Cross-region synchronous power monitoring system and method based on time series database

Abstract

The application discloses a cross-region synchronous power monitoring system and method based on a time series database, and relates to the technical field of data analysis.The method comprises the following steps: acquiring historical abnormal events and storing them into a time series database, extracting first and second event features therefrom, calculating a perception response duration and a recovery treatment duration, and taking the two as third event features; step-by-step construction of corresponding event sequence models to obtain a complete event sequence model set of all abnormal events; screening of event pairs by traversing the set, analysis of causality based on the event pairs, determination of a causal relationship, and generation of a causal link pair set; traversal of the causal link pair set, marking of associated causal chain pairs and grouping; establishment of a mapping and storage into a knowledge base, marking of consequence types and average recovery durations of each situation element; real-time monitoring of situation element abnormalities, query of the knowledge base to find common consequences, calculation of an estimated maximum recovery time, and early warning if the estimated maximum recovery time exceeds a threshold value.The application realizes cross-region synchronous power monitoring.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to a cross-regional synchronous power monitoring system and method based on a time-series database. Background Technology

[0002] As the power system develops towards cross-regional interconnection and large-scale new energy grid connection, cross-regional power dispatching models such as "West-to-East Power Transmission" and "North-South Interconnection" are becoming increasingly mature. The safe and stable operation of cross-regional power grids has become a core requirement for ensuring energy supply. As a key link in power grid operation and maintenance, power monitoring needs to process massive amounts of timestamped operational data in real time. Its data synchronization, accuracy of abnormal event analysis, and cross-regional correlation early warning capabilities affect the efficiency of power grid fault handling and overall power supply reliability.

[0003] However, traditional cross-regional power monitoring methods often face the following problems when dealing with the integration of multi-source time-series data, comprehensive analysis of abnormal events, and cross-regional causal correlation early warning: First, the integration and synchronization of multi-source data is difficult. The monitoring data sources of cross-regional power grids are scattered and have heterogeneous formats. Traditional monitoring often adopts a single-region independent storage mode and lacks a unified storage and cross-regional data synchronization mechanism based on time-series databases, resulting in inconsistent data timestamps and making it difficult to support unified analysis across regions. Second, the cross-regional causal correlation and early warning capabilities are weak. Traditional monitoring is mostly limited to event analysis within a single region and does not consider the electrical topology correlation of cross-regional power grids. It is unable to screen cross-regional event pairs that are close in time and spatially related to determine causal relationships. It also lacks situation element mapping and recovery time prediction based on historical event patterns. Early warning relies only on a single abnormal indicator trigger, which is prone to false alarms, missed alarms, or delayed early warning timing, making it difficult to meet the operation and maintenance needs of cross-regional power grids for "early detection, early handling, and accurate recovery". Summary of the Invention

[0004] The purpose of this invention is to provide a cross-regional synchronous power monitoring system and method based on a time-series database to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a cross-regional synchronous power monitoring method based on a time-series database, the method comprising the following steps:

[0006] Historical abnormal events are acquired and stored in a time-series database. First and second event features are extracted from these databases. The sensing response time and recovery handling time are calculated based on timestamps and used as the third event feature. The first event feature represents the inherent attributes of the abnormal event, including the fault source, event type, and the electrical connection area or equipment directly associated with the event. The second event feature represents the first observable impact on the power grid monitoring system before any corrective operations are implemented, based on the operational indicator status, including electrical quantity indicators, equipment status indicators, and system business indicators.

[0007] Based on the characteristics of the first, second and third events corresponding to each historical abnormal event, the corresponding event sequence model is constructed step by step to obtain a complete set of event sequence models for all abnormal events.

[0008] Traverse the complete event sequence model set, and for each event sequence model that is the main event, perform time proximity filtering and spatial correlation filtering to filter out event pairs composed of the main event and candidate events. Based on this, perform causal orientation analysis to determine causal relationships, and gather all the determined causal relationships in the form of link pairs with the main event as the cause and the candidate event as the effect to generate a causal link pair set.

[0009] Based on the causal link pair set, the associated causal link pairs with the same event type and electrical correlation in the scope of influence are marked and grouped; the situation element set and deviation state index set are extracted from the group, a mapping is established and stored in the knowledge base, and the consequence type and average recovery time of each situation element are marked; the anomaly of situation elements is monitored in real time, the knowledge base is queried to find common consequences, the estimated longest recovery time is calculated, and an early warning is issued if the time exceeds the threshold.

[0010] Historical abnormal events are acquired and stored in a time-series database. First and second event features are extracted from these databases. The perception response time and recovery handling time are calculated based on timestamps and used as the third event feature. Specific steps include:

[0011] Establish connections with the data sources of each monitored area to obtain historical abnormal events of the power monitoring equipment in each monitored area, and store them based on a time-series database. The data sources of the monitored areas include SCADA systems, protection information management systems, fault recorders, and sequence event recorders (SOEs).

