Complex terrain photovoltaic array fault detection system based on machine learning

Abstract

The invention discloses a complex terrain photovoltaic array fault detection system based on machine learning. The detection system carries out grouping and fault detection on a plurality of sub-arrays in the photovoltaic array through the grouping module and the analysis module. The grouping module divides the sub-arrays with similar environmental conditions into the same group based on the geographic position, the topographic feature and the power generation trend of each sub-array so as to form a plurality of array groups. And the grouping module is used for training a grouping model by utilizing a machine learning algorithm so as to realize dynamic and accurate grouping. And each array group applies the adaptive fault detection parameters to perform real-time monitoring on the array group. Wherein the grouping module comprises an information unit A and an information unit B which are respectively used for storing geographical and installation information of the sub-arrays and real-time power generation data so as to integrate environment and power generation conditions.

Description

Technical Field

[0001] The present invention belongs to the technical field of new energy application, and in particular, relates to a complex terrain photovoltaic array fault detection system based on machine learning. Background Art

[0002] With the popularization of photovoltaic power generation technology, more and more large photovoltaic arrays are installed on complex terrains, such as mountains, hills, and roofs of buildings. However, the differences in geographical location, orientation, slope, and shielding conditions brought about by complex terrain significantly affect the power generation efficiency of photovoltaic panels, posing challenges to traditional fault detection systems. In such scenarios, the output characteristics of photovoltaic panels are no longer uniform. Therefore, if fixed detection standards are adopted, it is likely to lead to misjudgment or missed judgment. In addition, with the expansion of the scale of photovoltaic power generation systems, traditional methods that rely on manual detection or a single standard are difficult to meet the needs of efficient and accurate fault detection. The fault detection system based on machine learning can dynamically adjust the detection standards according to the differences in geographical and power generation characteristics of each sub-array. Grouping and fault detection of photovoltaic arrays in complex terrains through machine learning not only improves the accuracy of fault detection, but also reduces maintenance costs, realizes intelligent and automated management of photovoltaic arrays, and provides technical support for the promotion and application of photovoltaic power generation systems in complex scenarios.

[0003] According to the related public technologies, the technical solution with the publication number CN116388669A proposes a photovoltaic panel foreign body detection method based on the Swin Transformer algorithm, which uses a high-definition camera to identify foreign bodies on the photovoltaic panel, calculates their location and degree of contamination, and then makes corresponding countermeasures. The technical solution with the publication number WO2015070482A1 proposes a low voltage ride-through detection system suitable for large photovoltaic power stations in high-altitude areas. By using mobile detection equipment carried in containers, large photovoltaic power stations can be continuously detected. The technical solution with the publication number US20130285670A1 proposes a photovoltaic fault detection system. By connecting multiple detectors to the output end of the photovoltaic system, the location of the specific module with faults in the photovoltaic array can be determined.

[0004] The above technical solutions all propose various detection technology solutions for photovoltaic systems. However, due to the complexity of the scene of photovoltaic systems in complex terrain, the accuracy and detection efficiency of large-scale faults using a single detection algorithm still need to be improved.

[0005] The foregoing discussion of the background art is intended only to facilitate an understanding of the present invention. This discussion does not acknowledge or admit that any of the material referred to is part of the common general knowledge. Summary of the invention

[0006] The object of the present invention is to provide a complex terrain photovoltaic array fault detection system based on machine learning. The detection system groups and detects faults on multiple sub-arrays in a photovoltaic array through a grouping module and an analysis module. The grouping module divides sub-arrays with similar environmental conditions into the same group based on the geographical location, terrain characteristics and power generation trend of each sub-array to form multiple array groups. The grouping module uses a machine learning algorithm to train a grouping model to achieve dynamic and accurate grouping. Each array group applies adaptive fault detection parameters to monitor the array group in real time. Among them, the grouping module includes information unit A and information unit B, which respectively store the geographical and installation information of the sub-array, as well as real-time power generation data, to integrate environmental and power generation conditions.

