Power consumer electricity utilization potential safety hazard identification, early warning and disposal method and system

Abstract

The invention discloses a method and a system for identifying, early warning and disposing potential safety hazards of power utilization of power consumers. The method comprises the following steps: acquiring multi-source power data; according to the multi-source electric power data, based on a hidden danger recognition model, spatio-temporal data association and a device-line-user three-level positioning mechanism, performing recognition, early warning, anomaly backtracking and positioning on the hidden danger of electricity utilization of the user; and inputting potential safety hazard identification early warning and positioning related information into the power utilization safety risk disposal large model to generate a disposal strategy. According to the invention, through adoption of a method of processing data through multi-source data fusion and a three-level positioning mechanism, the data treatment capability and the hidden danger disposal capability are enhanced, the risk identification and classification accuracy is improved, and the early warning response time is shortened.

Description

Technical Field

[0001] The present invention relates to the field of electricity safety hazard supervision, and in particular to a method and system for identifying, warning and handling electricity safety hazard of power users. Background Art

[0002] The method and system for identifying and warning safety hazards in electricity consumption by power users refers to the realization of full-process management from bottom-level data collection, cloud-based data calculation and analysis, to real-time monitoring and display on the client side, which improves the ability to identify and handle hidden dangers and risks, is conducive to building a complete safety risk identification and warning capability for power users, and solves the problem of lack of safety risk identification and warning software modules for power users.

[0003] Currently, research on methods and systems for identifying and warning of power safety hazards for power users primarily focuses on the following aspects: 1) Configuration and optimization of electrical safety devices. Properly configured electrical safety devices are crucial for identifying and warning of high-risk circuit safety hazards. These devices typically utilize sensing conductors to detect cable length or temperature changes and mechanical damage on the surfaces of electrical component casings and internal components. 2) Application of fault warning and monitoring systems. By collecting and analyzing large amounts of monitoring data, predictive models are constructed using data mining and machine learning techniques to identify hidden abnormal behaviors or trends. 3) Risk assessment methods based on the entire asset lifecycle. These methods quantitatively assess the likelihood of risk occurrence and the degree of risk impact, comprehensively assessing risk values ​​and levels. The degree of risk impact is determined by analyzing factors such as grid safety, casualties, social image, and direct economic losses, and the risk level is then classified based on the likelihood of risk occurrence. 4) Integration of intelligent and automated technologies. Using devices such as smart sensors, smart meters, and remote terminal units, real-time monitoring and data analysis of various power system parameters are possible. Automated control systems and intelligent dispatching algorithms enable optimized power system scheduling and rapid fault response. This will improve the safety and reliability of the power system and reduce the risks and losses of safety accidents.

[0004] Currently, safety risk assessments for power users often consider single structural factors, lacking the integrated processing and visualization of heterogeneous information such as power data and equipment status information. Furthermore, they primarily rely on qualitative assessments, lacking quantitative analysis for safety risk classification and identification. Currently, safety risk identification focuses primarily on identifying and evaluating features at or after a risk occurs. Technologies for quantifying, identifying, and providing early warnings for early features are insufficiently comprehensive, preventing early identification and grading of safety risks and enabling users to adjust their electricity usage measures in a timely manner. Furthermore, the functionality of the safety risk assessment software currently used by power users is relatively limited, primarily focusing on equipment monitoring and display on the power supply side. There is a lack of safety risk identification and warning software and system platforms that consider multi-source information fusion and multi-modal data perception. Therefore, the need for governments, power grid companies, and power users themselves to identify and warn of electricity safety risks is becoming increasingly urgent. Advanced technologies are needed to improve electricity safety management and ensure the safe and reliable operation of the power supply and utilization system. Summary of the Invention

[0005] In order to solve the problems of low identification accuracy, unscientific risk assessment mechanism, low disposal efficiency and poor system integration in the existing technology of power user safety risk identification, early warning and disposal, the present invention proposes a method and system for identifying, early warning and handling safety hazards of power users.

[0006] The present invention provides a method for identifying, warning and handling potential safety hazards of power users, including:

[0007] Acquire multi-source power data;

[0008] Based on multi-source power data, the system can identify, warn, trace back, and locate power safety hazards to users based on hidden danger identification models, spatiotemporal data association, and a three-level positioning mechanism of equipment, lines, and users.

[0009] Input the relevant information on potential safety hazards identification, warning and location into the large-scale model for power safety risk management to generate a management strategy;

[0010] The hidden danger identification model is trained with multi-source power data as input and safety hidden danger identification and cross-layer early warning as output;

[0011] Among them, the large model for handling electricity safety risks is trained with safety hazard identification, warning and positioning information as input and handling strategies as output.

[0012] Preferably, the acquiring of multi-source data includes:

[0013] Collect data from multiple sources;

[0014] A two-level data governance architecture, edge-side and master-station, is used to standardize the data framework and integrate multi-source power data.

[0015] Perform data quality verification and data quality assessment on multi-source power data;

[0016] The data quality check includes at least one or more of the following: format check, consistency check, integrity check or logic check.

[0017] Preferably, the method of identifying, warning, tracing back and locating user power safety hazards based on multi-source power data, a hidden danger identification model, spatiotemporal data association and a three-level positioning mechanism of equipment-line-user includes:

[0018] Based on multi-source power data, the system can identify and issue early warnings for users’ power safety hazards based on hidden danger identification models.

[0019] Conduct fault analysis based on multi-source power data and spatiotemporal data correlation to identify affected areas and perform anomaly retrospective analysis to analyze the cause of the power outage;

[0020] Based on multi-source power data and identified hidden dangers, a three-level positioning mechanism of equipment-line-user is adopted to locate users' power safety hazards.

[0021] Preferably, the method of identifying and warning users' power safety hazards based on multi-source power data and a hidden danger identification model includes:

[0022] Performing layered extraction of hidden danger features on multi-source power data to obtain layered multi-source power data;

[0023] Inputting layered multi-source power data into the hidden danger identification model to obtain cross-layer safety hidden danger identification and early warning results;

[0024] The level of the hidden danger feature includes at least one or more of the following: a device-level feature level, a circuit-level feature level, or a system-level feature level;

[0025] The hidden danger identification model is trained by taking layered multi-source power data as input and cross-layer safety hidden danger identification and early warning results as output.

[0026] Preferably, the step of inputting the hierarchical multi-source power data into the hidden danger identification model to obtain cross-layer safety hidden danger identification and early warning results includes:

[0027] Granger causality test or knowledge graph matching is used to mine the equipment-loop causal chain of hierarchical multi-source power data to obtain the equipment-loop causal chain;

[0028] Using causal relationship discovery methods and hybrid reasoning mechanisms, we can find other related causal chains at each node based on the existing device-loop causal chain to build a dynamic causal graph.

[0029] Make dynamic risk assessment based on dynamic cause-effect diagram;

[0030] Conducting risk conduction simulation based on Bayesian network according to hierarchical multi-source power data to obtain conduction path prediction results;

[0031] The results of conduction path prediction and dynamic risk assessment are used as the results of cross-layer safety hazard identification and early warning.

[0032] Preferably, the step of mining the device-loop causal chain using Granger causality test or knowledge graph matching on the hierarchical multi-source power data to obtain the device-loop causal chain includes:

[0033] The Granger causality test method is used to determine the causal relationship between hierarchical multi-source power data and obtain the equipment-loop causal chain;

[0034] Or input hierarchical multi-source power data, use the knowledge graph matching method, and search historical similar cases through the semantic big model to obtain the equipment loop causal chain.

