A method for verifying anomalies in multi-source operation data of photovoltaic power plants

By identifying the abnormal behavior types and correlations of multi-source operation data of photovoltaic power plants, generating event review configurations, and verifying the consistency of physical and communication states, the problem of unified review of abnormal detection of multi-source operation data of photovoltaic power plants is solved, and accurate anomaly identification and impact range determination are achieved.

CN123069207APending Publication Date: 2026-11-10GUANGDONG YANJI TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611312962.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-27
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing methods for detecting anomalies in multi-source operation data of photovoltaic power plants are difficult to verify uniformly when faced with different sources of anomalies. This leads to misjudgment of abnormal equipment data, expansion of communication anomalies, and discontinuity in data correlation, making it impossible to accurately identify the source of anomalies and the scope of their impact.

Method used

By acquiring multi-source operational data, data timing attributes, equipment measurement point relationships, electrical wiring relationships, control response relationships, and communication channel relationships, the abnormal behavior types are identified, abnormal candidate events are generated, and the configuration is matched and verified according to the association relationship type to generate an event-related operational status set. The consistency of physical operation, consistency of status timing, and common cause of communication are verified to determine the abnormal source type and the scope of impact.

Benefits of technology

It enables accurate anomaly verification of multi-source operation data of photovoltaic power plants, reduces misjudgments, distinguishes between changes in the actual status of equipment, anomalies at single measurement points, and communication anomalies, and ensures the accuracy of data quality and the scope of communication impact.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123069207A_ABST
    Figure CN123069207A_ABST
Patent Text Reader

Abstract

This invention relates to the field of photovoltaic power plant automation monitoring and power operation data processing technology, and particularly to a method for verifying anomalies in multi-source operation data of photovoltaic power plants. The method acquires multi-source operation data, data time-series attributes, and relationships between equipment measuring points, electrical wiring, control responses, and communication channels; identifies the anomaly types of target measuring points and forms an event verification configuration; determines a set of related measuring points for verification according to the type of association, and forms a set of event-related operating states based on the time of anomaly occurrence; verifies the operating constraints, control response sequence, and data refresh status of measuring points on shared communication channels of equipment and upstream and downstream measuring points, forming a set of verification evidence states; and determines the source and scope of anomaly impact based on evidence combinations, unavailable states, and evidence conflicts, and updates the corresponding states. This invention can distinguish between equipment anomalies, measuring point anomalies, communication anomalies, and composite anomalies, reducing misjudgments of anomaly sources caused by time-series differences and missing evidence.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power plant automation monitoring and power operation data processing technology, and in particular to a method for verifying anomalies in multi-source operation data of photovoltaic power plants. Background Technology

[0002] In the field of automated monitoring of photovoltaic power plants, existing solutions typically collect telemetry data, remote signaling data, equipment operating parameters, and communication status through data acquisition and monitoring control systems, measurement and control devices, remote control devices, and communication gateways. They also use operating thresholds, data change rates, data quality bits, or preset time intervals to identify data exceeding limits, data mutations, data freezes, and data missing. Some solutions also identify suspicious measurements based on electrical wiring relationships and power balance, or identify communication link and data refresh anomalies based on heartbeat status, data reception time, and data change time.

[0003] Existing solutions often have fixed processing procedures for equipment fault diagnosis, measurement data identification, and communication anomaly detection. When the same anomaly may be caused by changes in the actual state of the host equipment, abnormal acquisition of target measurement points, or abnormality of the shared communication channel, there is a lack of unified anomaly objects and verification criteria between different processing procedures. At the same time, telemetry data, teleindication data, control commands, execution feedback, and communication status have different sampling periods, communication delays, and equipment response times. When using a unified time slice or fixed time interval to organize data, it is easy to identify data that has not yet reached the response time as anomaly, or to include data that has lost its event correlation in the verification.

[0004] Furthermore, existing topology-related measurements are typically verified around fixed electrical ranges, while communication anomaly detection mainly focuses on the data refresh status of measurement points within the same communication channel. These two approaches lack a continuous connection in terms of the selection criteria for verification of related measurement points, the validity of evidence, and the scope of anomaly impact. When inconsistencies arise between physical operating status, control response status, and communication status, or when some verification data is missing, situations can easily occur where abnormal equipment data is mistakenly classified as bad data, single-point anomalies are expanded to communication anomalies, and different anomalies coexisting are attributed to a single source. Therefore, it is necessary to address the problem of difficulty in uniformly verifying the source and scope of anomalies when the timing of multi-source operating data from photovoltaic power plants differs and the source of anomalies is unclear. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for verifying anomalies in multi-source operation data of photovoltaic power plants, comprising: S100: Acquire multi-source operating data, data timing attributes, equipment measurement point relationships, electrical wiring relationships, control response relationships, and communication channel relationships; identify the abnormal behavior types of target measurement points, generate abnormal candidate events, and match the associated relationship types according to the abnormal behavior types to generate event review configurations; S200. Based on the event review configuration, determine the review-related measurement point set according to the association relationship type; organize the target measurement point data, review-related measurement point data, equipment status and communication status according to the abnormal occurrence time of the abnormal candidate event and the data time sequence attribute, and generate the event-related operating status set. S300. Based on the event-related operational status set, respectively verify the operational constraints of the same equipment measurement points and the change sequence of control commands and equipment operating parameters, and the data refresh status of measurement points with disconnected electrical circuits in the shared communication channel, and generate physical operational consistency status, status timing consistency status, communication common cause status and unavailable status to form a set of verification evidence status. S400. Based on the combination relationship, unavailable status and evidence conflict of the review evidence status set, determine the anomaly source type and anomaly impact range; update the device anomaly status, data quality status or communication impact status according to the anomaly source type, and generate anomaly review results.

[0006] Furthermore, the abnormal candidate event includes an event identifier, a target measurement point identifier, a host device identifier, an abnormal behavior type, an abnormal occurrence time, an abnormal duration, and an initial data quality status; the event review configuration includes an association type, a review measurement point category, data time sequence attributes, and a review evidence type.

