Substation field safety control system and method, storage medium and computer equipment

Abstract

The invention discloses a substation site safety control system and method, a storage medium and computer equipment, and the method comprises the steps that a sensing layer collects multi-source heterogeneous data of a substation site in real time, and the multi-source heterogeneous data comprises equipment state data, personnel position data and external environment data; the data and intelligent layer performs fusion processing on the multi-source heterogeneous data, constructs a digital twinborn model synchronized with a physical entity of the substation site in real time based on a fusion processing result, and performs security risk prediction on the digital twinborn model through an AI engine, the safety risk prediction comprises fault prediction of an equipment state and behavior prediction of personnel entering a dangerous area; and the application and control layer generates a control instruction according to a prediction result of the AI engine, and issues the control instruction to corresponding control equipment on the site of the transformer substation so as to carry out risk reminding and / or risk control based on the control equipment.

Description

Technical Field

[0001] This application relates to the field of safety control technology, and in particular to a substation field safety control system and method, storage medium, and computer equipment. Background Technology

[0002] With the rapid construction of ultra-high-voltage transmission networks and the in-depth advancement of smart grids, substations, as core hubs of the power system, are becoming increasingly larger and more complex. Numerous operating devices, complex electromagnetic environments, and frequent interactions between personnel and machinery create a high-risk working environment. Traditional safety management models heavily rely on manual inspections and fixed monitoring systems, making it difficult to achieve real-time, accurate, and proactive control of safety risks. This has become a bottleneck restricting the improvement of intelligent operation and maintenance and the inherent safety level of substations.

[0003] Currently, most existing substation safety control systems rely on 3D visualization models and electronic fence technology. These systems construct a 3D model of the substation, delineate hazardous areas within the model, and use positioning technology to monitor the real-time location of inspection personnel and equipment. An alarm is triggered upon intrusion. However, this approach has significant limitations: First, it is essentially a "passive response" mode, only issuing alarms when a risk has already occurred or is about to occur, lacking the ability to predict potential equipment failures, inspection personnel behavior trends, and environmental changes that may trigger risks. Second, the system's functionality is relatively limited, with various monitoring data (such as equipment status, personnel location, and environmental parameters) forming "information silos," making multi-dimensional correlation analysis impossible and insufficient to address complex systemic risks arising from the coupling of multiple factors. Summary of the Invention

[0004] In view of this, this application provides a substation field safety control system and method, storage medium, and computer equipment. Through a multi-layered architecture design, it achieves full-process automation and intelligence from data acquisition and analysis to control, greatly improving the efficiency and accuracy of substation field safety control. The multi-source heterogeneous data acquisition of the perception layer comprehensively covers all types of information at the substation site, ensuring no safety blind spots. The data fusion processing and digital twin model construction of the data and intelligence layers provide strong technical support for safety risk prediction, enabling early detection of potential safety hazards. The application and control layers quickly generate and issue control commands based on the prediction results, achieving timely response and effective control of safety risks, minimizing the possibility of safety accidents, ensuring the safe and stable operation of the substation, and providing reliable protection for the lives of personnel. The embodiments of this application form an intelligent closed-loop control from perception, decision-making to execution, fundamentally improving the safety management efficiency and intelligence level of substation sites.

[0005] According to one aspect of this application, a substation field safety control system is provided, comprising: The perception layer is used to collect multi-source heterogeneous data from the substation site in real time, including equipment status data, personnel location data, and external environment data. The data and intelligence layer is used to fuse the multi-source heterogeneous data, construct a digital twin model that is synchronized in real time with the physical entity of the substation site based on the fusion processing result, and perform safety risk prediction on the digital twin model through an AI engine. The safety risk prediction includes fault prediction of equipment status and behavior prediction of personnel entering dangerous areas. The application and control layer is used to generate control commands based on the prediction results of the AI ​​engine, and send the control commands to the corresponding control equipment at the substation site, so as to perform risk alerts and / or risk control based on the control equipment.

[0006] According to another aspect of this application, a substation on-site safety control method is provided, comprising: The sensing layer collects multi-source heterogeneous data from the substation site in real time, including equipment status data, personnel location data, and external environment data. The data and intelligence layer performs fusion processing on the multi-source heterogeneous data, constructs a digital twin model that is synchronized in real time with the physical entity at the substation site based on the fusion processing results, and performs safety risk prediction on the digital twin model through an AI engine. The safety risk prediction includes fault prediction of equipment status and prediction of personnel behavior when entering dangerous areas. The application and control layer generates control commands based on the prediction results of the AI ​​engine and sends the control commands to the corresponding control equipment at the substation site to perform risk alerts and / or risk control based on the control equipment.

[0007] According to another aspect of this application, a storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the above-described substation field safety control method.

[0008] According to another aspect of this application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the above-described substation on-site safety control method.

[0009] By employing the aforementioned technical solutions, this application provides a substation field safety control system and method, storage medium, and computer equipment. Through a multi-layered architecture design, it achieves full-process automation and intelligence from data acquisition and analysis to control, significantly improving the efficiency and accuracy of substation field safety control. The multi-source heterogeneous data acquisition at the perception layer comprehensively covers all types of information at the substation site, ensuring no safety blind spots. The data fusion processing and digital twin model construction at the data and intelligence layers provide strong technical support for safety risk prediction, enabling early detection of potential safety hazards. The application and control layers rapidly generate and issue control commands based on the prediction results, achieving timely response and effective control of safety risks, minimizing the possibility of safety accidents, ensuring the safe and stable operation of the substation, and providing reliable protection for the lives of personnel. This application's embodiments form an intelligent closed-loop control system from perception and decision-making to execution, fundamentally improving the safety management efficiency and intelligence level of substation sites.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0011] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This paper shows a schematic diagram of the structure of a substation field safety control system provided in an embodiment of this application; Figure 2 A flowchart illustrating a substation field safety control method provided in an embodiment of this application is shown. Figure 3 A schematic diagram of the device structure of a computer device provided in an embodiment of this application is shown. Detailed Implementation

[0012] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0013] This embodiment provides a substation field safety control system, such as Figure 1 As shown, the system includes: The perception layer is used to collect multi-source heterogeneous data from the substation site in real time, including equipment status data, personnel location data, and external environment data. The data and intelligence layer is used to fuse the multi-source heterogeneous data, construct a digital twin model that is synchronized in real time with the physical entity of the substation site based on the fusion processing result, and perform safety risk prediction on the digital twin model through an AI engine. The safety risk prediction includes fault prediction of equipment status and behavior prediction of personnel entering dangerous areas. The application and control layer is used to generate control commands based on the prediction results of the AI ​​engine, and send the control commands to the corresponding control equipment at the substation site, so as to perform risk alerts and / or risk control based on the control equipment.