[0035] Preferably, the method of using the causal relationship discovery method and the hybrid reasoning mechanism to construct a dynamic causal graph by finding other causal chains related to each node of the existing device-loop causal chain includes:

[0036] Based on the existing equipment-loop causal chain related node data, transfer entropy is used to quantify the causal strength between equipment parameters and loop risks, and a priori causal rule base is constructed based on the knowledge graph to discover the causal relationship of related node data;

[0037] The Granger causality test method is used to analyze the correlation between equipment, circuits, and systems in historical power outages. Based on the existing equipment-circuit causal chain node data and the equipment-circuit-system correlation, a semantic big model is used to analyze inspection reports, user repair records, and economic load demand to extract causal triples between related node data.

[0038] Other causal chains are formed based on the causal relationships and causal triples of the relevant node data, and the causal chains are interwoven to form a dynamic causal graph.

[0039] Preferably, making a dynamic risk assessment based on a dynamic cause-effect diagram includes:

[0040] Based on the hierarchical multi-source power data related to safety hazards in the dynamic causal diagram, the subjective weight of each preset risk indicator is calculated using the hierarchical analysis method, and the objective weight of each risk indicator is calculated using the entropy weight method;

[0041] Based on the hierarchical multi-source power data related to safety hazards in the dynamic causal diagram, the value of each risk indicator is calculated using the calculation formula of each preset risk indicator;

[0042] Based on the value of each risk indicator and the subjective weight and objective weight of each risk indicator, a weighted calculation method is used to calculate the dynamic risk assessment value;

[0043] Conduct dynamic risk assessment of potential safety hazards based on preset risk levels and dynamic risk evaluation.

[0044] Preferably, the risk conduction simulation based on the Bayesian network according to the hierarchical multi-source power data is performed to obtain the conduction path prediction result, including:

[0045] Simulating risk transmission paths based on Bayesian networks according to hierarchical multi-source power data or searching for risk transmission paths based on transmission paths of power grid topology according to hierarchical multi-source power data;

[0046] The path risk of the risk transmission path is calculated according to the weighting method to obtain the transmission path prediction result.

[0047] Preferably, the fault analysis based on the correlation of multi-source power data and spatiotemporal data to identify the affected area and perform anomaly retrospective analysis to determine the cause of the power outage includes:

[0048] Integrate and connect multi-source power data;

[0049] Based on the integrated and connected multi-source power data, spatiotemporal correlation matching and model intelligent algorithm analysis are performed to find the affected areas and trace back the causes of the power outage.

[0050] Preferably, the method of performing spatiotemporal correlation matching and model intelligent algorithm analysis based on the integrated and connected multi-source power data to identify the affected area and trace back the cause of the power outage includes:

[0051] Based on the integrated and connected multi-source power data, the outage location is mapped to the specific coordinates of the power grid map, and the outage time is recorded according to the timeline;

[0052] Conduct fault analysis along the grid topology based on the specific coordinates and duration of the outage to identify the affected area;

[0053] Based on the specific coordinates and time of the power outage, we can trace back to the source and find the substation that had the power outage first, and then infer the cause of the power outage.

[0054] Preferably, locating potential safety hazards of user electricity use based on multi-source power data and a three-level positioning mechanism of equipment, lines, and users includes:

[0055] Directly locate faulty equipment based on multi-source power data and identified hidden dangers, and perform device-level location of user power safety hazards;

[0056] Based on multi-source power data and identified hidden dangers, topology analysis is performed to obtain associated lines related to the fault, and line-level positioning of user power safety hazards is performed;

[0057] Based on multi-source power data and identified hidden dangers, compared with marketing data, a list of affected users is drawn up, and user-level positioning of power safety hazards is performed.

[0058] Preferably, the multi-source power data includes at least one or more of the following: user profile data, electricity usage behavior data, high-risk circuit data, equipment operation data, PMS system data, 95598 planned power outage data, scheduling support system data, power outage handling data or safety and compliance data.

[0059] Based on the same inventive concept, the present invention also provides a system for identifying, warning and handling hidden dangers in electricity safety of power users, comprising: a module for acquiring multi-source power data, an intelligent analysis module and a handling module;

[0060] The multi-source power data acquisition module is used to acquire multi-source power data;

[0061] The intelligent analysis module is used to identify, warn, trace back, and locate user power safety hazards based on multi-source power data, a hidden danger identification model, spatiotemporal data association, and a three-level positioning mechanism of equipment, lines, and users;

[0062] The treatment module is used to input information related to safety hazard identification, warning and location into the power safety risk treatment model to generate a treatment strategy;

[0063] The hidden danger identification model is trained with multi-source power data as input and safety hidden danger identification and cross-layer early warning as output;

[0064] Among them, the large model for handling electricity safety risks is trained with safety hazard identification, warning and positioning information as input and handling strategies as output.

[0065] Preferably, the module for acquiring multi-source power data is specifically used to:

[0066] Collect data from multiple sources;

[0067] A two-level data governance architecture, edge-side and master-station, is used to standardize the data framework and integrate multi-source power data.

[0068] Perform data quality verification and data quality assessment on multi-source power data;

[0069] The data quality check includes at least one or more of the following: format check, consistency check, integrity check or logic check.

[0070] Preferably, the intelligent analysis module is specifically used to:

[0071] Based on multi-source power data, the system can identify and issue early warnings for users’ power safety hazards based on hidden danger identification models.

[0072] Conduct fault analysis based on multi-source power data and spatiotemporal data correlation to identify affected areas and perform anomaly retrospective analysis to analyze the cause of the power outage;

[0073] Based on multi-source power data and identified hidden dangers, a three-level positioning mechanism of equipment-line-user is adopted to locate users' power safety hazards.

[0074] Preferably, the intelligent analysis module identifies and warns users of potential power safety hazards based on multi-source power data and a potential safety hazard identification model, including:

[0075] Performing layered extraction of hidden danger features on multi-source power data to obtain layered multi-source power data;

[0076] Inputting layered multi-source power data into the hidden danger identification model to obtain cross-layer safety hidden danger identification and early warning results;

[0077] The level of the hidden danger feature includes at least one or more of the following: a device-level feature level, a circuit-level feature level, or a system-level feature level;

[0078] The hidden danger identification model is trained by taking layered multi-source power data as input and cross-layer safety hidden danger identification and early warning results as output.

[0079] Preferably, the intelligent analysis module inputs the hierarchical multi-source power data into the hidden danger identification model to obtain cross-layer safety hidden danger identification and early warning results, including:

[0080] Granger causality test or knowledge graph matching is used to mine the equipment-loop causal chain of hierarchical multi-source power data to obtain the equipment-loop causal chain;

[0081] Using causal relationship discovery methods and hybrid reasoning mechanisms, we can find other related causal chains at each node based on the existing device-loop causal chain to build a dynamic causal graph.

[0082] Make dynamic risk assessment based on dynamic cause-effect diagram;

[0083] Conducting risk conduction simulation based on Bayesian network according to hierarchical multi-source power data to obtain conduction path prediction results;

[0084] The results of conduction path prediction and dynamic risk assessment are used as the results of cross-layer safety hazard identification and early warning.

[0085] Preferably, the intelligent analysis module mines the device-loop causal chain using Granger causality test or knowledge graph matching on the hierarchical multi-source power data to obtain the device-loop causal chain, including:

[0086] The Granger causality test method is used to determine the causal relationship between hierarchical multi-source power data and obtain the equipment-loop causal chain;

[0087] Or input hierarchical multi-source power data, use the knowledge graph matching method, and search historical similar cases through the semantic big model to obtain the equipment loop causal chain.

