Photovoltaic performance evaluation system based on cloud computing

Abstract

The invention relates to the technical field of photovoltaic performance evaluation, in particular to a photovoltaic performance evaluation system based on cloud computing, and the system comprises an electrical thermal response interpretation module, an abnormal aggregation distribution module, a rhythm deviation track module, a voltage accumulation deviation module and a risk early warning level judgment module. According to the method, multi-dimensional time sequence fusion is carried out on thermoelectric characteristics and voltage changes through cooperative comparison of a thermal response ratio and an electrical signal, the linkage anomaly identification capability of heterogeneous parameters is greatly improved, spatial structure distribution and number mapping are focused, hierarchical division of an anomaly aggregation phenomenon is realized, dynamic risk insight of a spatial dense area is enhanced, and the method is suitable for mass production. According to the method, a dynamic offset mode is constructed for running data across multiple days and minutes, hidden anomalies such as continuous offset and reverse turning in historical and real-time change trends are accurately captured, voltage trends and accumulated features are refined and filed, a foundation is laid for multi-node and multi-period risk positioning through classification and summarization, and a multi-source synchronous abnormal fragment alignment processing mode is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic performance evaluation, and in particular to a photovoltaic performance evaluation system based on cloud computing. Background Art

[0002] Photovoltaic performance evaluation mainly involves the comprehensive monitoring, analysis and evaluation of the operating status, power output, conversion efficiency and environmental adaptability of photovoltaic power generation systems. This field combines sensor technology, data acquisition systems, meteorological models, electrical modeling and data analysis algorithms to accurately evaluate the actual performance of photovoltaic modules and systems, identify potential faults, improve power generation efficiency, and provide a basis for system optimization and operation and maintenance decisions.

[0003] Among them, the cloud computing photovoltaic performance evaluation system is a system that uses the cloud computing platform to remotely collect, centrally process and intelligently analyze the operating data of the photovoltaic power generation system. It obtains photovoltaic components, grid interfaces and environmental parameter data through distributed data acquisition terminals, and uploads them to the cloud for real-time processing and comprehensive evaluation. Its purpose is to improve the intelligence and automation level of photovoltaic system performance evaluation, support remote monitoring, fault diagnosis, operation optimization and operation and maintenance decision-making, and is widely used in scenarios such as photovoltaic power stations, distributed photovoltaic systems and intelligent energy management platforms.

[0004] Existing technologies generally rely on independent monitoring of a single signal and decentralized data collection, lacking in-depth analysis of parameter fusion and spatial correlation, resulting in the inability to link and identify multiple types of abnormal events between heterogeneous data, making it difficult to track and locate spatial risk points. They rely on periodic collection and one-way trend judgment for a long time, making it difficult to reflect continuous offsets and trend changes in a timely manner, and weak anomalies are easily covered by historical and real-time changes. Voltage data lacks the classification and organization of cumulative features, and the distribution relationship between multiple nodes fails to reveal the risk propagation path. There is a lack of coordinated alignment and interpretation of multi-source data. Risk identification in complex environments is prone to omissions or misjudgments, and performance evaluation remains at the stage of rough judgment and single-point alarm, affecting the reliability of intelligent management and proactive operation and maintenance. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and propose a photovoltaic performance evaluation system based on cloud computing.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a photovoltaic performance evaluation system based on cloud computing, the system comprising: The electrical thermal response interpretation module, based on a cloud computing platform, analyzes the current drop and temperature rise rate data uploaded by PV modules, calculates the ratio and compares it with the same period of the previous day. It identifies the trend of the ratio change, selects key time periods, performs time alignment, and obtains synchronous thermal and electrical anomaly characteristics. The abnormal aggregation distribution module analyzes the abnormal fragment distribution of the components in the region based on the synchronous thermoelectric abnormal characteristics, calculates the cumulative number of abnormal fragments, determines the degree of aggregation, screens the key number intervals, optimizes the correspondence between the numbers and spatial coordinates, and generates regional abnormal aggregation distribution characteristics; The rhythm deviation trajectory module analyzes the current output of photovoltaic modules for multiple consecutive days based on the abnormal cluster distribution characteristics of the region, compares the current sequence with the historical current sequence, calculates the current deviation trajectory, determines the continuous deviation section and its direction change, and outputs the dynamic characteristics of current deviation based on the module number mark position; The voltage cumulative deviation module analyzes the voltage data of each node based on the current offset dynamic characteristics, calculates the voltage change trend, screens the voltage segments with continuous offset and cumulative change characteristics, determines whether the cumulative change is abnormal, and obtains the voltage cumulative offset characteristics.

[0007] The improvements of the present invention are that the synchronous thermoelectric anomaly characteristics include abnormal response ratio type, associated voltage change characteristics, and key time labels; the regional abnormal cluster distribution characteristics include cluster distribution type, spatial density, and number group identification results; the current offset dynamic characteristics include continuous offset category, change direction attribute, and number mapping information; the voltage cumulative offset characteristics include cumulative amplitude attribute, classification results, and node response grouping.

[0008] The present invention is improved in that the electrical thermal response judgment module includes: The ratio sequence calculation submodule is based on the cloud computing platform. It analyzes the current drop amplitude and temperature rise rate data uploaded by the photovoltaic modules, calculates the ratio of the current drop amplitude and temperature rise rate at each time point, compares the ratio changes at consecutive time points, and obtains the ratio sequence trend. The key time period screening submodule determines the changing trend of the ratio sequence in each time period based on the ratio sequence trend and the ratio sequence of the same time period of the previous day, and screens the time period where the ratio undergoes key changes to obtain the key interval of ratio change; The thermoelectric anomaly feature extraction submodule analyzes the voltage change in the corresponding time period based on the ratio change key interval, determines the synchronization features of the ratio change and the voltage change, performs time alignment, and obtains the synchronous thermoelectric anomaly features.

