that indicates that one sub-event type is the cause of another sub-event type.

[0049] In this embodiment, the causal logical relationship between sub-event types is determined based on the chronological order and co-occurrence frequency, specifically as follows: For each sub-event sequence segment, count the number of times the two sub-event types appear sequentially and the conditional probability: if the probability of the second sub-event type appearing within a preset time window after the first sub-event type appears is higher than the first threshold, and the first sub-event type always takes precedence over the second sub-event type in time in all temporal relationships, then it is determined that there is a candidate causal logical relationship between the first sub-event type and the second sub-event type. For the identified candidate causal logical relationships, the stability and consistency in different sub-event sequence segments are further tested: when the candidate causal logical relationships of the first sub-event type and the second sub-event type are both valid in sub-event sequence segments exceeding the second preset proportion, and there is no reverse causal evidence, the current candidate causal logical relationship is confirmed as the final causal logical relationship.

[0050] The working principle and beneficial effects of the above technical solution are as follows: By automatically analyzing historical event records, it can transform complex and disordered raw data into a logically clear and structurally rigorous event development chain. Its core advantage lies in the fact that it not only arranges events according to chronological order, but also intelligently identifies key state transition nodes by quantitatively analyzing the time intervals and state transition possibilities between events, thereby accurately dividing the continuous event flow into sub-event sequences with different stage significance. Furthermore, it integrates temporal correlation and causal logic to construct an event relationship diagram that can reflect both the chronological order and reveal the internal causal driving relationship. This makes the event chain that is finally mined no longer a simple timeline, but a profound reflection of the internal dynamic mechanism of event evolution, greatly improving the depth and accuracy of the analysis of the root causes and development paths of historical security events, and providing a solid and reliable basis for subsequent situational analysis.

[0051] Example 4: Based on Example 1, this example provides an unconventional situation online simulation device and a dynamic modeling module, including: Device configuration unit, used for: The system obtains the collection requirements for different types of parameters based on the management terminal, analyzes the collection requirements, and determines the collection cycle and single collection quantity for different types of parameters. Based on the acquisition cycle and the amount of data acquired in a single acquisition, the parameters of each sensor node pre-deployed in the hydropower station are adapted. The data acquisition unit is used for: Based on the parameter adaptation results, each sensor node is controlled in parallel to collect equipment status parameters and environmental monitoring parameters in the hydropower station, and the collected equipment status parameters and environmental monitoring parameters are marked with source tags based on the location source of each sensor node. Upload the device status parameters and environmental monitoring parameters marked with the source to the data center.

[0052] In this embodiment, the management terminal refers to the control device or interface used to set and transmit data acquisition requirements.

[0053] In this embodiment, the acquisition requirements refer to the specific specifications for data acquisition, including instructions such as parameter categories and acquisition frequency.

[0054] In this embodiment, the acquisition period refers to the time interval between two data acquisition operations.

[0055] In this embodiment, the number of data points or parameters acquired in a single acquisition refers to the number of data points or parameters acquired in each acquisition process.

[0056] In this embodiment, a sensor node refers to a physical sensor device deployed in a hydropower station for real-time data acquisition.

[0057] In this embodiment, parameter adaptation refers to adjusting the configuration parameters of the sensor node according to the acquisition requirements to match the specified acquisition period and single acquisition amount.

[0058] In this embodiment, equipment status parameters refer to indicators that reflect the operating status of industrial equipment, such as temperature, vibration, or rotational speed.

[0059] In this embodiment, environmental monitoring parameters refer to indicators that reflect the environmental conditions around the hydropower station, such as humidity, dust concentration, or light intensity.

[0060] In this embodiment, the location source refers to the physical or logical location information of the sensor node, which is used to identify the source of data acquisition.

[0061] In this embodiment, source tagging refers to adding an identifier to the collected data to record the location information of the source sensor node.

[0062] In this embodiment, a data center refers to a central server or system used to store, process, and manage collected data.