[0014] This application provides a substation field safety control system that, through a multi-layered architecture, achieves comprehensive perception, intelligent analysis, and precise control of the substation's on-site safety status. Specifically, the substation field safety control system mainly consists of a perception layer, a data and intelligence layer, and an application and control layer.

[0015] At the perception layer: This layer undertakes the crucial task of real-time acquisition of multi-source heterogeneous data from the substation site. This multi-source heterogeneous data encompasses various aspects of information. Equipment status data reflects the operating status of various electrical equipment within the substation, such as parameters like temperature, voltage, and current, helping to promptly identify potential equipment faults. Personnel location data accurately pinpoints the specific locations of on-site personnel, providing fundamental information for ensuring personnel safety. External environmental data includes weather conditions and changes in the surrounding environment, factors that may impact the safe operation of the substation. By comprehensively collecting this multi-source heterogeneous data, a rich and accurate information source is provided for subsequent safety control.

[0016] In the data and intelligence layer, this layer is primarily responsible for fusing and processing multi-source heterogeneous data collected by the perception layer. Due to the diverse sources and formats of the data, this fusion processing integrates these scattered data into a unified, organic whole that can be recognized by the digital twin system. Next, based on the fusion processing results, a digital twin model synchronized in real-time with the physical entities at the substation site is constructed. This digital twin model reflects the actual situation on-site in real time. Simultaneously, an AI engine is used to predict safety risks on the digital twin model. Safety risk prediction includes two important aspects: first, predicting equipment failures by analyzing historical and real-time operational data to anticipate potential failure types and times, allowing for timely maintenance; second, predicting personnel behavior entering hazardous areas by predicting the likelihood of personnel entering hazardous areas based on personnel location data and pre-set safety rules, thus issuing early warnings.

[0017] In terms of the application and control layer: This layer can generate corresponding control commands based on the prediction results of the AI ​​engine in the data and intelligence layer. These control commands are specifically targeted, formulating corresponding countermeasures for different safety risk situations. After generating the control commands, they can be sent to the corresponding control equipment at the substation site. These control devices include, but are not limited to, alarm devices, automatic control switches, etc. Through these control devices, risk alerts and / or risk control can be performed based on the control commands. For example, when a potential equipment failure is predicted, the control equipment can issue an alarm to remind personnel to inspect it; when a potential personnel entry into a dangerous area is predicted, the control equipment can promptly prevent personnel from entering or issue a strong warning signal.

[0018] By applying the technical solution of this embodiment and through a multi-layered architecture design, the entire process from data acquisition and analysis to control is automated and intelligent, greatly improving the efficiency and accuracy of on-site safety control in substations. The multi-source heterogeneous data acquisition at the perception layer comprehensively covers all types of information at the substation site, ensuring no safety blind spots. The data fusion processing and digital twin model construction at the data and intelligence layers provide strong technical support for safety risk prediction, enabling the early detection of potential safety hazards. The application and control layers quickly generate and issue control commands based on the prediction results, achieving timely response and effective control of safety risks, minimizing the possibility of safety accidents, ensuring the safe and stable operation of the substation, and providing reliable protection for the lives of personnel. This embodiment forms an intelligent closed-loop control system from perception and decision-making to execution, fundamentally improving the safety management efficiency and intelligence level of substation sites.

[0019] Optionally, in this embodiment, the data and intelligence layer is configured to: for each data source, determine the communication protocol corresponding to the data source, and determine a data format conversion strategy based on the communication protocol; convert the real-time collected data corresponding to the data source into standard format data recognizable by the digital twin system according to the data format conversion strategy; perform quality assessment on the standard format data based on a preset data quality assessment model; and input the standard format data that meets the quality requirements after assessment into the digital twin system to synchronously update the digital twin model based on the standard format data.

[0020] In this embodiment, data processing and model updating are crucial aspects of the data and intelligence layer of the substation field safety control system, directly impacting the accuracy and effectiveness of subsequent safety risk predictions. Specifically, the data and intelligence layer achieves synchronous updates of the digital twin model through the following process: Since the data collected at the substation site comes from multiple different data sources, each data source may use different communication protocols for data transmission. Communication protocols define the rules and formats for data transmission, and the data formats under different protocols vary greatly. Therefore, the first step is to accurately determine the communication protocol corresponding to each data source. After clarifying the communication protocol, a preset data format conversion strategy corresponding to that protocol can be further determined. Here, the data format conversion strategy can be formulated based on the characteristics and requirements of the communication protocol, capable of converting real-time acquired data in various formats into a standard format that the digital twin system can recognize and understand. For example, some devices use the Modbus protocol to transmit data, and its data format differs from the format required by the digital twin system. Therefore, a data format conversion strategy corresponding to the Modbus protocol can be determined. Through this data format conversion strategy, real-time acquired data under the Modbus protocol can be converted into standard format data, thereby ensuring that the data can flow and be processed smoothly within the digital twin system.

[0021] Even after data has been converted to a standard format, it cannot be used directly because data quality may vary. To ensure that the data input into the digital twin system is reliable and accurate, a quality assessment can be performed on the standard format data based on a pre-defined data quality assessment model. This model can check the data from multiple dimensions, such as completeness (checking for missing key information), accuracy (determining if the data matches reality), and timeliness (ensuring the data is recently collected). Through comprehensive evaluation, data that meets the quality requirements is selected, providing a reliable guarantee for subsequent model updates.

[0022] After data quality assessment, standard-format data that meets the quality requirements is input into the digital twin system. The digital twin model is a virtual mapping of the substation site. To ensure it accurately reflects the actual situation in real time, the model is synchronously updated based on newly input data, keeping it consistent with the real-world substation conditions. Through this synchronous updating, the digital twin model can promptly capture changes in equipment status, personnel locations, and the external environment at the substation site, providing the latest and most accurate data foundation for subsequent safety risk predictions based on the model.

