Power grid operation situation analysis system combined with knowledge graph

Abstract

The invention discloses a power grid operation situation analysis system combined with a knowledge graph, which relates to the technical field of intelligent power grids and comprises a power grid data acquisition module, a power grid feature extraction module, a monitoring information event analysis module, a power grid graph construction module and a power grid event intelligent research and judgment module. The power grid data acquisition module acquires power grid data, the power grid feature extraction module extracts features of the power grid data, the monitoring information event analysis module outputs an overhaul label, a fault label, an associated equipment list and a fault list, and the power grid graph construction module generates a power grid knowledge graph and outputs a fault propagation path and an out-of-limit threshold. And the power grid event intelligent research and judgment module pushes abnormal data and fault information. According to the method, automatic cleaning of massive monitoring signals, intelligent fault research and judgment and cross-station propagation path analysis are realized through natural language processing and machine learning technologies, manual intervention is reduced, the fault recognition accuracy is improved, graded alarms are pushed in real time, and the power grid situation awareness and emergency disposal efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and in particular to a power grid operation status analysis system combined with a knowledge graph. Background Art

[0002] The grid's operational status is a comprehensive reflection of the real-time operating conditions of the power system, encompassing the dynamic interaction of power supply and demand, equipment status, the stability of renewable energy access, and external influencing factors. Its core objective is to maintain a real-time balance between power supply and demand, ensuring that power generation capacity meets load requirements and avoiding both power shortages and surpluses. The operational status of grid equipment, including the load, failure rate, and aging of transmission lines and transformers, directly impacts the stability and reliability of the system. The rapid development of renewable energy significantly impacts grid operational status. The volatility of wind and photovoltaic power increases the pressure on frequency and voltage regulation, but intelligent control and regulation systems can optimize their access and improve stability. At the same time, external factors can also cause disturbances to the grid, requiring corresponding response capabilities. In terms of technical support, knowledge graph technology integrates multidimensional data to construct a hierarchical model, providing an intuitive understanding of grid status. Intelligent monitoring systems collect and analyze data in real time, predict risks, and issue warnings. Combined with electronic dashboards, management visualization and transparency are achieved, improving decision-making efficiency. The stability of the grid's operational status requires a comprehensive consideration of supply and demand, equipment, renewable energy, and external factors. Technical support is used to optimize management, ensure safe and efficient operation, and support economic and social development.

[0003] To address the inefficiencies and misjudgments caused by the inefficient processing of massive amounts of monitoring signals during power grid operation status analysis, as well as the reliance on manual experience, existing technologies rely on manual screening and static rule engines. However, this approach can result in the inability to effectively correlate unstructured data, and static rules that are unable to dynamically adapt to changes in power grid topology. This can lead to missed critical alarms, misjudgments of multiple fault correlations, and ineffective analysis of cross-site fault propagation paths. To address these issues, a power grid operation status analysis system combining knowledge graphs is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a power grid operation status analysis system combined with a knowledge graph to solve the problems raised in the above background technology.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is: a power grid operation status analysis system combined with knowledge graph, including a power grid data acquisition module, a power grid feature extraction module, a monitoring information event analysis module, a power grid graph construction module and a power grid event intelligent analysis module; The power grid data acquisition module collects and pre-processes power grid data, wherein the power grid data includes telesignaling data, telemetry data, OMS system data and D5000 system data; The power grid feature extraction module extracts telesignaling features, maintenance event features and telemetry features from the preprocessed power grid data; The monitoring information event analysis module combines remote signaling features, maintenance event features, and telemetry features to build a maintenance event analysis model, a power grid fault analysis model, and a power grid fault clustering model, and outputs equipment maintenance and debugging information tags, fault time tags, a list of associated devices, and a list of associated faults; The power grid map construction module constructs a power grid knowledge map based on equipment maintenance and debugging information tags, fault time tags, associated equipment lists, and associated fault lists, outputs fault propagation paths and voltage over-limit probability thresholds, and performs graded warnings. The power grid event intelligent analysis module analyzes power grid data in real time, generates disposal plans, and pushes abnormal data and power grid faults.

[0006] A further improvement of the technical solution of the present invention is that: in the power grid data acquisition module, the power grid data acquisition and preprocessing process includes: Receive telesignaling data from power grid equipment of all voltage levels, from 10kV to 500kV, through the real-time service interface of the dispatching data center. The telesignaling data includes alarm information on accidents, anomalies, displacements, and over-limits. Specific fields include the equipment key, occurrence time, voltage level, alarm flag, and switch open / close status. Regularly obtain telemetry data of power grid equipment through the minute-by-minute measurement service interface of the scheduling data center. The telemetry data includes measurement parameters such as active power, reactive power, voltage amplitude, current value, and oil temperature. OMS system data includes maintenance application data and trip record data, and D5000 system data includes monitoring operation data and card placement data; By connecting the dispatch data center to the OMS system service interface, maintenance application data and trip record data are obtained in real time. The maintenance application data includes the power outage equipment, maintenance period and approval time, and the trip record data includes the fault type, trip device association information and trip switch; By connecting the dispatch data center to the D5000 system service interface, real-time monitoring operation data and card placement data are obtained. The monitoring operation data includes operation time, operated device and operation result, and the card placement data includes card placement type, device ID and operation time. By verifying the logic rules of unreasonable parallel connection of main transformers and inappropriate switch positions, redundancy and error signals in remote signaling data are eliminated to generate standardized equipment status data. Identify outliers in telemetry data based on the 3σ principle, and fill in missing telemetry data by linear interpolation of adjacent time points to generate continuous telemetry time series data. Using bidirectional long short-term memory network word segmentation and conditional random field entity recognition technology, unstructured maintenance application data and trip record data are converted into device-related structured data to obtain trip record analysis results; The validity of operation data and card placement data is verified by matching operation results with real-time status. The relationship between card placement type and equipment maintenance status is analyzed using a support vector machine model to generate a standardized operation log.