[0007] The present invention adopts the following technical solution: a complex terrain photovoltaic array fault detection system based on machine learning, the detection system is configured to connect to an existing photovoltaic power generation system and perform group detection on multiple sub-arrays in the photovoltaic array according to power generation data of multiple sub-arrays to form multiple array groups, and perform fault detection on each array group using a corresponding fault detection model; the detection system includes:

[0008] The grouping module is configured to group multiple sub-arrays in the photovoltaic power generation system to form one or more array groups based on the following two conditions:

[0009] (1) The geographical location and terrain characteristics of each photovoltaic module in each subarray;

[0010] (2) The real-time power generation trend of each subarray in the same period in the past;

[0011] An analysis module is configured to perform fault detection on one or more array groups using a fault detection program to determine whether there is an abnormality in the working state of each array group;

[0012] The grouping module applies a machine learning algorithm to train a grouping model, and applies the grouping model to group the plurality of sub-arrays to establish the plurality of array groups.

[0013] Preferably, the grouping module includes an information unit A; the information unit A includes:

[0014] a geographic data storage component configured to store geographic location data of each photovoltaic module;

[0015] an installation data storage component configured to store installation parameter data of each photovoltaic module;

[0016] The environmental data collection component is configured to collect real-time environmental data.

[0017] Preferably, the grouping module further includes a B information unit; the B information unit includes:

[0018] A power generation data collection component is configured to collect power generation data of each of the sub-arrays; the power generation data at least includes daily power generation, and average output power and average output voltage based on time series;

[0019] The power generation data storage component is configured to store the historical power generation data of each of the sub-arrays.

[0020] Preferably, the grouping module further includes a grouping unit; the grouping unit is configured to train the grouping model based on the latest information data provided by the information unit A and the information unit B.

[0021] Preferably, the grouping module includes disbanding the established array groups and performing grouping again in a subsequent stage.

[0022] Preferably, the grouping module includes using a clustering algorithm to perform the following grouping steps to group the multiple sub-arrays:

[0023] S100: preprocessing the information data of each subarray;

[0024] S200: randomly selecting subarrays for grouping to form an initial array group and initializing cluster centers;

[0025] S300: Calculating the similarity between each photovoltaic sub-array and each cluster center;

[0026] S400: allocating each subarray to the array group to which the nearest cluster center belongs according to the minimum distance principle;

[0027] S500: updating cluster centers according to the photovoltaic sub-arrays in each array group;

[0028] S600: Repeat steps S300 to S500 until the change of the cluster center is less than a preset threshold.

[0029] The beneficial effects achieved by the present invention are:

[0030] 1. The detection system of this technical solution dynamically groups data based on multiple dimensions such as geographic location and power generation trends. The system can adjust the detection standards for complex terrain conditions, reduce misjudgments and missed judgments, and improve the accuracy of fault detection. In complex terrain scenarios, the system can flexibly adapt to geographical and environmental differences and achieve accurate sub-array status analysis.

[0031] 2. The detection system of this technical solution uses machine learning algorithms to automatically update the weights of each feature in the grouping model when the environment or system conditions change. When there are changes in weather or partial occlusion, for example, the system can adaptively adjust the grouping to maintain detection efficiency.

[0032] 3. The detection system of this technical solution uses automation technology to reduce dependence on manual inspections, significantly reduce maintenance costs, and can adapt to fault detection work in large-scale photovoltaic systems under complex terrain applications.

