Emergency disposal measure chain coupling analysis system

Abstract

The invention provides an emergency disposal measure chain coupling analysis system, and the system is characterized in that a disaster model service chain construction and management module is used for collecting various disaster models, packaging the disaster models, and storing the packaged disaster models in a plurality of storage nodes of a model storage cluster in a scattered manner; the knowledge base construction and accident evolution diagram generation module is used for generating an accident evolution diagram based on a pre-constructed knowledge base; the accident evolution diagram is used for representing a plurality of accident evolution processes with incidence relation and combined action; the real-time monitoring and dynamic deduction module is used for collecting various data of emergency scenes in the power industry in real time, and constructing a dynamic deduction model in combination with the accident evolution diagram and output provided by various disaster models corresponding to accidents; and the decision suggestion module is used for simulating the dynamic deduction model by using the domino effect and determining the optimal emergency decision suggestion information of the actual scene according to the simulation result. According to the invention, various associated disasters can be comprehensively processed through multi-module cooperation, and complex disasters and emergency situations can be handled.

Description

Technical Field

[0001] The present invention belongs to the technical field of power emergency disposal, and in particular relates to an emergency disposal measure chain coupling analysis system. Background Art

[0002] Currently, existing emergency response systems often focus on a single type of emergency scenario. For example, fire emergency response systems focus on alarming, extinguishing, and evacuating fires; earthquake emergency response systems primarily address earthquake monitoring and post-earthquake rescue resource deployment. These emergency response systems have indeed played a significant role in their respective specific emergency scenarios, effectively responding to relatively independent emergencies.

[0003] However, real-world emergency situations are far more complex than a single disaster. Numerous disasters or emergencies are closely interrelated and interact synergistically, forming complex disaster chains. For example, earthquakes often trigger fires and landslides, and floods can easily lead to mudslides. These secondary disasters can further exacerbate the damage, posing even greater threats to infrastructure and human safety. Relying solely on a single emergency response system will inevitably be significantly less effective in these complex, multifaceted emergency scenarios. Emergency response measures for different disaster types lack coordination and integration, making it impossible to effectively interrupt and address the disaster chain as a whole. Therefore, an emergency response system is urgently needed. Summary of the Invention

[0004] In view of this, the present invention aims to provide an emergency response measure chain coupling analysis system, which can propose corresponding emergency measures for a variety of emergency scenarios with correlation and common effects, coordinate emergency resources for different types of disasters, and use the correlation between various types of disasters to conduct comprehensive analysis and decision-making, so as to realize the organic integration of multi-scenario emergency measures, thereby significantly improving the effectiveness of emergency management and minimizing disaster losses.

[0005] In order to achieve the above object, the technical solution provided by the present invention is as follows:

[0006] An emergency response measure chain coupling analysis system, comprising:

[0007] Disaster model service chain construction and management module, knowledge base construction and accident evolution diagram generation module, real-time monitoring and dynamic deduction module, and decision-making recommendation module;

[0008] The disaster model service chain construction and management module is used to collect various disaster models, encapsulate the disaster models according to the unified model encapsulation specification, and store them in multiple storage nodes of the model storage cluster;

[0009] The knowledge base construction and accident evolution diagram generation module is used to generate an accident evolution diagram based on the pre-built knowledge base; the accident evolution diagram is used to represent the evolution process of multiple accidents with correlation and joint effects;

[0010] The real-time monitoring and dynamic deduction module is used to collect various data on power industry emergency scenarios in real time. Based on the collected data, combined with the accident evolution diagram and the output provided by various disaster models corresponding to the accident, a dynamic deduction model is constructed;

[0011] The decision-making recommendation module is used to simulate the dynamic deduction model using the domino effect, and determine the optimal emergency decision-making recommendation information for the actual scenario based on the simulation results.

[0012] Furthermore, the disaster model service chain construction and management module includes:

[0013] Model collection unit, used to collect various disaster models and encapsulate them according to unified model encapsulation specifications;

[0014] A distributed storage unit is used to construct a model storage cluster using a distributed file system and to store the encapsulated disaster model files in multiple storage nodes in a dispersed manner;

[0015] The orchestration rule unit is configured with a rule-based model orchestration call engine and an orchestration call rule library. It is used to trigger the model orchestration call engine after receiving disaster warning or power facility failure information, and call the disaster model based on the orchestration call rule library.

[0016] Furthermore, in the orchestration rule unit, the process of formulating the orchestration call rule base includes:

[0017] Conduct demand analysis on emergency scenarios in the power industry, determine emergency scenario information and disaster model requirements for each emergency scenario;

[0018] Based on the results of the demand analysis, determine the rule elements and rule logic; the rule elements are used to specify the calling conditions of various disaster models; the rule logic is used to specify the calling order of various disaster models;

[0019] Convert rule elements and rule logic into rule text, organize and test the rule text, and form an orchestration call rule library based on the rule text that passes the test.

[0020] Furthermore, the knowledge base construction and accident evolution graph generation modules include:

[0021] A knowledge base construction unit is used to construct and store a knowledge base based on the relevant knowledge of emergency repair in the power industry;

[0022] The accident evolution graph generation unit is used to generate the accident evolution graph based on the data in the knowledge base and continuously update the accident evolution graph.

[0023] Furthermore, in the knowledge base construction unit, the relevant knowledge of emergency repair in the power industry is classified, and each category is used as a node. After associating each node, the knowledge base is constructed and stored.

[0024] Furthermore, the real-time monitoring and dynamic deduction module includes:

[0025] A data acquisition unit, used to collect real-time data on power industry emergency scenarios and information fed back by emergency personnel through a sensor network;

[0026] A data processing unit, used for fusing the collected multi-source data based on a data fusion algorithm;

[0027] The dynamic deduction model construction unit is used to construct a dynamic deduction model based on real-time fusion data, combined with the accident evolution diagram and the output provided by various disaster models corresponding to the accident, using a time series-based prediction algorithm.