[0007] A further improvement of the technical solution of the present invention is that: in the power grid feature extraction module, the process of extracting telesignaling features, maintenance event features, and telemetry features includes: Telesignaling features include alarm type codes and switch opening and closing state vectors. Alarm information is mapped to numerical labels. Conditional judgment logic is used to generate corresponding code vectors based on the alarm flag bits. The alarm type code is obtained, and the binary state value of the switch opening and closing state field in the telesignaling data is extracted to generate the real-time switch opening and closing state vector of the device. Maintenance event features include a maintenance window period flag and an equipment maintenance flag. Based on the maintenance application data and telesignaling data from the OMS system, it is determined whether the current time is within the approved maintenance application period. Combined with the switch opening status, a Boolean maintenance window period flag is generated, with 0 indicating a non-maintenance period and 1 indicating a maintenance period. Based on the card placement data and maintenance window period flag from the D5000 system, a support vector machine model is used to analyze the relationship between the card placement type and the equipment status, and the equipment maintenance flag is output. The card placement types include maintenance cards and test cards. Telemetry features include active power fluctuation rate and voltage over-limit frequency. Based on the preprocessed active power time series data, a time window is set, the mean active power in the window is extracted, and the average absolute deviation of the active power and its mean at each time point in the window is calculated to obtain the active power fluctuation rate. Based on the preprocessed voltage amplitude time series data, the rated voltage is set. With the rated voltage as the benchmark, the over-limit range is defined. The number of times the voltage exceeds 1.1 times the rated voltage and is lower than 0.9 times the rated voltage per unit time is counted to obtain the voltage over-limit frequency.

[0008] A further improvement of the technical solution of the present invention is that: in the monitoring information event analysis module, the process of constructing a maintenance event analysis model and outputting equipment maintenance and debugging information tags includes: Using the switch state vector and the maintenance window flag as input features, a decision tree model is used to build a maintenance event analysis model based on determining whether the current time falls within the approved maintenance application period and checking whether the switch state is open. The root node of the decision tree prioritizes splitting the maintenance window period flag to filter out the equipment in the maintenance period. The child node splits the switch opening and closing status. If the switch opening condition is met within the approved maintenance application period, the power grid equipment is determined to be in the maintenance and debugging state. 0 represents the non-maintenance and debugging state, and 1 represents the maintenance and debugging state. Finally, the leaf node outputs the equipment maintenance and debugging information label.

[0009] A further improvement of the technical solution of the present invention is that: in the monitoring information event analysis module, the process of constructing a power grid fault analysis model and outputting a fault time label and a list of associated devices includes: Using the alarm type code, active power fluctuation rate, and trip record analysis results as input features, a fault mode rule base is established based on historical fault cases. The input features are compared one by one with the conditions in the fault mode rule base. If the alarm type code, active power fluctuation rate, and trip device association information meet the rule conditions, the corresponding fault tag is triggered. The alarm type code, active power fluctuation rate, and trip device primary key are converted into input feature vectors. The support vector classification algorithm is used to find the optimal classification hyperplane based on the principle of maximum margin. The fault category labels are divided according to the projection position of the input feature vector on the hyperplane. Combined with the topological relationship of the power grid model, the fault impact path is determined, and the device primary key and topologically associated devices in the trip record are extracted to form a list of associated devices; According to the alarm occurrence time and the fault mode rule matching result, the fault start timestamp is marked, and the fault time label and the associated device list including the fault source device and the topologically associated devices in the single fault event are output.

[0010] A further improvement of the technical solution of the present invention is that: in the monitoring information event analysis module, the process of constructing a power grid fault clustering model and outputting a list of associated faults includes: The fault start timestamp, trip device primary key, and fault category label are used as input features to define the time range of fault association. Two fault events with timestamps that do not exceed the length of the time window are considered to be associated events within the same time window. Check whether the primary keys of the tripped devices are the same. If they are the same, they are grouped into the same device fault cluster. If they are different, they are not grouped into the same device fault cluster. According to the power grid topology model, if the faulty devices belong to the same bay and the same substation, they are grouped into the same cluster. Based on the topological connection relationship, cross-station device faults are grouped into the same fault cluster. A decision tree clustering algorithm is adopted. The root node of the decision tree uses the time window as the first splitting condition to screen out fault events within the same time window. The child nodes are split according to the primary key of the tripping device to distinguish faults of different source devices. The fault subclusters are generated layer by layer according to the topological level to construct a power grid fault clustering model. Each leaf node represents a fault cluster, and an associated fault list and fault cluster label are output. The associated fault list is a list of multiple fault events within the same time window, the same device and the same topological association.