[0012] The first and second event features of each historical anomaly are extracted from the time-series database. The first event feature represents the inherent attributes of the anomaly, including the fault source, event type, and the directly associated electrical connection area or equipment at the time of the event. The second event feature represents the first observable impact on the power grid monitoring system before any corrective operations are implemented, based on operational indicators, including electrical quantity indicators, equipment status indicators, and system business indicators. Electrical quantity indicators include power or current anomalies of relevant power equipment and voltage anomalies of nodes, where nodes represent common connection points such as busbars in the power grid. Equipment status indicators include switch changes, protection device action signals, and equipment alarm signals. System business indicators include SCADA or EMS system data refresh status and command transmission timeliness.

[0013] Based on timestamps, the temporal performance characteristics of each abnormal event are calculated, including the perception response time and recovery handling time. The specific calculation steps are as follows:

[0014] The duration from the start time of the initial event of the fault source to the cutoff time of capturing and presenting the corresponding second event feature is denoted as the perception response duration.

[0015] The recovery processing time is defined as the time from the moment when the corresponding second event characteristic appears until the moment when the electrical quantity indicators, equipment status indicators, and system business indicators are all restored to the preset safe and stable threshold range through automatic or manual operation.

[0016] The calculated perception response time and recovery processing time are used as the third event feature for each abnormal event.

[0017] Based on the first, second, and third event characteristics corresponding to each historical anomaly, a corresponding event sequence model is constructed step by step to obtain a complete set of event sequence models for all anomalies. The specific steps include:

[0018] For each historical anomaly, its first event feature and second event feature are connected by a directed edge to generate the basic event sequence model of the event, wherein the direction of the directed edge is from the first event feature to the second event feature;

[0019] Based on the generated event sequence model, the second event feature and the third event feature are connected by directed edges to generate a complete event sequence model, wherein the direction of the directed edges is from the second event feature to the third event feature;

[0020] Each generated complete event sequence model is added to the event sequence model set to form a complete event sequence model set covering all historical abnormal events.

[0021] The complete event sequence model set is traversed. For each event sequence model that serves as the main event, temporal proximity filtering and spatial correlation filtering are performed to filter out event pairs consisting of the main event and candidate events. Based on this, causal orientation analysis is performed to determine causal relationships. All determined causal relationships are aggregated in the form of link pairs with the main event as the cause and the candidate event as the effect, generating a causal link pair set. The specific steps include:

[0022] Traverse the complete set of event sequence models. For each event sequence model, denoted as the main event M, perform time proximity filtering and spatial correlation filtering. The specific operations are as follows:

[0023] Proximity filtering: In the time series database, find all other event sequence models whose start timestamp differs from the main event M by a preset time window, and denot them as candidate events N;

[0024] Spatial correlation filtering: From candidate events N, further filter out events whose influence range within the first event feature overlaps with or is electrically connected to the influence range of the main event M;

[0025] For the event pairs consisting of the selected main event M and candidate events N, causal orientation analysis is performed. The specific steps are as follows:

[0026] Perform a timing determination by comparing the start timestamps of the main event M and the candidate event N. If the start time of the main event M is earlier than the start time of the candidate event N, then it is considered that there is a timing premise from the main event M to the candidate event N.

[0027] Perform logical judgment and analyze the characteristics of the second event of the main event M, that is, the impact of the main event M, to determine whether it provides a direct cause or necessary condition for the occurrence of the candidate event N.

[0028] If both the temporal premise and logical possibility are satisfied, then it is determined that there is a causal relationship from the main event M to the candidate event N;

[0029] The causal relationship obtained for each event pair will be represented as a link pair (M, N), where M is the cause and N is the effect.

[0030] All such causal link pairs are aggregated to generate a causal link pair set.

[0031] Based on the causal link pair set, related causal chain pairs with the same event type and electrically related impact range are marked and grouped; from the group, a set of situation element and deviation state index set are extracted, a mapping is established and stored in the knowledge base, and the consequence type and average recovery time of each situation element are marked; anomalies of situation elements are monitored in real time, the knowledge base is queried to find common consequences, the estimated longest recovery time is calculated, and an early warning is issued if the time exceeds the threshold. The specific steps include:

[0032] Iterate through the set of causal link pairs. For each causal link pair CP1 and CP2, the causal event is M1 and the effect event is N1 in causal link pair CP1; and the causal event is M2 and the effect event is N2 in causal link pair CP2.

[0033] If causal events M1 and M2 have the same event type and their scope of influence has an electrical or topological connection, then causal link pair CP1 and causal link pair CP2 are marked as associated causal link pairs.

[0034] Group all causal link pairs that have the same event type and electrical association to form associated causal chain groups;

[0035] For each causal chain group, analyze all causal events within the group, combine historical operating data, extract their common preconditions, including power grid structure characteristics, operating state characteristics before the event, and environmental condition characteristics, to form a situation element set F; extract the common state index anomalies in the second event characteristics of all result events within the group to form a deviation state index set U.

[0036] Establish a mapping relationship between the situation element set F and the deviation state index set U;

[0037] Repeat the above steps to process all related causal chain pairs, generate mapping pairs, and store them in the situation mapping knowledge base;

[0038] Based on the generated situation mapping knowledge base, each situation element f is labeled with all the types of consequences it has caused in history, and the average recovery and handling time corresponding to each situation element f is recorded.

