Fault hierarchical diagnosis method and device for photovoltaic inverter

Abstract

The present application relates to the technical field of computer, especially to a fault grading diagnosis method and device of photovoltaic inverter, the present application obtains the operation data of photovoltaive inverter, wherein the operation data includes current data, voltage data and communication state data, maps the operation data to the preset photovoltaic inverter distribution network diagram to obtain the fault analysis network diagram, wherein the fault analysis network diagram includes a plurality of position nodes corresponding to photovoltaic inverters one by one, there is a distance edge between two adjacent position nodes, the distance edge is used to represent the distance between two adjacent photovoltaic inverters, the feature extraction is carried out on the fault analysis network diagram to obtain the overall feature of photovoltaic inverter, the overall feature of photovoltaic inverter is graded diagnosed to obtain the fault grade of photovoltaic inverter, thus, the present application accurately diagnoses the fault grade of photovoltaic inverter.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a fault classification and diagnosis method and apparatus for photovoltaic inverters. Background Technology

[0002] Distributed photovoltaic (PV) systems are widely used in industrial and commercial areas, residential buildings, and rural PV power stations due to their advantages of flexible layout, local power generation, and low energy loss. As the core power conversion device in a distributed PV system, the PV inverter plays a crucial role in converting the direct current (DC) generated by the PV modules into alternating current (AC) and integrating it into the power grid. Its operating status directly determines the power generation efficiency and operational safety of the entire PV system.

[0003] Currently, most photovoltaic inverter fault diagnosis relies solely on threshold judgments based on single data points from real-time equipment operation. This involves detecting basic information such as whether real-time current and voltage exceed limits, and whether communication is interrupted to determine if the equipment is faulty. This diagnostic method has significant drawbacks: First, it ignores the spatial distribution and correlation of inverters in a distributed photovoltaic system. Adjacent and regional inverters are easily affected by environmental and grid fluctuations, making misjudgments based on single-device data prone to occur. Second, it fails to consider the long-term temporal variation patterns of the equipment, making it impossible to distinguish between instantaneous data fluctuations, minor aging faults, and severe organic faults, resulting in extremely low fault classification accuracy. Third, relying solely on real-time static data diagnosis fails to capture the evolution trend of faults, making it difficult to achieve refined fault classification. This leads to maintenance personnel being unable to accurately match maintenance plans, resulting in over-maintenance of minor faults and missed or delayed handling of major faults, seriously affecting the stable and efficient operation of distributed photovoltaic systems.

[0004] Based on this, the present invention proposes a fault classification diagnosis method and device for photovoltaic inverters to solve the problem of how to accurately diagnose the fault levels of photovoltaic inverters. Summary of the Invention

[0005] To address the problem of accurately diagnosing the fault levels of photovoltaic inverters, this invention provides a method and apparatus for fault classification and diagnosis of photovoltaic inverters.

[0006] In a first aspect, embodiments of the present invention provide a fault classification and diagnosis method for a photovoltaic inverter. The method is applied to a controller of a distributed photovoltaic system. The system includes the controller, multiple photovoltaic modules, and photovoltaic inverters corresponding to each photovoltaic module. The photovoltaic inverters are connected to the power grid. Acquire the operating data of the photovoltaic inverter; the operating data includes current data, voltage data, and communication status data; The operating data is mapped onto a preset photovoltaic inverter distribution network diagram to obtain a fault analysis network diagram. The fault analysis network diagram includes multiple location nodes that correspond one-to-one with the photovoltaic inverters. There is a distance edge between two adjacent location nodes, which is used to represent the distance between two adjacent photovoltaic inverters. Feature extraction is performed on the fault analysis network diagram to obtain the overall characteristics of the photovoltaic inverter; The overall characteristics of the photovoltaic inverter are classified and diagnosed to obtain the fault level of the photovoltaic inverter.