[0088] Preferably, the intelligent analysis module uses a causal relationship discovery method and a hybrid reasoning mechanism to construct a dynamic causal graph based on finding other causal chains related to each node of the existing device-loop causal chain, including:

[0089] Based on the existing equipment-loop causal chain related node data, transfer entropy is used to quantify the causal strength between equipment parameters and loop risks, and a priori causal rule base is constructed based on the knowledge graph to discover the causal relationship of related node data;

[0090] The Granger causality test method is used to analyze the correlation between equipment, circuits, and systems in historical power outages. Based on the existing equipment-circuit causal chain node data and the equipment-circuit-system correlation, a semantic big model is used to analyze inspection reports, user repair records, and economic load demand to extract causal triples between related node data.

[0091] Other causal chains are formed based on the causal relationships and causal triples of the relevant node data, and the causal chains are interwoven to form a dynamic causal graph.

[0092] Preferably, the intelligent analysis module makes a dynamic risk assessment based on the dynamic cause-effect diagram, including:

[0093] Based on the hierarchical multi-source power data related to safety hazards in the dynamic causal diagram, the subjective weight of each preset risk indicator is calculated using the hierarchical analysis method, and the objective weight of each risk indicator is calculated using the entropy weight method;

[0094] Based on the hierarchical multi-source power data related to safety hazards in the dynamic causal diagram, the value of each risk indicator is calculated using the calculation formula of each preset risk indicator;

[0095] Based on the value of each risk indicator and the subjective weight and objective weight of each risk indicator, a weighted calculation method is used to calculate the dynamic risk assessment value;

[0096] Conduct dynamic risk assessment of potential safety hazards based on preset risk levels and dynamic risk evaluation.

[0097] Preferably, the intelligent analysis module performs conduction path simulation based on risk conduction simulation of Bayesian network according to hierarchical multi-source power data to obtain conduction path prediction results, including:

[0098] Simulating risk transmission paths based on Bayesian networks according to hierarchical multi-source power data or searching for risk transmission paths based on transmission paths of power grid topology according to hierarchical multi-source power data;

[0099] The path risk of the risk transmission path is calculated according to the weighting method to obtain the transmission path prediction result.

[0100] Preferably, the intelligent analysis module performs fault analysis based on the correlation of multi-source power data and spatiotemporal data to identify the affected area and conducts anomaly retrospective analysis to determine the cause of the power outage, including:

[0101] Integrate and connect multi-source power data;

[0102] Based on the integrated and connected multi-source power data, spatiotemporal correlation matching and model intelligent algorithm analysis are performed to find the affected areas and trace back the causes of the power outage.

[0103] Preferably, the intelligent analysis module performs spatiotemporal correlation matching and model intelligent algorithm analysis based on the integrated and connected multi-source power data to find the affected area and trace back the cause of the power outage, including:

[0104] Based on the integrated and connected multi-source power data, the outage location is mapped to the specific coordinates of the power grid map, and the outage time is recorded according to the timeline;

[0105] Conduct fault analysis along the grid topology based on the specific coordinates and duration of the outage to identify the affected area;

[0106] Based on the specific coordinates and time of the power outage, we can trace back to the source and find the substation that had the power outage first, and then infer the cause of the power outage.

[0107] Preferably, the intelligent analysis module locates potential safety hazards of user electricity use based on multi-source power data and a three-level positioning mechanism of equipment, lines, and users, including:

[0108] Directly locate faulty equipment based on multi-source power data and identified hidden dangers, and perform device-level location of user power safety hazards;

[0109] Based on multi-source power data and identified hidden dangers, topology analysis is performed to obtain associated lines related to the fault, and line-level positioning of user power safety hazards is performed;

[0110] Based on multi-source power data and identified hidden dangers, compared with marketing data, a list of affected users is drawn up, and user-level positioning of power safety hazards is performed.

[0111] Preferably, the multi-source power data includes at least one or more of the following: user profile data, electricity usage behavior data, high-risk circuit data, equipment operation data, PMS system data, 95598 planned power outage data, scheduling support system data, power outage handling data or safety and compliance data.

[0112] Compared with the prior art, the present invention has the following beneficial effects:

[0113] The present invention provides a method and system for identifying, warning, and addressing power safety hazards for power users. The method comprises: acquiring multi-source power data; identifying, warning, tracing back anomalies, and locating power safety hazards for users based on a hazard identification model, spatiotemporal data correlation, and a three-level positioning mechanism for equipment, lines, and users based on the multi-source power data; and inputting information related to hazard identification, warning, and location into a large-scale power safety risk management model to generate a management strategy. By employing multi-source data fusion processing and a three-level positioning mechanism, the present invention enhances data governance and hazard management capabilities, improves the accuracy of risk identification and classification, and shortens warning response time. BRIEF DESCRIPTION OF THE DRAWINGS

[0114] Figure 1 This is a flow chart of the method for identifying, warning and handling potential safety hazards of electricity users of the present invention;

[0115] Figure 2 A schematic diagram of the system architecture of the method for identifying, warning and handling potential safety hazards of power users of the present invention;

[0116] Figure 3 Schematic diagram of the system for identifying, warning and handling potential safety hazards of power users according to the present invention. DETAILED DESCRIPTION

[0117] The following is a further detailed description of the specific embodiments of the present invention with reference to the accompanying drawings.

[0118] The present invention provides a method for identifying, warning and handling potential safety hazards of power users. The flow chart is as follows: Figure 1 As shown:

[0119] S1. Acquire multi-source power data;

[0120] S2. Based on multi-source power data, the system identifies, issues warnings, and performs anomaly tracing and location tracking for power safety hazards to users, using a hidden danger identification model, spatiotemporal data association, and a three-level positioning mechanism for equipment, lines, and users.

[0121] S3. Input the information related to safety hazard identification, warning and location into the large-scale model for power safety risk management to generate a management strategy;

[0122] The hidden danger identification model is trained with multi-source power data as input and safety hidden danger identification and cross-layer early warning as output;

[0123] Among them, the large model for handling electricity safety risks is trained with safety hazard identification, warning and positioning information as input and handling strategies as output.

[0124] Step S1 specifically includes:

[0125] Comprehensively monitor electrical, dynamic, and load data such as power user files, power supply circuits, and operating status of energy-consuming equipment, evaluate and manage user monitoring data, and improve the accuracy and breadth of data monitoring.

[0126] 1. Multi-source data monitoring

[0127] 1. Multi-source data collection

[0128] (1) User profile data: including user account number, name, address, contact person, importance level, voltage level, operating capacity, etc.;

[0129] a) User identity information: including company name, unified social credit code, user electricity account number, etc.;

[0130] b) Power supply contract information: electricity use address, power supply scope, power supply capacity, operating capacity, self-provided emergency power supply capacity, voltage level, power supply capacity, nature of electricity use, etc.;

[0131] c) Equipment records: models, commissioning time, and maintenance records of high-voltage equipment such as transformers, circuit breakers, and relay protection devices;

[0132] (2) Electricity usage data:

[0133] a) Load characteristics: hourly / minute-level power consumption curve, maximum demand, power factor;

[0134] b) Abnormal power consumption records: event logs such as overload, voltage sag, and harmonic exceeding the standard; voltage sag waveforms, zero-sequence current mutations, and circuit breaker trip signals;

[0135] (3) High-risk circuit data: including the operating power of the main electrical equipment in the production load circuit and the security load circuit, load data, power quality data, accident power outage data, equipment configuration data and action record data, power environment data, etc.; the fire protection circuit is mainly composed of fire detectors, sprinkler systems and fire extinguishers; control system circuits such as PLC and DCS systems, if a failure occurs, it may cause the production process to get out of control.

[0136] (4) Equipment operation data:

[0137] a) Electrical parameters: real-time monitoring data such as three-phase voltage / current, temperature, insulation resistance, busbar node circuit breaker on / off, and temperature rise in the high-voltage system cabinet.