[0028] Furthermore, if the time series-based forecasting algorithm is a seasonal autoregressive integrated moving average algorithm, a dynamic deduction model is constructed using the time series-based forecasting algorithm, including:

[0029] Receive real-time data collected by the data acquisition unit and obtain the accident evolution diagram and the results provided by various disaster models corresponding to the accident;

[0030] Preprocess the collected data and divide the preprocessed time series data into training set and test set;

[0031] The SARIMA model is fitted using the training set data, and the model parameters are estimated using the maximum likelihood estimation method;

[0032] Convert the accident evolution graph and the results provided by various disaster models corresponding to the accident into features that can be used as model input, and fuse these features with time series data as extended input features to train the SARIMA model;

[0033] The trained model is evaluated using the test set data, and the final model is used as the dynamic deduction model.

[0034] Furthermore, the data processing unit further includes:

[0035] Regular data backup subunit, used to regularly back up important data to off-site storage devices or cloud storage services;

[0036] The data recovery subunit is used to restore the data required for system operation from the backup data in the event of data loss or damage.

[0037] Furthermore, the decision suggestion module includes:

[0038] The chain-breaking decision model unit is used to identify key nodes in the accident evolution graph using a key node identification algorithm, simulate the domino effect propagation process in combination with the disaster model, and evaluate the severity of the domino effect;

[0039] The emergency resource allocation decision unit is used to allocate relevant emergency resources according to the output results of the chain-breaking decision model.

[0040] Furthermore, the execution process of the chain-breaking decision model unit includes:

[0041] Collect node data of the accident evolution diagram and data required by the disaster model, and pre-process the collected data;

[0042] Based on the degree centrality algorithm, the node data of the pre-processed accident evolution graph is used as input to calculate the centrality index of each node, sort the nodes according to the index value, and select the nodes with index values greater than the set value as key nodes;

[0043] According to different disaster types and characteristic parameters, a disaster impact model on nodes is established based on the pre-processed disaster model data;

[0044] Based on the physical characteristics and operating rules of the power system, define the propagation rules of the domino effect and determine the propagation conditions and probability of accidents;

[0045] Initialize the simulation, set the initial conditions for the disaster, including the disaster type, characteristic parameters and affected nodes, and initialize the states of all nodes in the accident evolution diagram;

[0046] Iterative propagation simulation is performed based on the influence model. In each time step, the state of each node and the state of adjacent nodes are checked according to the domino effect propagation rule. For nodes that meet the propagation conditions, the states of their adjacent nodes are updated according to the corresponding probability, and the propagation process is recorded. The above steps are repeated until no new node states change or the preset simulation time is reached.

[0047] Calculate predefined evaluation index values based on simulation results;

[0048] Analyze the role of key nodes in the domino effect propagation process based on the evaluation index value, and evaluate the impact of key node failure on the system;

[0049] Determine the chain-breaking decision information based on the impact assessment results.

[0050] In summary, the present invention provides an emergency response measures chain coupling analysis system, including a disaster model service chain construction and management module, a knowledge base construction and accident evolution graph generation module, a real-time monitoring and dynamic deduction module, and a decision recommendation module; the disaster model service chain construction and management module is used to collect various disaster models, and after encapsulating the disaster models according to a unified model encapsulation specification, the disaster models are dispersedly stored in multiple storage nodes of the model storage cluster; the knowledge base construction and accident evolution graph generation module is used to generate an accident evolution graph based on a pre-built knowledge base; the accident evolution graph is used to represent the evolution process of multiple accidents with related relationships and joint effects; the real-time monitoring and dynamic deduction module is used to collect various data of emergency scenarios in the power industry in real time, and based on the collected data, combined with the accident evolution graph and the output provided by various disaster models corresponding to the accident, a dynamic deduction model is constructed; the decision recommendation module is used to simulate the dynamic deduction model using the domino effect, and determine the optimal emergency decision recommendation information for the actual scenario based on the simulation results. The emergency response measures chain coupling analysis system of the present invention can cope with complex composite emergency scenarios, change the situation where a single emergency response system has low utility in the face of complex disaster chains, and can comprehensively handle multiple related disasters through multi-module collaboration. At the same time, it generates accident evolution diagrams, builds dynamic deduction models and conducts simulations to determine the optimal emergency decision-making recommendation information for actual scenarios, effectively blocking the disaster chain and responding to complex disasters and emergencies as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 This is a system architecture diagram of an emergency response measure chain coupling analysis system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the purposes, features, and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0054] See also Figure 1 , an embodiment of the present invention provides an emergency response measure chain coupling analysis system, comprising:

[0055] Disaster model service chain construction and management module, knowledge base construction and accident evolution diagram generation module, real-time monitoring and dynamic deduction module, and decision-making recommendation module;

[0056] The disaster model service chain construction and management module is used to collect various disaster models, encapsulate the disaster models according to the unified model encapsulation specification, and store them in multiple storage nodes of the model storage cluster;

[0057] The knowledge base construction and accident evolution diagram generation module is used to generate an accident evolution diagram based on the pre-built knowledge base; the accident evolution diagram is used to represent the evolution process of multiple accidents with correlation and joint effects;

[0058] The real-time monitoring and dynamic deduction module is used to collect various data on power industry emergency scenarios in real time. Based on the collected data, combined with the accident evolution diagram and the output provided by various disaster models corresponding to the accident, a dynamic deduction model is constructed;

[0059] The decision-making recommendation module is used to simulate the dynamic deduction model using the domino effect, and determine the optimal emergency decision-making recommendation information for the actual scenario based on the simulation results.