[0009] The present invention is improved in that the abnormal aggregation distribution module includes: The number space matching submodule analyzes the correspondence between the photovoltaic module number and the spatial position based on the synchronous thermoelectric anomaly characteristics, identifies the number of each photovoltaic module, and performs data pairing between the number and the spatial coordinate to obtain the number space correspondence; The abnormal accumulation calculation submodule calculates the total number of abnormal fragments that appear in the spatial range of each number based on the number-space correspondence, analyzes the spatial distribution of the accumulated number of abnormal fragments, compares the cumulative differences under each number, and determines the distribution trend of abnormal fragments with the same area number to obtain the accumulated total number of abnormal fragments; The clustering interval screening submodule screens the cumulative number in each numbered interval according to the cumulative total number of abnormal fragments, compares the spatial distribution data of the numbered intervals, calculates the continuous distribution status of the abnormal fragments in the numbered intervals, calculates the abnormal aggregation degree of the numbered intervals, and generates regional abnormal aggregation distribution characteristics.

[0010] The present invention is improved in that the rhythm deviation trajectory module includes: The current sequence acquisition submodule analyzes the minute-by-minute current data of the photovoltaic modules collected over multiple consecutive days based on the abnormal cluster distribution characteristics of the region. Combined with the module number index, the daily current data is sorted in chronological order. By performing daily classification and archiving operations, the data integrity and continuity are compared to obtain a current sequence set. The deviation segment identification submodule calls the current sequence set, calculates the absolute difference between the current per minute on the current operating day and the current per minute on the same period in history, selects minute segments with continuous daily deviations, determines continuous deviations and direction changes, calculates the characteristic value of the deviation segment, and obtains the current deviation segment; The current offset feature output submodule analyzes the segment range of the current deviation segment, optimizes the component number index, compares the segment feature with the actual position of the component, integrates each type of data, and outputs the current offset dynamic feature.

[0011] The present invention is improved in that the voltage accumulation deviation module includes: The voltage trend calculation submodule analyzes the voltage data continuously collected from each node based on the dynamic characteristics of the current offset, calculates the voltage change amplitude of each node at adjacent collection moments, determines the continuous change of voltage in each collection stage, analyzes the cumulative trend of voltage change in each stage, and obtains the voltage change trend; The continuous offset screening submodule determines whether the voltage increase and decrease directions in each acquisition stage of the voltage change trend are continuously consistent, analyzes the coverage interval and direction continuity of the continuous offset stage, compares the offset direction characteristics between different segments, and screens typical segments with continuous voltage offset to obtain continuous offset segments; The segment abnormality judgment submodule analyzes the voltage change interval of the continuous offset segment, calculates the continuity and fluctuation distribution of the voltage change direction in each segment, determines whether its change performance deviates from the normal state, and obtains the voltage cumulative offset feature.

[0012] The present invention is improved in that the system further comprises: The risk warning and classification module, based on the voltage cumulative offset characteristics, matches the current offset dynamic characteristics and synchronous thermoelectric anomaly characteristics within the same time period, determines the synchronous distribution of abnormal segments in time and component number, analyzes monitoring logs, and combines them with climate fluctuation data to obtain array risk classification indicators; The array risk grading index includes risk level category, risk interval identifier, and associated influencing factors.

[0013] The present invention is improved in that the risk warning and grading module includes: The voltage offset detection submodule analyzes the voltage variation curve of the photovoltaic modules within the same monitoring period based on the voltage cumulative offset characteristics, determines the differences in voltage fluctuation trends between modules, selects the module numbers with abnormal variation trends, and generates a voltage trend distribution; The current and thermoelectric synchronization determination submodule compares the current fluctuation characteristics of the voltage trend distribution corresponding to the period with the thermoelectric anomaly signal, determines the synchronization of the current and thermoelectric signals, filters the data segments with synchronization anomalies, and generates a synchronization anomaly distribution; The array risk grading submodule calculates the abnormal distribution of the synchronization anomaly distribution components within the fluctuation range of monitoring logs and climate data, analyzes the synchronization distribution range of each component and the abnormal data characteristics, optimizes the data correspondence, and obtains the array risk grading index.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, through the coordinated comparison of thermal response ratio and electrical signal, the thermoelectric characteristics and voltage changes are integrated into multi-dimensional time series, which greatly improves the linkage anomaly recognition ability of heterogeneous parameters, focuses on spatial structure distribution and number mapping, realizes hierarchical division of abnormal aggregation phenomena, strengthens dynamic risk insight into spatial dense areas, and constructs dynamic offset patterns across multi-day and minute-level operating data to accurately capture hidden anomalies such as continuous offset and reverse turning in historical and real-time change trends. Voltage trends and cumulative characteristics are refined and archived, and classification and aggregation lay the foundation for multi-node and multi-period risk positioning. The multi-source synchronous abnormal fragment alignment processing method brings global risk grading under the interaction of time, numbering and environmental factors, and improves the refined early warning level of large-scale photovoltaic systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a system flow chart of the present invention; Figure 2 This is a flow chart of the electrical thermal response judgment module in the present invention; Figure 3 This is a flow chart of the abnormal aggregation distribution module in the present invention; Figure 4 This is a flow chart of the rhythm deviation trajectory module in the present invention; Figure 5 This is a flow chart of the voltage accumulation deviation module in the present invention; Figure 6 This is a flow chart of the risk warning and grading module in the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the devices or elements referred to must have a specific direction, be constructed and operate in a specific direction, and therefore should not be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0018] Example See also Figure 1 The present invention provides a technical solution: a photovoltaic performance evaluation system based on cloud computing includes: The electrical thermal response interpretation module, based on a cloud computing platform, analyzes the current drop amplitude and temperature rise rate data uploaded by photovoltaic modules, calculates the ratio of the two at each time point, compares the current ratio sequence with the ratio sequence of the same period of the previous day, determines the changing trend of the ratio over multiple time periods, selects the time period with the key ratio change, and simultaneously analyzes the voltage change information, performs time alignment, and obtains the synchronous thermal and electrical anomaly characteristics; The abnormal aggregation distribution module uses the synchronous thermoelectric anomaly characteristics, calls the photovoltaic module number and spatial position, analyzes the distribution of abnormal fragments of the components in the same area, calculates the cumulative number of abnormal fragments under each number, determines the degree of aggregation of abnormal fragments within the number interval, selects the number interval with the key aggregation phenomenon, optimizes the correspondence between the number and the spatial coordinates, and generates regional abnormal aggregation distribution characteristics; The rhythm deviation trajectory module analyzes the PV module's current output sequence per minute for multiple consecutive days based on the regional abnormal cluster distribution characteristics. Using the cloud platform, it compares the current output per minute on the current day with the historical current sequence for the same period. It calculates the current deviation trajectory, identifies the continuous deviation sections and their direction changes, and combines the module number index mark position to output the dynamic characteristics of the current deviation. The voltage cumulative deviation module analyzes the voltage data of each node based on the dynamic characteristics of current offset, calculates the voltage change trend, screens voltage segments with continuous offset and cumulative change characteristics, determines whether the cumulative change of the segments is abnormal, adjusts the segment classification and archiving method, and obtains the voltage cumulative offset characteristics; The risk warning and classification module is based on the voltage cumulative offset characteristics, matching the current offset dynamic characteristics and synchronous thermoelectric anomaly characteristics within the same period, determining the synchronous distribution of abnormal fragments in time and component number, analyzing the monitoring logs of photovoltaic components on the cloud platform, and combining climate fluctuation data to determine the degree of impact, thereby obtaining the array risk classification index.