[0007] Secondly, embodiments of the present invention provide a fault classification and diagnosis device for a photovoltaic inverter. The device is applied to a controller of a distributed photovoltaic system. The system includes the controller, multiple photovoltaic modules, and photovoltaic inverters corresponding to each photovoltaic module. The photovoltaic inverters are connected to the power grid. The acquisition module is configured to acquire the operating data of the photovoltaic inverter; the operating data includes current data, voltage data, and communication status data. The first data processing module is configured to map the operating data onto a preset photovoltaic inverter distribution network diagram to obtain a fault analysis network diagram. The fault analysis network diagram includes multiple location nodes that correspond one-to-one with the photovoltaic inverters. There is a distance edge between two adjacent location nodes, which is used to represent the distance between two adjacent photovoltaic inverters. The second data processing module is configured to extract features from the fault analysis network diagram to obtain the overall features of the photovoltaic inverter. The third data processing module is configured to perform hierarchical diagnosis of the overall characteristics of the photovoltaic inverter to obtain the fault level of the photovoltaic inverter.

[0008] Thirdly, embodiments of the present invention also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the method described in any embodiment of the present invention.

[0009] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the methods described in any embodiment of the present invention.

[0010] This invention provides a method and apparatus for fault classification and diagnosis of photovoltaic (PV) inverters. First, the system collects real-time operating data from each PV inverter within a distributed PV system, including current, voltage, and communication status data. Then, the collected multi-dimensional operating data is mapped onto a pre-constructed PV inverter distribution network diagram to generate a corresponding fault analysis network diagram. This PV inverter distribution network diagram is built based on the actual installation layout of the power plant equipment and includes several independent location nodes. Each location node corresponds one-to-one with a PV inverter on-site. A distance edge is configured between two adjacent nodes, accurately quantifying the actual physical installation spacing between different PV inverters and restoring the spatial topology distribution characteristics of the PV array. After the network diagram is constructed, deep feature mining and extraction are performed on the fault analysis network diagram to obtain the overall characteristics of the PV inverters. These overall characteristics are used to characterize the overall state of the PV inverters in the distributed PV system. Thus, this invention overcomes the limitations of traditional single-parameter diagnosis, combines the spatial distribution relationship of equipment to carry out fault analysis, effectively avoids the problem of misdiagnosis caused by environmental disturbances and power grid fluctuations, and can accurately diagnose the fault level of photovoltaic inverters. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart of a fault classification and diagnosis method for a photovoltaic inverter according to one embodiment is shown; Figure 2 This is a hardware architecture diagram of an electronic device provided in an embodiment of the present invention; Figure 3 A structural diagram of a fault classification and diagnosis device for a photovoltaic inverter according to one embodiment is shown. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0014] Please refer to Figure 1This invention provides a fault classification and diagnosis method for photovoltaic inverters, including: Step 100: Obtain the operating data of the photovoltaic inverter; the operating data includes current data, voltage data, and communication status data; Step 102: Map the operating data to the preset photovoltaic inverter distribution network diagram to obtain the fault analysis network diagram; wherein, the fault analysis network diagram includes multiple location nodes that correspond one-to-one with the photovoltaic inverters, and there is a distance edge between two adjacent location nodes, which is used to represent the distance between two adjacent photovoltaic inverters. Step 104: Extract features from the fault analysis network diagram to obtain the overall features of the photovoltaic inverter; Step 106: Perform a graded diagnosis of the overall characteristics of the photovoltaic inverter to obtain the fault level of the photovoltaic inverter.

[0015] In this embodiment, the system first collects real-time operating data from each photovoltaic inverter within the distributed photovoltaic system, including current, voltage, and communication status data. Then, the collected multi-dimensional operating data is mapped onto a pre-constructed photovoltaic inverter distribution network diagram to generate a corresponding fault analysis network diagram. This photovoltaic inverter distribution network diagram is built based on the actual installation layout of the power station equipment and includes several independent location nodes. Each location node corresponds one-to-one with a photovoltaic inverter on site, and a corresponding distance edge is configured between two adjacent nodes. This distance edge accurately quantifies the actual physical installation spacing between different photovoltaic inverters, restoring the spatial topology distribution characteristics of the photovoltaic array. After the network diagram is constructed, deep feature mining and extraction are performed on the fault analysis network diagram to obtain the overall characteristics of the photovoltaic inverters. These overall characteristics are used to characterize the overall state of the photovoltaic inverters in the distributed photovoltaic system. Thus, this invention overcomes the limitations of traditional single-parameter diagnosis, combining the spatial distribution relationships of equipment for fault analysis, effectively avoiding diagnostic misjudgments caused by environmental disturbances and grid fluctuations, and accurately diagnosing the fault level of photovoltaic inverters.