[0063] The beneficial effects of the above technical solution are as follows: by dynamically analyzing the acquisition requirements, the appropriate acquisition cycle and single acquisition volume are automatically determined, thereby optimizing the configuration of sensor nodes and achieving efficient data acquisition. Secondly, the status and environmental parameters of each sensor node are acquired in parallel, and the source is identified according to the location, ensuring that the data is traceable and accurately uploaded to the data center. This improves the flexibility and efficiency of data acquisition, reduces resource waste, and provides timely and reliable data support for online simulation of unconventional situations in hydropower stations.

[0064] Example 5: Based on Example 1, this example provides an unconventional situation online simulation device and a dynamic modeling module, including: The hydropower station analysis unit is used to perform a global scan of the hydropower station and construct a dynamic state-space model of the hydropower station. The state-space model includes state equations and observation equations. The state equations characterize the dynamic law of the evolution of the internal state vector of the hydropower station over time, and the observation equations are used to establish the mapping relationship between the internal state vector of the hydropower station and the observation vectors directly measured by the sensors. Filter configuration unit, used for: The system obtains visualized retrospective scenarios of historical security events, extracts structured event data corresponding to the visualized retrospective scenarios, and extracts initial estimates and error covariance matrices of key state variables before and after the occurrence of historical security events based on the structured event data corresponding to the visualized retrospective scenarios. The filter is initialized and configured based on the initial estimates of the key state variables and the error covariance matrix. Dynamic model building unit, used for: Based on the initial configuration results, a recursive filtering estimation loop is performed on the dynamic state-space model of the hydropower station, specifically as follows: In each cycle, the state of the hydropower station at the current moment is predicted based on the state equation, and the equipment state parameters and environmental monitoring parameters of the hydropower station are collected after the prediction. The prediction results are corrected based on equipment status parameters and environmental monitoring parameters, and the optimal state prediction value of the hydropower station at the current moment is obtained based on the corrected results. A dynamic model of the state evolution of a hydropower station is obtained based on the recursive filtering estimation results.

[0065] In this embodiment, the state-space model of the hydropower station dynamics refers to a mathematical model used to describe the dynamic behavior of the state of the hydropower station changing over time, including state equations and observation equations.

[0066] In this embodiment, the state equation refers to the mathematical equation that describes the evolution of the internal state vector of the hydropower station over time, and is used to characterize the dynamic law of state change.

[0067] In this embodiment, the observation equation refers to the mathematical equation that establishes the mapping relationship between the internal state vector of the hydropower station and the observation vector directly measured by the sensor.

[0068] In this embodiment, a global scan of the hydropower station is performed to construct a dynamic state-space model of the hydropower station, specifically as follows: Define state vector Where: H(k) is the upstream water level of the reservoir (unit: m), reflecting the head change of the hydropower station; w(k) is the turbine generator speed (unit: rad / s), reflecting the unit's operating status; T represents the transpose; the observation vector is defined. ,in, This represents the measurement value of the water level sensor at time k; This represents the measurement value of the speed sensor at time k; k represents the sequence number of the time. The equation of state describes the changes in water level and rotational speed over time: ; ; in, This represents the upstream water level of the reservoir at time k+1; This represents the inflow rate at time k. Indicates outbound flow; This represents the time step from time k to time k+1; This represents the sum of water level changes caused by factors not considered in the deterministic water balance equation during the time interval from k to k+1 (i.e., process noise, where factors include, for example, wind and waves causing an instantaneous rise in water level of 0.5 cm, evaporation causing a drop in water level of 0.1 cm, etc.). This represents the combined contribution of all factors not considered in the qualitative torque balance equation to the change in unit speed from time k to time k+1 (i.e., process noise, including but not limited to: when the turbine is flowing, the low-frequency pressure pulsation caused by the draft tube vortex will generate additional random torque on the runner, resulting in slight fluctuations in speed). This represents the rotational speed of the hydro-generator unit at time k+1; This represents the hydraulic torque of the turbine at time k. This represents the generator torque at time k; Indicates the moment of inertia of the generator unit; The observation equation is: ; ; in, This indicates the noise level observed during water level monitoring. This indicates the noise level observed during rotational speed measurement.

[0069] In this embodiment, the state vector refers to a set of variables that represent the internal state of the hydropower station, such as equipment operating parameters or environmental conditions.

[0070] In this embodiment, the observation vector refers to a set of observation data obtained directly by the sensor, such as equipment status parameters or environmental monitoring parameters.

