Power transmission line state evaluation and early warning system, method and device based on cloud edge collaboration

Abstract

This invention relates to the field of transmission line condition monitoring and fault location technology, and discloses a transmission line condition assessment and early warning system, method, and device based on cloud-edge collaboration. The system includes multiple edge monitoring terminals and a cloud-based condition assessment and early warning platform. The cloud determines candidate fault intervals, the current location result level, and target nodes based on callable waveform records and line node topology; it determines the reception processing status of the waveform record corresponding to the target node based on reception start information, reception end information, and whether a callable waveform record has been formed; and it keeps the fault event active while receiving data or after reception but before a callable waveform record has been formed; when the preliminary closure conditions are met and the current callable waveform record supports at least one location path, it selects a location path and generates the location result and early warning information. This invention can reduce the risk of premature closure of fault events while critical waveforms are still being transmitted or processed.

Description

Technical Field

[0001] This invention relates to the field of transmission line condition monitoring and fault location technology, specifically to a transmission line condition assessment and early warning system, method and equipment based on cloud-edge collaboration. Background Technology

[0002] Multiple monitoring nodes are typically installed along transmission lines. Each node's edge monitoring terminal collects fault traveling waves, transient current waveforms, power frequency waveforms, or other line status signals, and uploads the waveforms and node information to a cloud platform. Based on the line topology, waveform generation time, and node location, the cloud platform generates a fault event for the same abnormal process on the line and outputs results such as anomaly confirmation, fault range, or fault distance.

[0003] Waveforms from multiple nodes involved in the same fault event may not necessarily arrive at the cloud simultaneously. Due to the influence of communication links, waveform data volume, and cloud processing progress, some waveforms may already be available for use by the fault location program, while others may still be being received, or may have been received but are still undergoing integrity verification, event association, and index building. These states can be distinguished by reception start information, reception end information, waveform integrity status, and processing records.

[0004] Current processing typically determines whether to perform location based on a fixed waiting time or a "waveform available / no waveform" status, without simultaneously considering whether nodes that have not yet formed callable records will change the current candidate fault range, form cross-node location combinations, or improve the location result level. This processing is prone to two situations: continuously waiting for nodes that are irrelevant to the current location result, or treating critical waveforms that are being transmitted or processed as missing data, causing the fault event to terminate prematurely.

[0005] Fields such as receive start, receive end, and integrity only reflect the processing stage of a waveform; they cannot directly indicate whether the waveform is worth waiting for. If control events are only based on the processing stage, the system may still wait for nodes that will not change the positioning result. If positioning is only performed based on currently available waveforms, it cannot identify key waveforms that are forming and may change the boundary or positioning path. Therefore, it is necessary to place the potential contribution of a node to the current positioning result and its waveform processing progress in the same decision chain.

[0006] Furthermore, even after a fault event generates a location result, waveforms belonging to the same physical fault process may still be received. If all late waveforms trigger the complete location process, it will increase redundant calculations and repeated warnings; if all late waveforms are simply archived, data that could narrow down the fault range or generate distance location data may be missed. Summary of the Invention

[0007] The technical problem this invention aims to solve is that, in the case of asynchronous uploading and processing of waveforms from multiple nodes, existing systems do not combine whether a node meets the positioning contribution conditions required for the current positioning result with the receiving and processing status of the node's waveform. This easily leads to the identification of critical waveforms that are still being transmitted or processed as missing data, causing fault events to close prematurely before the positioning conditions have stabilized.

[0008] The target node referred to in this invention is a node whose topological location and the acquisition capability of the corresponding terminal satisfy at least one of the following conditions: boundary re-judgment condition, cross-node combination condition, or location result level improvement condition. Therefore, it is necessary to track the expected received items or waveform record reception and processing status of the node before the fault event is closed.

[0009] This invention does not solve the aforementioned problem by extending the uniform waiting time. Its approach is to first filter waiting objects that meet the positioning contribution conditions based on the node topology location and the corresponding terminal's acquisition capabilities. Then, it determines whether the expected received item or waveform record of the object has already entered the receiving or processing process. Finally, it determines the fault event status based on whether the currently available waveform records meet the positioning conditions. Therefore, both the waiting objects and the closure conditions correspond to the actual positioning requirements of the current fault event.

[0010] To achieve the above objectives, this application adopts the following technical solution: a power transmission line status assessment and early warning system based on cloud-edge collaboration, comprising multiple edge monitoring terminals set at different monitoring nodes of the power transmission line, and a cloud-based status assessment and early warning platform that is communicatively connected to the edge monitoring terminals;

[0011] The edge monitoring terminal is used to generate and upload waveform records associated with line identifiers, node identifiers, terminal time, and waveform identifiers;

[0012] The cloud-based status assessment and early warning platform is used to: establish a line node topology; associate waveform records with fault events according to the line identifier and terminal time; determine candidate fault intervals and target nodes based on the callable waveform records and the line node topology; and after the waveform records corresponding to the target nodes are added to the fault events, they can change the boundaries of the candidate fault intervals, form cross-node arrival time difference positioning combinations, or improve the positioning result level.

[0013] Based on the reception start information, reception end information, and whether a callable waveform record has been formed, determine the reception processing status of the expected reception item or waveform record corresponding to the target node, and keep the fault event enabled when receiving or when receiving is completed but a callable waveform record has not yet been formed;

[0014] If there is no target node in either of the above two receiving and processing states, and the currently available waveform record supports at least one location path, the fault event is initially closed, the location path is selected according to the topological location of the node corresponding to the available waveform record, and the location result and warning information are generated.