[0033] 4. Each working part in the detection system of this technical solution adopts a modular design, and the system maintenance and upgrade can be achieved by individually optimizing and replacing the working modules therein, reducing subsequent use costs and upgrade costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention can be further understood from the following description in conjunction with the accompanying drawings. The components in the figures are not necessarily drawn to scale, but the emphasis is placed on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0035] Description of the accompanying drawings: 1-photovoltaic power generation system; 11-photovoltaic panel; 12-photovoltaic sub-array; 13-photovoltaic array; 14-inverter; 15-connector; 16-grouping module; 18-analysis module; 21-sub-array main line; 31-left laying surface; 32-right laying surface; 500-computing architecture; 502-bus; 504-processor; 506-main memory; 508-read-only memory; 510-storage device; 512-display; 514-input device; 516-cursor control device; 518-network device;

[0036] Figure 1 is a schematic diagram of the layout of the detection system described in an embodiment of the present invention;

[0037] Figure 2 is a schematic diagram of the architecture of the detection system described in an embodiment of the present invention;

[0038] Figure 3 A schematic diagram of the layout of a photovoltaic array under complex terrain in an embodiment of the present invention;

[0039] Figure 4 This is a comparison chart of detection accuracy after the detection system is used in an embodiment of the present invention;

[0040] Figure 5 FIG. 1 is a schematic diagram of the architecture of a computer system used in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with its embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. For those skilled in the art, after reviewing the following detailed description, other systems, methods and / or features of the present embodiment will become apparent. It is intended that all such additional systems, methods, features and advantages are included in this specification. Included within the scope of the present invention and protected by the appended claims. Additional features of the disclosed embodiments are described in the following detailed description, and these features will be apparent from the following detailed description.

[0042] The same or similar reference numerals in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right" and the like indicating the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or component referred to must have a specific orientation. The invention is constructed and operated in a specific orientation, so the terms describing the positional relationship in the drawings are only used for exemplary description and cannot be understood as a limitation of this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0043] Example 1: Exemplary, as shown in the attached Figure 1 As shown, a complex terrain photovoltaic array fault detection system based on machine learning is exemplarily proposed, the detection system is configured to connect to an existing photovoltaic power generation system 1 and perform group detection on multiple sub-arrays according to power generation data of multiple photovoltaic sub-arrays in the photovoltaic power generation system to form multiple array groups, and perform fault detection on each array group using a corresponding fault detection model; the detection system includes:

[0044] The grouping module 16 is configured to group multiple sub-arrays in the photovoltaic power generation system to form one or more array groups based on the following two conditions:

[0045] (3) The geographical location and terrain characteristics of each photovoltaic module in each subarray;

[0046] (4) Real-time power generation trend of each subarray over the same period in the past;

[0047] The analysis module 18 is configured to use a fault detection program to perform fault detection on one or more array groups to determine whether there is an abnormality in the working state of each array group;

[0048] The grouping module 16 applies a machine learning algorithm to train a grouping model, and applies the grouping model to group the plurality of sub-arrays to establish the plurality of array groups.

[0049] Preferably, the grouping module 16 includes an information unit A; the information unit A includes:

[0050] a geographic data storage component configured to store geographic location data of each photovoltaic module;

[0051] an installation data storage component configured to store installation parameter data of each photovoltaic module;

[0052] The environmental data collection component is configured to collect real-time environmental data.

[0053] Preferably, the grouping module 16 further includes a B information unit; the B information unit includes:

[0054] A power generation data collection component is configured to collect power generation data of each of the sub-arrays; the power generation data at least includes daily power generation, and average output power and average output voltage based on time series;

[0055] The power generation data storage component is configured to store the historical power generation data of each of the sub-arrays.

[0056] Preferably, the grouping module 16 further includes a grouping unit; the grouping unit is configured to train the grouping model based on the latest information data provided by the information unit A and the information unit B.

[0057] Preferably, the grouping module includes disbanding the established array groups and performing grouping again in a subsequent stage.

