Photovoltaic power station infrared image feature extraction and abnormal diagnosis method

Abstract

The application provides an infrared image feature extraction and abnormal diagnosis method for a photovoltaic power station, and relates to the technical field of data processing, comprising: dividing each photovoltaic array into a first sub-array, a second sub-array and a third sub-array according to the physical connection structure of the photovoltaic array; obtaining temperature monitoring data of the first sub-array, the second sub-array and the third sub-array in a plurality of continuous sampling periods based on a control ball deployed on the site of the photovoltaic power station to obtain a temperature distribution change sequence; performing hot spot risk judgment on the temperature distribution change sequence based on a hot spot risk judgment rule to obtain risk identification data; in response to the risk identification data confirming the hot spot risk, generating a load reduction control instruction and sending the instruction to a corresponding inverter, and in response to the hot spot risk being eliminated, generating a full load recovery instruction and sending the instruction to the corresponding inverter. The application can accurately distinguish between local heat accumulation of the photovoltaic array and overall environmental temperature fluctuation, and realize reliable judgment and automatic load reduction control of the hot spot risk.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method for extracting infrared image features and diagnosing anomalies in photovoltaic power plants. Background Technology

[0002] During long-term outdoor operation, photovoltaic (PV) power plants can experience hot spot effects due to factors such as partial shading and cell damage. This leads to abnormally high temperatures in certain areas of the modules, affecting power generation efficiency and posing safety hazards. Existing hot spot detection methods mainly fall into two categories: one is based on electrical parameter monitoring, which involves building physical circuits around the PV modules to collect parameters such as output current and voltage. For large-scale PV power plants, this requires deploying a large number of sensors, resulting in high costs and decreasing detection efficiency as the scale of the power plant increases. The other method uses infrared image analysis. Specifically, a monitoring sphere equipped with an infrared thermal imager performs real-time imaging of the PV array, converting the surface temperature distribution of the modules into temperature monitoring data. Hot spot detection is achieved by identifying areas with abnormal temperatures, without the need for external physical circuits.

[0003] A photovoltaic (PV) array consists of multiple components connected in series according to a physical connection sequence. A monitoring sphere is installed and deployed in a fixed manner at the PV power plant site, continuously acquiring infrared image frames of the PV array according to a pre-set sampling period. However, existing detection methods typically set a fixed threshold for the absolute temperature value in a single image frame; exceeding the threshold is considered a hotspot risk. However, the overall temperature level of the PV array is affected by factors such as light intensity and ambient temperature, resulting in significant fluctuations at different sampling times. Therefore, existing technologies are prone to generating numerous false alarms when the overall ambient temperature rises, and prone to missed alarms when the early temperature accumulation of hotspots is small. This makes it difficult to effectively distinguish between two fundamentally different sources of temperature change: overall ambient temperature fluctuations and localized heat accumulation.

[0004] Therefore, how to accurately distinguish between local heat accumulation in photovoltaic arrays and overall ambient temperature fluctuations under continuous data acquisition by a monitoring sphere, and how to reliably assess hot spot risks and automatically reduce load, has become an urgent problem to be solved. Summary of the Invention

[0005] This application provides a method for infrared image feature extraction and anomaly diagnosis of photovoltaic power plants to solve the problems mentioned in the background art.

[0006] This application provides a method for infrared image feature extraction and anomaly diagnosis of photovoltaic power plants, including: dividing each photovoltaic array into a first subarray, a second subarray, and a third subarray according to the physical connection structure of the photovoltaic array; Temperature monitoring data of the first, second, and third subarrays in multiple consecutive sampling periods are obtained by using a control ball deployed at the photovoltaic power station site to obtain a temperature distribution change sequence. Based on the hot spot risk assessment rules, the temperature distribution change sequence is assessed for hot spot risk to obtain risk identification data. In response to the risk identification data confirming the hot spot risk, a load reduction control command is generated and sent to the corresponding inverter, and in response to the elimination of the hot spot risk, a full load restoration command is generated and sent to the corresponding inverter.

[0007] Optionally, in one possible implementation, dividing each photovoltaic array into a first subarray, a second subarray, and a third subarray according to the physical connection structure of the photovoltaic array includes: The arrangement positions of each component in the photovoltaic array are determined according to the physical connection order. All components of the photovoltaic array are divided into three equal segments according to their arrangement positions. The components located in the first segment are designated as the first sub-array, the components located in the third segment are designated as the third sub-array, and the components located in the second segment are designated as the second sub-array. In response to the fact that the number of components in each segment cannot be equally divided, the remaining components are assigned to the second subarray.

[0008] Optionally, in one possible implementation, the acquisition of temperature monitoring data of the first, second, and third subarrays over multiple consecutive sampling periods based on a control sphere deployed at the photovoltaic power plant site to obtain a temperature distribution change sequence includes: The infrared image frames collected by the infrared thermal imager mounted on the control ball according to the preset sampling period are retrieved, and the time period subarray temperature difference reference range of the photovoltaic array is obtained from the acquisition record of the control ball according to the acquisition time period corresponding to the infrared image frame. Based on the comparison between the overall temperature level difference data between the first subarray, the second subarray and the third subarray in the infrared image frame and the reference interval of the temperature difference between the subarrays in the time period, the temperature monitoring data of the current sampling period is obtained. The temperature monitoring data obtained from each sampling period are arranged in chronological order to obtain a temperature distribution change sequence.

[0009] Optionally, in one possible implementation, obtaining the temperature monitoring data for the current sampling period based on the comparison result between the overall temperature level difference data between the first subarray, the second subarray, and the third subarray in the infrared image frame and the reference interval for the temperature difference between the subarrays in the time period includes: In response to the overall temperature level difference data being within the temperature difference reference range of the subarray in the time period, the infrared image frame is determined as a valid image frame, and the temperature values ​​are merged according to the division of the first subarray, the second subarray, and the third subarray to obtain the temperature monitoring data of the current sampling period. In response to the fact that the overall temperature level difference data is not in the temperature difference reference range of the subarray during the time period, the infrared image frame is identified as an interference image frame, and the temperature monitoring data of the previous effective sampling period and the next effective sampling period are interpolated to obtain the temperature monitoring data of the current sampling period.

[0010] Optionally, in one possible implementation, the step of performing hot spot risk assessment on the temperature distribution change sequence based on hot spot risk assessment rules to obtain risk identification data includes: Based on the temperature distribution change sequence, the subarray temperature difference data and array temperature discrete data corresponding to each sampling period are obtained; In response to the fact that the subarray temperature difference data in the current sampling period is greater than the subarray temperature difference data in the previous sampling period, and the array temperature discrete data in the current sampling period is greater than the array temperature discrete data in the previous sampling period, the current sampling period is marked as a double index rising period, and the count of consecutively marked double index rising periods is accumulated. In response to any data in the current sampling period, either the subarray temperature difference data or the array temperature discrete data, not being greater than the corresponding data in the previous sampling period, the dual-index rising period count is cleared. In response to the continuous marking of the dual-index rising cycle count reaching a preset cycle threshold, the photovoltaic array is confirmed to have a hot spot risk, and the risk identification data is obtained.

[0011] Optionally, in one possible implementation, obtaining the subarray temperature difference data and array temperature discrete data corresponding to each sampling period based on the temperature distribution change sequence includes: The representative temperature values ​​of the first subarray and the third subarray in the current sampling period are retrieved and compared to obtain the temperature difference data of the subarrays in the current sampling period. Based on the deviation between the temperature value of each component in the photovoltaic array during the current sampling period and the overall temperature reference value of the photovoltaic array, the discrete temperature data of the array during the current sampling period is determined.

[0012] Optionally, in one possible implementation, the step of marking the current sampling period as a dual-index rising period in response to the subarray temperature difference data being greater than the subarray temperature difference data of the previous sampling period and the array temperature discrete data of the current sampling period being greater than the array temperature discrete data of the previous sampling period includes: By comparing the temperature representative values ​​of the first subarray, the second subarray, and the third subarray in the current sampling period with the temperature representative values ​​corresponding to the previous sampling period, the temperature representative value change data of each subarray is obtained. Using the temperature representative value change data of the second subarray as an internal reference, the temperature representative value change data of the first subarray is compared with the temperature representative value change data of the second subarray to obtain a first comparison result. The temperature representative value change data of the third subarray is compared with the temperature representative value change data of the second subarray to obtain a second comparison result. Based on the first comparison result and the second comparison result, the source of the rise in the subarray temperature difference data in the current sampling period is determined, the rise source determination data is obtained, and it is decided whether to mark the current sampling period as a dual-index rise period.

[0013] Optionally, in one possible implementation, determining the source of the rise in the subarray temperature difference data of the current sampling period based on the first comparison result and the second comparison result, obtaining rise source determination data, and deciding whether to mark the current sampling period as a dual-index rise period includes: In response to the fact that both the first comparison result and the second comparison result indicate that the temperature representative value change data of the first subarray and the third subarray are equal to the temperature representative value change data of the second subarray, it is determined that the source of the rise in the temperature difference data of the subarray in the current sampling period is the synchronous change of the overall temperature level of each component of the photovoltaic array, and the current sampling period is not marked as a dual-index rise period. In response to the first comparison result or the second comparison result indicating that the temperature representative value change data of the first subarray or the third subarray is not equal to the temperature representative value change data of the second subarray, it is determined that the rise in the subarray temperature difference data in the current sampling period is due to local heat accumulation, and the current sampling period is marked as the dual index rise period.