[0015] Further, the positioning contribution conditions are determined according to the following rules: when a candidate node is located within a candidate fault interval and is between a node corresponding to a callable waveform record and a boundary node, the candidate node is determined as a boundary re-judgment node, and the boundary is re-determined according to a preset boundary judgment rule after it forms a callable waveform record; when a candidate node and at least one node corresponding to a callable waveform record are located on different topology sides, and the corresponding terminal has the ability to synchronously acquire the target waveform type, the candidate node is determined as a cross-node combination node; when the current positioning result level is anomaly confirmation and the topological position of the candidate node meets the input conditions for forming two boundary nodes, or when the current positioning result level is interval positioning and the candidate node meets the cross-node combination conditions, the candidate node is determined as a level improvement node.

[0016] Furthermore, the cloud-based status assessment and early warning platform establishes expected reception items for target nodes that have not yet formed actual waveform records, and determines the reception processing status of the expected reception items or waveform records according to the following rules: when no reception start information matching the target node and fault event is obtained, it is determined that reception has not yet started; when reception start information has been obtained but reception end information has not been obtained, it is determined that reception is in progress; when reception end information has been obtained but a callable waveform record has not yet been formed, it is determined that reception is completed but a callable waveform record has not yet been formed; when the waveform record can be called by fault location processing, it is determined that a callable waveform record has been formed.

[0017] Furthermore, the cloud-based status assessment and early warning platform determines the physical event time range based on the terminal time of the waveform record associated with the fault event; when the physical event time range ends and each target node satisfies any of the following conditions, the fault event is initially closed: a callable waveform record has been formed; it is determined to be unavailable; it is in a state where reception has not yet started and a preset waiting end condition is met; wherein, target nodes in a state of receiving or receiving but not yet forming a callable waveform record are not determined to be unavailable.

[0018] Furthermore, the waveform record is also associated with a time synchronization state and a waveform integrity state; when the time synchronization state meets the preset synchronization conditions and the waveform integrity state indicates that the waveform is complete, the waveform record is qualified for time difference of arrival positioning and waveform feature analysis; when the time synchronization state does not meet the preset synchronization conditions, but the waveform integrity state indicates that the waveform is complete, the waveform record is only qualified for waveform feature analysis; when the waveform integrity state indicates that the waveform is incomplete, the waveform record serves as a verification record and is not used for cross-node time difference of arrival positioning.

[0019] Furthermore, the cloud-based status assessment and early warning platform selects the positioning path according to the following rules: when at least two nodes with callable waveform records that are qualified for time difference of arrival positioning are located on different topological sides of the candidate fault interval, cross-node time difference of arrival positioning is selected; when the callable waveform records of multiple nodes can determine the boundary of the candidate fault interval but do not meet the aforementioned conditions, topological interval positioning is selected; when only one node's callable waveform record is qualified for waveform feature analysis, and other target nodes are determined to be unavailable, or are in a state where reception has not yet started and the preset waiting end condition is met, single-node waveform analysis is selected.

[0020] Furthermore, when the cloud-based status assessment and early warning platform receives a new waveform record after the initial closure of a fault event, it determines whether the new waveform record belongs to the fault event based on the line identifier and terminal time. After the late waveform record belonging to the fault event forms a callable waveform record, if the late waveform record, when added to the fault event, can form a new cross-node arrival time difference positioning combination, replace at least one boundary node of the candidate fault interval, or improve the positioning result level, then the positioning path and positioning result are updated; otherwise, the late waveform record is used as a supplementary record and the original positioning result remains unchanged.

[0021] This invention also provides a cloud-edge collaborative method for power transmission line status assessment and early warning, executed by a cloud-based status assessment and early warning platform. The method includes: acquiring line node topology and waveform records uploaded by edge monitoring terminals, and associating the waveform records with fault events; determining candidate fault intervals and the current location result level based on callable waveform records and line node topology, and identifying candidate nodes that meet at least one of the following location contribution conditions as target nodes: located within the candidate fault interval and between the node corresponding to the callable waveform record and the boundary node; located on a different topology side from the node corresponding to the callable waveform record and the corresponding terminal has synchronous acquisition capability; its topological location and the acquisition capability of the corresponding terminal meet location input conditions higher than the current location result level; determining the expected reception item or waveform record reception processing status corresponding to the target node based on reception start information, reception end information, and whether a callable waveform record has been formed; keeping the fault event open while receiving or receiving but before a callable waveform record has been formed; and when there are no target nodes in either of the above two states, and the current callable waveform record supports at least one location path, initially closing the fault event, selecting a location path based on the topological location of the node corresponding to the callable waveform record, and generating location results and early warning information.

[0022] Furthermore, the positioning contribution conditions are determined according to the following rules: when a candidate node is located within a candidate fault interval and is between a node corresponding to a callable waveform record and a boundary node, the candidate node is determined as a boundary re-judgment node; when a candidate node and at least one node corresponding to a callable waveform record are located on different topology sides, and the corresponding terminal has the ability to synchronously acquire the target waveform type, the candidate node is determined as a cross-node combination node; when the current positioning result level is anomaly confirmation and the topological position of the candidate node meets the input conditions for forming two boundary nodes, or when the current positioning result level is interval positioning and the candidate node meets the cross-node combination conditions, the candidate node is determined as a level improvement node.

[0023] The present invention also provides a transmission line condition assessment and early warning device based on cloud-edge collaboration, including a communication interface, a processor and a memory, wherein the memory stores a computer program; when the processor executes the computer program, the device executes a transmission line condition assessment and early warning method based on cloud-edge collaboration.

[0024] The technical effects and advantages of this invention are as follows:

[0025] This invention determines whether a candidate node meets the positioning contribution condition based on its topological location and the acquisition capability of the corresponding terminal. It only waits for target nodes that meet the positioning contribution condition, avoiding prolonged fault event processing time due to irrelevant nodes not yet uploading waveforms. In one embodiment, when a new callable waveform, candidate fault interval, or positioning result level changes, the platform can redetermine the target node set to ensure that the waiting objects correspond to the current positioning requirements.

