A photovoltaic power generation communication exception judgment method and system based on scada data

Abstract

The application discloses a kind of photovoltaic power generation communication exception judgment method and system based on scada data, it is related to new energy station data acquisition technique and monitoring technical field, including: through edge computing configuration is with 5s as collection interval, the scada state data of photovoltaic power generation unit is collected, and through pre-processing is stored in time series database;According to the effective power generation time of station place definition sunrise to sunset period of station belonging to region;Based on the scada state data of time series database call, judge whether photovoltaic inverter is in communication exception;Classified alarm is carried out to communication exception;While supporting the on-site confirmation and subclass record of a small amount of equipment exception under the same box transformer;The application has the beneficial effect that in combination with edge computing technology, reduce data transmission bandwidth occupation and processing delay, improve real-time performance, accurately distinguish between temporary failure and long-term abnormality, improve the accuracy and reliability of the abnormality identification of monitoring system.

Description

Technical Field

[0001] This invention relates to the fields of data acquisition and monitoring technology for new energy power plants, specifically to a method and system for judging communication anomalies in photovoltaic power generation based on SCADA data. Background Technology

[0002] In photovoltaic (PV) power generation systems, the communication stability between PV inverters and other power generation units and the monitoring system directly affects the operation and maintenance efficiency and power generation revenue of the site. Currently, communication anomaly detection methods based on SCADA data often rely on single data features or simple time thresholds, which have the following problems: First, they do not fully integrate edge computing for local data preprocessing, increasing transmission bandwidth pressure and cloud computing load; second, they easily misjudge equipment malfunctions, planned maintenance, or occasional data fluctuations as communication anomalies, resulting in insufficient accuracy; and third, they cannot effectively distinguish between communication channel anomalies and equipment malfunctions, hindering rapid on-site troubleshooting. Therefore, this paper proposes a highly accurate communication anomaly detection method that combines time-series databases and edge computing technology. Summary of the Invention

[0003] To address the aforementioned technical issues, this paper provides a method and system for judging communication anomalies in photovoltaic power generation based on SCADA data. This technical solution solves the problems of transmission bandwidth pressure and cloud computing load, which prevent accurate judgment of communication anomalies and hinder rapid on-site response and troubleshooting.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for judging communication anomalies in photovoltaic power generation based on SCADA data includes: By configuring edge computing to collect SCADA status data of photovoltaic power generation units at 5-second intervals, including: real-time power data, real-time irradiance data, real-time grid-side voltage data, real-time fault data, and real-time planned maintenance data; the data is then pre-processed and stored in a time-series database. The effective power generation time is defined as the period from sunrise to sunset in the area where the power station is located, according to the location of the power station. Based on the SCADA status data retrieved from the time-series database, it is determined whether the photovoltaic inverter is in a non-power generation state or a data offline state. Combined with fault and maintenance signals, communication abnormalities are determined within the effective power generation time. The communication anomaly is classified into different alarm levels, and different alarm levels are pushed when the anomaly lasts for 1 day, 3 days and 7 days respectively. Supports on-site confirmation and subclass recording of a small number of equipment malfunctions under the same transformer.

[0005] Preferably, the data is collected based on edge computing configuration. The edge computing node is deployed locally at the photovoltaic power station and has the ability to collect, preprocess, and compute data in real time. The preprocessing includes removing missing values ​​and abnormal pulse values ​​from the data and standardizing the data. The SCADA status data is obtained through the acquisition link and acquisition mechanism that reuses the wind turbine SCADA data. The acquisition link includes RS485 bus, industrial Ethernet, and 4G / 5G wireless transmission channel.

[0006] Preferably, based on the definition of effective power generation time as the sunrise to sunset period of the area where the power station is located, the following method is used to determine the effective power generation time: the latitude and longitude information of the power station is stored in advance, and the public data interface of the meteorological department is connected to obtain the daily sunrise and sunset times of the area where the power station is located.

