A fault diagnosis method, electronic device and storage medium for a photovoltaic string

Abstract

The present invention relates to the technical field of photovoltaic fault diagnosis. The present invention discloses a fault diagnosis method, an electronic device, and a storage medium for a photovoltaic string. The fault diagnosis method for a photovoltaic string includes: acquiring IV data of the photovoltaic string; preprocessing the IV data, converting the preprocessed IV data into an IV curve image and making a data set; training the data set to obtain a photovoltaic fault type recognition model, and combining the photovoltaic fault type recognition model with a photovoltaic fault degree discrimination method to obtain a photovoltaic fault diagnosis model; performing fault diagnosis on the IV curve image according to the photovoltaic fault diagnosis model to obtain a fault diagnosis result of the photovoltaic string. The present invention can realize the identification of fault types and the discrimination of fault degrees, provide a guiding direction for guiding the operation and maintenance of photovoltaic strings, and accurately provide different warning prompts and fault handling suggestions for users according to different fault types and different fault severities.

Description

Technical Field

[0001] Embodiments of the present invention relate to the technical field of photovoltaic fault diagnosis, and in particular, to a fault diagnosis method, an electronic device, and a storage medium for a photovoltaic string. Background Art

[0002] With the development and maturity of the photovoltaic industry, efficiently identifying and handling photovoltaic string faults has become a key strategy to improve power station output and reduce investment risks, and at the same time guides the development of operation and maintenance. After a photovoltaic string fails, it significantly affects the power generation efficiency, and sometimes even brings a series of safety problems. The operating state of a photovoltaic string directly affects the power generation efficiency of a photovoltaic power generation system. By analyzing the current and voltage data (IV data) generated during the power generation process of the photovoltaic string, fault diagnosis of the photovoltaic string can be achieved.

[0003] However, when using IV data for fault diagnosis in existing photovoltaic string fault diagnosis methods, they often only focus on fault classification and ignore the different severity levels of faults. For example, different degrees of shading have different effects on the power generation efficiency of a photovoltaic string. Obviously, the urgency of dealing with these faults, the warning prompts and treatment suggestions for users should be different, but the existing technology lacks a photovoltaic string fault diagnosis method that integrates the discrimination of fault severity, and the urgency of dealing with these faults, the warning prompts and treatment suggestions for users are not accurate and appropriate enough. Summary of the Invention

[0004] The present invention provides a fault diagnosis method, an electronic device, and a storage medium for a photovoltaic string, which solve the problem that the urgency of dealing with faults, the warning prompts and treatment suggestions for users in the existing method are not accurate and appropriate enough, and can realize the identification of fault types and the discrimination of fault severity, and accurately and appropriately provide different warning prompts and fault treatment suggestions for users.

[0005] According to one aspect of the present invention, a fault diagnosis method for a photovoltaic string is provided. The fault diagnosis method for a photovoltaic string includes:

[0006] Obtain the IV data of the photovoltaic string;

[0007] Preprocess the IV data, convert the preprocessed IV data into an IV curve image and make a data set;

[0008] Train the data set to obtain a photovoltaic fault type recognition model, and the photovoltaic fault type recognition model is combined with a photovoltaic fault severity discrimination method to obtain a photovoltaic fault diagnosis model;

[0009] Perform fault diagnosis on the IV curve image according to the photovoltaic fault diagnosis model to obtain the fault diagnosis result of the photovoltaic string.

[0010] Optionally, the obtaining of the IV data of the photovoltaic string includes:

[0011] Obtaining historical fault IV data of the photovoltaic string under different fault types and different fault degrees;

[0012] Obtaining the IV data when the photovoltaic string is in normal use.

[0013] Optionally, the preprocessing of the IV data and the conversion of the preprocessed IV data into an IV curve image and the production of a data set include:

[0014] Performing data cleaning on the IV data, removing incorrect data, and obtaining qualified IV data;

[0015] Drawing the qualified IV data into an IV curve, and then through normalization or size adjustment or not displaying the IV curve coordinate axes, making it into an IV curve image;

[0016] Classifying the IV curve images according to different fault types and fault degrees and producing a data set.

[0017] Optionally, the photovoltaic fault degree discrimination method includes:

[0018] Calculating the distance value between the actual maximum power point and the virtual maximum power point on the IV curve in the IV curve image, and using the distance value as the discrimination threshold for the photovoltaic fault degree;

[0019] Discriminating the photovoltaic fault degree according to the discrimination threshold of the photovoltaic fault degree.

