A power failure event intelligent judgment method and system based on stream computing

Abstract

The application discloses a power failure event intelligent judgment method and system based on flow calculation, and belongs to the technical field of distribution network flow calculation data processing. The method comprises the following steps: real-time access to multi-source power failure correlation data, performing event regularization processing on the multi-source power failure correlation data, and generating a power failure candidate event set; mapping the power failure candidate event set to a distribution network topology relationship chain, carrying out topology correlation analysis, completing power failure range merging, and outputting a power failure range result; matching the power failure range result with a historical power failure mode, completing event confirmation, and obtaining a fault point output result based on topology reverse convergence research and judgment. The application solves the problems that the prior art is prone to confusing single-point communication anomalies with real power failures and that the certainty of fault point determination is insufficient, can realize real-time research and judgment of distribution network power failure events, and improves the accuracy of power failure range identification and fault positioning.

Description

Technical Field

[0001] This invention relates to the field of power distribution network flow calculation data processing technology, and in particular to a method and system for intelligent judgment of power outage events based on flow calculation. Background Technology

[0002] With the continuous influx of massive amounts of terminal, event, curve, and frozen data into metering automation systems, power outage-related information has become increasingly multi-sourced, continuous, and strongly time-series correlated. In existing metering platforms, terminal power outage alarms, meter events, curve changes, frozen data results, and equipment file information originate from different data links, resulting in differences in data arrival order, representation, and correlation granularity. Simultaneously, the power distribution network has a clear topological hierarchy and branch structure, and the propagation process of power outage events along the topology is fundamentally different from that of single-point anomalies or communication anomalies. For such services, relying solely on a single alarm record or isolated state change is insufficient to accurately reflect the true extent and location of the power outage.

[0003] Existing distributed platforms already possess real-time streaming computing capabilities, which can trigger computing jobs by subscribing to curve data, freezing data, and real-time reporting events, and can be applied to intelligent analysis scenarios of power outage events. However, how to uniformly organize terminal power outage alarm streams, distribution network topology relationships, and historical power outage patterns, and form a streaming processing method for power outage event judgment, still requires specialized design based on the characteristics of power outage services.

[0004] Existing technologies typically employ a power outage monitoring approach based on real-time event subscription: receiving terminal power outage alarms, real-time reported events, and related metering data, using these as inputs for streaming computing, combining them with basic archival information for online analysis, and outputting power outage assessment results. The core of this approach lies in leveraging a streaming computing engine to improve event processing timeliness, enabling rapid identification of power outage-related anomalies upon event arrival. Event data, curve data, frozen data, and archival data are stored in adapted data carriers, and event-related data and real-time data are jointly accessed during real-time analysis, supporting power outage monitoring, anomaly identification, and status analysis, possessing a basic processing chain of "receiving events—triggering computation—outputting results."

[0005] However, existing technologies have the following shortcomings: 1) The judgment of power outage events is mostly based on single-point alarm triggering or simple event aggregation, which is insufficient in the structured correlation processing between terminal power failure alarms, meter status changes, equipment ownership relationships, and power supply path relationships, and lacks a unified event regularization mechanism; for duplicate alarms, conflict alarms, isolated alarms, and out-of-order arrival data records, there is a lack of dedicated processing links for power outage scenarios. 2) The utilization of distribution network topology is insufficient, often failing to incorporate the relationships between terminal nodes, meter nodes, transformer nodes, line nodes, and switch segments into a unified power outage propagation analysis process, making it difficult to perform branch-level merging, boundary-level interception, and range-level confirmation of power outage events; judging solely based on the number of alarms or temporal proximity can easily confuse local anomalies with actual power outages. 3) The utilization of historical power outage handling results, historical propagation paths, and historical boundary patterns is insufficient, lacking an event confirmation and fault point convergence analysis mechanism based on historical power outage patterns, making it impossible to establish a stable correspondence between current events and historically known power outage scenarios, resulting in insufficient certainty in fault point judgment results. Therefore, a method and system for intelligent judgment of power outage events based on stream computing is needed. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention provides a method and system for intelligent judgment of power outage events based on stream computing. This system can uniformly access and organize terminal power outage alarm data, real-time reported event data, curve data, frozen data, equipment file data, and topology mapping data. Based on this, it performs topology correlation analysis, power outage range merging, historical pattern matching, event confirmation, and fault point determination. The specific technical solution is as follows: A method for intelligently determining power outage events based on stream computing includes the following steps: Real-time access to multi-source power outage correlation data, event normalization processing of multi-source power outage correlation data, and generation of a set of power outage candidate events; The set of power outage candidate events is mapped to the distribution network topology relationship chain, topology correlation analysis is carried out and the power outage range is merged, and the power outage range result is output. The power outage range results are matched with historical power outage patterns to complete event confirmation, and the fault point output results are obtained based on topology reverse convergence analysis.