[0011] A further improvement of the technical solution of the present invention is that: in the power grid map construction module, the process of constructing the power grid knowledge map and outputting the fault propagation path and the voltage over-limit probability threshold includes: Build a power grid knowledge graph using equipment maintenance and debugging information tags, fault time tags, associated equipment lists, associated fault lists, and voltage over-limit frequencies as input; The power grid knowledge graph uses the primary key of the tripping device as the core to construct device node attributes, including device maintenance and debugging information tags and voltage over-limit frequency. Based on the event sequence in the associated fault list, the fault propagation relationship between devices is established, and the initial expert rule weights are set based on historical fault data. If the equipment is under maintenance, the weight of its associated path is multiplied by the attenuation coefficient. The voltage over-limit frequency threshold is set. If the voltage over-limit frequency exceeds the corresponding threshold, the weight of the relevant path increases, and the weight of the propagation path generated by the cross-site equipment due to the grid topology connection is increased by 20%; The multiple fault events in the associated fault list are sorted by timestamp to generate a fault propagation path. The path trigger probability is obtained based on the sum of the weights of each node in the path. The expert experience coefficient is introduced to dynamically adjust the voltage over-limit probability threshold based on the voltage over-limit frequency and the path trigger probability. If the real-time voltage exceeds the voltage over-limit probability threshold, a graded warning is issued.

[0012] A further improvement of the technical solution of the present invention is that: in the power grid map construction module, the process of performing graded warning includes: The power grid map construction module receives fault propagation paths, voltage over-limit probability thresholds, and real-time voltage data. It divides the alarm levels into emergency, major, and general alarms based on the path trigger probability. It matches the corresponding path trigger probability threshold ranges to different alarm levels. Based on the voltage over-limit frequency and path weight, it dynamically adjusts the voltage over-limit probability threshold. If the real-time voltage exceeds the voltage over-limit probability threshold, an alarm of the corresponding level is triggered. Cross-site equipment failures affect substations through the power grid topology, increasing the risk of cascading failures. If the path trigger probability corresponds to the emergency alarm threshold range, the real-time voltage exceeds the voltage over-limit probability threshold, and the fault propagation path includes cross-site equipment, then it is determined to be an emergency alarm. If one of the following conditions is met: the path trigger probability is within the emergency alarm and major alarm threshold range and exceeds the voltage over-limit probability threshold, or a single device is overloaded and the real-time voltage exceeds the voltage over-limit probability threshold, then it is determined to be a major alarm. If the path trigger probability corresponds to the general alarm threshold range and the real-time voltage does not exceed the voltage over-limit probability threshold, then it is determined to be a general alarm.

[0013] A further improvement of the technical solution of the present invention is that: in the intelligent analysis module for power grid events, the process of real-time analysis of power grid data and generation of treatment plans includes: Based on the fault propagation path and grid topology model, the loss load is automatically calculated using the equipment association relationship and expert rule weights in the grid knowledge graph. , generate the load shedding plan, the calculation process is as follows: ; in, represents the active power of the i-th load point, represents the priority weight of the i-th load point, is the total number of load points; Combined with a wiring diagram simulation drill, the risk points of the load shedding plan are analyzed, and the strategy is dynamically optimized. By integrating expert rules and machine learning, auxiliary decision-making that takes into account both safety and efficiency is output to support dispatchers to respond quickly.

[0014] A further improvement of the technical solution of the present invention is that: in the intelligent power grid event analysis module, the process of pushing abnormal data and power grid faults includes: Based on the alarm level and voltage over-limit probability threshold, hierarchical alarms are triggered dynamically. Abnormal data is classified by voltage level and over-limit frequency. High-priority abnormalities are pushed first to user terminals within the corresponding dispatching jurisdiction and fault impact range. Fault information is divided into areas according to dispatching authority and pushed to mobile terminals and on-duty terminals. The fault information includes fault location, related fault list and load shedding plan, and low-latency transmission is achieved through message queues, so that alarms can be delivered in real time.

[0015] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art: The present invention provides a power grid operation status analysis system combined with a knowledge graph. It uses natural language processing technology to parse unstructured maintenance applications and tripping records, and dynamically constructs a knowledge graph by integrating multi-source data. It significantly improves the processing efficiency of massive monitoring signals, solves the problems of missed detection and misjudgment in manual screening, realizes the automatic cleaning and correlation analysis of an average of 10,000 alarm information per day, and reduces the repetitive work of dispatchers.