[0071] In this embodiment, the scenario of visually reviewing historical security events refers to the scenario of visually replaying and analyzing past security events to extract key information before and after the event.

[0072] In this embodiment, key state variables refer to state variables that play an important role in historical safety events, such as equipment failure indicators or environmental anomaly parameters.

[0073] In this embodiment, the initial estimate refers to the initial guess of the state variables during filter initialization, which is used to start the filtering process.

[0074] In this embodiment, the error covariance matrix refers to a matrix that represents the statistical characteristics of the state estimation error and is used to quantify the uncertainty of the estimation.

[0075] In this embodiment, the filter refers to an algorithm used to estimate state variables, such as a Kalman filter, which optimizes the state estimate through prediction and correction.

[0076] In this embodiment, the recursive filtering estimation loop refers to an iterative process that predicts the state in each cycle and makes corrections based on new observation data to gradually optimize the estimation.

[0077] In this embodiment, the optimal state prediction value refers to the most accurate state estimate obtained through the filtering process at the current moment, which is used to reflect the true state of the hydropower station.

[0078] In this embodiment, the dynamic model refers to a mathematical model that describes the state evolution of a hydropower station. It is constructed based on the filtering estimation results and is used to deduce changes in the situation.

[0079] The beneficial effects of the above technical solution are as follows: By constructing a dynamic state space model of the hydropower station through global scanning, the mapping relationship between the internal state evolution and sensor observations can be accurately described. Secondly, by extracting the initial values ​​of key states based on the visual review of historical safety events, the filter configuration can be optimized. Through recursive filtering estimation loop, the scenario state can be predicted and corrected in real time to obtain the optimal state prediction value, thereby establishing an accurate dynamic model, improving the accuracy and real-time performance of situation simulation, enhancing the early warning and response capabilities for unconventional events, providing reliable support for industrial safety decision-making, and reducing the risk of false alarms and missed alarms.

[0080] Example 6: Based on Example 1, this example provides an online simulation device for unconventional situations, including an online simulation module comprising: The simulation unit is used to perform numerical simulation of the state evolution process of a hydropower station based on a dynamic model, and to obtain dynamic change data of state parameters. The risk bifurcation node identification unit is used to identify risk bifurcation nodes in the hydropower station based on the dynamic change data of the state parameters and according to a preset method. The oscillation node identification unit is used to monitor the oscillation mode of the state variable in real time based on the dynamic change data of the state parameters, and to identify the corresponding oscillation node when the oscillation amplitude and oscillation frequency exceed the preset stability domain threshold. The reasoning and analysis unit is used to take the risk bifurcation node and oscillation node as key state points of the evolution path, and analyze the causal relationship and temporal sequence relationship between the key state points based on the multivariate reasoning engine to determine the evolution path of the deduced system state. The map generation unit is used to construct and output a future evolution map of the situation based on the evolution path.

[0081] In this embodiment, the dynamic model refers to a mathematical model that describes the evolution of state parameters in a hydropower station over time.

[0082] In this embodiment, a risk bifurcation node refers to a key decision point or state point that may lead the system to develop along different risk paths during the state evolution process.

[0083] In this embodiment, oscillation mode refers to a specific fluctuation pattern or vibration characteristic exhibited by a state variable during dynamic changes.

[0084] In this embodiment, the preset stability threshold refers to the pre-set boundary values ​​of oscillation amplitude and frequency used to determine whether the system state is within a stable range.

[0085] In this embodiment, the multivariate inference engine refers to an intelligent analysis algorithm or system that can simultaneously analyze the complex relationships between multiple state variables.

[0086] In this embodiment, the temporal sequence relationship refers to the order in which different security events or state changes occur on the timeline and their logical association.

[0087] In this embodiment, the future evolution map of the situation refers to a visual map that graphically displays the possible future development paths of the system state.