[0020] Optionally, the calculation formula of the distance value is as follows:

[0021]

[0022] Wherein, is the distance value between the actual maximum power point and the virtual maximum power point in the IV curve image, V mpp_px is the horizontal axis coordinate of the actual maximum power point in the IV curve image, I mpp_px is the vertical axis coordinate of the actual maximum power point in the IV curve image, V T_px is the horizontal axis coordinate of the virtual maximum power point in the IV curve image, I T_px is the virtual maximum power point P T is the vertical axis coordinate of the virtual maximum power point P in the IV curve image.

[0023] Optionally, the calculation formula of the horizontal axis coordinate of the actual maximum power point in the IV curve image is as follows:

[0024]

[0025] Among them, V mpp is the voltage at the actual maximum power point, V min is the minimum value in the voltage array, V max is the maximum value in the voltage array, and H is the height of the IV curve image;

[0026] The calculation formula for the coordinate of the actual maximum power point on the vertical axis of the IV curve image is as follows:

[0027]

[0028] Among them, I mpp is the current at the actual maximum power point, I min is the minimum value in the current array, I max is the maximum value in the current array, and W is the width of the IV curve image;

[0029] The calculation formula for the coordinate of the virtual maximum power point on the horizontal axis of the IV curve image is as follows:

[0030]

[0031] Among them, ;

[0032] The calculation formula for the coordinate of the virtual maximum power point on the vertical axis of the IV curve image is as follows:

[0033]

[0034] Among them, is the short - circuit current.

[0035] Optionally, the training of the dataset to obtain the photovoltaic fault type recognition model includes:

[0036] Training the dataset based on a convolutional neural network model to obtain the photovoltaic fault type recognition model.

[0037] Optionally, the fault types include at least one of occlusion, hot spot, glass fragmentation, open circuit, short circuit, and aging;

[0038] The fault degrees include at least one of minor fault, medium fault, and severe fault.

[0039] According to another aspect of the present invention, an electronic device is provided, and the electronic device includes:

[0040] One or more processors;

[0041] A memory for storing one or more programs;

[0042] When the one or more programs are executed by the one or more processors, the one or more processors implement the fault diagnosis method of the photovoltaic string as described in any embodiment of the present invention.

[0043] According to another aspect of the present invention, there is also provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the fault diagnosis method of the photovoltaic string as described in any embodiment of the present invention.

[0044] The technical solution of the embodiment of the present invention provides a fault diagnosis method for a photovoltaic string that integrates fault degree discrimination. By acquiring the IV data of the photovoltaic string, preprocessing the IV data, converting the preprocessed IV data into an IV curve image and making a data set; training the data set to obtain a photovoltaic fault type recognition model, and combining the photovoltaic fault type recognition model with the photovoltaic fault degree discrimination method to obtain a photovoltaic fault diagnosis model; performing fault diagnosis on the IV curve image according to the photovoltaic fault diagnosis model to obtain the fault diagnosis result of the photovoltaic string, and giving corresponding warning prompts and processing suggestions; it can realize the identification of fault types and the discrimination of fault degrees, providing a guiding direction for guiding the operation and maintenance of photovoltaic strings. According to different fault types and different fault severities, accurate and appropriate warning prompts and fault handling suggestions are given to users. In summary, the present invention solves the problem that when the existing photovoltaic string fault diagnosis method uses IV data for fault diagnosis, it often only focuses on fault classification and ignores the different severities of faults, and the urgency of handling these faults, as well as the warning prompts and processing suggestions given to users, are not accurate and appropriate enough.

[0045] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Brief Description of the Drawings

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0047] Figure 1 is a flowchart of a fault diagnosis method for a photovoltaic string provided according to an embodiment of the present invention;

[0048] Figure 2 is a flowchart of the use of a photovoltaic fault diagnosis model provided according to an embodiment of the present invention;

[0049] Figure 3 It is a flowchart of another fault diagnosis method for a photovoltaic string provided according to an embodiment of the present invention;

[0050] Figure 4 It is a flowchart of yet another fault diagnosis method for a photovoltaic string provided according to an embodiment of the present invention;

[0051] Figure 5 It is a schematic diagram of the IV curve and key parameters of a photovoltaic string provided according to an embodiment of the present invention;

[0052] Figure 6 It is a schematic diagram of an IV data scanning process provided according to an embodiment of the present invention;

[0053] Figure 7 It is a schematic diagram of the fault degree distinguished by the D value provided according to an embodiment of the present invention;