[0016] In one embodiment of the present invention, after performing a graded diagnosis of the overall characteristics of the photovoltaic inverter to obtain the fault level of the photovoltaic inverter, the method further includes: Obtain historical operating data of photovoltaic inverters; Based on historical and operational data, determine the operational change data; The operational change data is mapped onto a preset photovoltaic inverter distribution network diagram to obtain a data change network diagram; Feature extraction is performed on the data change network diagram to obtain the change characteristics of the photovoltaic inverter; The variation characteristics of the photovoltaic inverter and the overall characteristics of the photovoltaic inverter are fused to obtain the fused characteristics; Based on the integration characteristics, the fault level of the photovoltaic inverter is updated.

[0017] In this embodiment, historical operating data of the corresponding photovoltaic inverter stored in the photovoltaic system's backend database is first retrieved. Based on this, the collected inverter operating data is compared and differentially analyzed with the retrieved historical operating data to quantify the operating change data that characterizes the fluctuation patterns of the equipment's operating status. Subsequently, the calculated operating change data is mapped onto the topology of a pre-constructed photovoltaic inverter distribution network, generating a data change network diagram that reflects the temporal fluctuations and regional linkage characteristics of the equipment. Further deep feature mining and extraction processing is performed on the data change network diagram to obtain the change characteristics of the photovoltaic inverter. Finally, the overall characteristics representing the equipment's steady-state operating condition and the change characteristics representing the equipment's dynamic evolution trend are deeply fused to generate a multi-dimensional fused feature that combines spatial correlation and temporal change characteristics. Based on the fused feature, the degree of equipment fault is re-quantified and assessed, and the fault level of the photovoltaic inverter is iteratively corrected and accurately updated, effectively eliminating diagnostic misjudgments caused by instantaneous data fluctuations and significantly improving the accuracy of photovoltaic inverter fault classification diagnosis.

[0018] In one embodiment of the present invention, the variation characteristics include the time-series variation characteristics of operating parameters, the linkage variation characteristics of adjacent device status, and the overall regional operation fluctuation characteristics.

[0019] In this embodiment, the historical operating data of the photovoltaic inverter includes historical current data, historical voltage data, and historical communication status data. This data can completely record the power output and data transmission conditions of the inverter in past operating cycles. Based on the comparison between historical and real-time operating data, the calculated operating change data, including current change data, voltage change data, and communication status change data, can quantify the dynamic fluctuation amplitude and state deviation of the equipment's operating parameters. The inverter change characteristics extracted based on this include the time-series change characteristics of the operating parameters of a single device, the state linkage change characteristics of adjacent devices, and the regional overall operating fluctuation characteristics of the photovoltaic array. This allows for a comprehensive characterization of the equipment fault evolution pattern from multiple dimensions, including individual time-series evolution, neighborhood linkage coupling, and regional cluster fluctuations.

[0020] In one embodiment of the present invention, the fusion feature is determined by the following formula: In the formula, As a feature of fusion, For dual-modal adaptive attention weighting factors, As an overall feature, As a characteristic of change, This is the residual correction factor. This is a dual-modal characteristic residual correction term. The redundancy penalty coefficient is... This is a feature redundancy orthogonal penalty term. It is a non-linear activation function. These are preset parameters.

