Key process-oriented power grid operation behavior risk intelligent identification and early warning system

Abstract

The application discloses a power grid operation behavior risk intelligent identification and early warning system for key processes, and relates to the technical field of power grid operation safety management. The system first acquires multi-source operation data of the execution state of the target key process in the power grid operation area, and constructs an operation event set with time sequence continuity and process correlation through spatial region mapping processing or process stage coupling processing. Then, the interaction change relationship between operation behavior parameters is analyzed to identify the behavior misfit risk in the execution process of the key process. Further, the behavior linkage state in the execution process of the target key process is analyzed according to the corresponding cooperative misfit state of the behavior misfit risk, and the corresponding behavior abnormality identification result is generated. Finally, process rollback early warning or current process review early warning is executed, and the corresponding behavior risk early warning information is output. The continuous identification and hierarchical early warning of the key process behavior risk are realized.

Description

Technical Field

[0001] This invention relates to the field of power grid operation safety management technology, and in particular to an intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes. Background Technology

[0002] With the continuous expansion of the power grid, the increasing variety of work types, and the growing complexity of on-site operations, the personnel activities, operational behaviors, and process coordination involved in power grid operations are exhibiting more complex and changing characteristics. Especially during critical processes such as voltage testing, power outages and restorations, maintenance, live-line testing, emergency repairs, and equipment inspections, personnel must complete corresponding actions according to established operating procedures, safety regulations, and operational requirements. Deviations in action execution, incorrect sequence of actions, disordered coordination, or abnormal action trajectories can easily lead to risks such as electric shock, accidental contact with live parts, equipment malfunction, short circuits, and work interruptions. Therefore, accurate identification and timely early warning of risks in power grid operations are crucial for ensuring operational safety and improving on-site risk control capabilities. Existing technologies typically rely on camera equipment, positioning devices, 3D modeling systems, or on-site sensors to model and collect data on the power grid operation area, equipment distribution, and personnel positions. They also identify personnel movement trajectories, entry into the operation area, operational actions, and equipment operating status, and assess on-site risks by combining these with preset rules or risk models.

[0003] For example, Chinese invention patent CN119558659B discloses a method for intelligent positioning and risk management in power operation environments based on a 3D model. By constructing a 3D model of the power operation environment, marking the positions of workers and equipment in real time, recording and predicting personnel trajectories, identifying non-standard behaviors, and combining sensor data for risk analysis, it achieves comprehensive assessment and early warning of risks at power operation sites. It can play a certain role in identifying obvious boundary violations, abnormal actions, or abnormal states. However, in continuous process scenarios, the risk of operational behavior is not only reflected in whether a single behavior is abnormal, but also in whether the timing of the behavior is reasonable, whether the corresponding equipment state matches, and whether the related state response is adequate.

[0004] Taking a substation bay power outage maintenance scenario as an example, voltage testing should be conducted only after power outage confirmation is completed, grounding wire installation should follow voltage testing, and maintenance should be carried out only after safety measures are in place. Furthermore, monitoring, operation, and coordination must remain consistent around the same equipment and timeframe. Any discrepancies in the data status of key processes, whether in connection, correspondence, or coordination, could lead to continuous risks. In this context, risk identification involves not only the time synchronization of multi-source data but also comprehensive analysis of the statuses of different processes.

[0005] Therefore, existing methods do not sufficiently analyze the correlation between relevant work behaviors and process states, resulting in an incomplete characterization of potential behavioral risks caused by mismatch in the correlation between behavior execution states. This, in turn, affects the effectiveness of risk identification and early warning, and restricts the application effect of the intelligent safety supervision system in the refined supervision and closed-loop risk management of the entire work process. Summary of the Invention

[0006] Therefore, embodiments of the present invention provide an intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes. The technical solution is as follows: The multi-source process data coupling module is used to acquire multi-source operation data corresponding to the execution status of the target key process within the power grid operation area, so as to perform spatial region mapping processing or process stage coupling processing to obtain a set of operation events with temporal continuity and process correlation; the target key process includes at least the voltage testing operation process, the power outage and restoration operation process, and the maintenance operation process.

[0007] The process behavior correlation analysis module is used to analyze the interaction and change relationship between the execution status and the corresponding work behavior parameters based on the set of work events, in order to identify the risk of behavioral mismatch during the execution of the target key process.

[0008] The linkage status anomaly identification module is used to analyze the linkage status of the target key process in the execution process based on the collaboration mismatch status corresponding to the behavior mismatch risk, and generate the corresponding behavior anomaly identification results; the behavior linkage status includes at least one or more of the following: the time sequence relationship of the operation behavior, the corresponding relationship of the operation subject, and the matching relationship of the operation status.

[0009] The behavioral risk warning output module is used to issue warnings for process rollback or current process verification based on the results of abnormal behavior identification, and to output corresponding behavioral risk warning information when the preset warning conditions are met.

[0010] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: 1. This invention first acquires multi-source operation data corresponding to the execution status of key operations such as voltage testing, power outages / restorations, and maintenance within the power grid operation area through a multi-source process data coupling module. This data is then processed through spatial region mapping or process stage coupling to form a set of operation events with temporal continuity and process correlation. This unifies the previously scattered personnel, equipment, tools, and process stage data into a single key process chain, addressing the problem that existing methods can only make judgments based on a single moment, location, or event, making it difficult to comprehensively represent the connection relationship between preceding and following processes. Secondly, the process behavior correlation analysis module analyzes the interactive changes between operation behavior parameters in the set of operation events to identify behavioral mismatch risks, improving the ability to identify mismatches in the execution correlation status of key processes. Furthermore, the linkage status anomaly identification module analyzes the behavioral linkage status during the execution of the target key process based on the collaborative mismatch status corresponding to the behavioral mismatch risk and generates behavioral anomaly identification results. This helps to discover continuity risks caused by connection deviations, corresponding deviations, or collaborative deviations in the data status of preceding and following key processes. Finally, the behavioral risk warning output module executes process rollback warnings or current process verification warnings based on the behavioral anomaly identification results, and outputs corresponding behavioral risk warning information, thereby improving the pertinence of risk identification results and the effectiveness of warning outputs, and thus enhancing the application effect of the intelligent safety supervision system in the refined supervision and closed-loop risk control of the entire operation process.

