Electric power system safety risk prediction method based on digital twinning and related device

Abstract

The invention provides the technical field of electric power systems, and particularly provides an electric power system safety risk prediction method based on digital twinning and a related device. According to the method and the device, the operator data, the tool data, the plan ticket data and the system environment data related to the operation of the power system can be acquired in real time, and the twinborn state of the power safety digital twinborn is updated, so that the power safety digital twinborn can be linked with the real situation of the physical world in real time; and multi-source heterogeneous data such as operators, operation plans, operation tools, system environments and the like are deeply fused, and comprehensive data support is provided for risk analysis. The risk coupling model can perform real-time deduction on continuously changing safety risks in the operation process based on the updated twinborn body state, so that advanced early warning and quantification of the safety risks can be realized, and potential safety risks possibly involved in the field operation process of the power system can be comprehensively and accurately predicted.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a method and related apparatus for predicting power system security risks based on digital twins. Background Technology

[0002] The safe and stable operation of power systems is crucial for ensuring socio-economic development and the normal functioning of daily life, and the safety management of on-site operations is a vital link in ensuring the safe and stable operation of power systems. Because power systems encompass multiple stages such as transmission, transformation, distribution, and consumption, the on-site operating environment is complex and high-risk. Therefore, effectively identifying and managing safety risks in on-site operations has become a critical issue that urgently needs to be addressed for the safe operation of power systems. However, existing technologies are insufficient to comprehensively and accurately identify potential safety risks during on-site operations in power systems. Summary of the Invention

[0003] The purpose of this application is to address at least one of the aforementioned technical deficiencies, particularly the technical deficiency in the prior art that makes it difficult to comprehensively and accurately identify potential safety risks during on-site operations of power systems.

[0004] Firstly, this application provides a power system security risk prediction method based on digital twins, including:

[0005] Real-time acquisition of data related to power system operations, including personnel data, tool data, project ticket data, and system environment data;

[0006] Based on the operator data, tool data, project ticket data, and system environment data, the state of the power safety digital twin is updated so that the power safety digital twin reflects the attributes of each entity and the relationships between entities in real time; wherein, the entities of the power safety digital twin include operators, tools, project tickets, and system environment related to the power system operation;

[0007] In response to the twin state update operation, the updated twin state is input into a pre-constructed risk coupling model, so that the risk coupling model performs dynamic risk prediction based on the twin state and outputs a comprehensive risk value;

[0008] Based on the comprehensive risk value, operational decision-making information is generated to guide operators in their on-site work.

[0009] In some embodiments, the twin state includes an entity state and a relation state;

[0010] The step of updating the digital twin status of the power safety digital twin based on the operator data, the tool data, the project ticket data, and the system environment data includes:

[0011] Based on the operator data, the tool data, the plan ticket data, and the system environment data, an entity attribute change event is determined; wherein, the entity attribute change event includes the target entity identifier, the attribute identifier, and the changed attribute value;

[0012] Based on the worker data, the tool data, and the project ticket data, an entity relationship change event is determined; wherein, the entity relationship change event includes the associated entity identifier and the relationship operation identifier;

[0013] Based on the entity attribute change event and the entity relationship change event, the entity state and relationship state of the power safety digital twin are updated.

[0014] In some embodiments, updating the entity state and relationship state of the power safety digital twin based on the entity attribute change event and the entity relationship change event includes:

[0015] The target entity is determined based on the target entity identifier, and the entity attributes of the target entity are updated according to the attribute identifier and the changed attribute value, so as to update the entity status of the power safety digital twin.

[0016] The entity relationship change event is verified based on the associated entity identifier and the preset entity coupling rules; wherein, the entity coupling rules include at least one of the following: personnel qualification verification rules, personnel status verification rules, tool status verification rules, environmental verification rules, and plan ticket verification rules.

[0017] If the verification passes, the relationship status of the power safety digital twin is updated according to the associated entity identifier and the relationship operation identifier;

[0018] If the verification fails, an alarm message is generated based on the verification result and pushed to the alarm terminal.

[0019] In some embodiments, the risk coupling model is a dynamic Bayesian network model, wherein the nodes of the risk coupling model include various security risk factors, and the risk coupling model records the conditional probability relationships of each node.

[0020] In some embodiments, generating operational decision information based on the comprehensive risk value to guide on-site operations by workers includes:

[0021] The real-time risk level is determined based on the comprehensive risk value and the preset risk level division range;

[0022] If the real-time risk level is higher than the preset alarm level, then a direct risk node is determined in each node of the risk coupling model, and N key factor candidate nodes are selected based on the direct risk node and the preset power system security domain knowledge base; where N is a positive integer.

[0023] The updated twin state is input into the risk coupling model to obtain the actual probability value of the direct risk node being in a high-risk state;

[0024] For each of the N key factor candidate nodes, the state information corresponding to the key factor candidate node in the updated twin state is adjusted to a preset safety value, and the adjusted twin state is input into the risk coupling model to obtain the assumed probability value that the direct risk node is in a high-risk state.

[0025] Based on the N hypothetical probability values ​​and the N candidate nodes of key factors, key risk factors are determined, and the operation decision information is generated based on the key risk factors and the preset decision rule base.

[0026] In some embodiments, determining key risk factors based on N hypothetical probability values ​​and N key factor candidate nodes includes:

[0027] Calculate the probability difference between each of the N hypothetical probability values ​​and the actual probability values;

[0028] Select the probability difference that is greater than or equal to the preset contribution threshold from the N probability differences as the target difference;

[0029] The key risk factors are determined based on the candidate nodes of key factors corresponding to the target difference.