[0039] The status of each situation element during network operation is captured and monitored in real time. When one or more situation elements are detected to be in an abnormal state: query the situation mapping knowledge base to find the common consequence event types caused by all activated situation elements.

[0040] The maximum historical average recovery time corresponding to all activated status elements is taken as the estimated maximum recovery time. When the estimated maximum recovery time exceeds the preset acceptable interruption time threshold, an early warning is immediately issued to the relevant area.

[0041] After anomaly handling, the data is sent back to the time-series database.

[0042] The cross-regional synchronous power monitoring system based on a time-series database includes: a historical abnormal event processing module, an event sequence model construction module, a causal analysis module, a related causal chain grouping module, and an early warning module. The historical abnormal event processing module acquires historical abnormal events and stores them in the time-series database, extracts first and second event features, calculates the sensing response time and recovery handling time, and uses these as the third event feature. The event sequence model construction module constructs corresponding event sequence models step-by-step to obtain a complete set of event sequence models for all abnormal events. The causal analysis module iterates through this set to filter out event pairs, performs causal orientation analysis to determine causal relationships, and generates a set of causal link pairs. The related causal chain grouping module iterates through the set of causal link pairs, marks related causal chain pairs and groups them, establishes mappings and stores them in a knowledge base, and marks the consequence type and average recovery time of each situation element. The early warning module monitors abnormal situation elements in real time, queries the knowledge base to find common consequences, calculates the estimated longest recovery time, and issues an early warning if the time exceeds a threshold.

[0043] The historical abnormal event processing module includes a multi-source data access unit, a time-series data storage unit, an event feature extraction unit, and a time-series performance calculation unit. The multi-source data access unit is used to establish stable connections with data sources in each monitored area to acquire historical abnormal events of the power monitoring equipment. The time-series data storage unit is used to sort the acquired historical abnormal events by timestamp and store them in a time-series database. The event feature extraction unit is used to extract the first and second event features of each historical abnormal event from the time-series database. The time-series performance calculation unit is used to calculate the third event feature based on the timestamp.

[0044] The event sequence model construction module includes a basic event sequence model generation unit, a complete event sequence model generation unit, and an event sequence model set management unit. The basic event sequence model generation unit generates a basic model for each historical abnormal event by connecting the first and second event features with directed edges. The complete event sequence model generation unit completes the event's full lifecycle logic by connecting the second and third event features with directed edges based on the basic model. The event sequence model set management unit aggregates all complete event sequence models to form a complete event sequence model set.

[0045] The causal analysis module includes a model set traversal unit, a temporal proximity filtering unit, a spatial correlation filtering unit, a causal orientation analysis unit, and a causal link set generation unit. The model set traversal unit is used to access each model in the complete event sequence model set one by one and define it as the main event. The temporal proximity filtering unit is used to filter candidate events. The spatial correlation filtering unit is used to filter events from the candidate events whose influence range intersects with or is electrically connected to the main event. The causal orientation analysis unit is used to verify the causal relationship of event pairs composed of the main event and candidate events through temporal and logical judgment. The causal link set generation unit is used to summarize the event pairs that are determined to have a causal relationship and generate a causal link set.

[0046] The associated causal chain grouping module includes an associated causal chain pair marking unit, an extraction unit, a mapping establishment unit, a recovery time marking unit, and an early warning unit. The associated causal chain pair marking unit is used to mark link pairs that are electrically related due to the same event type and the scope of influence, and group them into the same associated causal chain group. The extraction unit is used to extract the situation element set and the deviation state index set for each associated causal chain group. The mapping establishment unit is used to establish mapping relationships and store the established mapping relationships in the knowledge base. The recovery time marking unit is used to mark the type of consequence event that each situation element in the knowledge base has caused in the past, and calculate and store the corresponding average recovery and handling time.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] 1. By traversing the event sequence model set, performing time proximity filtering and spatial correlation filtering, and then constructing a causal link set through temporal and logical judgment, the invention extracts situational elements and deviation indicators based on the associated causal chain group to establish a knowledge base. Combined with historical data to mark the consequence type and average recovery time, the invention realizes cross-regional event causal correlation analysis and recovery time prediction based on historical patterns. Unlike the existing technology that is limited to single-region analysis and relies on a single abnormal indicator to trigger early warning, this invention considers the electrical correlation characteristics of cross-regional power grids, avoiding the problems of false alarms, missed alarms or timing delays in traditional early warning.

[0049] 2. This invention constructs a unified storage and cross-regional synchronization mechanism for multi-source data based on a time-series database, and achieves cross-regional data alignment by timestamp. Unlike the existing technology of independent storage in a single region and fragmented data, this invention can solve the problems of inconsistent data timestamps and fragmented key information in traditional cross-regional monitoring, and provide complete and synchronous data source support for unified anomaly analysis of cross-regional power grids. Attached Figure Description

[0050] Figure 1 This is a flowchart illustrating the cross-regional synchronous power monitoring method based on a time-series database according to the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] like Figure 1 As shown, this invention provides a technical solution: a cross-regional synchronous power monitoring method based on a time-series database, which includes the following steps:

[0053] Historical abnormal events are acquired and stored in a time-series database. First and second event features are extracted from these databases. The sensing response time and recovery handling time are calculated based on timestamps and used as the third event feature. The first event feature represents the inherent attributes of the abnormal event, including the fault source, event type, and the electrical connection area or equipment directly associated with the event. The second event feature represents the first observable impact on the power grid monitoring system before any corrective operations are implemented, based on the operational indicator status, including electrical quantity indicators, equipment status indicators, and system business indicators.