[0054] Figure 8 It is a schematic diagram of an image after processing some fault types provided according to an embodiment of the present invention;

[0055] Figure 9 It is an overall flowchart of obtaining and constructing a photovoltaic fault diagnosis model from IV data provided according to an embodiment of the present invention;

[0056] Figure 10 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present invention. Detailed implementation manners

[0057] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0058] It should be noted that the terms "first", "second", etc. in the description, claims and above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily limit to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0059] Figure 1 is a flowchart of a fault diagnosis method for a photovoltaic string according to an embodiment of the present invention. Refer to Figure 1 According to an embodiment of the present invention, a fault diagnosis method for a photovoltaic string is provided. This method can be executed by a fault diagnosis device of the photovoltaic string. The diagnosis device can be integrated into an electronic device and can be implemented by software and / or hardware. The fault diagnosis method for the photovoltaic string includes:

[0060] S110. Obtain the IV data of the photovoltaic string.

[0061] Specifically, the IV data of the photovoltaic string can be obtained through the power station level and the inverter level. The power station level means processing all the photovoltaic strings connected to the inverters under the power station, and the inverter level means processing the photovoltaic strings under a single or optionally multiple inverters. The logic for obtaining the IV data of the photovoltaic string is as follows: After the gateway receives the IV data scanning task request, it sends an IV data scanning instruction to the inverter to notify it to start scanning. After waiting for the inverter to complete the scanning, the IV data scanning address of the inverter is queried through the relevant protocol. According to the obtained IV data scanning address, the actual IV data is obtained from each inverter address. The obtained IV data is saved and sent to the web page or the front end of the local IV diagnosis program for display.

[0062] S120. Preprocess the IV data, convert the preprocessed IV data into an IV curve image and make a data set.

[0063] Specifically, the obtained IV data is preprocessed, classified according to different fault types and degrees, and converted into image form to make a dataset. First, the original IV data is preprocessed, including data cleaning and removing incorrect data to obtain qualified IV data. Then, the qualified IV data is plotted into an IV curve, and after normalization, size adjustment, and the process of not displaying the IV curve coordinate axes, an IV curve image is made. Finally, the above IV curve images are classified according to different fault types and degrees to form a dataset, which is used for the training of the photovoltaic fault type recognition model.

[0064] The dataset for training the photovoltaic fault type recognition model is the image after IV curve processing, and the processing methods include but are not limited to: normalization, fixing the image size, and the process of not displaying the IV curve coordinate axes. The purpose is to only focus on the shape of the IV curve without considering the size of the numerical value itself, to cope with the problem of different IV curve numerical values under different inverter models and different usage scenarios.

[0065] S130. Train the dataset to obtain a photovoltaic fault type recognition model, and combine the photovoltaic fault type recognition model with the photovoltaic fault degree discrimination method to obtain a photovoltaic fault diagnosis model.

[0066] Specifically, a convolutional neural network model can be used to train the dataset to obtain a photovoltaic fault type recognition model, which can be replaced by other image processing models or the superposition of multiple models to achieve a similar effect. The photovoltaic fault degree discrimination method can use the power analysis method of the photovoltaic string to achieve the discrimination of the fault degree. For example, the fault degree can be divided into three levels: minor, medium, and severe. Train the dataset to obtain a photovoltaic fault type recognition model, and then combine the photovoltaic fault degree discrimination method to obtain the final photovoltaic fault diagnosis model. Figure 2 It is a flowchart of the use of a photovoltaic fault diagnosis model provided by an embodiment of the present invention. Refer to Figure 2 , Figure 2 It shows the whole process of obtaining the final photovoltaic fault diagnosis model by combining the recognition of the photovoltaic fault type and the discrimination of the photovoltaic fault degree.

[0067] S140. Perform fault diagnosis on the IV curve image according to the photovoltaic fault diagnosis model to obtain the fault diagnosis result of the photovoltaic string.

[0068] Specifically, the fault diagnosis result includes the identification of the fault type and the discrimination of the fault degree. The IV curve image is subjected to fault diagnosis according to the pre-trained photovoltaic fault type identification model to first obtain the fault type identification result of the photovoltaic string, and then the fault degree of the IV curve image is discriminated by the photovoltaic fault degree discrimination method, and finally the fault diagnosis result of the photovoltaic string is obtained. Among them, the fault degree of the photovoltaic string can also be discriminated first, and then the fault type of the photovoltaic string can be identified.