[0016] The present invention provides a power grid operation status analysis system combined with a knowledge graph. Based on the dual diagnosis mode of the scheduling expert rule base and the support vector machine algorithm, it dynamically adjusts the weight of the fault propagation path, accurately identifies the correlation of cross-site equipment faults, improves the accuracy of fault analysis, and effectively avoids the risk of misjudgment of new fault modes by traditional static rule bases.

[0017] The present invention provides a power grid operation status analysis system combined with a knowledge graph. Through real-time voltage over-limit frequency statistics and path trigger probability calculation, it dynamically generates three-level alarm thresholds, and combines with the mobile terminal authority push mechanism to achieve emergency faults reaching the dispatch terminal within 5 seconds, shortening the emergency response time and ensuring the safety and stability of power grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0019] Figure 1 A block diagram of the present invention. DETAILED DESCRIPTION

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0021] Examples, such as Figure 1 As shown, the present invention provides a power grid operation status analysis system combined with knowledge graph, including a power grid data acquisition module, a power grid feature extraction module, a monitoring information event analysis module, a power grid graph construction module and a power grid event intelligent analysis module; The power grid data acquisition module collects and pre-processes power grid data, wherein the power grid data includes telesignaling data, telemetering data, OMS system data and D5000 system data. Through the real-time service interface of the dispatching data center, the telesignaling data of power grid equipment of all voltage levels from 10kV to 500kV is received. The telesignaling data includes alarm information of accidents, abnormalities, displacements and over-limits. The specific fields are the equipment main key, occurrence time, voltage level, alarm flag and switch opening and closing status. Through the minute-by-minute measurement service interface of the dispatching data center, the telemetering data of power grid equipment is regularly obtained. The telemetering data includes measurement parameters such as active power, reactive power, voltage amplitude, current value and oil temperature. The OMS system data includes maintenance application data and trip record data. The D5000 system data includes monitoring operation data and card placement data. Through the dispatching data center docking with the OMS system service interface, the maintenance application data and trip record data are obtained in real time. The maintenance application data includes power outage equipment, maintenance period and approval time, trip record data packet Including fault type, tripping device related information and tripping switch, through the dispatching data center to connect to the D5000 system service interface, real-time monitoring operation data and card data are obtained. The monitoring operation data includes operation time, operated equipment and operation results, and the card data includes card type, equipment ID and operation time. Through the logical rule verification of unreasonable parallelization of main transformers and mismatched switch positions, redundancy and error signals in the telemetering data are eliminated to generate standardized equipment status data. Based on the 3σ principle, outliers in the telemetry data are identified. Missing telemetry data are filled by linear interpolation of adjacent time points to generate continuous telemetry time series data. Bidirectional long short-term memory network word segmentation and conditional random field entity recognition technology are used to convert unstructured maintenance application data and trip record data into structured data associated with equipment, obtain trip record analysis results, verify the validity of operation data and card data by matching operation results with real-time status, combine support vector machine model to analyze the correlation between card type and equipment maintenance status, and generate standardized operation logs; The power grid feature extraction module extracts telesignaling features, maintenance event features and telemetering features from the preprocessed power grid data. The telesignaling features include the alarm type code and the switch opening and closing state vector. The alarm information is mapped to a numerical label. Through conditional judgment logic, the corresponding coding vector is generated according to the alarm flag bit, the alarm type code is obtained, the binary state value of the switch opening and closing state field in the telesignaling data is extracted, and the real-time switch opening and closing state vector of the equipment is generated. The maintenance event features include the maintenance window period flag and the equipment maintenance flag. Based on the maintenance application data and telesignaling data of the OMS system, it is judged whether the current time is within the period of approval of the maintenance application. Combined with the switch opening state, a Boolean maintenance window period flag is generated, 0 represents the non-maintenance period, and 1 represents the maintenance period. Based on the card placement data and maintenance window period mark of the D5000 system, the support vector machine model is used to analyze the correlation between the card placement type and the equipment status, and the equipment maintenance mark is output. The card placement types include maintenance cards and test cards. The telemetry characteristics include active power fluctuation rate and voltage over-limit frequency. Based on the pre-processed active power time series data, a time window is set, the active power mean in the window is extracted, and the average absolute deviation of the active power and its mean at each time point in the window is counted to obtain the active power fluctuation rate. Based on the pre-processed voltage amplitude time series data, the rated voltage is set. With the rated voltage as the benchmark, the over-limit range is defined. The number of times the voltage exceeds 1.1 times the rated voltage and is lower than 0.9 times the rated voltage per unit time is counted to obtain the voltage over-limit frequency. The monitoring information event analysis module combines the remote signaling features, the maintenance event features and the telemetering features, constructs a maintenance event analysis model, a power grid fault analysis model and a power grid fault clustering model, outputs a device maintenance debugging information label, a fault time label, an associated device list and an associated fault list, takes a switch split and combination state vector and a maintenance window period flag as input features, adopts a decision tree model, constructs a maintenance event analysis model based on a judgment of whether the current time falls within the time range of the maintenance application approval and a check of whether the switch split and combination state is split, the decision tree root node preferentially splits the maintenance window period flag, filters out the devices in the maintenance period, the child nodes split the switch split and combination state, if the conditions of being within the time range of the maintenance application approval and the switch split are simultaneously satisfied, it is determined that the power grid device is in a maintenance debugging state, 0 represents a non-maintenance debugging state, and 1 represents a maintenance debugging state, and finally the leaf node outputs the device maintenance debugging information label, takes an alarm type code, an active power