[0014] Optionally, in one possible implementation, the step of confirming the presence of hot spot risk in the photovoltaic array and obtaining the risk identification data in response to the continuous marking of the dual-index rising cycle count reaching a preset cycle threshold further includes: The growth rate of the subarray temperature difference data between each adjacent sampling period of the photovoltaic array is obtained from the acquisition records of the control ball, and the numerical range of the growth rate of the subarray temperature difference data corresponding to each adjacent sampling period is determined as the subarray temperature difference growth reference interval. In response to the fact that the growth rate of the subarray temperature difference data between the current sampling period and the previous sampling period exceeds the reference range for the growth of the subarray temperature difference, the current sampling period is marked as an abnormal growth period. In response to the existence of a sampling period that is simultaneously marked as the dual-index rise period in the continuously marked abnormal growth period, it is confirmed that the growth rate of the subarray temperature difference data exceeds the normal range of the photovoltaic array and the source of the rise of the subarray temperature difference data has ruled out the synchronous change of the overall temperature level of each component of the photovoltaic array. The risk level of the risk identification data is determined to be high risk level, and high risk identification data is obtained.

[0015] Optionally, in one possible implementation, the step of generating a load reduction control command and sending it to the corresponding inverter in response to the risk identification data confirming hot spot risk includes: Based on the risk identification data, the target photovoltaic array with hot spot risk and the target inverter corresponding to the target photovoltaic array are determined, the total number of photovoltaic arrays currently connected to the target inverter and the number of photovoltaic arrays with confirmed hot spot risk are obtained, and the risk array ratio data is obtained. The current output current of the target inverter is obtained, and the load reduction is determined according to the corresponding proportion of the current output current based on the risk array ratio data to obtain the load reduction target current. The load reduction target current is encapsulated into the load reduction control command and sent to the target inverter. In response to receiving the execution confirmation returned by the target inverter, the load reduction operation status is recorded.

[0016] Optionally, in one possible implementation, the step of generating a full-load recovery command in response to the hot spot risk elimination and sending it to the corresponding inverter includes: The temperature distribution change sequence of the photovoltaic array corresponding to the risk identification data is continuously acquired and updated in real time. Based on the temperature distribution change sequence, the subarray temperature difference data and array temperature discrete data of the photovoltaic array corresponding to the risk identification data in the current sampling period are obtained. In response to the fact that the temperature difference data of the subarray in the current sampling period is not greater than the temperature difference data of the subarray in the previous sampling period, and the discrete data of the array temperature in the current sampling period is not greater than the discrete data of the array temperature in the previous sampling period, the current sampling period is marked as a period in which both indicators do not rise, and the count of consecutively marked periods in which both indicators do not rise is accumulated. In response to the continuous marking of the dual indicators not rising cycle count reaching the preset elimination threshold, the hot spot risk is confirmed to be eliminated, the full load recovery command is generated and sent to the corresponding inverter, and in response to receiving the execution confirmation returned by the corresponding inverter, the full load recovery operation status is recorded.

[0017] Optionally, in one possible implementation, after obtaining the high-risk identification data, the method further includes: The high-risk identification data is sent to the operation and maintenance terminal, and the operation and maintenance scheduling request sent by the operation and maintenance terminal based on the high-risk identification data is received. Obtain the photovoltaic array location identifier carried in the operation and maintenance scheduling request, and match the photovoltaic array location identifier with the photovoltaic array location identifiers with high risk level in the high risk identification data to obtain location matching data; Based on the location matching data and the operation type identifier carried in the operation and maintenance scheduling request, a corresponding scheduling control instruction is generated and sent to the target scheduling inverter. In response to receiving the execution confirmation returned by the target scheduling inverter, the execution result is fed back to the operation and maintenance terminal.

[0018] Optionally, in one possible implementation, generating a corresponding scheduling control command based on the location matching data and the operation type identifier carried in the operation and maintenance scheduling request and sending it to the target scheduling inverter includes: In response to a successful match in the location matching data, the operation type identifier carried in the operation and maintenance scheduling request is obtained, and the scheduling operation type is determined based on the operation type identifier; in response to a failed match in the location matching data, an invalid scheduling prompt is returned to the operation and maintenance terminal. In response to the scheduling operation type being manual load reduction, a manual load reduction command is generated and sent to the target scheduling inverter; in response to the scheduling operation type being manual recovery, a manual full load recovery command is generated and sent to the target scheduling inverter.

[0019] The method for infrared image feature extraction and anomaly diagnosis of photovoltaic power plants provided in this application has the following beneficial effects: 1. This application divides each photovoltaic array into a first subarray, a second subarray, and a third subarray according to their physical connection sequence. The comparison of representative temperature values ​​between the first and third subarrays is used as the subarray temperature difference data. The deviation of the temperature values ​​of each component within the array from the overall temperature reference value is used as the array temperature discrete data. Hot spot risk is assessed through a dual-indicator linkage mechanism. The subarray temperature difference data captures the temperature asymmetry at the beginning and end of the series structure, while the array temperature discrete data reflects the overall concentration of temperature distribution among the components. The two indicators exhibit different characteristics when changing in the same direction during overall ambient temperature fluctuations, but show a continuous synchronous increase during localized heat accumulation. This dual-indicator linkage mechanism effectively distinguishes between two fundamentally different sources of temperature change, reducing the false alarm rate.

[0020] 2. This application uses the temperature representative value change data of the second subarray as an internal reference. By comparing the temperature representative value change data of the first and third subarrays with that of the second subarray, it determines the source of the increase in the subarray temperature difference data. This judgment benchmark comes from within the series structure and does not rely on any preset parameters. When the overall ambient temperature changes synchronously, the three segments show consistent change data. When there is local heat accumulation, the first and last segments and the middle segment show differentiation. By utilizing the endogenous characteristics of the series structure, it achieves accurate elimination of overall synchronous changes, further improving the reliability of hot spot risk judgment. In addition, in response to the hot spot risk confirmation, this application performs corresponding proportional load reduction control on the inverter output current based on the proportion data of the risk array. While protecting the hot spot array, it also takes into account the power generation efficiency of the normal array under the same inverter. After the hot spot risk is eliminated, it automatically generates a command to restore full load, realizing a complete closed loop from detection to control. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the method for extracting infrared image features and diagnosing anomalies in photovoltaic power plants provided in this application embodiment; Figure 2 This is a topology diagram of the implementation scheme of the photovoltaic power station infrared image feature extraction and anomaly diagnosis method provided in the embodiments of this application; Figure 3 This is a flowchart of the model training process for the method of infrared image feature extraction and anomaly diagnosis of photovoltaic power plants provided in the embodiments of this application; Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0023] The technical solutions of this application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0024] like Figure 2As shown, the implementation topology of the method applied in this application includes an integrated management platform, a control console, a storage device, a photovoltaic panel inspection PTZ camera, and a mobile terminal for the responsible person. The photovoltaic panel inspection PTZ camera, as described in this application, is equipped with an infrared thermal imager and is deployed in a fixed manner at the photovoltaic power station site. It continuously collects infrared image frames from the photovoltaic array according to a preset sampling period. The collected data is stored in the storage device and then transmitted to the control console for processing. The processing results are uploaded to the integrated management platform. The integrated management platform pushes risk identification data and alarm information to the mobile terminal for the responsible person via a wireless network. The mobile terminal for the responsible person, as described in this application, is responsible for receiving hot spot risk alarm information and sending maintenance scheduling requests to the system in high-risk scenarios. The above implementation topology constitutes a complete closed loop of the method of this application, from data acquisition, hot spot risk assessment, and load reduction control command issuance to the response of the maintenance terminal.

[0025] Based on the above implementation topology, the specific implementation steps of the photovoltaic power plant infrared image feature extraction and anomaly diagnosis method of this application are as follows: See Figure 1 This is a flowchart illustrating the method for infrared image feature extraction and anomaly diagnosis of photovoltaic power plants provided in this application embodiment. Figure 1 The execution entity of the method shown can be a software and / or hardware device. The execution entity of this application can include, but is not limited to, at least one of the following: user equipment, network equipment, etc. User equipment can include, but is not limited to, computers, smartphones, personal digital assistants (PDAs), and the aforementioned electronic devices. Network equipment can include, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers. Cloud computing is a type of distributed computing, consisting of a super virtual computer composed of a group of loosely coupled computers. This embodiment does not limit this. Steps 100 to 400 are included: Step 100: Divide each photovoltaic array into a first subarray, a second subarray, and a third subarray according to the physical connection structure of the photovoltaic array.

[0026] It should be noted that the photovoltaic array is composed of multiple photovoltaic modules connected in series according to a physical connection sequence, and the arrangement position of each module in the array is fixed. In this step, based on the physical connection structure of the photovoltaic array, the entire row of modules is divided into three sub-arrays: the first segment, the middle segment, and the last segment. The second sub-array is located between the first and third sub-arrays and is used as an internal reference segment in step 320.

[0027] In some embodiments, step 100 includes steps 110 to 120: Step 110: Determine the arrangement position of each component in the photovoltaic array according to the physical connection order. Divide all the components of the photovoltaic array into three equal segments according to their arrangement positions. The components located in the first segment are designated as the first sub-array, the components located in the third segment are designated as the third sub-array, and the components located in the second segment are designated as the second sub-array.

[0028] Step 120: In response to the fact that the number of components in each segment cannot be divided equally, the remaining components are assigned to the second subarray.

[0029] It is easy to understand that by assigning the surplus components to the second subarray, the number of components in the first and third subarrays is kept equal, thereby ensuring that the representative temperature values ​​of the first and third subarrays in step 311 are comparable.

[0030] Preferably, after step 100 is completed, each component in the photovoltaic array has been assigned to the first subarray, the second subarray, or the third subarray, and the three-segment division result remains unchanged throughout steps 200 to 400.