[0019] The characteristics of synchronous thermoelectric anomalies include abnormal response ratio type, associated voltage change characteristics, and key time labels. The regional anomaly cluster distribution characteristics include cluster distribution type, spatial density, and number group identification results. The dynamic characteristics of current offset include continuous offset category, change direction attribute, and number mapping information. The voltage cumulative offset characteristics include cumulative amplitude attribute, classification results, and node response grouping. The array risk grading indicators include risk level category, risk interval identification, and associated influencing factors.

[0020] A ratio sequence refers to a chronological sequence of the ratios between the current drop amplitude and the temperature rise rate. A ratio change key refers to a key period within the ratio sequence where a significant change or abrupt change occurs. A module number refers to the unique identification code for each PV module in the system. An anomaly segment refers to a time period during which an anomaly in the operating data is detected. The degree of clustering refers to the spatial or numerical distribution intensity of multiple anomaly segments. A clustering phenomenon key refers to a key numbered interval within the distribution of anomaly segments where anomalies are highly concentrated. A current output sequence refers to a sequence of actual minute-by-minute output current data recorded by a PV module over multiple consecutive days. A current deviation trajectory refers to a record of the deviation change process formed by comparing the current current with historical current data for the same period. A direction change refers to a turning point in the deviation trajectory from increasing to decreasing or from decreasing to increasing in the time series. Nodes refer to monitoring points in the PV system for data collection or electrical connections. Cumulative change characteristics refer to the characteristics of voltage data that exhibit continuous cumulative deviations over a certain period. A voltage segment refers to a time segment of voltage monitoring data that has been identified as exhibiting a specific deviation trend. A classification and archiving method refers to the method of categorizing and archiving voltage anomaly segments based on segment characteristics. The degree of impact refers to the size or severity of the impact of abnormal array distribution on the overall performance and risk level of photovoltaic modules.

[0021] See also Figure 2 , the electrical thermal response interpretation module includes: The ratio sequence calculation submodule is based on the cloud computing platform. It analyzes the current drop amplitude and temperature rise rate data uploaded by the photovoltaic modules, calculates the ratio of the current drop amplitude and temperature rise rate at each time point, compares the ratio changes at consecutive time points, and obtains the ratio sequence trend. Based on the cloud computing platform, the current drop amplitude and temperature rise rate data uploaded by the photovoltaic components are processed. First, the minute-by-minute current and temperature monitoring data of the specified photovoltaic components during operation are extracted from the database. Each set of data is traversed in chronological order. The current change amplitude at each two consecutive sampling moments is first determined. The drop amplitude at that time point is obtained by directly subtracting the current values ​​at adjacent time points. The current change amplitude statistics of the entire sequence are completed in sequence. Then, the temperature data in the same time period are used to obtain the temperature rise rate sequence. The current change amplitude at each time point is paired with the temperature rise rate at the same time point, and the ratio of the two is calculated. The ratio data of all time points are gradually filled to form a complete ratio sequence. In actual application, the current change amplitude of a photovoltaic component from 8:00 to 8 :04's current monitoring data are 5.0, 4.8, 4.6, 4.3, and 4.0, and the temperature monitoring data are 30, 31, 33, 35, and 37. First, the current drop amplitudes per minute are calculated to be 0.2, 0.2, 0.3, and 0.3, and the temperature rise rates are 1, 2, 2, and 2, respectively. Then, the ratio of each group of current changes to the temperature rise rates is calculated to obtain a ratio sequence of 0.2, 0.1, 0.15, and 0.15. The ratio data are further compared according to the time points, and the ratios of two adjacent time points are compared to determine whether the ratio change exceeds a certain value. If it is greater than 0.05, it is marked as a significant change. All time points where changes occur are marked separately and organized into a trend sequence of ratio changes. The ratio sequence trend is obtained as the basis for subsequent analysis.

[0022] The key time period screening submodule determines the changing trend of the ratio sequence in each time period based on the ratio sequence trend and the ratio sequence of the same period of the previous day, and screens the time period where the ratio undergoes key changes to obtain the key interval of ratio change; Compare the ratio series of the same time period of the current day and the previous day one by one. First, call the data of the corresponding time period of the same component on the previous day from the historical data of the cloud platform, extract the ratio series of the current day and the previous day, and compare them one by one by time points. For the ratio of each time point, directly calculate the difference between the two. For the time point with a difference exceeding 0.05, it is determined that a key change has occurred. All time points determined to be key changes and their surrounding time periods are classified as key intervals of ratio changes. Assuming that in actual operation, the current sequence is 0.2, 0.1, 0.15, 0.15. The previous day's sequence was 0.18, 0.1, 0.12, 0.13. Subtracting the two sequences, the results are 0.02, 0, 0.03, 0.02. No point exceeds 0.05. If the current third point becomes 0.25, the difference will be 0.02, 0, 0.13, 0.02. The third point is classified as a key change. The third time point and the moments before and after it are classified as the key interval. All periods where the ratio changes significantly are summarized and output in chronological order to form the key interval of ratio change.