fluctuation rate and a trip record analysis result as input features, establishes a fault mode rule library based on historical fault cases, compares the input features with the conditions in the fault mode rule library one by one, if the alarm type code, the active power fluctuation rate and the trip device association information satisfy the rule conditions, corresponding fault labels are triggered, the alarm type code, the active power fluctuation rate and the trip device primary key are converted into an input feature vector, a support vector classification algorithm is adopted, an optimal classification hyperplane is found based on the interval maximization principle, the fault category label is divided according to the projection position of the input feature vector on the hyperplane, the fault influence path is determined in combination with the power grid model topological relationship, the device primary key and the topological associated device in the trip record are extracted to form the associated device list, the fault start timestamp is marked according to the alarm occurrence time and the fault mode rule matching result, the fault time label and the associated device list containing the fault source device and the topological associated device within a single fault event are output, takes the fault start timestamp, the trip device primary key and the fault category label as input features, defines the time range of the fault association, considers two fault events with a time stamp interval not more than the length of the time window as associated events within the same time window, checks whether the trip device primary keys are the same, if the same, the same device fault cluster is formed, if different, the same device fault cluster is not formed, according to the power grid topological model, if the fault devices belong to the same interval and the same transformer substation, they are classified into the same cluster, through the topological connection relationship, the cross-station device fault is included in the same fault cluster, a decision tree clustering algorithm is adopted, the decision tree root node takes the time window as the first split condition, filters out the fault events within the same time window, the child nodes split according to the trip device primary key, different source device faults are distinguished, the topological hierarchy is refined layer by layer, a fault sub-cluster is generated, a power grid fault clustering model is constructed, each leaf node represents a fault cluster, and an associated fault list and a fault cluster label are output, the associated fault list is a multiple fault event list containing the same time window, the same device and the same topological association within the same time window; The power grid graph construction module constructs a power grid knowledge graph based on the device maintenance and debugging information label, the fault time label, the associated device list and the associated fault list, outputs the fault propagation path and the voltage overrun probability threshold, and performs hierarchical early warning. The power grid knowledge graph is constructed with the device maintenance and debugging information label, the fault time label, the associated device list, the associated fault list and the voltage overrun frequency as inputs. The power grid knowledge graph takes the tripping device primary key as the core, constructs the device node attribute, and the device node attribute includes the device maintenance and debugging information label and the voltage overrun frequency. Based on the event sequence in the associated fault list, the fault propagation relationship between devices is established. According to the historical fault data, the initial expert rule weight is set. If the device is in the maintenance period, the associated path weight is multiplied by the attenuation coefficient. The voltage overrun frequency threshold is set. If the voltage overrun frequency exceeds the corresponding threshold, the related path weight increases, and the propagation path of the cross-station device due to the connection of the power grid topology has a weight increase of 20%. The multiple fault events in the associated fault list are sorted by timestamp to generate the fault propagation path. Based on the sum of the weights of each node in the path, the path trigger probability is obtained. The expert experience coefficient is introduced. According to the voltage overrun frequency and the path trigger probability, the voltage overrun probability threshold is dynamically adjusted. If the real-time voltage exceeds the voltage overrun probability threshold, hierarchical early warning is performed. The power grid graph construction module receives the fault propagation path, the voltage overrun probability threshold and the real-time voltage data, and divides the emergency alarm, the important alarm and the general alarm levels according to the path trigger probability. The corresponding path trigger probability threshold range is matched for different alarm levels. Based on the voltage overrun frequency and the path weight, the voltage overrun probability threshold is dynamically adjusted. If the real-time voltage exceeds the voltage overrun probability threshold, the corresponding level alarm is triggered. The cross-station device fault affects the substation through the power grid topology, which increases the risk of cascading failure. If the path trigger probability corresponds to the emergency alarm threshold range, the real-time voltage exceeds the voltage overrun probability threshold, and the fault propagation path contains the cross-station device, it is determined as an emergency alarm. If the path trigger probability is in the emergency alarm and important alarm threshold range and exceeds the voltage overrun probability threshold, one of the single device overload and the real-time voltage exceeding the voltage overrun probability threshold is met, it is determined as an important alarm. If the path trigger probability corresponds to the general alarm threshold range and the real-time voltage does not exceed the voltage overrun probability threshold, it is determined as a general alarm. The power grid event intelligent research and judgment module analyzes the power grid data in real time, generates a treatment scheme, and pushes the abnormal data and the power grid fault. According to the fault propagation path and the power grid topology model, the device association relationship and the expert rule weight in the power grid knowledge graph are used to automatically calculate the loss load , generate a load shedding scheme, and the calculation process is as follows: ; Wherein, represents the active power of the i-th load point, represents the priority weight of the i-th load point, The system is based on the total number of load points, combined with a wiring diagram simulation drill, analyzes the risk points of the load shedding plan, dynamically optimizes the strategy, and outputs auxiliary decisions that take into account both safety and efficiency through the integration of expert rules and machine learning. It supports dispatchers to respond quickly, and dynamically triggers graded alarms based on the alarm level and voltage over-limit probability threshold. Abnormal data is classified by voltage level and over-limit frequency, and high-priority abnormalities are pushed preferentially to user terminals within the corresponding dispatching jurisdiction and the fault impact range. Fault information is divided into areas according to dispatching authority and pushed to mobile terminals and on-duty terminals. The fault information includes fault location, related fault list and load shedding plan, and low-latency transmission is achieved through message queues, so that alarms can be delivered in real time.