[0058] Preferably, the grouping module 16 includes using a clustering algorithm to perform the following grouping steps to group the multiple sub-arrays:

[0059] S100: preprocessing the information data of each subarray;

[0060] S200: randomly selecting subarrays for grouping to form an initial array group and initializing cluster centers;

[0061] S300: Calculating the similarity between each photovoltaic sub-array and each cluster center;

[0062] S400: allocating each subarray to the array group to which the nearest cluster center belongs according to the minimum distance principle;

[0063] S500: updating cluster centers according to the photovoltaic sub-arrays in each array group;

[0064] S600: Repeat steps S300 to S500 until the change of the cluster center is less than a preset threshold.

[0065] As attached Figure 2 As shown, the layout diagram of the detection system is exemplarily illustrated; wherein the photovoltaic power generation system 1 is a device that uses sunlight energy to generate electrical energy; the smallest independent unit of the photovoltaic power generation system is a single photovoltaic panel 11, and multiple photovoltaic panels 11 can be combined into a photovoltaic sub-array 12; two or more groups of photovoltaic sub-arrays 12 are combined to form a photovoltaic array 13; multiple photovoltaic arrays 13 can be combined to form an overall photovoltaic power generation system; the photovoltaic power generation system in this technical solution refers not only to photovoltaic power stations installed on the ground, but also to various types of photovoltaic power generation systems installed on complex terrains such as building roofs, water surfaces, mountains, and hillsides, such as building integrated photovoltaic systems (BIPV).

[0066] Further, in an exemplary embodiment, two or more photovoltaic panels 11 are connected in series circuit to form a photovoltaic sub-array 12; in the case of large-scale solar power generation, dozens to hundreds of photovoltaic sub-arrays 12 may be provided; the plurality of photovoltaic sub-arrays 12 are connected to the inverter 14 through the sub-array main line 21 connector 15, so as to transmit the produced power to the inverter 14 through a 6-channel or 12-channel line, and the inverter 14 is used to convert the direct current generated by the photovoltaic panel into, for example, 220V alternating current to provide to the power load;

[0067] In a preferred embodiment, the sub-array main line 21 of each photovoltaic sub-array 12 is connected to the grouping module 16 through a branch circuit, and the grouping module 16 performs the grouping and detection steps in the detection system.

[0068] Further, as attached Figure 3 As shown, the implementation method of dynamic grouping is exemplarily demonstrated in an embodiment of a photovoltaic array with two slopes; this method may be an abstract implementation scenario on a hillside or on a roof, and is only described here as a descriptive example, not as a restrictive implementation means of the present technical solution.

[0069] Optionally, a photovoltaic sub-array can be defined as a plurality of photovoltaic modules in a single row, for example, a photovoltaic sub-array can be formed by 4 photovoltaic modules of 1×4, or it can be in the form of 1×6 or 1×8, etc.; in more photovoltaic sub-array definitions, it can be formed by 2×2, 2×4, 3×3, 3×4, 4×4, etc., to define a photovoltaic sub-array; this definition method can be defined by technicians when applying the present technical solution according to the geographical environment, lighting environment or line conditions, and is not a limitation of the present technical solution.

[0070] Further, as attached Figure 3 In the embodiment, there is a left paving surface 31 and a right paving surface 32; there is an angle between the two paving surfaces, so that the orientations of the photovoltaic modules on the two paving surfaces are also correspondingly different; three sub-arrays are arranged on the two paving surfaces, namely 31a, 31b and 31c on the left paving surface 31; and 32a, 32b and 32c on the right paving surface 32.

[0071] When the sun 34 is in the Figure 3 In the position shown, that is, at the side of the left paving surface 31, the left paving surface 31 can obtain relatively good illumination; although during paving, pre-calculation can be performed to set the orientation angles of each sub-array on the right paving surface 32 as much as possible according to the movement trajectory of the sun 34 to obtain the maximum sunlight condition, however, due to the actual complex terrain or the limitation of the paving ground, the right paving surface 32 can never provide the same light receiving condition as the left paving surface 31;

[0072] In the attached Figure 3 Under the working condition, the three sub-arrays 31a, 31b and 31c of the left paving surface 31 all have relatively similar illumination conditions. At this time, the angles at which the three sub-arrays receive sunlight are slightly different and the three are relatively close, so they can be grouped into the same array group.