[0007] Furthermore, the matching relationship types based on the abnormal behavior types include: numerical limit exceeding or data mutation matching device measurement point relationships and electrical wiring relationships; data freezing or data missing matching communication channel relationships; illegal remote signaling state transition matching device measurement point relationships, electrical wiring relationships, and control response relationships; and inconsistent control feedback matching control response relationships and communication channel relationships. Based on the matching results, select the same equipment measurement points, upstream and downstream electrical related measurement points, execution feedback measurement points, and measurement points with non-connected electrical circuits within the shared communication channel to form the verification related measurement point set, and read the corresponding control commands.

[0008] Furthermore, generating the event-related operational status set includes: using the time of the anomaly occurrence as a time reference, determining the effective time range of the target measurement point data and the verification related measurement point data according to the sampling period and data update time, determining the effective time range of the equipment status according to the equipment response time, and determining the effective time range of the communication status according to the communication delay; Data within the corresponding valid time range is written into the event-associated running status set, and an unavailable flag is configured for data that exceeds the corresponding valid time range.

[0009] Furthermore, generating the physical operational consistency status includes: comparing the voltage, current, or power changes of the target measuring point with the voltage, current, or power changes of the measuring points on the same equipment; verifying the power balance of upstream and downstream electrically related measuring points according to the switch status; and reading equipment alarms and protection actions. When the abnormal behavior of the target measuring point meets both the change of the measuring point of the same device and the power balance, the physical operation consistency state is recorded as supporting the change of device state; when the abnormal behavior of the target measuring point does not meet either, it is recorded as not supporting the change of device state; when the corresponding data has an unavailable flag, it is recorded as unavailable.

[0010] Furthermore, generating the state timing consistency state includes: arranging remote control commands or remote adjustment settings, execution feedback, remote signaling status, and equipment operating parameters according to the control response relationship; comparing the change time of execution feedback, remote signaling status, and equipment operating parameters with the equipment response time, and recording the comparison result as conforming to the control response relationship, not conforming to the control response relationship, or unavailable.

[0011] Furthermore, generating the common cause state of the communication includes: extracting multiple measurement points where the electrical circuits within the shared communication channel are not connected; comparing the data reception time, data change time, data quality bits, and data freeze state of the multiple measurement points, and reading the heartbeat status data and protocol message status; When multiple measurement points simultaneously experience data not refreshing, data freezing, or abnormal data quality bits within their corresponding valid time range, and the heartbeat status data or protocol message status is abnormal, it is recorded as a shared communication reason; if they do not simultaneously occur, it is recorded as no shared communication reason; if the corresponding data has an unavailable flag, it is recorded as unavailable.

[0012] Furthermore, determining the type of anomaly source includes: When the physical operation consistency status supports device status changes, the communication common cause status does not have a shared communication cause, and the review evidence status corresponding to the review evidence types listed in the event review configuration does not have an unavailable status or evidence conflict, it is recorded as a device abnormality; When the physical operation consistency status does not support device status changes, the communication common cause status does not have a shared communication reason, and the review evidence status corresponding to the review evidence types listed in the event review configuration does not have an unavailable status or evidence conflict, it is recorded as a measurement point abnormality; When the common cause state of communication has a shared communication cause, the physical operation consistency state does not support changes in device state, and there is no conflict of evidence, it is recorded as a communication anomaly. When the physical operational consistency state supports device state changes and the common cause state of communication has a shared communication cause, it is recorded as a composite anomaly. If the review evidence status corresponding to the review evidence type listed in the event review configuration includes an unavailable status, or if there is a conflict in the review evidence status set, it is recorded as pending review.

[0013] Furthermore, updating the device abnormal state, the data quality state, or the communication impact state includes: When equipment malfunctions, the actual malfunction data of the target measuring point should be retained and the equipment malfunction status should be updated. If a measurement point is abnormal, the data quality status of the target measurement point will be updated to invalid or downgraded for use. In the event of a communication anomaly, the measurement points that commonly experience the anomaly within the shared communication channel will be written into the anomaly impact range and the communication impact status will be updated. The composite anomaly corresponds to updating the device's abnormal status and the communication impact status; Data awaiting review should be kept in a questionable quality state and trigger a re-collection or delayed review.

[0014] Furthermore, generating the anomaly review result includes: recording the electrical wiring status version and communication status snapshot at the time the anomaly occurred; and writing the anomaly candidate event, the review associated measurement point set, the effective time range corresponding to various types of data in the event associated operating status set, the review evidence status set, the anomaly source type, the anomaly impact range, and the status update result into the anomaly review record. When performing the delayed review, the electrical wiring status version and communication status snapshot in the abnormal review record are called to regenerate the event-associated running status set.

[0015] The key innovations of this invention include: (1) Establish a correspondence between the abnormal performance type of the target measurement point and the relationship between the equipment measurement point, the electrical wiring relationship, the control response relationship and the communication channel relationship. The abnormal performance type limits the association type, the measurement point category and the evidence type of this review call, forming an event review configuration that changes with the abnormal candidate event.

[0016] (2) Based on the time of occurrence of the abnormal candidate event, combined with the sampling period, data update time, communication delay and equipment response time, the effective time range of the target measurement point data, the verification related measurement point data, the equipment status and the communication status are respectively defined; within the effective time range, the data refresh status of the physical operation constraints, the control response sequence and the shared communication channel measurement points are verified together to form a set of verification evidence status including physical operation consistency status, status timing consistency status, communication common cause status and unavailable status.

[0017] (3) The evidence combination, unavailable status and evidence conflict in the review evidence status set are used as the classification conditions for the source of anomalies, distinguishing between equipment anomalies, measurement point anomalies, communication anomalies, compound anomalies and pending review, and making the source of anomalies correspond to the equipment anomaly status, the target measurement point data quality status or the communication impact status of the measurement point in the shared communication channel, respectively; when the review is delayed, the electrical wiring status version and communication status snapshot at the time of the anomaly are called to reorganize the event-related running status set.