[0022] Firstly, the multi-source data is accessed in real time through the power grid data acquisition module, the module obtains the remote signaling data and remote measurement data of 10kV-500kV voltage levels from the dispatching data center, and synchronously interfaces the maintenance application of the OMS system, the trip record and the monitoring operation of the D5000 system, and the data of the placard, the redundant signals of the remote signaling are removed by using the logic rules of unreasonable parallel of the main transformer and non-corresponding switch position, the abnormal values of the remote measurement are cleaned based on the 3σ principle, the missing data is filled, and the standardized time series data set is generated, then the power grid feature extraction module performs structured processing on the preprocessed data, for the remote signaling data, the alarm type is mapped to a numerical label, and the binary vector of the switch opening and closing state is extracted, for the unstructured maintenance application of the OMS system, the bidirectional long short-term memory network word segmentation and conditional random field entity recognition technology are adopted to extract the equipment entity and the maintenance window period mark, for the remote measurement data, the active power fluctuation rate and the voltage out-of-limit frequency are calculated through the sliding window to form a dynamic feature matrix, then the monitoring information event analysis module constructs an intelligent analysis model, determines the equipment maintenance state through the decision tree model, outputs the maintenance debugging information label, classifies the alarm type, active fluctuation rate and trip record through the support vector machine classification algorithm and historical fault rule library, outputs the fault time label and associated equipment list, and uses the decision tree clustering algorithm to cluster the fault events according to the 5-minute time window and the topological level to generate a multiple fault list containing cross-station association, subsequently, the power grid graph construction module dynamically generates a knowledge graph, integrates the maintenance label, fault label and clustering result with the main key of the tripped equipment as the core node to construct the power grid knowledge graph, dynamically attenuates the path weight according to the equipment maintenance state, and enhances the cross-station fault propagation path weight combined with the voltage out-of-limit frequency to generate a fault propagation path with weight, at the same time, the voltage out-of-limit threshold is dynamically adjusted based on the path triggering probability and expert coefficient to trigger a three-level alarm, finally, the power grid event intelligent research and judgment module realizes real-time response and push, generates a load shedding scheme automatically according to the fault propagation path and out-of-limit threshold output by the knowledge graph, and verifies the disposal risk through the primary wiring diagram simulation exercise, the abnormal data is classified and pushed to the D5000 system according to the voltage level and out-of-limit frequency, and the fault information is pushed to the mobile terminal according to the dispatching authority, the emergency alarm reaches within 5 seconds, the important alarm is pushed to the regional person in charge, and the general alarm is recorded in the log.

[0023] The above merely describes specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which shall be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A power grid operation status analysis system combined with knowledge graph, characterized by: It includes power grid data acquisition module, power grid feature extraction module, monitoring information event analysis module, power grid map construction module and power grid event intelligent analysis module; The power grid data acquisition module collects and pre-processes power grid data, wherein the power grid data includes telesignaling data, telemetry data, OMS system data and D5000 system data; The power grid feature extraction module extracts telesignaling features, maintenance event features and telemetry features from the preprocessed power grid data; The monitoring information event analysis module combines remote signaling features, maintenance event features, and telemetry features to build a maintenance event analysis model, a power grid fault analysis model, and a power grid fault clustering model, and outputs equipment maintenance and debugging information tags, fault time tags, a list of associated devices, and a list of associated faults; The power grid map construction module constructs a power grid knowledge map based on equipment maintenance and debugging information tags, fault time tags, associated equipment lists, and associated fault lists, outputs fault propagation paths and voltage over-limit probability thresholds, and performs graded warnings. The power grid event intelligent analysis module analyzes power grid data in real time, generates disposal plans, and pushes abnormal data and power grid faults.