[0073] On the other hand, since the left paving surface 31 itself has a larger angle with the sunlight at this moment, the three sub-arrays 32a, 32b and 32c above it also have a larger angle with the sunlight, and the difference in the angle between each sub-array of the three 32a, 32b and 32c and the sunlight is also large, so it is not appropriate to classify these three sub-arrays into the same array group.

[0074] The above description only abstractly describes a simplified application scenario of complex terrain. In actual implementation scenarios, there may be more complex terrain expressions and more complex grouping strategies may be adopted.

[0075] Embodiment 2: This embodiment should be understood to include at least all the features of any of the above embodiments, and further improve upon them;

[0076] Preferably, the information unit A is used to collect and record information of fixed dimensions of the photovoltaic module, such as the geographical location of the photovoltaic module, such as horizontal height, slope, orientation, etc.; and installation parameters of the photovoltaic module, such as tilt angle, array position; such information is basically fixed when the photovoltaic module is installed;

[0077] Preferably, the information unit B is used to collect and record real-time variable information, which belongs to variable dimensions; including real-time environmental data, such as light intensity, wind speed, temperature, etc.; and photovoltaic module power generation data, such as daily power generation, average output power, historical power generation data, etc.; environmental data can affect the power generation data of photovoltaic modules, and at the same time, the production process of photovoltaic modules themselves, the speed of aging over time and the changes caused by aging are also different, so they all belong to variable dimension information;

[0078] Furthermore, in a preferred embodiment, the grouping module performs cluster analysis on the multiple sub-arrays based on the above fixed dimension information and variable dimension information based on a clustering algorithm.

[0079] In a preferred embodiment, the aggregate analysis can be performed using the following steps:

[0080] S100: Preprocessing the information data of each subarray; specifically, the preprocessing includes:

[0081] Normalization processing,In order to ensure that data of different dimensions have the same magnitude, all input data are first normalized so that the value range of each parameter of the subarray falls between [0,1].

[0082] Feature weighting: For each parameter in information unit A and information unit B, the corresponding weight W can be given in advance i and U j , where i is a positive integer; W i is the weight corresponding to the i-th parameter in information unit A; U j is the weight corresponding to the jth parameter in information unit B; the weight value of each parameter can be optimized and set according to terrain conditions or the laying design of the photovoltaic system by machine learning.

[0083] Preferably, the initial model is trained by collecting a large amount of experimental data or geographic feature data (such as location, orientation) and power generation feature data (such as real-time output power, historical power generation) in actual applications; the model is preliminarily trained using these data, the initial weight of each feature pair is set, and then the performance is evaluated based on the current grouping results and fault detection accuracy.

[0084] Preferably, the machine learning algorithm can use grid search or Bayesian optimization to optimize the weight combination through experiments; different weight combinations are substituted into the model to maximize the accuracy of fault detection or minimize the false alarm rate.

[0085] Preferably, the grouping model is updated using an optimal weight combination to ensure automatic adaptation to differences in geographic and power generation characteristics under dynamic terrain conditions, thereby optimizing the weight value of each feature.

[0086] S200: Initialize cluster centers, randomly select K initial cluster centers from the input data to represent the initial array group; the initial array group is represented by G1, G2, ..., G K ;

[0087] Since the information of the initial grouping is uncertain, the algorithm randomly selects K subarrays from the photovoltaic subarrays as the initial clustering centers; this random selection is to provide a preliminary grouping reference point in the data space, and each clustering center contains two parts of information, namely fixed dimension information and variable dimension information.