[0069] According to the fault diagnosis result of the photovoltaic string, corresponding early warning prompts and processing suggestions are given for different fault types and fault degrees, and early warnings of different emergency levels can be given when users use the photovoltaic string products. For example, a fault with a severe degree level is given a red early warning, prompting the user to process it as soon as possible, and informing that the fault degree seriously affects the power generation efficiency of the photovoltaic string, and then giving fault handling suggestions according to the fault type.

[0070] The technical solution of the embodiment of the present invention provides a fault diagnosis method for a photovoltaic string that integrates the discrimination of the fault degree. By obtaining the IV data of the photovoltaic string, preprocessing the IV data, converting the preprocessed IV data into an IV curve image and making a data set; training the data set to obtain a photovoltaic fault type identification model, and combining the photovoltaic fault degree discrimination method with the photovoltaic fault type identification model to obtain a photovoltaic fault diagnosis model; performing fault diagnosis on the IV curve image according to the photovoltaic fault diagnosis model to obtain the fault diagnosis result of the photovoltaic string, and giving corresponding early warning prompts and processing suggestions; it can realize the identification of the fault type and the discrimination of the fault degree, providing a guidable direction for guiding the operation and maintenance of the photovoltaic string, and accurately and appropriately giving different early warning prompts and fault handling suggestions to users according to different fault types and different fault severities. In summary, the present invention solves the problem that when the existing photovoltaic string fault diagnosis method uses IV data for fault diagnosis, it often only focuses on fault classification and ignores the different severities of faults, and the emergency levels of handling these faults, the early warning prompts and processing suggestions given to users are not accurate and appropriate enough.

[0071] On the basis of the above embodiments, the embodiment of the present invention also refines step S110, which will be specifically described below, but is not a limitation to the present invention.

[0072] Figure 3 is a flowchart of another fault diagnosis method for a photovoltaic string provided by an embodiment of the present invention. Refer to Figure 3 , obtaining the IV data of the photovoltaic string includes:

[0073] S111. Obtain the historical fault IV data of the photovoltaic string under different fault types and different fault degrees.

[0074] S112, obtaining IV data of the photovoltaic string when it is in normal use.

[0075] Specifically, obtaining IV data of PV strings is divided into two stages. The first stage obtains historical fault IV data of PV strings under different fault types and fault degrees, which is used for training the PV fault type recognition model and determining the threshold for distinguishing the degree of PV faults. The second stage obtains IV data of PV strings during normal use, and uses the PV fault diagnosis model to diagnose faults for running equipment.

[0076] On the basis of the above embodiments, the embodiment of the present invention further refines step S120, which is described in detail below, but is not intended to limit the present invention.

[0077] Figure 4 is a flowchart of another photovoltaic string fault diagnosis method provided according to an embodiment of the present invention, referring to Figure 4 , preprocessing the IV data, converting the preprocessed IV data into IV curve images and making data sets include:

[0078] S121. Clean the IV data, remove erroneous data, and obtain IV data that meets the requirements.

[0079] S122, drawing the IV data that meets the conditions into an IV curve, and then making an IV curve image through normalization or size adjustment or non-displaying the IV curve coordinate axis.

[0080] S123. Classify the IV curve images according to different fault types and fault degrees and create a data set.

[0081] Optionally, the fault type includes: at least one of: shading, hot spot, glass breakage, open circuit, short circuit and aging; the fault degree includes: at least one of: minor fault, moderate fault and severe fault.

[0082] Specifically, first, the original IV data is preprocessed, including data cleaning, eliminating erroneous data, and obtaining qualified IV data. Then, the qualified IV data is drawn into an IV curve and then normalized, resized, and the IV curve coordinate axis is not displayed to produce an IV curve image. Finally, the above IV curve images are classified according to different fault types and fault degrees to produce a fault data set, which is used to train the photovoltaic fault type recognition model. Among them, the fault types include shading, hot spots, glass breakage, open circuit, short circuit and aging. The degree of fault is divided into three levels: mild, moderate and severe.

[0083] Optionally, the method for determining the degree of photovoltaic fault includes: calculating the distance value between the actual maximum power point and the virtual maximum power point on the IV curve image, and using the distance value as the discrimination threshold for the degree of photovoltaic fault; and determining the degree of photovoltaic fault according to the discrimination threshold for the degree of photovoltaic fault.

[0084] Optionally, the discrimination threshold for the degree of photovoltaic fault includes at least one of a minor threshold, a medium threshold, and a severe threshold.