[0021] In this embodiment, the present invention no longer relies on manually setting fixed fusion weights. Instead, it adaptively generates modal attention weights through changes in overall features and changing features. The dual-modal adaptive attention weight factor is used to compensate for subtle fault features lost during the fusion process, retaining information on weak faults and early gradual faults, and preventing missed detections caused by excessive smoothing of feature fusion. When the equipment operates stably and temporal fluctuations are small, the weight automatically approaches 1, trusting static spatial features. When the equipment experiences continuous fluctuations and significant fault evolution, the weight automatically decreases, strengthening dynamic temporal change features to adapt to fault diagnosis needs under different operating conditions. Secondly, an innovative feature redundancy orthogonal penalty term is introduced, which is the cosine similarity between the two modal features, used to measure the information overlap between static and dynamic features. The higher the similarity and the greater the redundancy, the stronger the penalty, automatically eliminating duplicate and invalid information and improving the discriminativeness and specificity of the fused features. The redundancy penalty coefficient ranges from 0.2 to 0.5, and the residual correction coefficient ranges from 0.1 to 0.3.

[0022] In one embodiment of the present invention, feature extraction is performed on the fault analysis network diagram to obtain the overall features of the photovoltaic inverter, including: Feature extraction is performed on the fault analysis network diagram to obtain individual unit operation characteristics and neighborhood spatial correlation characteristics; The overall characteristics are determined based on the characteristics of individual unit operation and the characteristics of neighborhood spatial association; Among them, the individual operation characteristics are used to characterize the independent real-time operation health status of a single photovoltaic inverter; the neighborhood spatial correlation characteristics are used to characterize the operational linkage differences and spatial coupling correlation status between the target inverter and adjacent inverters in the same area.

[0023] In this embodiment, the node data and topological association information in the fault analysis network diagram are first parsed and extracted to obtain two types of features: refined individual operating characteristics and neighborhood spatial association characteristics. Individual operating characteristics primarily characterize the independent real-time operational health status of a single photovoltaic inverter, relying on the device's current current, voltage, and communication status parameters to reflect the inverter's current independent operating conditions and abnormal situations. Neighborhood spatial association characteristics characterize the operational linkage differences and spatial coupling association states between the target inverter and adjacent inverters in the same region. Combined with the spatial topological relationship of the inverter array, the operational linkage patterns and operating condition coupling characteristics between regional devices are explored. Based on this, the extracted individual operating characteristics and neighborhood spatial association characteristics are adaptively fused and integrated to obtain the overall characteristics of the photovoltaic inverter that combine individual device operating characteristics and regional cluster association characteristics.

[0024] In one embodiment of the present invention, the overall features are determined by the following formula: In the formula, As an overall feature, It is a non-linear activation function. For adaptive individual weight coefficients, For individual operation confidence factor, Characterized by single-unit operation. It is the set of topological neighborhood nodes of the photovoltaic inverter. As a feature of neighborhood spatial association, Spatial distance weights Here, j is the spatial topological attention coefficient, and j is the index. These are preset parameters. The actual physical installation distance between inverter i and inverter j. This represents the maximum installation spacing of inverters within the current photovoltaic sub-region.

[0025] In this embodiment, the invention first introduces a single-unit operation confidence factor, which adaptively identifies the reliability of real-time acquired data, automatically weakens distorted and abnormally fluctuating invalid data, and improves the purity of single-unit features. Secondly, it innovatively designs a spatial topological attention coefficient, no longer relying solely on physical distance to determine correlation strength, but also combining device operating state similarity to filter effective neighborhood associations, filtering irrelevant neighborhood interference, and strengthening the true regional cluster coupling characteristics. Simultaneously, it configures a dynamic adaptive single-unit weight coefficient, which balances the proportion of single-unit features and neighborhood features according to the degree of device operating fluctuation, achieving differentiated fusion where abnormal devices emphasize their own fault characteristics and stable devices emphasize regional correlation characteristics. Finally, it introduces nonlinear mapping processing to effectively mine implicit coupling fault information between multiple parameters. This effectively solves the technical shortcomings of traditional feature extraction, such as noise sensitivity, mixed correlation features, and low fault feature identification, improving the accuracy of fault level diagnosis for photovoltaic inverters.