[0023] The thermoelectric anomaly feature extraction submodule analyzes the voltage changes in the corresponding time period based on the ratio change key interval, determines the synchronization characteristics of the ratio change and voltage change, performs time alignment, and obtains the synchronous thermoelectric anomaly features; Based on the obtained ratio change key interval, the voltage data needs to be further processed. For the time period corresponding to the ratio change key interval, the voltage monitoring data of the same component in this period is extracted, and the voltage data is matched at each time point in the key interval in turn. The voltage change value is retrieved at the time point, and the trend of voltage change and ratio change is compared. All time points with a voltage change amplitude greater than 0.5 and occurring at the same time as the ratio change amplitude are identified as synchronous changes. For example, the voltage data in the key interval are 38.0, 37.2, 36.9, and 36.5, and the ratio data are 0.2, 0.25, 0.15, and 0.15. The change amplitude of the voltage and ratio at each point is judged. The first point has a voltage change of 0.8 and a ratio change of 0.05, which is a synchronous change. The second point has a voltage change of 0.3 and a ratio change of 0.1, which is not synchronous. After screening, all the points with synchronous changes are recorded. Combined with the time point, ratio change, and voltage change, all sections with synchronous anomalies are marked. The ratio, associated voltage data, and time points in the section are summarized as synchronous thermoelectric anomaly characteristics.

[0024] See also Figure 3 , the abnormal aggregation distribution module includes: The number-space matching submodule analyzes the correspondence between the PV module number and spatial position based on the synchronous thermoelectric anomaly characteristics, identifies the number of each PV module, and matches the data of the number and spatial coordinates to obtain the number-space correspondence; Retrieve the unique number of the photovoltaic module associated with each thermoelectric anomaly record, retrieve the module number corresponding to each abnormal feature in the data system one by one, associate all abnormal features with their numbers, and then find the spatial coordinate information corresponding to the number in the photovoltaic power station area spatial layout database. The system locates the actual geographical coordinates of the component in the area layout table according to the number, and uses a one-to-one pairing method to form a unique mapping relationship between the number and its specific spatial location data. Then, according to the mapping table, all uploaded abnormal features are sorted and archived according to the number and spatial coordinates. In the following example, if the components numbered A101, A102, A103, and A104 correspond to coordinates (10, 20), (10, 21), (11, 20), and (11, 21), respectively, if A101 and A103 are recorded as synchronous thermoelectric anomalies in the same period, the numbers and coordinates of A101 and A103 are bound to the abnormal events and stored respectively, and sorted by number in the data list to ensure that each abnormal record has a number and spatial position. Further, in subsequent operations, the spatial coordinates can be directly located by the number, completing the data pairing of the number and spatial coordinates, and obtaining the number-space correspondence.

[0025] The abnormal accumulation calculation submodule calculates the total number of abnormal fragments that appear in the spatial range of each number based on the number-space correspondence, analyzes the spatial distribution of the accumulated number of abnormal fragments, compares the cumulative differences under each number, and determines the distribution trend of abnormal fragments with the same number in the same area to obtain the cumulative total number of abnormal fragments; Traverse all the identified photovoltaic module numbers in the spatial area, count the number of abnormal fragments for each number in the abnormal feature archive library, set cumulative counting parameters for each number, and the system retrieves the appearance time of the abnormal feature corresponding to each number one by one. The number of abnormal fragments that appear in different time periods of the same number is accumulated to form the abnormal cumulative value of each number. In the abnormal cumulative statistical process, the spatial unit is divided according to the actual photovoltaic field. For example, every 10 numbers are a spatial unit. All numbers are classified according to the spatial unit, and the cumulative values ​​of the numbers in each spatial unit are arranged. The numbers with cumulative values ​​that differ by more than 3 times in the same spatial unit are marked and judged as divided. The area where the distribution trend shifts. For example, the spatial units are numbered A101-A110. If the cumulative abnormal fragments of A101 are 5 times, A102 are 2 times, A103 are 6 times, and A104 are 1 time, the system will mark the numbers with a cumulative number of 5 and 6 times as high frequency, and the numbers with 2 and 1 times as low frequency. The high-frequency numbers will be further checked to see if they are continuously distributed in the spatial coordinates. If A101 and A103 are in the same row, the spatial coordinates corresponding to the row will be recorded as the abnormal concentration area. Finally, the total number of cumulative abnormal fragments in all spatial units will be counted, and the number intervals with a cumulative number higher than 1.5 times the average value will be marked as abnormal clusters, and the cumulative total number of abnormal fragments will be output.