[0026] This invention utilizes reception start information, reception end information, and whether a retrievable waveform record has been formed to distinguish between reception processing states such as not yet started, receiving in progress, reception completed but no retrievable waveform record formed, and retrievable waveform record formed. This allows for the separation of data being transmitted or processed from truly missing data. The reception start information indicates that the waveform has entered the communication reception process; the reception end information indicates that network reception is complete; and the formation of a retrievable waveform record reflects whether the data has undergone necessary processing and can be retrieved for fault location.

[0027] This invention uses the positioning contribution conditions satisfied by the target node and its receiving and processing status together for fault event closure control. This ensures that the waveform of the target node in the receiving or processing stage is not excluded before a callable record is formed, reducing the risk of premature positioning results while critical waveforms are still being processed. This linkage is not simply an extension of the waiting time, but rather a delayed closure after the target node has entered the receiving or processing process. For nodes that do not meet the positioning contribution conditions or for non-receiving nodes whose waiting conditions have ended, the event can still form results in a timely manner based on the current callable record. Therefore, positioning integrity and event processing timeliness can be coordinated according to the current node combination.

[0028] This invention selects the positioning path based on time synchronization status, waveform integrity, and node topology location. This allows records with insufficient synchronization but complete waveforms to still be used for waveform feature analysis. Even when existing data is insufficient for distance positioning, interval positioning or anomaly confirmation results can still be output. By separating the eligibility for use from the positioning path assessment, a waveform's inability to participate in cross-node time difference calculations does not mean it has completely lost its technical value. Waveforms with insufficient synchronization but complete data can still be used for amplitude, polarity, frequency, or wavefront feature analysis. Even when multiple nodes can determine boundaries but cannot form a synchronized combination, interval results can still be output, thus avoiding the direct discarding of existing data due to insufficient precise ranging conditions.

[0029] This invention assesses the contribution of late-arriving waveforms to the overall positioning process, repositioning only when the positioning structure changes; otherwise, it saves the waveforms, reducing redundant calculations and warnings. It also retains late-arriving waveforms that can improve the positioning results. These waveforms undergo structural change testing within a temporary event set, and the positioning results are only updated when boundaries, positioning levels, or cross-node combinations change. This method separates the saving and repositioning of late-arriving data: all waveforms belonging to the original event are traceable, but only waveforms that alter the on-site positioning information trigger complete calculations and new warnings. Attached Figure Description

[0030] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts:

[0031] Figure 1 This is a schematic diagram of the overall architecture of the cloud-edge collaborative power transmission line condition assessment and early warning system of the present invention;

[0032] Figure 2 This is a schematic diagram illustrating how the present invention determines the target node based on candidate fault intervals and current positioning results;

[0033] Figure 3 This is a schematic diagram of the receiving and processing state transition of the waveform record corresponding to the target node in this invention;

[0034] Figure 4 This is a flowchart illustrating the fault event initiation, preliminary closure, and location path selection of the present invention.

[0035] Figure 5 This is a flowchart illustrating the selective processing of delayed waveform records received after the initial closure of a fault event, as per the present invention. Detailed Implementation

[0036] The embodiments of the present invention will now be described with reference to the accompanying drawings. The functional units in the drawings are used to illustrate data processing relationships. Each functional unit can be implemented by an independent program module or by different processing procedures within the same program.

[0037] In this embodiment, the transmission line can be an overhead transmission line, a new energy power plant transmission line, or other power lines with multiple monitoring nodes along the line. The waveforms collected by the edge monitoring terminal can be fault traveling waves, transient current waveforms, power frequency waveforms, or other waveforms that can reflect abnormal line conditions.

[0038] The "physical event time range" referred to in this embodiment refers to the time range of an abnormal process on the same line determined based on the terminal time recorded by the edge monitoring terminal. In one embodiment, the terminal time of the first abnormal waveform record that meets the event association conditions in the current line is taken as the starting point, and the terminal time of the last waveform record that meets the event association conditions plus a preset silence duration is taken as the ending time. When a new matching waveform record is received within the silence duration, the terminal time of the last waveform record is updated and the ending time is postponed. The platform receiving time is used to determine the order of waveform arrival and processing progress, and does not replace the terminal time in determining the attribution of the fault event.

[0039] In this embodiment, the "candidate fault interval" refers to the line segment used to limit node retrieval and location processing in the current processing stage, and is not equivalent to the final location result; the "preliminary closure" refers to the platform forming a location result based on the currently available waveform records, but still allowing subsequent waveforms to be associated with the fault event. In one embodiment, the platform can save the set of waveform identifiers involved in the location during the preliminary closure as a snapshot of the event members corresponding to this location result.

[0040] The fault event includes at least an open state and a preliminary closed state. The open state allows waveform recording to be continuously associated with the fault event; the preliminary closed state has already formed a location result, but still allows selective processing of subsequently associated late waveforms. In some implementations, an archiving state can also be set, and archiving conditions can be triggered by manual confirmation, on-site verification, or exceeding the late waveform reception deadline.

[0041] In this embodiment, "target node unavailable" means that before the current fault event is initially closed, the target node is determined to be unable to continue generating callable waveform records based on the terminal's offline status, failed receiving task, terminated processing task, or preset waiting end conditions.

[0042] Example 1: Cloud-Edge Collaborative System Architecture

[0043] like Figure 1 As shown, the system includes a power transmission line 1, multiple monitoring nodes 2 set along the power transmission line 1, edge monitoring terminals 3 corresponding to the monitoring nodes 2, a communication network 4, and a cloud-based status assessment and early warning platform 5.

[0044] Each edge monitoring terminal 3 is bound to a line identifier, a node identifier, and a terminal identifier; when phase differentiation is required, the binding information also includes the phase. After detecting a preset trigger condition, the edge monitoring terminal 3 generates a waveform record 6. Each waveform record 6 includes at least the line identifier, node identifier, terminal time, waveform identifier, and waveform data.