[0026] In one embodiment of the present invention, a graded diagnosis of the overall characteristics of the photovoltaic inverter is performed to obtain the fault level of the photovoltaic inverter, including: The overall characteristics of the photovoltaic inverter are input into a preset fault level identification model to obtain the fault level of the photovoltaic inverter. The fault level identification model is obtained through machine learning model training. The fault levels of photovoltaic inverters include normal operation level, minor abnormality level, general fault level and severe fault level.

[0027] In this embodiment, the overall characteristics of the photovoltaic inverter, which integrate the individual operating status and regional spatial correlation features, are input into a pre-trained fault level identification model. Through the model's high-dimensional feature analysis and intelligent classification capabilities, the corresponding fault level of the photovoltaic inverter is accurately output. The fault level identification model is a machine learning classification model obtained through offline training and optimization using a massive number of labeled photovoltaic equipment samples, possessing excellent adaptability to operating conditions and fault differentiation capabilities. The fault levels of the photovoltaic inverter are divided into four levels according to the severity of the fault, from low to high: normal operation level, minor abnormality level, general fault level, and severe fault level, enabling refined classification and judgment of steady-state operating faults of the inverter.

[0028] like Figure 2 , Figure 3 As shown, this invention provides a fault classification and diagnosis device for photovoltaic inverters. The device can be implemented through software, hardware, or a combination of both. From a hardware perspective, as... Figure 2 The diagram shown is a hardware architecture diagram of an electronic device for a fault classification and diagnosis device of a photovoltaic inverter provided in an embodiment of the present invention. Except for... Figure 2In addition to the processor, memory, network interface, and non-volatile memory shown, the electronic device in the embodiment may also include other hardware, such as a forwarding chip responsible for processing packets. Taking software implementation as an example, such as... Figure 3 As shown, a device in a logical sense is formed by the CPU of the electronic device in which it is located reading the corresponding computer program from the non-volatile memory into the memory for execution.

[0029] like Figure 3 As shown in the figure, this embodiment provides a fault classification and diagnosis device for a photovoltaic inverter, comprising: The acquisition module 300 is configured to acquire the operating data of the photovoltaic inverter; the operating data includes current data, voltage data and communication status data. The first data processing module 302 is configured to map the operating data to a preset photovoltaic inverter distribution network diagram to obtain a fault analysis network diagram; wherein, the fault analysis network diagram includes multiple location nodes that correspond one-to-one with the photovoltaic inverters, and there is a distance edge between two adjacent location nodes, which is used to represent the distance between two adjacent photovoltaic inverters. The second data processing module 304 is configured to extract features from the fault analysis network diagram to obtain the overall features of the photovoltaic inverter. The third data processing module 306 is configured to perform hierarchical diagnosis of the overall characteristics of the photovoltaic inverter to obtain the fault level of the photovoltaic inverter.

[0030] In one embodiment of the present invention, the apparatus further includes a fourth data processing module, which is configured to perform the following operations: Obtain historical operating data of photovoltaic inverters; Based on historical and operational data, determine the operational change data; The operational change data is mapped onto a preset photovoltaic inverter distribution network diagram to obtain a data change network diagram; Feature extraction is performed on the data change network diagram to obtain the change characteristics of the photovoltaic inverter; The variation characteristics of the photovoltaic inverter and the overall characteristics of the photovoltaic inverter are fused to obtain the fused characteristics; Based on the integration characteristics, the fault level of the photovoltaic inverter is updated.

[0031] In one embodiment of the present invention, the variation characteristics include the time-series variation characteristics of operating parameters, the linkage variation characteristics of adjacent device status, and the overall regional operation fluctuation characteristics.

[0032] In one embodiment of the present invention, the fusion feature is determined by the following formula: In the formula, As a feature of fusion, For dual-modal adaptive attention weighting factors, As an overall feature, As a characteristic of change, This is the residual correction factor. This is a dual-modal characteristic residual correction term. The redundancy penalty coefficient is... This is a feature redundancy orthogonal penalty term. It is a non-linear activation function. These are preset parameters.