[0138] b) Environmental data: temperature and humidity collection, water immersion collection, video monitoring, smoke alarm, and firefighting facility status;

[0139] (5) PMS system data: transformer / circuit breaker service life, maintenance record data, equipment ID and grid GIS coordinate mapping data, etc.;

[0140] (6) 95598 planned power outage data: power outage scope, number of affected users, power outage application number;

[0141] (7) Dispatching support system data: grid topology, feeder connection relationship, protection device action sequence, SCADA data, load curve 15 minutes before the fault, relay protection settings, etc.;

[0142] (8) Power outage handling data: fault isolation time, emergency repair team response time, user power restoration confirmation record, etc.;

[0143] (9) Security and compliance data:

[0144] a) Safety protection records: equipment preventive test reports, inspection records, and fault handling files

[0145] b) Compliance documents: power business license, energy efficiency test report, and emergency plan filing;

[0146] 2. Unified integration of multi-source data

[0147] A two-level data governance architecture of "edge side-main site" is adopted to ensure the compatibility of multi-source heterogeneous data.

[0148] Data frame standardization

[0149] (1) Data unification: Unify the metadata field mapping table and device identity, adopt a three-level coding system of user-device-loop for devices, and apply dimensional conversion rules to data processing;

[0150] (2) Timing alignment mechanism: The edge side uses PTP (Precision Time Protocol) to synchronize sensor clocks, and the master side interpolates non-real-time data (such as daily inspection records) to generate 1-minute granularity time series data;

[0151] (3) Data format specification: standardize the transmission protocol, adopt MQTT (real-time data) + Apache Avro (batch data), and standardize the data storage format;

[0152] Multi-source data fusion

[0153] (1) Horizontal fusion (multi-source data association)

[0154] a) Rule mapping: Align the data of the scheduling system, electricity collection system, and PMS system through user ID, device code, and timestamp

[0155] b) Topology overlay: Map the power outage event to the grid topology and associate the loop risk level.

[0156] (2) Vertical fusion (time-space dimension):

[0157] a) Rule mapping: Align data from the scheduling system, electricity consumption collection system, and PMS system through user ID, device code, and timestamp.

[0158] b) Topology overlay: Mapping power outage events to the grid topology and correlating circuit risk levels;

[0159] 2. User Data Governance

[0160] 1. Data quality verification rules:

[0161] (1) Format check:

[0162] a) Unified coding rules: Equipment records must comply with the DL / T 700 series power equipment coding standards

[0163] b) Document specification: The contract text must be in OFD format and embedded with a digital signature;

[0164] c) Naming standards: There must be a logical correlation between the user's industry type and the user's username. For example, if the industry type is hospital and the user name is XX school, the naming is considered incorrect.

[0165] (2)Consistency check:

[0166] a) Cross-system comparison: User profile information must be consistent with the power grid marketing system and dispatching automation system data

[0167] b) Timing alignment: The power consumption curve data and the SCADA system acquisition frequency must match (error < 0.5%);

[0168] (3) Integrity check:

[0169] a) Required field check: The missing rate of fields such as safety tool inspection records and emergency plans in high-voltage user files must be less than 1%

[0170] b) Data relevance verification: Equipment failure records must be associated with corresponding maintenance work orders and processing results;

[0171] (4)Logical verification:

[0172] a) Capacity matching: The rated capacity of the transformer and the maximum demand declared by the user must meet the 1.2 times margin requirement

[0173] b) Time series compliance: The preventive test cycle shall not exceed the interval period specified in the power industry standard DL / T 596;

[0174] 2. Data quality assessment method:

[0175] The data quality scoring model is constructed based on industry characteristics and grid security requirements, achieving full-chain governance through multi-source data fusion analysis and dynamic closed-loop control. Based on game theory empowerment, it combines the complex correlations between equipment parameters, energy usage characteristics, and grid topology among different types of power users to form an integrated "assessment-monitoring-optimization" architecture.

[0176] First, a multi-dimensional indicator quantification system is used to deeply analyze user profiles. For core fields such as transformer capacity, industry classification, and reactive power configuration, industry-specific calculation rules are defined for completeness (such as the missing rate of required fields), accuracy (deviation between equipment parameters and SCADA measured values), and consistency (cross-system ledger matching). A Shapley value dynamic weighting algorithm is introduced to balance the weighting demands of the power grid digitalization department, marketing department, and equipment suppliers, ensuring that quality assessment results are consistent with expert experience and actual risks. For example, if the deviation between the rated capacity of an arc furnace and the real-time load in a user profile exceeds 20%, the system automatically increases the global weight of the accuracy indicator, triggering a special verification of high-energy-consuming enterprises.

[0177] Secondly, an edge-cloud collaborative architecture is adopted to enhance real-time performance: the industrial gateway on the plant side has a built-in lightweight rule engine, which performs millisecond-level verification of key parameters such as the current transformer ratio and load rate, and synchronizes abnormal data to the cloud through the 5G slicing network; the platform layer relies on the knowledge graph to build a device topology relationship network, and uses the GraphSAGE algorithm to explore hidden defects (such as power ownership conflicts caused by metering point connection errors), and combines with the industry benchmark library to intelligently fill in missing fields.

[0178] Then, an industry benchmarking analysis mechanism is designed to dynamically compare the power factor, harmonic content and other indicators of individual users with the average level of the same industry. Visual interfaces such as the capacity health matrix and industry anomaly heat map are generated through PowerBI. When the "harmonic control device" field is missing in a chemical user's file and the measured harmonic distortion rate exceeds the limit, the system automatically and synchronously optimizes the weight distribution logic of the scoring model, and dynamically feedbacks typical problems such as industry classification mislabeling and equipment parameter distortion.

[0179] At the data governance level, a unified device coding system and cross-system data alignment rules have established a high-quality power safety data foundation, significantly reducing the cost of duplicate data collection and manual verification. Furthermore, the deployment of edge computing terminals and lightweight models has achieved low-cost, high-efficiency user-side implementation.

[0180] Step S2 specifically includes S2-1 and S2-2:

[0181] S2-1: Hidden danger identification based on data mining

[0182] Based on the monitoring and integrated management of the above data, a hidden danger identification model is constructed to identify safety hazards in high-risk circuits of power users.

[0183] Hidden danger identification model construction:

[0184] 1. Hidden danger feature extraction and fusion

[0185] (1) Hierarchical feature extraction

[0186] a) Device-level features

[0187] Temperature gradient: Calculate the temperature difference between adjacent monitoring points of the busbar joint (threshold: ΔT>15℃ triggers an early warning).

[0188] Partial discharge pulse mode: Identify typical defect waveforms (such as corona discharge and surface discharge) through the CNN model.

[0189] Leakage current and fault arc identification: Extract abnormal waveform features based on the wavelet transform method, and judge by the leakage current size and the current threshold when the arc occurs.

[0190] b) Loop-level characteristics

[0191] Three-phase imbalance

[0192]

[0193] Where U unbalance Indicates the three-item imbalance, I a, I b, I c Refers to the three-phase currents a, b, and c respectively.

[0194] Based on the LSTM method, the deviation of time series data is predicted.

[0195] Insulation Health Index: This index integrates insulation resistance, dielectric loss factor (tanδ), and ambient humidity to establish a comprehensive health index based on game theory weighting.

[0196] Security load safety: Based on the security load identification results in the circuit, determine whether the security has emergency power backup power supply and whether the security load is connected to the control circuit.

[0197] c) System-level characteristics

[0198] Topology vulnerability: Analyzes the power grid topology to determine whether the loop is in the N-1 verification critical state.

[0199] N: represents the total number of components required for normal operation of the power system (such as N lines, N generators, etc.); N-1: refers to the number of components that can still meet the load demand and not cause cascading failures when any one component (1) fails or is planned to exit.

[0200] Backup power supply support capability: Evaluate the backup power supply switching success rate in the event of a fault based on the backup power supply capacity and switch performance.