[0030] In some embodiments, the method further includes:

[0031] In response to the model optimization instruction, the historical operation execution results of the power system are obtained, and the conditional probability relationship of the risk coupling model is updated based on the historical operation execution results.

[0032] Secondly, embodiments of this application provide a power system security risk prediction device based on digital twins, comprising:

[0033] The data acquisition module is used to acquire real-time data on personnel, tools, project tickets, and system environment related to power system operations.

[0034] The digital twin update module is used to update the state of the power safety digital twin based on the operator data, the tool data, the project ticket data, and the system environment data, so that the power safety digital twin reflects the attributes of each entity and the relationships between entities in real time; wherein, the entities of the power safety digital twin include operators, tools, project tickets, and system environment related to the power system operation;

[0035] The risk prediction module is used to respond to the twin state update operation by inputting the updated twin state into a pre-built risk coupling model, so that the risk coupling model can perform dynamic risk prediction based on the twin state and output a comprehensive risk value.

[0036] The task decision generation module is used to generate task decision information based on the comprehensive risk value to guide the on-site operations of the workers.

[0037] Thirdly, embodiments of this application provide a storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the power system security risk prediction method based on digital twins described in any of the above embodiments.

[0038] Fourthly, embodiments of this application provide a computer device, which includes: one or more processors, and a memory;

[0039] The memory stores computer-readable instructions, which, when executed by the one or more processors, perform the steps of the power system security risk prediction method based on digital twins as described in any of the above embodiments.

[0040] In the power system safety risk prediction method and related apparatus based on digital twins provided in some embodiments of this application, real-time acquisition of personnel data, tool data, work order data, and system environment data related to power system operations is possible. Based on this, the state of the power safety digital twin is updated, enabling the power safety digital twin to interact in real-time with the real-world situation and reflect the real-time status and interactions between personnel, tools, work orders, and the system environment. This deeply integrates multi-source heterogeneous data such as personnel, work plans, work tools, and the system environment, providing comprehensive data support for risk analysis. The risk coupling model can, based on the updated twin state, perform real-time simulations of constantly changing safety risks during operations, thereby transforming from the passive information management of existing technologies to the proactive risk prediction of this application. This achieves advanced early warning and quantification of safety risks and can comprehensively and accurately predict potential safety risks that may be involved in power system field operations. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is one of the flowcharts illustrating a power system security risk prediction method based on digital twins in one embodiment;

[0043] Figure 2 This is a second flowchart illustrating a digital twin-based power system security risk prediction method in one embodiment.

[0044] Figure 3 This is a schematic diagram of the structure of a power system security risk prediction device based on digital twins in one embodiment;

[0045] Figure 4 This is a diagram of the internal structure of a computer device in one embodiment. Detailed Implementation

[0046] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0047] The technical terms used in this application are explained below.

[0048] Digital twin: refers to a high-fidelity virtual model corresponding to a physical power system, which can achieve synchronous mapping, two-way interaction and collaborative evolution with the physical entity through real-time data-driven processes.

[0049] Full life cycle: refers to all stages of a physical power system from planning and design, construction, operation and maintenance to decommissioning and scrapping. This application focuses on safety risk management during the operation and maintenance phase.

[0050] Safety risks refer to the combination of the possibility and consequences of personal injury, equipment damage, or power grid accidents caused by factors such as personnel, equipment, environment, and management during power operations.

[0051] Multi-source heterogeneous data: refers to data with diverse sources and different formats. For example, it can include any one or any combination of job plans from the production management system, personnel qualification information from the human resources system, tool status data from Internet of Things sensors, text information from the work ticket system, environmental data from the geographic information system, etc.

[0052] Dynamic Bayesian network: A graphical model suitable for modeling and probabilistic inference of stochastic processes that change over time. It can handle uncertainty and temporal causal relationships and is suitable for risk evolution prediction.

[0053] As described in the background art, it is difficult for the prior art to comprehensively and accurately identify potential safety risks in the on-site operation process of the power system. Through the research of the inventor, it is found that this problem is mainly caused by information islands, lag in risk early warning, and static control measures. Among them, information islands refer to the fact that information such as personnel qualification information, job plans, tool calibration records, and historical work tickets of operators is scattered in different systems and it is difficult to conduct linkage analysis. The lag in risk early warning means that the prior art mostly relies on managers to evaluate job risks based on manual experience before the operation and lacks the dynamic prediction ability based on real-time data. Static control measures mean that once the safety measures are formulated, it is difficult to dynamically adjust them according to the changes in the system environment and personnel status during the operation process.

[0054] In addition, the prior art also has the problem of insufficient decision support. When an abnormality occurs during the operation process, the prior art mainly relies on pre-formulated emergency plans and manual decisions for decision analysis. This method not only has too low efficiency but also lacks scientific decision-making and cannot ensure the feasibility and accuracy of the decision.

[0055] To solve at least one of the above technical problems, the present application provides a power system safety risk prediction method and related device based on digital twin. By deeply integrating multi-source heterogeneous data such as operators, job plans, job tools, and system environment, and by real-time推演 of the continuously changing safety risks during the operation process, the present application can achieve early warning and quantification of safety risks and can comprehensively and accurately predict potential safety risks that may be involved in the on-site operation process of the power system.

[0056] The following describes the power system safety risk prediction method based on digital twin provided by the present application.