[0007] Preferably, the real-time access to multi-source power outage correlation data, and the event normalization processing of the multi-source power outage correlation data to generate a set of power outage candidate events, specifically includes: The multi-source power outage associated data includes terminal power outage alarm data, real-time reported event data, curve data, frozen data, device file data, and topology mapping data; Event standardization includes: unified identifier merging, event field standardization, duplicate record merging, conflict record resolution, missing field completion, and invalid record removal; The power outage candidate event set is then obtained based on equipment ownership, reporting sequence, and event adjacency.

[0008] Preferred options also include: When generating a set of candidate power outage events, isolated event records are kept separately and not directly merged into the existing event set; event records that only have a single-point alarm and do not form a response relationship between adjacent nodes are kept as objects to be verified.

[0009] Preferably, the step of mapping the set of power outage candidate events to the distribution network topology relationship chain, performing topology correlation analysis and completing the power outage range merging, and outputting the power outage range result specifically includes: The set of candidate power outage events is mapped to a topological relationship chain consisting of terminals, meters, transformer substations, switches, and line nodes. Path connection and path isolation are performed along the power supply path. The event timing and status are combined to perform correlation verification, and power outage correlation segments are formed. The power outage range is determined based on the power outage correlation segments.

[0010] Preferred options also include: When merging the outage range, nodes with missing file mappings but clear path relationships are retained to participate in the merging; event records with recovery status before outage status, contradictory upstream and downstream propagation order, or unclosed path relationships are removed from the current outage-related segments.

[0011] Preferably, the step of matching the power outage range result with historical power outage patterns, completing event confirmation, and obtaining the fault point output node based on topology reverse convergence analysis specifically includes: Extract the event features corresponding to the power outage range results. The event features include alarm combination structure, topology propagation path, node power failure sequence, event recovery sequence, and range boundary shape. The event characteristics are compared and matched with historical power outage patterns stored in the historical power outage event database; After a successful match, combined with historical pattern information, the fault location range is narrowed down by tracing back upstream of the power source along the topology within the power outage area, and candidate fault point segments and candidate fault point devices are obtained by gradually narrowing down the fault location range. When multiple candidate objects exist, disambiguation filtering is performed, and the final fault point judgment result is output.

[0012] Preferably, the reverse backtracking process includes: backtracking sequentially from the boundary of the power outage area to upstream nodes, identifying the boundary nodes of the power outage propagation, excluding devices that do not conform to the current propagation path based on the power outage coverage relationship, and finally retaining the candidate devices that are most consistent with the current power outage area.

[0013] Preferably, when multiple candidate objects exist, the disambiguation screening is performed, and the final fault point judgment result is output, including: When multiple candidate devices exist, the multiple candidate devices are disambiguated by combining the location of the breakpoint in the historical power outage mode, the difference in power loss coverage between upstream and downstream, and the order of restoration, and the final fault point judgment result is output.

[0014] Preferably, the historical power outage patterns consist of a set of structured patterns formed by confirmed historical power outage events, and each historical power outage pattern includes at least the corresponding topology propagation relationship, node change sequence, impact range boundary, and historical handling results.

[0015] A power outage event intelligent judgment system based on stream computing, which applies the method described, including a data access and normalization module, a topology association and range merging module, and a pattern matching and fault judgment module; The data access and normalization module is used to access multi-source power outage related data in real time, perform event normalization processing on the multi-source power outage related data, and generate a set of power outage candidate events. The topology association and range merging module is used to map the set of power outage candidate events to the distribution network topology relationship chain, perform topology association analysis and complete the power outage range merging, and output the power outage range result. The pattern matching and fault assessment module is connected to the topology association and range merging module. It is used to match the power outage range results with historical power outage patterns, complete event confirmation, and obtain the fault point output result based on topology reverse convergence assessment.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. In this invention, the power outage judgment process is divided into three consecutive processing links: event regularization, topology merging, and pattern confirmation. This enables the terminal power outage alarm, meter status change, equipment file relationship, and power supply path of the distribution network to be uniformly organized within the same streaming processing link, forming a clearly structured power outage event judgment process that is suitable for real-time streaming computing processing scenarios.