[0026] The aggregation interval screening submodule screens the cumulative number of each numbered interval based on the total number of abnormal fragments, compares the spatial distribution data of the numbered intervals, and calculates the continuous distribution status of the abnormal fragments in the numbered intervals using the formula: ; Calculate the abnormal aggregation degree of the number interval The abnormal concentration degree of the numbered interval is an indicator used to evaluate whether the abnormal phenomena of photovoltaic modules in a certain interval are dense and concentrated. It reflects the spatial characteristics of local risk aggregation and generates regional abnormal aggregation distribution characteristics. Indicates the Number in the range The cumulative number of abnormal fragments of PV panels, Representative The number of PV modules in the numbering interval, Indicates the The spatial position distribution differences corresponding to the number intervals, Indicates the The maximum number of consecutive occurrences of abnormal fragments within the number interval, Indicates the The number of components with abnormal fragments in the number interval, Indicates the The total number of PV modules in the numbering interval; Filter the cumulative number in each numbering interval, group all component numbers according to the interval, each group contains 10 numbers, for example, the first group is numbered from 001 to 010, and the second group is from 011 to 020, arranged in sequence, and analyze the cumulative number of abnormal fragments in each interval. First, calculate the sum of the cumulative number of abnormal fragments of all components under each interval number, then calculate the total number of components in the numbering interval, and then analyze the spatial coordinate data of all components in the interval. Use the spatial coordinate mean and variance to measure the distribution, and select the spatial variance as a reference for spatial distribution differences. Further, in each numbering interval, find out one by one whether the number of abnormal fragments of components with consecutive numbers is greater than zero. The numbers that are continuously greater than zero are regarded as continuous abnormal fragments. Count the maximum number of consecutive abnormal fragments in the interval, and at the same time count the number of numbers with the cumulative number of all abnormal fragments greater than zero in the interval. Substitute the above data into the formula. Taking the numbering interval 001-010 as an example, assuming that the cumulative number of abnormal fragments under the numbers is 4, 2, 5, 3, 1, 3, 2, 0, 1, 2, the cumulative total number of abnormal fragments in the interval is: ; The total number of components in the number range , spatial variance (e.g., by spatial coordinate statistics), the maximum number of consecutive occurrences of abnormal fragments (For example, numbers 003, 004, and 005 have consecutive abnormal segments), the number of abnormal components (That is, the number of numbers whose cumulative number is greater than zero is 8), substitute into the formula: ; If you get If the value is greater than 1.5, it means that the abnormal fragments in the numbered interval have a high degree of aggregation, which can be used as the key area for subsequent spatial optimization and screening of abnormal aggregation areas. The regional abnormal aggregation distribution characteristics are calculated, and the result is 1.946, indicating that there is a certain abnormal fragment aggregation trend in the numbered interval.

[0027] See also Figure 4 , the rhythm deviation trajectory module includes: The current sequence acquisition submodule analyzes the minute-by-minute current data of PV modules collected over multiple consecutive days based on the regional abnormal cluster distribution characteristics. Combined with the module number index, the daily current data is sorted in chronological order. By performing daily classification and archiving operations, the data integrity and continuity are compared to obtain the current sequence set. Retrieve all photovoltaic module numbers under the corresponding spatial distribution area, screen each module number one by one, extract the current measurement value of each minute of the number in the last three consecutive days from the historical operation data records, and the system binds the extracted current data to the number by day. Assuming that the module numbers are B201, B202, and B203, the current monitoring data from 8:00 to 8:05 every day in three days are B201: 4.6, 4.7, 4.8, 4.8, 4.7, 4.6, B202: 5.1, 5.2, 5.3, 5.3, 5.2, 5.1, B203: 4.2, 4.2, 4.3, 4.3, 4.2, 4.1, and traverse the current values ​​at each sampling moment in ascending time from the original database. All current sampling points are arranged in chronological order, and the system automatically archives them day by day, and stores the daily data separately by classification. Then, the integrity check of the daily current data within three days is performed by comparing whether the number of daily sampling points is consistent with the number of data points that should be collected within the theoretical period. If the missing data points on a certain day are greater than 1%, the data of that day are marked as incomplete, otherwise they are recorded as complete. All complete and continuous daily series are classified and stored in the current series set. During the data processing process, the continuity and integrity of the current data with the same number are judged by comparing the daily data series of each number within three consecutive days. If B201 has fewer than two sampling points on a certain day, the day is archived as an exception. If the data of B202 and B203 are collected completely every day, they are all included in the series set, and a current series set with complete data and indexed by number and archived daily is obtained.

[0028] The deviation segment identification submodule calls the current sequence set, calculates the absolute difference between the current per minute on the current operating day and the current per minute on the same period in history, and selects the minute segments with continuous daily deviations using the formula: ; Determine continuous deviation and direction change, and calculate the characteristic value of the deviation section , the current deviation segment is obtained, which represents the overall deviation of the current change of the PV module during the operation day relative to the historical performance, where No. Current output data of the current operating day in minutes, It is Current output data for the same period of the minute history, is the total number of minutes in the current running day, It is The outlier current characteristic parameters of the abnormal aggregation area, is the number of abnormal gathering points in the area to which the current component belongs, It is Fluctuation amplitude measurement parameters of minute historical current data over the same period (such as variation range or degree of dispersion); Call the current sequence set and calculate the absolute difference between the current per minute on the current operation day and the current per minute in the same period in history. For example, the currents collected by the A1 component from 9:00 to 9:02 on May 12, 2025 were 3.0, 3.2, and 3.1 amperes respectively, while the currents in the same period in history were 3.1, 3.3, and 3.0 amperes. The corresponding absolute difference is ampere, ampere, Ampere. When screening the minute segments with daily continuous deviations, if the absolute difference is not less than 0.05 ampere for three consecutive minutes, then the period will be included in the deviation segment. Count the abnormal gathering points in the area where the A1 component is located. Assuming that there are two abnormal gathering points, their outlier current characteristic parameters are 2.0 ampere and 1.5 ampere respectively. The sum of these two items is calculated to be 3.5 ampere, and the square root of it is obtained Ampere, a total of 1440 minutes of data are collected on the current day. Assuming that the standard deviation of each minute is 0.04 ampere, the total current fluctuation amplitude in the same period is 57.6 amperes. The denominator is , substitute the above parameters into the formula: ; The obtained segment GQ=0.00145, if the judgment reference interval is set to 0.002-0.003, then the segment is judged as not significantly deviated, and the deviation characteristics are judged by the abnormal aggregation points and historical fluctuations, and the output current deviation segment is determined.