[0057] In some embodiments, the power system security risk prediction method based on digital twins provided in this application can be applied to a power system security risk intelligent management and control platform. The power system security risk intelligent management and control platform may include computer equipment, edge computing terminals, and an intelligent analysis dashboard system. The computer equipment may be deployed with core software, such as digital twin models, risk coupling models, and decision optimization modules. It is understood that the computer equipment may be a cloud server or a local server, and this application does not impose specific limitations on this. The edge computing terminal may be integrated into substations, distribution rooms, or tools such as work vehicles and smart safety helmets. The edge computing terminal may also be worn by workers, thereby enabling real-time collection of power system environmental data, personnel location, and tool status data. The intelligent analysis dashboard system can provide decision support to managers at all levels, including a global risk view, personnel qualification status, and work plan safety ratings.

[0058] In some embodiments, the power system security risk prediction method based on digital twins provided in this application can be executed by computer equipment. For example... Figure 1 As shown, the method provided in this application may include the following steps:

[0059] S102: Real-time acquisition of data related to power system operations, including personnel data, tool data, project ticket data, and system environment data.

[0060] In this step, computer equipment can continuously acquire various real-time data related to power system operations through methods such as calling data interfaces and interacting with sensors, resulting in multi-source heterogeneous data. This multi-source heterogeneous data is the driving force behind safety risk prediction and can specifically include personnel data, tool data, project schedule data, and system environment data, ensuring that the digital twin can accurately and timely map the state of the physical entity.

[0061] Personnel data refers to various types of information associated with on-site personnel in the power system, including personnel qualification information, physiological status information, and location information. Personnel qualification information is used to determine the qualifications and capabilities of personnel to perform on-site power operations; this may include professional skill identifiers, certificate identifiers, training records, operating permissions, and relevant examination results. Personnel physiological status information is used to assess the physiological state of personnel during operations to evaluate operational accuracy and safety. For example, physiological status information may include fatigue level, health status, and blood pressure values. Personnel location information reflects the specific location of personnel at the power system work site, facilitating monitoring and management of their work scope and movement trajectories.

[0062] Tool and equipment data refers to information related to the tools and equipment used in power operations, such as the tool's model, specifications, usage status, performance parameters, applicable scenarios, and skill requirements. Work order data may include work plan information and work permit information. Work plan information records the planned arrangements for on-site operations in the power system, such as the start and end times, work content, and work area.

[0063] Work order information can serve as an important credential for standardizing work procedures and ensuring work safety during operations. For example, it may include work order number, work task, work location, work team members, safety measures, etc.

[0064] System environmental data refers to environmental information at the power system operation site, including meteorological data (such as temperature, humidity, wind speed, SF6 concentration, and rainfall), geographical information (such as altitude and topography of the operation site), and equipment operating status (such as voltage, current, and power parameters of power equipment). This environmental information has a significant impact on the safety and feasibility of power operations. For example, severe weather conditions may increase the safety risks for workers, and abnormal equipment operating conditions may indicate potential safety hazards.

[0065] It is understood that this application can acquire data in any way, and this application does not impose any specific restrictions on this. In some examples, this application can obtain personnel qualification information from personnel systems or training systems, obtain personnel physiological status information and personnel location information through base station positioning or smart bracelets, etc.; obtain information such as model, specifications, applicable scenarios, and skill requirements from tool management systems, and obtain the real-time usage status and performance parameters of tools through RFID (Radio Frequency Identification) or sensors; obtain planning ticket data from production management systems (MIS), and obtain system environmental data from field sensors by calling the meteorological early warning interface.

[0066] S104: Update the status of the power safety digital twin based on the data of operators, tools, plans and tickets, and system environment, so that the power safety digital twin reflects the attributes of each entity and the relationships between entities in real time; wherein, the entities of the power safety digital twin include operators, tools, plans and tickets, and system environment related to power system operations.

[0067] In this step, the acquired real-time data can be injected into the pre-constructed power safety digital twin, enabling the digital twin to dynamically update its status based on real-time personnel data, tool data, work plan data, and system environment data. Since the power safety digital twin treats all personnel, tools, work plans, and system environment related to power system operations as entities (i.e., nodes), the updated digital twin can reflect a complete snapshot of the overall safety posture, including the inherent attributes of all entities related to power system operations, the dynamic relationships between entities, and the resulting overall safety situation.

[0068] The entity's inherent attributes can reflect its individual state. Further, in some examples, when an entity corresponds to a worker, its inherent attributes may include qualifications, current working hours, and fatigue index; when an entity corresponds to tools, its inherent attributes may include calibration status, tool health, and current user identification; when an entity corresponds to a work plan, its inherent attributes may include work status and risk level; and when an entity corresponds to a work area, its inherent attributes may include SF6 concentration and temperature.

[0069] For example: the entity attributes corresponding to worker A can be {Qualification level: Advanced, Current working hours: 6.5, Fatigue index: 0.7}, the entity attributes corresponding to insulating gloves B can be {Verification status: Valid, Health level: 0.85, Current user: Personnel A}, the entity attributes corresponding to work plan C can be {Status: Permitted, Risk level: To be assessed}, and the entity attributes corresponding to the work area can be {SF6 concentration: 800ppm, Temperature: 35℃}.

[0070] Relationships between entities can reflect the dynamic connections between different entities at a specific moment, and can couple different entities. For example, the relationship (person A, insulated glove B, in use) means that person A is currently using insulated glove B; another example is the relationship (person A, work plan C, responsible for execution) which means that person A is responsible for executing work plan C; yet another example is the relationship (lightning warning E, work plan C, time overlap) which means that there may be lightning weather in the work time and work area of ​​work plan C.

[0071] It should be noted that the power safety digital twin of this application includes not only a three-dimensional model of the power equipment, but also a multi-dimensional, multi-scale, and multi-physics virtual mapping of four major entity models: operators, tools, work plans, and system environment, as well as their interrelationships.