[0011] 2. First, spatial operation data corresponding to the target key processes is acquired. Combined with equipment location data, area boundary data, and the spatial correspondence of personnel and tools, the target process spatial areas corresponding to key processes such as voltage testing, power outage / restoration, and maintenance are determined. This generates spatial mapping events and forms a spatial chain event set, enhancing the ability to represent the spatial attribution relationship of operational behaviors and the continuity of process attribution. Further, stage operation data corresponding to the target key processes is acquired, and stage coding is performed based on the dependencies between different key processes. Then, the process stage identifier is updated through bit addition or subtraction processing to determine the stage connection relationship between different target key processes and form a process connection event set. This allows for the continuous depiction of the forward / backward progression relationship, entry state, completion state, and regress state between processes such as voltage testing, power outage / restoration, and maintenance, enhancing the ability to describe changes in process stages and the correlation of process progression. This is beneficial for improving the accuracy and continuity of risk identification for key process behaviors.

[0012] 3. First, according to the preset time window and the process affiliation relationship corresponding to the target key process, adjacent operation events in the spatial chain event set or process connection event set are sorted and paired to generate candidate related event combinations. The starting process in the current process chain is excluded from the candidate related event combinations to avoid events without prior support directly participating in the association judgment. Further, the binary state corresponding to each candidate related event combination is read, and the current event is divided into normal candidate events or abnormal candidate events based on the binary state. This distinguishes events that meet the basic requirement of effective completion of the preceding event from events with reverse execution or connection anomalies, improving the processing efficiency of process behavior association analysis. For abnormal candidate events, the behavior object data, behavior state data, behavior time data, and behavior stage data are further read, along with the preceding object data, preceding completion state data, preceding completion time data, and preceding stage data of the corresponding preceding event. These are then summarized to form the process behavior adaptation result, enabling comprehensive adaptation analysis of abnormal candidate events from four dimensions: object, state, time sequence, and stage. Finally, based on the process behavior adaptation results, it is determined whether there is a risk of behavioral mismatch in the abnormal candidate events. This enables a more comprehensive identification of potential risks caused by object mismatch, state disconnection, time sequence discrepancy, or stage correspondence errors, thereby improving the accuracy and reliability of key process behavior risk identification.

[0013] 4. First, in the linkage status anomaly identification module, abnormal candidate events corresponding to behavioral mismatch risks are classified and processed: For abnormal candidate events not associated with adjacent work events, they are recorded as independent abnormal candidate events, and basic anomaly identification results are generated based on process affiliation information, current status information, and preset process chain information. This allows for independent judgment of isolated anomalies lacking support from adjacent work events. For abnormal candidate events associated with adjacent work events, the behavior occurrence time parameter, behavior subject identification parameter, and behavior status parameter are further extracted, and the time sequence offset, subject corresponding deviation, and status matching deviation are determined respectively. This allows for quantitative analysis of the linkage relationship in the execution of key processes from three dimensions: time sequence, subject, and status. Subsequently, based on the satisfaction of each parameter and preset conditions, corresponding behavioral anomaly identification results are generated, and process rollback warnings or current process verification warnings are further triggered, thereby improving the accuracy and closed-loop nature of risk identification and warning handling in key processes. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 A schematic diagram of the structure of the intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the timing chain of the target key process provided in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the construction of a spatial homing event set provided in an embodiment of the present invention; Figure 4 This is a logic diagram for identifying behavioral mismatch risks provided in embodiments of the present invention; Figure 5 A trend chart of state matching deviation provided for embodiments of the present invention; Figure 6 A comparison chart of process rollback judgment parameters and process verification judgment parameters provided in this embodiment of the invention; Figure 7 This is a time sequence diagram of 24-hour fault events in different substations provided in an embodiment of the present invention; Figure 8 A fault type distribution diagram provided for embodiments of the present invention; Figure 9 A heatmap showing the correlation between voltage, current, power, and energy consumption provided for embodiments of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0017] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0018] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0019] like Figure 7 The figure shows the time-series variation of fault events at different substations over 24 hours. The horizontal axis represents time (hours), and the vertical axis represents the number of fault events. The three broken lines correspond to substations 001, 002, and 003, respectively. It can be seen that there are significant fluctuations in the number of fault events at different stations throughout the day, with relatively high values ​​appearing in some periods. This reflects the time-series variation characteristics of different stations in terms of operating load, work activities, or risk exposure. Figure 8The diagram shows a distribution of fault types. The number of fault-free events is significantly higher than that of abnormal events such as overload, undervoltage, and power outages, indicating that most operating periods are in a normal state, while abnormal events are mainly concentrated in a few fault categories. Figure 9 The figure shows a heatmap of the correlation between voltage, current, power and energy consumption. The correlation between current, power and energy consumption is relatively high, while the correlation between voltage and the other parameters is relatively weak.

[0020] As illustrated in the diagrams above, in substation bay outage maintenance scenarios, key processes such as voltage detection, power restoration / outage, and maintenance are not only affected by the on-site work itself, but also correlated with the substation's operating status, fault type distribution, and changes in electrical parameters. Analyzing only a single location, action, or state makes it difficult to comprehensively determine which key process the current action belongs to, or whether the current process is based on the effective completion of preceding processes. Therefore, it is necessary to combine work behavior analysis with runtime sequence information, fault type information, and parameter correlation information to provide more comprehensive data support for key process behavior risk identification, linkage anomaly judgment, and early warning output. Based on this, this invention provides an intelligent identification and early warning system for power grid work behavior risks oriented towards key processes, such as... Figure 1 The diagram shown illustrates the structure of a power grid operation behavior risk intelligent identification and early warning system for key processes. This system may include: Multi-source process data coupling module: Used to perform spatial region mapping processing in scenario 1 to determine the spatial attribution relationship between work actions and target key processes such as voltage testing, power outage and restoration, and maintenance; and to perform process stage coupling processing in scenario 2 to determine the stage connection relationship and process advancement relationship between different target key processes. For example... Figure 2 The diagram illustrates the sequence chain of key target processes. These key processes, arranged sequentially along the timeline, include power outage confirmation, voltage testing, grounding, safety measures, maintenance, and recovery operations. Each process is connected by a forward arrow, representing the stage progression under normal operating conditions. Specifically, after power outage confirmation, voltage testing begins; after voltage testing, grounding and related power outage / restoration operations follow; maintenance can only commence after safety measures are implemented; and recovery operations are performed after maintenance is completed. The dashed arrows in the diagram indicate abnormal connections such as reverse order, jumps, or rollbacks, reflecting potential issues like incorrect process sequence, premature triggering, or rollback execution when key processes do not follow the preset sequence chain.