[0029] The current offset feature output submodule analyzes the segment range of the current deviation segment, optimizes the component number index, compares the segment features with the actual component position, and integrates each type of data to output the current offset dynamic features; Analyze the range of the current deviation segment. If the GQ of component A1 is continuously higher than 0.002 between 9:00 and 10:00 on May 12, 2025, this hour is identified as the deviation segment. The component number index is optimized, and all minute current data from 9:00 to 10:00 is associated with the A1 component number. The segment characteristics are compared with the actual physical location of the A1 component. The GQ, absolute difference, and fluctuation amplitude of each minute in the segment are calculated. The mean and maximum GQ between 9:00 and 10:00 are taken. For example, the mean GQ of the segment is 0.0025 and the maximum GQ is 0.003. The historical standard deviation and abnormal cluster point parameters of the same period in the segment are summarized. Various data are integrated to form segment dynamic characteristic data, including the segment start and end time, associated component number, mean and maximum GQ, and minute current absolute difference sequence. The dynamic characteristics of the current deviation are output.

[0030] See also Figure 5 , the voltage accumulation deviation module includes: The voltage trend calculation submodule analyzes the voltage data continuously collected from each node based on the dynamic characteristics of current offset, calculates the voltage change amplitude of each node at adjacent collection moments, determines the continuous change of voltage in each collection stage, analyzes the cumulative trend of voltage change in each stage, and obtains the voltage change trend; Call the continuous voltage acquisition sequence of each node, and organize the voltage sampling data corresponding to each node into a complete sequence in chronological order. For each node, directly subtract the voltage values ​​of adjacent time points to obtain the voltage change amplitude between each two time points. The voltage change amplitude in all time periods is counted in turn and archived according to the acquisition time. The system collects all voltage change amplitudes of each node in the whole cycle to determine whether the voltage in each time period increases or decreases. By directly comparing the voltage value at the current moment with the previous moment, if the current value is greater than the previous value, it is judged to be an increase, otherwise it is a decrease. In the full cycle data, the positive and negative directions of the voltage change in each time period are recorded in turn, and the continuous change intervals of each stage are sorted and merged. For the same node, if the voltage changes in five or more consecutive time periods are positive, it is classified as an upward trend stage. If they are all negative, it is classified as a downward trend stage. The cumulative change of voltage variation in each stage is summed up and the result is archived together with the duration of the stage, node number, and start and end time. In actual application, for example, the data collected at a node from 8:00 to 8:10 is 37.1, 37.3, 37.6, 37.8, 38.0, 37.7, 37.5, 37.2, 37.0, 36.8, and 36.5. The system calculates the voltage change value of each time period as 0.2, 0.3, 0.2, 0.2, -0.3, -0.2, -0.3, -0.2, -0.2, -0.3. The cumulative value of the continuous positive segment is 0.2+0.3+0.2+0.2=0.9, which lasts for five minutes and is classified as an upward trend. The cumulative value of the subsequent negative segment is -0.3-0.2-0.3-0.2-0.2-0.3=-1.3, which is classified as a downward trend segment. After completing the segmented cumulative statistics of all data, the voltage change trend of each node is obtained.

[0031] The continuous offset screening submodule determines whether the voltage increase and decrease directions are consistent in each acquisition stage during the voltage change trend. It analyzes the coverage interval and direction continuity of the continuous offset stage, compares the offset direction characteristics between different segments, and screens typical segments with continuous voltage offset to obtain continuous offset segments. According to the segmentation trend, the voltage change direction of each node in each continuous acquisition stage is manually marked, and the direction attributes of the changes in each stage are compared segment by segment. If the voltage change is positive at all time points in a certain stage, the direction of the stage is defined as positive, and if all are negative, it is negative. If the direction attribute switches in a stage, it is segmented again with the switching point as the boundary. The duration and coverage interval of each stage are counted, and the section with a stage duration of more than five minutes and no switching of the direction attribute is classified as a continuous offset stage. Further comparison is made between two adjacent stages to analyze the change characteristics of the direction between the stages. For example, if stage one is positive and stage two is negative, it is judged that If the direction switches, if the two stages have the same direction, the direction is considered continuous. For example, if the voltage change direction of node acquisition segment A is positive and lasts for eight minutes, and acquisition segment B is negative and lasts for six minutes, then A and B are two independent continuous offset segments. If the direction of segment C is consistent with that of segment B and is continuous, they are merged into a longer continuous offset segment. After all the acquired data are divided and compared with the direction attributes, the continuous offset segments are screened. Segments with a duration of more than five minutes, no direction switch, and a cumulative change greater than 0.5 are extracted. Finally, they are archived with the segment number, node number, time period start and end points, and direction attribute as labels to obtain the continuous offset segments.

[0032] The segment abnormality judgment submodule analyzes the voltage change interval of the continuous offset segment, calculates the continuity and fluctuation distribution of the voltage change direction in each segment, and determines whether its change performance deviates from the normal state using the formula: ; Get the voltage cumulative offset characteristics , used to quantify the deviation of voltage segments in continuous time periods, where Indicates the The voltage data at a time point is the voltage observation value recorded by the node at that time point. Indicates the The voltage data at each time point is used to compare Constructing time difference comparison, Indicates the The direction consistency discrimination factor of the segment is used to indicate the consistency of the voltage change direction within the time period, and the positive and negative signs are used for direction weighting. Indicates the The width of the interval fluctuation corresponding to the segment voltage change reflects the distribution range of the segment voltage during the change process. Indicates the The time span of the voltage segment reflects the duration of the voltage change in the segment. Indicates the total number of time segments in the segment; Arrange all voltage observation data in the filtered continuous offset segment in chronological order, and calculate the change of each adjacent time point. To ensure the comparability of different physical quantities in the characteristic calculation, and Normalization is performed. Specifically, the minimum and maximum normalization method can be used. For example, assuming that the voltage data collected by node A at 10:00, 10:05, and 10:10 are 230V, 232V, and 235V respectively, the first change is V, the second change is V, and at the same time judge the direction consistency of each voltage change. If both changes are positive, then the corresponding direction consistency judgment factor 、 When dealing with the changing fluctuation distribution, and The fluctuation width of the voltage segment is analyzed. If the range is directly used, then V, V, corresponding to time spans such as 10:00-10:05 and 10:05-10:10, both are 5 minutes, which can be assigned 、 ,at this time 、 、 , substitute the data into the formula for calculation: ; If the same historical fragments are in normal state Mostly distributed in interval, then this fragment , the value is higher than the normal range, and it is determined to be an abnormal offset fragment, The calculation logic is to couple the voltage change and directional consistency factors of each time period with the joint normalization of fluctuation width and time span through weighted accumulation, and integrate the multi-dimensional segment characteristics into a set of numerical values. The result shows that 0.63 reflects that the voltage cumulative offset of this segment is obvious, which is an abnormal performance, and is further used as a voltage cumulative offset feature for global analysis.