[0072] S106: In response to the twin state update operation, the updated twin state is input into the pre-built risk coupling model so that the risk coupling model can perform dynamic risk prediction based on the twin state and output a comprehensive risk value.

[0073] In this step, whenever the state of the power safety digital twin is updated, the computer equipment recalculates the risk posture of the affected entities based on the updated twin state, dynamically determining the latest comprehensive risk value according to the real-time situation of the physical entities. During the risk posture determination process, this application can input the current twin state (i.e., the updated twin state corresponding to this update) into the risk coupling model, and use the risk coupling model to perform risk reasoning, potential risk identification, and risk prediction, thereby outputting a comprehensive risk value.

[0074] It is understood that the risk coupling model can be implemented using any model, such as recurrent neural networks, Transformer time series models, etc., and this application does not impose any specific restrictions on it. The comprehensive risk value is a macro-level indicator derived from a comprehensive calculation of the states and relationships of all entities; for example, the comprehensive risk value can be 0.65.

[0075] S108: Generate operational decision-making information based on the comprehensive risk value to guide on-site operations for workers.

[0076] In this step, the comprehensive risk value reflects the overall safety risk level of on-site operations in the power system at a specific moment. After acquiring the comprehensive risk value, the computer equipment can generate corresponding operational decision information. This information aims to guide operators in taking appropriate safety measures, deciding whether to advance / postpone on-site operations, and whether to change tools, in order to reduce or avoid potential safety risks. For example, when the comprehensive risk value exceeds a certain threshold, indicating a significant safety risk in the current work environment, the operational decision information may include suggestions such as suspending operations, strengthening safety protection measures, and adjusting the work plan.

[0077] Furthermore, operational decision-making information can be customized based on the specific responsibilities of the operators and the characteristics of the working environment. For example, for operators responsible for high-altitude operations, the operational decision-making information can particularly emphasize the monitoring of weather conditions and corresponding countermeasures; for operators responsible for operating high-voltage equipment, the operational decision-making information can focus on the operating status of the equipment and the procedures for handling abnormalities.

[0078] In some examples, the generated operational decision information can be conveyed to workers through various means, such as intelligent analytics dashboards, mobile devices, or on-site broadcasting systems, to ensure they are promptly aware of the current safety risks and can take appropriate countermeasures. Simultaneously, this operational decision information can also provide decision support for managers, helping them better monitor and manage the safety risks of on-site operations in the power system.

[0079] This application enables real-time acquisition of data related to power system operations, including personnel data, tool data, work order data, and system environment data. Based on this data, the digital twin status of the power safety digital twin is updated, allowing it to interact in real-time with the physical world and reflect the real-time status and interactions between personnel, tools, work orders, and the system environment. It deeply integrates multi-source heterogeneous data from personnel, work plans, tools, and the system environment, providing comprehensive data support for risk analysis. The risk coupling model, based on the updated twin status, can perform real-time simulations of constantly changing safety risks during operations. This transforms the passive information management of existing technologies into the proactive risk prediction of this application, achieving advanced early warning and quantification of safety risks, and comprehensively and accurately predicting potential safety risks that may be involved in power system field operations.

[0080] In some embodiments, the twin state may include entity state and relation state. Entity state refers to the entity's own attributes, such as the aforementioned qualification level, current working hours, fatigue count, verification status, and health. Relationship state refers to the association between entities.

[0081] In this embodiment, updating the state of the power safety digital twin based on operator data, tool data, project ticket data, and system environment data may include:

[0082] Step A1: Based on the operator data, tool data, project schedule data, and system environment data, determine the entity attribute change event; wherein, the entity attribute change event includes the target entity identifier, attribute identifier, and the changed attribute value;

[0083] Step A3: Based on the operator data, tool data, and project ticket data, determine the entity relationship change events; whereby, the entity relationship change events include the associated entity identifier and the relationship operation identifier;

[0084] Step A5: Update the entity state and relationship state of the power safety digital twin based on entity attribute change events and entity relationship change events.

[0085] Specifically, computer equipment can continuously sense events based on operator data, tool data, plan and ticket data, and system environment data. These events can be categorized into entity attribute change events and entity relationship change events. Entity attribute change events refer to events related to changes in the attributes of an entity itself, such as an upgrade in an operator's qualification level, a change in the calibration status of tools, or an adjustment to the risk level of a work plan. Event information for entity attribute change events can include the target entity identifier, the attribute identifier, and the changed attribute value. The target entity identifier is the identifier of the entity whose attribute needs to be changed; the attribute identifier indicates the specific attribute that needs to be changed; and the changed attribute value refers to the specific numerical value or state of the changed attribute. For example, if event 1 is the update of worker A's fatigue level from 0.6 to 0.7, then the target entity identifier is the entity identifier corresponding to worker A, the attribute identifier is the attribute identifier corresponding to the fatigue level, and the changed attribute value is 0.7.

[0086] Entity relationship change events refer to events related to changes in the relationships between entities, such as workers changing the tools they use or changes in the work plan that workers are responsible for executing. Event information for entity relationship change events can include associated entity identifiers and relationship operation identifiers. Associated entity identifiers refer to the identifiers of multiple entities involved in the relationship. Relationship operation identifiers indicate the type of change in the relationship between these entities, including establishing and dissolving the relationship. For example, if event 2 is that worker A retrieves insulating gloves B from the tool cabinet, then the associated entity identifiers for event 2 are the entity identifiers corresponding to worker A and insulating gloves B, and the relationship operation identifier is the identifier corresponding to establishing the relationship. As another example, if event 3 is that work plan C is completed, then the associated entity identifiers for event 3 are the entity identifier corresponding to work plan C, the entity identifier corresponding to the worker executing work plan C, and the entity identifier corresponding to the tool used by the worker, and the relationship operation identifier is the identifier corresponding to dissolving the relationship.