[0031] Step 200: Based on the deployment ball at the photovoltaic power station site, acquire temperature monitoring data of the first subarray, the second subarray, and the third subarray in multiple consecutive sampling periods to obtain the temperature distribution change sequence.

[0032] It should be noted that existing hot spot detection methods mainly fall into two categories: one is based on electrical parameter monitoring, which requires the construction of a large number of physical sensing circuits around the photovoltaic modules. For large-scale photovoltaic power plants, the deployment cost is high and the detection efficiency decreases with increasing scale. The other is based on pixel-by-pixel analysis of infrared images, which relies on fixed thresholds to judge the absolute temperature value in a single frame of image. This method is prone to generating a large number of false alarms when the overall ambient temperature rises, and prone to missed alarms when the temperature accumulation of hot spots is small in the early stages. This application eliminates the need for building external physical circuits and pixel-by-pixel image processing. Instead, it utilizes the temperature distribution characteristics of the subarray inherent in the series structure for judgment, fundamentally avoiding the limitations of the above two methods.

[0033] It is worth mentioning that the monitoring ball is installed and deployed at the photovoltaic power station site in a fixed manner. The infrared thermal imager on it continuously captures infrared image frames of the photovoltaic array according to a preset sampling cycle. The collected data is stored in the storage device and then transmitted to the control console. After the control console completes the extraction and processing of temperature monitoring data, it uploads the risk identification data to the integrated management platform. The integrated management platform then pushes the alarm information to the mobile terminal of the person in charge through the wireless network, that is, the operation and maintenance terminal mentioned in this application.

[0034] However, during long-term outdoor operation, the PTZ camera may be affected by shading or direct sunlight, causing overall distortion of the temperature value of a subarray region in a single frame of infrared image. Directly incorporating the distorted frame data into the temperature distribution change sequence would introduce spurious fluctuations, interfering with the hotspot risk assessment in step 300. Therefore, before merging the temperature values, this step first performs a validity assessment of each infrared image frame based on the time-segment subarray temperature difference benchmark interval. Steps 210 to 222 are executed iteratively for each sampling period until temperature monitoring data from multiple consecutive sampling periods have been obtained. Then, step 230 arranges the data to obtain the temperature distribution change sequence.

[0035] It is worth noting that hot spots appear in infrared images as small, bright areas with significantly higher grayscale values ​​than the surrounding areas and clear edges, mostly distributed at the edges of components. The temperature anomalies of genuine hot spots only affect local components and do not cause a shift in the overall temperature level of the entire subarray; however, external interference (such as shading or strong light) can cause an overall temperature shift in a certain subarray area, resulting in abnormal differences in the overall temperature levels between the three subarrays. It is precisely based on these fundamental differences in the spatial distribution patterns of genuine hot spots and external interference that we can use whether the overall temperature level difference data between the three subarrays falls within the time-limited subarray temperature difference baseline range as a validity criterion, distinguishing between interfering frames and valid frames.

[0036] In some embodiments, step 200 includes steps 210 to 230: Step 210: Retrieve infrared image frames collected by the infrared thermal imager mounted on the control ball according to the preset sampling period, and obtain the time period subarray temperature difference reference range of the photovoltaic array from the acquisition record of the control ball according to the acquisition time period corresponding to the infrared image frame.

[0037] The preset sampling period is the collection interval time that the personnel pre-configure for the monitoring ball. It can be a shorter interval set according to the peak power generation period of the photovoltaic power station, or a longer interval set according to the off-peak power generation period. The specific interval is determined by the operation and maintenance personnel according to the actual monitoring needs.

[0038] It is worth mentioning that, because the installation position of the surveillance PTZ camera is fixed, the shooting angle and lighting conditions remain consistent across different dates at the same time. Utilizing this characteristic, historical sampling period data with the same time frame as the current acquisition period are extracted from the PTZ camera's data collection records. The numerical range corresponding to this historical data is then used as the baseline interval for the temperature difference of the subarray during the time period. This ensures that the baseline interval accurately reflects the normal acquisition characteristics of the current time period, rather than using a uniform interval throughout the day, thus effectively avoiding misjudgments caused by natural temperature variations during the time period.

[0039] In some embodiments, step 210, which involves obtaining the time-segment subarray temperature difference reference interval of the photovoltaic array from the acquisition record of the control ball based on the acquisition time segment corresponding to the infrared image frame, includes steps 211 and 212: Step 211: Retrieve the overall temperature level difference data between the first subarray, the second subarray, and the third subarray in each historical sampling period that is the same as the current sampling time from the acquisition record of the control ball, and determine the numerical range corresponding to the overall temperature level difference data as the reference interval for the temperature difference of the subarray in the time period.

[0040] The overall temperature level difference data refers to the data obtained by statistically analyzing the differences between the overall temperature levels of all components in the first, second, and third subarrays within the current sampling period. The overall temperature level can be the mean or the median of the temperature values ​​of all components in each subarray, which is determined by the maintenance personnel based on the actual situation.

[0041] Understandably, since photovoltaic arrays are composed of series-connected modules, the overall temperature levels of the first, second, and third subarrays are similar under normal sampling conditions within the same time period. Therefore, the overall temperature level difference data among the three subarrays exhibits a stable distribution throughout the historical sampling period. Determining the numerical range corresponding to the historical overall temperature level difference data as the benchmark interval for subarray temperature differences during the current sampling period reflects the pattern of temperature level differences among the three subarrays under normal shooting conditions, providing a basis for determining the validity of step 220.

[0042] Step 212: In response to the absence of a historical sampling period with the same time in the acquisition records of the control ball during the current acquisition period, obtain the overall temperature level difference data between the first subarray, the second subarray, and the third subarray within the historical sampling period closest to the current acquisition time in the acquisition records of the control ball, and determine the numerical range corresponding to the overall temperature level difference data as the reference interval for the temperature difference of the subarrays during the time period.

[0043] It is easy to understand that step 212 handles the case of the first acquisition by the surveillance sphere or the case of missing acquisition records. Since the most recent historical sampling period is closest to the lighting conditions of the current acquisition period, the overall temperature level difference data corresponding to the most recent historical sampling period is used as a substitute. This ensures that the temperature difference benchmark range of the subarray can be obtained in any acquisition period, and the judgment process of step 220 will not be interrupted due to missing records.

[0044] It should be noted that steps 211 and 212 together ensure that the temperature difference reference range of the subarray during any acquisition period has corresponding historical data support. Step 211 covers the case where the acquisition records are complete, and step 212 covers the case where the acquisition records are missing. The two work together to ensure that the execution of step 210 is not constrained by the completeness of the acquisition records.

[0045] Step 220: Based on the comparison results between the overall temperature level difference data between the first subarray, the second subarray and the third subarray in the infrared image frame and the temperature difference benchmark interval of the subarray in the time period, the temperature monitoring data of the current sampling period is obtained.

[0046] It should be noted that the overall temperature level difference data in step 220 is obtained by extracting the overall temperature levels of the first subarray, the second subarray, and the third subarray in the current infrared image frame according to the method defined in step 211, and then statistically analyzing the differences between the three. The generation method is consistent with the generation method of the historical overall temperature level difference data in step 211, thereby ensuring that the two are directly comparable.

[0047] In some embodiments, step 220 includes steps 221 and 222: Step 221: In response to the overall temperature level difference data being within the temperature difference reference range of the subarray during the time period, the infrared image frame is determined as a valid image frame, and the temperature values ​​are merged according to the division of the first subarray, the second subarray, and the third subarray to obtain the temperature monitoring data of the current sampling period.

[0048] It is easy to understand that the overall temperature level difference data is within the reference range of the temperature difference of the subarrays in the time period, indicating that the temperature level difference of the three subarrays in the current frame conforms to the historical normal acquisition pattern, and the infrared image frame is not affected by external interference. Therefore, reliable temperature monitoring data for the current sampling period can be obtained by directly dividing and merging the temperature values ​​according to the three subarrays.

[0049] Step 222: In response to the fact that the overall temperature level difference data is not in the temperature difference reference range of the subarray during the time period, the infrared image frame is identified as an interference image frame. The temperature monitoring data of the previous effective sampling period and the next effective sampling period are interpolated to obtain the temperature monitoring data of the current sampling period.

[0050] It's easy to understand that if the overall temperature level difference data exceeds the baseline range of temperature difference in the subarray for a given time period, it indicates overall distortion in the current frame. Directly using the current frame data would introduce spurious fluctuations into the temperature distribution change sequence. By interpolating the current sampling period with temperature monitoring data from the previous and next effective sampling periods, the missing data corresponding to the interfering frames can be filled in without interrupting the temperature distribution change sequence, thus maintaining the continuity of the temperature distribution change sequence.

[0051] Step 230: Arrange the temperature monitoring data obtained from each sampling period in chronological order to obtain a temperature distribution change sequence.

[0052] Preferably, step 200 uses the temperature difference reference interval of the time subarray to determine the validity of the infrared image frame, and incorporates the identification and removal of interference frames into the data acquisition process. This ensures that the temperature monitoring data of each sampling period in the temperature distribution change sequence comes from valid image frames or reliable data that has been interpolated, thus guaranteeing the accuracy of the hot spot risk judgment in step 300 from the data source.

[0053] Step 300: Based on the hot spot risk judgment rules, perform hot spot risk judgment on the temperature distribution change sequence to obtain risk identification data.