[0054] Based on the characteristics of the first, second and third events corresponding to each historical abnormal event, the corresponding event sequence model is constructed step by step to obtain a complete set of event sequence models for all abnormal events.

[0055] Traverse the complete event sequence model set, and for each event sequence model that is the main event, perform time proximity filtering and spatial correlation filtering to filter out event pairs composed of the main event and candidate events. Based on this, perform causal orientation analysis to determine causal relationships, and gather all the determined causal relationships in the form of link pairs with the main event as the cause and the candidate event as the effect to generate a causal link pair set.

[0056] Based on the causal link pair set, the associated causal link pairs with the same event type and electrical correlation in the scope of influence are marked and grouped; the situation element set and deviation state index set are extracted from the group, a mapping is established and stored in the knowledge base, and the consequence type and average recovery time of each situation element are marked; the anomaly of situation elements is monitored in real time, the knowledge base is queried to find common consequences, the estimated longest recovery time is calculated, and an early warning is issued if the time exceeds the threshold.

[0057] Historical abnormal events are acquired and stored in a time-series database. First and second event features are extracted from these databases. The perception response time and recovery handling time are calculated based on timestamps and used as the third event feature. Specific steps include:

[0058] Establish connections with the data sources of each monitored area to obtain historical abnormal events of the power monitoring equipment in each monitored area, and store them based on a time-series database. The data sources of the monitored areas include SCADA systems, protection information management systems, fault recorders, and sequence event recorders (SOEs).

[0059] The first and second event features of each historical anomaly are extracted from the time-series database. The first event feature represents the inherent attributes of the anomaly, including the fault source, event type, and the directly associated electrical connection area or equipment at the time of the event. The second event feature represents the first observable impact on the power grid monitoring system before any corrective operations are implemented, based on operational indicators, including electrical quantity indicators, equipment status indicators, and system business indicators. Electrical quantity indicators include power or current anomalies of relevant power equipment and voltage anomalies of nodes, where nodes represent common connection points such as busbars in the power grid. Equipment status indicators include switch changes, protection device action signals, and equipment alarm signals. System business indicators include SCADA or EMS system data refresh status and command transmission timeliness.

[0060] Based on timestamps, the temporal performance characteristics of each abnormal event are calculated, including the perception response time and recovery handling time. The specific calculation steps are as follows:

[0061] The duration from the start time of the initial event of the fault source to the cutoff time of capturing and presenting the corresponding second event feature is denoted as the perception response duration.

[0062] The recovery processing time is defined as the time from the moment when the corresponding second event characteristic appears until the moment when the electrical quantity indicators, equipment status indicators, and system business indicators are all restored to the preset safe and stable threshold range through automatic or manual operation.

[0063] The calculated perception response time and recovery processing time are used as the third event feature for each abnormal event.

[0064] Based on the first, second, and third event characteristics corresponding to each historical anomaly, a corresponding event sequence model is constructed step by step to obtain a complete set of event sequence models for all anomalies. The specific steps include:

[0065] For each historical anomaly, its first event feature and second event feature are connected by a directed edge to generate the basic event sequence model of the event, wherein the direction of the directed edge is from the first event feature to the second event feature;

[0066] Based on the generated event sequence model, the second event feature and the third event feature are connected by directed edges to generate a complete event sequence model, wherein the direction of the directed edges is from the second event feature to the third event feature;

[0067] Each generated complete event sequence model is added to the event sequence model set to form a complete event sequence model set covering all historical abnormal events.

[0068] The complete event sequence model set is traversed. For each event sequence model that serves as the main event, temporal proximity filtering and spatial correlation filtering are performed to filter out event pairs consisting of the main event and candidate events. Based on this, causal orientation analysis is performed to determine causal relationships. All determined causal relationships are aggregated in the form of link pairs with the main event as the cause and the candidate event as the effect, generating a causal link pair set. The specific steps include:

[0069] Traverse the complete set of event sequence models. For each event sequence model, denoted as the main event M, perform time proximity filtering and spatial correlation filtering. The specific operations are as follows:

[0070] Proximity filtering: In the time series database, find all other event sequence models whose start timestamp differs from the main event M by a preset time window, and denot them as candidate events N;

[0071] Spatial correlation filtering: From candidate events N, further filter out events whose influence range within the first event feature overlaps with or is electrically connected to the influence range of the main event M;

[0072] For the event pairs consisting of the selected main event M and candidate events N, causal orientation analysis is performed. The specific steps are as follows:

[0073] Perform a timing determination by comparing the start timestamps of the main event M and the candidate event N. If the start time of the main event M is earlier than the start time of the candidate event N, then it is considered that there is a timing premise from the main event M to the candidate event N.

[0074] Perform logical judgment and analyze the characteristics of the second event of the main event M, that is, the impact of the main event M, to determine whether it provides a direct cause or necessary condition for the occurrence of the candidate event N.