[0045] Each tower node can be equipped with edge monitoring terminals 3 corresponding to phases A, B, and C, each terminal having an independent terminal identifier. The platform stores a binding table between terminal identifiers and lines, nodes, and phases. When the waveform event association unit 52 reads the terminal identifier in the waveform record 6, it can verify the node identifier and phase from the binding table to prevent incorrect combination of waveforms from different nodes or different phases due to terminal configuration errors.

[0046] Edge monitoring terminal 3 uploads waveform records 6 via communication network 4. The cloud generates a reception start message upon receiving the first data packet of the waveform; it generates a reception end message upon receiving the terminal's end marker, or upon receiving a data packet matching the expected data length. The waveform integrity can be determined by comparing the expected data length, data segment sequence number, or number of data packets with the actual reception result.

[0047] For example, a waveform record is expected to have a data length of M sampling points and is divided into K data segments. The cloud accumulates the received segments according to the waveform identifier. When the segment sequence number is consecutive, the cumulative number of sampling points reaches M, and the verification passes, a complete state is formed. When the end marker has been received but the segment is missing, the reception completion and incomplete state are recorded, and the missing segment information is retained for communication retransmission or manual review.

[0048] The cloud-based status assessment and early warning platform 5 includes a line topology management unit 51, a waveform event association unit 52, a candidate fault interval determination unit 53, a target node determination unit 54, a processing status determination unit 55, a fault event control unit 56, a location path selection unit 57, and an early warning output unit 58; in the implementation of processing late waveforms, it also includes a late waveform processing unit 59.

[0049] The line topology management unit 51 stores the node sequence, adjacency relationship, and line segment of each monitoring node 2. The node topology can be represented by a node sequence table, adjacency table, or node mileage table. In branch lines, the platform first determines the topology path where the current candidate fault section is located, and then determines the order and topology side of the nodes on that path.

[0050] For branch lines, the platform prioritizes selecting a unique topology path that simultaneously includes existing abnormal nodes, candidate nodes, and the current candidate fault range. If multiple feasible paths exist, the branch to which the fault belongs is determined based on the anomaly confirmation results of each branch side of the branch connecting nodes. If a unique propagation path still cannot be determined, cross-node distance positioning is not performed, and the platform reverts to topology range positioning.

[0051] The waveform event association unit 52 first excludes events from other lines according to the line identifier, and then compares the terminal time with the physical event time range of the existing fault event 7. If the terminal time falls within the range or the continuous triggering range, the waveform record 6 is added to the existing fault event 7; if there is no matching event, a new fault event 7 is created.

[0052] The conditions for associating waveform records with fault events include at least the consistency of line identification and the terminal time being within the preset event association time window. When multiple abnormal processes exist on the same line in adjacent time periods, matching is also performed based on at least one of the following: phase, trigger type, initial abnormal node set, continuous trigger identification, and waveform characteristics, in order to avoid merging adjacent fault processes into the same fault event.

[0053] Fault event 7 saves the event identifier, the line it belongs to, the physical event time range, the waveform record set, the node set, and the event status. To support the processing of late waveforms, it can also save the original positioning result, the updated positioning result, and the set of waveform identifiers involved in the positioning.

[0054] The event status field at least indicates whether it is open or initially closed, and the event member set can store waveform identifiers and their node identifiers. In one implementation, the platform can generate a result version number for each positioning result and record the waveform identifier set used in that version; when a late waveform triggers an update, the new version can remain associated with the original version.

[0055] The candidate fault interval determination unit 53 forms candidate fault intervals and current positioning result levels based on the currently available waveform records and line node topology; the target node determination unit 54 determines whether the candidate node meets the positioning contribution conditions based on the topological location of the candidate node and the acquisition capability of the corresponding terminal.

[0056] In one implementation, when the current fault event includes only the available waveform record of one abnormal node, the nearest configured monitoring nodes on both sides of the topology path containing the abnormal node are used as temporary search boundaries. When the current fault event includes multiple abnormal nodes, the two outermost abnormal nodes on the same topology path are used as temporary search boundaries. If no configured monitoring node exists on a certain topology side, the corresponding line endpoint or branch connection node is used as the temporary search boundary for that side. The temporary search boundaries are used to limit the search range of candidate nodes and are not output as the final fault interval.

[0057] The processing status determination unit 55 reads the reception start information, reception end information, integrity verification result, and event association status to determine the reception processing status of the waveform record corresponding to the target node. The fault event control unit 56 decides accordingly whether to keep the event open or allow initial closure. The positioning path selection unit 57 selects a positioning path based on the waveform's purpose and the node's topology location, and the early warning output unit 58 generates the positioning result 8 and early warning information.

[0058] The cloud-based status assessment and early warning platform 5 can be deployed on one or more computing devices. The computing device includes a communication interface, a processor, and memory. The computer program in the memory is executed by the processor to perform the aforementioned data processing.

[0059] Example 2: Determination of the target node

[0060] like Figure 2 As shown, nodes N1 to N6 are sequentially set along the transmission line. For the current fault event, callable waveform records for nodes N2 and N3 have been obtained. The current candidate fault range is N2 to N5, and the current location result level is range location. N2 and N5 are located at the two boundary directions of the candidate fault range, while N3 and N4 are located inside the range. Note that N5 is the second boundary of the current temporary search range and does not necessarily mean that the waveform of N5 has been recorded as a callable record.

[0061] In this example, N5 can be determined according to the aforementioned temporary retrieval boundary rules, but this does not mean that N5 has formed a callable waveform. The temporary retrieval interval is used to restrict the retrieval of target nodes; after the actual waveform of N5 forms a callable record, the platform then determines whether it supports the current boundary, forms an arrival time difference combination, or meets the distance positioning input conditions.