[0085] Optionally, the calculation formula for the distance value is as follows:

[0086]

[0087] where, is the distance value between the actual maximum power point and the virtual maximum power point on the IV curve image, V mpp_px is the horizontal axis coordinate of the actual maximum power point on the IV curve image, I mpp_px is the vertical axis coordinate of the actual maximum power point on the IV curve image, V T_px is the horizontal axis coordinate of the virtual maximum power point on the IV curve image, I T_px is the vertical axis coordinate of the virtual maximum power point P T on the IV curve image.

[0088] Optionally, the calculation formula for the horizontal axis coordinate of the actual maximum power point on the IV curve image is as follows:

[0089]

[0090] where, V mpp is the voltage of the actual maximum power point, V min is the minimum value in the voltage array, V max is the maximum value in the voltage array, H is the height of the IV curve image;

[0091] The calculation formula for the vertical axis coordinate of the actual maximum power point on the IV curve image is as follows:

[0092]

[0093] where, I mpp is the current of the actual maximum power point, I min is the minimum value in the current array, I max is the maximum value in the current array, W is the width of the IV curve image;

[0094] The calculation formula for the horizontal axis coordinate of the virtual maximum power point on the IV curve image is as follows:

[0095]

[0096] Among them, ;

[0097] The calculation formula for the coordinate of the virtual maximum power point on the vertical axis of the IV curve image is as follows:

[0098]

[0099] Among them, is the short - circuit current.

[0100] The method for dividing the degree of photovoltaic fault is introduced as follows. Figure 5 is a schematic diagram of the IV curve and key parameters of a photovoltaic string provided according to an embodiment of the present invention. The schematic diagram of the IV curve and key parameters of the photovoltaic string is as Figure 5 shown:

[0101] Figure 5 The open - circuit voltage V oc in is the maximum voltage obtained from the photovoltaic string, which appears in the zero - current state. The short - circuit current I sc is the current passing through the photovoltaic string when the voltage across the battery is zero. The actual maximum power point P mpp is defined as the point on the IV curve where the output power of the photovoltaic string is the largest. The current corresponding to the actual maximum power point is the actual maximum power point current I mpp , and the voltage corresponding to the actual maximum power point is the actual maximum power point voltage V mpp . P T represents the virtual maximum power point, and its value is the product of the open - circuit voltage and the short - circuit current. When the IV curve is normalized and fixed as an image of a unified size, the distance value between the corresponding points of P mpp and P T on the image is used as the threshold for dividing the fault degree. The formula for converting the IV curve parameters to distance coordinates is as follows:

[0102]

[0103]

[0104]

[0105]

[0106] Among them, H is the height of the IV curve image, and W is the width of the IV curve image (the unit can be the size of the picture or pixels, etc., as long as it is a unified fixed size). I min , V min are respectively the minimum values in the current and voltage arrays. I max , V max are respectively the maximum values in the current and voltage arrays. Vmpp_px and I mpp_px are the horizontal and vertical coordinates of point P mpp on the IV curve image respectively, and V T_px and I T_px are the horizontal and vertical coordinates of point P T on the IV curve image respectively. Then the actual maximum power point P mpp and the virtual maximum power P T have a distance value D on the image, and the formula is as follows:

[0107]

[0108] Using the value of D as the discrimination threshold for the degree of photovoltaic fault, through the analysis of the IV data of a large number of photovoltaic strings with different fault types and different fault degrees in the early stage, the mild threshold , the medium threshold and the severe threshold are determined.

[0109] The specific analysis process is as follows: Classify the historical fault IV data collected under different fault types and different fault degrees according to the degree of fault simulation (such as occlusion area, coverage area, broken glass and resistance size connected, etc.), and perform data analysis on the classified IV data after obtaining the D value, including but not limited to analyzing the average value, maximum and minimum values of the D value, and obtaining the three-level discrimination threshold that meets the conditions. For example, the three-level discrimination threshold can be the micro threshold , the medium threshold and the severe threshold . Then use the historical fault IV data under different fault types and different fault degrees for verification, and finally determine the discrimination threshold for the degree of photovoltaic fault after verification.

[0110] In the method for discriminating the degree of photovoltaic fault, the distance value between the virtual maximum power point and the actual maximum power point on the IV curve image is used as the discrimination threshold for the degree of photovoltaic fault. The unit of the distance value can be pixels or inches, etc. Using the obtained mild threshold , the medium threshold and the severe threshold these three-level thresholds to discriminate the degree of fault. Finally, combine the recognition of photovoltaic fault types and the discrimination of the degree of photovoltaic fault to obtain the final photovoltaic fault diagnosis model.