[0021] Scene 1 First, spatial operation data corresponding to the target key processes is acquired. This data can be obtained through video acquisition equipment, personnel positioning equipment, equipment object databases, and area boundary configuration files. Personnel location data can be obtained from on-site video monitoring equipment combined with image recognition and positioning algorithms, or from ultra-wideband (UWB) positioning tags, Bluetooth positioning tags, or BeiDou positioning terminals worn on safety helmets, work badges, or belts. For example, at a certain moment, worker A's location coordinates are (12.4, 8.1), and worker B's location coordinates are (15.2, 6.9). Equipment object location data can be obtained from the substation's primary equipment ledger, GIS element coordinates, or the substation's 3D model. For example, the center coordinates of the No. 1 line bay disconnect switch equipment are (12.0, 8.3), and the center coordinates of the grounding disconnect switch equipment are (15.0, 7.1). Tool location data can be obtained through video recognition, visual tag recognition, or tool-carrying terminal positioning. For example, the coordinates of the voltage detector are (12.1, 8.0), the coordinates of the grounding clamp are (14.8, 7.0), and the coordinates of the maintenance toolbox are (13.9, 5.8). Area boundary data can be obtained through in-station electronic fences, polygonal area markings, or preset coordinate boundaries. For example, the voltage testing area can be set as a circular area with a radius of 1.2 meters centered on the center point of the disconnect switch, the power outage and restoration area can be set as a circular area with a radius of 1.0 meter centered on the center point of the grounding disconnect switch, and the maintenance area can be set as a rectangular boundary [10, 16] × [5, 9]. Process space identification data is used to distinguish which key process different spatial areas belong to. For example, YD-01 represents the voltage testing space identification, TSD-02 represents the power outage and restoration space identification, and JX-03 represents the maintenance space identification.

[0022] After acquiring the spatial operation data, the target process spatial area corresponding to the target key process is determined based on the equipment object location data and area boundary data. For example, using the center coordinates (12.0, 8.3) of the No. 1 line bay switch as a reference, and combining the effective operating range and safety distance requirements of the voltage detector, the voltage detection process spatial area is expanded by 1.2 meters in all directions; using the center coordinates (15.0, 7.1) of the grounding switch as a reference, and combining the typical operating space of the grounding clamp and operating rod, the power outage and restoration process spatial area is expanded by 1.0 meter; and the maintenance process spatial area is formed by the boundary of the maintenance work surface, so that different key processes each have an identifiable spatial range.

[0023] Next, before determining the spatial attribution relationship between the current operation and the corresponding key target process based on the spatial correspondence between the operator location data, tool location data and the target process spatial area, a unified coordinate system within the station can be constructed. For example, the southwest corner reference point of the area where the No. 1 line bay of the substation is located can be used as the origin of the coordinate system, with the horizontal direction as the X-axis and the vertical direction as the Y-axis, to establish a two-dimensional plane coordinate system, and the location data of operators, tools and equipment objects can be uniformly mapped to this coordinate system.

[0024] Based on this, the corresponding target process space area is determined according to the center coordinates of the equipment object and the boundary range of the area. For example, the voltage testing process space area is constructed with the coordinates of the disconnect switch object (12.0, 8.3) as the center and a radius of 1.2 meters; the power outage and restoration process space area is constructed with the coordinates of the grounding disconnect switch object (15.0, 7.1) as the center and a radius of 1.0 meter; and the maintenance process area is constructed with the boundary of the maintenance work surface. Subsequently, the position data of the operators and the position data of the tools are substituted into a unified coordinate system for spatial correspondence analysis. For example, when the position of operator A is (12.4, 8.1) and the position of the voltage tester is (12.1, 8.1), the corresponding spatial correspondence analysis is performed. When both the current action and the grounding clamp are located within the voltage testing process space centered on the disconnector (8.0), and the distance between them and the disconnector's coordinates (12.0, 8.3) is less than 1.2 meters, it can be determined that there is a spatial correspondence between the current action and the voltage testing process, and its spatial classification is determined as the voltage testing operation process. When the position of worker B is (15.2, 6.9) and the position of the grounding clamp is (14.8, 7.0), and both are located within the power outage / restoration process space, the current action can be classified as the power outage / restoration operation process. If a person carries maintenance tools into the maintenance process area, the action can be classified as the maintenance process. Conversely, if worker A carries a voltage detector but is located within the maintenance process area, or if the distance between the tool and the target equipment exceeds the preset range, the spatial correspondence is not established.

[0025] After determining the spatial attribution relationships, the spatial operation data is categorized and integrated to generate corresponding spatial mapping events. For example, at 08:31:12, the location of worker A, the location of the voltage detector, the location of the disconnect switch, and the voltage detection area identifier together constitute a voltage detection spatial mapping event; at 08:33:45, the location of worker B, the location of the grounding clamp, the location of the grounding disconnect switch, and the power outage / restoration area identifier together constitute a power outage / restoration spatial mapping event; and at 08:38:20, the location of worker C, the location of the maintenance tool, the location of the maintenance object, and the maintenance area identifier together constitute a maintenance spatial mapping event. Subsequently, according to the occurrence time sequence and process attribution sequence of the spatial mapping events, each event is continuously arranged to form a spatial chain event set with spatial location association characteristics and continuous process attribution characteristics.

[0026] like Figure 3 The schematic diagram illustrating the construction of a spatially linked event set demonstrates the spatial division of different target key processes and the distribution of their associated objects in a power outage maintenance scenario within the same substation bay. The area within the dashed box on the left represents the voltage testing space, containing disconnect switches, worker A, and a voltage detector, indicating that voltage testing activities revolve around the corresponding equipment and tools. The area within the dashed box on the right represents the power outage / restoration space, containing grounding switches, worker B, and an operating lever, indicating that power outage / restoration-related activities revolve around the corresponding equipment and operating tools. The area within the dashed box at the bottom represents the maintenance space, containing maintenance equipment and tools, indicating that maintenance activities revolve around the maintenance object. The monitoring personnel in the upper middle section are located outside multiple process areas, representing the on-site monitoring relationships for different key processes. By classifying and integrating the distribution of workers, tools, and equipment in different process spatial areas, voltage testing spatial mapping events, power outage / restoration spatial mapping events, and maintenance spatial mapping events can be formed. These can then be further structured into a spatially linked event set according to their occurrence time sequence and process affiliation sequence.

[0027] Through the spatial region mapping processing in Scenario 1, the data on personnel, equipment, tools, and area boundaries collected in a scattered manner during the substation interval power outage maintenance process can be uniformly converted into a set of spatially linked events corresponding to the power testing operation, power outage and restoration operation, and maintenance operation.