[0054] It should be noted that, as Figure 3 As shown, the hot spot risk assessment rule collects image data and corresponding label data from surveillance cameras deployed at the photovoltaic power station site. The image data is input into a deep learning network for feature extraction. After model prediction, the loss is calculated together with the label data, and the model parameters are continuously updated. After multiple rounds of training, a mature rule for hot spot risk assessment is formed. During the training process of the hot spot risk assessment rule, it was found that existing detection methods based on setting a fixed threshold based on the absolute temperature value of a single frame image have significant limitations: the overall temperature level of the photovoltaic array fluctuates greatly at different sampling times due to factors such as light intensity and ambient temperature. When the overall ambient temperature rises, the temperature of all components in the array rises synchronously, and the fixed threshold method will produce a large number of false alarms in this case; while in the early stages of a hot spot, the local heat accumulation is small, and the component temperature has not yet exceeded the fixed threshold, making it easy to miss detections. Therefore, the hot spot risk assessment rule does not use the comparison between the absolute temperature value and a fixed threshold as the basis for judgment. Instead, it utilizes the inherent characteristics of the photovoltaic array series structure, namely, the local heat accumulation caused by hot spots will lead to a continuous increase in the temperature asymmetry between the first and last sub-arrays, and at the same time, the temperature distribution dispersion between the components in the array will increase synchronously. Therefore, this application judges the hot spot risk through the linkage of two indicators, thereby distinguishing the actual hot spot accumulation from the overall fluctuation of the ambient temperature.

[0055] In some embodiments, step 300 includes steps 310 to 340: Step 310: Obtain the subarray temperature difference data and array temperature discrete data corresponding to each sampling period based on the temperature distribution change sequence.

[0056] It should be noted that the subarray temperature difference data reflects the degree of temperature asymmetry between the first and third subarrays, while the array temperature discrete data reflects the overall deviation of the temperature values ​​of each module within the photovoltaic array from the overall temperature reference value. These two sets of data describe the temperature distribution of the photovoltaic array from different dimensions: the subarray temperature difference data captures the temperature asymmetry at the beginning and end of the series structure, while the array temperature discrete data captures the concentration of temperature distribution among the modules. Together, they form the data basis for the dual-indicator linkage judgment in step 320.

[0057] It should be noted that step 310 is executed cyclically for each sampling period in the temperature distribution change sequence, and the subarray temperature difference data and array temperature discrete data corresponding to each sampling period are obtained in turn, providing data input for the period-by-period judgment in step 320.

[0058] In some embodiments, step 310 includes steps 311 and 312: Step 311: Retrieve the representative temperature values ​​of the first subarray and the third subarray in the current sampling period and compare them to obtain the subarray temperature difference data in the current sampling period.

[0059] The representative temperature value refers to a statistically significant value obtained after analyzing the temperature values ​​of all components within the subarray. It can be the average or the median of the temperature values ​​of all components within the subarray, determined by the maintenance personnel based on the actual situation. The subarray temperature difference data refers to the difference between the representative temperature values ​​of the first subarray and the representative temperature values ​​of the third subarray.

[0060] Understandably, hot spots are typically caused by localized shading, cell damage, or other factors, and are spatially distributed around the edges of the module. Since photovoltaic arrays are connected in series according to their physical connection sequence, the location of the hot spot naturally divides the array into two segments. The overall temperature level of the segment containing the hot spot will continuously rise during the heat accumulation process, leading to a continuously increasing difference in the representative temperature values ​​between the first and third sub-arrays. Therefore, using the difference in the representative temperature values ​​between the first and third sub-arrays as the sub-array temperature difference data can effectively capture the temperature asymmetry between the beginning and end caused by the hot spot.

[0061] Step 312: Based on the deviation between the temperature value of each component in the photovoltaic array during the current sampling period and the overall temperature reference value of the photovoltaic array, determine the discrete array temperature data for the current sampling period.

[0062] The overall temperature reference value refers to the average temperature value of all components within the photovoltaic array, calculated from the temperature values ​​of each component retrieved in step 311. The array temperature discrete data refers to the overall deviation of the temperature values ​​of each component within the photovoltaic array from the overall temperature reference value, reflecting the concentrated or discrete state of the temperature distribution of the components within the array.

[0063] It should be noted that when the overall ambient temperature rises, the temperature values ​​of each component within the photovoltaic array rise synchronously, and the deviation between the temperature values ​​of each component and the overall temperature reference value does not increase significantly, keeping the array temperature dispersion data relatively stable. However, when hot spots appear, the temperature value of the component containing the hot spot will remain higher than the overall temperature reference value, causing the deviation of the temperature values ​​of each component relative to the overall temperature reference value to increase, and the array temperature dispersion data will rise accordingly. Therefore, the array temperature dispersion data can effectively reflect the local temperature anomalies caused by hot spots, and together with the sub-array temperature difference data, it forms the basis for a dual-indicator linkage judgment.

[0064] Step 320: In response to the fact that the subarray temperature difference data in the current sampling period is greater than the subarray temperature difference data in the previous sampling period, and the array temperature discrete data in the current sampling period is greater than the array temperature discrete data in the previous sampling period, mark the current sampling period as a double index rising period, and accumulate the count of consecutively marked double index rising periods.

[0065] It should be noted that an isolated increase in the subarray temperature difference data does not directly indicate hot spot risk. In other words, when the ambient temperature changes, even a slight difference in light intensity between the locations of the first and third subarrays may cause a temporary increase in the subarray temperature difference data. However, in this case, the temperature distribution of the components within the array remains relatively concentrated, and the discrete array temperature data will not increase synchronously. Conversely, an isolated increase in the discrete array temperature data may stem from random fluctuations in individual components and does not represent a systemic heat accumulation process. Only when the subarray temperature difference data and the discrete array temperature data increase simultaneously within the same sampling period can it indicate the existence of continuous local heat accumulation within the array, serving as a valid signal of hot spot risk. Therefore, this step uses the simultaneous increase of both data within the same sampling period as a marking condition. Sampling periods that meet this condition are marked as dual-indicator rise periods, and the count of consecutively marked dual-indicator rise periods is accumulated, providing a quantitative basis for confirming hot spot risk in step 340 after reaching the preset period threshold.

[0066] In some embodiments, step 320 includes steps 321 to 323: Step 321: Compare the temperature representative values ​​of the first subarray, the second subarray, and the third subarray in the current sampling period with the temperature representative values ​​corresponding to the previous sampling period to obtain the temperature representative value change data of each subarray.

[0067] The temperature representative value change data for each subarray refers to the difference between the temperature representative value of the first subarray, the second subarray, and the third subarray in the current sampling period and the temperature representative value in the previous sampling period, respectively denoted as the temperature representative value change data of the first subarray, the temperature representative value change data of the second subarray, and the temperature representative value change data of the third subarray.

[0068] Step 322: Using the temperature representative value change data of the second subarray as an internal reference, compare the temperature representative value change data of the first subarray with the temperature representative value change data of the second subarray to obtain the first comparison result. Compare the temperature representative value change data of the third subarray with the temperature representative value change data of the second subarray to obtain the second comparison result.

[0069] The first comparison result refers to the judgment on whether the temperature representative value change data of the first subarray is equal to the temperature representative value change data of the second subarray, and the second comparison result refers to the judgment on whether the temperature representative value change data of the third subarray is equal to the temperature representative value change data of the second subarray.

[0070] It should be noted that step 322 uses the second subarray as an internal reference because the second subarray is located between the first and third subarrays. Before the hot spot spreads to the middle section, the temperature change of the second subarray best represents the temperature fluctuation benchmark of the entire array under normal operating conditions. If the overall ambient temperature rises, the representative temperature values ​​of the first, second, and third subarrays will rise synchronously, and the temperature representative value changes of the three subarrays should be consistent with each other. If the hot spot forms in the first or last section, the temperature representative value changes of the first or third subarray will be higher than those of the second subarray, resulting in unequal conclusions in the first or second comparison results. Therefore, by using the temperature representative value change data of the second subarray as an internal reference, no external threshold needs to be introduced; the two sources can be distinguished solely by the internal comparison between the three subarrays within the array.

[0071] Step 323: Based on the first comparison result and the second comparison result, determine the source of the rise in the subarray temperature difference data of the current sampling period, obtain the rise source determination data, and decide whether to mark the current sampling period as a dual index rise period.

[0072] It should be noted that the data for determining the source of the rise is a conclusion drawn from the combined results of the first and second comparisons, including two determinations: the source of the rise being the synchronous change in the overall temperature level of each component of the photovoltaic array and the source of the rise being local heat accumulation.

[0073] In some embodiments, step 323 includes steps A1 and A2: Step A1: In response to the fact that both the first comparison result and the second comparison result indicate that the temperature representative value change data of the first subarray and the third subarray are equal to the temperature representative value change data of the second subarray, it is determined that the source of the increase in the subarray temperature difference data in the current sampling period is the synchronous change of the overall temperature level of each component of the photovoltaic array, and the current sampling period is not marked as a dual index increase period.

[0074] It's easy to understand that the first and second comparison results are equal, meaning that the temperature representative values ​​of the first, second, and third subarrays change at the same rate relative to the previous sampling period, indicating that the three subarrays are heating synchronously. With synchronous heating, the temperature difference between the first and third subarrays will not increase due to heat accumulation. The increase in the subarray temperature difference data originates from overall ambient temperature fluctuations rather than localized heat accumulation. Therefore, the current sampling period is not marked as a dual-indicator rising period to avoid false alarms.

[0075] Step A2: In response to the first comparison result or the second comparison result indicating that the temperature representative value change data of the first subarray or the third subarray is not equal to the temperature representative value change data of the second subarray, determine that the source of the increase in the subarray temperature difference data in the current sampling period is local heat accumulation, and mark the current sampling period as the dual index rise period.

[0076] It is easy to understand that if the first or second comparison results are unequal, it means that the temperature representative value change of the first or third subarray deviates from that of the second subarray, indicating that additional heat accumulation has occurred in the first or last segment, independent of the overall temperature rise. This additional heat accumulation is an early manifestation of the hot spot effect in a series structure. Therefore, the current sampling period is marked as a dual-index rising period and included in the continuous marking count.