[0062] In one implementation, the target node determination unit 54 determines whether a candidate node meets the boundary re-judgment condition, cross-node combination condition, or hierarchical improvement condition based on the candidate node's topological location and the corresponding terminal's acquisition capability. If a candidate node meets at least one condition, it is determined as a target node. After the target node forms an actual callable waveform record, positioning processing is performed according to the corresponding boundary judgment rules or positioning path input conditions.

[0063] The candidate node set can consist of nodes within the candidate fault interval, adjacent nodes in the two boundary directions, and nodes that can form an effective topology path with existing callable waveform nodes. The platform can also pre-check whether candidate nodes are configured with the corresponding phase terminals and whether the terminals have the target waveform type and synchronous sampling capability; nodes that do not meet the basic acquisition conditions will not enter the contribution test.

[0064] Once the waveform of node N5 is recorded and can be retrieved, it can be combined with the qualified waveforms of N2 or N3 to form a cross-node arrival time difference positioning combination. Therefore, N5 belongs to the target node. Here, "different topology sides" refers to two nodes located in different boundary directions of the candidate fault interval, or the topology path formed by the two nodes crosses the current line segment to be located.

[0065] Node N4 is located within the candidate fault range and between the callable waveform record corresponding node N3 and the second boundary node N5, and is therefore identified as a boundary re-judgment node. After the actual waveform of N4 forms a callable record, the platform re-determines the second boundary according to the preset boundary judgment rules; when N4 meets the boundary judgment rules, N4 replaces N5, and the candidate fault range is adjusted from N2 to N5 to N2 to N4.

[0066] In one implementation, the platform compares the anomaly confirmation results of candidate nodes with those of adjacent nodes along the topological path corresponding to the candidate fault interval. If the anomaly confirmation result of a candidate node is consistent with that of the first boundary-side node but inconsistent with that of the second boundary-side node, the candidate node replaces the second boundary node; otherwise, the candidate node replaces the first boundary node. If a candidate node cannot generate a valid anomaly confirmation result, boundary replacement is not performed.

[0067] When the current result is only an anomaly confirmation, if the waveform of a candidate node can form two valid boundaries after being added, then the node belongs to the target node that improves the result to interval positioning; when the current result is interval positioning, if the synchronous complete waveform of the candidate node can form a time difference of arrival positioning combination with the existing waveform, then the node belongs to the target node that improves the result to distance positioning.

[0068] Although node N1 has not yet provided a waveform, it is assumed that its addition will not change the boundaries of N2 to N5, form a new cross-node combination, or improve the positioning result level. Therefore, it is not considered as a target node to prevent the initial closure of the fault event.

[0069] In one implementation, when the level of newly added callable waveforms, candidate fault intervals, or location results changes, or when a target node is determined to be unavailable, the platform can re-execute the target node contribution judgment and update the target node set accordingly.

[0070] Example 3: Reception and Processing Status of Target Node Waveform Recording

[0071] like Figure 3 As shown, the processing state determination unit 55 divides the expected received items or waveform records corresponding to the target node into four receiving processing states, P1 to P4, based on the communication receiving record and the data processing record.

[0072] P1 indicates that reception has not yet started. When the platform has not obtained reception start information matching the target node and the current fault event, it is recorded as P1. P1 only indicates that the reception process has not been observed at present, and does not directly indicate that the target node is unavailable. When the target node has not yet formed an actual waveform record, the platform establishes an expected reception item based on the fault event identifier, line identifier, node identifier, terminal identifier, and physical event time range; after receiving waveform header information or the first data packet matching the expected reception item, it writes a waveform identifier and forms reception start information.

[0073] P2 indicates reception in progress. When the platform has generated a reception start message but not yet a reception end message, it is recorded as P2. In this state, waveform data has not yet fully arrived, and the fault event remains active.

[0074] P3 indicates that reception is complete but a callable waveform record has not yet been generated. The platform has generated reception completion information, but the waveform is still undergoing data segmentation verification, decoding, event association, analysis index establishment, or accessible data object generation, and the record is P3.

[0075] P4 represents a retrieveable waveform record. A record is designated as P4 when the waveform data is complete, includes waveform identification and fault event association information, and can be read by the fault location process. P4 does not require persistence to a specific database; it only needs to be accessible to the location process.

[0076] Under normal circumstances, after receiving the first data packet, the process moves from P1 to P2; after receiving the data, it moves from P2 to P3; and after integrity verification and event association are completed and the conditions for reading are met, it moves from P3 to P4. Status changes are written to the processing record corresponding to the waveform identifier, and the fault event control unit 56 updates the event status based on the processing record.

[0077] The processing log can save the state entry time and state exit time. Updates to the receiving processing state can be triggered by messages such as receiving start, receiving end, integrity verification completion, and event association completion, or the target node in state P1, P2, or P3 can be checked according to a preset period.

[0078] When the target node is in P2 or P3, it is not determined to be unavailable. If communication reception is explicitly stopped, or the processing task is terminated and cannot continue to generate a callable waveform, the platform first removes the record from P2 or P3 and records the exception handling result, and then determines that the target node is unavailable based on the terminal offline information, failure information, and waiting conditions.

[0079] When the target node is in P1, the platform decides whether to continue waiting based on preset waiting termination conditions. Waiting termination conditions may include: the physical event time range has ended; the waiting time since the end of the physical event has reached the preset upper limit of the terminal or the line; the terminal was offline or in a communication abnormal state during the event; or the node no longer has positioning contribution after the candidate fault range has been updated.

[0080] In cases of abnormal P2 or P3 durations, the platform first checks whether the receiving or processing task is still running. If the task is still running, the event remains active; if the task fails, is canceled, or exceeds its task-level timeout and cannot be recovered, the reason for the failure is recorded, and the task is treated as an unavailable node. Therefore, the event-level wait limit will not directly override the currently executing receiving and processing tasks.