[0075] If both the temporal premise and logical possibility are satisfied, then it is determined that there is a causal relationship from the main event M to the candidate event N;

[0076] The causal relationship obtained for each event pair will be represented as a link pair (M, N), where M is the cause and N is the effect.

[0077] All such causal link pairs are aggregated to generate a causal link pair set.

[0078] Based on the causal link pair set, related causal chain pairs with the same event type and electrically related impact range are marked and grouped; from the group, a set of situation element and deviation state index set are extracted, a mapping is established and stored in the knowledge base, and the consequence type and average recovery time of each situation element are marked; anomalies of situation elements are monitored in real time, the knowledge base is queried to find common consequences, the estimated longest recovery time is calculated, and an early warning is issued if the time exceeds the threshold. The specific steps include:

[0079] Iterate through the set of causal link pairs. For each causal link pair CP1 and CP2, the causal event is M1 and the effect event is N1 in causal link pair CP1; and the causal event is M2 and the effect event is N2 in causal link pair CP2.

[0080] If causal events M1 and M2 have the same event type and their scope of influence has an electrical or topological connection, then causal link pair CP1 and causal link pair CP2 are marked as associated causal link pairs.

[0081] Group all causal link pairs that have the same event type and electrical association to form associated causal chain groups;

[0082] For each causal chain group, analyze all causal events within the group, combine historical operating data, extract their common preconditions, including power grid structure characteristics, operating state characteristics before the event, and environmental condition characteristics, to form a situation element set F; extract the common state index anomalies in the second event characteristics of all result events within the group to form a deviation state index set U.

[0083] Establish a mapping relationship between the situation element set F and the deviation state index set U;

[0084] Repeat the above steps to process all related causal chain pairs, generate mapping pairs, and store them in the situation mapping knowledge base;

[0085] Based on the generated situation mapping knowledge base, each situation element f is labeled with all the types of consequences it has caused in history, and the average recovery and handling time corresponding to each situation element f is recorded.

[0086] The status of each situation element during network operation is captured and monitored in real time. When one or more situation elements are detected to be in an abnormal state: query the situation mapping knowledge base to find the common consequence event types caused by all activated situation elements.

[0087] The maximum historical average recovery time corresponding to all activated status elements is taken as the estimated maximum recovery time. When the estimated maximum recovery time exceeds the preset acceptable interruption time threshold, an early warning is immediately issued to the relevant area.

[0088] In Example 1: By establishing a stable connection with the data sources of each area to be monitored, these data sources include monitoring and data acquisition responsible for real-time collection of power grid operation data, protection information management that records the action details of protection devices, fault recorders that capture electrical waveforms at the moment of a fault, and sequence event recording devices that record the order of events with high precision. Past power anomaly events are obtained from these data sources, and all events are organized according to the time dimension and stored in a time-series database to ensure that the data can be quickly retrieved in chronological order.

[0089] Features of each historical anomaly are extracted from the time-series database: The first event feature focuses on the attributes of the anomaly itself. For example, an anomaly caused by a line insulation problem has a specific transmission line as the source of the fault, the event type is line insulation fault, and the directly associated electrical area is the two substations connected by the line. The second event feature focuses on the first observable impact after the fault occurs, based on the power grid operation indicators. For example, the first impact caused by the above-mentioned line insulation fault is the voltage anomaly of the busbar of the associated substation (electrical quantity indicator), accompanied by the change signal of the line switch (equipment status indicator). The third event feature is calculated through time relationships, namely, the perception response time from the start of the fault to the capture of the second event feature, and the recovery and handling time from the presentation of the second event feature to the restoration of the power grid's various indicators to a safe and stable state.

[0090] Construct an event sequence model and form a set. For each historical abnormal event, first connect the first event feature and the second event feature with a directed edge, with the directed edge pointing from the first event feature to the second event feature, to form a basic event sequence model, reflecting the logic of "event attributes triggering initial impact". Then, on the basic model, connect the second event feature and the third event feature with a directed edge, with the directed edge pointing from the second event feature to the third event feature, to complete the relationship from "initial impact to handling effectiveness", generating a complete event sequence model. Summarize all complete models to form a model set covering all historical abnormal events.

[0091] Event pair screening and causal analysis are performed. The event sequence model set is traversed, and each model is taken as the main event. Candidate events that are close to the main event in time are first screened out. Then, events that are electrically related to the scope of influence of the main event are further screened out to form event pairs. Causal determination is performed on the event pairs: if the main event occurs earlier than the candidate event, and the second event feature of the main event is the direct cause of the candidate event, then the two are determined to have a causal relationship. Causal link pairs are formed in the form of "main event as cause and candidate event as effect". All link pairs are summarized to generate a causal link pair set.

[0092] Traverse the set of causal link pairs, mark link pairs that are electrically related due to the same event type and the scope of influence as related causal link pairs and group them. For example, multiple link pairs of "line insulation fault triggering protection action in adjacent area" are grouped into the same group. Extract common situational elements from each group, such as weather conditions and line load status before the fault occurs, as well as common deviation status indicators, such as voltage abnormality amplitude and switch change duration. Establish the mapping relationship between situational elements and deviation indicators, store them in the knowledge base, and mark the types of consequences and average recovery time of each situational element in the past.