[0018] The following are its main beneficial effects: (1) The abnormal behavior type directly constrains the correlation relationship and the selection of verification measurement points, so that numerical limits, data mutation, data freezing, data missing and inconsistent control feedback call the verification object corresponding to its abnormal source respectively, reducing the confusion of evidence sources caused by fixed topology range or full measurement points participating in verification.

[0019] (2) For different timing sequences of telemetry data, teleindication data, control commands, execution feedback and communication status response, the effective time range of each type of data is limited, and data that exceeds the effective time range is marked as unusable. This can prevent data that has not yet reached the device response time from being identified as abnormal in advance, and can also prevent data that has lost event association from continuing to participate in physical, timing or communication verification.

[0020] (3) By combining evidence, conflicting evidence and unavailable evidence, the actual changes in equipment status, single measurement point acquisition anomalies and shared communication channel anomalies can be distinguished under the same abnormal manifestation; the corresponding differentiated status update allows the actual abnormal data of the equipment to be retained, the abnormal data of the measurement point is used as invalid or downgraded in subsequent processing, the impact of communication anomalies is limited to the measurement points that have common anomalies, and remains in a pending review state when the evidence is insufficient. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating a method for verifying anomalies in multi-source operation data of a photovoltaic power plant, as provided in an embodiment of this application. Detailed Implementation

[0022] Example 1: Refer to Figure 1 This is a flowchart illustrating a method for verifying anomalies in multi-source operation data of a photovoltaic power plant, provided by an embodiment of the present invention. The process may include at least steps S100-S400: S100: Acquire multi-source operating data, data timing attributes, equipment measurement point relationships, electrical wiring relationships, control response relationships, and communication channel relationships; identify the abnormal behavior types of target measurement points, generate abnormal candidate events, and match the associated relationship types according to the abnormal behavior types to generate event review configurations; S200. Based on the event review configuration, determine the review-related measurement point set according to the association relationship type; organize the target measurement point data, review-related measurement point data, equipment status and communication status according to the abnormal occurrence time of the abnormal candidate event and the data time sequence attribute, and generate the event-related operating status set. S300. Based on the event-related operational status set, respectively verify the operational constraints of the same equipment measurement points and the change sequence of control commands and equipment operating parameters, and the data refresh status of measurement points with disconnected electrical circuits in the shared communication channel, and generate physical operational consistency status, status timing consistency status, communication common cause status and unavailable status to form a set of verification evidence status. S400. Based on the combination relationship, unavailable status and evidence conflict of the review evidence status set, determine the anomaly source type and anomaly impact range; update the device anomaly status, data quality status or communication impact status according to the anomaly source type, and generate anomaly review results.

[0023] S100: Acquire multi-source operating data, data timing attributes, equipment measurement point relationships, electrical wiring relationships, control response relationships, and communication channel relationships; identify the abnormal behavior types of target measurement points, generate abnormal candidate events, and match the associated relationship types according to the abnormal behavior types to generate event review configurations. In this embodiment, a photovoltaic power plant monitoring server is used as the execution entity. The photovoltaic power plant monitoring server establishes data connections with the Supervisory Control and Data Acquisition (SCADA) system, inverter controller, measurement and control device, protection device, remote control device, and communication gateway. The SCADA system provides telemetry data, remote signaling data, remote control commands, remote adjustment setpoints, and execution feedback. The inverter controller and protection device provide equipment operating parameters, equipment alarms, and protection actions. The remote control device and communication gateway provide data reception time, data change time, data quality bits, heartbeat status data, and protocol message status. The photovoltaic power plant monitoring server receives the above data according to the acquisition cycle and retains the measurement point identifier and time identifier of each data in the original data source.

[0024] In this embodiment, multi-source operating data includes voltage, current, active power, reactive power, and operating status of inverters, combiner devices, transformers, buses, branches, and switchgear in the photovoltaic plant. It also includes remote control commands, remote adjustment settings, execution feedback, equipment alarms, and protection actions. The data timing attributes include the sampling period, data reception time, and data change time of each measuring point, as well as the communication delay of the communication channel and the device response time of the host device to the remote control command or remote adjustment setting. These timing attributes are not uniformly converted to the same sampling frequency in this step, but are associated with the corresponding measuring point, device, or communication channel, so that S200 can establish the effective time range of the data respectively.

[0025] The device measurement point relationship is read from the photovoltaic power plant engineering configuration. The device measurement point relationship represents the affiliation between the measurement point and the host equipment, as well as the correspondence between voltage, current, power, switch status and equipment alarm within the same host equipment. The electrical wiring relationship is formed by the main wiring model of the power plant and the current switch status. When the switch status changes, the monitoring server re-identifies the busbars, branches and equipment connected to the electrical circuit of the target measurement point according to the closed branch, avoiding the use of static wiring diagrams to replace the actual electrical connection status at the time of the event.

[0026] The control response relationship is pre-configured in the control response rules. The control response relationship indicates the order and corresponding time range of changes that should occur in the execution feedback, remote signaling status, and equipment operating parameters after a remote control command or remote adjustment setpoint is issued. For example, after an inverter shutdown command is issued, the execution feedback first returns to the command receiving status, and the inverter operation remote signaling then changes from the operating status to the shutdown status. The active power and AC current decrease within the equipment response time. If one of the data does not change in this order, the monitoring server does not directly identify the unchanged data as a equipment fault, but instead treats the inconsistent control feedback as an abnormal behavior type for subsequent review.

[0027] The communication channel relationship is established based on the correspondence between measuring points, measurement and control devices, remote control devices, and data acquisition channels in the engineering configuration. When the same data acquisition channel carries multiple measuring points, the monitoring server records the shared communication channel of the multiple measuring points. For communication status verification, measuring points with disconnected electrical circuits are also selected from the shared communication channel. The disconnected electrical circuits mean that the host equipment corresponding to the measuring point is not in the same bus, the same branch, or the direct upstream and downstream electrical connection range. This constraint is used to distinguish between synchronous changes caused by real electrical disturbances and common data anomalies caused by the shared communication channel.