[0060] In this embodiment, the disaster model service chain construction and management module collects various disaster models, packages them according to unified specifications, and then stores them in a decentralized manner. This allows the system to include models for a wide range of different disaster types, overcoming the limitations of single emergency response systems that focus on a single disaster type. For example, the system includes models for earthquakes, fires, landslides, and other disasters, allowing the system to comprehensively consider the relationships between different disasters.

[0061] The knowledge base construction and accident evolution diagram generation module generates an accident evolution diagram based on a pre-built knowledge base. This diagram can represent the evolution of multiple related and interrelated accidents. Through this diagram, the system can clearly display the development of disaster chains, such as the specific evolutionary path of earthquake-induced fires and landslides. This allows emergency response to no longer view a single disaster in isolation, thereby improving the ability to respond to complex and complex emergency scenarios.

[0062] In addition, the real-time monitoring and dynamic deduction module of this embodiment collects various types of data of emergency scenarios in the power industry in real time, and constructs a dynamic deduction model in combination with the accident evolution diagram and the output of the disaster model. This means that the system can integrate and dynamically analyze information on different types of disasters according to actual conditions. For example, when a flood occurs, it combines the flood model and the mudslide model that may be triggered, and deduces based on the real-time monitoring data, so that emergency measures for different types of disasters can work together in this dynamic model. At the same time, the decision-making recommendation module uses the domino effect to simulate the dynamic deduction model, and determines the optimal emergency decision-making recommendation information for the actual scenario based on the simulation results. This simulation-based decision-making method can comprehensively consider the mutual influence of each link in the disaster chain, formulate emergency strategies as a whole, effectively block the disaster chain, achieve the coordination and connection of emergency measures for different types of disasters, and comprehensively respond to complex disasters and emergencies.

[0063] In one embodiment, the disaster model service chain construction and management module includes:

[0064] The model collection unit is used to collect various disaster models and encapsulate the disaster models according to the unified model packaging specifications.

[0065] For disaster models in different fields, we leverage containerization technologies (such as Docker) to package the models and their dependent operating environments into independent container images. We also standardize the model's input and output interfaces, clarifying the meaning and value range of input parameters, as well as fixed output fields, to ensure the model's portability and compatibility across different computing environments. This approach overcomes the limitations of single emergency response systems, which typically target only a single type of disaster. Instead, the system encompasses models for multiple different types of disasters, comprehensively considering the interrelationships between them and laying the foundation for subsequent responses to complex, complex emergency scenarios.

[0066] The distributed storage unit is used to build a model storage cluster using a distributed file system and store the encapsulated disaster model files in multiple storage nodes in a dispersed manner.

[0067] The distributed storage unit uses a distributed file system to construct a model storage cluster, distributing the encapsulated disaster model files across multiple storage nodes. Specifically, the Ceph distributed file system is employed, utilizing distributed hash table (DHT) technology to rapidly locate and retrieve model files, improving storage and access efficiency. Furthermore, a version management mechanism is established for each disaster model, recording different model versions. When a model is optimized or updated, the system automatically generates a new version and saves historical versions for retrospective use, ensuring the stability and reliability of model applications. This storage method enables the orderly management and efficient access of a large number of disaster models, meeting the diverse model requirements in complex emergency scenarios.

[0068] The orchestration rule unit is configured with a rule-based model orchestration call engine and an orchestration call rule library. It is used to trigger the model orchestration call engine after receiving disaster warning or power facility failure information, and call the disaster model based on the orchestration call rule library.

[0069] The orchestration call rule base is developed based on multiple factors, including the combination of disaster types, the operating status of power facilities, and historical emergency response experience. For example, when a typhoon and heavy rain disaster are detected, the orchestration call engine automatically matches and calls a model for typhoon wind impacts on power lines, a model for heavy rain-induced flooding of power facilities, and a model for predicting line short-circuit faults, and determines the order in which each model is called and the data exchange process. This intelligent orchestration call mechanism enables the system to coordinate the call of multiple related disaster models based on the actual disaster situation, achieving coordination and integration of emergency measures for different disaster types at the model level, analyzing and responding to the disaster chain as a whole, and improving the effectiveness of emergency response.

[0070] In one embodiment, in the orchestration rule unit, the process of formulating the orchestration call rule base includes:

[0071] S11: Conduct demand analysis on emergency scenarios in the power industry to determine emergency scenario information and disaster model requirements for each emergency scenario.

[0072] Conducting a demand analysis of emergency scenarios in the power industry aims to determine emergency scenario information and the disaster model requirements for each scenario. This requires in-depth research on various disasters that affect power facilities, including analysis of disaster characteristics, occurrence patterns, and damage to power facilities, as well as research on the structure of the power system. For example, in a typhoon disaster emergency scenario, it is necessary to clarify information such as the typhoon's wind force level and movement path, as well as matching disaster model requirements such as models for the impact of typhoon wind force on power lines and models for the impact of typhoon-induced secondary disasters on power facilities. This avoids the limitations of a single emergency response system in complex disaster scenarios and provides direction for subsequent coordinated responses to multiple disasters.

[0073] S12: Based on the results determined by the demand analysis, determine the rule elements and rule logic; the rule elements are used to specify the calling conditions of various disaster models; the rule logic is used to specify the calling order of various disaster models.

[0074] Based on the results of the demand analysis, the rule elements and rule logic are determined. The rule elements specify the conditions for calling the disaster model, and the rule logic specifies the order of calling. The formulation of the rules fully refers to multiple factors, such as the combination of disaster types, the operating status of power facilities, and historical emergency response experience. Taking the concurrent disaster of typhoons and heavy rain as an example, the rule elements will set trigger conditions such as typhoon wind force level and heavy rain rainfall. The rule logic clearly states that the typhoon wind force affecting the power line model will be called first, followed by the heavy rain-induced flooding of power facilities model and the line short-circuit fault prediction model. This ensures that the various disaster models operate in sequence and in coordination in complex disaster chain scenarios, realizing the effective connection of emergency measures for different disaster types.