[0093] The system monitors the status of various power grid elements in real time. When an anomaly is detected in a certain element, such as weather conditions and load status related to "line insulation fault" in the knowledge base, the system immediately queries the knowledge base to find the common consequence type corresponding to the abnormal element. At the same time, it calculates the estimated longest recovery time. If the estimated longest recovery time exceeds the preset acceptable threshold, the system immediately issues an early warning to the operation and maintenance personnel in the relevant area, reminding them to take timely measures to prevent the fault from escalating and to ensure the safe and stable operation of the cross-regional power grid.

[0094] The cross-regional synchronous power monitoring system based on a time-series database includes: a historical abnormal event processing module, an event sequence model construction module, a causal analysis module, a related causal chain grouping module, and an early warning module. The historical abnormal event processing module acquires historical abnormal events and stores them in the time-series database, extracts first and second event features, calculates the sensing response time and recovery handling time, and uses these as the third event feature. The event sequence model construction module constructs corresponding event sequence models step-by-step to obtain a complete set of event sequence models for all abnormal events. The causal analysis module iterates through this set to filter out event pairs, performs causal orientation analysis to determine causal relationships, and generates a set of causal link pairs. The related causal chain grouping module iterates through the set of causal link pairs, marks related causal chain pairs and groups them, establishes mappings and stores them in a knowledge base, and marks the consequence type and average recovery time of each situation element. The early warning module monitors abnormal situation elements in real time, queries the knowledge base to find common consequences, calculates the estimated longest recovery time, and issues an early warning if the time exceeds a threshold.

[0095] The historical abnormal event processing module includes a multi-source data access unit, a time-series data storage unit, an event feature extraction unit, and a time-series performance calculation unit. The multi-source data access unit is used to establish stable connections with data sources in each monitored area to acquire historical abnormal events of the power monitoring equipment. The time-series data storage unit is used to sort the acquired historical abnormal events by timestamp and store them in a time-series database. The event feature extraction unit is used to extract the first and second event features of each historical abnormal event from the time-series database. The time-series performance calculation unit is used to calculate the third event feature based on the timestamp.

[0096] The event sequence model construction module includes a basic event sequence model generation unit, a complete event sequence model generation unit, and an event sequence model set management unit. The basic event sequence model generation unit generates a basic model for each historical abnormal event by connecting the first and second event features with directed edges. The complete event sequence model generation unit completes the event's full lifecycle logic by connecting the second and third event features with directed edges based on the basic model. The event sequence model set management unit aggregates all complete event sequence models to form a complete event sequence model set.

[0097] The causal analysis module includes a model set traversal unit, a temporal proximity filtering unit, a spatial correlation filtering unit, a causal orientation analysis unit, and a causal link set generation unit. The model set traversal unit is used to access each model in the complete event sequence model set one by one and define it as the main event. The temporal proximity filtering unit is used to filter candidate events. The spatial correlation filtering unit is used to filter events from the candidate events whose influence range intersects with or is electrically connected to the main event. The causal orientation analysis unit is used to verify the causal relationship of event pairs composed of the main event and candidate events through temporal and logical judgment. The causal link set generation unit is used to summarize the event pairs that are determined to have a causal relationship and generate a causal link set.

[0098] The associated causal chain grouping module includes an associated causal chain pair marking unit, an extraction unit, a mapping establishment unit, a recovery time marking unit, and an early warning unit. The associated causal chain pair marking unit is used to mark link pairs that are electrically related due to the same event type and the scope of influence, and group them into the same associated causal chain group. The extraction unit is used to extract the situation element set and the deviation state index set for each associated causal chain group. The mapping establishment unit is used to establish mapping relationships and store the established mapping relationships in the knowledge base. The recovery time marking unit is used to mark the type of consequence event that each situation element in the knowledge base has caused in the past, and calculate and store the corresponding average recovery and handling time.

[0099] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A cross-regional synchronous power monitoring method based on a time-series database, characterized in that: The method includes the following steps: Historical abnormal events are acquired and stored in a time-series database. First and second event features are extracted from these databases. The sensing response time and recovery handling time are calculated based on timestamps and used as the third event feature. The first event feature represents the inherent attributes of the abnormal event, including the fault source, event type, and the electrical connection area or equipment directly associated with the event. The second event feature represents the first observable impact on the power grid monitoring system before any corrective operations are implemented, based on the operational indicator status, including electrical quantity indicators, equipment status indicators, and system business indicators. Based on the characteristics of the first, second and third events corresponding to each historical abnormal event, the corresponding event sequence model is constructed step by step to obtain a complete set of event sequence models for all abnormal events. Traverse the complete event sequence model set, and for each event sequence model that is the main event, perform time proximity filtering and spatial correlation filtering to filter out event pairs composed of the main event and candidate events. Based on this, perform causal orientation analysis to determine causal relationships, and gather all the determined causal relationships in the form of link pairs with the main event as the cause and the candidate event as the effect to generate a causal link pair set. Based on the causal link pair set, the associated causal link pairs with the same event type and electrical correlation in the scope of influence are marked and grouped; the situation element set and deviation state index set are extracted from the group, a mapping is established and stored in the knowledge base, and the consequence type and average recovery time of each situation element are marked; the anomaly of situation elements is monitored in real time, the knowledge base is queried to find common consequences, the estimated longest recovery time is calculated, and an early warning is issued if the time exceeds the threshold.

