Photovoltaic module fault diagnosis and location method coupled with data drive and image recognition

Abstract

The present invention relates to a photovoltaic module fault diagnosis and location method that couples data-driven and image recognition. This method first identifies faults in a photovoltaic array using a data-driven method, collects infrared images of photovoltaic modules of the faulty photovoltaic array based on the photovoltaic array fault identification results, and then uses image recognition methods to identify and locate faults in the infrared images of the photovoltaic modules. This method can accurately identify and locate photovoltaic module faults, effectively improve the accuracy of photovoltaic module fault identification, quickly locate faulty photovoltaic modules, and effectively reduce the difficulty of photovoltaic module fault handling and operation and maintenance. Compared with the prior art, the present invention has the advantages of increasing the proportion of fault data in the training data set, low sample correlation used in model training, high fault location accuracy, and sufficient extraction of infrared image features.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic component fault diagnosis, and in particular to a photovoltaic component fault judgment and positioning method coupled with data drive and image recognition. Background Art

[0002] The goal of PV module fault diagnosis is to develop efficient and accurate methods and systems to identify and locate faults in PV modules. Because PV modules are exposed to complex environments for long periods of time and are susceptible to various factors that can lead to performance degradation or failure, fault diagnosis technology is crucial to ensuring the stable operation of PV systems.

[0003] In the field of photovoltaic module fault diagnosis, existing technologies face the following challenges:

[0004] (1) Fault data is easily submerged in non-fault data, affecting the training effect of the fault diagnosis model. In practical applications, photovoltaic modules are in normal operation most of the time, and the frequency of faults is relatively low, resulting in a relative scarcity of fault data. When these limited fault data are surrounded by a large amount of non-fault data, the model training process is easily disturbed, making it difficult to accurately capture fault characteristics, thus affecting the accuracy and reliability of fault diagnosis.

[0005] (2) PV panels are usually installed in fixed locations. This characteristic leads to high correlation in training data, which affects the model training effect. Since the installation location and environmental conditions of PV panels are relatively fixed, their operating data often show a high degree of temporal and spatial correlation. This correlation may cause overfitting during the model training process, that is, the model fits the training data too well, while the generalization ability to new data is reduced. This will not only reduce the accuracy of the fault diagnosis model, but also may cause it to perform poorly in practical applications.

[0006] Chinese patent application publication number CN113139955A discloses a photovoltaic module fault identification method and system based on dual-light images. Although 3,000 infrared images were used to train the ResNet model, the problem of sample class imbalance was not taken into account. In the actual operation of photovoltaic power stations, the frequency of faults is low. If the proportion of fault samples in the training set is too small, it is easy to lead to insufficient feature learning and increase the fault missed detection rate.

[0007] Therefore, there is an urgent need for a photovoltaic module diagnosis method to improve the accuracy and reliability of photovoltaic module fault diagnosis. Summary of the Invention

[0008] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a photovoltaic component fault judgment and positioning method coupled with data drive and image recognition, so as to solve or partially solve the problem that the low proportion of fault data and the high sample correlation degree lead to unsatisfactory model training effect, thereby affecting fault judgment and positioning.

[0009] The purpose of the present invention can be achieved by the following technical solutions:

[0010] One aspect of the present invention provides a photovoltaic module fault diagnosis and location method coupled with data drive and image recognition, comprising the following steps:

[0011] The target PV array's historical power generation and environmental data are obtained and array fault categories are labeled. Based on the obtained fault category labels, the power generation and environmental data are balanced to obtain a sample set with low correlation and a fault-to-non-fault ratio that meets the preset requirements. The first-stage binary classification model and the second-stage multi-classification model are trained separately.