[0088] In this embodiment, the causal and temporal relationships between the key state points are analyzed based on a multivariate inference engine. The processing flow of the multivariate inference engine includes: The specific input to the multivariate inference engine is: a sequence of key state points, each key state point containing node type (risk bifurcation node or oscillation node), occurrence timestamp, associated state parameter name and value, and spatial location label of the node; and a list of the original chronological order of each key state point. The processing logic of the multivariate inference engine is as follows: First, the input key state points are sorted according to timestamps to form an initial timeline. Second, for every two adjacent key state points, the change trajectory of all state parameters between the adjacent key state points is extracted, and the direction of change (increasing, decreasing, or remaining unchanged) and magnitude of change of each parameter are determined. Then, for the same state parameter, if the change trend of the current parameter after a key state point is statistically consistent with the change trend before subsequent key state points (i.e., the same change pattern appears in multiple simulation cycles), then it is determined that there is a causal driving relationship between the two key state points. For the determination of the temporal sequence relationship, the order on the timeline is directly adopted, supplemented by the weighting of the time interval between each state point. Finally, all confirmed causal relationships and temporal sequence relationships are summarized to form a complete evolution path description from the initial key state point to the final key state point. The output format of the multivariate inference engine is: a directed structure description of an evolution path, which specifically includes: a list of nodes (each node contains all the information of the original key state point) and a list of edges (each edge contains the start node identifier, the end node identifier, the relationship type (causal relationship or temporal sequence), and the relationship strength value (a value between 0 and 1)).

[0089] The beneficial effects of the above technical solution are: by dynamically simulating and accurately capturing the state evolution process, intelligently identifying risk bifurcation points and abnormal oscillation nodes, determining the causal and temporal relationships between key state points, clearly deducing the evolution path of the system state, and generating an intuitive situation evolution map, the forward-looking nature of risk warning and the accuracy of situation assessment are significantly improved, providing key decision support for the proactive prevention and control of unconventional events.

[0090] Example 7: Based on Example 6, this example provides an unconventional situation online simulation device. In the graph generation unit, the evolution path is mapped into a directed graph structure, where nodes represent risk bifurcation nodes and oscillation nodes, and edges represent causal and temporal relationships between nodes. The directed graph structure is visualized and rendered using a graph drawing algorithm to generate the future evolution graph of the situation.

[0091] In this embodiment, a directed graph structure refers to a graphical model that uses arrowed lines to connect nodes to represent the directional relationships between elements.

[0092] In this embodiment, the graph drawing algorithm refers to a computer graphics algorithm used to automatically calculate node positions, lay out connections, and generate visual graphics.

[0093] In this embodiment, visualization rendering refers to the computational process of converting abstract data structures into graphical images that can be viewed intuitively.

[0094] The beneficial effects of the above technical solution are: it transforms complex evolution paths into intuitive directed graphs, clearly presents the causal and temporal relationships between key risk nodes through visualization rendering, and generates dynamic evolution maps that make potential risk paths and system evolution trends readily apparent, significantly improving the intuitiveness of situational awareness and decision-making efficiency, and providing key basis for risk warning and intervention path planning.

[0095] Example 8: Based on Example 1, this example provides an unconventional situation online simulation device and a risk identification module, including: The sequence extraction unit is used to extract continuous path sequences representing the evolution of the system state over time in the future situation evolution map; The sampling unit is used to sample the system state variables in the continuous path sequence and determine the target state value of the system state variables at discrete time points. The risk step point judgment unit is used to analyze system variables based on target state values ​​and determine whether there is a risk step point. The synchronous display unit is used to highlight the risk step point when there is a risk step point, and overlay and render it onto the corresponding position of the future evolution map of the situation to generate a fused view. At the same time, the fused view is sent to the visualization terminal for synchronous display.

[0096] In this embodiment, a continuous path sequence refers to a complete trajectory extracted from the future evolution map of the situation, which describes the continuous change of the system state from the current moment to a certain future moment.

[0097] In this embodiment, the target state value refers to the specific value of one or more key system state variables at a specific future point in time, obtained by discretely sampling a continuous path sequence.

[0098] In this embodiment, the fusion view refers to the process of overlaying and rendering the identified risk step points onto the original situational future evolution map in the form of highlighted markers, forming a comprehensive display interface that integrates future predictions and key risk information.