[0017] 2. In this invention, a distribution network topology relationship chain is introduced in the power outage range determination stage. The event records are processed by path concatenation, path isolation, correlation verification and fragment merging. The power outage range result directly corresponds to the power supply branch, transformer area boundary and affected node set, avoiding the merging of discontinuous topology fragments into the same power outage event. The quantitative coverage index is used to distinguish between real power outages and single-point communication anomalies.

[0018] 3. In this invention, historical power outage patterns are introduced during the event confirmation stage. The alarm combination structure, propagation path, power outage sequence, recovery sequence, and range boundary of the current event are matched and analyzed. Pattern filtering is completed through a similarity formula. Furthermore, the fault point convergence judgment is performed by combining the topology backtracking confidence calculation formula. This ensures that the power outage event confirmation result and the fault point judgment result have quantifiable data basis and unified processing logic, thereby improving the accuracy of fault location.

[0019] 4. In this invention, the historical power outage event database can continuously iterate and expand the pattern samples according to the on-site handling results, and has self-learning ability to adapt to changes in the operation of the power distribution network. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0021] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0022] 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, not all, of the embodiments of the present invention. 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.

[0023] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0024] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0025] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0026] Example 1: like Figure 1 As shown, a method for intelligently determining power outage events based on stream computing includes the following steps: S1. Real-time access to multi-source power outage correlation data, perform event normalization processing on the multi-source power outage correlation data, and generate a set of power outage candidate events; details are as follows: Specifically, in this step, multi-source data related to power outage determination is first received. This multi-source data includes terminal power outage alarm data, real-time reported event data, curve data, frozen data, equipment file data, and distribution network topology mapping data. Specifically, terminal power outage alarm data and real-time reported event data are used to characterize the status change information on the terminal side; curve data and frozen data are used to provide metering-side change information corresponding to the power outage event; equipment file data is used to characterize the affiliation relationships between terminals, meters, transformer substations, lines, and switching equipment; and distribution network topology mapping data is used to characterize the power supply path relationships.

[0027] Specifically, a unified identifier mapping process is performed on all types of data received, converting data from different sources and with different field formats into a uniformly processable event data structure. This unified identifier mapping process includes device identifier mapping, terminal-meter association mapping, user-transformer area association mapping, and transformer area-line association mapping. Subsequently, event parsing and field normalization are performed on the event data structure to extract the alarm occurrence time, alarm recovery time, event type identifier, event status identifier, associated device identifier, and associated topology node identifier.

[0028] Furthermore, data cleaning processing is performed on the standardized event records. This data cleaning processing includes merging duplicate records, correcting conflicting records, removing invalid records, and completing missing fields. Specifically, for duplicate power outage alarm records generated by the same device in adjacent reporting cycles, they are merged according to device identifier and event status. For event records of the same device with contradictory statuses within the same time period, they are corrected according to the original reporting order and the evolution relationship of event status. For event records that are missing key related fields but can be supplemented through file relationships, field completion is performed using device file data.

[0029] After data normalization, the event records are grouped according to equipment affiliation, reporting sequence, and event adjacency to generate multiple power outage candidate event sets. The event adjacency relationship refers to the positional relationship of the topology nodes mapped to by the event records after normalization, within the distribution network topology chain consisting of terminals, meters, transformer substations, switches, and line nodes, being located on the same power supply path and directly connected, or belonging to the same power supply branch or transformer substation defined by the same switch segment. Determining whether two event records have the event adjacency relationship includes: based on the attribution relationship between terminals, meters, distribution areas, lines, and switches represented by the equipment file data, and the power supply path connection relationship represented by the topology mapping data, determining the connection level and power supply unit to which the corresponding nodes of the two event records belong in the distribution network topology relationship chain; when the two nodes are directly connected on the power supply path (i.e., there are no other nodes between them), or when the two nodes belong to the same power supply branch or the same distribution area defined by the same switch segment, it is determined that the two event records have an event adjacency relationship; The process of dividing the corresponding nodes into the power outage candidate event set specifically includes: grouping multiple event records that have event adjacency relationships between their corresponding nodes, belong to the same equipment, and are reported within the same preset time window with consecutive timing into the same power outage candidate event set; and dividing event records that do not have event adjacency relationships between their corresponding nodes, are isolated by switch boundaries, or belong to different transformer areas into different power outage candidate event sets. The power outage candidate event set is used to represent clusters of abnormal events with continuous reporting associations within the same local power supply relationship.