[0201] 2. Hidden danger identification and risk assessment

[0202] (1) Equipment-loop causal chain mining

[0203] Granger causality test: Analyze whether a temperature increase leads to a decrease in insulation resistance (p<0.05 determines a causal relationship, and p is a key indicator for determining whether the test result is statistically significant).

[0204] Knowledge graph matching: Retrieve historical similar cases through a large semantic model (e.g., "connector overheating → insulation carbonization → short circuit")

[0205] (2) Dynamic causal graph construction

[0206] a) Causal relationship discovery

[0207] Transfer entropy is used to quantify the causal strength between equipment parameters and circuit risks (such as the contribution of partial discharge signals to insulation degradation).

[0208] Build a priori causal rule base based on knowledge graph (for example: "circuit breaker aging → opening delay → short-circuit current exceeds the standard")

[0209] b) Hybrid reasoning mechanism

[0210] Data-driven: Granger causality test analyzes the correlation between equipment, circuits, and systems in historical power outage events.

[0211] Knowledge-driven: Through the semantic big model, inspection reports, user repair records, and economic load requirements are analyzed to extract the causal triple of "defect description → fault consequence".

[0212] (3) Dynamic risk assessment

[0213] Construction of risk indicator system: Determine the risk indicator system based on the standards and specifications of the power grid company.

[0214] Risk Assessment: Calculate the total risk value caused by a safety hazard based on subjective and objective weights. The subjective weights are calculated using the Analytic Hierarchy Process (AHP), while the objective weights are analyzed using the Entropy Weight Method (EWM).

[0215] Risk level classification:

[0216] Total Value at Risk grade Disposal response time ≥0.9 Level 1 Immediate response 0.7≤Risk<0.9 Level 2 within 15 minutes 0.5≤Risk<0.7 Level 3 Within 2 hours

[0217] (4) Causal transmission risk warning

[0218] a) Transmission path prediction: risk transmission simulation based on dynamic Bayesian networks, transmission path search based on grid topology, and path risk calculation using a weighted approach.

[0219] b) Hazard graph generation: Output risk association relationships based on the knowledge graph and Figure 2 The "one map of the power grid" is a schematic diagram of the system architecture of the method for identifying, warning and handling safety hazards in electricity use by power users of the present invention, on which the locations of high-risk circuits, types of hazards and conduction paths are superimposed.

[0220] c) Cross-layer warning threshold

[0221] Conduction level Warning trigger conditions Equipment → Loop Equipment failure probability > 60% and circuit load rate > 80% Loop → System Loop risk index>0.7 and system reserve capacity<critical load External → System Meteorological disaster level ≥ orange and system vulnerability index > 0.5

[0222] S2-2: Abnormal topology backtracking based on a single grid map

[0223] 1. Data integration

[0224] (1) Geographic location information: GPS coordinates of substations, utility poles, and cable wells (accurate to meters);

[0225] (2) Topological architecture data: power supply line connection relationship (which devices are grouped together) and power supply direction;

[0226] 2. Data integration

[0227] (1) Acquisition system minute-level data: voltage zero time, which transformer stopped working;

[0228] (2) Marketing archive data: factory / community electricity contract, contact information, and importance level (e.g., a hospital is marked as a first-level important user);

[0229] 3. Data location and backtracking calculation

[0230] (1) Data association matching

[0231] a) Spatial alignment: Map the outage location (e.g., "Distribution box No. X, Road XX") to the specific coordinates of the power grid map;

[0232] b) Timeline construction: record the sequence of events by minute;

[0233] (2) Model intelligent algorithm analysis

[0234] a) Fault analysis: From the fault point (e.g., a burned-out transformer), follow the grid topology to identify the affected area;

[0235] b) Source tracing back: tracing back from the point of power outage to the possible cause;

[0236] 4. Accurate positioning and backtracking

[0237] (1) Three-level positioning mechanism

[0238] a) Level 1 positioning (device level): directly locks the faulty device, for example: "No. 2 main transformer of XX substation trips";

[0239] b) Secondary positioning (line level): Analyze related lines, for example: "Due to a fault in cable 3, power supply lines A, B, and C are cut off."

[0240] c) Level 3 positioning (user level): Combined with marketing data, a list of affected users is created: "The power outage affected one hospital, three factories, and two residential communities."

[0241] (2) Visual Backtracking Dashboard

[0242] a) Map display: The fault point flashes red as a warning, and the affected area is displayed in orange. You can view: photos of the faulty equipment, maintenance history, and surrounding surveillance videos;

[0243] b) Timeline review: Generate an event timeline, for example: 13:00 construction team digs up the cable → 13:05 line short circuit → 13:06 circuit breaker trips → 13:08 user reports a repair;

[0244] In summary, S2 is based on cross-departmental and cross-level collaborative data sharing. The power user safety hazard identification and risk warning solution improves the timeliness and accuracy of risk warnings by integrating multi-source heterogeneous data, and optimizes the effectiveness of the traditional risk prevention and control system. Early risk warnings can help industrial users avoid economic losses of tens of millions caused by unplanned outages.

[0245] Leveraging semantic big models and knowledge graph technology, the system has achieved a significant upgrade from single-electrochemical monitoring to multi-dimensional risk linkage across "equipment, environment, and behavior." By building a dynamic rules engine and edge computing architecture, the system can adapt to the differentiated needs of regional power grids, providing a unified technical framework for preventing and controlling power safety risks in complex environments.

[0246] Through the spatiotemporal fusion and data correlation analysis of power grid topology data and operation status monitoring data, the accuracy and speed of fault location can be effectively improved, solving the pain points of traditional manual troubleshooting, which is inefficient and prone to missing hidden risks.

[0247] Step S3 specifically includes:

[0248] Closed-loop disposal of user electricity risk:

[0249] By building a large-scale model for electricity safety risk management, the system can intelligently identify risk types, associated devices, and users, and generate management strategies. This large-scale model involves structured data parsing, semi- and unstructured knowledge graph generation, and semantic understanding and reasoning to ensure the accuracy and comprehensiveness of early warning information. Risk warning information is based on S2-1 risk warning-related information, and the large-scale model enables the association of management information.

[0250] 1. Knowledge graph construction

[0251] (1) Structured: such as equipment records, user files, line topology, etc.

[0252] (2) Semi-structured: such as equipment maintenance logs, fault reports, etc.

[0253] (3) Unstructured: such as text reports, user feedback, social media data, etc.

[0254] (4) Semantic data: domain knowledge base, ontology model, etc. used for semantic understanding and reasoning.

[0255] 2. Emergency response and knowledge computing engine support

[0256] After an alert is generated, the system calls upon the knowledge computing engine for in-depth analysis. Through representation learning, relational reasoning, attribute reasoning, and event reasoning, the engine quickly locates the source of risk and generates emergency response plans. For example, for a high-voltage equipment overload alert, the engine analyzes equipment operating data, infers the cause of the overload (such as a sudden load surge or equipment failure), and recommends adjusting the load or activating backup equipment. At the same time, the system automatically matches relevant response processes (such as inspections, preliminary remediation, firefighting, and rescue) based on the emergency plan knowledge graph, and updates the response status in real time through MySQL and Neo4j databases.

[0257] (1) Core functions of the knowledge computing engine

[0258] Representation learning: Embed entities such as devices, users, and routes and their relationships into a low-dimensional vector space to facilitate subsequent reasoning and analysis.

[0259] Relational reasoning: Analyze the relationships between entities (such as the association between devices and lines, and the dependency between users and devices) to infer the risk propagation path.

[0260] Attribute reasoning: Infer the specific cause of the risk based on the attributes of equipment, lines, and users (such as current, voltage, load, etc.).