[0088] S300: Calculate the relationship between each photovoltaic sub-array and each cluster center G1, G2, ..., G K The similarity of ; weighted Euclidean distance is used, combined with fixed dimension information and variable dimension information, and the following calculation formula is used for calculation:

[0089]

[0090] In the above formula, x p,i is the feature of the i-th fixed dimension of the p-th subarray; y p,j is the j-th variable dimension feature of the p-th subarray; G k,i and G k,j They are the features of the i-th fixed dimension and the j-th variable dimension of the k-th cluster center respectively.

[0091] S400: assigning subarrays to clusters, that is, assigning each subarray to the array group to which the nearest cluster center G belongs according to the similarity calculated in S300 and the minimum distance principle, wherein the distance refers to the Euclidean distance.

[0092] S500: Update the cluster center according to the photovoltaic sub-arrays in each array group; the new cluster center is the weighted average of all sub-arrays in the group, that is:

[0093]

[0094] In the above formula, │num k │ is the number of sub-arrays in the kth array group.

[0095] S600: Repeat steps S300 to S500 until the change of the cluster center is less than a preset threshold ε, that is:

[0096] ‖C k ′-C k ‖<ε;

[0097] In the above formula, C k ′ represents the kth cluster center in the current iteration, C k represents the kth cluster center in the previous iteration; ‖C k ′-C k ‖ represents the modulus of the two; the threshold ε is set by relevant technical personnel based on the number of sub-arrays, system computing power or detection accuracy and other requirements.

[0098] Finally, after the grouping results are output, the divided array groups are used as input for the next step of fault detection.

[0099] In a preferred embodiment, after obtaining multiple array groups, the detection criteria of each array group are different according to their respective locations and environmental differences; under complex terrain conditions, the following parameters can be preferably used to achieve more efficient fault detection, and the detection criteria can be adaptively adjusted according to the specific conditions within the group;

[0100] Benchmark power generation parameters within a group: Based on the differences in geographical location, slope, orientation, etc. of each group, specific benchmarks are set for parameters such as power generation, output power, and output voltage. The detection benchmark for each group is set based on its daily power generation level and typical power generation data under similar environmental conditions. During detection, if a subarray within a group deviates from the set benchmark, it may indicate a fault.

[0101] Environmental response parameters: Due to different geographical locations, each group responds differently to light, wind speed, and temperature. Real-time monitoring of each group’s environmental response (such as the immediate impact of changes in light intensity and temperature on power generation) can help identify abnormal responses. For example, if the output of a subarray is low under consistent light intensity, it may be caused by shading or equipment failure.

[0102] Generation trend comparison parameters: Analyze the short-term and long-term generation trends of each sub-array group, and compare the current real-time data with historical data; this can detect cumulative generation declines or abnormal fluctuations (such as a continuous decline in generation in a group).

[0103] Relative balance parameters within the group: In each array group, when the sub-arrays that make up the array group are the same, the relative differences in the output power and output voltage of the array group at different time sequences can be compared; this relative detection is based on the sameness of the sub-arrays within the group to avoid the impact of errors caused by different geographical locations. If the output of a sub-array within the group is significantly different from that of other sub-arrays at other times, it can be marked as abnormal for further detection.

[0104] As attached Figure 4As shown in the figure, in the experimental photovoltaic power generation system, a large photovoltaic power generation system is set up and each photovoltaic sub-array in the photovoltaic system is set at different positions and lighting conditions, and the experiment is carried out for 480 hours. By generating controllable fault phenomena in the photovoltaic power generation system, the performance of the detection system is tested. By comparing the fault detection rate with the old detection system, its accuracy has been significantly improved.