[0028] Scene 2 First, acquire the stage operation data corresponding to the target key process. Stage operation data can be obtained from multiple sources. Process stage status data can be generated jointly by electronic work tickets, mobile work terminals, safety monitoring terminals, and video recognition results. For example, it can identify status signals such as voltage testing starting, grounding wire approaching, and maintenance tools entering at a certain moment. Process switching indicator data can be obtained from process switching button records, voice command recognition results, or terminal click records. For example, the terminal records switching from power outage confirmation to voltage testing, and from voltage testing to grounding wire installation. Process completion status data can be obtained through work ticket confirmation information, manual terminal confirmation results, video completion judgment results, or sensor signal trigger results. For example, power outage confirmation = completion, voltage testing = completion. Process rollback status data can be obtained through scheduling commands, monitoring confirmations, error correction commands, or terminal rollback operation records. For example, maintenance rollback to voltage testing confirmation. In a real-world scenario, examples could be given such that: 08:25:10, the operation terminal uploads confirmation of power outage completion; 08:31:12, video recognition and terminal recording together indicate that the power verification has entered the execution state; 08:33:45, the grounding clamp enters the vicinity of the target equipment, and the terminal records preparation for grounding wire installation; 08:38:20, maintenance tools enter the work area, and the maintenance record indicates that the maintenance has entered the pending execution state. These records collectively constitute the stage operation data.

[0029] After acquiring the phased operation data, it is necessary to perform phase coding based on the sequential dependencies between the voltage testing, power outage / restoration, and maintenance operations to form phase identifiers. The sequential dependencies here mean that, in a normal substation interval power outage maintenance scenario, power outage confirmation should precede voltage testing, voltage testing should precede power outage / restoration or grounding wire installation, and maintenance can only begin after power outage / restoration and safety measures are implemented. For example, the voltage testing operation can be coded as 01, the power outage / restoration operation as 02, and the maintenance operation as 03; simultaneously, the entry status can be coded as 1, the execution status as 2, the completion status as 3, and the rollback status as 4. Thus, voltage testing entry can be represented as 01-1, voltage testing completion as 01-3, power outage / restoration entry as 02-1, and maintenance rollback as 03-4. In this way, the phased operation data, originally existing in text or log format, is uniformly converted into continuously comparable phase identifiers.

[0030] After completing the stage coding, the process stage identifier is incremented or decremented to update the entry, execution, completion, or rollback status of the corresponding target key process. The incrementing operation can be understood as: when a key process is detected to have entered a new stage, an increment or decrement operation is performed on the corresponding stage status bit; the decrement operation can be understood as: when a key process is detected to have exited the current stage, needs to roll back, or has its entered state revoked, a decrement or clear operation is performed on the corresponding stage status bit. For example, the initial status bits of the voltage testing process can be set to 0000, where the first bit represents the entry status, the second bit represents the execution status, the third bit represents the completion status, and the fourth bit represents the rollback status. When the voltage testing starts at 08:31:12, the entry status bit is set from 0 to 1, forming 1000; when the voltage testing continues at 08:32:40, the execution status bit is set from 0 to 1, forming 1100; and when the voltage testing is confirmed to be completed at 08:33:20, the completion status bit is set from 0 to 1, forming 1110.

[0031] If the start of the power outage / restoration process is detected at 08:33:45, the status bit of the power outage / restoration process can be updated from 0000 to 1000, indicating that its entry state has been established. If the power outage / restoration process is detected to have entered the execution state before the power verification is completed, the status bit of the power outage / restoration process can be directly updated from 0000 to 1100, while the completion status bit corresponding to the power verification process remains 0. This indicates that the power outage / restoration process has been prematurely initiated. Conversely, if the maintenance / repair process has already entered the execution state, for example, with its status bit at 1100, and then a rollback command to return to re-verify is received, the entry and execution status bits of the maintenance / repair process can be cleared to 0 or decremented by 1, causing it to revert from 1100 to 0000. At the same time, the rollback status bit of the power verification process can be set to 1 or incremented by 1 to indicate that it has returned to the power verification confirmation stage.

[0032] Through the above bit addition and bit subtraction processing, the progress, completion and rollback of key processes at different stages can be transformed into continuously updatable stage status results, which makes it easier to identify situations such as process advance, abnormal jump or process rollback.

[0033] After the status update is completed, the stage connection relationships between different target key processes are determined based on the updated process stage identifiers. Here, the stage connection relationship is used to determine whether the current key process and its predecessor key process meet the effective handover conditions. For example, if power outage confirmation is completed, power verification is completed, and power interruption / restoration is initiated, it indicates that the power interruption / restoration process is based on the completion of power verification, and the stage connection relationship is valid. If power outage confirmation is completed, power verification is in progress, and power interruption / restoration is initiated, it indicates that the power interruption / restoration process is not based on the effective completion of power verification, and the stage connection relationship is abnormal. If maintenance is initiated but safety measures are not completed, it indicates that the maintenance process has a stage jump or is being advanced prematurely. Therefore, stage connection relationships can be categorized as normal progress relationships, advanced progress relationships, waiting relationships, jump relationships, or rollback relationships.

[0034] Subsequently, the stage operation data is integrated based on the stage connection relationships to form a set of process connection events. This integration process can be understood as: linking stage events on the same key process chain according to time and dependency relationships. For example, integrating "power outage confirmation completed," "power testing begins," "power testing completed," "power outage / restoration begins," "power outage / restoration completed," and "maintenance / repair begins" into a continuous process connection chain; if an anomaly occurs at any step, an abnormal connection event is marked at the corresponding location. The resulting set of process connection events includes both normally progressing process chains and those that are advanced, regressed, or skipped, thus possessing both stage status connection characteristics and process progression correlation characteristics.

[0035] By using the process stage coupling processing in Scenario 2, the process advancement data, switching data, completion data, and rollback data recorded in a scattered manner during the substation interval power outage maintenance process can be organized into a set of process connection events with continuous stage meaning. This clarifies whether the current behavior is based on the effective completion of the preceding process, and whether there are situations such as early advancement, stage jump, execution interruption, or process rollback between different key processes.

[0036] Preferably, when it is necessary to perform correlation analysis based on the spatial affiliation relationship between operators, tools, and equipment to determine whether the current behavior belongs to the voltage testing, power outage / restoration, or maintenance work process, the spatial affiliation event set formed in Scenario 1 can be invoked. When it is necessary to perform correlation analysis based on the sequential progression, completion, and stage succession relationships between key processes to determine whether the current behavior is based on the effective completion of the preceding process, the process connection event set formed in Scenario 2 can be invoked. If both conditions are met simultaneously, the spatial affiliation event set and the process connection event set are invoked synchronously to jointly analyze the process affiliation and stage succession relationships of the current behavior. For example, based on the maintenance work ticket and on-site procedures, grounding can be allowed only after voltage testing is completed, and maintenance can only begin after grounding is completed. When it is simultaneously detected that an operator is carrying a voltage tester in the voltage testing space area, and the corresponding stage marker indicates that voltage testing has not yet been completed, it can be determined that the behavior meets the spatial affiliation conditions but does not meet the stage succession conditions.