[0087] Upon sensing entity attribute change events and entity relationship change events, the computer device can update the entity state and relationship state of the power safety digital twin accordingly to achieve twin update. For example, when the power safety digital twin uses a graph database to record data, nodes represent entities and edges represent relationships. When event 2 is sensed, the computer device can perform the following operations:

[0088] (1) Create a relation edge: between the personnel node A and the insulating glove node B, create or update an edge labeled USE and set the attribute start_time = now().

[0089] (2) Update node status: Set the current_tool attribute of the personnel A node to the ID of glove B; update the status attribute of the glove B node to in_use and the user attribute to the ID of personnel A.

[0090] The above update process can be atomic, ensuring the consistency of relations and states.

[0091] In this embodiment, through event perception and digital twin update mechanisms, various dynamic changes during power system operations can be accurately captured. Whether it's subtle adjustments to an entity's own attributes or the establishment and dissolution of relationships between entities, all can be recorded and updated promptly and accurately. This enhances the real-time performance and accuracy of the power safety digital twin, facilitating a more comprehensive and accurate discovery of potential safety hazards and risk patterns, and providing strong protection for the long-term safe operation of the power system.

[0092] In some embodiments, the entity state and relationship state of the power safety digital twin are updated based on entity attribute change events and entity relationship change events, including:

[0093] Step A51: Determine the target entity based on the target entity identifier, and update the entity attributes of the target entity according to the attribute identifier and the changed attribute values ​​to update the entity status of the power safety digital twin;

[0094] Step A53: Verify entity relationship change events based on the associated entity identifier and preset entity coupling rules; wherein, the entity coupling rules include at least one of the following: personnel qualification verification rules, personnel status verification rules, tool status verification rules, environmental verification rules, and plan ticket verification rules;

[0095] Step A55: If the verification passes, update the relationship status of the power security digital twin based on the associated entity identifier and the relationship operation identifier;

[0096] Step A57: If the verification fails, generate an alarm message based on the verification result and push the alarm message to the alarm terminal.

[0097] In this embodiment, when it is necessary to establish / de-establish relationships between entities, this application can first trigger a preset entity coupling rule for verification, and only update the relationship state of the power safety digital twin if the verification passes. If the verification fails, the relationship state is not updated, and an alarm is triggered to avoid potential security risks caused by unreasonable changes in entity relationships, thereby further improving the safety of on-site operations in the power system.

[0098] Specifically, entity coupling rules may include at least one of the following: personnel qualification verification rules, personnel status verification rules, tool and equipment status verification rules, environmental verification rules, and work plan / document verification rules. Personnel verification rules verify the qualification level, certificate validity period, and training records of operators to ensure they possess the qualifications and abilities to perform the corresponding tasks. Personnel status verification rules verify the current physiological state of operators to prevent safety accidents caused by poor operator condition. Tool and equipment status verification rules verify the verification status, health, and service life of tools and equipment to ensure they are in a safe and reliable working condition. Environmental verification rules verify environmental parameters such as SF6 concentration, temperature, humidity, and oxygen content in the work area to ensure the work environment meets safety requirements. Work plan / document verification rules verify the status, risk level, work time, and work area of ​​the work plan to ensure the work plan is reasonable, feasible, and complies with safety regulations. Through the verification of these entity coupling rules, changes to entity relationships in the power safety digital twin can be ensured to be reasonable and safe, thereby further improving the safety and reliability of on-site operations in the power system.

[0099] For example, when the entity coupling rule is: IF (Create Relationship(Personnel, Plan, "Execution")) AND (Personnel.Qualification Level < Plan.Required Qualification) THEN Trigger Event ("Qualification Mismatch Alarm"), if an attempt is made to establish the relationship "Personnel A executes Plan C", the computer device will automatically trigger the entity coupling rule for checking. Under the action of the entity coupling rule, if the condition is true, the relationship establishment fails, and the computer device will send an alarm command to the business chain to prevent the issuance of work tickets.

[0100] In some embodiments, the risk coupling model is a Dynamic Bayesian Network (DBN) model. The nodes of the risk coupling model include various safety risk factors, and the risk coupling model records the conditional probability relationships of each node. Because the DBN model has advantages in uncertainty representation, causal interpretation, and few-shot learning, it can better handle the complex and ever-changing safety risk factors and their interrelationships in power system field operations compared to other models.

[0101] In the DBN model, various safety risk factors, such as insufficient personnel qualifications, expired tools and equipment, severe weather, and complex plans, are abstracted as nodes, and the conditional probability relationships between each node are recorded. These conditional probability relationships between nodes characterize the causal connections and the degree of mutual influence among these risk factors. For example, the conditional probability relationships between nodes can be determined based on historical accident data and expert experience.

[0102] In this embodiment, the updated twin state can be used as evidence input into the DBN, enabling the DBN model to perform probabilistic reasoning and calculate a series of risk probability distributions for future states, ultimately outputting a quantitative and forward-looking comprehensive risk value.

[0103] In some embodiments, such as Figure 2 As shown, operational decision-making information is generated based on the comprehensive risk value to guide on-site operations for workers, including:

[0104] Step B1: Divide the risk range according to the comprehensive risk value and the preset risk level to determine the real-time risk level;

[0105] Step B3: If the real-time risk level is higher than the preset alarm level, then determine the direct risk node in each node of the risk coupling model, and select N key factor candidate nodes based on the direct risk node and the preset power system security domain knowledge base; where N is a positive integer.