[0012] Acquire infrared image data of the target photovoltaic array and label component fault categories, and train an attention-based multi-classification image recognition and diagnosis model;

[0013] Acquire real-time power generation and environmental data of the target photovoltaic array, and use the first-stage model to determine whether the target photovoltaic array is faulty to achieve fault diagnosis;

[0014] In response to a target photovoltaic array failure, using the second-stage model to obtain a predicted photovoltaic array failure type based on the real-time power generation and environmental data;

[0015] In response to a target photovoltaic array failure, real-time infrared image data of the target photovoltaic array is acquired, and a predicted photovoltaic component failure type is obtained using the image recognition diagnostic model;

[0016] A coupled judgment is performed based on the photovoltaic array fault type and the photovoltaic component fault type to achieve fault location.

[0017] As a preferred technical solution, the balanced processing of power generation and environmental data based on fault category labels to screen out samples with low correlation and a fault-to-non-fault ratio that meets preset requirements includes the following steps:

[0018] Calculate the Pearson correlation coefficient between samples, select and retain samples with correlation coefficients less than the preset value, and adjust the ratio of faulty to non-faulty samples to 2:1;

[0019] Perform data normalization on the screened samples;

[0020] By randomly deleting samples from the screened samples, or randomly selecting samples from the screened samples for supplementation, samples matching the preset number are obtained as the sample set.

[0021] As a preferred technical solution, the sample and The Pearson correlation coefficient is calculated using the following formula:

[0022] ,

[0023] in, For samples and The covariance of and The samples and The variance of .

[0024] As a preferred technical solution, the training process of the first-stage model and the second-stage model includes the following steps:

[0025] Dividing the sample set into a binary classification sample set and a multi-classification sample set;

[0026] Based on the binary classification sample set, the XGBoost model is trained with the goal of maximizing the prediction accuracy of whether the photovoltaic module is faulty to obtain a first-stage model;

[0027] Based on the multi-classification sample set, the XGBoost model is trained with the goal of maximizing the prediction accuracy of array fault category labels to obtain a second-stage model.

[0028] As a preferred technical solution, the training process of the image recognition diagnosis model includes the following steps:

[0029] Adjusting the infrared image data to a preset size and performing a normalization operation to obtain an image data set;

[0030] Based on the image dataset, with the goal of maximizing the prediction accuracy of photovoltaic module fault type labels, a ResNet50-CBAM model based on the self-attention mechanism was trained to obtain an image recognition and diagnosis model.

[0031] As a preferred technical solution, the power generation and environmental data include phase current, phase, radiation intensity and photovoltaic panel temperature.

[0032] As a preferred technical solution, the array fault categories include normal operation, short circuit fault, aging fault, open circuit fault and shadow obstruction.

[0033] As a preferred technical solution, the component failure categories include component hot spots, component cracking, diode failure, module offline failure, vegetation obstruction and no failure.

[0034] Another aspect of the present invention provides an electronic device comprising: one or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the aforementioned photovoltaic component fault judgment and positioning method coupled with data drive and image recognition.

[0035] Another aspect of the present invention provides a computer-readable storage medium comprising one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the aforementioned photovoltaic component fault judgment and positioning method coupled with data drive and image recognition.

[0036] Compared with the prior art, the present invention has at least one of the following beneficial effects:

[0037] (1) Increasing the proportion of fault data in the training data set: The present invention obtains a sample set with low correlation and a fault-to-non-fault ratio that meets the preset requirements by screening, and trains the first-stage model and the second-stage model respectively. By ensuring that the proportion of fault data in the training data set meets the preset ratio, the problem of unsatisfactory model training results caused by the low proportion of fault data is effectively alleviated.

[0038] (2) The samples used in model training have low correlation: The present invention calculates the Pearson coefficient between samples, screens samples with coefficients less than a preset value, and ensures the number of final samples, thereby effectively improving the differences between training samples and alleviating the problem that a large amount of homogeneous data on photovoltaic modules affects the model training effect.

[0039] (3) High fault location accuracy: The present invention first determines whether there is a fault, and then integrates power generation and environmental data and infrared images to determine the type of fault. The model is divided into two stages to perform fault judgment and location respectively. After confirming the existence of a fault, the fault type is predicted. Since there is no interference from non-fault data, the prediction will be more accurate.