[0030] Preferably, when generating the power outage candidate event set, isolated event records are retained separately and not directly merged into the existing event set; event records that only have a single-point alarm and do not form a response relationship between adjacent nodes are retained as objects to be verified. These objects to be verified do not belong to the power outage candidate event set and are not included in subsequent topology association analysis, power outage range merging, and fault point assessment processes. They are only retained separately as events to be confirmed and do not participate in the power outage event judgment process of this invention. For example, when multiple terminals in the same area successively report power outage alarms during continuous reporting, and the curve data of the corresponding electricity meters synchronously show characteristics of voltage loss or stoppage, the multiple event records are merged into the same power outage candidate event set; when a terminal alone reports a power outage alarm, but its adjacent power supply nodes do not report corresponding events, the event record is retained as an independent candidate object.

[0031] S2. Map the set of candidate power outage events to the distribution network topology chain, perform topology correlation analysis, merge power outage ranges, and output the power outage range results; specifically including: Specifically, in this step, the set of candidate power outage events generated in step S1 is mapped to the distribution network topology chain, the terminal node, meter node, user node, branch node, switch node, line node and transformer substation node corresponding to each event record are determined, and the correspondence between the set of candidate events and the power supply path is established.

[0032] Specifically, path expansion processing is performed on the candidate event set along the topological relationship chain to identify the upstream and downstream positional relationships, branch boundary positional relationships, and main branch positional relationships of the event records in the power supply path. For event records located on the same power supply path and with continuous topological positions, path concatenation processing is performed; for event records located on different power supply branches or isolated by switch boundaries, path isolation processing is performed.

[0033] After path expansion is completed, a correlation verification process is performed by combining the occurrence sequence, duration, and recovery status of each event record to determine whether abnormal events on different nodes belong to the same power outage propagation process. The correlation verification process includes: comparing the power outage sequence of upstream and downstream nodes, verifying the consistency of the states of multiple nodes within a branch, and verifying the closure relationship between recovery events and power outage events. Event records that satisfy the propagation relationship of the same power supply path are merged into the same power outage correlation segment.

[0034] Furthermore, based on the set of topology nodes and line segments covered by the power outage-related fragments, the affected user range, equipment range, and branch range are extracted to form the corresponding power outage range results. Specifically, for event sets spanning multiple branches, they are split and merged according to the topology boundary; for event sets that are discontinuous in the topology, no forced concatenation is performed.

[0035] Preferably, when performing the power outage range merging, for nodes with missing file mappings but clear path relationships, their current topological position is retained for merging; for event records where the recovery state precedes the power outage state, the upstream and downstream propagation order is contradictory, or the path relationship cannot be closed, the current power outage-related fragments are removed.

[0036] For example, when multiple downstream terminals corresponding to an upstream switch node of a line branch continuously experience power outage alarms, and no corresponding anomalies occur in adjacent branches outside the branch, the event records within the branch are merged into the same power outage associated segment, and the transformer area, users, and equipment covered by the branch are determined as the power outage range; when two event sets are located after different switch segments, they form two independent power outage associated segments respectively.

[0037] S3. Match the power outage range results with historical power outage patterns, complete event confirmation, and obtain the fault point output results based on topology reverse convergence analysis.

[0038] Specifically, in this step, event features corresponding to each power outage range result obtained in step S200 are extracted. These event features include alarm combination structure, topology propagation path, node power outage sequence, event recovery sequence, and range boundary morphology. Then, the event features are matched and analyzed against historical power outage patterns in a historical power outage event database. The historical power outage patterns consist of a set of structured patterns formed by confirmed historical power outage events, and each historical power outage pattern includes at least the corresponding topology propagation relationship, node change sequence, affected range boundary, and historical handling results.

[0039] Specifically, structural alignment processing is performed on the current power outage range results and historical power outage patterns. Path segment comparison and boundary segment comparison are conducted at the equipment level, branch level, and transformer area level to screen out target patterns that are consistent with the propagation relationship of the current event. For historical patterns with consistent matching results, their corresponding key node locations, breakpoint locations, and handling conclusions are used as the basis for confirming and processing the current event.