[0033] See also Figure 6 , the risk warning and grading module includes: The voltage offset detection submodule analyzes the voltage variation curve of the photovoltaic modules within the same monitoring cycle based on the voltage cumulative offset characteristics, determines the differences in voltage fluctuation trends between modules, selects the module numbers with abnormal variation trends, and generates a voltage trend distribution; The voltage change curves of each PV module within each monitoring cycle are called up, and the voltage data collected from each module within the same monitoring cycle are sorted by number. For each module number, the voltage curve data for the corresponding period is extracted, and the trend of the voltage change curves of all modules within the same cycle is compared in sequence. By calculating the change amplitude of the voltage value of each module within the same time period, the overall rise, fall, or fluctuation characteristics of its voltage curve are judged. The voltage change differences between different modules at the same time or time interval are compared. In the comparison process, the module voltage change amplitude is compared with the average change value of the modules during the same period as the benchmark. If the voltage change amplitude of a module is greater than 1.2 times the average value of the same group, it is judged to have an abnormal change trend. Conversely, if it is less than 0.8 times the average value, it is judged to have a trend below normal. In this example, if the voltage of module A drops by 2.0V in 10 minutes and the average drop within the group is 1.3V, then module A is judged to have an abnormal trend. All numbers judged to be abnormal are archived, and the module numbers with abnormal voltage change trends are screened out to generate a voltage trend distribution.

[0034] The current and thermoelectric synchronization determination submodule compares the current fluctuation characteristics of the voltage trend distribution corresponding to the period with the thermoelectric anomaly signal, determines the synchronization of the current and thermoelectric signals, filters the data segments with synchronization anomalies, and generates a synchronization anomaly distribution; Call the filtered abnormal component numbers, retrieve the current monitoring sequence for each number during the abnormal voltage trend period, and compare the current data with the thermoelectric abnormal signal one by one. For each time point, directly judge whether the current fluctuation amplitude and the thermoelectric abnormal response appear synchronously. By setting a threshold, if the absolute value of the current fluctuation exceeds 0.3A and the thermoelectric abnormal signal appears at the same time, it is determined to be a synchronous abnormality. If the two are not synchronized, the data segment is discarded. In this example, if the voltage trend of component B is abnormal from 8:00 to 8:05, the current data fluctuation amplitude is 0.35A, 0.37A, 0.32A, 0.29A, and 0.4A respectively, and the thermoelectric abnormal signal is activated at the same time, then the 5-minute data segment is marked as a synchronous abnormality. After completing the pairing judgment of all abnormal component numbers and the current and thermoelectric signals of the corresponding time periods, all synchronous abnormal data segments are summarized to generate a synchronous abnormality distribution.

[0035] The array risk grading submodule calculates the anomaly distribution of synchronization anomalies involving components within the fluctuation range of monitoring logs and climate data, analyzes the synchronization distribution range of each component and the characteristics of the anomaly data, optimizes the data correspondence, and obtains the array risk grading index; Based on the component number and abnormal period in the synchronous anomaly distribution, we extract operational anomaly data for the corresponding period from the monitoring log. We then search the meteorological database to extract climate parameters such as ambient temperature, irradiance, and wind speed during the same period. We then compare the degree of overlap between the occurrence time of each component anomaly and the climate fluctuation range. We then mark the intervals where the climate parameter fluctuation exceeds the set threshold and the abnormal data appears simultaneously. For each component, we calculate the distribution range of the synchronous anomaly data segments across all monitoring periods. We also calculate the duration and number of abnormal segments covered by each component. We then set a risk classification threshold based on the correspondence between each component's abnormal coverage rate and climate fluctuation. For example, components with an abnormal coverage rate greater than 10% and a climate fluctuation range greater than 1.5 times the historical average are classified as high risk, those with an abnormal coverage rate between 5% and 10% are classified as medium risk, and those with an abnormal coverage rate below 5% are classified as low risk. In this example, component C has a cumulative synchronous anomaly segment of 30 minutes during the monitoring period, accounting for 12% of the total duration. Furthermore, the wind speed and temperature fluctuations during this segment are higher than the set baseline. Therefore, component C is classified as high risk. All component numbers, risk levels, distribution intervals, and climate impact parameters are archived to obtain the array risk classification index.

[0036] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. Photovoltaic performance evaluation system based on cloud computing, characterized by: The system comprises: The electrical thermal response interpretation module, based on a cloud computing platform, analyzes the current drop and temperature rise rate data uploaded by PV modules, calculates the ratio and compares it with the same period of the previous day. It identifies the trend of the ratio change, selects key time periods, performs time alignment, and obtains synchronous thermal and electrical anomaly characteristics. The abnormal aggregation distribution module analyzes the abnormal fragment distribution of the components in the region based on the synchronous thermoelectric abnormal characteristics, calculates the cumulative number of abnormal fragments, determines the degree of aggregation, screens the key number intervals, optimizes the correspondence between the numbers and spatial coordinates, and generates regional abnormal aggregation distribution characteristics; The rhythm deviation trajectory module analyzes the current output of photovoltaic modules for multiple consecutive days based on the abnormal cluster distribution characteristics of the region, compares the current sequence with the historical current sequence, calculates the current deviation trajectory, determines the continuous deviation section and its direction change, and outputs the dynamic characteristics of current deviation based on the module number mark position; The voltage cumulative deviation module analyzes the voltage data of each node based on the current offset dynamic characteristics, calculates the voltage change trend, screens the voltage segments with continuous offset and cumulative change characteristics, determines whether the cumulative change is abnormal, and obtains the voltage cumulative offset characteristics.