[0261] Event reasoning: Combine historical event data to predict risk development trends and generate emergency response plans.

[0262] (2) Emergency Response Process

[0263] Risk source location: Analyze equipment operation data, line operation data, etc. through the knowledge calculation engine to quickly locate the source of risk.

[0264] Disposal plan generation: Generate recommended plans based on risk type and cause.

[0265] Processing status update: The processing status is updated in real time through MySQL and Neo4j databases.

[0266] Multi-department collaboration: Based on the emergency plan knowledge graph, relevant responsible departments (such as operation and maintenance department, fire department, etc.) are automatically notified. Disposal suggestions and real-time data support are provided.

[0267] (3) Technical support

[0268] Knowledge graph construction: Use Neo4j to store entity relationships and support fast query and reasoning.

[0269] Real-time data processing: Use stream processing technologies (such as Apache Kafka and Flink) to process device operation data in real time.

[0270] Database support:

[0271] MySQL: Stores structured data such as equipment status, disposal records, etc.

[0272] Neo4j: stores knowledge graph data and supports complex relational reasoning.

[0273] Semantic understanding and reasoning:

[0274] Use Natural Language Processing (NLP) techniques to parse semi- / unstructured data (such as fault reports).

[0275] Use graph neural networks (GNNs) for relational reasoning and risk prediction.

[0276] 3. Ensure resource scheduling and multi-resource collaboration

[0277] During emergency response, the system dispatches grid emergency resources, electricity consumption emergency resources, and other collaborative resources based on analysis from the knowledge graph and knowledge computing engine. For example, in response to a short-circuit warning on a power line, the system dispatches backup lines and emergency power generation equipment to ensure uninterrupted power supply and coordinates maintenance teams to quickly repair the faulty line. Resource scheduling plans are optimized using the knowledge computing engine to maximize resource utilization efficiency. The system also incorporates historical experience from the emergency plan case library to provide the optimal response path.

[0278] (1) Core functions of resource scheduling and coordination

[0279] Resource matching: Match the most appropriate emergency resources based on risk type and disposal requirements.

[0280] Resource optimization: Optimize resource scheduling plans through the knowledge computing engine to ensure maximum resource utilization efficiency.

[0281] Multi-resource collaboration: Coordinate multiple resources such as power grid emergency resources, power emergency resources, and maintenance teams to achieve efficient collaboration.

[0282] Reference to historical experience: Combine historical experience in the emergency plan case library to provide the optimal response path.

[0283] (2) Resource scheduling and collaboration process

[0284] Risk Analysis and Needs Assessment:

[0285] Analyze risk types and impact scope through the knowledge calculation engine to assess resource requirements.

[0286] Resource matching and scheduling: Matching available resources according to demand.

[0287] Reference to historical experience: Combine historical experience in the emergency plan case library to provide the optimal response path.

[0288] 4. Information confirmation and personnel interaction

[0289] During the response process, operations personnel interact with the knowledge graph through the system to confirm the response results and record any special requirements. For example, to alert a user of a potential electrical safety hazard, operations personnel will inspect the user's site, confirm the hazard's rectification status, and record suggestions for improving the user's electrical behavior through the system. The system dynamically updates the knowledge graph and knowledge computing engine based on the input from personnel, ensuring the accuracy of subsequent warnings and responses.

[0290] (1) Core functions of information confirmation and personnel interaction

[0291] Information confirmation: Operation and maintenance personnel confirm the disposal results through the system to ensure that the risks have been eliminated or controlled.

[0292] Special requirement records: record the specific needs or improvement suggestions of users or sites.

[0293] Knowledge graph update: Dynamically update the knowledge graph and knowledge calculation engine based on the information entered by personnel.

[0294] Interactive support: Provides a friendly user interface and interactive mode to facilitate quick operation by operation and maintenance personnel.

[0295] (2) Information confirmation and personnel interaction process

[0296] Confirmation of disposal results:

[0297] Operation and maintenance personnel view early warning information and disposal plans through the system.

[0298] Conduct on-site inspections to check the rectification of hidden dangers on equipment, lines or user sides.

[0299] Confirm the disposal results in the system (such as hidden dangers have been eliminated and equipment has returned to normal).

[0300] Special requirements record:

[0301] Record the specific needs of users or sites (such as adding backup power and adjusting power usage hours).

[0302] Record suggestions for improving users’ electricity usage behavior (such as reducing the simultaneous use of high-power devices).

[0303] Knowledge Graph Updates:

[0304] The knowledge graph is dynamically updated based on the information entered by the operation and maintenance personnel.

[0305] 5. Experience storage and knowledge accumulation

[0306] After the risk is resolved, the system will enter the warning's response plan, fault cause, resource scheduling, and other information into the emergency plan case library to form standardized experience. Through the knowledge computing engine's path calculation and comparison sorting functions, the system can automatically extract key experience and optimize the association relationships and inference rules in the knowledge graph. For example, for high-voltage equipment overload warnings, the system will summarize best practices for load adjustment and incorporate them into the knowledge graph to provide reference for similar risks in the future. Ultimately, these experiences will further enhance the system's intelligent warning and response capabilities through representation learning and relational reasoning.

[0307] (1) Core functions of experience storage and knowledge accumulation

[0308] Case standardization: Standardize information such as disposal plans, fault causes, and resource scheduling and enter them into the case library.

[0309] Key experience extraction: Extract key experiences through the knowledge computing engine and optimize the knowledge graph.

[0310] Knowledge graph update: Update the association relationships and inference rules in the knowledge graph.

[0311] Improved intelligent capabilities: Improve the system's early warning and response capabilities through representation learning and relational reasoning.

[0312] (2) Experience storage and knowledge accumulation process

[0313] Case standardization:

[0314] Standardize the information of this warning, such as the disposal plan, fault cause, resource scheduling, etc.

[0315] Enter the emergency plan case library to form reusable experience.

[0316] Key experience extraction: Extract key experiences through the path calculation and comparison sorting functions of the knowledge calculation engine.

[0317] Knowledge graph update: Update the association relationships and inference rules in the knowledge graph.

[0318] Improved intelligence capabilities:

[0319] Through representation learning, key experiences are embedded into the vector space of the knowledge graph.

[0320] Through relational reasoning, the warning rules and disposal plan recommendation logic are optimized.

[0321] Case library management:

[0322] Classify and label the case library to facilitate quick retrieval and matching. Regularly clear outdated or invalid cases to ensure the timeliness of the case library.

[0323] In summary, the data-driven handling strategy generation technology in step S3 further transforms emergency handling from "experience-based" to "intelligent decision-making". For example, in the event of a power supply circuit failure for an important user, the system can simultaneously push load shedding plans, backup power supply switching paths, and emergency repair resource scheduling suggestions, forming a complete risk handling closed loop.

[0324] Example 2

[0325] Based on the same inventive concept, the present invention also provides a system for identifying, warning and handling potential safety hazards of power users, such as Figure 3 As shown, it includes: a module for acquiring multi-source power data, an intelligent analysis module and a disposal module;

[0326] The multi-source power data acquisition module is used to acquire multi-source power data;

[0327] The intelligent analysis module is used to identify, warn, trace back, and locate user power safety hazards based on multi-source power data, a hidden danger identification model, spatiotemporal data association, and a three-level positioning mechanism of equipment, lines, and users;

[0328] The treatment module is used to input information related to safety hazard identification, warning and location into the power safety risk treatment model to generate a treatment strategy;

[0329] The hidden danger identification model is trained with multi-source power data as input and safety hidden danger identification and cross-layer early warning as output;

[0330] Among them, the large model for handling electricity safety risks is trained with safety hazard identification, warning and positioning information as input and handling strategies as output.