[0077] It should be noted that steps 321 to 323 introduce the temperature representative value change data of the second subarray as an internal reference, changing the judgment of the source of the rise in the subarray temperature difference data in step 320 from comparison with an external threshold to internal comparison between the three subarray segments within the array. This does not rely on any preset parameters, but only utilizes the inherent characteristics of the serial structure itself to complete the judgment, making the marking results of the dual index rise cycle more accurate and reliable.

[0078] In other embodiments, step 320 may be specifically implemented as steps B1 to B3 as follows: Step B1: Compare the temperature representative values ​​of the first subarray, the second subarray, and the third subarray in the current sampling period with the temperature representative values ​​corresponding to the previous sampling period to obtain the temperature representative value change data of each subarray.

[0079] Step B2: In response to the fact that the temperature representative values ​​of the first subarray, the second subarray, and the third subarray are all increasing and the increase is within the preset synchronization error range, it is confirmed that the temperature representative values ​​of the first subarray, the second subarray, and the third subarray are rising in an overall consistent manner. It is determined that the source of the increase in the subarray temperature difference data in the current sampling period is the synchronous change of the overall temperature level of each component of the photovoltaic array, and the current sampling period is not marked as a dual-index rising period.

[0080] The preset synchronization error range is the maximum deviation range between the representative temperature values ​​of each subarray that is allowed to vary in advance by personnel. It is used to tolerate slight inconsistencies caused by individual differences in components and can be determined by operation and maintenance personnel based on the actual operating conditions of the photovoltaic power station.

[0081] Step B3: In response to the fact that the increase in the temperature representative value change data of the first subarray, the second subarray and the third subarray exceeds the preset synchronization error range, it is determined that the source of the increase in the subarray temperature difference data in the current sampling period is local heat accumulation, and the current sampling period is marked as the dual index rise period.

[0082] It should be noted that steps B1 to B3 and steps 321 to A2 have the same judgment objective: to distinguish whether the increase in the molecular array temperature difference data originates from overall synchronous change or local heat accumulation. The difference lies in the following: steps 321 to A2 use the change data of the representative temperature value of the second subarray as an internal reference, directly comparing whether the change data of each segment are equal to complete the judgment, without introducing any preset parameters; steps B1 to B3 use a preset synchronization error range as the judgment boundary, judging whether the increase in the change data of the representative temperature values ​​of the three subarrays is within the preset synchronization error range, requiring maintenance personnel to pre-set the preset synchronization error range. Maintenance personnel can choose one of the implementation methods according to the actual situation.

[0083] Step 330: In response to any data in the subarray temperature difference data or array temperature discrete data of the current sampling period not being greater than the corresponding data of the previous sampling period, clear the dual index rising period count.

[0084] It should be noted that the dual-index rise cycle count records the number of sampling cycles in which the condition of simultaneous rise of both indices is met consecutively. If any drop occurs in the subarray temperature difference data or the array temperature discrete data in the current sampling cycle, it indicates an interruption in the heat accumulation process, and the previously accumulated continuous upward trend is no longer continuous. Therefore, the dual-index rise cycle count is reset to zero, and the statistics begin again. This reset operation ensures that the criterion for reaching the preset cycle threshold in step 340 is a continuous and uninterrupted rise of both indices, rather than intermittent and sporadic rises, thus effectively preventing misjudgments caused by transient fluctuations.

[0085] Step 340: In response to the continuous marking of the dual index rising cycle count reaching the preset cycle threshold, the risk of hot spots in the photovoltaic array is confirmed, and risk identification data is obtained.

[0086] The preset cycle threshold is the minimum number of consecutive dual-indicator rise cycles that personnel pre-set for hot spot risk confirmation. It can be determined by the operation and maintenance personnel based on the actual operating conditions of the photovoltaic power station and their tolerance for false alarm rates. The larger the preset cycle threshold, the higher the requirement for the continuity of the heat accumulation process and the lower the false alarm rate, but the response speed to early hot spots will be correspondingly reduced; the smaller the preset cycle threshold, the faster the response speed, but the correspondingly higher the false alarm rate.

[0087] Understandably, the hot spot effect is a gradual accumulation process; the local temperature needs to undergo multiple sampling cycles of continuous thermal accumulation to go from a slight anomaly to a significant increase. Requiring the count of both indicators to reach a preset cycle threshold before confirming hot spot risk means that only when the condition of simultaneous increases in both indicators is met for multiple consecutive sampling cycles can risk identification data be obtained, thus distinguishing occasional fluctuations from genuine hot spot accumulation.

[0088] In some embodiments, step 340, in response to the dual-index rising cycle count of the continuously marked markers reaching a preset cycle threshold, confirms the presence of hot spot risk in the photovoltaic array and obtains risk identification data, and further includes steps 341 to 343: Step 341: Obtain the growth rate of the subarray temperature difference data between adjacent sampling periods of the photovoltaic array from the acquisition records of the control ball, and determine the numerical range of the growth rate of the subarray temperature difference data corresponding to each adjacent sampling period as the reference interval for the growth of the subarray temperature difference.

[0089] The growth rate of subarray temperature difference data refers to the difference between the subarray temperature difference data in two adjacent sampling periods, reflecting the growth rate of the subarray temperature difference data within a single sampling period interval. The subarray temperature difference growth reference interval is defined as the range of values ​​corresponding to the growth rates of all historical subarray temperature difference data obtained from the monitoring sphere's data acquisition records for each adjacent historical sampling period of the photovoltaic array. This range reflects the growth pattern of the subarray temperature difference data under normal operating conditions.

[0090] It should be noted that step 340 has already confirmed the hot spot risk by counting the dual-indicator rise cycle to the preset cycle threshold. Steps 341 to 343 further determine the severity of the hot spot risk based on this, namely, whether the increase in the subarray temperature difference data has exceeded the historical range of the photovoltaic array under normal operating conditions, and whether the sampling period for the increase exceeding the historical range simultaneously meets the condition of the dual-indicator rise. Only when both conditions are met can it be confirmed that the hot spot accumulation has entered the accelerated stage, and the risk level is upgraded to a high-risk level.

[0091] Step 342: In response to the fact that the growth rate of the subarray temperature difference data between the current sampling period and the previous sampling period exceeds the reference range for the growth of the subarray temperature difference, the current sampling period is marked as an abnormal growth period.

[0092] It is easy to understand that the growth rate of the subarray temperature difference data exceeds the reference range for the growth of the subarray temperature difference data, indicating that the growth rate of the subarray temperature difference data in the current sampling period has exceeded the growth range of the photovoltaic array under normal historical operating conditions. The heat accumulation process has shown signs of acceleration, and the current sampling period is marked as an abnormal growth period, providing a basis for marking the cross-validation in step 343.

[0093] Step 343: In response to the presence of sampling periods that are simultaneously marked as dual-indicator rise periods in the continuously marked abnormal growth periods, it is confirmed that the growth rate of the subarray temperature difference data exceeds the normal range of the photovoltaic array and the source of the rise in the subarray temperature difference data has been ruled out as synchronous change in the overall temperature level of each component of the photovoltaic array. The risk level of the risk identification data is determined to be high risk level, and high risk identification data is obtained.

[0094] It should be noted that step 343 requires an overlap between the abnormal growth cycle and the dual-indicator rise cycle. This is because the abnormal growth cycle only indicates an abnormal growth rate in the subarray temperature difference data, but this abnormal growth rate could also stem from temporary environmental interference. The dual-indicator rise cycle, on the other hand, has already been confirmed in step 320 to show that the subarray temperature difference data and the array temperature discrete data are rising simultaneously, and step 323 has ruled out interference from overall synchronous changes. When both markers simultaneously hit the same sampling period, it indicates that the growth rate of the subarray temperature difference data in that sampling period exceeds the historical normal range, and the source of the growth is indeed local heat accumulation rather than environmental interference. Therefore, the risk level of the risk identification data is upgraded to a high-risk level, resulting in high-risk identification data.

[0095] Preferably, steps 341 to 343 combine the reference range of the subarray temperature difference growth with the cross-validation of the dual index rise cycle, further distinguishing between ordinary risk and high risk based on the hot spot risk confirmed in step 340, so that the risk identification data carries risk level information, providing a basis for the differentiated response of the operation and maintenance terminal in steps 344 to 346.

[0096] In some embodiments, steps 344 to 346 are further included after step 343: Step 344: Send the high-risk identification data to the operation and maintenance terminal, and receive the operation and maintenance scheduling request sent by the operation and maintenance terminal based on the high-risk identification data.

[0097] It should be noted that the high-risk identification data is pushed to the responsible person's mobile terminal, i.e., the operation and maintenance terminal described in this application, via a wireless network through the integrated management platform. After receiving the high-risk identification data, the operation and maintenance personnel can determine whether manual intervention is required based on the photovoltaic array location identifier, risk level, and other information carried in the high-risk identification data, and send an operation and maintenance scheduling request to the system. The key difference between step 344 and ordinary risk alarms is that ordinary risks only trigger the system to automatically reduce load, while high risks also trigger active intervention by operation and maintenance personnel. The two responses complement each other, ensuring timely handling during the accelerated accumulation phase of hot spots.

[0098] Step 345: Obtain the photovoltaic array location identifier carried in the operation and maintenance scheduling request, and match the photovoltaic array location identifier with the photovoltaic array location identifiers with high risk levels in the high risk identification data to obtain location matching data.