[0037] Specifically, firstly, adjacent work events are sorted and paired according to the preset time window and the corresponding process affiliation of the target key process to generate candidate related event combinations. The preset time window is used to limit the effective time range for the current process event and the preceding process event to participate in the correlation analysis, so as to avoid events with excessively long time intervals or lack of continuity being incorrectly paired. The setting method can be obtained based on the actual time interval distribution of preceding and following process events in the historical samples of similar maintenance operations. For example, the time difference range between the completion of voltage testing and the start of grounding wire installation, and between the completion of grounding and the start of maintenance, can be statistically analyzed, and the corresponding statistical interval can be used as the preset time window; it can usually be set to 60 seconds to 300 seconds. For example, the preset time window between voltage testing and grounding can be set to 120 seconds, and the preset time window between grounding and maintenance can be set to 180 seconds. The process affiliation can be obtained through maintenance process templates, work ticket flow rules, or pre-configured process chains. Candidate related event combinations include the current event corresponding to the current process and the preceding event corresponding to the preceding process that satisfies the preset process affiliation relationship with the current process. If the current process is the starting process in the current process chain, such as the power outage confirmation itself, it will not be included in the candidate related event combinations because there is no preceding process to refer to.

[0038] Next, the binary status corresponding to each candidate related event combination is read. The binary status is not the final risk conclusion, but rather a preliminary indicator of whether the preceding sequence connection is valid. Specifically, it is obtained as follows: if the current event corresponds to a process that satisfies a preset preceding process relationship, the preceding event's completion status is valid, and the current event occurs after the preceding event's completion time, then the binary status is recorded as 1; if the current event occurs before the preceding event is completed, or the preset attribution relationship between the current event and the preceding process is not valid, then the binary status is recorded as 0. When the binary status is 1, it indicates that the current event is based on the valid completion of the preceding event, the current process is in a sequential execution state, and the current event is recorded as a normal candidate event; when the binary status is 0, it indicates that the current event is not based on the valid completion of the preceding event, the current process is in a reverse execution state or an abnormal connection state, and the current event is recorded as an abnormal candidate event.

[0039] For abnormal candidate events, it is necessary to further read the behavior object data, behavior status data, behavior time data, and behavior stage data corresponding to the current event, and simultaneously read the preceding object data, preceding completion status data, preceding completion time data, and preceding stage data of the corresponding preceding event. Behavior object data can come from the equipment object database, video recognition results, or work order binding information, such as the No. 1 line bay switch, grounding switch, and maintenance workface object; behavior status data can come from work terminal status reports, equipment status interfaces, or video recognition judgments, such as voltage testing in progress, voltage testing completed, and grounding preparation; behavior time data is the event occurrence time, such as 08:31:12; behavior stage data can come from stage coding results, such as voltage testing stage recorded as 02, and power outage / restoration stage recorded as 03. Preceding event data corresponds to the object, status, completion time, and stage information of the previous process.

[0040] After reading this data, process behavior adaptation is performed to obtain object correspondence results, state connection results, timing judgment results, and stage correspondence results, which are then summarized to obtain the process behavior adaptation results. The object correspondence process involves comparing the behavior object data of the current event with the preceding object data of the previous event to see if a preset object association relationship is met. This preset object association relationship can be provided by equipment ledgers, bay topology relationships, or process templates. For example, the voltage testing object and the grounding object should belong to the same line bay or the same power outage maintenance unit. If the current event object is the grounding switch of line bay 1, and the preceding event object is the disconnector of line bay 1, then the object correspondence result is valid; if the current event object jumps to line bay 2, then the object correspondence result is invalid. The state connection process involves comparing the behavior state data of the current event with the preceding completion state data of the previous event to see if a preset state handover condition is met. For example, if the current event is grounding wire installation and the preceding event is voltage testing, then the preceding completion state must be voltage testing completed; if the preceding event is still voltage testing in progress, then the state connection result is invalid. The time sequence processing involves comparing the timing data of the current event with the completion time data of the preceding event to determine if the current event occurred after the completion of the preceding event. If the current event occurs earlier than the completion time of the preceding event, the timing judgment result is invalid. The stage correspondence processing involves comparing the timing data of the current event with the stage data of the preceding event to determine if they conform to the stage progression order in the preset work chain. For example, the power testing stage should precede the power outage / restoration stage, and the power outage / restoration stage should precede the maintenance stage. If maintenance occurs directly after power outage confirmation, the stage correspondence result is invalid.

[0041] After completing object matching, state connection, time sequence, and stage matching processing, the results of these four types of processing are summarized to form the process behavior adaptation result. If the process behavior adaptation result indicates that at least one of the following is true: object mismatch, state disconnect, time sequence mismatch, or stage mismatch, then the current abnormal candidate event is determined to have a behavior mismatch risk. If the process behavior adaptation result indicates that object matching, state connection, time sequence, and stage matching are all satisfied, then the current abnormal candidate event is determined not to have a behavior mismatch risk. By incorporating object relationships, state relationships, time relationships, and stage relationships into the analysis, the potential behavioral risks caused by process connection mismatches during substation bay power outage maintenance are more fully revealed.

[0042] Linkage status anomaly identification module: When an abnormal candidate event corresponding to a behavior mismatch risk is not associated with an adjacent work event, the abnormal candidate event can be recorded as an independent abnormal candidate event. Then, a basic anomaly identification result can be generated based on the process attribution information, current status information, and preset process chain information. Here, the process attribution information indicates which category the current event belongs to: power testing, power outage / restoration, or maintenance. The current status information indicates whether the corresponding process is in the entry, execution, completion, or rollback state when the current event occurs. The preset process chain information can be obtained from maintenance work tickets, standard operating procedure configuration tables, operation ticket templates, or on-site procedure documents. The basic anomaly identification result can be generated as follows: first, determine the position of the process corresponding to the current event in the preset process chain. If the process corresponding to the current event is not the starting process, and the system has not retrieved a corresponding preceding process event, but the current status information indicates that the process has entered the execution or completion state, then the basic anomaly identification result is generated directly. If the current event itself belongs to the starting process, such as a power outage confirmation event, then no anomaly result can be generated or it can only be marked as an observation state.