2. The cross-regional synchronous power monitoring method based on a time-series database according to claim 1, characterized in that: Historical abnormal events are acquired and stored in a time-series database. First and second event features are extracted from these databases. The perception response time and recovery handling time are calculated based on timestamps and used as the third event feature. Specific steps include: Establish connections with the data sources of each monitored area to obtain historical abnormal events of the power monitoring equipment in each monitored area, and store them based on a time-series database. The data sources of the monitored areas include SCADA systems, protection information management systems, fault recorders, and sequence event recorders (SOEs). The first and second event features of each historical anomaly are extracted from the time-series database. The first event feature represents the inherent attributes of the anomaly, including the fault source, event type, and the directly associated electrical connection area or equipment at the time of the event. The second event feature represents the first observable impact on the power grid monitoring system before any corrective operations are implemented, based on operational indicators, including electrical quantity indicators, equipment status indicators, and system business indicators. Electrical quantity indicators include power or current anomalies of relevant power equipment and voltage anomalies of nodes, where nodes represent common connection points such as busbars in the power grid. Equipment status indicators include switch changes, protection device action signals, and equipment alarm signals. System business indicators include SCADA or EMS system data refresh status and command transmission timeliness. Based on timestamps, the temporal performance characteristics of each abnormal event are calculated, including the perception response time and recovery handling time. The specific calculation steps are as follows: The duration from the start time of the initial event of the fault source to the cutoff time of capturing and presenting the corresponding second event feature is denoted as the perception response duration. The recovery processing time is defined as the time from the moment when the corresponding second event characteristic appears until the moment when the electrical quantity indicators, equipment status indicators, and system business indicators are all restored to the preset safe and stable threshold range through automatic or manual operation. The calculated perception response time and recovery processing time are used as the third event feature for each abnormal event.

3. The cross-regional synchronous power monitoring method based on a time-series database according to claim 2, characterized in that: Based on the first, second, and third event characteristics corresponding to each historical anomaly, a corresponding event sequence model is constructed step by step to obtain a complete set of event sequence models for all anomalies. The specific steps include: For each historical anomaly, its first event feature and second event feature are connected by a directed edge to generate the basic event sequence model of the event, wherein the direction of the directed edge is from the first event feature to the second event feature; Based on the generated event sequence model, the second event feature and the third event feature are connected by directed edges to generate a complete event sequence model, wherein the direction of the directed edges is from the second event feature to the third event feature; Each generated complete event sequence model is added to the event sequence model set to form a complete event sequence model set covering all historical abnormal events.

4. The cross-regional synchronous power monitoring method based on a time-series database according to claim 3, characterized in that: The complete event sequence model set is traversed. For each event sequence model that serves as the main event, temporal proximity filtering and spatial correlation filtering are performed to filter out event pairs consisting of the main event and candidate events. Based on this, causal orientation analysis is performed to determine causal relationships. All determined causal relationships are aggregated in the form of link pairs with the main event as the cause and the candidate event as the effect, generating a causal link pair set. The specific steps include: Traverse the complete set of event sequence models. For each event sequence model, denoted as the main event M, perform time proximity filtering and spatial correlation filtering. The specific operations are as follows: Proximity filtering: In the time series database, find all other event sequence models whose start timestamp differs from the main event M by a preset time window, and denot them as candidate events N; Spatial correlation filtering: From candidate events N, further filter out events whose influence range within the first event feature overlaps with or is electrically connected to the influence range of the main event M; For the event pairs consisting of the selected main event M and candidate events N, causal orientation analysis is performed. The specific steps are as follows: Perform a timing determination by comparing the start timestamps of the main event M and the candidate event N. If the start time of the main event M is earlier than the start time of the candidate event N, then it is considered that there is a timing premise from the main event M to the candidate event N. Perform logical judgment and analyze the characteristics of the second event of the main event M, that is, the impact of the main event M, to determine whether it provides a direct cause or necessary condition for the occurrence of the candidate event N. If both the temporal premise and logical possibility are satisfied, then it is determined that there is a causal relationship from the main event M to the candidate event N; The causal relationship obtained for each event pair will be represented as a link pair (M, N), where M is the cause and N is the effect. All such causal link pairs are aggregated to generate a causal link pair set.