[0105] Embodiment 3: This embodiment should be understood to include at least all the features of any of the above embodiments, and further improve upon them;

[0106] For example, as shown in the attached Figure 5 As shown, the implementation of the computer system 500 used by the analysis module in the detection system is described; the computer system 500 can be applied to the data storage, calculation and result output process of each working module in the recognition and judgment system;

[0107] Illustratively, computer system 500 includes a bus 502 or other communication mechanism for communicating information, one or more processors 504 coupled to bus 502 for processing information; processor 504 may be, for example, one or more general-purpose microprocessors;

[0108] The computer system 500 also includes a main memory 506, such as a random access memory (RAM), a cache, and / or other dynamic storage device, coupled to the bus 502 for storing information and instructions to be executed by the processor 504; the main memory 506 may also be used to store temporary variables or other intermediate information during the execution of instructions to be executed by the processor 504; these instructions, when stored in a storage medium accessible to the processor 504, present the computer system 500 as a special-purpose machine customized to perform the operations specified in the instructions;

[0109] The computer system 500 may also include a read-only memory (ROM) 508 or other static storage device coupled to the bus 502 for storing static information and instructions for the processor 504; a storage device 510 such as a disk, an optical disk, or a USB drive (flash drive) will be coupled to the bus 502 for storing information and instructions;

[0110] And further, coupled to the bus 502 may also include a display 122 for displaying various information, data, media, etc., an input device 514 for allowing a user of the computer system 500 to control, manipulate, and / or interact with the computer system 500;

[0111] A preferred way to interact with the management system may be through a cursor control device 516, such as a computer mouse or similar control / navigation mechanism;

[0112] Furthermore, the computer system 500 may also include a network device 518 coupled to the bus 502; wherein the network device 518 may include, for example, a wired network card, a wireless network card, a switching chip, a router, a switch, and other components;

[0113] In general, the terms "engine", "component", "system", "database", etc., as used herein, may refer to logic embodied in hardware or firmware, or to a collection of software instructions, possibly with entry and exit points, written in a programming language such as Java, C, or C++; software components may be compiled and linked into executable programs, installed in a dynamic link library, or may be written in an interpreted programming language (e.g., BASIC, Perl, or Python); it should be understood that software components may be called from other components or from themselves, and / or may be called in response to detected events or interrupts;

[0114] Software components configured to execute on a computing device may be provided on a computer-readable medium, such as a compact disc, digital video disc, flash drive, diskette, or any other tangible medium, or as a digital download (and may be initially stored in a compressed or installable format that requires installation, decompression, or decryption prior to execution); such software code may be stored in part or in whole on a memory device of the executing computing device for execution by the computing device; software instructions may be embedded in firmware, such as an EPROM; it is also understood that hardware components may be composed of connected logic units (such as gates and flip-flops), and / or may be composed of programmable units (such as programmable gate arrays or processors);

[0115] Computer system 500 includes the techniques described herein that may be implemented using custom hardwired logic, one or more ASICs or FPGAs, firmware, and / or program logic that, in combination with the computer system, renders computer system 500 a special-purpose computing device;

[0116] According to one or more embodiments, the techniques herein may be performed by computer system 500 in response to processor 504 executing one or more sequences of one or more instructions contained in main memory 506; such instructions may be read into main memory 506 from another storage medium, such as storage device 510; execution of the sequences of instructions contained in main memory 506 causes processor 504 to perform the process steps described herein; in alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions;

[0117] As used herein, the term "non-transitory media" and similar terms refer to any media that store data and / or instructions that cause a machine to operate in a specific manner; such non-transitory media may include non-volatile media and / or volatile media; non-volatile media include, for example, optical or magnetic disks, such as storage device 510; volatile media include dynamic memory, such as main memory 506;

[0118] Among them, common forms of non-transitory media include, for example, floppy disks, diskettes, hard disks, solid-state drives, magnetic tapes or any other magnetic data storage medium, CD-ROMs, any other optical data storage medium, any physical medium having a pattern of holes, RAM, PROM and EPROM, FLASH-EPROM, NVRAM, any other memory chip or cartridge, and network versions thereof;

[0119] Non-transient media are distinct from transmission media but may be used in conjunction with transmission media; transmission media participate in the transmission of information between non-transient media; for example, transmission media include coaxial cables, copper wires, and optical fibers, including the wires that make up bus 502; transmission media may also take the form of sound waves or light waves, such as radio waves and infrared data communications.