[0099] Among them, the photovoltaic array location identifier refers to the unique number or coordinate information that identifies the deployment location of each photovoltaic array in the photovoltaic power plant, which is pre-entered by the operation and maintenance personnel during the initial system configuration. The location matching data refers to the matching conclusion obtained by comparing the photovoltaic array location identifier carried in the operation and maintenance scheduling request with the photovoltaic array location identifiers with high-risk levels in the high-risk identification data, including two results: successful matching and failed matching.

[0100] It is easy to understand that step 345 performs location matching verification before executing the scheduling operation. The purpose is to prevent operation errors by maintenance personnel. If the photovoltaic array location identifier carried in the maintenance scheduling request is inconsistent with the high-risk photovoltaic array location identifier recorded in the high-risk identification data, it means that the target array of the maintenance scheduling request is not a high-risk array at present. In this case, the scheduling operation should not be executed to avoid unnecessary interference to non-risk arrays.

[0101] Step 346: Generate the corresponding scheduling control command based on the location matching data and the operation type identifier carried in the operation and maintenance scheduling request, and send it to the target scheduling inverter. In response to receiving the execution confirmation returned by the target scheduling inverter, feed back the execution result to the operation and maintenance terminal.

[0102] The operation type identifier refers to the scheduling operation category selected by the maintenance personnel on the maintenance terminal, including two types: manual load reduction and manual recovery. The target scheduling inverter is the inverter corresponding to the target photovoltaic array confirmed by the location matching data.

[0103] In some embodiments, step 346 includes steps C1 and C2: Step C1: In response to a successful location matching data match, obtain the operation type identifier carried in the operation and maintenance scheduling request, and determine the scheduling operation type based on the operation type identifier; in response to a failed location matching data match, return an invalid scheduling prompt to the operation and maintenance terminal.

[0104] It is easy to understand that if the location matching data is successful, it means that the target array of the operation and maintenance scheduling request is consistent with the high-risk photovoltaic array recorded in the high-risk identification data, and the scheduling operation is legal and valid. Therefore, the operation type identifier is obtained to determine the type of scheduling operation. If the location matching data fails, it means that the target array of the operation and maintenance scheduling request is not within the range of the high-risk identification data. An invalid scheduling prompt is returned to the operation and maintenance terminal to remind the operation and maintenance personnel to verify the operation target and avoid misoperation.

[0105] Step C2: In response to the scheduling operation type being manual load reduction, a manual load reduction command is generated and sent to the target scheduling inverter; in response to the scheduling operation type being manual recovery, a manual full load recovery command is generated and sent to the target scheduling inverter.

[0106] It's easy to understand that the generation logic for manual load reduction commands and manual full-load recovery commands is the same as the logic for the system automatically generating load reduction control commands in step 412. The difference lies in the triggering source: the load reduction control commands in step 412 are automatically generated by the system based on risk identification data, while the manual load reduction commands in step C2 are proactively initiated by operations and maintenance personnel based on high-risk identification data. These two load reduction paths complement each other: automatic system load reduction covers ordinary risk scenarios, while manual load reduction covers high-risk scenarios where operations and maintenance personnel need to actively intervene.

[0107] It should be noted that steps C1 and C2, through the combination of location matching verification and operation type branching, not only ensure the accuracy of the operation target of the operation and maintenance scheduling request, but also cover two complete scheduling paths: manual load reduction and manual recovery. This enables operation and maintenance personnel to make precise manual intervention on the target scheduling inverter in high-risk scenarios.

[0108] Preferably, steps 343 to 346 introduce an active intervention path for operation and maintenance personnel outside the system's automatic control process. High-risk identification data serves as a trigger condition, location matching data serves as a verification of the legality of the operation, and operation type identification serves as the basis for execution. Together, these three elements constitute a complete scheduling closed loop for human-machine collaboration in high-risk scenarios, effectively compensating for the limitations of purely automatic control in handling extreme hot spot scenarios.

[0109] Preferably, step 300 distinguishes the actual hot spot accumulation from the overall fluctuation of the ambient temperature through dual-indicator linkage judgment, introduces the internal reference of the second sub-array through steps 321 to A2 to further eliminate the interference of overall synchronous change, subdivides the hot spot risk into different levels through steps 341 to 343, and establishes a human-machine collaborative scheduling closed loop in high-risk scenarios through steps 344 to 346, thus realizing a complete judgment chain from hot spot risk identification to differentiated response.

[0110] Step 400: In response to the hot spot risk confirmed by the risk identification data, a load reduction control command is generated and sent to the corresponding inverter, and in response to the elimination of the hot spot risk, a full load recovery command is generated and sent to the corresponding inverter.

[0111] It's important to note that a photovoltaic (PV) array consists of multiple modules connected in series to an inverter, which converts the direct current (DC) generated by the PV array into alternating current (AC) output. When a hot spot risk occurs, the module containing the hot spot becomes a high-impedance node in the series circuit due to localized overheating. Current flowing through the hot spot module exacerbates its temperature rise, creating a positive feedback heat accumulation process. Reducing the inverter's output current, i.e., reducing the operating current flowing through the hot spot module, slows down the heat accumulation rate, thus protecting it from further damage. It's worth mentioning that an inverter typically connects multiple PV arrays. If only some PV arrays are at risk of hot spots, a full reduction in the inverter's output current would cause a decrease in the power generation of all PV arrays connected to the same inverter, resulting in unnecessary power loss. Therefore, this step reduces the target inverter's output current proportionally based on the proportion of PV arrays at risk of hot spots to the total number of PV arrays connected to the target inverter, protecting the hot spot arrays while minimizing the impact on the power generation efficiency of the normal arrays.

[0112] In some embodiments, step 400 includes steps 410 and 420: Step 410: In response to the risk identification data confirming the hot spot risk, generate a load reduction control command and send it to the corresponding inverter.

[0113] It should be noted that step 410 is the conversion link from the hot spot risk judgment result in step 300 to the inverter control action. The risk identification data serves as the triggering basis, and the load reduction control command serves as the output result. The two are connected through steps 411 to 413 to complete the complete chain from risk positioning to command generation and then to execution confirmation.

[0114] In some embodiments, step 410 includes steps 411 to 413: Step 411: Based on the risk identification data, determine the target photovoltaic array with hot spot risk and the target inverter corresponding to the target photovoltaic array, obtain the total number of photovoltaic arrays currently connected to the target inverter and the number of photovoltaic arrays with confirmed hot spot risk, and obtain the risk array proportion data.

[0115] Here, the target photovoltaic array refers to the photovoltaic array confirmed to have hot spot risk in the risk identification data of step 300, and the target inverter refers to the inverter currently connected to the target photovoltaic array. The risk array proportion data refers to the proportion of photovoltaic arrays with confirmed hot spot risk out of the total number of photovoltaic arrays currently connected to the target inverter, which is calculated from the total number of photovoltaic arrays connected to the target inverter and the number of photovoltaic arrays with confirmed hot spot risk among them.

[0116] Understandably, the total number of photovoltaic arrays connected to the target inverter and the connection relationship of each array are pre-entered by the operation and maintenance personnel during the initial system configuration. Step 411 retrieves the connection relationship of the target inverter from the system configuration, compares it with the hot spot risk array recorded in the risk identification data, and counts the number of photovoltaic arrays that have been confirmed to have hot spot risks, thereby obtaining the risk array ratio data.

[0117] Step 412: Obtain the current output current of the target inverter, determine the corresponding ratio of load reduction based on the risk array ratio data, obtain the load reduction target current, encapsulate the load reduction target current into a load reduction control command and send it to the target inverter.

[0118] Among them, the load reduction target current refers to the target current value obtained by reducing the current output current according to the proportion of the risk array data, and is calculated from the current output current and the risk array proportion data.

[0119] It should be noted that in this embodiment of the application, after obtaining the current output current of the target inverter in step 412, the current output current can be recorded as the output current before load derating, and the output current before load derating is used as a comparison benchmark in step D1.

[0120] It should be noted that the load reduction of the current output current is based on the proportion of risk arrays, making the load reduction proportional to the number of hot spot risk arrays. The more risk arrays there are, the greater the impact of heat accumulation on the entire series circuit, and the greater the load reduction; the fewer risk arrays there are, the smaller the load reduction, thus maximizing the preservation of the power generation capacity of non-risk arrays.

[0121] Step 413: In response to receiving the execution confirmation from the target inverter, record the load reduction operation status.

[0122] It should be noted that after recording the load reduction operation status in step 413, the temperature distribution change sequence of the target photovoltaic array is continuously monitored and updated in real time to determine whether the load reduction operation has achieved the expected effect of suppressing heat accumulation. If the temperature difference data of the subarray of the hot spot array fails to fall back within the preset response time period after the load reduction is executed, it indicates that the load reduction is insufficient, and a secondary load reduction control command needs to be generated to further reduce the output current; if the temperature difference data of the subarray has fallen back within the preset response time period, it indicates that the load reduction operation is effective, and the process proceeds to the risk elimination judgment process in step 420.

[0123] In some embodiments, step 413 includes steps D1 to D3: Step D1: Continuously acquire real-time updated data of the temperature distribution change sequence of the target photovoltaic array, compare the subarray temperature difference data of the target photovoltaic array within the preset response period after load reduction with the subarray temperature difference data of the target photovoltaic array recorded at the corresponding time of the output current before load reduction, and obtain the load reduction response comparison data.

[0124] The preset response period is a pre-defined observation duration for the load reduction effect, used to determine whether the suppression effect of the load reduction operation on the heat accumulation process appears within a reasonable time. It can be determined by the operation and maintenance personnel based on the actual operating conditions of the photovoltaic power station. The load reduction response comparison data refers to the conclusion obtained by comparing the subarray temperature difference data of the target photovoltaic array within the preset response period with the subarray temperature difference data of the target photovoltaic array recorded at the corresponding time of the output current before load reduction. It includes two results: one where the temperature has decreased and the other where it has not.