[0043] When an abnormal candidate event corresponding to a behavior mismatch risk is associated with an adjacent work event, parameter extraction is required for the current event and its preceding events. Parameter extraction can be performed from event log tables, video recognition results, personnel location records, work terminal logs, work order information, and status databases. The behavior occurrence time parameter is the specific time value of the current event, usually represented by a timestamp, such as 2026-04-13 08:31:12; the preceding event completion time parameter is the time value when the preceding event was marked as completed, such as 2026-04-13 08:30:05; both are time-based numerical values. The behavior subject identifier parameter is the identifier of the executing subject corresponding to the current event, which can be represented by employee number, work badge number, tag number, or role code, such as P-A01; the preset associated subject identifier parameter is the responsible subject identifier pre-bound in the work order or task assignment table for the current process, such as P-A01 indicating that it should be performed by personnel A, and P-B02 indicating that it should be supervised by personnel B. These parameters are discrete identifier data. Behavioral status parameters represent the execution status when the current event is triggered. They can be represented by status codes or text status values, such as voltage testing in progress, voltage testing completed, and grounding preparation in progress. Target status parameters represent the target state that the current process should achieve after the previous event is completed. For example, if the previous event is voltage testing completed, the target state of the current process could be that grounding wire installation is allowed. These parameters belong to discrete state data.

[0044] After extracting these parameters, three deviations can be determined. The timing offset represents the degree of time deviation of the current event relative to the preceding event. It is obtained by subtracting the completion time parameter of the preceding event from the occurrence time parameter of the current event. For example, if the grounding start time of the current event is 08:33:45 and the voltage testing completion time of the preceding event is 08:33:10, then the timing offset is 35 seconds. If the current event occurs before the completion time of the preceding event, the timing offset can be negative, such as -15 seconds, indicating premature triggering. The subject-corresponding deviation represents the degree of deviation between the current event's executing subject and the preset associated subject. It is obtained by comparing the behavior subject identifier parameter with the preset associated subject identifier parameter: if they match, the subject-corresponding deviation is recorded as 0; if they do not match, the subject-corresponding deviation is recorded as 1. If it is necessary to further distinguish the degree of role deviation, it can be further assigned a value of 2 or 3.

[0045] The state matching deviation represents the degree of deviation between the actual state of the current event and the target state corresponding to the completion of the preceding event. It is obtained by first reading the behavioral state parameters of the current event and the target state parameters corresponding to the completion of the preceding event, and then normalizing the calculation based on the proportions of consistent, missing, and conflicting states to obtain a state matching deviation between 0 and 1. A value around 0.1 or close to 0 indicates that the current state and the target state are basically consistent, which can be considered a normal state match. The closer the value is to 1, the more significant the state mismatch. Figure 5 It is evident that the overall level of normal samples remained low, while the level of abnormal samples increased significantly during the grounding, safety measures, and maintenance phases.

[0046] like Figure 5 The chart showing the trend of state matching deviation is illustrated. The horizontal axis represents key process stages such as power outage confirmation, voltage testing, grounding, safety measures, maintenance, and recovery operations, while the vertical axis represents the state matching deviation. The blue line represents the trend of state matching deviation for samples with normal states at each process stage, remaining at a low level overall, indicating a high degree of consistency between the current state and the preset target state under normal operating conditions. The orange line represents the trend of state matching deviation for samples with abnormal states, showing a significant increase in deviations during grounding, safety measures, and maintenance stages, exceeding the state abnormality threshold represented by the green dashed line in some stages. This chart illustrates the significant differences in state matching deviation at different stages during the advancement of key processes, especially at process nodes with strict continuity requirements, where state mismatches are more likely to occur. Trend analysis of state matching deviation can provide quantitative support for identifying abnormal state results, thereby improving the system's ability to identify risks associated with state deviations in key processes.

[0047] The preset conditions corresponding to the three deviation quantities are used to determine whether the current process is within the allowable range. The preset conditions for timing offset can be obtained from historical work records, on-site procedures, and process cycle statistics. For example, it can be stipulated that the timing offset must not be earlier than the completion time of the previous process, and that triggering within 180 seconds after completion is considered normal. The preset conditions for subject-related deviation can be obtained from work tickets, work assignment tables, and permission lists. For example, it can be stipulated that the executing subject must be equal to the preset responsible subject or belong to the set of allowed substitute subjects. The preset conditions for state matching deviation can be obtained from the process inheritance rule table. For example, it can be stipulated that after the power testing is completed, the target state should be allowed to ground; after the grounding is completed, the target state should be allowed to enter for maintenance. Based on this, if the timing offset, subject-related deviation, and state matching deviation all meet the corresponding preset conditions, a normal sequential linkage result is generated; if only the timing offset does not meet the conditions, a timing abnormal result is generated; if only the subject-related deviation does not meet the conditions, a subject abnormal result is generated; if only the state matching deviation does not meet the conditions, a state abnormal result is generated; if at least two of the three do not meet the corresponding preset conditions, a linkage abnormal result is generated. In this way, abnormal candidate events can be further refined from whether they are abnormal to what kind of abnormality they are.

[0048] Behavioral risk warning output module: When the behavioral anomaly identification result is characterized as a timing anomaly or a linkage anomaly, the current target key process is first determined to be in a process rollback pending confirmation state. This pending confirmation state indicates that the premise for continuing the current process may be insufficient, but whether a true rollback is necessary requires further judgment based on the process connection status and the process rollback determination parameters. The process connection status can be comprehensively obtained through the current process identifier corresponding to the current target key process, the preset preceding process identifier, the preceding process completion status information, and the current process trigger time information. This status is used to characterize whether the current process is based on the effective completion of the preset preceding process. The process rollback determination parameters are used to quantify the necessity of rollback of the current process relative to the preset preceding process. They are obtained by first extracting the timing offset, the preceding process completion status parameters, and the process connection relationship parameters corresponding to the current target key process, then weighting these three types of parameters with rollback coefficients and summing them.

[0049] Preferably, the completion status parameter of the preceding process and the process connection relationship parameter can both be represented by continuous normalized values ​​between 0 and 1. The completion status parameter of the preceding process is used to characterize the degree of completion of the preceding process. 0 indicates incomplete, 1 indicates complete, and the larger the value between 0 and 1, the higher the degree of completion or connection.

[0050] Weighted rollback factor processing refers to multiplying different parameters by a pre-set rollback factor to reflect the strength of each parameter's influence on the necessity of rollback in a process. For example, if the timing offset is 0.8, the completion status parameter of the preceding process is 0.3, and the process connection parameter is 0.4, the first rollback factor can be set to 0.5, the second rollback factor to 0.3, and the third rollback factor to 0.2. Then, 0.8×0.5, 0.3×0.3, and 0.4×0.2 are calculated respectively, and the three results are added together to obtain the process rollback judgment parameter. In this way, when the timing offset is large, the parameter value will increase; when the completion status of the preceding process is insufficient or the process connection is poor, the parameter value will also increase, thus more intuitively reflecting whether the current process needs to be rolled back. If the process connection status indicates that the current target key process is not based on the effective completion of the preset preceding process, and the process rollback judgment parameters meet the preset rollback conditions, then the corresponding process rollback warning information will be output, such as prompting to rollback to the power testing process for reconfirmation or to rollback to the power outage confirmation process for reverification.