[0040] After pattern matching is completed, confirmation processing is performed on the current power outage event, and reverse backtracking is performed along the power supply topology within the current power outage area to narrow down the fault location range layer by layer, obtaining candidate fault point segments and candidate fault point devices. The reverse backtracking process includes: backtracking sequentially from the boundary of the power outage area to upstream nodes, identifying the boundary nodes of the power outage propagation, excluding devices that do not conform to the current propagation path based on the power outage coverage relationship, and finally retaining the candidate devices that are most consistent with the current power outage area; wherein, the exclusion includes: when the power outage coverage of a candidate device does not cover the downstream power outage nodes of the device, or when the upstream and downstream power outage sequence of the candidate device does not match the power outage sequence of the power outage area, the candidate device is excluded; backtracking stops after reaching the boundary node level by level, and the candidate devices that have not been excluded at the end of the backtracking are the candidate devices that are most consistent with the current power outage area. For example, when the power outage coverage of candidate device X includes all downstream power outage nodes and the power outage coverage of candidate device Y does not include some downstream power outage nodes, candidate device Y is excluded and candidate device X is retained. When tracing back to the boundary node, if the power outage coverage of the boundary node includes a non-power outage area and does not match the power outage range, it is excluded and tracing back stops. The candidate device retained at the end of the tracing back is the candidate device that is most consistent with the current power outage range.

[0041] Furthermore, when multiple candidate devices exist, the disambiguation process is performed on these candidate devices by considering the location of the breakpoint in the historical power outage pattern, the difference in power loss coverage between upstream and downstream, and the order of restoration, to output the final fault point determination result. The disambiguation process includes: retaining candidate devices whose breakpoint location matches the location of the breakpoint in the historical power outage pattern; retaining candidate devices whose power loss coverage differs least from the power outage area; retaining candidate devices whose restoration order matches the historical power outage pattern; when a candidate device is located upstream of multiple power loss nodes, that candidate device is determined as the final fault point determination result; the candidate devices retained after disambiguation are output as the final fault point determination result. For example, when the breakpoint location of candidate device P is consistent with the breakpoint location in the historical power outage pattern, but inconsistent with candidate device Q, candidate device P is retained; when the breakpoint locations of candidate device P and candidate device Q are both consistent, the difference between their power outage coverage area and the power outage range is compared, and the one with the smaller difference is retained; when a candidate device is located at the common upstream position of multiple power outage nodes, the candidate device or its adjacent critical boundary device is directly identified as the fault point judgment result. For current events that do not match a complete historical pattern but have formed a stable topology segment, their power outage range result is retained, and the corresponding candidate fault segment is output.

[0042] Preferably, the historical power outage patterns in the historical power outage event database are continuously updated using historically handled events; whenever the current event is confirmed and a handling conclusion is obtained, its event propagation relationship, scope boundary information and fault point results are written into the historical power outage event database for subsequent pattern expansion.

[0043] For example, if the power outage sequence, recovery sequence, and boundary pattern of the current event on a certain branch are consistent with a certain type of branch power outage event in the past, and a certain upstream switchgear is located at the common upstream position of multiple power outage nodes, then the section corresponding to the switchgear is identified as a candidate section for the fault point, and the switchgear or its adjacent key boundary equipment is further identified as the fault point judgment result.

[0044] Example 2: A power outage event intelligent judgment system based on stream computing, comprising the following methods: The data access and standardization module is used to receive terminal power failure alarm data, real-time reported event data, curve data, frozen data, device file data and topology mapping data, and perform identification merging, event parsing, field standardization, duplicate record merging, conflict record resolution, missing field completion and invalid record removal on the access data, and generate a set of power failure candidate events. The topology association and range merging module, connected to the data access and normalization module, is used to map the set of power outage candidate events to the distribution network topology relationship chain, perform power supply path serialization, path isolation, association verification and fragment merging, and output the power outage range result; The pattern matching and fault assessment module is connected to the topology association and range merging module. It is used to extract the alarm combination structure, topology propagation path, node power failure sequence, recovery sequence and range boundary shape corresponding to the power outage range result, match them with the power outage patterns in the historical power outage event database, and perform event confirmation, fault section convergence and fault equipment screening, and output the power outage event judgment result.

[0045] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0046] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0047] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0048] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for intelligently determining power outage events based on stream computing, characterized in that, Includes the following steps: Real-time access to multi-source power outage correlation data, event normalization processing of multi-source power outage correlation data, and generation of a set of power outage candidate events; The set of power outage candidate events is mapped to the distribution network topology relationship chain, topology correlation analysis is carried out and the power outage range is merged, and the power outage range result is output. The power outage range results are matched with historical power outage patterns to complete event confirmation, and the fault point output results are obtained based on topology reverse convergence analysis.