[0081] The platform determines whether a task is still running based on at least one of the following: the most recent progress update time, changes in the amount of data received, the heartbeat of the receiving or processing task, the maximum receiving duration, the maximum processing duration, and the number of retries. If the task-level time limit is exceeded and no changes in the amount of data or the task heartbeat are detected in a preset number of consecutive checks, it is determined that the receiving or processing task can no longer generate callable waveform records.

[0082] After the waveform enters P4, its eligibility for use is determined based on its time synchronization status and waveform integrity. The time synchronization status can be determined by satellite time synchronization, network time synchronization, or terminal clock calibration results. When the synchronization conditions are met and the waveform is complete, it enters the time difference of arrival (TDOA) positioning candidate set; when the waveform is complete but the synchronization conditions are not met, it enters the waveform feature analysis set; when the waveform is incomplete, it is saved as a verification record.

[0083] Example 4: Initial Closure of Fault Event and Selection of Location Path

[0084] like Figure 4 As shown, in step S101, the cloud-based status assessment and early warning platform 5 acquires waveform record 6 and associates waveform record 6 with fault event 7 according to line identifier and terminal time.

[0085] Step S102: Determine the candidate fault range and location result level based on the currently available waveform records and line node topology. If only line anomaly can be confirmed, the location result level is anomaly confirmation; if two boundary nodes can be identified, it is range location; if reference nodes, line direction, and distance along the line can be further obtained, it is distance location.

[0086] Step S103: Determine the target node according to the positioning contribution conditions described in Embodiment 2; Step S104: Determine the expected receiving item or waveform record receiving and processing status corresponding to each target node according to the rules described in Embodiment 3.

[0087] Step S105: Determine if a target node in state P2 or P3 exists. If it exists, keep fault event 7 enabled and continue reading the receiving and processing status; if it does not exist, proceed to step S106.

[0088] Step S106: If no target node exists in state P2 or P3, determine whether the physical event time range has ended, and whether each target node satisfies any of the following conditions: in state P4; determined to be unavailable; in state P1 and meets the preset waiting end condition. If the above conditions are met, proceed to step S107; otherwise, return to step S104 and continue updating the receiving and processing status of the target node.

[0089] Step S107: Determine the eligibility for each callable waveform record based on its time synchronization status and waveform integrity status. If the time synchronization status meets the preset synchronization conditions and the waveform is complete, it is determined to be eligible for time difference of arrival (TDOA) positioning and waveform feature analysis; if the time synchronization status does not meet the preset synchronization conditions but the waveform is complete, it is determined to be eligible only for waveform feature analysis; incomplete waveforms are saved as verification records. Continuous triggering status, communication delay, and reception time can be used as verification information.

[0090] Step S108: Determine whether the currently available waveform records support at least one positioning path. If not, keep the fault event enabled and return to step S104 to update the status; if the receiving or processing process has terminated and positioning conditions are still not met, proceed to abnormal event handling. If supported, proceed to step S109.

[0091] Step S109 involves initially closing fault event 7 and selecting a location path based on the topological position of the corresponding node relative to the candidate fault interval according to the available waveform record. In one implementation, the waveform identifier and node identifier involved in this location can be written into the event member set to form an event member snapshot at the moment of initial closure.

[0092] In the implementation of saving event member snapshots, the same result version's location processing can read the same waveform identifier set; the waveform arriving after the initial closure can be judged separately by the late waveform processing unit 59 without directly modifying the currently output result version. When at least two nodes qualified for time difference of arrival (TDOA) location are located on different topological sides of the candidate fault interval, cross-node TDOA location is selected; when multiple nodes can determine the interval boundary but do not have the conditions for synchronous location, topological interval location is selected; when only one node has a complete waveform and other target nodes are unavailable or waiting to finish, single-node waveform analysis is selected.

[0093] For example, when the in-phase waveforms of N2 and N5 are complete and meet the synchronization conditions, they are located on different topological sides of the candidate fault interval, and the platform selects cross-node arrival time difference for location; when N2, N4 and N5 can determine the abnormal boundary, but at least one end of the time synchronization condition is insufficient, the platform outputs the topological interval from N2 to N4 or from N2 to N5; when only one complete waveform of N3 is retained, the platform performs single-node feature analysis and outputs abnormal confirmation, without expressing the single-node result as cross-node distance.

[0094] Cross-node time-of-arrival (TOA) positioning can employ existing dual- or multi-terminal traveling wave positioning algorithms. The platform obtains characteristic arrival times from qualified waveforms and calculates the fault location by combining the line topology, segment length, and propagation parameters. The improvement of this invention focuses on the positioning input set and event closure conditions, without limiting the specific ranging formula.

[0095] For linear sections, the platform reads the effective path length L and propagation parameter v between the two nodes from the line ledger. It extracts the arrival times t_A and t_B of the same characteristic wavefront from the qualified waveforms at both ends, and calculates the distance along the line from node A using x = [L + v(t_A - t_B)] / 2. Here, x is the distance along the line from the fault point to node A, L is the effective path length between node A and node B, v is the equivalent propagation speed of the waveform along the effective path, and t_A and t_B are the arrival times of the same characteristic wavefront at nodes A and B, respectively. If the calculation result is less than 0 or greater than L, the path result is marked as abnormal, and the system reverts to the topology interval location.

[0096] For branch lines, the platform first determines the branch to which the fault belongs and the segment sequence between the two nodes according to the aforementioned topology path selection rules, and uses the sum of the lengths of each segment as the effective path length; if a unique propagation path cannot be determined, cross-node distance positioning is not performed, and the platform falls back to topology interval positioning.

[0097] The topology interval location outputs the line segment between the first boundary node and the second boundary node. In one implementation, the platform compares the anomaly confirmation results of each node along the topology path corresponding to the candidate fault interval, according to the node arrangement order, and determines the adjacent nodes where the anomaly confirmation results change as the boundary nodes of the fault interval; when there are multiple changing positions, a set of boundary nodes matching the current fault event is selected based on phase, trigger time, and waveform direction characteristics. If a unique boundary combination cannot be determined, the current candidate fault interval is retained and a verification mark is output.