[0075] S13: Convert the rule elements and rule logic into rule texts, organize and test the rule texts, and form an orchestration call rule library based on the rule texts that pass the test.

[0076] By testing the rule base using historical data, the correctness and effectiveness of the rules are verified, and the rules are optimized and adjusted based on the test results. For example, based on the typhoon and rainstorm coupling rules, the trigger conditions, model selection, and call sequence set in the rule text are tested to ensure that in actual emergency scenarios, when the system calls the disaster model based on the rule base, it can effectively block and respond to the disaster chain as a whole, providing a reliable basis for emergency decision-making.

[0077] In one embodiment, the knowledge base construction and accident evolution graph generation module includes:

[0078] The knowledge base construction unit is used to construct and store the knowledge base based on the relevant knowledge of emergency repair in the power industry.

[0079] The relevant knowledge of the power industry covers a wide range, including actual operation data of the power system (such as equipment information, operation records, maintenance history, fault reports, etc. of power facilities), historical event data, industry standards and specifications (such as installation, maintenance and inspection standards of power equipment), and disaster warning information (such as information on natural disasters such as earthquakes, floods, and typhoons). This knowledge is comprehensively sorted out and classified, with each category as a node (such as power facilities, personnel, events, locations, etc.), and associations between different nodes are established (such as "connected," "affected," "responsible," etc.). A knowledge base is constructed and stored through node association. This method integrates a large amount of information related to emergency repairs in the power industry, changing the previous problems of scattered information and lack of systematicity. It provides a rich and orderly knowledge resource for the subsequent generation of accident evolution diagrams and emergency decision-making, which helps to deal with complex disaster chain scenarios.

[0080] The accident evolution graph generation unit is used to generate the accident evolution graph based on the data in the knowledge base and continuously update the accident evolution graph.

[0081] When a power system failure occurs or a disaster warning is received, the system uses the affected power facility as the initial node and searches the knowledge base for all relationships associated with that node. Based on these relationships, the system then expands outward to find other related nodes. Based on historical accident cases and disaster impact models, the system determines the likely fault type of the associated nodes in the current disaster scenario and adds these nodes to the accident evolution graph. Simultaneously, a rule-based inference engine (such as Drools) is used to determine the influence weights and propagation directions between nodes, based on emergency repair technical specifications and expert experience. Combining real-time monitoring data with analysis results, the accident evolution graph is dynamically updated and visualized to present the accident development process. In this way, the accident evolution graph can reflect the development of the disaster chain in power industry emergency scenarios in real time, providing decision makers with a clear understanding of the accident situation. This allows emergency response to move beyond the isolation of a single disaster type and instead analyze and respond to the disaster chain holistically. This effectively addresses the lack of coordination and integration of emergency measures for different disaster types in existing technologies.

[0082] In one embodiment, in the knowledge base construction unit, the knowledge related to emergency repair in the power industry is classified, each category is used as a node, and the knowledge base is constructed and stored after associating the nodes.

[0083] In this embodiment, the knowledge base construction unit categorizes the relevant knowledge for emergency repairs in the power industry, with each category serving as a node. The relevant knowledge for the power industry includes data on the actual operation of the power system (such as equipment information and operation records of power facilities), historical event data, industry standards and specifications, disaster warning information, and the like. This knowledge is divided into nodes by category, such as power facilities (towers, transformers, lines, etc.), personnel (repair personnel, dispatchers, etc.), events (fault events, disaster events, etc.), and locations (substation locations, line directions, etc.). This makes the previously complex and disordered knowledge organized, facilitating subsequent management and use. This organized knowledge structure changes the previous situation where emergency knowledge was dispersed, difficult to find, and difficult to utilize, providing a solid knowledge foundation for the system to respond to complex emergency scenarios and avoiding the problem of low effectiveness caused by insufficient knowledge reserves in a single emergency response system.

[0084] After associating each node, the knowledge base is constructed and stored by establishing relationships between different nodes (such as "connected," "affected," and "responsible," etc.), organically connecting different categories of knowledge nodes to form a complete knowledge network. For example, the "connected" relationship between transmission lines and towers, and the "affected" relationship of disasters on power facilities, make the knowledge in the knowledge base no longer isolated, but an interconnected and mutually influential whole. This knowledge collaboration method can better reflect the inherent connections between various factors in the disaster chain when generating accident evolution diagrams, helping to achieve coordinated and integrated emergency response measures for different disaster types, effectively blocking and responding to the disaster chain as a whole, and solving the problem of lack of coordination among emergency response measures for different disaster types in existing technologies.

[0085] In one embodiment, the real-time monitoring and dynamic deduction module includes:

[0086] The data acquisition unit is used to collect data on emergency scenarios in the power industry and information fed back by emergency personnel in real time through a sensor network.

[0087] In power industry emergency scenarios, a variety of sensors are deployed at key locations such as towers, transmission lines, and substations, depending on facility type and monitoring needs. For example, vibration sensors are deployed on towers to monitor stability, while temperature and sag sensors are installed on transmission lines to monitor operational status. Meteorological sensors are also used to collect environmental data. Furthermore, mobile applications or mini-programs are developed to provide emergency personnel with real-time feedback on the progress of on-site emergency response and resource usage. This multi-channel data collection approach overcomes the limitations of a single data source and can capture all kinds of dynamic information in power emergency scenarios in real time and comprehensively. This provides a rich and accurate data foundation for subsequent emergency decision-making, effectively responding to complex and changing disaster chain scenarios.

[0088] The data processing unit is used to perform fusion processing on the collected multi-source data based on the data fusion algorithm.