[0023] In a specific embodiment, the core data source may include: (1) Sensor network data: various sensors deployed in key equipment and environment of substations, such as temperature, humidity, vibration, SF6 gas concentration, partial discharge, etc., continuously outputting equipment status and environmental parameters. (2) Inspection robot data: inspection robots equipped with high-definition visible light cameras, infrared thermal imagers and sound sensors provide multi-dimensional dynamic data such as equipment appearance, thermodynamic status and abnormal noise. (3) Existing industrial control system data: integrating the data streams of existing power monitoring and data acquisition systems (SCADA) and anti-misoperation systems. These data include equipment operating parameters (such as voltage, current, active power) and operating instructions, which are an important part of ensuring system stability. (4) External environment data: accessing external data sources such as meteorological stations and earthquake monitoring networks to obtain information such as meteorological conditions and disaster warnings, providing a basis for macro-risk assessment.

[0024] To unify the processing of data from different devices using different communication protocols, a layered protocol conversion and fusion strategy can be adopted. For Internet of Things (IoT) devices, their data communication is based on lightweight application layer protocols such as MQTT and CoAP. The MQTT protocol uses a publish / subscribe model, suitable for environments with limited bandwidth and unstable networks; the CoAP protocol is similar to HTTP, but based on the UDP protocol, specifically designed for devices with limited resources such as CPU, RAM, and Flash, to achieve efficient information exchange. Furthermore, for short-range communication protocols such as Bluetooth, ZigBee, and LoRa, device data must first access a gateway, which then converts the data stream into a protocol supported by the TCP / IP protocol suite (such as MQTT) before uploading it to the cloud or local platform. For traditional industrial control systems, dedicated interfaces and parsing engines are used to deeply parse protocols such as Modbus, IEC 61850, and DNP3 to ensure the compliance and security of data packets. Through this protocol fusion and conversion mechanism, all data streams are standardized, ensuring the uniformity and compatibility of the digital twin system input.

[0025] This application's embodiments achieve seamless integration of data from different data sources by determining communication protocols and data format conversion strategies, solving the problem of incompatibility between multi-source heterogeneous data and enabling various types of data to be effectively utilized in the digital twin system. The pre-set data quality assessment model strictly controls data quality, preventing low-quality data from interfering with the digital twin model and ensuring its accuracy and reliability. Synchronous updates to the digital twin model based on data meeting quality requirements ensure that the model can reflect the actual situation at the substation in real time, providing solid data support for safety risk prediction. This significantly improves the level and efficiency of on-site safety control at the substation and effectively reduces the risk of safety accidents.

[0026] In this embodiment of the application, optionally, the data and intelligence layer performs quality assessment on the standard format data of the target data source based on the following method: determining the total number of fields in the target data source, and based on the standard format data of the target data source, identifying the number of fields assigned values, calculating a first ratio of the number of fields assigned values ​​to the total number of fields according to a field assignment quality assessment model, and determining the field assignment quality of the standard format data based on the relationship between the first ratio and a first preset ratio threshold, wherein the preset data quality assessment model includes the field assignment quality assessment model; and / or, determining the total number of records in the target data source, and based on the standard format data of the target data source, identifying the number of records with complete field assignments, calculating a second ratio of the number of records with complete field assignments to the total number of records according to an assignment integrity quality assessment model, and determining the standard format data based on the relationship between the second ratio and a second preset ratio threshold. The data assignment integrity quality is assessed by: (1) determining the total amount of data from the target data source, and (2) identifying the number of data items that meet the data correctness requirements based on the standard format data from the target data source, (3) calculating a third ratio of the number of data items that meet the data correctness requirements to the total amount of data based on the data correctness quality assessment model, and (4) determining the data correctness quality of the standard format data based on the relationship between the third ratio and a third preset ratio threshold, wherein the preset data quality assessment model includes the data correctness quality assessment model; and (5) calculating the kurtosis value corresponding to the standard format data from the target data source based on the data kurtosis quality assessment model, and (6) determining the data kurtosis quality of the standard format data based on the relationship between the kurtosis value and a preset kurtosis threshold, wherein the preset data quality assessment model includes the data kurtosis quality assessment model.

[0027] In this embodiment, in the data and intelligence layer of the substation field safety control system, quality assessment of the standard format data from the target data source is a crucial step in ensuring data reliability and validity. Accurate data quality assessment ensures the high precision and reliability of subsequent operations such as building digital twin models and predicting safety risks based on this data. Specifically, for the standard format data from the target data source, quality assessment can be performed using one or more of the following methods (the target data source can be any one of multiple data sources; this embodiment is based on the standard format data from one of these data sources for quality assessment): The first method involves first determining the total number of fields in the target data source. Next, based on standard format data, identify the number of fields that have already been assigned values, i.e., the number of fields with actual content. Then, use a field assignment quality assessment model to calculate a first ratio of the number of assigned fields to the total number of fields. This ratio reflects the degree to which the data is populated at the field level. Finally, compare this first ratio with a pre-set first preset ratio threshold. If the first ratio is greater than or equal to the first preset ratio threshold, it indicates that most fields have been assigned values, and the field assignment quality is high; conversely, it indicates that many fields are not assigned values, and the field assignment quality is poor. This method allows for a precise assessment of the quality of field assignment in standard format data.

[0028] The second method involves first determining the total number of records in the target data source, for example, counting the "rows" of standard format data. Then, identify the number of records in the standard format data with all fields fully assigned values, i.e., the number of records where all fields in a "row" have values. Using a value assignment integrity quality assessment model, calculate a second ratio between the number of records with fully assigned fields and the total number of records. This ratio reflects the completeness of the data records. Compare this second ratio with a second preset threshold. If the second ratio reaches or exceeds the second preset threshold, it means that most records have all fields fully assigned values, indicating good value assignment integrity quality; if it is below the threshold, it indicates that many records have incomplete field assignments, indicating a problem with value assignment integrity quality. This method is used to assess the quality of standard format data in terms of record value assignment integrity.

[0029] The third method involves first determining the total amount of data from the target data source, i.e., the overall scale of the data. Then, based on standard format data, the number of data points that meet the data correctness requirements is selected. These correctness requirements can be set according to the actual conditions at the substation and relevant standards, such as whether the data is within a reasonable range. Next, a data correctness quality assessment model is used to calculate a third ratio between the number of data points meeting the correctness requirements and the total amount of data. This ratio reflects the proportion of correct data in the dataset. This third ratio is compared to a third preset ratio threshold. When the third ratio is greater than or equal to the third preset ratio threshold, it indicates high data correctness; conversely, it indicates low data correctness. This method evaluates the quality of standard format data in terms of correctness.