5. The cross-regional synchronous power monitoring method based on a time-series database according to claim 4, characterized in that: Based on the causal link pair set, the associated causal link pairs with the same event type and electrical correlation in the scope of influence are marked and grouped; the situation element set and deviation state index set are extracted from the group, a mapping is established and stored in the knowledge base, and the consequence type and average recovery time of each situation element are marked. Real-time monitoring of anomalies in situational elements, querying the knowledge base to identify common consequences, calculating the estimated maximum recovery time, and issuing an alert if the time exceeds a threshold. Specific steps include: Iterate through the set of causal link pairs. For each causal link pair CP1 and CP2, the causal event is M1 and the effect event is N1 in causal link pair CP1; and the causal event is M2 and the effect event is N2 in causal link pair CP2. If causal events M1 and M2 have the same event type and their scope of influence has an electrical or topological connection, then causal link pair CP1 and causal link pair CP2 are marked as associated causal link pairs. Group all causal link pairs that have the same event type and electrical association to form associated causal chain groups; For each causal chain group, analyze all causal events within the group, combine historical operating data, extract their common preconditions, including power grid structure characteristics, operating state characteristics before the event, and environmental condition characteristics, to form a situation element set F; extract the common state index anomalies in the second event characteristics of all result events within the group to form a deviation state index set U. Establish a mapping relationship between the situation element set F and the deviation state index set U; Repeat the above steps to process all related causal chain pairs, generate mapping pairs, and store them in the situation mapping knowledge base; Based on the generated situation mapping knowledge base, each situation element f is labeled with all the types of consequences it has caused in history, and the average recovery and handling time corresponding to each situation element f is recorded. The status of each situation element during network operation is captured and monitored in real time. When one or more situation elements are detected to be in an abnormal state: query the situation mapping knowledge base to find the common consequence event types caused by all activated situation elements. The maximum historical average recovery time corresponding to all activated status elements is taken as the estimated maximum recovery time. When the estimated maximum recovery time exceeds the preset acceptable interruption time threshold, an early warning is immediately issued to the relevant area.

6. A cross-regional synchronous power monitoring system based on a time-series database, applied to the cross-regional synchronous power monitoring method based on a time-series database as described in any one of claims 1-5, characterized in that: The system includes: a historical anomaly event processing module, an event sequence model construction module, a causal analysis module, a related causal chain grouping module, and an early warning module. The historical anomaly event processing module acquires historical anomaly events and stores them in a time-series database, extracts first and second event features, calculates the perception response time and recovery handling time, and uses these as the third event feature. The event sequence model construction module constructs corresponding event sequence models step-by-step to obtain a complete set of event sequence models for all anomaly events. The causal analysis module iterates through this set to filter out event pairs, performs causal orientation analysis to determine causal relationships, and generates a set of causal link pairs. The related causal chain grouping module iterates through the set of causal link pairs, marks related causal chain pairs, and groups them; establishes mappings and stores them in a knowledge base, marking the consequence type and average recovery time of each situation element. The early warning module monitors anomalies in situation elements in real time, queries the knowledge base to find common consequences, calculates the estimated longest recovery time, and issues an early warning if the time exceeds a threshold.

7. The cross-regional synchronous power monitoring system based on a time-series database according to claim 6, characterized in that: The historical abnormal event processing module includes a multi-source data access unit, a time-series data storage unit, an event feature extraction unit, and a time-series performance calculation unit; the multi-source data access unit is used to establish a stable connection with the data sources of each monitored area to obtain historical abnormal events of the power monitoring equipment; The time-series data storage unit is used to sort the acquired historical abnormal events by timestamp and store them in the time-series database; the event feature extraction unit is used to extract the first and second event features of each historical abnormal event from the time-series database. The timing performance calculation unit is used to calculate the characteristics of the third event based on the timestamp.

8. The cross-regional synchronous power monitoring system based on a time-series database according to claim 7, characterized in that: The event sequence model construction module includes a basic event sequence model generation unit, a complete event sequence model generation unit, and an event sequence model set management unit. The basic event sequence model generation unit generates a basic model for each historical abnormal event by connecting the first and second event features with directed edges. The complete event sequence model generation unit completes the event's full lifecycle logic by connecting the second and third event features with directed edges based on the basic model. The event sequence model set management unit aggregates all complete event sequence models to form a complete event sequence model set.

9. The cross-regional synchronous power monitoring system based on a time-series database according to claim 8, characterized in that: The causal analysis module includes a model set traversal unit, a temporal proximity filtering unit, a spatial correlation filtering unit, a causal orientation analysis unit, and a causal link set generation unit. The model set traversal unit is used to access each model in the complete event sequence model set one by one and define it as the main event. The temporal proximity filtering unit is used to filter candidate events. The spatial correlation filtering unit is used to filter events from the candidate events whose influence range intersects with or is electrically connected to the main event. The causal orientation analysis unit is used to verify the causal relationship of event pairs composed of the main event and candidate events through temporal and logical judgment. The causal link set generation unit is used to summarize the event pairs that are determined to have a causal relationship and generate a causal link set.

10. The cross-regional synchronous power monitoring system based on a time-series database according to claim 9, characterized in that: The associated causal chain grouping module includes an associated causal chain pair marking unit, an extraction unit, a mapping establishment unit, a recovery time marking unit, and an early warning unit. The associated causal chain pair marking unit is used to mark link pairs that are electrically related due to the same event type and the scope of influence, and group them into the same associated causal chain group. The extraction unit is used to extract the situation element set and the deviation state index set for each associated causal chain group. The mapping establishment unit is used to establish mapping relationships and store the established mapping relationships in the knowledge base. The recovery time marking unit is used to mark the type of consequence event that each situation element in the knowledge base has caused in the past, and calculate and store the corresponding average recovery and handling time.

Images