[0028] Anomaly identification is initiated jointly by periodic scanning and state change triggers. Periodic scanning targets data not refreshing, data freezing, and data missing. State change triggers target numerical limits exceeding limits, data mutations, illegal remote signaling state transitions, and inconsistent control feedback. Numerical limits exceeding limits are identified according to the operating threshold corresponding to the device's operating state. Data mutations are identified according to the amount or rate of change of adjacent valid sample values. Data freezing is identified according to the difference between the data change time and the current time. Data missing is identified according to the difference between the data reception time and the current time. Illegal remote signaling state transitions are identified according to the allowed state transition relationships. Inconsistent control feedback is identified according to the control response relationship and the device response time.

[0029] Taking the AC active power measurement point of the inverter as an example, when the output of the measurement point suddenly drops from normal to near zero within a sampling period, and the inverter shutdown remote signal has not changed, the monitoring server records the AC active power measurement point as the target measurement point, the inverter to which the measurement point belongs as the host device, and records the abnormal behavior type as data mutation. If there is a shutdown remote control command at the same time, the command issuance time and execution feedback are checked at the same time, and the abnormal behavior type is handled according to the inconsistency of control feedback, instead of selecting the verification path solely based on the power drop magnitude.

[0030] The abnormal candidate event includes an event identifier, target measurement point identifier, host device identifier, abnormal behavior type, abnormal occurrence time, abnormal duration, and initial data quality status. The abnormal occurrence time is the time when the data first meets the abnormal identification conditions, and the abnormal duration is formed by the period during which the abnormal state is continuously maintained. If the same target measurement point repeatedly meets the same abnormal conditions while the abnormal state is not resolved, the monitoring server updates the abnormal duration of the original abnormal candidate event and does not create the event again. If the abnormal behavior type changes, the previous behavior type is retained in the original event and the current behavior type is updated, so that S200 can reselect the associated measurement point for review according to the current abnormal behavior.

[0031] The event review configuration matches the type of abnormal behavior with the type of correlation, the category of the review measurement point, the data time sequence attribute, and the type of review evidence. Value exceeding the limit or data mutation is matched with the relationship between the device measurement point and the electrical wiring relationship. Data freezing or data missing is matched with the relationship between the communication channel. Illegal remote signaling state transition is matched with the relationship between the device measurement point, the electrical wiring relationship, and the control response relationship. Inconsistent control feedback is matched with the control response relationship and the communication channel relationship. The monitoring server also writes the physical operation consistency status, state time sequence consistency status, or communication common cause status that need to be verified into the event review configuration.

[0032] When an abnormal behavior type can match multiple related relationship types, the event review configuration retains each related relationship type and its corresponding review evidence type, and does not merge the judgment conclusions in advance in the S100 stage; when the equipment measurement point relationship, electrical wiring relationship, control response relationship or communication channel relationship is missing, the monitoring server still retains the abnormal candidate event, and records the review evidence type affected by the missing relationship as pending verification in the event review configuration. When selecting the related measurement point for review in S200, an unusable flag is configured according to the actual available data to avoid directly classifying the target measurement point as a measurement point abnormal due to missing engineering configuration.

[0033] After S100 is completed, the abnormal candidate event provides S200 with the target measurement point, host device, abnormal behavior type and abnormal occurrence time. The event review configuration limits the types of associations that S200 needs to call, the types of measurement points to be reviewed and the data time sequence attributes. Subsequent steps do not perform indiscriminate scanning of all measurement points on the entire site, but instead form a review data range corresponding to the abnormal behavior around this abnormal candidate event.

[0034] S200. Based on the event review configuration, determine the review-related measurement point set according to the association relationship type; organize the target measurement point data, review-related measurement point data, equipment status, and communication status according to the abnormal occurrence time of the abnormal candidate event and the data time sequence attribute, and generate the event-related operating status set: After receiving the candidate abnormal event and event review configuration generated by S100, the photovoltaic plant monitoring server first checks the electrical wiring status and communication channel status corresponding to the time of the abnormality. If the switch status or communication configuration has changed after the abnormality occurred, the monitoring server calls the switch status and communication channel relationship saved at the time of the abnormality, without using the current relationship to replace the historical status. The electrical wiring status version and communication status snapshot called are retained with the running status set associated with this event for use in abnormality review records and delayed review.

[0035] The set of verification related measurement points is selected hierarchically according to the relationship types listed in the event verification configuration. The equipment measurement point relationship corresponds to the selection of voltage, current, power, operating status and equipment alarm measurement points of the host equipment where the target measurement point is located; the electrical wiring relationship corresponds to the selection of upstream or downstream electrical related measurement points directly connected to the host equipment from the electrical wiring status at the time of the event, along the closed switch and operating branch; the control response relationship corresponds to the selection of control command, execution feedback measurement points, remote signaling status measurement points and equipment operating parameters reflecting the execution results; the communication channel relationship corresponds to the selection of multiple measurement points that share the communication channel with the target measurement point and whose electrical circuits are not connected.

[0036] The selection of measurement points for the same equipment does not require all measurement points of the host equipment to be written into the verification associated measurement point set. Instead, measurement points with corresponding operational constraints are selected based on the physical quantities represented by the target measurement point. When the target measurement point is the AC active power of the inverter, the verification associated measurement points include DC input current, AC current, AC voltage, operation remote signaling, and equipment alarms. When the target measurement point is the switch position remote signaling, the verification associated measurement points include the corresponding remote control command, execution feedback, branch current, and upstream and downstream power. Temperature or cumulative power generation that does not form operational constraints with the target anomaly are not included in this physical operation verification.