[0007] Preferably, the specific rules for determining the data communication status of the photovoltaic inverter based on the SCADA status data retrieved from the time-series database are as follows: The photovoltaic inverter is considered to be in a non-power generation state when its real-time power data is less than 1% of its installed capacity. The offline data status is characterized by the real-time power data, real-time irradiance data, and real-time grid-side voltage data remaining unchanged for more than 15 minutes. A normal communication condition is defined as at least one of the real-time power data, real-time irradiance data, and real-time grid-side voltage data changing 6 times within 1 minute. If the communication anomaly occurs when the current time is within the valid power generation period and the target inverter simultaneously meets the "no power generation state" and "data offline state," then the system will further retrieve fault data and planned maintenance data for the inverter within one minute prior to the "data offline start time" from the time-series database. If the search results show no corresponding fault record and no valid maintenance plan, it is determined that there is a communication anomaly between the photovoltaic inverter power generation unit and the monitoring system. When one of the following conditions is met, it is determined that the no power generation and data anomaly are caused by non-communication factors: the system retrieves equipment fault data; the system retrieves a valid maintenance plan. Communication Anomaly Persistence Status: When the photovoltaic inverter is in a communication anomaly state and the current time is not within the effective power generation time, the communication anomaly state will be maintained. Based on the collected SCADA status data, within the effective power generation time interval, the data communication status judgment logic of the photovoltaic inverter power generation unit is executed to identify data communication anomalies.

[0008] Preferably, the criterion based on "data remains unchanged" is as follows: retrieve the real-time power data sequence, real-time irradiance data sequence, and real-time grid-side voltage data sequence of the target inverter within the most recent 15 minutes from the time-series database. The absolute value of the difference between two adjacent sets of data is less than the preset error threshold of the corresponding data. The preset error threshold is: the error threshold of real-time power data ≤ 0.1kW, the error threshold of real-time irradiance data ≤ 5W / ㎡, and the error threshold of real-time grid-side voltage data ≤ 0.5kV.

[0009] Preferably, the criterion for determining that at least one item in the data changes 6 times within 1 minute is as follows: within any consecutive 1-minute judgment window, at least one of the real-time power data, real-time irradiance data, and real-time grid-side voltage data is judged as follows: if the absolute value of the numerical difference between data points in adjacent acquisition cycles is greater than the dead zone threshold preset for this type of data, it is counted as a valid change; when the cumulative number of valid changes reaches the preset number threshold, it is determined that the "communication is normal" condition is met. The dead zone threshold is configured as follows: for real-time power data, the dead zone threshold is not less than 0.5kW; for real-time irradiance data, the dead zone threshold is not less than 20W / ㎡; and for real-time grid-side voltage data, the dead zone threshold is not less than 1kV.

[0010] Preferably, based on graded alarms for communication anomalies, when a device is determined to be in a "data communication anomaly" state, if the current time is not within the effective power generation time interval, the "data communication anomaly" state will continue to be maintained; communication anomaly alarms are generated and managed: when a communication anomaly is determined, the anomaly start timestamp is recorded in the time series database, and a duration timer is started; alarms of the corresponding level are triggered and pushed according to different level thresholds reached based on the duration; when the photovoltaic inverter power generation unit is detected to have returned to a "normal communication" state, the corresponding alarm is automatically terminated.

[0011] Preferably, based on different threshold levels for duration, when communication anomalies last for 1 day, a Level 3 alarm is automatically generated and pushed to the monitoring center; when communication anomalies last for 3 days, a Level 2 alarm is automatically escalated and pushed; when communication anomalies last for 7 days, a Level 1 alarm is automatically escalated and pushed; data communication is determined to have recovered when one of the following conditions is met: the monitoring module detects that the inverter's data communication status has switched from inactive to active, and communication recovery is manually confirmed; the "data communication anomaly" status is cleared, the timer is reset to zero, and the corresponding alarm automatically ends. The alarm information includes the photovoltaic inverter equipment number with communication abnormality, the transformer substation ownership identifier, the start time of the communication abnormality, the current duration, and the alarm level. The alarm information is pushed through the cloud monitoring platform, the local monitoring screen at the site, and the mobile terminal APP of the site personnel.