[0030] The fourth type: Kurtosis reflects the peak or spiky nature of the data distribution. Using a kurtosis quality assessment model, the kurtosis value corresponding to standard-format data from a target data source can be calculated. This kurtosis value reflects the morphological characteristics of the data distribution. Then, the calculated kurtosis value is compared with a preset kurtosis threshold. Based on the comparison result, the data kurtosis quality can be judged. For example, if the kurtosis value is within the preset range, it indicates that the data distribution is relatively normal and the data kurtosis quality is good; if it exceeds the range, it may indicate an abnormal data distribution and poor data kurtosis quality. This method assesses data quality from the perspective of data distribution morphology.

[0031] This application's embodiments cover multiple aspects, including field assignment quality, assignment integrity quality, data accuracy quality, and data kurtosis quality. It can comprehensively and meticulously evaluate the quality of standard-format data from the target data source, avoiding the potential bias of a single evaluation dimension. By comparing with preset thresholds, it is possible to clearly determine whether the data quality meets the standards, providing a clear basis for subsequent data processing and application. This evaluation method helps to promptly identify problems in the data, such as missing fields, incomplete records, data errors, and abnormal distributions. This allows for correction and optimization before the data enters the digital twin model, improving the accuracy and reliability of the digital twin model, thereby enhancing the performance and safety of the entire substation's on-site safety control system and ensuring the stable operation of the substation.

[0032] In this embodiment of the application, optionally, the data and intelligence layer is further configured to: obtain the equipment status data time series of the target device from the digital twin model, and input the equipment status data time series into the AI ​​engine to obtain the future equipment status and potential fault risks of the target device; obtain the personnel location data of maintenance personnel from the digital twin model, determine the real-time location and movement trajectory of the maintenance personnel based on the personnel location data, and determine the predicted trajectory of the maintenance personnel through the AI ​​engine based on the real-time location and movement trajectory, perform collision detection between the predicted trajectory and the dangerous area vector library defined in the digital twin model to obtain the behavioral risk of the maintenance personnel entering the dangerous area within a preset time period in the future; construct an environmental data time series based on the external environment data mapped to the spatial layout of the digital twin model, and input the environmental data time series into the AI ​​engine to obtain extreme environment prediction results, and analyze the potential fault risks of specific equipment in the substation based on the extreme environment prediction results.

[0033] In this embodiment, the data and intelligence layer of the substation field safety control system not only processes and assesses data quality, but also undertakes the important task of risk prediction and assessment using digital twin models and AI engines. By deeply mining various types of data in the digital twin model and combining them with the powerful analytical capabilities of the AI ​​engine, the data and intelligence layer can identify potential safety risks in advance, providing strong protection for the safe and stable operation of the substation.

[0034] A digital twin model is a precise virtual mapping of a substation, storing rich status data of the target equipment. Here, the target equipment can be various electrical devices in the substation. For each target device, the time series of its equipment status data is first extracted from the digital twin model. This data includes operational status information at different times, such as temperature, pressure, and current. The time series of equipment status data is then input into an AI engine. Leveraging its powerful data analysis and machine learning capabilities, the AI ​​engine can perform in-depth mining and analysis of the equipment status data time series. By learning the patterns and trends of data changes, the AI ​​engine can predict the equipment status of the target device in the future and identify potential fault risks. For example, if the equipment temperature data shows a continuous upward trend over a period of time, the AI ​​engine may predict that the equipment will experience an overheating fault at some point in the future, thus issuing an early warning so that staff can take timely measures for inspection and maintenance to prevent the fault from occurring.

[0035] Digital twin models not only record the status information of the target equipment but also include the location data of maintenance personnel. By extracting this data from the digital twin model, the real-time location and movement trajectory of the maintenance personnel can be accurately determined. The AI ​​engine utilizes this real-time information, combined with its prediction algorithms, to infer the predicted trajectory of the maintenance personnel over a future period. Simultaneously, the digital twin model defines a hazardous area vector library, which may include areas near high-voltage equipment, flammable and explosive zones, etc. Collision detection is performed between the predicted trajectory of the maintenance personnel and the hazardous area vector library. If an overlap is found, it means that the maintenance personnel are likely to enter the hazardous area within a preset time period, thus determining their behavioral risk. This detection method can identify potential safety risks faced by maintenance personnel in advance, promptly reminding them to adjust their routes to avoid entering hazardous areas and ensuring their personal safety.

[0036] Substation operation can also be significantly affected by the external environment, such as extreme weather and temperature changes. Therefore, external environmental data can be mapped onto the spatial layout of a digital twin model, visually demonstrating the relationship between the external environment and substation equipment. Based on this mapping data, an environmental data time series is constructed, recording changes in the external environment at different points in time. This time series is then input into an AI engine, which analyzes the trends and patterns of environmental data changes and predicts potential extreme environmental conditions, such as heavy rain, heavy snow, and high temperatures. Based on these extreme environment predictions, the potential failure risks of specific equipment in the substation under extreme environments can be further analyzed. For example, in high-temperature environments, the heat dissipation performance of some equipment may be affected, leading to overheating and malfunctions; in heavy rain, the substation's waterproofing facilities may be tested, posing safety hazards such as electrical leakage. Through this analysis, preventative measures can be taken in advance to reduce the impact of extreme environments on substation equipment and ensure the safe operation of the substation. Here, "specific equipment" refers to equipment significantly affected by extreme weather.

[0037] This application fully leverages the comprehensiveness and accuracy of digital twin models, enabling the acquisition of data on equipment, personnel, and the external environment, providing a rich information foundation for risk prediction. Combined with the powerful analytical capabilities of the AI ​​engine, it achieves accurate predictions of equipment status, personnel behavior, and extreme environments across multiple dimensions, allowing for the early detection of potential safety risks and providing forward-looking decision-making support for substation safety management. Through collision detection and risk analysis, it enables the development of corresponding countermeasures for different risk scenarios, effectively reducing the probability of accidents and improving the safety and reliability of substations.

[0038] Optionally, in this embodiment, the data and intelligence layer is further configured to: extract standard format data sets from at least two data sources from the digital twin model, calculate the Pearson correlation coefficient between each pair of standard format data sets, determine the correlation between the two standard format data sets based on the Pearson correlation coefficient, wherein the data sources include at least two of equipment status data, personnel behavior data, and external environment data; when the prediction result of the AI ​​engine indicates a security risk, determine the data source on which the security risk prediction is based, and determine the associated data source whose correlation with the data source is greater than a preset correlation threshold, and generate a linkage control instruction based on the associated data source, wherein the linkage control instruction is used to provide risk warnings and / or risk control to the control device corresponding to the associated data source.