[0033] In one embodiment of the present invention, the second data processing module 304 is configured to perform the following operations: Feature extraction is performed on the fault analysis network diagram to obtain individual unit operation characteristics and neighborhood spatial correlation characteristics; The overall characteristics are determined based on the characteristics of individual unit operation and the characteristics of neighborhood spatial association; Among them, the individual operation characteristics are used to characterize the independent real-time operation health status of a single photovoltaic inverter; the neighborhood spatial correlation characteristics are used to characterize the operational linkage differences and spatial coupling correlation status between the target inverter and adjacent inverters in the same area.

[0034] In one embodiment of the present invention, the overall features are determined by the following formula: In the formula, As an overall feature, It is a non-linear activation function. For adaptive individual weight coefficients, For individual operation confidence factor, Characterized by single-unit operation. It is the set of topological neighborhood nodes of the photovoltaic inverter. As a feature of neighborhood spatial association, Spatial distance weights Here, j is the spatial topological attention coefficient, and j is the index. These are preset parameters. The actual physical installation distance between inverter i and inverter j. This represents the maximum installation spacing of inverters within the current photovoltaic sub-region.

[0035] In one embodiment of the present invention, the third data processing module 206 is configured to perform the following operations: The overall characteristics of the photovoltaic inverter are input into a preset fault level identification model to obtain the fault level of the photovoltaic inverter. The fault level identification model is obtained through machine learning model training. The fault levels of photovoltaic inverters include normal operation level, minor abnormality level, general fault level and severe fault level.

[0036] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on a fault classification and diagnosis device for a photovoltaic inverter. In other embodiments of the present invention, a fault classification and diagnosis device for a photovoltaic inverter may include more or fewer components than illustrated, or combine some components, or split some components, or arrange the components differently. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0037] The information interaction and execution process between the modules in the above-mentioned device are based on the same concept as the method embodiment of the present invention, and the specific details can be found in the description in the method embodiment of the present invention, and will not be repeated here.

[0038] This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements a fault classification and diagnosis method for a photovoltaic inverter according to any embodiment of this invention.

[0039] This invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform a fault classification and diagnosis method for a photovoltaic inverter according to any embodiment of this invention.

[0040] Specifically, a system or apparatus equipped with a storage medium may be provided, on which software program code implementing the functions of any of the embodiments described above is stored, and the computer (or CPU or Mpu) of the system or apparatus may read and execute the program code stored in the storage medium.

[0041] In this case, the program code read from the storage medium can itself implement the function of any of the above embodiments, and therefore the program code and the storage medium storing the program code constitute part of the present invention.

[0042] Examples of storage media used to provide program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as Cd-ROM, Cd-R, Cd-Rw, DVD-ROM, DVD-Ram, DVD-Rw, DVD+Rw), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer via a communication network.

[0043] Furthermore, it should be clear that not only can the program code read by the computer be executed, but also the operating system or other components operating on the computer can be instructed based on the program code to perform some or all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0044] Furthermore, it is understood that the program code read from the storage medium is written to the memory set in the expansion board inserted into the computer or to the memory set in the expansion module connected to the computer. Then, based on the instructions of the program code, the CPU or other device installed on the expansion board or expansion module executes some and all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0045] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0046] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as ROM, RAM, magnetic disk, or optical disk.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fault classification and diagnosis method for a photovoltaic inverter, characterized in that, The method is applied to the controller of a distributed photovoltaic system. The system includes the controller, multiple photovoltaic modules, and photovoltaic inverters corresponding to each photovoltaic module. The photovoltaic inverters are connected to the power grid and include: Acquire the operating data of the photovoltaic inverter; the operating data includes current data, voltage data, and communication status data; The operating data is mapped onto a preset photovoltaic inverter distribution network diagram to obtain a fault analysis network diagram. The fault analysis network diagram includes multiple location nodes that correspond one-to-one with the photovoltaic inverters. There is a distance edge between two adjacent location nodes, which is used to represent the distance between two adjacent photovoltaic inverters. Feature extraction is performed on the fault analysis network diagram to obtain the overall characteristics of the photovoltaic inverter; The overall characteristics of the photovoltaic inverter are classified and diagnosed to obtain the fault level of the photovoltaic inverter.