[0331] Preferably, the module for acquiring multi-source power data is specifically used to:

[0332] Collect data from multiple sources;

[0333] A two-level data governance architecture, edge-side and master-station, is used to standardize the data framework and integrate multi-source power data.

[0334] Perform data quality verification and data quality assessment on multi-source power data;

[0335] The data quality check includes at least one or more of the following: format check, consistency check, integrity check or logic check.

[0336] Preferably, the intelligent analysis module is specifically used to:

[0337] Based on multi-source power data, the system can identify and issue early warnings for users’ power safety hazards based on hidden danger identification models.

[0338] Conduct fault analysis based on multi-source power data and spatiotemporal data correlation to identify affected areas and perform anomaly retrospective analysis to analyze the cause of the power outage;

[0339] Based on multi-source power data and identified hidden dangers, a three-level positioning mechanism of equipment-line-user is adopted to locate users' power safety hazards.

[0340] Preferably, the intelligent analysis module identifies and warns users of potential power safety hazards based on multi-source power data and a potential safety hazard identification model, including:

[0341] Performing layered extraction of hidden danger features on multi-source power data to obtain layered multi-source power data;

[0342] Inputting layered multi-source power data into the hidden danger identification model to obtain cross-layer safety hidden danger identification and early warning results;

[0343] The level of the hidden danger feature includes at least one or more of the following: a device-level feature level, a circuit-level feature level, or a system-level feature level;

[0344] The hidden danger identification model is trained by taking layered multi-source power data as input and cross-layer safety hidden danger identification and early warning results as output.

[0345] Preferably, the intelligent analysis module inputs the hierarchical multi-source power data into the hidden danger identification model to obtain cross-layer safety hidden danger identification and early warning results, including:

[0346] Granger causality test or knowledge graph matching is used to mine the equipment-loop causal chain of hierarchical multi-source power data to obtain the equipment-loop causal chain;

[0347] Using causal relationship discovery methods and hybrid reasoning mechanisms, we can find other related causal chains at each node based on the existing device-loop causal chain to build a dynamic causal graph.

[0348] Make dynamic risk assessment based on dynamic cause-effect diagram;

[0349] Conducting risk conduction simulation based on Bayesian network according to hierarchical multi-source power data to obtain conduction path prediction results;

[0350] The results of conduction path prediction and dynamic risk assessment are used as the results of cross-layer safety hazard identification and early warning.

[0351] Preferably, the intelligent analysis module mines the device-loop causal chain using Granger causality test or knowledge graph matching on the hierarchical multi-source power data to obtain the device-loop causal chain, including:

[0352] The Granger causality test method is used to determine the causal relationship between hierarchical multi-source power data and obtain the equipment-loop causal chain;

[0353] Or input hierarchical multi-source power data, use the knowledge graph matching method, and search historical similar cases through the semantic big model to obtain the equipment loop causal chain.

[0354] Preferably, the intelligent analysis module uses a causal relationship discovery method and a hybrid reasoning mechanism to construct a dynamic causal graph based on finding other causal chains related to each node of the existing device-loop causal chain, including:

[0355] Based on the existing equipment-loop causal chain related node data, transfer entropy is used to quantify the causal strength between equipment parameters and loop risks, and a priori causal rule base is constructed based on the knowledge graph to discover the causal relationship of related node data;

[0356] The Granger causality test method is used to analyze the correlation between equipment, circuits, and systems in historical power outages. Based on the existing equipment-circuit causal chain node data and the equipment-circuit-system correlation, a semantic big model is used to analyze inspection reports, user repair records, and economic load demand to extract causal triples between related node data.

[0357] Other causal chains are formed based on the causal relationships and causal triples of the relevant node data, and the causal chains are interwoven to form a dynamic causal graph.

[0358] Preferably, the intelligent analysis module makes a dynamic risk assessment based on the dynamic cause-effect diagram, including:

[0359] Based on the hierarchical multi-source power data related to safety hazards in the dynamic causal diagram, the subjective weight of each preset risk indicator is calculated using the hierarchical analysis method, and the objective weight of each risk indicator is calculated using the entropy weight method;

[0360] Based on the hierarchical multi-source power data related to safety hazards in the dynamic causal diagram, the value of each risk indicator is calculated using the calculation formula of each preset risk indicator;

[0361] Based on the value of each risk indicator and the subjective weight and objective weight of each risk indicator, a weighted calculation method is used to calculate the dynamic risk assessment value;

[0362] Conduct dynamic risk assessment of potential safety hazards based on preset risk levels and dynamic risk evaluation.

[0363] Preferably, the intelligent analysis module performs conduction path simulation based on risk conduction simulation of Bayesian network according to hierarchical multi-source power data to obtain conduction path prediction results, including:

[0364] Simulating risk transmission paths based on Bayesian networks according to hierarchical multi-source power data or searching for risk transmission paths based on transmission paths of power grid topology according to hierarchical multi-source power data;

[0365] The path risk of the risk transmission path is calculated according to the weighting method to obtain the transmission path prediction result.

[0366] Preferably, the intelligent analysis module performs fault analysis based on the correlation of multi-source power data and spatiotemporal data to identify the affected area and conducts anomaly retrospective analysis to determine the cause of the power outage, including:

[0367] Integrate and connect multi-source power data;

[0368] Based on the integrated and connected multi-source power data, spatiotemporal correlation matching and model intelligent algorithm analysis are performed to find the affected areas and trace back the causes of the power outage.

[0369] Preferably, the intelligent analysis module performs spatiotemporal correlation matching and model intelligent algorithm analysis based on the integrated and connected multi-source power data to find the affected area and trace back the cause of the power outage, including:

[0370] Based on the integrated and connected multi-source power data, the outage location is mapped to the specific coordinates of the power grid map, and the outage time is recorded according to the timeline;

[0371] Conduct fault analysis along the grid topology based on the specific coordinates and duration of the outage to identify the affected area;

[0372] Based on the specific coordinates and time of the power outage, we can trace back to the source and find the substation that had the power outage first, and then infer the cause of the power outage.

[0373] Preferably, the intelligent analysis module locates potential safety hazards of user electricity use based on multi-source power data and a three-level positioning mechanism of equipment, lines, and users, including:

[0374] Directly locate faulty equipment based on multi-source power data and identified hidden dangers, and perform device-level location of user power safety hazards;

[0375] Based on multi-source power data and identified hidden dangers, topology analysis is performed to obtain associated lines related to the fault, and line-level positioning of user power safety hazards is performed;

[0376] Based on multi-source power data and identified hidden dangers, compared with marketing data, a list of affected users is drawn up, and user-level positioning of power safety hazards is performed.

[0377] Preferably, the multi-source power data includes at least one or more of the following: user profile data, electricity usage behavior data, high-risk circuit data, equipment operation data, PMS system data, 95598 planned power outage data, scheduling support system data, power outage handling data or safety and compliance data.

[0378] In summary, the present invention provides a method and system for identifying, warning, and addressing power safety hazards for power users, comprising: acquiring multi-source power data; identifying, warning, tracing back anomalies, and locating power safety hazards for users based on a hazard identification model, spatiotemporal data correlation, and a three-level device-line-user positioning mechanism based on the multi-source power data; and inputting information related to hazard identification, warning, and location into a large-scale power safety risk management model to generate a management strategy. By employing multi-source data fusion processing, hierarchical feature extraction, Granger causality testing, knowledge graph matching, and a three-level positioning mechanism, the present invention enhances data governance and hazard management capabilities, improves risk identification and classification accuracy, and shortens warning response time.