[0106] Step B5: Input the updated twin state into the risk coupling model to obtain the actual probability value of the direct risk node being in a high-risk state;

[0107] Step B7: For each of the N key factor candidate nodes, adjust the state information corresponding to that key factor candidate node in the updated twin state to a preset safety value, and input the adjusted twin state into the risk coupling model to obtain the assumed probability value that the direct risk node is in a high-risk state.

[0108] Step B9: Based on N hypothetical probability values ​​and N key factor candidate nodes, determine the key risk factors, and generate operational decision information based on the key risk factors and the preset decision rule base.

[0109] In this embodiment, the computer equipment can perform risk classification based on the comprehensive risk value output by the DBN model and determine the real-time risk level corresponding to the power system. It is understood that the total number of risk levels and the risk level division range corresponding to each risk level can be determined according to the actual situation, and this application does not impose specific limitations on this. For ease of description, the following embodiments are illustrated using three risk levels: low, medium, and high.

[0110] When the real-time risk level exceeds the preset alarm level, such as a medium / high real-time risk level, the system enters the intelligent decision generation stage. In this stage, the computer equipment does not simply issue an alarm; instead, it identifies the key factors causing the risk spike and, based on these key risk factors, matches specific, actionable countermeasures from a preset decision rule base, thereby obtaining operational decision information. In this way, optimal safety control measures and decision-making schemes can be automatically generated or recommended, achieving intelligent decision-making and improving the accuracy and effectiveness of risk response.

[0111] Specifically, computer equipment can identify key risk factors using the concept of controlled variables. First, the computer equipment can determine the direct risk node R in the DBN model that directly leads to the overall risk value. Then, based on the direct risk node and the power system safety knowledge base, it can screen all parent or ancestor nodes that may significantly affect R from each node in the DBN model, using these as candidate key factor nodes X1, ..., Xn. These candidate key factor nodes can correspond to personnel qualifications, equipment status, environmental indicators, etc.

[0112] This application inputs the current twin state, State1, as evidence into the DBN model and runs an inference algorithm to calculate the actual probability value P_benchmark of node R being in a high-risk state, assuming all factors remain unchanged. This value serves as the benchmark risk probability. Next, this application can perform hypothetical scenario simulations, that is, for each key factor candidate node Xi (i=1, ..., n), the following operations are performed:

[0113] Step B71: Construct counterfactual evidence, that is, adjust the state information corresponding to Xi in State1 to a preset safe value to obtain State_i. For example, if Xi corresponds to personnel qualification, then set the personnel qualification information of State1 to "advanced and no violations"; if Xi corresponds to weather, then set the system environment data of State1 to "sunny and windless".

[0114] Step B73: Keep other state information unchanged, that is, except for Xi, the state of all other factors: Xj: (j≠i): remains the current true state.

[0115] Step B75: Input State_i into the DBN model, calculate the probability that node R is in a high-risk state at this time, and obtain the assumed probability value P_assuming_Xi_ideal corresponding to Xi.

[0116] After obtaining N hypothetical probability values, this application determines the degree of influence of each key factor candidate node on node R, thereby identifying key risk factors from the N key factor candidate nodes.

[0117] It is understood that the decision rule base can be set according to specific circumstances, and this application does not impose specific restrictions on it. For example, when the key risk factor is mismatched personnel qualifications, the operational decision information may be to replace the personnel with personnel Y who have certain qualifications; when the key risk factor is that the tools are about to expire, the operational decision information may be to retrieve the spare tool Z from the warehouse and send the current tool for inspection; when the key risk factor is a lightning warning, the operational decision information may be to postpone the original plan to a certain time and minute.

[0118] In some embodiments, if the real-time risk level is lower than a preset alarm level, for example, if the real-time risk level is low, the computer equipment can determine that the job should be executed as originally planned. In this case, the process directs to "Execute the job plan" and sends this signal back to the work order system or on-site personnel.

[0119] In some embodiments, key risk factors are determined based on N hypothetical probability values ​​and N key factor candidate nodes, including:

[0120] Step B91: Calculate the probability difference between each of the N hypothetical probability values ​​and the actual probability values;

[0121] Step B93: Select the probability difference that is greater than or equal to the preset contribution threshold from the N probability differences as the target difference;

[0122] Step B95: Determine the key risk factors based on the candidate nodes of key factors corresponding to the target difference.

[0123] In this embodiment, for each candidate node Xi of a key factor, this application can calculate the probability difference ΔPi between the assumed probability value and the actual probability value corresponding to Xi, i.e., ΔPi = P_benchmark - P_assuming_Xi_ideal. This probability difference reflects the degree of contribution of the risk factor corresponding to Xi to the final risk. The larger the probability difference, the greater the contribution of the risk factor to the current high risk, and the more critical the risk factor.

[0124] After obtaining N probability differences, this application can select probability differences that are greater than or equal to a preset contribution threshold as target differences, and designate the risk factors corresponding to the target differences as key risk factors. In some examples, the computer device can sort the N probability differences in descending order and filter them according to the sorting results and the preset contribution threshold.

[0125] In this embodiment, by quantitatively analyzing the probability differences, the key factors that have the most significant impact on power system security risks can be accurately identified. This data-driven decision-making approach not only avoids the subjectivity and bias of human judgment but also significantly improves the accuracy and efficiency of risk identification, ensuring the scientific and practical nature of the decision.

[0126] In some embodiments, the method provided in this application may further include the following steps:

[0127] In response to model optimization instructions, the system obtains historical operation execution results of the power system and updates the conditional probability relationships of the risk coupling model based on these results.