2. The photovoltaic performance evaluation system based on cloud computing according to claim 1, characterized in that: The synchronous thermoelectric anomaly characteristics include abnormal response ratio type, associated voltage change characteristics, and key time labels; the regional abnormal cluster distribution characteristics include cluster distribution type, spatial density, and number group identification results; the current offset dynamic characteristics include continuous offset category, change direction attribute, and number mapping information; the voltage cumulative offset characteristics include cumulative amplitude attribute, classification results, and node response grouping.

3. The photovoltaic performance evaluation system based on cloud computing according to claim 1, characterized in that: The electrical thermal response judgment module includes: The ratio sequence calculation submodule is based on the cloud computing platform. It analyzes the current drop amplitude and temperature rise rate data uploaded by the photovoltaic modules, calculates the ratio of the current drop amplitude and temperature rise rate at each time point, compares the ratio changes at consecutive time points, and obtains the ratio sequence trend. The key time period screening submodule determines the changing trend of the ratio sequence in each time period based on the ratio sequence trend and the ratio sequence of the same time period of the previous day, and screens the time period where the ratio undergoes key changes to obtain the key interval of ratio change; The thermoelectric anomaly feature extraction submodule analyzes the voltage change in the corresponding time period based on the ratio change key interval, determines the synchronization features of the ratio change and the voltage change, performs time alignment, and obtains the synchronous thermoelectric anomaly features.

4. The photovoltaic performance evaluation system based on cloud computing according to claim 1, characterized in that: The abnormal aggregation distribution module includes: The number space matching submodule analyzes the correspondence between the photovoltaic module number and the spatial position based on the synchronous thermoelectric anomaly characteristics, identifies the number of each photovoltaic module, and performs data pairing between the number and the spatial coordinate to obtain the number space correspondence; The abnormal accumulation calculation submodule calculates the total number of abnormal fragments that appear in the spatial range of each number based on the number-space correspondence, analyzes the spatial distribution of the accumulated number of abnormal fragments, compares the cumulative differences under each number, and determines the distribution trend of abnormal fragments with the same area number to obtain the accumulated total number of abnormal fragments; The clustering interval screening submodule screens the cumulative number in each numbered interval according to the cumulative total number of abnormal fragments, compares the spatial distribution data of the numbered intervals, calculates the continuous distribution status of the abnormal fragments in the numbered intervals, calculates the abnormal aggregation degree of the numbered intervals, and generates regional abnormal aggregation distribution characteristics.

5. The photovoltaic performance evaluation system based on cloud computing according to claim 1, characterized in that: The rhythm deviation trajectory module includes: The current sequence acquisition submodule analyzes the minute-by-minute current data of the photovoltaic modules collected over multiple consecutive days based on the abnormal cluster distribution characteristics of the region. Combined with the module number index, the daily current data is sorted in chronological order. By performing daily classification and archiving operations, the data integrity and continuity are compared to obtain a current sequence set. The deviation segment identification submodule calls the current sequence set, calculates the absolute difference between the current per minute on the current operating day and the current per minute on the same period in history, selects minute segments with continuous daily deviations, determines continuous deviations and direction changes, calculates the characteristic value of the deviation segment, and obtains the current deviation segment; The current offset feature output submodule analyzes the segment range of the current deviation segment, optimizes the component number index, compares the segment feature with the actual position of the component, integrates each type of data, and outputs the current offset dynamic feature.

6. The photovoltaic performance evaluation system based on cloud computing according to claim 1, characterized in that: The voltage accumulation deviation module includes: The voltage trend calculation submodule analyzes the voltage data continuously collected from each node based on the dynamic characteristics of the current offset, calculates the voltage change amplitude of each node at adjacent collection moments, determines the continuous change of voltage in each collection stage, analyzes the cumulative trend of voltage change in each stage, and obtains the voltage change trend; The continuous offset screening submodule determines whether the voltage increase and decrease directions in each acquisition stage of the voltage change trend are continuously consistent, analyzes the coverage interval and direction continuity of the continuous offset stage, compares the offset direction characteristics between different segments, and screens typical segments with continuous voltage offset to obtain continuous offset segments; The segment abnormality judgment submodule analyzes the voltage change interval of the continuous offset segment, calculates the continuity and fluctuation distribution of the voltage change direction in each segment, determines whether its change performance deviates from the normal state, and obtains the voltage cumulative offset feature.

7. The photovoltaic performance evaluation system based on cloud computing according to claim 1, characterized in that: The system further comprises: The risk warning and classification module, based on the voltage cumulative offset characteristics, matches the current offset dynamic characteristics and synchronous thermoelectric anomaly characteristics within the same time period, determines the synchronous distribution of abnormal segments in time and component number, analyzes monitoring logs, and combines them with climate fluctuation data to obtain array risk classification indicators; The array risk grading index includes risk level category, risk interval identifier, and associated influencing factors.

8. The photovoltaic performance evaluation system based on cloud computing according to claim 7, characterized in that: The risk warning and grading module includes: The voltage offset detection submodule analyzes the voltage variation curve of the photovoltaic modules within the same monitoring period based on the voltage cumulative offset characteristics, determines the differences in voltage fluctuation trends between modules, selects the module numbers with abnormal variation trends, and generates a voltage trend distribution; The current and thermoelectric synchronization determination submodule compares the current fluctuation characteristics of the voltage trend distribution corresponding to the period with the thermoelectric anomaly signal, determines the synchronization of the current and thermoelectric signals, filters the data segments with synchronization anomalies, and generates a synchronization anomaly distribution; The array risk grading submodule calculates the abnormal distribution of the synchronization anomaly distribution components within the fluctuation range of monitoring logs and climate data, analyzes the synchronization distribution range of each component and the abnormal data characteristics, optimizes the data correspondence, and obtains the array risk grading index.

Images