[0039] In this embodiment, within the data and intelligence layer of the substation field safety control system, to more comprehensively and deeply mine the value of data and achieve accurate risk prediction and effective linkage control, this layer also undertakes two key tasks. On the one hand, by analyzing the correlation between different data sources, it provides a basis for understanding the intrinsic connections between data; on the other hand, when a safety risk is predicted, it generates linkage control commands based on data correlation to ensure the safe and stable operation of the substation.

[0040] Specifically, the digital twin model, as a virtual mapping of the substation's on-site environment, aggregates a rich variety of data, covering equipment status data, personnel behavior data, and external environmental data. Standardized format data from at least two different data sources are extracted from the digital twin model. The standardized format data from each data source forms a standardized data set. These standardized data sets are data sets that have been formalized and are convenient for subsequent analysis. Next, the Pearson correlation coefficient is calculated between each pair of standardized format data sets. The Pearson correlation coefficient is a statistic used to measure the degree of linear correlation between two variables, with a value ranging from -1 to 1. When the correlation coefficient is close to 1, it indicates a strong positive correlation between the two data sets, meaning that an increase in one data set often accompanies an increase in the other. When the correlation coefficient is close to -1, it indicates a strong negative correlation, meaning that an increase in one data set leads to a decrease in the other. When the correlation coefficient is close to 0, it indicates a weak linear correlation between the two data sets. By calculating the Pearson correlation coefficient, the correlation between standardized format data sets from different data sources can be accurately determined, thereby understanding the intrinsic relationships between different types of data.

[0041] During operation, the AI ​​engine can predict various situations in the substation. When the prediction indicates a safety risk, it first clarifies which data sources the prediction is based on. For example, if a device is predicted to malfunction, the data sources might include historical operating data and real-time monitoring data. After identifying the data sources, it further searches for related data sources with a correlation greater than a preset correlation threshold. This preset correlation threshold is a standard value set based on actual needs and experience; only data sources with a correlation exceeding this value are considered related data sources. Finally, it generates linkage control commands based on these related data sources. The purpose of these linkage control commands is to provide risk alerts and / or risk control to the control equipment corresponding to the related data sources. For example, if the related data source involves personnel operation data, a risk alert can be sent to relevant personnel, informing them that their current operation may be risky; if the related data source involves equipment control data, the equipment can be automatically adjusted, such as reducing operating power or activating backup equipment, to reduce safety risks and ensure the safe operation of the substation.

[0042] In terms of data correlation analysis, this application's embodiments, by calculating the Pearson correlation coefficient between standard format data sets from different data sources, can deeply explore the intrinsic connections between data, providing strong support for understanding the complex relationships between equipment, personnel, and the environment in substations. This helps to discover potential influencing factors and patterns, providing a more comprehensive basis for subsequent risk prediction and decision-making. In terms of safety risk response, when the AI ​​engine predicts a safety risk, it can quickly identify the relevant data source and generate linkage control instructions. This linkage control method based on data correlation can timely and accurately provide risk warnings and controls to related equipment and personnel, effectively reducing the probability and scope of safety risks.

[0043] Optionally, in this embodiment, the data and intelligence layer is further configured to: acquire raw point cloud data of the substation collected by 3D laser scanning and SLAM motion measurement technology, and preprocess the raw point cloud data to obtain a preprocessed point cloud, wherein the preprocessing includes at least one of noise removal, outlier removal, and downsampling; register the preprocessed point cloud from different perspectives using an iterative nearest-point algorithm, and after completing the point cloud registration, calculate the root mean square error of the point pair corresponding to the registration result, and obtain a quantitative evaluation result of the registration result based on the root mean square error of the point pair; when the quantitative evaluation result meets preset requirements, construct an initial digital twin model in a unified coordinate system based on the registration result, and synchronize the fusion processing result to the initial digital twin model in real time to obtain a digital twin model that is synchronized in real time with the physical entity of the substation site.

[0044] In this embodiment, the data and intelligence layer can specifically construct a digital twin model that is synchronized in real time with the physical entities at the substation site in the following manner.

[0045] 3D laser scanning and SLAM (Simultaneous Localization and Mapping) mobile measurement technology are important means of acquiring spatial information of substations. 3D laser scanning can quickly and accurately acquire large amounts of point cloud data of object surfaces, while SLAM mobile measurement technology can determine its own position in real time and build a map of the surrounding environment during movement. The combination of the two can comprehensively and meticulously collect raw point cloud data of the substation. However, raw point cloud data often contains a large amount of noise and outliers, which can affect the accuracy and efficiency of subsequent modeling. Therefore, preprocessing of the raw point cloud data is necessary. Preprocessing includes at least one of the following operations: noise removal, outlier removal, and downsampling. Noise removal aims to eliminate random noise points caused by environmental interference, equipment errors, etc.; outlier removal removes points that significantly deviate from the normal data range, which may be due to measurement errors or external interference; downsampling reduces the amount of data while maintaining the characteristics of the point cloud data, improving the efficiency of subsequent processing. Through preprocessing, a higher-quality preprocessed point cloud more suitable for modeling can be obtained.

[0046] Since point cloud data is collected from different perspectives, these preprocessed point clouds are not uniform in spatial coordinates, requiring registration to build a complete model. Specifically, the Iterative Closest Point (TLP) algorithm can be applied for point cloud registration. This algorithm continuously finds the closest point pairs between two preprocessed point clouds and calculates their transformation relationship, gradually aligning the preprocessed point clouds from different perspectives to the same coordinate system. After point cloud registration is completed, to evaluate the accuracy and quality of the registration, the root mean square error (RMSE) of the corresponding point pairs can be calculated. The RMSE reflects the average distance deviation between corresponding point pairs after registration; the smaller the error, the more accurate the registration result. Based on the calculated RMSE, a quantitative evaluation of the registration result can be obtained, thereby determining whether the registration meets the expected requirements.