[0012] Preferably, based on supporting on-site confirmation and subclass recording of a small number of equipment anomalies under the same transformer box, the system monitors the communication status of all inverters under the same transformer box topology node in real time. When the number of inverters triggering "data communication anomaly" alarms in that node is less than 1%, an additional diagnostic prompt message "suspected individual equipment failure, requires on-site confirmation by site personnel" is automatically generated. The alarm information is pushed through at least one of the following: cloud monitoring platform, local monitoring screen at the site, and mobile terminal APP. Site personnel confirm whether the communication anomaly is caused by a fault in the equipment itself. When the fault diagnosis module confirms that the fault is caused by a defect in the equipment itself, the database marks the subclass of the corresponding alarm record as "communication anomaly caused by equipment failure". The system stores the marked structured data in the historical database, including equipment serial number, production batch information, IEC 62443-3 standard fault type code, and environmental operating condition parameters, which are used to generate equipment reliability analysis reports to optimize fault prediction algorithms, evaluate supplier quality, and allocate spare parts inventory.

[0013] This application also provides a photovoltaic power generation communication anomaly detection system based on SCADA data, specifically including: Data acquisition module: configured to acquire SCADA status data of photovoltaic power generation units at 5-second intervals and transmit it to the data processing module; Data processing module: configured to define the effective power generation time and determine the communication status of the photovoltaic inverter according to preset rules: no power generation, offline data, normal communication, abnormal communication, and communication abnormality hold-on status; Alarm push module: Configured to record the start time of communication abnormality, push alarms of corresponding levels according to the duration of abnormality, automatically end alarms when communication returns to normal, and push data to the field station when preset conditions are met; Storage module: Configured to store collected SCADA status data, communication status judgment results, alarm information, and abnormal fault records.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention proposes a method and system for judging photovoltaic power generation communication anomalies based on SCADA data. By combining edge computing technology, it realizes local acquisition and preprocessing of SCADA data, reducing data transmission bandwidth occupation and processing latency, and improving real-time performance. Based on multi-dimensional data characteristics, it judges the data communication anomalies of photovoltaic inverter power generation units, accurately distinguishes between short-term faults and long-term anomalies, and improves the accuracy and reliability of anomaly identification in the monitoring system. By filtering out a large number of instantaneous alarms, it significantly reduces the number of invalid alarms that operation and maintenance personnel need to handle, promptly detects missing key monitoring data and communication interruptions, and ensures the effectiveness of system monitoring. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a block diagram of the system structure of the present invention. Detailed Implementation

[0016] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0017] Reference Figure 1 and Figure 2 As shown, a method for judging communication anomalies in photovoltaic power generation based on SCADA data includes: S1. By configuring edge computing to collect SCADA status data of photovoltaic power generation units at 5-second intervals, including: real-time power data, real-time irradiance data, real-time grid-side voltage data, real-time fault data, and real-time planned maintenance data; the data is then pre-processed and stored in a time-series database. S2. The effective power generation time is defined as the sunrise to sunset period in the area where the power station is located, based on the location of the power station. S3. Based on the SCADA status data retrieved from the time-series database, determine whether the photovoltaic inverter is in a non-power generation state or a data offline state, and combine the fault and maintenance signals to determine the communication abnormality within the effective power generation time. S4. Implement graded alarms for communication anomalies, and push different level alarms when the anomaly lasts for 1 day, 3 days and 7 days respectively. S5. Supports on-site confirmation and subclass recording of a small number of equipment malfunctions under the same transformer substation.

[0018] In step S1, the edge computing node is deployed locally at the photovoltaic power station and has the ability to collect, preprocess, and compute data in real time. The preprocessing includes removing missing values ​​and abnormal pulse values ​​from the data and standardizing the data. The SCADA status data is obtained through the acquisition link and acquisition mechanism of the wind turbine SCADA data reuse. The acquisition link includes RS485 bus, industrial Ethernet, and 4G / 5G wireless transmission channel.

[0019] The edge computing algorithm offers the advantage of performing data analysis at 5-second intervals and determining 15-minute windows directly at the data source. This eliminates the need to upload massive amounts of 5-second interval data (a 100MW power plant generates approximately 170 million data points daily) to the cloud for computation, avoiding network transmission delays (typically ranging from minutes to hours). This allows rules like "data offline for 15 minutes" to be determined instantly when the conditions are met, enabling maintenance personnel to perceive communication interruptions almost in real-time and gain valuable time for emergency repairs.