[0099] The working principle and beneficial effects of the above technical solution are as follows: By automatically analyzing the future evolution path, it can accurately capture the sudden risk of system state changes during continuous processes. Its core advantage lies in the fact that it does not statically assess risk, but dynamically scans the entire prediction timeline. Through dense sampling and quantitative analysis of the changing trends of state variables, it intelligently identifies critical moments when risks rise sharply. This method can effectively avoid the problem of missing key risk signals due to sparse data sampling, ensuring early detection of potential crises. Once a risk step-point is identified, it will be immediately highlighted in the global situation map. The intuitive fusion view greatly reduces the cognitive load of operators, enabling decision-makers to see at a glance when and where the risk will undergo a qualitative change, thus providing key information for taking precise and timely intervention measures and significantly improving the speed and effectiveness of emergency response.

[0100] Example 9: Based on Example 8, this example provides an unconventional situation online simulation device and a risk step point judgment unit, including: The analysis sub-unit is used to calculate the first derivative sequence of the system state variables using a numerical difference algorithm based on the target state value, and then calculate the second derivative sequence of the system state variables again using a numerical difference algorithm based on the first derivative sequence. The risk step point determination sub-unit is used to scan the second derivative sequence, identify local extreme points in the second derivative sequence where the absolute value of each value is greater than a preset mutation threshold, and use the local extreme points as risk step points.

[0101] In this embodiment, the first derivative sequence refers to the sequence formed by the rate of change of the system state variables over time.

[0102] In this embodiment, the second derivative sequence refers to the sequence formed by the rate of change (i.e., acceleration) of the system state variable itself.

[0103] In this embodiment, a local extremum point refers to a data point in the second derivative sequence where the absolute value of the value exhibits a local peak (maximum or minimum).

[0104] In this embodiment, the mutation threshold refers to a preset numerical threshold used to determine whether the change in the second derivative is drastic enough to indicate the occurrence of a risk step event.

[0105] The working principle and beneficial effects of the above technical solution are as follows: by analyzing the acceleration of the rate of change of the system state, it is possible to extremely keenly capture the critical point when the system state is about to undergo a drastic change or sudden change, and to accurately identify the early warning signal that the risk is not accumulated linearly, but suddenly and sharply increased, thereby achieving an early insight into potential crises and buying valuable time for intervention measures.

[0106] Example 10: This example provides a method for online simulation of unconventional situations in hydropower stations, such as... Figure 3 As shown, it includes: Step 1: Retrieve historical security event data from the built-in historical situation model library, and perform intelligent analysis on the historical security event data to generate a visual review scenario of the historical security events; Step 2: Collect equipment status parameters and environmental monitoring parameters in the hydropower station, and perform integrated analysis on the visualized review scenario and equipment status parameters and environmental monitoring parameters to construct a dynamic model of the hydropower station's state evolution; Step 3: Based on the dynamic model, determine the risk bifurcation nodes and oscillation nodes of the hydropower station, and perform online reasoning analysis on the risk bifurcation nodes and oscillation nodes of the hydropower station to obtain the future evolution map of the situation; Step 4: Based on the future evolution map of the situation, identify the risk step points in the hydropower station, and simultaneously visualize the future evolution map of the situation and the risk step points.

[0107] The working principle and beneficial effects of the above technical solution are as follows: By integrating historical safety event data and real-time industrial parameters, it enables intelligent analysis and online simulation of the hydropower station's status. Based on lessons learned from historical data, it generates visualized recap scenarios and constructs a dynamic model by combining real-time equipment status and environmental monitoring parameters to simulate the state evolution of the hydropower station. It can dynamically identify risk bifurcation nodes and oscillation nodes, predict the future evolution of the status through online reasoning, and ultimately determine the risk step point and visualize it in real time, thereby helping operators to provide early warnings and make decisions. It is conducive to improving the predictability and response efficiency of industrial safety management, reducing the probability of unexpected events, enhancing risk identification capabilities through data-driven analysis, optimizing emergency response processes, and supporting continuous improvement of safety strategies.