[0128] In this embodiment, to provide closed-loop feedback and optimization for the DBN model, this application can update the conditional probability table of the DBN model during operation using newly acquired historical job execution results, allowing the DBN model to continuously optimize from actual results. Through this continuous self-learning, the prediction accuracy of the DBN model will become increasingly higher, and its grasp of risk evolution patterns will become more and more accurate.

[0129] The following describes the power system security risk prediction device based on digital twins provided in the embodiments of this application. The power system security risk prediction device based on digital twins described below can be referred to in correspondence with the power system security risk prediction method based on digital twins described above.

[0130] In some embodiments, such as Figure 3 As shown, this application provides a power system security risk prediction device 300 based on digital twins, comprising:

[0131] Data acquisition module 302 is used to acquire in real time data on personnel, tools, plans and tickets, and system environment related to power system operations;

[0132] The digital twin update module 304 is used to update the digital twin status of the power safety digital twin based on the operator data, the tool data, the plan ticket data, and the system environment data, so that the power safety digital twin reflects the attributes of each entity and the relationships between entities in real time; wherein, the entities of the power safety digital twin include operators, tools, plan tickets, and system environment related to the power system operation;

[0133] The risk prediction module 306 is used to respond to the twin state update operation by inputting the updated twin state into a pre-built risk coupling model, so that the risk coupling model performs dynamic risk prediction based on the twin state and outputs a comprehensive risk value.

[0134] The task decision generation module 308 is used to generate task decision information based on the comprehensive risk value to guide the on-site operations of the workers.

[0135] In some embodiments, the twin state includes an entity state and a relationship state. The twin update module 304 of this application includes:

[0136] The first event determination unit is used to determine an entity attribute change event based on the operator data, the tool data, the plan ticket data, and the system environment data; wherein, the entity attribute change event includes a target entity identifier, an attribute identifier, and a changed attribute value;

[0137] The second event determination unit is used to determine entity relationship change events based on the operator data, the tool data, and the plan ticket data; wherein, the entity relationship change event includes associated entity identifier and relationship operation identifier;

[0138] The twin update unit is used to update the entity state and relationship state of the power safety digital twin based on the entity attribute change event and the entity relationship change event.

[0139] In some embodiments, the twin update unit of this application includes:

[0140] An entity attribute update unit is used to determine a target entity based on the target entity identifier, and update the entity attributes of the target entity according to the attribute identifier and the changed attribute value, so as to update the entity status of the power safety digital twin.

[0141] The verification unit is used to verify the entity relationship change event according to the associated entity identifier and the preset entity coupling rules; wherein, the entity coupling rules include at least one of personnel qualification verification rules, personnel status verification rules, tool status verification rules, environmental verification rules, and plan ticket verification rules;

[0142] The relationship status update unit is used to update the relationship status of the power safety digital twin based on the associated entity identifier and the relationship operation identifier when the verification passes.

[0143] The alarm unit is used to generate an alarm message based on the verification result when the verification fails, and to push the alarm message to the alarm terminal.

[0144] In some embodiments, the risk coupling model is a dynamic Bayesian network model, wherein the nodes of the risk coupling model include various security risk factors, and the risk coupling model records the conditional probability relationships of each node.

[0145] In some embodiments, the job decision generation module 308 of this application includes:

[0146] The risk level determination unit is used to determine the real-time risk level based on the comprehensive risk value and the preset risk level division range;

[0147] The candidate node determination unit is used to determine the direct risk node among the nodes of the risk coupling model if the real-time risk level is higher than the preset alarm level, and to select N key factor candidate nodes based on the direct risk node and the preset power system security domain knowledge base; where N is a positive integer.

[0148] The actual probability value acquisition unit is used to input the updated twin state into the risk coupling model to obtain the actual probability value of the direct risk node being in a high-risk state.

[0149] Assuming the probability value acquisition unit is used to adjust the state information corresponding to the candidate node of the key factor in the updated twin state to a preset safety value for each of the N key factor candidate nodes, and input the adjusted twin state into the risk coupling model to obtain the assumed probability value that the direct risk node is in a high-risk state;

[0150] The task decision generation unit is used to determine key risk factors based on N hypothetical probability values ​​and N key factor candidate nodes, and to generate task decision information based on the key risk factors and a preset decision rule base.

[0151] In some embodiments, the job decision generation unit of this application includes:

[0152] A probability difference calculation unit is used to calculate the probability difference between each of the N hypothetical probability values ​​and the actual probability values.

[0153] The target difference filtering unit is used to filter out the probability differences that are greater than or equal to a preset contribution threshold from the N probability differences as target differences;

[0154] The key risk factor determination unit is used to determine the key risk factors based on the key factor candidate nodes corresponding to the target difference.

[0155] In some embodiments, the power system security risk prediction device 300 based on digital twin of this application further includes:

[0156] The model update module is used to respond to model optimization instructions, obtain historical operation execution results of the power system, and update the conditional probability relationship of the risk coupling model based on the historical operation execution results.

[0157] In one embodiment, this application also provides a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the power system security risk prediction method based on digital twins as described in any embodiment.

[0158] In one embodiment, this application also provides a computer device storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the power system security risk prediction method based on digital twins as described in any embodiment.

[0159] Indicatively, Figure 4 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of this application. In one example, the computer device can be a server. (Refer to...) Figure 4 The computer device 900 includes a processing component 902, which further includes one or more processors, and memory resources represented by memory 901 for storing instructions executable by the processing component 902, such as application programs. The application programs stored in memory 901 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 902 is configured to execute instructions to perform the steps of the digital twin-based power system security risk prediction method described in any of the above embodiments.