[0020] In step S2, the sunrise and sunset times of the area where the station is located are determined in the following way: the latitude and longitude information of the station is stored in advance, and the public data interface of the meteorological department is connected to obtain the daily sunrise and sunset times of the area where the station is located.

[0021] In step S3, the data communication status of the photovoltaic inverter power generation unit is defined as follows: The photovoltaic inverter is considered to be in a non-power generation state when its real-time power data is less than 1% of its installed capacity. The offline data status is characterized by the real-time power data, real-time irradiance data, and real-time grid-side voltage data remaining unchanged for more than 15 minutes. A normal communication condition is defined as at least one of the real-time power data, real-time irradiance data, and real-time grid-side voltage data changing 6 times within 1 minute. If the communication anomaly occurs when the current time is within the valid power generation period and the target inverter simultaneously meets the "no power generation state" and "data offline state," then the system will further retrieve fault data and planned maintenance data for the inverter within one minute prior to the "data offline start time" from the time-series database. If the search results show no corresponding fault record and no valid maintenance plan, it is determined that there is a communication anomaly between the photovoltaic inverter power generation unit and the monitoring system. When one of the following conditions is met, it is determined that the no power generation and data anomaly are caused by non-communication factors: the system retrieves equipment fault data; the system retrieves a valid maintenance plan. Communication Anomaly Persistence Status: When the photovoltaic inverter is in a communication anomaly state and the current time is not within the effective power generation time, the communication anomaly state will be maintained. Based on the collected SCADA status data, within the effective power generation time interval, the data communication status judgment logic of the photovoltaic inverter power generation unit is executed to identify data communication anomalies.

[0022] The criterion for determining "constantly unchanged data" is as follows: retrieve the real-time power data sequence, real-time irradiance data sequence, and real-time grid-side voltage data sequence of the target inverter within the most recent 15 minutes from the time-series database. The absolute value of the difference between two adjacent sets of data is less than the preset error threshold of the corresponding data. The preset error threshold is: error threshold of real-time power data ≤ 0.1kW, error threshold of real-time irradiance data ≤ 5W / ㎡, and error threshold of real-time grid-side voltage data ≤ 0.5kV.

[0023] The criterion for determining that at least one data item changes 6 times within 1 minute is as follows: within any consecutive 1-minute judgment window, at least one of the real-time power data, real-time irradiance data, and real-time grid-side voltage data is judged as follows: if the absolute value of the numerical difference between data points in adjacent acquisition cycles is greater than the preset dead zone threshold for this type of data, it is counted as a valid change; when the cumulative number of valid changes reaches the preset number threshold, it is determined that the "communication is normal" condition is met. The dead zone threshold is configured as follows: for real-time power data, the dead zone threshold is not less than 0.5kW; for real-time irradiance data, the dead zone threshold is not less than 20W / ㎡; and for real-time grid-side voltage data, the dead zone threshold is not less than 1kV.

[0024] The benefits of multi-dimensional assessment of communication status include strong anti-interference capabilities, resisting fluctuations in a single data point or short-term interference. For example, even if a power value suddenly changes due to interference, the "data offline" (unchanged for 15 minutes) and "no continuous activity" states will prevent misjudgment. The states are clearly self-describing: each inverter has a set of explicit status labels at any given time (e.g., [generating power, data online, no fault] or [not generating power, data offline, communication abnormal]), not just a simple "normal / abnormal" binary label. This provides a rich data foundation for advanced analysis (such as a comprehensive view of power plant health) and refined operation and maintenance. It also facilitates rule optimization: when misjudgments or omissions occur, maintenance personnel can clearly trace back which dimension's conditions were too lenient or strict (e.g., should the 15-minute "data offline" threshold be adjusted to 10 minutes?), thereby enabling precise parameter optimization and making the algorithm more accurate with repeated use.