[0111] Optionally, training the dataset to obtain a photovoltaic fault type recognition model includes: training the dataset based on a convolutional neural network model to obtain a photovoltaic fault type recognition model.

[0112] Specifically, the prepared dataset is used to train a Convolutional Neural Network (CNN) model, which includes an input layer, four sets of convolutional layers, a flattening layer, a fully connected layer, and an output layer. After comparison and verification, the best model for fault type recognition is saved, and finally a photovoltaic fault type recognition model is obtained.

[0113] Optionally, the IV curve image is fault diagnosed according to the photovoltaic fault diagnosis model, and the fault diagnosis results of the photovoltaic string include:

[0114] The fault type recognition result of the photovoltaic string is obtained by fault diagnosing the IV curve image according to the photovoltaic fault type recognition model;

[0115] The fault degree of the IV curve image is discriminated according to the photovoltaic fault degree discrimination method, and finally the fault diagnosis result of the photovoltaic string is obtained.

[0116] Specifically, according to the diagnosis result of the photovoltaic fault diagnosis model, corresponding warning prompts and treatment suggestions are given for different fault types and fault degrees.

[0117] The logic of obtaining the IV data of the photovoltaic string in step S110 is as follows:

[0118] First, the instrument accuracy needs to be measured. The accuracy of the current and voltage measuring instruments for the photovoltaic inverter string or array should be ≤ 1.0%. Then, after the gateway receives the IV data scanning task request, it temporarily stops other instruction communications with the inverter, and then sends an IV data scanning instruction to the inverter to notify it to start scanning. Wait for the inverter to complete the scanning. If the inverter scanning fails, the reason for its failure will be queried through relevant protocols. If the scanning is successful, the IV scanning address of the inverter will be queried through relevant protocols. According to the obtained IV scanning address, the actual IV data is obtained from each inverter address. The obtained IV data is sent to the web page or the front end of the local IV diagnosis program for display. Figure 6 It is a schematic diagram of an IV data scanning process provided according to an embodiment of the present invention. The schematic diagram of the IV data scanning process is as Figure 6 shown.

[0119] The IV data of the photovoltaic string can be obtained in the ways of power station level and inverter level. The power station level means processing all the photovoltaic strings connected to the inverters under the power station, and the inverter level means processing the photovoltaic strings under a single or optionally multiple inverters.

[0120] The introduction of obtaining the IV data of the PV string at the power station level is as follows: Communication between the IV diagnostic program and the Ethernet card: Using the network port and optical fiber, send the IV scan instruction to the Ethernet card end through a specific port number. After the Ethernet card receives the IV scan instruction sent from the specific port number, it immediately sends the shutdown instruction to the data collector through the serial port, and at the same time starts to send instructions to the inverter to start the IV scan and collect the IV data of the inverter, and uploads it to the background of the IV diagnostic program in real time. After all the data is sent, immediately send the startup instruction to the data collector through the serial port to resume normal real-time data collection and upload.

[0121] The introduction of obtaining the IV data of the PV string at the inverter level is as follows: The communication between the diagnostic program and the Ethernet card is through clicking the "IV curve" function module in the cloud. The server immediately sends the IV data scan instruction to the local Ethernet card through the specified protocol. After the Ethernet card receives the IV scan instruction sent from the cloud, it immediately interrupts the real-time data communication with the inverter, and at the same time starts to send instructions to the inverter to start the IV scan and collect the IV data of the inverter, and uploads it to the server background in real time. After all the data is sent, immediately resume the real-time data communication with the inverter for normal real-time data collection and upload.

[0122] Step S120 includes the process of preprocessing the data, cleaning the collected IV data of the PV string and removing the incorrect data. The screening rules are as follows: data with negative voltage or current, current greater than 5% of the short-circuit current (irradiance mutation), abnormal curve slope, voltage regression or data with the same voltage, etc. The included fault types are occlusion, hot spot, glass fragmentation, open circuit, short circuit, aging, etc. The fault degree is divided into three levels: minor, medium and severe. Among them, the corresponding relationship between the fault type and the fault degree is shown in Table 1:

[0123] Table 1 Corresponding relationship between fault type and fault degree

[0124]

[0125] When making the dataset, convert the IV time-series data into IV curve images of a fixed size, perform normalization processing, and do not display the IV data coordinate axes. Here, the processing of erasing the coordinate axes and normalizing controls the IV data between 0 and 1. This processing can address the numerical difference problems of different environments and different model strings. Scaling the IV curve image into a fixed-size shape is to find the variation range of the power point in the image as the threshold for judging the degree of fault. According to the photovoltaic fault degree discrimination method, the pixel distance value between the actual maximum power point and the virtual maximum power point can be used to judge the degree of fault. In this way, the discrimination only focuses on the shape change of the IV curve and is not affected by the data size of the power value itself. Such a judgment method is applicable to strings of different environments and different models.