[0037] The upstream and downstream electrical connection measurement points are determined based on the closed switches and branch operating status at the time of the event. When the target measurement point is located on the AC side of the inverter, the monitoring server selects the power measurement point on the busbar or the low-voltage side of the box-type transformer along the inverter output line. When the target measurement point is located on the high-voltage side of the transformer, the input power on the low-voltage side and the output power on the high-voltage side are selected respectively. If the corresponding switch is in the open position when the event occurs, the isolated branch measurement point is not used as the associated object for power balance, so as to avoid using data that are not in the same electrical circuit for physical operation consistency verification.

[0038] The selection of measurement points within the shared communication channel simultaneously verifies the communication path and electrical wiring relationship. The monitoring server prioritizes measurement points that are transmitted from the same measurement and control device, remote control device, or data acquisition channel, but whose host devices are located on different branches or different buses. If all existing measurement points within the shared communication channel are located in the same electrical circuit, these measurement points can still be retained in the event-related operating status set, but their communication common cause status configuration is unavailable because their synchronous changes cannot exclude real electrical disturbances. S300 does not form a conclusion on the existence of shared communication reasons based solely on this group of measurement points.

[0039] After the verification and correlation measurement points are determined, the monitoring server uses the time of the anomaly occurrence as the time base to establish the effective time range of the target measurement point data, the verification and correlation measurement point data, the equipment status, and the communication status. The target measurement point and similar telemetry data are extended before and after the time of the anomaly occurrence according to their respective sampling cycles and data update times. The equipment status is extended after the time of the anomaly occurrence according to the equipment response time. The communication status covers the communication process before and after the anomaly occurrence according to the communication delay and message reception cycle. The time ranges are related to each other, but the same start and end times are not mandatory.

[0040] In scenarios of inconsistent control feedback, the remote control command time serves as the start time for status timing verification. The effective time range of the execution feedback is limited by the command issuance time and communication delay, while the effective time range of the remote signaling status and device operating parameters is limited by the command issuance time and device response time. If the execution feedback returns within the communication delay range, but the remote signaling status remains unchanged after the device response time ends, the execution feedback is retained as valid data, and the remote signaling status is considered to have not changed in the response order, rather than being excluded due to inconsistent timestamps.

[0041] In scenarios involving data freezing or missing data, a different time organization method is adopted. The monitoring server reads the last normal change time, the last normal reception time, and the time of the anomaly occurrence of the target measurement point, and extracts the data reception status within the same time range from multiple measurement points in the shared communication channel. If a measurement point has been in a frozen state for a long time before the anomaly of the target measurement point occurs, the measurement point is not used as supporting data for this common anomaly, but is configured with an unavailable flag. Only measurement points that newly appear within the time range corresponding to the target event and do not refresh, are frozen, or have abnormal data quality bits will participate in the communication common cause status verification.

[0042] Data entering the event-related running status set retains its original time stamp and data quality bit. The monitoring server does not use interpolation results to replace missing telemetry, heartbeat status, or execution feedback. When there is a short-term gap between adjacent valid telemetry data and the event review configuration requires verification of the power change direction, in one implementation, the original sampled values ​​on both sides of the gap can be retained and the middle section can be marked as unusable. S300 verifies the change direction based on the valid sampled values, but does not calculate the precise power balance based on the gap section.

[0043] For duplicate messages, out-of-order messages, or data recalls after link recovery, the monitoring server first checks the data order based on the data source identifier, data time, and reception time. Duplicate data retains only one valid value from the corresponding data source, and out-of-order data is assigned to the corresponding event's valid time range according to the data time. The batch data generated by the recall after link recovery retains both reception time and data time. If multiple historical data arrive at the same reception time, the reception time will not be interpreted as multiple devices undergoing state changes simultaneously.

[0044] The event-related operational status set organizes target measurement point data, verification-related measurement point data, equipment status, and communication status according to the verification evidence type, and configures an unavailable flag for objects that exceed the valid time range of the corresponding event, lack time identifiers, or cannot verify the data source; S300 extracts physical operation verification data, status timing verification data, and communication common factor verification data from the event-related operational status set respectively. The electrical wiring status version and communication status snapshot do not directly participate in numerical comparison, but limit the validity of the selected measurement point and channel relationship when the event occurs.

[0045] S300. Based on the event-related operational status set, verify the operational constraints of the same equipment measurement points and the sequence of changes in control commands and equipment operating parameters, as well as the data refresh status of measurement points with disconnected electrical circuits within the shared communication channel, respectively. Generate physical operational consistency status, status timing consistency status, communication common cause status, and unavailability status to form a verification evidence status set: S300 is executed by the review and analysis program in the photovoltaic plant monitoring server. The review and analysis program calls the corresponding review evidence type according to the event review configuration and reads data with a valid time range from the event-related running status set. Review evidence types not included in the review configuration of this event are not considered as necessary evidence for the anomaly source classification. If they are included in the configuration but the corresponding data has an unavailable flag, an unavailable state is formed in the review evidence status set. Subsequently, S400 decides whether to keep it pending review based on whether the evidence is complete and whether there is a conflict.

[0046] The physical operation consistency verification first compares the direction and time of change of the target measuring point and the measuring point of the same equipment, and then verifies the operation constraints between upstream and downstream electrical related measuring points according to the electrical wiring status. When the target measuring point is active power, the monitoring server reads the current, voltage and operation telemetry of the same equipment. When the power suddenly drops and is accompanied by a drop in current, a change in operating status or an equipment alarm, a state supporting the change of the same equipment is formed. When the target power suddenly drops but the current, voltage and operating status remain unchanged, a state not supporting the change of the same equipment is formed.

[0047] Upstream and downstream operational constraints are established based on the switch status and electrical circuit at the time of the event. When effective power measurement points are configured at both ends of the branch, the verification analysis program compares the power balance deviation formed by the branch's input power, output power, and allowable losses. In the transformer scenario, the power change is verified to have a transmission relationship according to the power on the high and low voltage sides and the equipment operating status. When the power balance deviation is within the preset balance deviation range and the direction of change is consistent with the target measurement point, it is recorded as the upstream and downstream change support state. When the switch is in the open position or the associated measurement point has an unavailable flag, the branch is not used to form a support or non-support conclusion.