[0025] In step S4, based on the graded alarm for communication anomalies, when a device is determined to be in a "data communication anomaly" state, if the current time is not within the effective power generation time interval, the "data communication anomaly" state will continue to be maintained; communication anomaly alarms are generated and managed: when a communication anomaly is determined, the anomaly start timestamp is recorded in the time series database, and a duration timer is started; alarms of the corresponding level are triggered and pushed according to different level thresholds reached based on the duration; when the photovoltaic inverter power generation unit is detected to have returned to the "normal communication" state, the corresponding alarm is automatically terminated.

[0026] Based on different threshold levels for duration, when communication anomalies last for 1 day, a Level 3 alarm is automatically generated and pushed to the monitoring center; when communication anomalies last for 3 days, a Level 2 alarm is automatically escalated and pushed; when communication anomalies last for 7 days, a Level 1 alarm is automatically escalated and pushed. Data communication is considered restored when one of the following conditions is met: the monitoring module detects that the inverter's data communication status has switched from inactive to active, and communication is manually confirmed to have been restored; the "data communication anomaly" status is cleared, the timer is reset to zero, and the corresponding alarm automatically ends. The alarm information includes the photovoltaic inverter equipment number with communication abnormality, the transformer substation ownership identifier, the start time of the communication abnormality, the current duration, and the alarm level. The alarm information is pushed through the cloud monitoring platform, the local monitoring screen at the site, and the mobile terminal APP of the site personnel.

[0027] The time-series database, as the core database of the system, is deeply integrated with edge computing. Its basic process is as follows: Data is first acquired through the acquisition layer and stored in the time-series database deployed on the edge computing nodes. This database has the ability to write and query time-series data with high performance, and supports data compression and retention strategies. The edge computing engine runs detection algorithms on the edge nodes, reads the latest data in the time-series database in real time, performs multi-dimensional status judgments (such as no power generation, offline data, active data, etc.), and generates events (such as the start / end of communication anomalies) and alarms. Then, key events, alarms, and aggregated data (such as minute-by-minute statistics) are synchronized to the cloud time-series database to support global monitoring and big data analysis. The cloud platform provides a visual interface to display the power plant status and alarm list, and supports alarm push function, thereby realizing a closed loop of the entire process from data acquisition to analysis and alarm management.

[0028] The core function of time-series databases is to efficiently store and process time-series data. Key features include: high write throughput optimization designed for high-frequency write scenarios, stably supporting write pressures of tens of thousands of data points per second; intelligent data compression significantly reducing storage space usage and long-term storage costs through specialized algorithms such as differential encoding and run-length encoding; sliding window analysis, based on built-in window functions and continuous query capabilities, directly supports dynamic condition judgments such as "no data change within 15 minutes"; time-dimensional aggregation automatically calculates statistical indicators such as averages and maximum values ​​at intervals of minutes and hours, simplifying report generation; elastic retention strategies support tiered storage configurations (e.g., retaining original data for 30 days and hourly aggregated data for 1 year), achieving a balance between storage efficiency and data availability; and event-related storage unifies the storage of algorithm-generated events (such as communication anomaly start and end markers) with the original time-series data, forming a complete analytical context.

[0029] In step S5, based on the support for on-site confirmation and subclass recording of a small number of equipment anomalies under the same transformer box, the system monitors the communication status of all inverters under the same transformer box topology node in real time. When the number of inverters that trigger the "data communication anomaly" alarm in the node is less than 1%, an additional diagnostic prompt message "suspected individual equipment failure, requires on-site confirmation by site personnel" is automatically generated. The alarm information is pushed through at least one of the following: cloud monitoring platform, site local monitoring screen, and mobile terminal APP. Site personnel confirm whether the communication anomaly is caused by the equipment itself. When the fault diagnosis module confirms that the fault is caused by the equipment itself, the database marks the subclass of the corresponding alarm record as "communication anomaly caused by equipment failure". The system stores the marked structured data in the historical database, including equipment serial number, production batch information, IEC 62443-3 standard fault type code and environmental operating condition parameters, which are used to generate equipment reliability analysis reports to optimize fault prediction algorithms, evaluate supplier quality, and allocate spare parts inventory.