[0047] When the quantitative evaluation results of the registration meet the preset requirements, it indicates that the preprocessed point clouds from different perspectives have been accurately aligned. At this point, an initial digital twin model in a unified coordinate system can be constructed based on the registration results. This initial digital twin model is a preliminary mapping of the substation's physical entities in virtual space, but further fusion and processing of various data are needed to achieve real-time synchronization. Therefore, the fusion processing results can be synchronized to the initial digital twin model in real time, and real-time data from the substation site, such as equipment status data and environmental data, can be continuously updated into the digital twin model, enabling it to reflect the real-time state changes of the substation's physical entities. In this way, a digital twin model that is synchronized in real time with the physical entities at the substation site is finally obtained. This model can provide an intuitive and accurate virtual reference for the substation's operation and maintenance management, helping staff to promptly identify potential problems and make decisions.

[0048] In this embodiment of the application, optionally, the data and intelligence layer is further configured to: compare the similarity between the rendered image of the digital twin model and the real image based on structural similarity index and peak signal-to-noise ratio index to obtain the visual fidelity corresponding to the digital twin model; calculate the transmission performance value corresponding to the digital twin model based on latency index and jitter index; determine whether the synchronization quality between the digital twin model and the physical entity meets preset requirements based on the visual fidelity and the transmission performance value; if the synchronization quality does not meet the preset requirements, generate a synchronization strategy adjustment instruction and send the synchronization strategy adjustment instruction to a preset management terminal.

[0049] In this embodiment, within the intelligent management and operation system of a substation, the digital twin model, as a precise mapping of the physical entity in virtual space, is crucial in ensuring its synchronization quality with the physical entity. The data and intelligence layer undertakes the key task of guaranteeing this synchronization quality, ensuring that the digital twin model can accurately reflect the state of the physical entity in real time through a series of scientifically sound indicator evaluation and strategy adjustment mechanisms. Specifically, the data and intelligence layer assesses synchronization quality through the following steps: Visual fidelity is a key indicator for measuring how well a digital twin model visually matches a real physical scene. Structural similarity metrics assess the similarity between two images from three core aspects: brightness, contrast, and structure. Brightness considers whether the overall brightness of the images is similar; contrast focuses on whether the differences between bright and dark areas are consistent; and structure emphasizes the similarity of shapes, textures, and other structural information of objects in the images. By comprehensively evaluating these three aspects, structural similarity metrics can reflect the degree of similarity between images in a relatively comprehensive and detailed manner. Peak signal-to-noise ratio (PSNR) measures image quality by calculating the ratio of the mean squared error between the original image (real image) and the processed image (digital twin rendered image) to the maximum signal value of the original image. It primarily focuses on the noise level and signal fidelity of the image; a higher ratio indicates better image quality and a higher similarity to the real image. Combining these two metrics to compare the similarity between the rendered image of the digital twin model and the real image yields an accurate and comprehensive visual fidelity assessment result, which intuitively reflects the degree to which the digital twin model visually reproduces the physical entity.

[0050] Transmission performance plays a decisive role in the synchronization between the digital twin model and the physical entity, directly affecting whether data can be transmitted in a timely and stable manner. Latency measures the time required for data to be collected from the physical entity, transmitted to the digital twin model, and processed, encompassing the time consumed in multiple stages such as data acquisition, transmission link transmission, and model processing. If the latency is too long, the digital twin model cannot obtain the latest status information of the physical entity in a timely manner, leading to a discrepancy between its reflected status and the actual situation, affecting real-time monitoring and decision-making in the substation. Jitter reflects the fluctuation of latency during data transmission, i.e., the instability of latency. In actual transmission, latency may fluctuate significantly due to factors such as network congestion and signal interference. Excessive jitter means unstable data transmission latency, which may cause chaotic data updates in the digital twin model, such as data updating too quickly at some times and too slowly at others, severely affecting synchronization quality. By comprehensively considering latency and jitter metrics, the transmission performance value corresponding to the digital twin model is calculated. This value can comprehensively and accurately evaluate the efficiency and stability of data transmission, providing an important basis for judging the synchronization quality between the digital twin model and the physical entity.

[0051] Visual fidelity reflects the degree of similarity between the digital twin model and the real physical scene in terms of visual presentation, while transmission performance reflects the efficiency and stability of data transmission. These two indicators evaluate the synchronization quality between the digital twin model and the physical entity from different perspectives. The preset requirement can be a standard threshold set according to the actual operational needs and business scenarios of the substation. Only when both visual fidelity and transmission performance meet or exceed this preset requirement is the synchronization quality between the digital twin model and the physical entity considered satisfactory. For example, in some substation monitoring scenarios with extremely high real-time and accuracy requirements, it may be necessary to set a high visual fidelity standard to ensure that the digital twin model can clearly and accurately present the details of the physical entity; at the same time, a low latency and jitter standard may be set to ensure timely and stable data transmission, enabling the digital twin model to reflect the state changes of the physical entity in real time. Through this comprehensive judgment method, the synchronization status between the digital twin model and the physical entity can be comprehensively and accurately evaluated, providing a clear direction for subsequent adjustments and optimizations.

[0052] When it is determined that the synchronization quality between the digital twin model and the physical entity does not meet the preset requirements, it indicates a problem in visual presentation or data transmission, requiring targeted adjustments and optimizations. At this point, the data and intelligence layer can automatically generate synchronization strategy adjustment instructions based on the specific visual fidelity and transmission performance values. These instructions can include various adjustment measures; for example, if visual fidelity is low, the instructions can adjust rendering parameters, optimize image rendering algorithms, and improve the quality of rendered images. If transmission performance is unsatisfactory, the instructions can optimize data acquisition frequency, reduce unnecessary data acquisition, or improve data transmission protocols to enhance data transmission efficiency and stability. After generating the synchronization strategy adjustment instructions, they are sent to a preset management terminal, which can be an operator's terminal or an automated control device. Operators can promptly adjust the system based on the received instructions, or automated control devices can automatically execute corresponding operations, thereby quickly improving the synchronization quality between the digital twin model and the physical entity.

[0053] Furthermore, as Figure 1 The specific implementation of the system, as described in this application, provides a substation on-site safety control method, such as... Figure 2 As shown, the method includes: Step 101: The sensing layer collects multi-source heterogeneous data from the substation site in real time, including equipment status data, personnel location data, and external environment data.

[0054] Step 102: The data and intelligence layer performs fusion processing on the multi-source heterogeneous data, constructs a digital twin model that is synchronized in real time with the physical entity at the substation site based on the fusion processing result, and performs safety risk prediction on the digital twin model through an AI engine. The safety risk prediction includes fault prediction of equipment status and prediction of personnel behavior when entering dangerous areas.