[0040] (4) Fully extract infrared image features: The present invention uses the ResNet50-CBAM model based on the self-attention mechanism to construct an image recognition diagnosis model, which can fully extract features in the image to achieve accurate prediction of the fault type.

[0041] (5) The hierarchical structure is clear, providing a cascade fault identification and location process algorithm from photovoltaic array failure to photovoltaic module failure, accurately locating the fault of the photovoltaic module in the faulty photovoltaic array, effectively reducing the fault location time of photovoltaic power generation operation and maintenance, and improving the efficiency of photovoltaic power generation operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Schematic diagram of the flow of a photovoltaic module fault judgment and positioning method coupled with data drive and image recognition in an embodiment;

[0043] Figure 2 A flowchart of a two-stage data-driven modeling dataset screening process in an embodiment;

[0044] Figure 3 A flowchart of photovoltaic array fault diagnosis based on data-driven two-stage modeling in an embodiment;

[0045] Figure 4 Schematic diagram of an electronic device in an embodiment. DETAILED DESCRIPTION

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0047] In the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0048] Example 1

[0049] In order to solve the problems existing in the above-mentioned prior art, this embodiment provides a photovoltaic module fault diagnosis and positioning method coupled with data drive and image recognition, see Figure 1 , the method comprises the following steps:

[0050] In step S1, the historical power generation and environmental data of the target photovoltaic array are obtained and array fault categories are marked. The power generation and environmental data are balanced based on the fault category labels, and a sample set with low correlation and a fault-to-non-fault ratio that meets the preset requirements is screened out. The first-stage binary classification model and the second-stage multi-classification model are trained respectively.

[0051] For details, see Figure 2 and Figure 3 , this step includes the following sub-steps:

[0052] Step S101: Collect historical operation data of the photovoltaic power station and mark the historical data as faults. Including phase 1 current, phase 2 current, phase 1 voltage, phase 2 voltage, irradiation intensity and photovoltaic panel temperature (single sampling) and other characteristics; and the data is marked with fault categories, that is, fault labeling, to obtain the array fault category , including normal operation, short circuit fault, aging fault, open circuit fault and shadow obstruction fault.

[0053] Step S102: Feature selection processing: Based on the fault category labels, the power generation and environmental data are balanced to obtain a sample set with low correlation and a fault-to-non-fault ratio that meets preset requirements.

[0054] First, the data collected in step S101 Select and balance the data according to the fault labels and fault categories. The data balancing process uses the Pearson correlation coefficient, and the formula is: ,in For samples and The covariance of and The samples and The variance of .

[0055] Then, samples with smaller correlation coefficients are selected. Specifically, this embodiment performs balancing processing for each type of fault, deletes samples with Pearson correlation coefficients greater than 0.6, and retains all samples with correlation coefficients less than or equal to 0.6. The proportion of data for each fault type is adjusted to an appropriate ratio, that is, a ratio that conforms to the probability of fault occurrence (specifically, the proportion of each fault in the original data set can be used), and the ratio of fault data to non-fault data is 2:1, to obtain a data set. .

[0056] Step S103: Data preprocessing. , using the formula Perform normalization processing. When the number of existing samples is greater than or equal to the threshold (such as 210867), they are randomly deleted from the filtered samples ( m -210867) samples. If the number of existing samples is less than the threshold, a random sample (210867- m ) samples are supplemented to obtain samples that match the preset number as a sample set, and the sample set is divided into a fault and non-fault binary classification data set and fault multi-classification dataset , where the dataset Only fault samples are included in the dataset. For each dataset, it is divided into a training set for parameter selection and model training, and a test set for model evaluation and output.

[0057] Preferably, when incremental samples are collected, the data set is updated using steps S101 to S103.

[0058] Step S104: train the first stage model for binary classification and the second stage model for multi-classification. Use the XGBoost model to train the fault and non-fault binary classification dataset pre-processed in step S103. Perform the first stage model training to obtain the first stage model . Use the XGBoost model to classify the fault multi-classification dataset preprocessed in step S103 Conduct the second stage model training to obtain the second stage model The first stage model and the second-stage model Together they constitute a data-driven photovoltaic array fault diagnosis model .