[0108] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An online simulation device for unconventional situations, characterized in that, include: The cognitive reasoning module is used to retrieve historical safety event data of hydropower stations from the built-in historical situation model library, and to perform intelligent analysis on the historical safety event data to generate a visual review scenario of the historical safety events. The dynamic modeling module is used to collect equipment status parameters and environmental monitoring parameters in hydropower stations, and to integrate and analyze the visualized replay scenarios with the equipment status parameters and environmental monitoring parameters to construct a dynamic model of the state evolution of hydropower stations. The online simulation module is used to determine the risk bifurcation nodes and oscillation nodes of the hydropower station based on the dynamic model, and to perform online reasoning analysis on the risk bifurcation nodes and oscillation nodes of the hydropower station to obtain the future evolution map of the situation. The risk identification module is used to determine the risk step points in the hydropower station based on the future evolution map of the situation, and to simultaneously visualize the future evolution map of the situation and the risk step points.

2. The unconventional situation online simulation device according to claim 1, characterized in that, The cognitive reasoning module includes: The data extraction unit is used to extract historical safety event data of hydropower stations from the historical situation model library; The mining unit is used to perform event chain mining on historical safety event data of hydropower stations to identify multiple ordered sub-events corresponding to a safety event. The link construction unit is used to construct a complete evolution link of historical security events based on the event logic relationship of multiple ordered sub-events; The visualization retrospective scenario generation unit is used to generate visualization retrospective scenarios of historical security events based on the complete evolution chain.

3. The unconventional situation online simulation device according to claim 2, characterized in that, The excavation unit includes: The attribute extraction subunit is used to extract the event type attribute and timestamp attribute of each event record in the historical security event data to generate a basic event set; The initial event sequence generation sub-unit is used to arrange the event records of the basic event set according to the timestamp attribute in a preset order to generate the initial event sequence; The event flow analysis subunit is used for: Perform event flow analysis on the initial event sequence, the event flow analysis including: calculating the time interval and state transition probability between adjacent events; The time interval difference is compared with a preset time threshold, and the state transition probability is compared with a preset probability threshold. When the time interval difference is greater than the preset time threshold, or when the state transition probability is less than the preset probability threshold, it is determined that there is a state transition point between the current adjacent events. The sub-unit is used to divide the initial event sequence into multiple consecutive sub-event sequence segments based on the state transition points. The type determination subunit is used to identify the event type in each sub-event sequence segment based on the event type attribute, and to determine the type of each sub-event in each sub-event sequence segment; Event relationship diagram construction sub-units are used for: The sub-event types are used as nodes, and temporal association edges are established between adjacent nodes according to the time order of the sub-event sequence segments. Traverse each sub-event sequence segment to determine the temporal order and co-occurrence frequency of event types in each sub-event sequence segment, and determine the causal logical relationship between sub-event types based on the temporal order and co-occurrence frequency; Based on the causal logical relationship between sub-event types, construct the causal logical edge of the corresponding node; An event relationship graph is constructed based on the nodes, temporal correlation edges, and causal logic edges of each sub-event sequence segment. The event chain corresponding to the starting point to the target endpoint is determined based on the event relationship graph, and multiple ordered sub-events are determined based on the event chain.

4. The unconventional situation online simulation device according to claim 1, characterized in that, The dynamic modeling module includes: Device configuration unit, used for: The system obtains the collection requirements for different types of parameters based on the management terminal, analyzes the collection requirements, and determines the collection cycle and single collection quantity for different types of parameters. Based on the acquisition cycle and the amount of data acquired in a single acquisition, the parameters of each sensor node pre-deployed in the hydropower station are adapted. The data acquisition unit is used for: Based on the parameter adaptation results, each sensor node is controlled in parallel to collect equipment status parameters and environmental monitoring parameters in the hydropower station, and the collected equipment status parameters and environmental monitoring parameters are marked with source tags based on the location source of each sensor node. Upload the device status parameters and environmental monitoring parameters marked with the source to the data center.