[0055] Step 103: The application and control layer generates control commands based on the prediction results of the AI ​​engine, and sends the control commands to the corresponding control equipment at the substation site to perform risk alerts and / or risk control based on the control equipment.

[0056] Optionally, step 102, "fusing the multi-source heterogeneous data and constructing a digital twin model that is synchronized in real time with the physical entities at the substation site based on the fusion processing results," includes: For each data source, determine the communication protocol corresponding to the data source, and determine the data format conversion strategy based on the communication protocol. According to the data format conversion strategy, convert the real-time collected data corresponding to the data source into standard format data that can be recognized by the digital twin system. The standard format data is evaluated for quality based on a preset data quality assessment model. The standard format data that meets the quality requirements after evaluation is input into the digital twin system so that the digital twin model can be updated synchronously based on the standard format data.

[0057] Optionally, the step of "performing a quality assessment of the standard format data based on a preset data quality assessment model" includes: The total number of fields in the target data source is determined, and based on the standard format data of the target data source, the number of fields to be assigned values ​​is identified. A first ratio of the number of fields to be assigned values ​​to the total number of fields is calculated according to a field assignment quality assessment model. The field assignment quality of the standard format data is determined based on the relationship between the first ratio and a first preset ratio threshold, wherein the preset data quality assessment model includes the field assignment quality assessment model; and / or, Determine the total number of records in the target data source, and based on the standard format data of the target data source, identify the number of records with complete field assignments. Calculate a second ratio between the number of records with complete field assignments and the total number of records according to the assignment integrity quality assessment model. Determine the assignment integrity quality of the standard format data based on the relationship between the second ratio and a second preset ratio threshold, wherein the preset data quality assessment model includes the assignment integrity quality assessment model; and / or, The total amount of data from the target data source is determined, and based on the standard format data from the target data source, the number of data points that meet the data correctness requirements is identified. According to the data correctness quality assessment model, a third ratio of the number of data points meeting the data correctness requirements to the total amount of data is calculated. Based on the relationship between the third ratio and a third preset ratio threshold, the data correctness quality of the standard format data is determined. The preset data quality assessment model includes the data correctness quality assessment model itself; and / or, According to the data kurtosis quality assessment model, the kurtosis value corresponding to the standard format data of the target data source is calculated, and the data kurtosis quality of the standard format data is determined according to the relationship between the kurtosis value and the preset kurtosis threshold. The preset data quality assessment model includes the data kurtosis quality assessment model.

[0058] Optionally, the "predicting security risks on the digital twin model using an AI engine" mentioned in step 102 includes: The device status data time series of the target device is obtained from the digital twin model, and the device status data time series is input into the AI ​​engine to obtain the future device status and potential failure risks of the target device; The personnel location data of the maintenance personnel are obtained from the digital twin model. The real-time location and movement trajectory of the maintenance personnel are determined based on the personnel location data. Based on the real-time location and movement trajectory, the predicted trajectory of the maintenance personnel is determined by the AI ​​engine. The predicted trajectory is then compared with the dangerous area vector library defined in the digital twin model to obtain the behavioral risk of the maintenance personnel entering the dangerous area within a preset time period in the future. Based on external environmental data mapped to the spatial layout of the digital twin model, an environmental data time series is constructed, and the environmental data time series is input into the AI ​​engine to obtain extreme environment prediction results. Based on the extreme environment prediction results, the potential failure risks of specific equipment in the substation are analyzed.

[0059] Optionally, the method further includes: The data and intelligence layer extracts standard format data sets from at least two data sources from the digital twin model, calculates the Pearson correlation coefficient between each pair of standard format data sets, and determines the correlation between the two standard format data sets based on the Pearson correlation coefficient. The data sources include at least two of the following: equipment status data, personnel behavior data, and external environment data. When the AI ​​engine's prediction indicates a security risk, the data source on which the security risk prediction is based is determined, and an associated data source with a correlation greater than a preset correlation threshold is identified. A linkage control instruction is generated based on the associated data source, wherein the linkage control instruction is used to provide risk warnings and / or risk control to the control device corresponding to the associated data source.

[0060] Optionally, step 102, "constructing a digital twin model that is synchronized in real time with the physical entities at the substation site based on the fusion processing results," includes: The original point cloud data of the substation is acquired by 3D laser scanning and SLAM mobile measurement technology, and the original point cloud data is preprocessed to obtain a preprocessed point cloud. The preprocessing includes at least one of noise removal, outlier removal and downsampling. By using the iterative nearest point algorithm, the preprocessed point cloud from different perspectives is registered. After the point cloud registration is completed, the root mean square error of the point pair corresponding to the registration result is calculated. Based on the root mean square error of the point pair, the quantitative evaluation result of the registration result is obtained. Once the quantitative evaluation results meet the preset requirements, an initial digital twin model in a unified coordinate system is constructed based on the registration results, and the fusion processing results are synchronized to the initial digital twin model in real time to obtain a digital twin model that is synchronized in real time with the physical entity of the substation site.

[0061] Optionally, the method further includes: The data and intelligence layer compares the similarity between the rendered image of the digital twin model and the real image based on structural similarity index and peak signal-to-noise ratio index to obtain the visual fidelity of the digital twin model. Based on latency and jitter metrics, the transmission performance value corresponding to the digital twin model is calculated. Based on the visual fidelity and the transmission performance value, determine whether the synchronization quality between the digital twin model and the physical entity meets the preset requirements; If the synchronization quality does not meet the preset requirements, a synchronization strategy adjustment instruction is generated and sent to the preset management terminal.

[0062] It should be noted that other corresponding descriptions of the functional units involved in the substation field safety control method provided in this application embodiment can be found in the following references. Figure 1 The corresponding descriptions in the system will not be repeated here.

[0063] This application also provides a computer device, which may specifically be a personal computer, a server, a network device, etc. Figure 3 As shown, the computer device includes a bus, a processor, memory, and a communication interface, and may also include an input / output interface and a display device. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores location information. The network interface allows communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the various method embodiments.