[0126] According to the proposed photovoltaic fault degree discrimination method, use the value of D as the discrimination threshold. Through the analysis of the D values of a large number of faulty IV data in the early stage, the mild threshold , the medium threshold and the severe threshold are determined. Figure 7 is a schematic diagram of the fault degree distinguished by the D value provided by an embodiment of the present invention. The schematic diagrams of the fault degree corresponding to different D values are as shown in Figure 7 (taking the occlusion type as an example).

[0127] When training the photovoltaic fault type recognition model with CNN in step S130, first, read the IV curve images in the dataset and uniformly adjust them to a size of 300x300 pixels. Then normalize the IV curve images (divide the pixel values by 255. The pixel values of an image are usually represented as integers in the range of 0 to 255 because a pixel value represents the color intensity or gray level of a point in the image. In the RGB color model, the value range of each color channel (red, green, blue) is 0 to 255, which represents the intensity or brightness of that channel). The normalized image data is very useful for subsequent image processing and machine learning tasks because it can reduce the dynamic range of the data and avoid precision loss caused by numerical overflow or too small numerical values during the calculation process. In addition, the normalized data is usually easier to compare and operate in the algorithm, especially when tasks such as calculating distances, similarities, or feature extraction are required. After normalization, add the image data to the list of feature set X, and add the label information to the list of y_label, completing the construction of the feature set and the label set. Then store a mapping table y_label_map to record the correspondence between the fault labels and their one-hot encodings. Use a function called counting_data to first count how many images there are in total in the dataset and the number of images with different labels. Then use LabelEncoder to perform integer encoding on the labels, and then use to_categorical to convert them into one-hot encodings so that the model can handle multi-classification problems. Use the train_test_split function to divide the dataset into a training set and a test set, with a ratio of 90% for the training set and 10% for the test set. Then, use the Sequential model of Keras to build the CNN architecture. Specifically, it includes the following:

[0128] (1) Input layer (InputLayer), specifying the size of the input IV curve image as 300x300 pixels, Figure 8 It is a schematic diagram of an image after processing some fault types provided by an embodiment of the present invention. Among them, the IV fault images in some datasets after normalization and without showing the coordinate axes are as Figure 8 shown.

[0129] (2) Convolutional layer (Conv2D), there are a total of 4 groups of convolutional layers, each followed by a max pooling layer (MaxPooling2D). The first convolutional layer uses 32 3x3 convolutional kernels, and the activation function is ReLU. Then there is a 2x2 max pooling layer, which is used to reduce the spatial resolution of the feature map, while reducing the number of parameters and the amount of computation; the second convolutional layer uses 64 3x3 convolutional kernels, the activation function is ReLU, and the max pooling layer is the same as the first group; the third and fourth convolutional layers both use 128 3x3 convolutional kernels, the activation function is ReLU, and the max pooling layer is also the same as the first group.

[0130] (3) A flattening layer (Flatten) that converts the multi-dimensional feature map into a one-dimensional array so that it can be input into the fully connected layer.

[0131] (4) A fully connected layer (Dense) with 512 neuron nodes, and the activation function is ReLU. This layer is usually used to extract high-level features.

[0132] (5) The output layer (Dense) is also a fully connected layer, and the number of its neurons is the same as the number of classes (determined by the size of the y_label_map dictionary), and the activation function is softmax, which is used for multi-classification tasks.

[0133] After defining the above model structure, when training the model, categorical_crossentropy is used as the loss function. This is a commonly used loss function in multi-classification problems and is suitable for the case where the labels are one-hot encoded. It calculates the cross-entropy loss between the predicted probability distribution and the true label distribution. Adam is used as the optimizer. Adam is an optimization algorithm with an adaptive learning rate that combines gradient descent and momentum methods and usually performs well in deep learning. Accuracy is used as the evaluation metric to compile the model. It is a direct measure of the model's accuracy and represents the proportion of samples that the model correctly predicts. The fit function is called to train the model, specifying 35 training epochs and a batch size of 128, and at the same time using 10% of the data for validation during training. The fit function is the main function in Keras for training models, and it accepts training data, labels, number of training epochs (epochs), batch size (batch size), and validation data as parameters. After training, the model will be named and saved according to the current date and configuration parameters. At the same time, the show_history function is used to plot and display the loss and accuracy curves during the training process. Finally, the trained model is evaluated using the test set, the test accuracy is output, and the label mapping is saved to a file for subsequent use.