[0379] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0380] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0381] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0382] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0383] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for identifying, warning and handling potential safety hazards of electricity users, characterized in that: include: Acquire multi-source power data; Based on multi-source power data, the system can identify, warn, trace back, and locate power safety hazards to users based on hidden danger identification models, spatiotemporal data association, and a three-level positioning mechanism of equipment, lines, and users. Input the relevant information on potential safety hazards identification, warning and location into the large-scale model for power safety risk management to generate a management strategy; The hidden danger identification model is trained with multi-source power data as input and safety hidden danger identification and cross-layer early warning as output; Among them, the large model for handling electricity safety risks is trained with safety hazard identification, warning and positioning information as input and handling strategies as output.

2. The method according to claim 1, wherein The acquiring of multi-source data includes: Collect data from multiple sources; A two-level data governance architecture, edge-side and master-station, is used to standardize the data framework and integrate multi-source power data. Perform data quality verification and data quality assessment on multi-source power data; The data quality check includes at least one or more of the following: format check, consistency check, integrity check or logic check.

3. The method according to claim 1, wherein Based on multi-source power data, the system identifies, warns, and traces back abnormalities and locates hidden dangers in user electricity safety, based on a hidden danger identification model, spatiotemporal data association, and a three-level positioning mechanism of equipment, lines, and users. This includes: Based on multi-source power data, the system can identify and issue early warnings for users’ power safety hazards based on hidden danger identification models. Conduct fault analysis based on multi-source power data and spatiotemporal data correlation to identify affected areas and perform anomaly retrospective analysis to analyze the cause of the power outage; Based on multi-source power data and identified hidden dangers, a three-level positioning mechanism of equipment-line-user is adopted to locate users' power safety hazards.

4. The method according to claim 3, wherein The method of identifying and warning users' power safety hazards based on multi-source power data and a hidden danger identification model includes: Performing layered extraction of hidden danger features on multi-source power data to obtain layered multi-source power data; Inputting layered multi-source power data into the hidden danger identification model to obtain cross-layer safety hidden danger identification and early warning results; The level of the hidden danger feature includes at least one or more of the following: a device-level feature level, a circuit-level feature level, or a system-level feature level; The hidden danger identification model is trained by taking layered multi-source power data as input and cross-layer safety hidden danger identification and early warning results as output.

5. The method according to claim 4, wherein The step of inputting the layered multi-source power data into the hidden danger identification model to obtain cross-layer safety hidden danger identification and early warning results includes: Granger causality test or knowledge graph matching is used to mine the equipment-loop causal chain of hierarchical multi-source power data to obtain the equipment-loop causal chain; Using causal relationship discovery methods and hybrid reasoning mechanisms, we can find other related causal chains at each node based on the existing device-loop causal chain to build a dynamic causal graph. Make dynamic risk assessment based on dynamic cause-effect diagram; Conducting risk conduction simulation based on Bayesian network according to hierarchical multi-source power data to obtain conduction path prediction results; The results of conduction path prediction and dynamic risk assessment are used as the results of cross-layer safety hazard identification and early warning.

6. The method according to claim 5, wherein The device-loop causal chain is mined by using Granger causality test or knowledge graph matching on hierarchical multi-source power data to obtain the device-loop causal chain, including: The Granger causality test method is used to determine the causal relationship between hierarchical multi-source power data and obtain the equipment-loop causal chain; Or input hierarchical multi-source power data, use the knowledge graph matching method, and search historical similar cases through the semantic big model to obtain the equipment loop causal chain.

7. The method according to claim 5, wherein The causal relationship discovery method and hybrid reasoning mechanism are used to construct a dynamic causal graph by finding other causal chains related to each node of the existing device-loop causal chain, including: Based on the existing equipment-loop causal chain related node data, transfer entropy is used to quantify the causal strength between equipment parameters and loop risks, and a priori causal rule base is constructed based on the knowledge graph to discover the causal relationship of related node data; The Granger causality test method is used to analyze the correlation between equipment, circuits, and systems in historical power outages. Based on the existing equipment-circuit causal chain node data and the equipment-circuit-system correlation, a semantic big model is used to analyze inspection reports, user repair records, and economic load demand to extract causal triples between related node data. Other causal chains are formed based on the causal relationships and causal triples of the relevant node data, and the causal chains are interwoven to form a dynamic causal graph.

8. The method according to claim 5, wherein The dynamic risk assessment based on the dynamic causal diagram includes: Based on the hierarchical multi-source power data related to safety hazards in the dynamic causal diagram, the subjective weight of each preset risk indicator is calculated using the hierarchical analysis method, and the objective weight of each risk indicator is calculated using the entropy weight method; Based on the hierarchical multi-source power data related to safety hazards in the dynamic causal diagram, the value of each risk indicator is calculated using the calculation formula of each preset risk indicator; Based on the value of each risk indicator and the subjective weight and objective weight of each risk indicator, a weighted calculation method is used to calculate the dynamic risk assessment value; Conduct dynamic risk assessment of potential safety hazards based on preset risk levels and dynamic risk evaluation.

9. The method according to claim 5, wherein The method of performing a conduction path simulation based on the risk conduction simulation of the Bayesian network according to the hierarchical multi-source power data to obtain a conduction path prediction result includes: Simulating risk transmission paths based on Bayesian networks according to hierarchical multi-source power data or searching for risk transmission paths based on transmission paths of power grid topology according to hierarchical multi-source power data; The path risk of the risk transmission path is calculated according to the weighting method to obtain the transmission path prediction result.

10. The method according to claim 3, wherein The fault analysis based on the correlation of multi-source power data and spatiotemporal data to identify the affected area and conduct anomaly retrospective analysis of the cause of the power outage includes: Integrate and connect multi-source power data; Based on the integrated and connected multi-source power data, spatiotemporal correlation matching and model intelligent algorithm analysis are performed to find the affected areas and trace back the causes of the power outage.

11. The method according to claim 10, wherein The method of performing spatiotemporal correlation matching and model intelligent algorithm analysis based on the integrated and connected multi-source power data to identify the affected areas and trace back the causes of the power outage includes: Based on the integrated and connected multi-source power data, the outage location is mapped to the specific coordinates of the power grid map, and the outage time is recorded according to the timeline; Conduct fault analysis along the grid topology based on the specific coordinates and duration of the outage to identify the affected area; Based on the specific coordinates and time of the power outage, we can trace back to the source and find the substation that had the power outage first, and then infer the cause of the power outage.

12. The method according to claim 3, wherein The method of locating user power safety hazards based on multi-source power data and the three-level positioning mechanism of equipment, lines, and users includes: Directly locate faulty equipment based on multi-source power data and identified hidden dangers, and perform device-level location of user power safety hazards; Based on multi-source power data and identified hidden dangers, topology analysis is performed to obtain associated lines related to the fault, and line-level positioning of user power safety hazards is performed; Based on multi-source power data and identified hidden dangers, compared with marketing data, a list of affected users is drawn up, and user-level positioning of power safety hazards is performed.

13. The method according to claim 1, wherein The multi-source power data includes at least one or more of the following: user profile data, power consumption behavior data, high-risk circuit data, equipment operation data, PMS system data, 95598 planned power outage data, scheduling support system data, power outage handling data or safety and compliance data.

14. A system for identifying, warning and handling potential safety hazards of power users, characterized by: include: Acquisition of multi-source power data module, intelligent analysis module and disposal module; The multi-source power data acquisition module is used to acquire multi-source power data; The intelligent analysis module is used to identify, warn, trace back, and locate user power safety hazards based on multi-source power data, a hidden danger identification model, spatiotemporal data association, and a three-level positioning mechanism of equipment, lines, and users; The treatment module is used to input information related to safety hazard identification, warning and location into the power safety risk treatment model to generate a treatment strategy; The hidden danger identification model is trained with multi-source power data as input and safety hidden danger identification and cross-layer early warning as output; Among them, the large model for handling electricity safety risks is trained with safety hazard identification, warning and positioning information as input and handling strategies as output.

Images