[0089] Faced with multi-source data from sensors, video surveillance, emergency personnel feedback, etc., they are first extracted, cleaned, converted and loaded, and then the same physical quantity data collected by different types of sensors are fused through data fusion algorithms (such as those based on Kalman filtering algorithms).

[0090] The dynamic deduction model construction unit is used to construct a dynamic deduction model based on real-time fusion data, combined with the accident evolution diagram and the output provided by various disaster models corresponding to the accident, using a time series-based prediction algorithm.

[0091] This unit can use time series forecasting algorithms, such as the Seasonal Autoregressive Integrated Moving Average (SARIMA) model, to predict the likely evolutionary path of accidents over the next period of time based on the current operating status of power facilities and disaster development trends. For example, based on the current rate of flood water rise, water flow direction, and terrain data surrounding the power facilities, it predicts the scope and extent of the flood's subsequent impact on power facilities and dynamically updates the node states and relationships in the accident evolution diagram. Furthermore, the system features online model updates, adjusting model parameters in real time as new monitoring data continuously arrives, ensuring the model remains consistent with the actual operating conditions of the power system and the development of disasters. This dynamic deduction model comprehensively considers multiple factors, simulating the development of disaster chains, and providing decision makers with accurate emergency situation forecasts, enabling more effective emergency decision-making and a holistic response to disaster chains. This addresses the existing issues of lack of coordination in emergency measures and the difficulty in responding to complex disaster chains.

[0092] In one embodiment, if the time series-based prediction algorithm is a seasonal autoregressive integrated moving average algorithm, a dynamic deduction model is constructed using the time series-based prediction algorithm, including:

[0093] S21: Receive the real-time collected data transmitted by the data collection unit, and obtain the accident evolution diagram and the results provided by various disaster models corresponding to the accident.

[0094] The data acquisition unit collects real-time data on power industry emergency scenarios through various means (such as sensor networks and mobile devices). The knowledge base construction and accident evolution diagram generation module generates accident evolution diagrams, while the disaster model service chain construction and management module provides various disaster model results. By integrating this data, a comprehensive and rich information foundation is provided for the construction of dynamic deduction models. This overcomes the one-sidedness of the model caused by a single data source, enabling the model to comprehensively consider multiple factors and more accurately simulate the development of the disaster chain in power emergency scenarios.

[0095] S22: Preprocess the collected data and divide the preprocessed time series data into a training set and a test set.

[0096] Dividing time series data into training and test sets (e.g., in an 8:2 ratio) is necessary for subsequent model training and evaluation. Reasonable data preprocessing and partitioning can remove noise and outliers from the data, fill missing values, and standardize the data to ensure that the data in the training and test sets are representative, providing guarantees for accurate model fitting and reliable evaluation, and avoiding inaccurate model training due to data quality issues, which can affect the scientific nature of emergency decision-making. Draw a line graph of the time series data to observe whether the data has obvious seasonal cycles. For example, power load data may show cyclical changes at specific times of the day, specific days of the week, or specific seasons of the year. Calculate the autocorrelation function (ACF) and partial autocorrelation function (PACF) to analyze the seasonal characteristics and correlation of the data. If the data is not stationary, perform difference processing to eliminate trends and seasonality, and use a unit root (such as the ADF test) to verify whether the differenced data has reached a stationary state.

[0097] S23: Fit the SARIMA model using the training set data and estimate the model parameters using the maximum likelihood estimation method.

[0098] The parameter range of the SARIMA model is determined. The parameters are (p, d, q)(P, D, Q)s, where (p, d, q) are the parameters of the ordinary autoregressive integrated moving average component, (P, D, Q) are the parameters of the seasonal autoregressive integrated moving average component, and s is the seasonal period. Information criteria (such as AIC and BIC) are used to select the optimal model parameter combination. By traversing all possible combinations within the parameter range and calculating the AIC or BIC value for each combination, the parameter combination that minimizes the information criterion value is selected as the optimal model. The model is then fitted to the training data, enabling it to learn the patterns and characteristics of the time series data. Maximum likelihood estimation, a commonly used parameter estimation method, can identify the model parameter values most likely to produce these data given the data. Through this process, a model that reflects the changing patterns of time series data in power industry emergency scenarios is initially constructed, providing a foundational framework for subsequent forecasting and deduction.

[0099] S24: Convert the accident evolution diagram and the results provided by various disaster models corresponding to the accident into features that can be used for model input, and fuse these features with time series data as extended input features for SARIMA model training.

[0100] Combining the disaster chain structure information reflected in the accident evolution diagram and the output of the disaster model with time series data enriches the model's input information. For example, by quantifying the severity of the accident as a numerical feature and representing the scope of impact as the identifier or weight of the relevant area, the model not only considers changes in the time series but also comprehensively considers the mutual influence of various factors in the disaster chain. This enhances the model's comprehensive analysis capabilities and better copes with complex power emergency scenarios.

[0101] S25: Use the test set data to evaluate the trained model, and use the final model as the dynamic deduction model.

[0102] The model is evaluated using test data, and error metrics (such as mean squared error, mean absolute error, and root mean square error) between the predicted and true values are calculated to objectively evaluate the model's performance and accuracy. Based on the evaluation results, the model is optimized to ensure the final dynamic deduction model has high reliability and accuracy. This provides a scientific and reliable basis for decision-making in power industry emergency scenarios, enabling effective analysis and response to disaster chains from a holistic perspective, addressing the lack of coordination in emergency measures and the difficulty in responding to complex disaster chains in existing technologies.

[0103] In one embodiment, the data processing unit further includes:

[0104] The regular data backup subunit is used to regularly back up important data to an off-site storage device or cloud storage service.