2. The intelligent judgment method for power outage events based on stream computing according to claim 1, characterized in that, The real-time access to multi-source power outage correlation data, the event normalization processing of the multi-source power outage correlation data, and the generation of a power outage candidate event set specifically include: The multi-source power outage associated data includes terminal power outage alarm data, real-time reported event data, curve data, frozen data, device file data, and topology mapping data; Event standardization includes: unified identifier merging, event field standardization, duplicate record merging, conflict record resolution, missing field completion, and invalid record removal; The power outage candidate event set is then obtained based on equipment ownership, reporting sequence, and event adjacency.

3. The intelligent judgment method for power outage events based on stream computing according to claim 2, characterized in that, Also includes: When generating a set of candidate power outage events, isolated event records are kept separately and are not directly merged into the existing event set. Event records that only show a single alarm point and do not form a response relationship between adjacent nodes are retained as objects to be verified.

4. The intelligent judgment method for power outage events based on stream computing according to claim 1, characterized in that, The process of mapping the set of candidate power outage events to the distribution network topology relationship chain, performing topology correlation analysis, merging power outage ranges, and outputting power outage range results specifically includes: The set of candidate power outage events is mapped to a topological relationship chain consisting of terminals, meters, transformer substations, switches, and line nodes. Path connection and path isolation are performed along the power supply path. The event timing and status are combined to perform correlation verification, and power outage correlation segments are formed. The power outage range is determined based on the power outage correlation segments.

5. The intelligent judgment method for power outage events based on stream computing according to claim 4, characterized in that, Also includes: When merging power outage areas, for nodes with missing file mappings but clear path relationships, their current topology position is retained for merging. For event records where the recovery state precedes the power outage state, the upstream and downstream propagation order is contradictory, or the path relationship cannot be closed, the current power outage-related segments are removed.

6. The intelligent judgment method for power outage events based on stream computing according to claim 1, characterized in that, The process of matching the power outage range results with historical power outage patterns to complete event confirmation and obtain the fault point output node based on topology reverse convergence analysis specifically includes: Extract the event features corresponding to the power outage range results. The event features include alarm combination structure, topology propagation path, node power failure sequence, event recovery sequence, and range boundary shape. The event characteristics are compared and matched with historical power outage patterns stored in the historical power outage event database; After a successful match, combined with historical pattern information, the fault location range is narrowed down by tracing back upstream of the power source along the topology within the power outage area, and candidate fault point segments and candidate fault point devices are obtained by gradually narrowing down the fault location range. When multiple candidate objects exist, disambiguation filtering is performed, and the final fault point judgment result is output.

7. The intelligent judgment method for power outage events based on stream computing according to claim 6, characterized in that, The reverse backtracking process includes: tracing back sequentially from the boundary of the power outage area to upstream nodes, identifying the boundary nodes of the power outage propagation, excluding devices that do not conform to the current propagation path based on the power outage coverage relationship, and finally retaining the candidate devices that are most consistent with the current power outage area.

8. The intelligent judgment method for power outage events based on stream computing according to claim 6, characterized in that, When multiple candidate objects exist, disambiguation filtering is performed, and the final fault point judgment result is output, including: When multiple candidate devices exist, the multiple candidate devices are disambiguated by combining the location of the breakpoint in the historical power outage mode, the difference in power loss coverage between upstream and downstream, and the order of restoration, and the final fault point judgment result is output.

9. The intelligent judgment method for power outage events based on stream computing according to claim 6, characterized in that, The historical power outage patterns consist of a set of structured patterns formed by confirmed historical power outage events. Each historical power outage pattern includes at least the corresponding topology propagation relationship, node change sequence, impact range boundary, and historical handling results.

10. A power outage event intelligent judgment system based on stream computing, characterized in that, The method applied to any one of claims 1 to 9 includes a data access and normalization module, a topology association and range merging module, and a pattern matching and fault assessment module; The data access and normalization module is used to access multi-source power outage related data in real time, perform event normalization processing on the multi-source power outage related data, and generate a set of power outage candidate events. The topology association and range merging module is used to map the set of power outage candidate events to the distribution network topology relationship chain, perform topology association analysis and complete the power outage range merging, and output the power outage range result. The pattern matching and fault assessment module is connected to the topology association and range merging module. It is used to match the power outage range results with historical power outage patterns, complete event confirmation, and obtain the fault point output result based on topology reverse convergence assessment.

Images