[0051] When the behavior anomaly identification result is characterized as a subject anomaly or a state anomaly, the current target key process is determined to be in a process verification pending confirmation state. This "process verification pending confirmation state" indicates that the current process does not necessarily need to be rolled back, but there may be local deviations at the object, subject, or state level, requiring verification at the current process node. At this time, it is necessary to obtain the object-corresponding state, subject-corresponding state, state-matching state, and process verification judgment parameters. Specifically, the object-corresponding state indicates whether the current behavior object is consistent with the preset process object; the subject-corresponding state indicates whether the current executing subject is consistent with the preset associated subject; and the state-matching state indicates whether the current executing state is consistent with the preset target state. The process verification judgment parameters are obtained by weighting the object deviation degree, subject deviation degree, and state deviation degree corresponding to the current target key process with verification coefficients and then summing them.

[0052] The weighted processing of verification coefficients is similar to that of backoff coefficients. It first assigns different verification coefficients to different deviations, then uses multiplication and addition to obtain a comprehensive result. For example, if the object deviation is 0.6, the main body deviation is 0.4, and the state deviation is 0.7, the first verification coefficient can be set to 0.4, the second to 0.3, and the third to 0.3. Then, 0.6 × 0.4, 0.4 × 0.3, and 0.7 × 0.3 are calculated respectively, and the results are added together to obtain the process verification judgment parameters. If at least one of the object's corresponding state, the main body's corresponding state, or the state matching state is deemed invalid, and the process verification judgment parameters meet the preset verification conditions, then the corresponding current process verification warning information is output, such as prompting a re-verification of the main body executing the current process or a reconfirmation of the current process's state conditions. Preset verification conditions refer to the conditions used to determine whether the current target key process needs to be verified and warned. These conditions can be set by the preset personnel based on the statistical distribution results of the object deviation degree, main body deviation degree and status deviation degree in historical maintenance operation samples, or they can be set in combination with maintenance operation ticket requirements, process acceptance rules and on-site safety control requirements.

[0053] By introducing weighted processing of rollback coefficients and weighted processing of verification coefficients, the originally scattered anomaly parameters can be further transformed into comparable comprehensive judgment parameters, thereby avoiding the use of a uniform processing method for all anomalies. For processes where the preceding basis is no longer valid, the process rollback judgment parameters can guide the system to output process rollback warnings; for processes that mainly exhibit local deviations in objects, subjects, or states, the process verification judgment parameters can output current process verification warnings. This enables a closer correspondence between warning results and anomaly types, process states, and handling paths, improving the pertinence, accuracy, and closed-loop nature of risk identification and warning handling in key processes.

[0054] It should be added that, such as Figure 4The diagram shown illustrates the logic for identifying behavioral mismatch risks. On the left, candidate related event combinations are first input. Then, binary status is read to preliminarily determine whether the current event and its preceding events are based on valid completion. Based on this, three types of parameters are further extracted and calculated: temporal offset, subject-correspondence deviation, and state-matching deviation. The temporal offset characterizes the deviation of the current event's occurrence time from the completion time of the preceding event; the subject-correspondence deviation characterizes the deviation between the current executing subject and the preset related subject; and the state-matching deviation characterizes the mismatch between the current executing state and the target state after the completion of the preceding event. The anomaly type determination unit in the diagram outputs a result indicating linkage anomaly, subject anomaly, state anomaly, or normal sequential linkage, based on the satisfaction of the above three types of parameters and preset conditions. Therefore, this diagram intuitively reflects that the present invention does not rely solely on static judgment based on a single event, but rather achieves detailed identification of behavioral anomalies in key processes through comprehensive analysis of the temporal, subject, and state relationships between preceding and following events.

[0055] like Figure 6 The chart comparing the process rollback judgment parameters and process verification judgment parameters shows that the horizontal axis represents different abnormal samples, and the vertical axis represents the judgment parameter values; blue bars represent process rollback judgment parameters, orange bars represent process verification judgment parameters, and dashed and dotted lines represent rollback thresholds and verification thresholds, respectively. It can be seen that the process rollback judgment parameters for some samples are higher than the rollback threshold, indicating that the current process corresponding to these samples is not based on the effective completion of the preset preceding process, making it suitable to trigger a process rollback warning; while the process verification judgment parameters for some samples are higher than the verification threshold, indicating that they are more suitable to enter the current process verification processing path. This chart reflects that the present invention does not adopt a uniform handling method for abnormal results, but rather outputs process rollback warning information or current process verification warning information separately based on the differences in the necessity of rollback and verification for different samples, thereby achieving more targeted warning and diversion handling.

[0056] In summary, by integrating spatial attribution identification, process stage connection identification, process behavior correlation analysis, linkage status anomaly identification, and differentiated early warning output in substation bay power outage maintenance scenarios, the system can not only identify where the current behavior occurs and which key process it belongs to, but also further determine whether the current behavior is based on the effective completion of the preceding process, whether it meets the preset subject corresponding conditions, and whether it meets the preset state acceptance conditions. Based on this, it can differentiate between timing anomalies, subject anomalies, state anomalies, and linkage anomalies, thereby triggering process rollback early warnings or current process verification early warnings respectively. This improves the accuracy and targetedness of identifying continuity risks, connection risks, and collaborative risks in key processes, thus enhancing the application effect of the intelligent safety supervision system in refined supervision and closed-loop risk management throughout the entire operation process.

[0057] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0058] This invention is described with reference to flowchart illustrations and / or block diagrams of systems, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0059] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0060] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0061] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0062] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A smart identification and early warning system for power grid operation risks in key processes, characterized in that, The system includes: The multi-source process data coupling module is used to acquire multi-source operation data corresponding to the execution status of the target key process within the power grid operation area, so as to perform spatial region mapping processing or process stage coupling processing to obtain a set of operation events with temporal continuity and process correlation. The key processes for the target include at least the voltage testing process, the power outage and restoration process, and the inspection and maintenance process. The process behavior correlation analysis module is used to analyze the interaction and change relationship between the operation behavior parameters corresponding to the execution state based on the set of operation events, so as to identify the risk of behavior mismatch during the execution of the target key process. The linkage state anomaly identification module is used to analyze the linkage state of the target key process during execution based on the collaborative mismatch state corresponding to the behavior mismatch risk, and generate the corresponding behavior anomaly identification result. The behavioral linkage state includes at least one or more of the following: the temporal relationship of work behavior, the corresponding relationship of work subject, and the matching relationship of work status; The behavioral risk warning output module is used to issue a process rollback warning or a current process verification warning based on the behavioral anomaly identification results, and to output the corresponding behavioral risk warning information when the preset warning conditions are met.

2. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 1, characterized in that, The multi-source process data coupling module includes: Acquire spatial operation data corresponding to the target key process, wherein the spatial operation data includes operator location data, equipment object location data, tool location data, area boundary data, and process space identification data; Based on the equipment location data and the area boundary data, determine the target process spatial area corresponding to at least one of the following: the voltage testing operation process, the power outage and restoration operation process, and the maintenance operation process. Based on the spatial correspondence between the worker location data, the tool location data, and the target process space area, determine the spatial attribution relationship between the current work behavior and the corresponding target key process. The spatial operation data is categorized and integrated according to the spatial affiliation relationship to generate corresponding spatial mapping events. Then, according to the occurrence time order and process affiliation order of the spatial mapping events, a set of spatial chain events with spatial location association characteristics and process affiliation continuity characteristics is formed.

3. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 1, characterized in that, The multi-source process data coupling module further includes: Acquire stage operation data corresponding to the target key process, wherein the stage operation data includes process stage status data, process switching identifier data, process completion status data, and process rollback status data; Based on the sequential dependencies between the voltage testing operation, the power outage and restoration operation, and the maintenance operation, the stage operation data is processed by stage coding to form corresponding process stage identifiers. The process stage identifier is incremented or decremented to update the entry status, execution status, completion status, or rollback status of the corresponding target key process. The stage connection relationship between different target key processes is determined based on the updated process stage identifier, and the stage operation data is integrated and processed based on the stage connection relationship to form a set of process connection events with stage status connection characteristics and process advancement association characteristics.

4. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 2 or 3, characterized in that, The process behavior correlation analysis module includes: According to the preset time window and the process affiliation relationship corresponding to the target key process, adjacent operation events in the spatial chain event set or process connection event set are sorted and paired to generate candidate associated event combinations. The candidate associated event combination includes the current event corresponding to the current process and the preceding event corresponding to the preceding process that satisfies the preset process affiliation relationship with the current process. When the current process is the starting process in the current process chain, it is not included in the candidate associated event combination. Read the binary status corresponding to each candidate associated event combination; If the identifier corresponding to the binary state is 1, it indicates that the current event is based on the valid completion of the preceding event, the current process is in the sequential execution state, and the current event is recorded as a normal candidate event; If the identifier corresponding to the binary state is 0, it indicates that the current event is not based on the effective completion of the preceding event, the current process is in reverse execution or connection abnormality state, and the current event is recorded as an abnormal candidate event.

5. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 4, characterized in that, The process behavior correlation analysis module also includes: Read the behavior object data, behavior status data, behavior time data and behavior stage data corresponding to the abnormal candidate event, and read the preceding object data, preceding completion status data, preceding completion time data and preceding stage data of the preceding event corresponding to the abnormal candidate event; Perform process behavior adaptation processing to obtain object correspondence results, state connection results, timing judgment results and stage correspondence results, and summarize them to obtain process behavior adaptation results; The process behavior adaptation processing includes: performing object correspondence processing between the behavior object data and the preceding object data; performing state connection processing between the behavior state data and the preceding completion state data; performing time sequence processing between the behavior time data and the preceding completion time data; and performing stage correspondence processing between the behavior stage data and the preceding stage data. When the process behavior adaptation result indicates that at least one of the following is true: object mismatch, state disconnection, time sequence mismatch, or stage correspondence error, then the abnormal candidate event is determined to have a risk of behavior mismatch. When the process behavior adaptation results show that the objects correspond, the states are connected, the time sequence is satisfied, and the stages are correctly matched, then it is determined that the abnormal candidate event does not have the risk of behavior mismatch.

6. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 5, characterized in that, The linkage status anomaly identification module includes: If the abnormal candidate event corresponding to the behavior mismatch risk is not associated with an adjacent operation event, the abnormal candidate event is recorded as an independent abnormal candidate event, and a corresponding basic abnormal identification result is generated based on the corresponding process affiliation information, current status information and preset process chain information. If the abnormal candidate event corresponding to the behavior mismatch risk is associated with an adjacent operation event, then the behavior occurrence time parameter, behavior subject identification parameter, and behavior status parameter corresponding to the adjacent operation event are extracted.

7. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 6, characterized in that, The linkage status anomaly identification module further includes: The timing offset is determined based on the difference between the occurrence time parameter of the current event and the completion time parameter of the preceding event. Based on the correspondence between the behavioral subject identifier parameters of the current event and the preset associated subject identifier parameters of the preceding event, determine the subject corresponding deviation amount; The state matching deviation is determined based on the matching relationship between the behavior state parameters of the current event and the corresponding target state parameters after the previous event is completed. Based on the time offset, the corresponding deviation of the subject, and the state matching deviation, linkage anomaly detection and processing are performed.

8. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 7, characterized in that, The linkage anomaly detection and processing includes: If the timing offset, the corresponding deviation of the subject, and the state matching deviation all meet the corresponding preset conditions, then a normal sequential linkage result is generated. If only the timing offset does not meet the corresponding preset condition, a timing anomaly result is generated; If only the deviation amount corresponding to the subject does not meet the corresponding preset condition, an abnormal result for the subject is generated; If only the state matching deviation amount does not meet the corresponding preset condition, then an abnormal state result is generated; If at least two of the time offset, the corresponding deviation of the subject, and the state matching deviation do not meet the corresponding preset conditions, a linkage anomaly result is generated.

9. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 8, characterized in that, The behavioral risk warning output module includes: When the behavior anomaly identification result is characterized as a time-series anomaly or a linkage anomaly, it is determined that the current target key process is in a process rollback pending confirmation state, and the corresponding process connection state and process rollback judgment parameters are obtained. If the process connection status indicates that the current target key process is not established on the basis of the effective completion of the preset preceding process, and the process rollback judgment parameter meets the preset rollback condition, then the corresponding process rollback warning information is output. The process rollback determination parameters are obtained by weighting the time offset corresponding to the current target key process, the completion status parameters of the previous process, and the process connection relationship parameters with rollback coefficients and then adding them together.

10. The intelligent identification and early warning system for power grid operation behavior risks oriented towards key processes as described in claim 8, characterized in that, The behavioral risk warning output module also includes: When the behavior anomaly identification result is characterized as a subject anomaly result or a state anomaly result, it is determined that the current target key process is in a process verification pending confirmation state, and the corresponding object corresponding state, subject corresponding state, state matching state and process verification judgment parameters are obtained; If at least one of the object-corresponding state, the subject-corresponding state, and the state-matching state is invalid, and the process verification judgment parameter meets the preset verification conditions, then the corresponding current process verification warning information is output. The process verification and judgment parameters are obtained by weighting the object deviation degree, the main body deviation degree, and the state deviation degree corresponding to the current target key process with verification coefficients and then adding them together.

Images