[0059] Step S2: Acquire infrared image data of the target photovoltaic array and label component fault categories, and train an attention-based multi-classification image recognition diagnosis model.

[0060] Specifically, this step includes the following sub-steps:

[0061] Step S201: Collecting infrared image data sets of photovoltaic modules , collect infrared image data of photovoltaic modules and mark the images to obtain module fault category labels , including Cell (component hot spot), Cracking (component cracking), Diode (diode failure), Offine-Module (module offline failure), Shadoing (vegetation shading) and No-Anomaly (no failure).

[0062] Step S202: Image preprocessing. The image data set collected in step S201 Crop the original image of size 40*24 to a size of 36*22, and then use the formula Normalize the image to obtain the image dataset .

[0063] Step S203: Image recognition model training. After the last convolutional layer of each basic block of ResNet50, the self-attention mechanism module CBAM is added to form the ResNet50-CBAM model. The ResNet50-CBAM model is used to implement the photovoltaic module fault data set pre-processed in step S202. Image recognition training to obtain the image recognition diagnosis model .

[0064] Step S3: Acquire the real-time power generation and environmental data of the target photovoltaic array, use the first-stage model to determine whether the target photovoltaic array is faulty, and implement fault judgment. When the prediction result is that the target photovoltaic array has failed, the second-stage model is used to obtain the predicted photovoltaic array fault type based on the real-time power generation and environmental data.

[0065] Specifically, this step includes the following sub-steps:

[0066] Step S301: Real-time collection of photovoltaic array power generation and environmental data , and using the formula Normalize the data to get the data .

[0067] Step S302: PV array fault determination. Using the first stage model obtained in step S1 Identify the data obtained by S301 for the photovoltaic fault array Perform fault diagnosis.

[0068] Step S303: If the prediction result is that the photovoltaic array has failed, the second stage model obtained in step S104 is used to The photovoltaic array data determined to be faulty in step S302 Perform fault identification.

[0069] Step S4: When the prediction result is that the target photovoltaic array has a fault, real-time infrared image data of the target photovoltaic array is obtained, and the predicted photovoltaic component fault type is obtained using the image recognition diagnosis model. A coupled judgment is performed based on the photovoltaic array fault type and the photovoltaic component fault type to achieve fault location.

[0070] Specifically, this step includes the following sub-steps:

[0071] Step S401, collecting images of photovoltaic components of the faulty photovoltaic array, collecting images of photovoltaic components of the faulty array determined in step S3, and cropping the infrared image to a picture of size 36*22, and using the formula Perform normalization processing.

[0072] Step S402: Photovoltaic module fault identification and location, using the image recognition model obtained in step S2 The photovoltaic module image pre-processed in step S401 is used to identify and locate photovoltaic module faults, and coupled with the array fault type obtained in step S303 to make a judgment.

[0073] In summary, this method first uses a data-driven approach to identify photovoltaic array faults. Based on the PV array fault identification results, infrared images of the faulty PV modules are collected. Then, image recognition methods are used to identify and locate the faults in these infrared images. This method can accurately identify and locate PV module faults, effectively improving the accuracy of PV module fault identification, rapidly locating the faulty PV module, and effectively reducing the difficulty of fault handling and maintenance for PV modules.

[0074] Example 2

[0075] On the basis of Example 1, Figure 4 As shown, this embodiment provides an electronic device, including: one or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include computer program instructions for executing the photovoltaic component fault judgment and positioning method coupled with data drive and image recognition as described in Example 1.

[0076] In a typical configuration, an electronic device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0077] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0078] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0079] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1A step that specifies a function in one or more boxes.

[0080] Example 3

[0081] This embodiment provides a computer-readable storage medium, including one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the photovoltaic component fault judgment and positioning method coupled with data drive and image recognition as described in Example 1.