[0030] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for judging communication anomalies in photovoltaic power generation based on SCADA data, characterized in that, include: S1. By configuring edge computing to collect SCADA status data of photovoltaic power generation units at 5-second intervals, including: real-time power data, real-time irradiance data, real-time grid-side voltage data, real-time fault data, and real-time planned maintenance data; pre-process and store the data in a time-series database. S2. The effective power generation time of the station is determined by astronomical calculations based on its geographical coordinates, from the local sunrise time to the sunset time. S3. Based on the SCADA status data retrieved from the time-series database, determine whether the photovoltaic inverter is in a non-power generation state or a data offline state, and combine the fault and maintenance signals to determine the communication abnormality within the effective power generation time. S4. Push different levels of alarms based on different durations; S5. Supports on-site confirmation and subclass recording of a small number of equipment malfunctions under the same transformer substation.

2. The photovoltaic power generation communication anomaly judgment method based on SCADA data according to claim 1, characterized in that: In step S1, the edge computing node is deployed locally at the photovoltaic power station and has the ability to collect, preprocess, and compute data in real time. The preprocessing includes removing missing values ​​and abnormal pulse values ​​from the data and standardizing the data. The SCADA status data is obtained through the acquisition link and acquisition mechanism of the wind turbine SCADA data reuse. The acquisition link includes RS485 bus, industrial Ethernet, and 4G / 5G wireless transmission channel.

3. The photovoltaic power generation communication anomaly judgment method based on SCADA data according to claim 1, characterized in that: In step S2, the sunrise and sunset times of the area where the station is located are determined in the following way: the latitude and longitude information of the station is stored in advance, and the public data interface of the meteorological department is connected to obtain the daily sunrise and sunset times of the area where the station is located.

4. The photovoltaic power generation communication anomaly judgment method based on SCADA data according to claim 1, characterized in that: In step S3, the data communication status of the photovoltaic inverter power generation unit is defined as follows: The photovoltaic inverter is considered to be in a non-power generation state when its real-time power data is less than 1% of its installed capacity. The offline data status is characterized by the real-time power data, real-time irradiance data, and real-time grid-side voltage data remaining unchanged for more than 15 minutes. A normal communication condition is defined as at least one of the real-time power data, real-time irradiance data, and real-time grid-side voltage data changing 6 times within 1 minute. If the communication anomaly occurs when the current time is within the effective power generation period and the target inverter simultaneously meets the "non-power generation state" and "data offline state", then the fault data and planned maintenance data of the inverter within 1 minute before the "data offline start time" are further retrieved from the time series database. If the search results show no corresponding fault record and no effective maintenance plan, it is determined that there is a communication anomaly between the photovoltaic inverter power generation unit and the monitoring system. The system determines that the power outage and data anomaly are caused by non-communication factors when one of the following conditions is met: the system retrieves equipment fault data; the system retrieves a valid maintenance plan. Communication Anomaly Persistence Status: When the photovoltaic inverter is in a communication anomaly state and the current time is not within the effective power generation time, the communication anomaly state will be maintained. Based on the collected SCADA status data, within the effective power generation time interval, the data communication status judgment logic of the photovoltaic inverter power generation unit is executed to identify data communication anomalies.

5. The method for judging photovoltaic power generation communication anomalies based on SCADA data according to claim 4, characterized in that, The criteria for determining "data remains unchanged" are as follows: retrieve the real-time power data sequence, real-time irradiance data sequence, and real-time grid-side voltage data sequence of the target inverter within the most recent 15 minutes from the time-series database. The absolute value of the difference between two adjacent sets of data is less than the preset error threshold of the corresponding data. The preset error threshold is: error threshold of real-time power data ≤ 0.1kW, error threshold of real-time irradiance data ≤ 5W / ㎡, and error threshold of real-time grid-side voltage data ≤ 0.5kV.