2. The power grid operation status analysis system combined with knowledge graph according to claim 1 is characterized by: In the power grid data acquisition module, the power grid data acquisition and preprocessing process includes: Receive telesignaling data from power grid equipment of all voltage levels, from 10kV to 500kV, through the real-time service interface of the dispatching data center. The telesignaling data includes alarm information on accidents, anomalies, displacements, and over-limits. Specific fields include the equipment key, occurrence time, voltage level, alarm flag, and switch open / close status. Regularly obtain telemetry data of power grid equipment through the minute-by-minute measurement service interface of the scheduling data center. The telemetry data includes measurement parameters such as active power, reactive power, voltage amplitude, current value, and oil temperature. OMS system data includes maintenance application data and trip record data, and D5000 system data includes monitoring operation data and card placement data; By connecting the dispatch data center to the OMS system service interface, maintenance application data and trip record data are obtained in real time. The maintenance application data includes the power outage equipment, maintenance period and approval time, and the trip record data includes the fault type, trip device association information and trip switch; By connecting the dispatch data center to the D5000 system service interface, real-time monitoring operation data and card placement data are obtained. The monitoring operation data includes operation time, operated device and operation result, and the card placement data includes card placement type, device ID and operation time. By verifying the logic rules of unreasonable parallel connection of main transformers and inappropriate switch positions, redundancy and error signals in remote signaling data are eliminated to generate standardized equipment status data. Identify outliers in telemetry data based on the 3σ principle, and fill in missing telemetry data by linear interpolation of adjacent time points to generate continuous telemetry time series data. Using bidirectional long short-term memory network word segmentation and conditional random field entity recognition technology, unstructured maintenance application data and trip record data are converted into device-related structured data to obtain trip record analysis results; The validity of operation data and card placement data is verified by matching operation results with real-time status. The relationship between card placement type and equipment maintenance status is analyzed using a support vector machine model to generate a standardized operation log.

3. The power grid operation status analysis system combined with knowledge graph according to claim 2 is characterized by: In the power grid feature extraction module, the process of extracting telesignaling features, maintenance event features, and telemetry features includes: Telesignaling features include alarm type codes and switch opening and closing state vectors. Alarm information is mapped to numerical labels. Conditional judgment logic is used to generate corresponding code vectors based on the alarm flag bits. The alarm type code is obtained, and the binary state value of the switch opening and closing state field in the telesignaling data is extracted to generate the real-time switch opening and closing state vector of the device. Maintenance event features include a maintenance window period flag and an equipment maintenance flag. Based on the maintenance application data and telesignaling data from the OMS system, it is determined whether the current time is within the approved maintenance application period. Combined with the switch opening status, a Boolean maintenance window period flag is generated, with 0 indicating a non-maintenance period and 1 indicating a maintenance period. Based on the card placement data and maintenance window period flag from the D5000 system, a support vector machine model is used to analyze the relationship between the card placement type and the equipment status, and the equipment maintenance flag is output. The card placement types include maintenance cards and test cards. Telemetry features include active power fluctuation rate and voltage over-limit frequency. Based on the preprocessed active power time series data, a time window is set, the mean active power in the window is extracted, and the average absolute deviation of the active power and its mean at each time point in the window is calculated to obtain the active power fluctuation rate. Based on the preprocessed voltage amplitude time series data, the rated voltage is set. With the rated voltage as the benchmark, the over-limit range is defined. The number of times the voltage exceeds 1.1 times the rated voltage and is lower than 0.9 times the rated voltage per unit time is counted to obtain the voltage over-limit frequency.

4. The power grid operation status analysis system combined with knowledge graph according to claim 3 is characterized by: In the monitoring information event analysis module, the process of constructing a maintenance event analysis model and outputting equipment maintenance and debugging information tags includes: Using the switch state vector and the maintenance window flag as input features, a decision tree model is used to build a maintenance event analysis model based on determining whether the current time falls within the approved maintenance application period and checking whether the switch state is open. The root node of the decision tree prioritizes splitting the maintenance window period flag to filter out the equipment in the maintenance period. The child node splits the switch opening and closing status. If the switch opening condition is met within the approved maintenance application period, the power grid equipment is determined to be in the maintenance and debugging state. 0 represents the non-maintenance and debugging state, and 1 represents the maintenance and debugging state. Finally, the leaf node outputs the equipment maintenance and debugging information label.

5. The power grid operation status analysis system combined with knowledge graph according to claim 4 is characterized by: In the monitoring information event analysis module, the process of constructing a power grid fault analysis model and outputting a fault time label and a list of associated devices includes: Using the alarm type code, active power fluctuation rate, and trip record analysis results as input features, a fault mode rule base is established based on historical fault cases. The input features are compared one by one with the conditions in the fault mode rule base. If the alarm type code, active power fluctuation rate, and trip device association information meet the rule conditions, the corresponding fault tag is triggered. The alarm type code, active power fluctuation rate, and trip device primary key are converted into input feature vectors. The support vector classification algorithm is used to find the optimal classification hyperplane based on the principle of maximum margin. The fault category labels are divided according to the projection position of the input feature vector on the hyperplane. Combined with the topological relationship of the power grid model, the fault impact path is determined, and the device primary key and topologically associated devices in the trip record are extracted to form a list of associated devices; According to the alarm occurrence time and the fault mode rule matching result, the fault start timestamp is marked, and the fault time label and the associated device list including the fault source device and the topologically associated devices in the single fault event are output.