[0105] In emergency scenarios in the power industry, multi-source data (including sensor monitoring data, emergency personnel feedback, video surveillance data, etc.) is crucial for emergency decision-making. Regularly backing up this important data to off-site storage devices or cloud storage services effectively prevents data loss due to local storage device failures, natural disasters (such as earthquakes and floods that could damage the local data center), or other unexpected situations. Combined with the original plan's strategy of using multiple databases (relational databases, time-series databases, object storage, etc.) to store different types of data, this further improves the data storage security system. For example, even if the local data center is damaged by a sudden disaster, the off-site backup can still ensure data integrity, providing data support for subsequent emergency response and analysis, avoiding the difficulties of emergency response due to data loss, and improving the system's ability to cope with complex situations.

[0106] The data recovery subunit is used to restore the data required for system operation from the backup data in the event of data loss or damage.

[0107] When data is lost or damaged, the data recovery subunit can quickly recover the data that the system relies on for operation from the backup data, so that other modules such as the real-time monitoring and dynamic deduction module, the knowledge base construction and accident evolution diagram generation module can continue to operate normally. For example, when the main storage device in the data processing unit fails and causes data loss, the data recovery subunit can quickly recover the data from the off-site backup, ensuring that the dynamic deduction model construction unit can build and deduce the model based on complete data, ensuring the accuracy and timeliness of emergency decision-making recommendations. This function solves the problem that the system cannot operate normally after data loss in the existing technology, realizes the integrity of data management and the continuity of system operation, enhances the stability and reliability of the entire emergency response system, and enables the system to better coordinate and respond to complex disaster chains and emergency scenarios.

[0108] In one embodiment, the decision suggestion module includes:

[0109] The chain-breaking decision model unit is used to identify key nodes in the accident evolution diagram using a key node identification algorithm, simulate the domino effect propagation process in combination with the disaster model, and evaluate the severity of the domino effect.

[0110] By calculating metrics such as degree centrality, betweenness centrality, and closeness centrality for each node, a comprehensive assessment of the node's importance in the accident propagation network is conducted. For example, substations that connect multiple transmission lines and provide power to a large number of loads within the power system topology are identified as key nodes. Based on these key nodes, a disaster model is used to simulate the domino effect propagation process, observing the propagation path and impact range of the accident within the power system network. Quantitative indicators (such as outage area, outage load, and economic losses) are then used to assess its severity. This approach addresses the previous situation in which emergency response efforts lacked sufficient understanding of the disaster chain and were unable to effectively interrupt the spread of disasters. It enables in-depth analysis of the mutual influence of each node in the disaster chain, providing a scientific basis for making targeted chain interruption decisions, and overall improving the system's ability to respond to complex disaster chains.

[0111] The emergency resource allocation decision unit is used to allocate relevant emergency resources according to the output results of the chain-breaking decision model.

[0112] Taking into account factors such as emergency resource constraints, time urgency, and cost-effectiveness, different chain-breaking strategies are modeled and evaluated. By solving optimization problems, the optimal chain-breaking solution is determined to minimize accident losses and chain reaction risks. For example, in the deployment of emergency personnel, a database of emergency personnel is established, and task allocation algorithms such as the Hungarian algorithm are used to match the most suitable repair personnel to each fault point, taking into account the time cost of personnel deployment. In the deployment of emergency vehicles, a mixed integer programming algorithm is used to develop emergency vehicle deployment plans, determining vehicle routes, task assignments, and arrival times at each fault point. In the deployment of emergency supplies, optimization algorithms such as genetic algorithms are used to optimize the emergency supply deployment plan, determining which storage depot to draw supplies from, which transportation method to use, and the allocated quantities. This precise allocation of emergency resources based on chain-breaking decision-making achieves resource-level coordination and integration of emergency response measures for different disaster types, ensuring the efficient use of emergency resources and minimizing disaster losses. It effectively addresses the irrational and lack of coordination in existing emergency resource allocation techniques.

[0113] In one embodiment, the execution process of the link-breaking decision model unit includes:

[0114] S31: Collect node data of the accident evolution diagram and data required by the disaster model, and preprocess the collected data.

[0115] Collect data related to the accident evolution diagram, identify nodes (such as power facilities and regions) and their connections, record node attributes (such as facility type and capacity) and the attributes of the edges between nodes (such as connection strength and transmission capacity), collect data required for disaster models, including characteristic parameters (such as magnitude, wind speed, and water level) of different disaster types (such as earthquakes, floods, and typhoons) and how they affect nodes, and preprocess the collected data. Preprocessing this data, such as removing noise and filling missing values, ensures data accuracy and usability, providing a reliable data foundation for subsequent analysis and simulation, and avoiding biased analysis results due to data quality issues, which could affect the accurate assessment of disaster chains.

[0116] S32: Based on the degree centrality algorithm, the node data of the preprocessed accident evolution graph is used as input, the centrality index of each node is calculated, the nodes are sorted according to the index value, and the nodes with index values greater than the set value are selected as key nodes.

[0117] By calculating metrics like node degree centrality, we can accurately identify critical nodes within the power system network, such as substations that connect multiple transmission lines and power a large load. Failure of these critical nodes could trigger severe chain reactions, so prioritizing them as key prevention and control targets helps us implement targeted measures, interrupt the spread of disaster chains, and improve the efficiency and effectiveness of emergency response.

[0118] S33: According to different disaster types and characteristic parameters, a disaster impact model on nodes is established based on the data required by the pre-processed disaster model.

[0119] Developing impact models for different disaster types (such as earthquakes, floods, and typhoons) accurately assesses the extent of damage to power infrastructure nodes. Furthermore, the propagation rules for the domino effect are defined based on the physical characteristics and operating rules of the power system, clarifying the conditions and probabilities for propagation. For example, a domino effect is triggered when node load exceeds a certain percentage of capacity. This makes the simulation more realistic, accurately simulating the development of a disaster chain, and providing a reliable basis for subsequent assessments and decision-making.