[0048] Equipment alarms and protection actions serve as the basis for physical operation verification. When the time of an equipment alarm or protection action falls within the valid time range of the corresponding equipment status and is consistent with the abnormal direction of the target measuring point, it supplements the status of the equipment change support. The absence of equipment alarms is not directly equivalent to the equipment being normal. Only when the abnormal behavior of the target measuring point does not conform to the change of the same equipment measuring point or the upstream and downstream power balance, is the physical operation consistency status recorded as not supporting the change of equipment status. When the abnormal behavior of the target measuring point conforms to both the above-mentioned changes and power balance, it is recorded as supporting the change of equipment status.

[0049] When the changes in the measurement points of the same equipment are inconsistent with the conclusions of the upstream and downstream power balance, the verification analysis program retains their respective verification results and does not force the physical operation consistency status to be classified as supported or not supported. For example, if the inverter power suddenly drops and is supported by the current drop of the same equipment, but the power on the low-voltage side of the box transformer does not change accordingly, if both sets of data are within the valid time range, the physical operation consistency status carries an evidence conflict flag, and S400 handles the conflict by combining the communication common cause status and the status timing consistency status.

[0050] The state timing consistency check arranges the remote control command or remote adjustment setpoint, execution feedback, remote signaling status, and equipment operating parameters according to the control response relationship, and compares the change time of each data with the equipment response time. After the remote control command is issued, if the execution feedback returns within the communication delay range, and the remote signaling status and equipment operating parameters change sequentially within the equipment response time, the state timing consistency status is recorded as conforming to the control response relationship. If the execution feedback returns successfully but the remote signaling status or equipment operating parameters do not change within the equipment response time, it is recorded as not conforming to the control response relationship.

[0051] If the control command is not received by the remote control device, or if the execution feedback is missing and the communication status is abnormal at the same time, the review and analysis program does not directly interpret the lack of change in remote signaling as a failure of the device to operate. Instead, it records the status timing consistency status as unavailable. If the execution feedback returns a failure, and the remote signaling status and the device operating parameters remain in their original states, and the order of changes matches the failure feedback, the status timing consistency status is not recorded as not conforming to the control response relationship. The specific manifestations of inconsistent control feedback in the abnormal candidate events are retained with the review evidence status set for S400 to identify whether there are other evidence conflicts.

[0052] For events involving numerical limits exceeding limits or data mutations that do not involve remote control commands or remote adjustment settings, the state timing consistency status is determined by the event review configuration to participate in the anomaly source classification. When the switch status, protection actions, and equipment operating parameters before and after the event have a clear sequential relationship, the review analysis program can still form a state timing consistency status based on the state transition sequence in the control response relationship. If the event review configuration does not list this status as necessary review evidence, its unavailable status will not trigger a separate review.

[0053] The verification object for the common cause status of communication is multiple test points whose electrical circuits are not connected within the shared communication channel. The verification and analysis program compares the data reception time, data change time, data quality bits, and data freeze status of the multiple test points, and reads the heartbeat status data and protocol message status within the effective time range of the same event. When multiple test points jointly experience data not refreshing, data freezing, or abnormal data quality bits, and the heartbeat status data is interrupted or the protocol message status is abnormal, the common cause status of communication is recorded as having a shared communication cause.

[0054] When only some of the measurement points on multiple shared communication channels show abnormalities, while the heartbeat status data and protocol message status are normal, the communication cause status is recorded as no shared communication cause. When the heartbeat status data is abnormal but the measurement points on the same channel are still refreshed normally according to their respective sampling cycles, the verification and analysis program retains the abnormal heartbeat status and does not determine the existence of a shared communication cause based solely on the heartbeat status. This is because abnormalities in the heartbeat monitoring program itself and abnormalities in production data transmission may have different scopes of application.

[0055] If the number of test points with disconnected electrical circuits in the shared communication channel is insufficient, or if multiple test points were already in a non-refreshing state before the anomaly occurred, the communication common cause state is recorded as unavailable. When the communication link is restored and a data general call is executed, if multiple test points only change simultaneously in the receiving time, and the data times correspond to different historical moments, the batch arrival is not recorded as a common data mutation, and the link restoration status in the protocol message is retained in the communication status for S400 to distinguish between link restoration and continuous communication anomaly.

[0056] The physical operation consistency state, the state timing consistency state, and the communication common cause state all retain the data range, state values, and unavailability flags on which they are based. The review and analysis program also identifies direct contradictions between the three types of states. When the physical operation consistency state supports changes in equipment state but the state timing consistency state does not conform to the control response relationship, if the candidate event of this anomaly belongs to control feedback inconsistency, then an evidentiary conflict is formed. When the physical operation consistency state supports changes in equipment state and the communication common cause state has a shared communication cause, then it is not considered a mutually exclusive conflict, but is retained as a combination of evidence for a composite anomaly.

[0057] The verification evidence status set consists of the above three types of status, corresponding validity, and evidence conflict. S400 no longer rereads the original operating data of the entire station, but classifies the source of anomalies according to the verification evidence status set and the necessary verification evidence listed in the event verification configuration. This process preserves the independent meaning of each piece of evidence, avoids physical operation changes being directly covered by communication anomaly status, and also avoids a single measurement point anomaly being forcibly interpreted as a communication failure when communication evidence is lacking.

[0058] S400. Based on the combination relationship, unavailability status, and evidence conflict of the verification evidence status set, determine the anomaly source type and anomaly impact range; update the device anomaly status, data quality status, or communication impact status according to the anomaly source type, and generate anomaly verification results: S400 is executed by the anomaly source determination and status update program. This program receives the verification evidence status set formed by S300 and checks whether all verification evidence types listed in the event verification configuration have obtained a valid status. The anomaly source classification is not sorted by the value of a single piece of evidence, but is matched according to the combination relationship between physical operation consistency status, status timing consistency status and communication common cause status. When the necessary verification evidence listed in the event verification configuration is unavailable or there is an evidence conflict, no definite equipment anomaly, measurement point anomaly or communication anomaly is output.