6. The power grid operation status analysis system combined with knowledge graph according to claim 5 is characterized by: In the monitoring information event analysis module, the process of building a power grid fault clustering model and outputting a list of related faults includes: The fault start timestamp, trip device primary key, and fault category label are used as input features to define the time range of fault association. Two fault events with timestamps that do not exceed the length of the time window are considered to be associated events within the same time window. Check whether the primary keys of the tripped devices are the same. If they are the same, they are grouped into the same device fault cluster. If they are different, they are not grouped into the same device fault cluster. According to the power grid topology model, if the faulty devices belong to the same bay and the same substation, they are grouped into the same cluster. Based on the topological connection relationship, cross-station device faults are grouped into the same fault cluster. A decision tree clustering algorithm is adopted. The root node of the decision tree uses the time window as the first splitting condition to screen out fault events within the same time window. The child nodes are split according to the primary key of the tripping device to distinguish faults of different source devices. The fault subclusters are generated layer by layer according to the topological level to construct a power grid fault clustering model. Each leaf node represents a fault cluster, and an associated fault list and fault cluster label are output. The associated fault list is a list of multiple fault events within the same time window, the same device and the same topological association.

7. The power grid operation status analysis system combined with knowledge graph according to claim 6 is characterized by: In the power grid map construction module, the process of constructing a power grid knowledge map and outputting fault propagation paths and voltage over-limit probability thresholds includes: Build a power grid knowledge graph using equipment maintenance and debugging information tags, fault time tags, associated equipment lists, associated fault lists, and voltage over-limit frequencies as input; The power grid knowledge graph uses the primary key of the tripping device as the core to construct device node attributes, including device maintenance and debugging information tags and voltage over-limit frequency. Based on the event sequence in the associated fault list, the fault propagation relationship between devices is established, and the initial expert rule weights are set based on historical fault data. If the equipment is under maintenance, the weight of its associated path is multiplied by the attenuation coefficient. The voltage over-limit frequency threshold is set. If the voltage over-limit frequency exceeds the corresponding threshold, the weight of the relevant path increases, and the weight of the propagation path generated by the cross-site equipment due to the grid topology connection is increased by 20%; The multiple fault events in the associated fault list are sorted by timestamp to generate a fault propagation path. The path trigger probability is obtained based on the sum of the weights of each node in the path. The expert experience coefficient is introduced to dynamically adjust the voltage over-limit probability threshold based on the voltage over-limit frequency and the path trigger probability. If the real-time voltage exceeds the voltage over-limit probability threshold, a graded warning is issued.

8. The power grid operation status analysis system combined with knowledge graph according to claim 7 is characterized by: In the power grid map construction module, the process of performing hierarchical warning includes: The power grid map construction module receives fault propagation paths, voltage over-limit probability thresholds, and real-time voltage data. It divides the alarm levels into emergency, major, and general alarms based on the path trigger probability. It matches the corresponding path trigger probability threshold ranges to different alarm levels. Based on the voltage over-limit frequency and path weight, it dynamically adjusts the voltage over-limit probability threshold. If the real-time voltage exceeds the voltage over-limit probability threshold, an alarm of the corresponding level is triggered. Cross-site equipment failures affect substations through the power grid topology, increasing the risk of cascading failures. If the path trigger probability corresponds to the emergency alarm threshold range, the real-time voltage exceeds the voltage over-limit probability threshold, and the fault propagation path includes cross-site equipment, then it is determined to be an emergency alarm. If one of the following conditions is met: the path trigger probability is within the emergency alarm and major alarm threshold range and exceeds the voltage over-limit probability threshold, or a single device is overloaded and the real-time voltage exceeds the voltage over-limit probability threshold, then it is determined to be a major alarm. If the path trigger probability corresponds to the general alarm threshold range and the real-time voltage does not exceed the voltage over-limit probability threshold, then it is determined to be a general alarm.

9. The power grid operation status analysis system combined with knowledge graph according to claim 8, characterized in that: In the power grid event intelligent analysis module, the process of real-time analysis of power grid data and generation of treatment plans includes: Based on the fault propagation path and grid topology model, the loss load is automatically calculated using the equipment association relationship and expert rule weights in the grid knowledge graph. , generate load shedding plan; Combined with a wiring diagram simulation drill, the risk points of the load shedding plan are analyzed, and the strategy is dynamically optimized. By integrating expert rules and machine learning, auxiliary decision-making that takes into account both safety and efficiency is output to support dispatchers to respond quickly.

10. The power grid operation status analysis system combined with knowledge graph according to claim 9, characterized in that: In the power grid event intelligent analysis module, the process of pushing abnormal data and power grid faults includes: Based on the alarm level and voltage over-limit probability threshold, hierarchical alarms are triggered dynamically. Abnormal data is classified by voltage level and over-limit frequency. High-priority abnormalities are pushed first to user terminals within the corresponding dispatching jurisdiction and fault impact range. Fault information is divided into areas according to dispatching authority and pushed to mobile terminals and on-duty terminals. The fault information includes fault location, related fault list and load shedding plan, and low-latency transmission is achieved through message queues, so that alarms can be delivered in real time.

Images