5. The unconventional situation online simulation device according to claim 1, characterized in that, The dynamic modeling module includes: The hydropower station analysis unit is used to perform a global scan of the hydropower station and construct a dynamic state-space model of the hydropower station. The state-space model includes state equations and observation equations. The state equations characterize the dynamic law of the evolution of the internal state vector of the hydropower station over time, and the observation equations are used to establish the mapping relationship between the internal state vector of the hydropower station and the observation vectors directly measured by the sensors. Filter configuration unit, used for: The system obtains visualized retrospective scenarios of historical security events, extracts structured event data corresponding to the visualized retrospective scenarios, and extracts initial estimates and error covariance matrices of key state variables before and after the occurrence of historical security events based on the structured event data corresponding to the visualized retrospective scenarios. The filter is initialized and configured based on the initial estimates of the key state variables and the error covariance matrix. Dynamic model building unit, used for: Based on the initial configuration results, a recursive filtering estimation loop is performed on the dynamic state-space model of the hydropower station, specifically as follows: In each cycle, the state of the hydropower station at the current moment is predicted based on the state equation, and the equipment state parameters and environmental monitoring parameters of the hydropower station are collected after the prediction. The prediction results are corrected based on equipment status parameters and environmental monitoring parameters, and the optimal state prediction value of the hydropower station at the current moment is obtained based on the corrected results. A dynamic model of the state evolution of a hydropower station is obtained based on the recursive filtering estimation results.

6. The unconventional situation online simulation device according to claim 1, characterized in that, The online simulation module includes: The simulation unit is used to perform numerical simulation of the state evolution process of a hydropower station based on a dynamic model, and to obtain dynamic change data of state parameters. The risk bifurcation node identification unit is used to identify risk bifurcation nodes in the hydropower station based on the dynamic change data of the state parameters and according to a preset method. The oscillation node identification unit is used to monitor the oscillation mode of the state variable in real time based on the dynamic change data of the state parameters, and to identify the corresponding oscillation node when the oscillation amplitude and oscillation frequency exceed the preset stability domain threshold. The reasoning and analysis unit is used to take the risk bifurcation node and oscillation node as key state points of the evolution path, and analyze the causal relationship and temporal sequence relationship between the key state points based on the multivariate reasoning engine to determine the evolution path of the deduced system state. The map generation unit is used to construct and output a future evolution map of the situation based on the evolution path.

7. The unconventional situation online simulation device according to claim 6, characterized in that, In the graph generation unit, the evolution path is mapped into a directed graph structure, where nodes represent risk bifurcation nodes and oscillation nodes, and edges represent causal and temporal relationships between nodes. The directed graph structure is visualized and rendered using a graph drawing algorithm to generate the future evolution graph of the situation.

8. The unconventional situation online simulation device according to claim 1, characterized in that, The risk identification module includes: The sequence extraction unit is used to extract continuous path sequences representing the evolution of the system state over time in the future situation evolution map; The sampling unit is used to sample the system state variables in the continuous path sequence and determine the target state value of the system state variables at discrete time points. The risk step point judgment unit is used to analyze system variables based on target state values ​​and determine whether there is a risk step point. The synchronous display unit is used to highlight the risk step point when there is a risk step point, and overlay and render it onto the corresponding position of the future evolution map of the situation to generate a fused view. At the same time, the fused view is sent to the visualization terminal for synchronous display.

9. The unconventional situation online simulation device according to claim 8, characterized in that, The risk step-point judgment unit includes: The analysis sub-unit is used to calculate the first derivative sequence of the system state variables using a numerical difference algorithm based on the target state value, and then calculate the second derivative sequence of the system state variables again using a numerical difference algorithm based on the first derivative sequence. The risk step point determination sub-unit is used to scan the second derivative sequence, identify local extreme points in the second derivative sequence where the absolute value of each value is greater than a preset mutation threshold, and use the local extreme points as risk step points.

10. A method for online simulation of unconventional situations in hydropower stations, characterized in that, include: Step 1: Retrieve historical security event data from the built-in historical situation model library, and perform intelligent analysis on the historical security event data to generate a visual review scenario of the historical security events; Step 2: Collect equipment status parameters and environmental monitoring parameters in the hydropower station, and perform integrated analysis on the visualized review scenario and equipment status parameters and environmental monitoring parameters to construct a dynamic model of the hydropower station's state evolution; Step 3: Based on the dynamic model, determine the risk bifurcation nodes and oscillation nodes of the hydropower station, and perform online reasoning analysis on the risk bifurcation nodes and oscillation nodes of the hydropower station to obtain the future evolution map of the situation; Step 4: Based on the future evolution map of the situation, identify the risk step points in the hydropower station, and simultaneously visualize the future evolution map of the situation and the risk step points.

Images