[0064] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0065] In one embodiment, a computer-readable storage medium is provided, which may be non-volatile or volatile, having stored thereon a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0066] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0067] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0068] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0069] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0070] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A substation field safety control system, characterized by, The application relates to a power substation safety risk prediction method and system. The method comprises the following steps: a perception layer is used to collect multi-source heterogeneous data in real time, wherein the multi-source heterogeneous data comprises device state data, personnel position data and external environment data; a data and intelligence layer is used to fuse the multi-source heterogeneous data, construct a digital twin model which is real-time synchronized with physical entities in the power substation based on the fusion result, and perform safety risk prediction on the digital twin model through an AI engine, wherein the safety risk prediction comprises fault prediction of device state and behavior prediction of personnel entering a dangerous area; 2. The system of claim 1, wherein, an application and control layer is used to generate control instructions according to the prediction result of the AI engine, and the control instructions are sent to corresponding control devices in the power substation to perform risk reminding and / or risk control based on the control devices. The data and intelligence layer is used to: for each data source, determine the communication protocol corresponding to the data source, determine a data format conversion strategy based on the communication protocol, and convert real-time collection data corresponding to the data source into standard format data recognizable by a digital twin system according to the data format conversion strategy; perform quality evaluation on the standard format data based on a preset data quality evaluation model; 3. The system of claim 2, wherein, input the standard format data meeting the quality requirements after the evaluation into the digital twin system to update the digital twin model based on the standard format data. The data and intelligence layer performs quality evaluation on the standard format data of a target data source in the following manner: determine the total number of fields of the target data source, identify the number of fields with values based on the standard format data of the target data source, calculate a first ratio of the number of fields with values to the total number of fields according to a field value assignment quality evaluation model, and determine the field value assignment quality of the standard format data according to the relationship between the first ratio and a first preset ratio threshold value, wherein the preset data quality evaluation model comprises the field value assignment quality evaluation model; and / or determine the total number of records of the target data source, identify the number of records with complete field value assignment based on the standard format data of the target data source, calculate a second ratio of the number of records with complete field value assignment to the total number of records according to a value completeness quality evaluation model, and determine the value completeness quality of the standard format data according to the relationship between the second ratio and a second preset ratio threshold value, wherein the preset data quality evaluation model comprises the value completeness quality evaluation model; and / or determine the total amount of data of the target data source, identify the number of data meeting data correctness requirements based on the standard format data of the target data source, calculate a third ratio of the number of data meeting data correctness requirements to the total amount of data according to a data correctness quality evaluation model, and determine the data correctness quality of the standard format data according to the relationship between the third ratio and a third preset ratio threshold value, wherein the preset data quality evaluation model comprises the data correctness quality evaluation model; and / or According to the data kurtosis quality evaluation model, the kurtosis value corresponding to the standard format data of the target data source is calculated, and the data kurtosis quality of the standard format data is determined according to the relationship between the kurtosis value and a preset kurtosis threshold. The preset data quality evaluation model includes the data kurtosis quality evaluation model.

4. The system of claim 1, wherein, The data and intelligent layer is further configured to: obtain device state data time series of the target device from the digital twin model, and input the device state data time series into the AI engine to obtain future device states and potential failure risks of the target device; obtain personnel location data of the maintenance personnel from the digital twin model, determine real-time positions and motion trajectories of the maintenance personnel according to the personnel location data, determine a predicted trajectory of the maintenance personnel based on the real-time positions and motion trajectories through the AI engine, perform collision detection between the predicted trajectory and a defined dangerous area vector library in the digital twin model to obtain a behavior risk of the maintenance personnel entering a dangerous area within a future preset time period; construct an environment data time series based on external environment data mapped into a space layout of the digital twin model, and input the environment data time series into the AI engine to obtain an extreme environment prediction result, and analyze potential failure risks of specific devices in the substation according to the extreme environment prediction result.

5. The system of claim 1, wherein, The data and intelligent layer is further configured to: extract a set of standard format data of at least two data sources from the digital twin model, calculate a Pearson correlation coefficient between each two sets of standard format data, and determine the correlation between the two sets of standard format data according to the Pearson correlation coefficient, wherein the data sources include at least two of device state data, personnel behavior data, and external environment data; when the prediction result of the AI engine indicates that there is a safety risk, determine the data source on which the safety risk prediction is based, and determine an associated data source having a correlation greater than a preset correlation threshold with the data source, and generate a linkage control instruction according to the associated data source, wherein the linkage control instruction is used for risk reminding and / or risk control on a control device corresponding to the associated data source.

6. The system of claim 1, wherein, The data and intelligent layer is further configured to: obtain original point cloud data of the substation collected through three-dimensional laser scanning and SLAM mobile measurement technology, and pre-process the original point cloud data to obtain pre-processed point cloud, wherein the pre-processing includes at least one of noise removal, outlier rejection, and downsampling; register the pre-processed point cloud from different perspectives through an iterative closest point algorithm, and calculate a point pair root mean square error corresponding to a registration result after completing point cloud registration, and obtain a quantitative evaluation result of the registration result according to the point pair root mean square error; when the quantitative evaluation result meets a preset requirement, construct an initial digital twin model under a unified coordinate system according to the registration result, and synchronize the fusion processing result to the initial digital twin model in real time to obtain a digital twin model that is real-time synchronized with physical entities in the substation site.

7. The system of claim 6, wherein, The data and intelligence layer is also used for: Based on the structural similarity index and the peak signal-to-noise ratio index, the rendering image of the digital twin model is compared with the real image in similarity to obtain the visual fidelity corresponding to the digital twin model; Based on the time delay index and the jitter amount index, the transmission performance value corresponding to the digital twin model is calculated; According to the visual fidelity and the transmission performance value, it is judged whether the synchronization quality of the digital twin model and the physical entity meets the preset requirement; If the synchronization quality does not meet the preset requirement, a synchronization strategy adjustment instruction is generated, and the synchronization strategy adjustment instruction is sent to a preset management terminal.

8. A substation field safety control method characterized by, Comprise: The perception layer collects multi-source heterogeneous data of the substation site in real time, wherein the multi-source heterogeneous data includes device state data, personnel position data and external environment data; The data and intelligence layer fuses the multi-source heterogeneous data, constructs a digital twin model which is real-time synchronized with the physical entity of the substation site based on the fusion processing result, and performs safety risk prediction on the digital twin model through an AI engine, wherein the safety risk prediction includes fault prediction of device state and behavior prediction of personnel entering dangerous area; The application and control layer generates a control instruction according to the prediction result of the AI engine, and issues the control instruction to the corresponding control device of the substation site, so as to perform risk warning and / or risk control based on the control device.

9. A storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to realize the method of claim 8.

10. A computer device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, The processor executes the computer program to realize the method of claim 8.

Images