[0125] It's easy to understand that the subarray temperature difference data of the target photovoltaic array recorded at the corresponding moment before load derating reflects the degree of temperature asymmetry in the hot spot array before load derating. After load derating, if heat accumulation is effectively suppressed, the subarray temperature difference data should decrease over time; if the load derating is insufficient and heat accumulation continues, the subarray temperature difference data will not decrease. Using the subarray temperature difference data recorded before load derating as a comparison benchmark, the actual effect of the load derating operation can be accurately determined.

[0126] Step D2: In response to the load reduction response comparison data indicating that the temperature difference data of the subarray of the target photovoltaic array has not dropped within the preset response period, the current output current of the target inverter is reacquired, and a secondary load reduction control command is generated and sent to the target inverter based on the risk array ratio data.

[0127] It is understandable that the subarray temperature difference data did not decrease within the preset response period, indicating that the current load reduction was insufficient to suppress the continued heat accumulation process. Based on the risk array proportion data, the current output current was re-determined for load reduction, resulting in a secondary load reduction target current. A secondary load reduction control command was then generated and sent to the target inverter to further reduce the operating current flowing through the hot spot components until heat accumulation was effectively suppressed.

[0128] Step D3: The response to the load reduction response comparison data shows that the temperature difference data of the subarray of the target photovoltaic array has fallen back within the preset response period, and the risk elimination judgment process is initiated.

[0129] It is easy to understand that the temperature difference data of the subarray has fallen back within the preset response period, indicating that the load reduction operation has effectively suppressed the heat accumulation process, and the temperature asymmetry of the target photovoltaic array has begun to decrease. At this time, the risk elimination judgment process in step 420 is entered, and the temperature distribution change sequence of the target photovoltaic array is continuously monitored and the data is updated in real time to determine whether the hot spot risk has been completely eliminated.

[0130] It should be noted that steps D1 to D3, by introducing load reduction effect verification within a preset response period, extend the load reduction control from a single command issuance to a complete process of closed-loop verification and iterative adjustment. If the load reduction is effective, the risk elimination judgment will be initiated; if the load reduction is insufficient, a second load reduction will be automatically triggered, enabling the load reduction control to have self-correcting capabilities and effectively address situations where a single load reduction is insufficient due to varying degrees of hotspot severity.

[0131] Preferably, step 410 directly correlates the load reduction magnitude with the number of hot spot risk arrays through the risk array proportion data, so as to protect the hot spot arrays while maximizing the power generation capacity of the non-risk arrays, and ensures the load reduction effect through closed-loop verification from step D1 to step D3, making the load reduction control both accurate and reliable.

[0132] Step 420: In response to the elimination of hot spot risk, a full load recovery command is generated and sent to the corresponding inverter.

[0133] It should be noted that the load reduction operation will decrease the output current of all photovoltaic arrays connected to the target inverter. If full load is not restored in time after the hot spot risk is eliminated, it will continue to cause unnecessary power generation losses. Therefore, after confirming that the hot spot risk has been eliminated, step 420 generates a full load restoration command and sends it to the corresponding inverter to restore the inverter output current to the level before the load reduction, thereby restoring the normal power generation capacity of the photovoltaic power station. It is worth mentioning that the judgment of hot spot risk elimination cannot rely solely on the observation results of a single sampling period. It is necessary to meet the condition that the two indicators have not increased for multiple consecutive sampling periods to confirm that the heat accumulation process has truly ended, and to avoid prematurely restoring full load if the hot spot temporarily eases and then re-aggravates.

[0134] In some embodiments, step 420 includes steps 421 to 423: Step 421: Continuously acquire the real-time update data of the temperature distribution change sequence of the photovoltaic array corresponding to the risk identification data, and obtain the subarray temperature difference data and array temperature discrete data of the photovoltaic array in the current sampling period corresponding to the risk identification data based on the real-time update data of the temperature distribution change sequence.

[0135] Among them, the real-time update data of the temperature distribution change sequence refers to the newly added temperature monitoring data obtained after the latest infrared image frame continuously collected in step 200 is processed by steps 210 to 230 after the load reduction is executed. The data is added to the temperature distribution change sequence in chronological order so that the temperature distribution change sequence is kept updated in real time.

[0136] It should be noted that step 421 is executed cyclically on the photovoltaic array corresponding to the risk identification data. In accordance with the methods defined in steps 311 and 312, the subarray temperature difference data and array temperature discrete data are extracted periodically from the real-time updated data of temperature distribution change sequence, providing data input for the periodic judgment in step 422.

[0137] Step 422: In response to the fact that the subarray temperature difference data in the current sampling period is not greater than the subarray temperature difference data in the previous sampling period, and the array temperature discrete data in the current sampling period is not greater than the array temperature discrete data in the previous sampling period, mark the current sampling period as a period in which both indicators do not rise, and accumulate the count of consecutively marked periods in which both indicators do not rise.

[0138] The "dual index non-rising period" refers to the sampling period in which both the subarray temperature difference data and the array temperature discrete data in the current sampling period are not greater than the corresponding data in the previous sampling period. The "dual index non-rising period count" refers to the number of sampling periods that continuously meet the dual index non-rising condition, which is accumulated by step 422 for each sampling period.

[0139] It is understandable that the judgment logic in step 422 is symmetrical to the marking logic of the rising cycle of the two indicators in step 320. Step 320 judges whether heat accumulation continues by accumulating the number of rising cycles of the two indicators, while step 422 judges whether heat accumulation continues to subside by accumulating the number of non-rising cycles of the two indicators. If both indicators do not rise at the same time, it indicates that the degree of temperature asymmetry of the subarray and the degree of temperature dispersion of the components are no longer increasing, and the heat accumulation process has stopped; if the condition of non-rising of the two indicators is met for multiple consecutive sampling cycles, the possibility of heat accumulation restarting after a brief interruption is further ruled out.

[0140] Step 423: In response to the continuous marking of the dual indicators not rising cycle count reaching the preset elimination threshold, confirm that the hot spot risk has been eliminated, generate a full load recovery command and send it to the corresponding inverter, and in response to receiving the execution confirmation returned by the corresponding inverter, record the full load recovery operation status.

[0141] The preset elimination threshold is the minimum number of consecutive cycles in which both indicators have not risen, required for the elimination of hot spot risk. This threshold can be determined by the operation and maintenance personnel based on the actual operating conditions of the photovoltaic power station. A higher preset elimination threshold requires a higher level of continuity in the reduction of heat accumulation and a more conservative judgment on restoring full load, but it can more effectively prevent hot spot recurrence after premature restoration to full load. A lower preset elimination threshold results in a faster response speed to restoring full load, which helps reduce power generation losses caused by load reduction.

[0142] It's easy to understand that when the count of the two indicators not rising reaches the preset elimination threshold, it means that the subarray temperature difference data and the array temperature discrete data have not increased within multiple consecutive sampling periods, the heat accumulation process has continued to subside, and it can be determined that the hot spot risk has been eliminated. A full-load recovery command is generated and sent to the corresponding inverter to restore the inverter output current to the level before load reduction. After receiving the execution confirmation from the corresponding inverter, the full-load recovery operation status is recorded, completing the complete closed loop of load reduction control.

[0143] Preferably, step 420 uses the count of the two indicators not rising for a preset elimination threshold as the trigger condition for restoring full load. This forms a symmetrical closed loop with the judgment logic of the count of the two indicators rising for a preset elimination threshold in step 320, ensuring that the judgment basis for restoring full load is consistent with the judgment basis for hot spot risk confirmation in terms of technical logic, so that the entire process control of load reduction and restoration of full load has inherent consistency.

[0144] Preferably, step 400 achieves precise proportional load reduction through risk array proportion data, introduces closed-loop verification and iterative adjustment of load reduction effect through steps D1 to D3, and establishes a hot spot risk elimination judgment mechanism based on dual-index non-rising cycle counting through steps 421 to 423. This expands the load reduction control from a single command issuance to a complete control closed loop integrating precise load reduction, effect verification, risk elimination judgment and full load restoration, effectively ensuring the reliability of hot spot risk handling and maximizing the power generation benefits of photovoltaic power plants.

[0145] See Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. The electronic device 40 includes: a processor 41, a memory 42, and a computer program; wherein, The memory 42 is used to store the computer program, and the memory may also be flash memory. The computer program is, for example, an application program or functional module that implements the above method.

[0146] The processor 41 is configured to execute the computer program stored in the memory to implement the various steps performed by the device in the above method. For details, please refer to the relevant descriptions in the preceding method embodiments.

[0147] Alternatively, the memory 42 can be either standalone or integrated with the processor 41.

[0148] When the memory 42 is a device independent of the processor 41, the device may further include: Bus 43 is used to connect the memory 42 and the processor 41.

[0149] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for infrared image feature extraction and anomaly diagnosis of photovoltaic power plants, characterized in that, include: Based on the physical connection structure of the photovoltaic array, each photovoltaic array is divided into a first subarray, a second subarray, and a third subarray. Temperature monitoring data of the first, second, and third subarrays in multiple consecutive sampling periods are obtained by using a control ball deployed at the photovoltaic power station site to obtain a temperature distribution change sequence. Based on the hot spot risk assessment rules, the temperature distribution change sequence is assessed for hot spot risk to obtain risk identification data. In response to the risk identification data confirming the hot spot risk, a load reduction control command is generated and sent to the corresponding inverter, and in response to the elimination of the hot spot risk, a full load restoration command is generated and sent to the corresponding inverter.