[0082] Computer-readable storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0083] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A photovoltaic module fault judgment and positioning method coupled with data drive and image recognition, characterized in that: The steps include: The target PV array's historical power generation and environmental data are obtained and array fault categories are labeled. Based on the obtained fault category labels, the power generation and environmental data are balanced to obtain a sample set with low correlation and a fault-to-non-fault ratio that meets the preset requirements. The first-stage binary classification model and the second-stage multi-classification model are trained separately. Acquire infrared image data of the target photovoltaic array and label component fault categories, and train an attention-based multi-classification image recognition and diagnosis model; Acquire real-time power generation and environmental data of the target photovoltaic array, and use the first-stage model to determine whether the target photovoltaic array is faulty to achieve fault diagnosis; In response to a target photovoltaic array failure, using the second-stage model to obtain a predicted photovoltaic array failure type based on the real-time power generation and environmental data; In response to a target photovoltaic array failure, real-time infrared image data of the target photovoltaic array is acquired, and a predicted photovoltaic component failure type is obtained using the image recognition diagnostic model; Based on the photovoltaic array fault type and the photovoltaic component fault type, a coupled judgment is performed to achieve fault location. The following steps are used to balance power generation and environmental data based on fault category labels and screen samples with low correlation and a fault-to-non-fault ratio that meets preset requirements: Calculate the Pearson correlation coefficient between samples, select and retain samples with correlation coefficients less than the preset value, and adjust the ratio of faulty to non-faulty samples to 2:1; Perform data normalization on the screened samples; By randomly deleting samples from the screened samples, or randomly selecting samples from the screened samples to supplement, we can obtain samples that match the preset number as a sample set. The training process of the image recognition diagnosis model includes the following steps: Adjusting the infrared image data to a preset size and performing a normalization operation to obtain an image data set; Based on the image dataset, with the goal of maximizing the prediction accuracy of photovoltaic module fault type labels, a ResNet50-CBAM model based on the self-attention mechanism was trained to obtain an image recognition and diagnosis model.

2. The photovoltaic module fault judgment and positioning method coupled with data drive and image recognition according to claim 1 is characterized in that: The Pearson correlation coefficient of the sample and is calculated using the following formula: ， in, For samples and The covariance of and The samples and The variance of .

3. The photovoltaic module fault judgment and positioning method coupled with data drive and image recognition according to claim 1 is characterized in that: The training process of the first-stage model and the second-stage model includes the following steps: Dividing the sample set into a binary classification sample set and a multi-classification sample set; Based on the binary classification sample set, the XGBoost model is trained with the goal of maximizing the prediction accuracy of whether the photovoltaic module is faulty to obtain a first-stage model; Based on the multi-classification sample set, the XGBoost model is trained with the goal of maximizing the prediction accuracy of array fault category labels to obtain a second-stage model.

4. The photovoltaic module fault judgment and positioning method coupled with data drive and image recognition according to claim 1, characterized in that: The power generation and environmental data include phase current, phase, irradiation intensity and photovoltaic panel temperature.

5. The photovoltaic module fault judgment and positioning method coupled with data drive and image recognition according to claim 1, characterized in that: The array fault categories include normal operation, short circuit fault, aging fault, open circuit fault and shadowing.

6. The photovoltaic module fault judgment and positioning method coupled with data drive and image recognition according to claim 1, characterized in that: The component failure categories include component hot spots, component cracking, diode failure, module offline failure, vegetation obstruction, and no failure.

7. An electronic device, characterized in that: include: One or more processors and a memory, wherein the memory stores one or more programs, wherein the one or more programs include instructions for executing the photovoltaic component fault judgment and positioning method coupled with data drive and image recognition as described in any one of claims 1-6.

8. A computer-readable storage medium, characterized in that It includes one or more programs for execution by one or more processors of an electronic device, and the one or more programs include instructions for executing the photovoltaic component fault judgment and positioning method coupled with data drive and image recognition as described in any one of claims 1-6.

Images