[0120] S34: Based on the physical characteristics and operating rules of the power system, define the propagation rules of the domino effect and determine the propagation conditions and probability of the accident.

[0121] Based on the physical characteristics and operating rules of the power system, define the propagation rules of the domino effect. For example, when one node fails, its neighboring nodes may be affected due to overload, voltage anomalies, and other factors, leading to a cascading failure. Determine the conditions and probability of propagation. For example, when a node's load exceeds a certain percentage of its capacity, a domino effect will be triggered.

[0122] S35: Initialize the simulation, set the initial conditions for the disaster, including the disaster type, characteristic parameters and affected nodes, and initialize the status of all nodes in the accident evolution diagram.

[0123] S36: Perform iterative propagation simulation based on the influence model. In each time step, check the status of each node and the status of adjacent nodes according to the domino effect propagation rule. For nodes that meet the propagation conditions, update the status of their adjacent nodes according to the corresponding probability, and record the propagation process. Repeat the above steps until no new node status changes or the preset simulation time is reached.

[0124] Based on the disaster impact model, the state of each node and its adjacent nodes is checked at each time step according to the domino effect propagation rules. For nodes that meet the propagation conditions, the states of adjacent nodes are updated according to the probability determined by the propagation rules, and the propagation process is fully recorded. This process is repeated until no new node states change or the preset simulation time is reached, dynamically displaying the propagation process of the disaster in the power system network.

[0125] Steps S34-S36 above comprehensively simulate and quantitatively assess the development of the disaster chain. Setting initial conditions and performing iterative propagation simulations dynamically demonstrates the spread of disasters within the power system network. Calculating evaluation indicators (such as outage area, outage load, and economic losses) intuitively reflects the severity of the domino effect, enabling decision-makers to clearly understand the scope and extent of the disaster's impact.

[0126] S37: Calculate a predefined evaluation index value based on the simulation results;

[0127] Define evaluation indicators and count the number of nodes that fail during the propagation of the domino effect. The more failed nodes there are, the more severe the domino effect is. Evaluate the functional loss of the power system due to node failure, such as the area of the power outage area, the amount of load loss, etc. Calculate the direct economic losses (such as equipment repair costs) and indirect economic losses (such as production losses caused by power outages) caused by the domino effect. Calculate the evaluation index value based on the simulation results. The weighted average method can be used to combine multiple indicators into a severity indicator.

[0128] S38: Analyze the role of key nodes in the domino effect propagation process based on the evaluation index value, and evaluate the impact of key node failure on the system;

[0129] S39: Determine the chain breaking decision information based on the impact assessment results.

[0130] Based on evaluation index values, the role of key nodes in the domino effect propagation process is analyzed, and the impact of key node failure on the system is assessed. Based on the domino effect impact assessment results, information for chain disconnection decisions is determined. This analysis of evaluation index values provides a deeper understanding of the role and impact of key nodes in disaster propagation, providing a basis for making appropriate chain disconnection decisions. For example, when the assessment results indicate that a domino effect could cause a widespread system collapse, certain key connections are promptly severed to prevent further propagation of the fault. Furthermore, when making chain disconnection decisions, the feasibility and impact of the decision are fully considered, ensuring that while minimizing domino effect losses, the normal operation of the system is not significantly impacted, effectively interrupting the disaster chain and optimizing emergency management.

[0131] Relevant emergency resources are deployed based on the output of the chain-breaking decision model. Different chain-breaking strategies are modeled and evaluated, taking into account factors such as emergency resource constraints, time urgency, and cost-effectiveness. By solving an optimization problem, the optimal chain-breaking solution is determined to minimize accident losses and chain reaction risks. In this embodiment of the present invention, the optimal emergency decision-making recommendations include the deployment of repair personnel, emergency vehicles, and emergency supplies.

[0132] A resource allocation optimization model was established, with the objective functions of minimizing emergency response time, maximizing resource utilization efficiency, and minimizing accident losses. This model considered constraints such as the type, quantity, distribution, transportation time, and task priority of emergency resources. Combined with the emergency task requirements output by the chain-breaking decision model, it rapidly calculated the optimal deployment plan for repair personnel, emergency vehicles, and emergency supplies. Furthermore, dynamic factors such as traffic conditions and weather were considered to adjust and optimize the resource allocation plan in real time.

[0133] The following uses the deployment of emergency repair personnel, emergency vehicles, and emergency supplies as examples:

[0134] Repair personnel deployment strategy: A repair personnel database is established, recording each repair personnel's skills, expertise, work experience, and location. Based on the technical difficulty, urgency, and required skills of each fault point in the accident evolution diagram, task allocation algorithms such as the Hungarian algorithm are used to match the most suitable repair personnel to each fault point. Taking into account the time cost of personnel deployment, priority is given to personnel who are close to the fault point and have matching skills. The optimal route is planned, and repair personnel are guided to the scene quickly in real time through a navigation system.

[0135] Optimizing Emergency Material Allocation: Build a model of emergency material storage depots, recording information such as material types, quantities, and storage locations. Based on accident evolution simulation results, estimate the types and quantities of emergency materials required at each fault point. Utilizing optimization algorithms such as genetic algorithms, optimize emergency material allocation plans while meeting material needs, taking into account factors such as material transportation costs and inventory management costs. This determines which depot to draw materials from, which transportation method to use, and the allocated quantities, ensuring efficient distribution of emergency materials and smooth repair work.