2. The method according to claim 1, characterized in that, The process of dividing each photovoltaic array into a first subarray, a second subarray, and a third subarray based on the physical connection structure of the photovoltaic array includes: The arrangement positions of each component in the photovoltaic array are determined according to the physical connection order. All components of the photovoltaic array are divided into three equal segments according to their arrangement positions. The components located in the first segment are designated as the first sub-array, the components located in the third segment are designated as the third sub-array, and the components located in the second segment are designated as the second sub-array. In response to the fact that the number of components in each segment cannot be equally divided, the remaining components are assigned to the second subarray.

3. The method according to claim 1, characterized in that, The temperature monitoring data of the first, second, and third subarrays in multiple consecutive sampling periods are obtained based on the control sphere deployed at the photovoltaic power station site, resulting in a temperature distribution change sequence, including: The infrared image frames collected by the infrared thermal imager mounted on the control ball according to the preset sampling period are retrieved, and the time period subarray temperature difference reference range of the photovoltaic array is obtained from the acquisition record of the control ball according to the acquisition time period corresponding to the infrared image frame. Based on the comparison between the overall temperature level difference data between the first subarray, the second subarray and the third subarray in the infrared image frame and the reference interval of the temperature difference between the subarrays in the time period, the temperature monitoring data of the current sampling period is obtained. The temperature monitoring data obtained from each sampling period are arranged in chronological order to obtain a temperature distribution change sequence.

4. The method according to claim 3, characterized in that, The temperature monitoring data for the current sampling period is obtained by comparing the overall temperature level difference data between the first, second, and third subarrays in the infrared image frame with the reference interval for the temperature difference between the subarrays in the specified time period. This includes: In response to the overall temperature level difference data being within the temperature difference reference range of the subarray in the time period, the infrared image frame is determined as a valid image frame, and the temperature values ​​are merged according to the division of the first subarray, the second subarray, and the third subarray to obtain the temperature monitoring data of the current sampling period. In response to the fact that the overall temperature level difference data is not in the temperature difference reference range of the subarray during the time period, the infrared image frame is identified as an interference image frame, and the temperature monitoring data of the previous effective sampling period and the next effective sampling period are interpolated to obtain the temperature monitoring data of the current sampling period.

5. The method according to claim 1, characterized in that, The hot spot risk assessment of the temperature distribution change sequence based on the hot spot risk assessment rule yields risk identification data, including: Based on the temperature distribution change sequence, the subarray temperature difference data and array temperature discrete data corresponding to each sampling period are obtained; In response to the fact that the subarray temperature difference data in the current sampling period is greater than the subarray temperature difference data in the previous sampling period, and the array temperature discrete data in the current sampling period is greater than the array temperature discrete data in the previous sampling period, the current sampling period is marked as a double index rising period, and the count of consecutively marked double index rising periods is accumulated. In response to any data in the current sampling period, either the subarray temperature difference data or the array temperature discrete data, not being greater than the corresponding data in the previous sampling period, the dual-index rising period count is cleared. In response to the continuous marking of the dual-index rising cycle count reaching a preset cycle threshold, the photovoltaic array is confirmed to have a hot spot risk, and the risk identification data is obtained.

6. The method according to claim 5, characterized in that, The step of obtaining subarray temperature difference data and array temperature discrete data corresponding to each sampling period based on the temperature distribution change sequence includes: The representative temperature values ​​of the first subarray and the third subarray in the current sampling period are retrieved and compared to obtain the temperature difference data of the subarrays in the current sampling period. Based on the deviation between the temperature value of each component in the photovoltaic array during the current sampling period and the overall temperature reference value of the photovoltaic array, the discrete temperature data of the array during the current sampling period is determined.

7. The method according to claim 5, characterized in that, The condition that the subarray temperature difference data in the current sampling period is greater than the subarray temperature difference data in the previous sampling period, and the array temperature discrete data in the current sampling period is greater than the array temperature discrete data in the previous sampling period, marks the current sampling period as a dual-index rising period, including: By comparing the temperature representative values ​​of the first subarray, the second subarray, and the third subarray in the current sampling period with the temperature representative values ​​corresponding to the previous sampling period, the temperature representative value change data of each subarray is obtained. Using the temperature representative value change data of the second subarray as an internal reference, the temperature representative value change data of the first subarray is compared with the temperature representative value change data of the second subarray to obtain a first comparison result. The temperature representative value change data of the third subarray is compared with the temperature representative value change data of the second subarray to obtain a second comparison result. Based on the first comparison result and the second comparison result, the source of the rise in the subarray temperature difference data in the current sampling period is determined, the rise source determination data is obtained, and it is decided whether to mark the current sampling period as a dual-index rise period.

8. The method according to claim 7, characterized in that, The step of determining the source of the rise in the subarray temperature difference data of the current sampling period based on the first comparison result and the second comparison result, obtaining the rise source determination data, and deciding whether to mark the current sampling period as a dual-index rise period includes: In response to the fact that both the first comparison result and the second comparison result indicate that the temperature representative value change data of the first subarray and the third subarray are equal to the temperature representative value change data of the second subarray, it is determined that the source of the rise in the temperature difference data of the subarray in the current sampling period is the synchronous change of the overall temperature level of each component of the photovoltaic array, and the current sampling period is not marked as a dual-index rise period. In response to the first comparison result or the second comparison result indicating that the temperature representative value change data of the first subarray or the third subarray is not equal to the temperature representative value change data of the second subarray, it is determined that the rise in the subarray temperature difference data in the current sampling period is due to local heat accumulation, and the current sampling period is marked as the dual index rise period.

9. The method according to claim 5, characterized in that, The step of responding to the continuous marking of the dual-index rising cycle count reaching a preset cycle threshold, confirming the presence of hot spot risk in the photovoltaic array, and obtaining the risk identification data, further includes: The growth rate of the subarray temperature difference data between each adjacent sampling period of the photovoltaic array is obtained from the acquisition records of the control ball, and the numerical range of the growth rate of the subarray temperature difference data corresponding to each adjacent sampling period is determined as the subarray temperature difference growth reference interval. In response to the fact that the growth rate of the subarray temperature difference data between the current sampling period and the previous sampling period exceeds the reference range for the growth of the subarray temperature difference, the current sampling period is marked as an abnormal growth period. In response to the existence of a sampling period that is simultaneously marked as the dual-index rise period in the continuously marked abnormal growth period, it is confirmed that the growth rate of the subarray temperature difference data exceeds the normal range of the photovoltaic array and the source of the rise of the subarray temperature difference data has ruled out the synchronous change of the overall temperature level of each component of the photovoltaic array. The risk level of the risk identification data is determined to be high risk level, and high risk identification data is obtained.

10. The method according to claim 1, characterized in that, The step of responding to the risk identification data confirming hot spot risk and generating a load reduction control command to send to the corresponding inverter includes: Based on the risk identification data, the target photovoltaic array with hot spot risk and the target inverter corresponding to the target photovoltaic array are determined, the total number of photovoltaic arrays currently connected to the target inverter and the number of photovoltaic arrays with confirmed hot spot risk are obtained, and the risk array ratio data is obtained. The current output current of the target inverter is obtained, and the load reduction is determined according to the corresponding proportion of the current output current based on the risk array ratio data to obtain the load reduction target current. The load reduction target current is encapsulated into the load reduction control command and sent to the target inverter. In response to receiving the execution confirmation from the target inverter, the load reduction operation status is recorded.

11. The method according to claim 10, characterized in that, The response to the elimination of hot spot risk generates a full-load recovery command and sends it to the corresponding inverter, including: The temperature distribution change sequence of the photovoltaic array corresponding to the risk identification data is continuously acquired and updated in real time. Based on the temperature distribution change sequence, the subarray temperature difference data and array temperature discrete data of the photovoltaic array corresponding to the risk identification data in the current sampling period are obtained. In response to the fact that the temperature difference data of the subarray in the current sampling period is not greater than the temperature difference data of the subarray in the previous sampling period, and the discrete data of the array temperature in the current sampling period is not greater than the discrete data of the array temperature in the previous sampling period, the current sampling period is marked as a period in which both indicators do not rise, and the count of consecutively marked periods in which both indicators do not rise is accumulated. In response to the continuous marking of the dual indicators not rising cycle count reaching the preset elimination threshold, the hot spot risk is confirmed to be eliminated, the full load recovery command is generated and sent to the corresponding inverter, and in response to receiving the execution confirmation returned by the corresponding inverter, the full load recovery operation status is recorded.

12. The method according to claim 9, characterized in that, After obtaining the high-risk identification data, the process also includes: The high-risk identification data is sent to the operation and maintenance terminal, and the operation and maintenance scheduling request sent by the operation and maintenance terminal based on the high-risk identification data is received. Obtain the photovoltaic array location identifier carried in the operation and maintenance scheduling request, and match the photovoltaic array location identifier with the photovoltaic array location identifiers with high risk level in the high risk identification data to obtain location matching data; Based on the location matching data and the operation type identifier carried in the operation and maintenance scheduling request, a corresponding scheduling control instruction is generated and sent to the target scheduling inverter. In response to receiving the execution confirmation returned by the target scheduling inverter, the execution result is fed back to the operation and maintenance terminal.

13. The method according to claim 12, characterized in that, The step of generating a corresponding scheduling control command based on the location matching data and the operation type identifier carried in the operation and maintenance scheduling request and sending it to the target scheduling inverter includes: In response to a successful match in the location matching data, the operation type identifier carried in the operation and maintenance scheduling request is obtained, and the scheduling operation type is determined based on the operation type identifier; in response to a failed match in the location matching data, an invalid scheduling prompt is returned to the operation and maintenance terminal. In response to the scheduling operation type being manual load reduction, a manual load reduction command is generated and sent to the target scheduling inverter; in response to the scheduling operation type being manual recovery, a manual full load recovery command is generated and sent to the target scheduling inverter.

Images