[0059] When the physical operation consistency status supports changes in device status, the common cause status of communication does not have a shared communication cause, and the review evidence status corresponding to the review evidence type listed in the event review configuration does not have an unavailable status or evidence conflict, the anomaly source type is recorded as device anomaly; for control feedback inconsistency events, the status timing consistency status must also be consistent with the changes in device status. If the order between execution feedback, remote signaling status and device operating parameters does not conform to the control response relationship, this event is not directly classified as device anomaly, but is recorded as pending review based on evidence conflict.

[0060] When the physical operation consistency status does not support changes in equipment status, the communication common cause status does not have a shared communication cause, and all necessary verification evidence is valid, the anomaly source type is recorded as a measurement point anomaly. Taking a sudden drop in the AC active power of an inverter as an example, if the current, voltage, operation remote signaling, and equipment alarms of the same equipment do not change, the upstream and downstream power balance does not support the actual power drop, and the measurement points that are not connected in the same channel are refreshed normally, then the scope of the anomaly is limited to the target measurement point and does not extend to the host equipment or the shared communication channel.

[0061] When there is a shared communication cause in the communication common cause state, the physical operation consistency state does not support the change of device state, and there is no evidence conflict in the verification evidence state set, the abnormal source type is recorded as communication abnormality; the abnormal impact range is formed based on the measurement points in the shared communication channel that commonly experience data not refreshing, data freezing, or data quality bit abnormality. Measurement points in the same channel that do not experience common abnormalities and continue to refresh normally are not included in the abnormal impact range to avoid extending the communication impact state to data objects that have not been affected by this abnormality.

[0062] When the physical operation consistency status supports changes in equipment status, and there is a shared communication cause in the communication common cause status, the anomaly source type is recorded as a composite anomaly. For example, if an inverter fault causes a synchronous decrease in active power and current, and equipment alarms and upstream and downstream power changes both support changes in equipment status, and the communication channel where the inverter is located experiences a heartbeat interruption accompanied by the freezing of multiple non-same electrical circuit measurement points, then the scope of the anomaly impact includes the host equipment where the real anomaly occurred and the measurement points with common anomalies in the shared communication channel. The equipment anomaly status and the communication impact status are retained separately.

[0063] If the necessary review evidence listed in the event review configuration includes an unavailable state, or if there is an evidence conflict among the physical operation consistency state, state timing consistency state, and communication common cause state that cannot be explained by the above combination relationship, the anomaly source type is recorded as pending review. Pending review does not mean deleting the original anomaly candidate event. The monitoring server maintains the suspicious data quality status of the target measurement point and triggers re-collection based on the data source corresponding to the unavailable evidence. If the data source cannot be recovered within the current communication cycle, a delayed review status is established.

[0064] Once the source type of the anomaly is determined, the status update program updates the operating status according to the source object. For equipment anomalies, the actual anomaly data of the target measurement point is retained, and the equipment anomaly status is added to the host equipment status. The data can still participate in equipment fault analysis. For measurement point anomalies, the data quality status of the target measurement point is updated to invalid or degraded. Data in an invalid state is not included in subsequent power statistics and operating status calculations. Data in a degraded state carries its data quality status when it is called.

[0065] Communication anomalies are updated to reflect the communication impact status of multiple measuring points within the affected area. The monitoring server does not modify the original sampled values ​​of these measuring points, but retains the source of the communication impact in the data quality status. Composite anomalies update the device anomaly status of the host device and the communication impact status of the measuring points within the affected area. Real device anomaly data is not uniformly marked as invalid due to the simultaneous existence of communication anomalies. If the data is pending review, the suspicious data quality status of the target measuring point is maintained. The corresponding unusable evidence is replaced when the newly collected data enters the delayed review.

[0066] The anomaly review results record the electrical wiring status version and communication status snapshot at the time of the anomaly, and associate them with the anomaly candidate events, the review associated measurement point set, the event associated operating status set, the effective time range corresponding to various types of data, the review evidence status set, the anomaly source type, the anomaly impact range, and the status update results. The records retain the technical status and data reference relationships involved in this review, and do not use alarm display, personnel handling opinions, or maintenance work orders as anomaly source classification conditions.

[0067] When performing delayed review, the monitoring server calls the electrical wiring status version and communication status snapshot in the abnormal review record, reselects valid data according to the original event review configuration, and generates an event-related running status set. The newly collected data only supplements the parts of the original review evidence status set that have the unavailable flag. The original valid evidence is retained if the data source has not been corrected. If the re-collection result eliminates the unavailable status or evidence conflict, the abnormal source classification and status update are re-executed, and the previous pending review status is updated with the new abnormal review result.

[0068] When reading measurement point data in the next acquisition cycle, the monitoring server synchronously reads the updated data quality status and communication impact status of S400; the data corresponding to equipment abnormalities retains the true abnormality attributes, and the data corresponding to measurement point abnormalities or communication abnormalities are processed in subsequent stages according to the status of invalid, downgraded use, or communication impact. After the abnormality status is cleared, the record of this abnormality review is still retained, so that when the same target measurement point abnormality occurs in the future, the electrical wiring status, communication status, and evidence combination used in the previous review can be verified.