[0160] The computer device 900 may also include a power supply component 903, a wired or wireless network interface 904, and an input / output (I / O) interface 905, wherein the power supply component 903 is configured to perform power management of the computer device 900, and the network interface 904 is configured to connect the computer device 900 to a network. The computer device 900 can operate on an operating system stored in memory 901, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or similar.

[0161] Those skilled in the art will understand that the internal structure of the computer device shown in this application is merely a block diagram of a portion of the structure related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0162] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. In this document, "a," "an," "the," "the," and "its" may also include plural forms unless the context clearly indicates otherwise. "Multiple" refers to at least two, such as 2, 3, 5, or 8, etc. "And / or" includes any and all combinations of the related listed items.

[0163] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0164] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A power system safety risk prediction method based on digital twinning, characterized in that, The method comprises: real-time acquisition of work personnel data, tool data, plan ticket data and system environment data related to power system operation; updating of a twin state of a power safety digital twin according to the work personnel data, the tool data, the plan ticket data and the system environment data, so that the power safety digital twin reflects self attributes of each entity and inter-entity correlations in real time; wherein entities of the power safety digital twin include work personnel, tools, plan tickets and system environment related to the power system operation; in response to the twin state updating operation, input of the updated twin state into a pre-constructed risk coupling model, so that the risk coupling model performs dynamic risk prediction based on the twin state and outputs a comprehensive risk value; generation of work decision information for guiding on-site work of work personnel according to the comprehensive risk value.

2. The method of claim 1, wherein, The twin state comprises entity states and relationship states. The updating of the twin state of the power safety digital twin according to the work personnel data, the tool data, the plan ticket data and the system environment data comprises: determination of entity attribute change events according to the work personnel data, the tool data, the plan ticket data and the system environment data; wherein the entity attribute change events include target entity identifiers, attribute identifiers and changed attribute values; determination of entity relationship change events according to the work personnel data, the tool data and the plan ticket data; wherein the entity relationship change events include associated entity identifiers and relationship operation identifiers; updating of entity states and relationship states of the power safety digital twin based on the entity attribute change events and the entity relationship change events.

3. The method of claim 2, wherein, The updating of the entity states and the relationship states of the power safety digital twin based on the entity attribute change events and the entity relationship change events comprises: determination of a target entity according to the target entity identifier, and updating of an entity attribute of the target entity according to the attribute identifier and the changed attribute value, to update the entity states of the power safety digital twin; verification of the entity relationship change events according to the associated entity identifiers and preset entity coupling rules; wherein the entity coupling rules include at least one of a personnel qualification verification rule, a personnel state verification rule, a tool state verification rule, an environment verification rule and a plan ticket verification rule; in the case of verification passing, updating of the relationship states of the power safety digital twin according to the associated entity identifiers and the relationship operation identifiers; in the case of verification failing, generation of an alarm prompt message according to a verification result, and pushing of the alarm prompt message to an alarm terminal.

4. The method according to any one of claims 1 to 3, characterized in that, The risk coupling model is a dynamic Bayesian network model, nodes of the risk coupling model include various safety risk factors, and the risk coupling model records conditional probability relationships of each node.

5. The method of claim 4, wherein, The generation of the work decision information for guiding on-site work of work personnel according to the comprehensive risk value comprises: determining a real-time risk level according to the comprehensive risk value and a preset risk level division interval; if the real-time risk level is higher than a preset alarm level, determining a direct risk node in each node of the risk coupling model, and screening N key factor candidate nodes according to the direct risk node and a preset power system safety domain knowledge base; wherein N is a positive integer; inputting the updated twin state into the risk coupling model to obtain an actual probability value of the direct risk node being in a high risk state; for each of the N key factor candidate nodes, adjusting state information corresponding to the key factor candidate node in the updated twin state to a preset safety value, and inputting the adjusted twin state into the risk coupling model to obtain a hypothetical probability value of the direct risk node being in a high risk state; determining a key risk factor according to N hypothetical probability values and N key factor candidate nodes, and generating the operation decision information based on the key risk factor and a preset decision rule base.

6. The method of claim 5, wherein, The method further comprises: in response to a model optimization instruction, obtaining a historical operation execution result of the power system, and updating a conditional probability relationship of the risk coupling model according to the historical operation execution result. comprises: a data acquisition module configured to acquire, in real time, operation personnel data, tool data, plan ticket data, and system environment data related to power system operation; 7. The method of claim 4, wherein, a twin updating module configured to update a twin state of a power safety digital twin according to the operation personnel data, the tool data, the plan ticket data, and the system environment data, so that the power safety digital twin reflects, in real time, self attributes of each entity and an association relationship between entities; wherein the entities of the power safety digital twin include operation personnel, tools, plan tickets, and system environment related to the power system operation; a risk prediction module configured to, in response to a twin state updating operation, input an updated twin state into a pre-constructed risk coupling model, so that the risk coupling model performs dynamic risk prediction based on the twin state and outputs a comprehensive risk value; 8.A power system safety risk prediction device based on digital twinning, characterized by, an operation decision generation module configured to generate operation decision information for guiding on-site operation of operation personnel according to the comprehensive risk value. The storage medium stores computer readable instructions, and the computer readable instructions, when executed by one or more processors, cause the one or more processors to perform the steps of the power system safety risk prediction method based on digital twin according to any one of claims 1 to 7. comprises: one or more processors, and a memory; ​ 9. A storage medium, characterized by ​ 10. A computer device, comprising: ​ ​ The memory stores computer readable instructions which, when executed by the one or more processors, perform the steps of the digital-twin-based power system security risk prediction method of any one of claims 1 to 7.

Images