[0098] Single-node waveform analysis is not used to generate distances along the line across nodes. In one implementation, the platform selects a pre-trigger baseline window and a post-trigger analysis window, and calculates waveform amplitude characteristics, RMS characteristics, or frequency characteristics, respectively; when the change in at least one characteristic of the post-trigger analysis window relative to the baseline window reaches a corresponding preset threshold, an anomaly confirmation result is output. The platform can also output the fault nature based on at least one of waveform polarity, wavefront direction, and frequency distribution.

[0099] In step S110, the early warning output unit 58 generates positioning result 8 and early warning information. Positioning result 8 at least saves the fault event identifier, waveform identifiers involved in positioning, boundary nodes, and positioning result level; when forming distance positioning, it also saves the reference node, line direction, distance along the line, and adjacent towers.

[0100] The platform queries the node mileage table based on reference nodes, line direction, and distance along the line, mapping continuous distances to nearby poles. Warning information can include the line in question, fault phase, location result level, two boundary nodes, reference node, line direction, distance along the line, and nearby poles. If distance positioning cannot be obtained, only established intervals or anomaly confirmation information are output, without filling in speculative distances.

[0101] Example 5: Selective Processing of Delayed Waveforms

[0102] like Figure 5 As shown, in step T101, the late waveform processing unit 59 obtains the line identifier, physical event time range, original candidate fault interval, original location result level, original location path and original location result of the initially closed fault event 7.

[0103] Step T102: Receive a new waveform record; Step T103: Compare the line identifier and terminal time of the new waveform. If the line identifiers match and the terminal time is within the original physical event time range or the continuous triggering range, it is identified as a late waveform record. The platform's received time is only used to confirm that the waveform arrived after the initial closure.

[0104] In step T104, after the late waveform forms a callable waveform record, it can be added to the temporary event member set, and the node combination and positioning result level judgment can be re-executed. If the newly added waveform forms a new cross-node arrival time difference positioning combination, replaces at least one boundary node, or improves the positioning result level, it is considered to have changed the current positioning structure.

[0105] The temporary event member set does not directly overwrite the original event result. The platform first generates a trial result according to the same purpose eligibility and path selection rules as the initial closure, and then compares the trial result with the original result item by item. In this way, whether a late waveform triggers relocation depends on the change in the result structure, rather than on the waveform arriving later than a certain fixed time.

[0106] In step T105, when the late waveform changes the positioning structure, the event node set is updated, the candidate fault range, the positioning result level and the positioning path are redefined, and the original positioning result, late waveform identifier, update reason and updated positioning result can be saved.

[0107] In one implementation, the updated event log can simultaneously retain the original result version, the late waveform identifier used in the trial calculation, the new result version, and the reason for the update. When subsequent expert reviews or on-site reviews modify the event labels, the original labels and review conclusions can be retained.

[0108] In step T106, if the delayed waveform does not change the positioning structure, it is added as a supplementary record to the original fault event, maintaining the original positioning path and result. The platform compares the positioning result level, boundary nodes, reference nodes, direction, and distance in the structured positioning result; if the positioning result level, boundary nodes, reference nodes, or line direction changes, the positioning structure is considered to have changed; only if the distance along the line changes, and the change in distance reaches the preset distance resolution or warning output accuracy, is the positioning structure considered to have changed. If the above fields do not change to meet the conditions, the platform can maintain the original warning result and determine whether to repeatedly output the same warning information based on the warning deduplication settings.

[0109] Through the above processing, while retaining the delayed waveform, the fault event is repositioned only for records whose positioning structure can be changed.

[0110] The above embodiments are used to illustrate the technical solution of the present invention. The data structure, communication method, and deployment location of the program modules of the node topology can be adjusted according to the actual system, but the relationship between the target node contribution judgment, waveform processing status identification, and fault event closure control remains unchanged.

[0111] The above embodiments are used to illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

Claims

1. A transmission line condition assessment and early warning system based on cloud-edge collaboration, characterized in that, It includes edge monitoring terminals installed at multiple monitoring nodes of the transmission line and a cloud-based status assessment and early warning platform that is communicatively connected to the edge monitoring terminals; The edge monitoring terminal is used to generate and upload waveform records associated with line identifiers, node identifiers, terminal time, and waveform identifiers; The cloud-based status assessment and early warning platform is used to: establish a line node topology and associate waveform records with fault events; determine candidate fault intervals and the current positioning result level based on the callable waveform records and the line node topology, and identify candidate nodes that meet at least one of the following positioning contribution conditions as target nodes: located within the candidate fault interval and between the node corresponding to the callable waveform record and the boundary node; located on a different topology side from the node corresponding to the callable waveform record and the corresponding terminal has synchronous acquisition capability; its topological location and the acquisition capability of the corresponding terminal meet positioning input conditions higher than the current positioning result level. Based on the reception start information, reception end information, and whether a callable waveform record has been formed, determine the reception processing status of the expected reception item or waveform record corresponding to the target node, and keep the fault event enabled when receiving or when receiving is completed but a callable waveform record has not yet been formed; If there is no target node in either of the above two states, and the currently available waveform record supports at least one location path, the fault event is initially closed. The location path is selected based on the topological location of the node corresponding to the available waveform record, and the location result and warning information are generated.