[0136] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. An emergency response measures chain coupling analysis system, characterized in that: include: Disaster model service chain construction and management module, knowledge base construction and accident evolution diagram generation module, real-time monitoring and dynamic deduction module, and decision-making recommendation module; The disaster model service chain construction and management module is used to collect various disaster models, encapsulate the disaster models according to a unified model encapsulation specification, and store them in multiple storage nodes of the model storage cluster in a dispersed manner; The knowledge base construction and accident evolution diagram generation module is used to generate an accident evolution diagram based on a pre-constructed knowledge base; the accident evolution diagram is used to represent the evolution process of multiple accidents that have correlations and joint effects; The real-time monitoring and dynamic deduction module is used to collect various data of emergency scenarios in the power industry in real time, and build a dynamic deduction model based on the collected data, combined with the accident evolution diagram and the output provided by various disaster models corresponding to the accident; The decision suggestion module is used to simulate the dynamic deduction model using the domino effect, and determine the optimal emergency decision suggestion information for the actual scenario based on the simulation results.

2. The emergency response measures chain coupling analysis system according to claim 1, characterized in that: The disaster model service chain construction and management module includes: A model collection unit, used to collect various disaster models and encapsulate the disaster models according to a unified model encapsulation specification; A distributed storage unit is used to construct a model storage cluster using a distributed file system and to store the encapsulated disaster model files in multiple storage nodes in a dispersed manner; The orchestration rule unit is configured with a rule-based model orchestration call engine and an orchestration call rule library, and is used to trigger the model orchestration call engine after receiving disaster warning or power facility failure information, and call the disaster model based on the orchestration call rule library.

3. The emergency response measures chain coupling analysis system according to claim 2, characterized in that: In the orchestration rule unit, the process of formulating the orchestration call rule base includes: Conduct demand analysis on emergency scenarios in the power industry, determine emergency scenario information and disaster model requirements for each emergency scenario; Determine the rule elements and rule logic based on the results determined by the demand analysis; the rule elements are used to specify the calling conditions of various disaster models; the rule logic is used to specify the calling order of various disaster models; The rule elements and the rule logic are converted into rule texts, the rule texts are collated and tested, and the orchestration call rule library is formed based on the rule texts that have passed the test.

4. The emergency response measures chain coupling analysis system according to claim 1, characterized in that: The knowledge base construction and accident evolution graph generation module includes: A knowledge base construction unit is used to construct and store a knowledge base based on the relevant knowledge of emergency repair in the power industry; The accident evolution graph generating unit is used to generate an accident evolution graph based on data in a knowledge base and continuously update the accident evolution graph.

5. The emergency response measures chain coupling analysis system according to claim 4, characterized in that: In the knowledge base construction unit, the knowledge related to emergency repair in the power industry is classified, each category is used as a node, and the knowledge base is constructed and stored after associating each node.

6. The emergency response measures chain coupling analysis system according to claim 1, characterized in that: The real-time monitoring and dynamic deduction module includes: A data acquisition unit, used to collect real-time data on power industry emergency scenarios and information fed back by emergency personnel through a sensor network; A data processing unit, used for fusing the collected multi-source data based on a data fusion algorithm; The dynamic deduction model construction unit is used to construct a dynamic deduction model based on a time series-based prediction algorithm based on real-time fusion data, combined with the accident evolution diagram and the outputs provided by various disaster models corresponding to the accident.

7. The emergency response measures chain coupling analysis system according to claim 6, characterized in that: If the time series-based prediction algorithm is a seasonal autoregressive integrated moving average algorithm, a dynamic deduction model is constructed using the time series-based prediction algorithm, including: Receiving the real-time collected data transmitted by the data collection unit, and obtaining the accident evolution diagram and the results provided by the various disaster models corresponding to the accident; Preprocess the collected data and divide the preprocessed time series data into training set and test set; The SARIMA model is fitted using the training set data, and the model parameters are estimated using the maximum likelihood estimation method; Converting the accident evolution graph and the results provided by the various disaster models corresponding to the accident into features that can be used as model input, and fusing these features with time series data as extended input features to train the SARIMA model; The trained model is evaluated using the test set data, and the final model is used as the dynamic deduction model.

8. The emergency response measures chain coupling analysis system according to claim 6, characterized in that: The data processing unit further includes: Regular data backup subunit, used to regularly back up important data to off-site storage devices or cloud storage services; The data recovery subunit is used to restore the data required for system operation from the backup data in the event of data loss or damage.

9. The emergency response measures chain coupling analysis system according to claim 1, characterized in that: The decision suggestion module includes: a chain-breaking decision model unit, configured to identify key nodes of the accident evolution diagram using a key node identification algorithm, simulate the domino effect propagation process in combination with a disaster model, and evaluate the severity of the domino effect; The emergency resource allocation decision unit is used to allocate relevant emergency resources according to the output results of the chain-breaking decision model.

10. The emergency response measures chain coupling analysis system according to claim 9, characterized in that: The execution process of the link-breaking decision model unit includes: Collecting node data of the accident evolution diagram and data required by the disaster model, and preprocessing the collected data; Based on the degree centrality algorithm, the node data of the pre-processed accident evolution graph is used as input to calculate the centrality index of each node, sort the nodes according to the index value, and select the nodes with index values greater than the set value as key nodes; According to different disaster types and characteristic parameters, a disaster impact model on nodes is established based on the pre-processed data required by the disaster model; Based on the physical characteristics and operating rules of the power system, define the propagation rules of the domino effect and determine the propagation conditions and probability of accidents; Initialize the simulation, set the initial conditions for the disaster, including the disaster type, characteristic parameters and affected nodes, and initialize the states of all nodes in the accident evolution graph; An iterative propagation simulation is performed based on the influence model. Within each time step, the state of each node and the state of its adjacent nodes are checked according to the domino effect propagation rule. For nodes that meet the propagation conditions, the states of their adjacent nodes are updated according to the corresponding probability, and the propagation process is recorded. The above steps are repeated until no new node states change or the preset simulation time is reached; Calculate predefined evaluation index values based on simulation results; Analyze the role of key nodes in the domino effect propagation process based on the evaluation index value, and evaluate the impact of key node failure on the system; Based on the impact assessment results, the link breaking decision information is determined.

Images