6. The photovoltaic power generation communication anomaly judgment method based on SCADA data according to claim 4, characterized in that: The criterion for determining that at least one item in the data changes 6 times within 1 minute is as follows: Within any consecutive 1-minute judgment window, at least one of the real-time power data, real-time irradiance data, and real-time grid-side voltage data is judged as follows: If the absolute value of the numerical difference between data points in adjacent acquisition cycles is greater than the preset dead zone threshold for this type of data, it is counted as a valid change; when the cumulative number of valid changes reaches the preset number threshold, it is determined that the "communication is normal" condition is met. The dead zone threshold is configured as follows: for real-time power data, the dead zone threshold is not less than 0.5kW; for real-time irradiance data, the dead zone threshold is not less than 20W / ㎡; and for real-time grid-side voltage data, the dead zone threshold is not less than 1kV.

7. The photovoltaic power generation communication anomaly judgment method based on SCADA data according to claim 1, characterized in that: In step S4, once the device is determined to be in a "data communication abnormality" state, if the current time is not within the effective power generation time interval, the "data communication abnormality" state will continue to be maintained; a communication abnormality alarm will be generated and managed: when the time is determined to be a communication abnormality, the abnormality start timestamp will be recorded in the time series database, and a duration timer will be started; according to the different level thresholds reached by the duration, alarms of the corresponding level will be triggered and pushed; when the photovoltaic inverter power generation unit is detected to have returned to the "communication normal" state, the corresponding alarm will automatically end.

8. The photovoltaic power generation communication anomaly judgment method based on SCADA data according to claim 7, characterized in that: The different threshold levels are as follows: when communication abnormality lasts for 1 day, a Level 3 alarm is automatically generated and pushed to the monitoring center; when communication abnormality lasts for 3 days, a Level 2 alarm is automatically escalated and pushed; when communication abnormality lasts for 7 days, a Level 1 alarm is automatically escalated and pushed. Data communication is considered to have recovered when one of the following conditions is met: the monitoring module detects that the inverter's data communication status has switched from an inactive state to an active state, and the communication recovery is confirmed manually; the "data communication abnormality" status is cleared, the timer is reset to zero, and the corresponding alarm ends automatically. The alarm information includes the photovoltaic inverter equipment number with communication abnormality, the transformer substation ownership identifier, the start time of the communication abnormality, the current duration, and the alarm level. The alarm information is pushed through the cloud monitoring platform, the local monitoring screen at the site, and the mobile terminal APP of the site personnel.

9. The method for judging photovoltaic power generation communication anomalies based on SCADA data according to claim 1, characterized in that: In step S5, the system monitors the communication status of all inverters under the same box-type transformer topology node in real time. When the number of inverters triggering the "data communication abnormality" alarm in that node is less than 1%, an additional diagnostic prompt message "suspected individual equipment failure, requires on-site confirmation by site personnel" is automatically generated. The alarm information is pushed through at least one of the following: cloud monitoring platform, local monitoring screen at the site, and mobile terminal APP. Site personnel confirm whether the communication abnormality is caused by a fault in the equipment itself. When the fault diagnosis module confirms that the fault is caused by a defect in the equipment itself, the database marks the corresponding alarm record as "communication abnormality caused by equipment failure". The system stores the marked structured data in the historical database, including the equipment serial number, production batch information, IEC 62443-3 standard fault type code, and environmental operating condition parameters, which are used to generate equipment reliability analysis reports to optimize fault prediction algorithms, evaluate supplier quality, and allocate spare parts inventory.

10. A photovoltaic power generation communication anomaly judgment system based on SCADA data, used to implement the photovoltaic power generation communication anomaly judgment method based on SCADA data as described in any one of claims 1-9, characterized in that, include: Data acquisition module: configured to acquire SCADA status data of photovoltaic power generation units at 5-second intervals and transmit it to the data processing module; Data processing module: configured to define the effective power generation time and determine the communication status of the photovoltaic inverter according to preset rules: no power generation, offline data, normal communication, abnormal communication, and communication abnormality hold-on status; Alarm push module: Configured to record the start time of communication abnormality, push alarms of corresponding levels according to the duration of abnormality, automatically end alarms when communication returns to normal, and push data to the field station when preset conditions are met; Storage module: Configured to store collected SCADA status data, communication status judgment results, alarm information, and abnormal fault records.

Images