2. The transmission line condition assessment and early warning system based on cloud-edge collaboration according to claim 1, characterized in that, The location contribution conditions are determined according to the following rules: When a candidate node is located within a candidate fault interval and is between a node corresponding to a callable waveform record and a boundary node, the candidate node is determined as a boundary re-judgment node, and the boundary is re-determined according to a preset boundary judgment rule after it forms a callable waveform record; when a candidate node and at least one node corresponding to a callable waveform record are located on different topology sides, and the corresponding terminal has the ability to synchronously acquire the target waveform type, the candidate node is determined as a cross-node combination node; when the current location result level is anomaly confirmation and the topological position of the candidate node meets the input conditions for forming two boundary nodes, or when the current location result level is interval positioning and the candidate node meets the cross-node combination conditions, the candidate node is determined as a level improvement node.

3. The transmission line condition assessment and early warning system based on cloud-edge collaboration according to claim 1, characterized in that, The cloud-based status assessment and early warning platform establishes expected reception items for target nodes that have not yet formed actual waveform records, and determines the reception processing status of the expected reception items or waveform records according to the following rules: when no reception start information matching the target node and fault event is obtained, it is determined that reception has not yet started; when reception start information has been obtained but reception end information has not been obtained, it is determined that reception is in progress; when reception end information has been obtained but a callable waveform record has not yet been formed, it is determined that reception is completed but a callable waveform record has not yet been formed; when the waveform record can be called by fault location processing, it is determined that a callable waveform record has been formed.

4. The transmission line condition assessment and early warning system based on cloud-edge collaboration according to claim 3, characterized in that, The cloud-based status assessment and early warning platform determines the physical event time range based on the terminal time of the waveform record associated with the fault event. When the physical event time range ends and each target node satisfies any of the following conditions, the fault event is initially closed: a callable waveform record has been formed; it is determined to be unavailable; it is in a state where reception has not yet started and a preset waiting end condition is met; wherein, target nodes in a state of receiving or receiving but not yet forming a callable waveform record are not determined to be unavailable.

5. The transmission line condition assessment and early warning system based on cloud-edge collaboration according to claim 1, characterized in that, The waveform record is also associated with a time synchronization state and a waveform integrity state. When the time synchronization state meets the preset synchronization conditions and the waveform integrity state indicates that the waveform is complete, the waveform record is qualified for time difference of arrival (TDOA) positioning and waveform feature analysis. When the time synchronization state does not meet the preset synchronization conditions, but the waveform integrity state indicates that the waveform is complete, the waveform record is only qualified for waveform feature analysis. When the waveform integrity state indicates that the waveform is incomplete, the waveform record serves as a verification record and is not used for cross-node TDOA positioning.

6. The transmission line condition assessment and early warning system based on cloud-edge collaboration according to claim 5, characterized in that, The cloud-based status assessment and early warning platform selects the location path according to the following rules: when at least two callable waveform records with time difference of arrival (TDOA) positioning qualifications are located on different topological sides of the candidate fault interval, cross-node TDOA positioning is selected; when the callable waveform records of multiple nodes can determine the boundary of the candidate fault interval but do not meet the aforementioned conditions, topological interval positioning is selected; when only one node's callable waveform record has waveform feature analysis qualifications, and other target nodes are determined to be unavailable, or are in a state that has not yet started receiving and meet the preset waiting termination conditions, single-node waveform analysis is selected.

7. The transmission line condition assessment and early warning system based on cloud-edge collaboration according to claim 1, characterized in that, When the cloud-based status assessment and early warning platform receives a new waveform record after the initial closure of a fault event, it determines whether the new waveform record belongs to the fault event based on the line identifier and terminal time of the new waveform record. After the late waveform record belonging to the fault event is formed into a callable waveform record, if the late waveform record, when added to the fault event, can form a new cross-node arrival time difference positioning combination, replace at least one boundary node of the candidate fault interval, or improve the positioning result level, then the positioning path and positioning result are updated; otherwise, the late waveform record is used as a supplementary record and the original positioning result remains unchanged.

8. A method for assessing and warning the condition of transmission lines based on cloud-edge collaboration, characterized in that, The process, executed by the cloud-based status assessment and early warning platform, includes: acquiring the topology of line nodes and waveform records uploaded by edge monitoring terminals, and associating the waveform records with fault events; determining candidate fault intervals and the current positioning result level based on the callable waveform records and the topology of line nodes, and identifying candidate nodes that meet at least one of the following positioning contribution conditions as target nodes: located within the candidate fault interval and between the node corresponding to the callable waveform record and the boundary node; located on a different topology side from the node corresponding to the callable waveform record and the corresponding terminal has synchronous acquisition capability; its topological location and the acquisition capability of the corresponding terminal meet positioning input conditions higher than the current positioning result level; determining the expected reception item or waveform record reception processing status corresponding to the target node based on reception start information, reception end information, and whether a callable waveform record has been formed, and keeping the fault event open when receiving or receiving but not yet forming a callable waveform record; when there are no target nodes in the above two states, and the current callable waveform record supports at least one positioning path, initially closing the fault event, selecting a positioning path based on the topological location of the node corresponding to the callable waveform record, and generating positioning results and early warning information.

9. The method for power transmission line condition assessment and early warning based on cloud-edge collaboration according to claim 8, characterized in that, The location contribution conditions are determined according to the following rules: when a candidate node is located within a candidate fault interval and is between a node corresponding to a callable waveform record and a boundary node, the candidate node is determined as a boundary re-judgment node; when a candidate node and at least one node corresponding to a callable waveform record are located on different topology sides, and the corresponding terminal has the ability to synchronously acquire the target waveform type, the candidate node is determined as a cross-node combination node; when the current location result level is anomaly confirmation and the topological position of the candidate node meets the input conditions for forming two boundary nodes, or when the current location result level is interval location and the candidate node meets the cross-node combination conditions, the candidate node is determined as a level improvement node.

10. A power transmission line condition assessment and early warning device based on cloud-edge collaboration, characterized in that, It includes a communication interface, a processor, and a memory, wherein the memory stores a computer program; when the processor executes the computer program, it causes the device to execute the power transmission line status assessment and early warning method based on cloud-edge collaboration as described in claim 8.

Images