[0120] Although the present invention has been described above with reference to various embodiments, it should be understood that many changes and modifications may be made without departing from the scope of the present invention. That is, the methods, systems and devices discussed above are examples. Various configurations may appropriately omit, replace or add various processes or components. For example, in alternative configurations, the method may be performed in an order different from the order described, and / or various components may be added, omitted and / or combined. Moreover, the features described with respect to certain configurations may be combined in various other configurations, such as different aspects and elements of the configurations may be combined in a similar manner. In addition, the elements therein may be updated as the technology develops, i.e., many elements are examples and do not limit the scope of the present disclosure or claims.

[0121] Specific details are given in the specification to provide a thorough understanding of the exemplary configurations including implementations. However, the configurations may be practiced without these specific details. For example, well-known circuits, processes, algorithms, structures, and techniques have been shown without unnecessary details to avoid obscuring the configurations. This description provides only example configurations and does not limit the scope, applicability, or configurations of the claims. On the contrary, the foregoing description of the configurations will provide those skilled in the art with an enabling description for implementing the described techniques. Various changes may be made to the functions and arrangements of the elements without departing from the spirit or scope of the present disclosure.

[0122] In summary, it is intended that the above detailed description is considered to be illustrative rather than restrictive, and it should be understood that the above embodiments should be understood to be only used to illustrate the present invention and not to limit the scope of protection of the present invention. After reading the contents of the present invention, the technician can make various changes or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. A complex terrain photovoltaic array fault detection system based on machine learning, characterized in that: The detection system is configured to connect to an existing photovoltaic power generation system and perform group detection on multiple sub-arrays in the photovoltaic array according to power generation data of the multiple sub-arrays to form multiple array groups, and perform fault detection on each array group using a corresponding fault detection model; the detection system includes: The grouping module is configured to group multiple sub-arrays in the photovoltaic power generation system to form one or more array groups based on the following two conditions: (1) The geographical location and terrain characteristics of each photovoltaic module in each subarray; (2) The real-time power generation trend of each subarray in the same period in the past; An analysis module is configured to perform fault detection on one or more array groups using a fault detection program to determine whether there is an abnormality in the working state of each array group; The analysis module applies a machine learning algorithm to train a grouping model, and applies the grouping model to group the plurality of sub-arrays to establish the plurality of array groups.

2. The detection system according to claim 1, characterized in that: The grouping module includes an information unit A; the information unit A includes: a geographic data storage component configured to store geographic location data of each photovoltaic module; an installation data storage component configured to store installation parameter data of each photovoltaic module; The environmental data collection component is configured to collect real-time environmental data.

3. The detection system according to claim 2, characterized in that: The grouping module also includes a second information unit; the second information unit includes: A power generation data collection component is configured to collect power generation data of each of the sub-arrays; the power generation data at least includes daily power generation, and average output power and average output voltage based on time series; The power generation data storage component is configured to store the historical power generation data of each of the sub-arrays.

4. The detection system according to claim 3, characterized in that: The grouping module also includes a grouping unit; the grouping unit is configured to train the grouping model based on the latest information data provided by the information unit A and the information unit B.

5. The detection system according to claim 4, characterized in that: The grouping module includes disbanding the established array groups and performing grouping again in a subsequent stage.

6. The detection system according to claim 5, characterized in that: The grouping module includes using a clustering algorithm to perform the following grouping steps to group the multiple sub-arrays: S100: preprocessing the information data of each subarray; S200: randomly selecting subarrays for grouping to form an initial array group and initializing cluster centers; S300: Calculate the similarity between each photovoltaic sub-array and each cluster center; S400: allocating each subarray to the array group to which the nearest cluster center belongs according to the minimum distance principle; S500: updating cluster centers according to the photovoltaic sub-arrays in each array group; S600: Repeat steps S300 to S500 until the change of the cluster center is less than a preset threshold.

Images