Claims

1. A method for verifying anomalies in multi-source operation data of photovoltaic power plants, characterized in that, include: S100: Acquire multi-source operating data, data timing attributes, equipment measurement point relationships, electrical wiring relationships, control response relationships, and communication channel relationships; Identify the abnormal behavior types of the target measurement points, generate abnormal candidate events, and match the correlation relationship types according to the abnormal behavior types to generate event review configurations; S200. Based on the event review configuration, determine the review-related measurement point set according to the association relationship type; organize the target measurement point data, review-related measurement point data, equipment status and communication status according to the abnormal occurrence time of the abnormal candidate event and the data time sequence attribute, and generate the event-related operating status set. S300. Based on the event-related operational status set, respectively verify the operational constraints of the same equipment measurement points and the change sequence of control commands and equipment operating parameters, and the data refresh status of measurement points with disconnected electrical circuits in the shared communication channel, and generate physical operational consistency status, status timing consistency status, communication common cause status and unavailable status to form a set of verification evidence status. S400. Based on the combination relationship, unavailable status and evidence conflict of the review evidence status set, determine the anomaly source type and anomaly impact range; update the device anomaly status, data quality status or communication impact status according to the anomaly source type, and generate anomaly review results.

2. The method according to claim 1, characterized in that, The abnormal candidate events include event identifier, target measurement point identifier, host device identifier, abnormal behavior type, abnormal occurrence time, abnormal duration, and initial data quality status; the event review configuration includes association type, review measurement point category, data time sequence attribute, and review evidence type.

3. The method according to claim 2, characterized in that, The types of correlations matched based on the abnormal behavior types include: numerical limits exceeding limits or data mutations are matched with device measurement point relationships and electrical wiring relationships; data freezing or data loss is matched with communication channel relationships; illegal remote signaling state transitions are matched with device measurement point relationships, electrical wiring relationships, and control response relationships; and inconsistent control feedback is matched with control response relationships and communication channel relationships. Based on the matching results, select the same equipment measurement points, upstream and downstream electrical related measurement points, execution feedback measurement points, and measurement points with non-connected electrical circuits within the shared communication channel to form the verification related measurement point set, and read the corresponding control commands.

4. The method according to claim 1, characterized in that, Generating the event-related operational status set includes: using the time of the anomaly occurrence as the time reference, determining the effective time range of the target measurement point data and the verification related measurement point data according to the sampling period and data update time, determining the effective time range of the equipment status according to the equipment response time, and determining the effective time range of the communication status according to the communication delay; Data within the corresponding valid time range is written into the event-associated running status set, and an unavailable flag is configured for data that exceeds the corresponding valid time range.

5. The method according to claim 1, characterized in that, Generating the physical operational consistency status includes: comparing the voltage, current, or power changes at the target measurement point with the voltage, current, or power changes at measurement points on the same equipment; verifying the power balance of upstream and downstream electrically related measurement points based on the switch status; and reading equipment alarms and protection actions. When the abnormal behavior of the target measuring point meets both the change of the measuring point of the same device and the power balance, the physical operation consistency state is recorded as supporting the change of device state; when the abnormal behavior of the target measuring point does not meet either, it is recorded as not supporting the change of device state; when the corresponding data has an unavailable flag, it is recorded as unavailable.

6. The method according to claim 1, characterized in that, Generating the state timing consistency status includes: arranging remote control commands or remote adjustment settings, execution feedback, remote signaling status, and equipment operating parameters according to the control response relationship; comparing the change time of execution feedback, remote signaling status, and equipment operating parameters with the equipment response time, and recording the comparison result as conforming to the control response relationship, not conforming to the control response relationship, or unavailable.

7. The method according to claim 1, characterized in that, Generating the common cause state of the communication includes: extracting multiple measurement points where the electrical circuits within the shared communication channel are not connected; comparing the data reception time, data change time, data quality bits, and data freeze state of the multiple measurement points; and reading the heartbeat status data and protocol message status. When multiple measurement points simultaneously experience data not refreshing, data freezing, or abnormal data quality bits within their corresponding valid time range, and the heartbeat status data or protocol message status is abnormal, it is recorded as a shared communication reason; if they do not simultaneously occur, it is recorded as no shared communication reason; if the corresponding data has an unavailable flag, it is recorded as unavailable.

8. The method according to claim 2, characterized in that, Determining the type of anomaly source includes: When the physical operation consistency status supports device status changes, the communication common cause status does not have a shared communication cause, and the review evidence status corresponding to the review evidence types listed in the event review configuration does not have an unavailable status or evidence conflict, it is recorded as a device abnormality; When the physical operation consistency status does not support device status changes, the communication common cause status does not have a shared communication reason, and the review evidence status corresponding to the review evidence types listed in the event review configuration does not have an unavailable status or evidence conflict, it is recorded as a measurement point abnormality; When the common cause state of communication has a shared communication cause, the physical operation consistency state does not support changes in device state, and there is no conflict of evidence, it is recorded as a communication anomaly. When the physical operational consistency state supports device state changes and the common cause state of communication has a shared communication cause, it is recorded as a composite anomaly. If the review evidence status corresponding to the review evidence type listed in the event review configuration includes an unavailable status, or if there is a conflict in the review evidence status set, it is recorded as pending review.

9. The method according to claim 8, characterized in that, Updating the device malfunction status, the data quality status, or the communication impact status includes: When equipment malfunctions, the actual malfunction data of the target measuring point should be retained and the equipment malfunction status should be updated. If a measurement point is abnormal, the data quality status of the target measurement point will be updated to invalid or downgraded for use. In the event of a communication anomaly, the measurement points that commonly experience the anomaly within the shared communication channel will be written into the anomaly impact range and the communication impact status will be updated. The composite anomaly corresponds to the update of both the device anomaly status and the communication impact status. Data awaiting review should be kept in a questionable quality state and trigger a re-collection or delayed review.

10. The method according to claim 9, characterized in that, Generating the anomaly review result includes: recording the electrical wiring status version and communication status snapshot at the time the anomaly occurred; and writing the anomaly candidate event, the review associated measurement point set, the effective time range corresponding to various types of data in the event associated operating status set, the review evidence status set, the anomaly source type, the anomaly impact range, and the status update result into the anomaly review record. When performing the delayed review, the electrical wiring status version and communication status snapshot in the abnormal review record are called to regenerate the event-associated running status set.