2. The method according to claim 1, characterized in that, After performing a graded diagnosis of the overall characteristics of the photovoltaic inverter to obtain the fault level of the photovoltaic inverter, the method further includes: Obtain historical operating data of photovoltaic inverters; Based on historical and operational data, determine the operational change data; The operational change data is mapped onto a preset photovoltaic inverter distribution network diagram to obtain a data change network diagram; Feature extraction is performed on the data change network diagram to obtain the change characteristics of the photovoltaic inverter; The variation characteristics of the photovoltaic inverter and the overall characteristics of the photovoltaic inverter are fused to obtain the fused characteristics; Based on the integration characteristics, the fault level of the photovoltaic inverter is updated.

3. The method according to claim 1, characterized in that, The change characteristics include the time-series change characteristics of operating parameters, the linkage change characteristics of adjacent equipment status, and the overall regional operation fluctuation characteristics.

4. The method according to claim 3, characterized in that, The fusion features are determined by the following formula: In the formula, As a feature of fusion, For dual-modal adaptive attention weighting factors, As an overall feature, As a characteristic of change, This is the residual correction factor. This is a dual-modal characteristic residual correction term. The redundancy penalty coefficient is... This is a feature redundancy orthogonal penalty term. It is a non-linear activation function. These are preset parameters.

5. The method according to claim 3, characterized in that, The feature extraction of the fault analysis network diagram yields the overall characteristics of the photovoltaic inverter, including: Feature extraction is performed on the fault analysis network diagram to obtain individual unit operation characteristics and neighborhood spatial correlation characteristics; The overall characteristics are determined based on the characteristics of individual unit operation and the characteristics of neighborhood spatial association; Among them, the individual operation characteristics are used to characterize the independent real-time operation health status of a single photovoltaic inverter; the neighborhood spatial correlation characteristics are used to characterize the operational linkage differences and spatial coupling correlation status between the target inverter and adjacent inverters in the same area.

6. The method according to claim 1, characterized in that, The overall characteristics are determined by the following formula: In the formula, As an overall feature, It is a non-linear activation function. For adaptive individual weight coefficients, For individual operation confidence factor, Characterized by single-unit operation. It is the set of topological neighborhood nodes of the photovoltaic inverter. As a feature of neighborhood spatial association, Spatial distance weights Here, j is the spatial topological attention coefficient, and j is the index. These are preset parameters. The actual physical installation distance between inverter i and inverter j. This represents the maximum installation spacing of inverters within the current photovoltaic sub-region.

7. The method according to claim 6, characterized in that, The method of classifying and diagnosing the overall characteristics of the photovoltaic inverter to obtain the fault level of the photovoltaic inverter includes: The overall characteristics of the photovoltaic inverter are input into a preset fault level identification model to obtain the fault level of the photovoltaic inverter. The fault level identification model is obtained through machine learning model training. The fault levels of photovoltaic inverters include normal operation level, minor abnormality level, general fault level and severe fault level.

8. A fault classification and diagnosis device for a photovoltaic inverter, characterized in that, The device is used as a controller for a distributed photovoltaic system. The system includes the controller, multiple photovoltaic modules, and photovoltaic inverters corresponding to each photovoltaic module. The photovoltaic inverters are connected to the power grid and include: The acquisition module is configured to acquire the operating data of the photovoltaic inverter; the operating data includes current data, voltage data, and communication status data. The first data processing module is configured to map the operating data onto a preset photovoltaic inverter distribution network diagram to obtain a fault analysis network diagram. The fault analysis network diagram includes multiple location nodes that correspond one-to-one with the photovoltaic inverters. There is a distance edge between two adjacent location nodes, which is used to represent the distance between two adjacent photovoltaic inverters. The second data processing module is configured to extract features from the fault analysis network diagram to obtain the overall features of the photovoltaic inverter. The third data processing module is configured to perform hierarchical diagnosis of the overall characteristics of the photovoltaic inverter to obtain the fault level of the photovoltaic inverter.

9. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed in a computer, causes the computer to perform the method described in any one of claims 1-7.

Images