[0134] Through the verification and testing process, a final photovoltaic fault diagnosis model is ultimately obtained. This model can classify IV curve images and identify different fault types. Since the current data of the series string open circuit is almost close to 0, the method of image processing curve is not applicable. Here, after the IV data preprocessing, this type is detected by setting the open circuit threshold, and other fault types are identified by the fault type identification model. Combining the above processes and methods, Figure 9 is an overall flowchart of obtaining and constructing a photovoltaic fault diagnosis model from IV data according to an embodiment of the present invention. Figure 9 It includes the overall process of the photovoltaic fault diagnosis model structure from IV data acquisition to model diagnosis.

[0135] Figure 10 FIG. shows a schematic structural diagram of an electronic device 1 that can be used to implement an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0136] As Figure 10 shown, the electronic device 1 includes at least one processor 11, and a memory communicatively connected to at least one processor 11, such as a read-only memory ROM 12, a random access memory RAM 13, etc. Among them, the memory stores a computer program executable by at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory ROM 12 or the computer program loaded from the storage unit 18 into the random access memory RAM 13. In the RAM 13, various programs and data required for the operation of the electronic device 1 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. The input / output I / O interface 15 is also connected to the bus 14.

[0137] Multiple components in the electronic device 1 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 1 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0138] Processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 executes the various methods and processes described above, for example, the fault diagnosis method for a photovoltaic string.

[0139] In some embodiments, the fault diagnosis method for a photovoltaic string can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto electronic device 1 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the fault diagnosis method for a photovoltaic string described above can be executed. Alternatively, in other embodiments, processor 11 can be configured to execute the fault diagnosis method for a photovoltaic string in any other suitable manner (e.g., by means of firmware).

[0140] The various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGA), application-specific integrated circuits (ASIC), application-specific standard products (ASSP), system-on-chip systems (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a dedicated or general-purpose programmable processor, and can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0141] The computer program for implementing the method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowchart and / or block diagram are implemented. The computer program can be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0142] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0143] To provide for interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide for interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, speech input, or tactile input).

[0144] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0145] A computing system may include a client and a server. The client and the server are generally far from each other and usually interact via a communication network. The client-server relationship is created by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0146] It should be understood that various forms of processes shown above can be used, with steps reordered, added or deleted. For example, the steps described in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is imposed herein.

[0147] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A photovoltaic string fault diagnosis method, characterized in that: include: Get IV data of PV strings; Preprocessing the IV data, converting the preprocessed IV data into an IV curve image and preparing a data set; The data set is trained to obtain a photovoltaic fault type recognition model, and the photovoltaic fault type recognition model is combined with a photovoltaic fault degree discrimination method to obtain a photovoltaic fault diagnosis model; Performing fault diagnosis on the IV curve image according to the photovoltaic fault diagnosis model to obtain a fault diagnosis result of the photovoltaic string; The photovoltaic fault degree determination method comprises: Calculating a distance value between an actual maximum power point on the IV curve and a virtual maximum power point on the IV curve image, wherein the distance value is used as a threshold for determining a photovoltaic fault degree; Determining the degree of photovoltaic faults according to the photovoltaic fault degree determination threshold; The distance value is calculated as follows: ； in, is the distance between the actual maximum power point and the virtual maximum power point on the IV curve image, V mpp_px is the coordinate of the actual maximum power point on the horizontal axis of the IV curve image, I mpp_px is the coordinate of the actual maximum power point on the vertical axis of the IV curve image, V T_px is the coordinate of the virtual maximum power point on the horizontal axis of the IV curve image, I T_px is the virtual maximum power point P T The coordinate of the vertical axis on the IV curve image; The calculation formula of the coordinate of the actual maximum power point on the horizontal axis of the IV curve image is as follows: ； Among them, V mpp is the voltage at the actual maximum power point, V min is the minimum value in the voltage array, V max is the maximum value in the voltage array, and H is the height of the IV curve image; The calculation formula of the coordinate of the actual maximum power point on the vertical axis of the IV curve image is as follows: ； Among them, I mpp is the current at the actual maximum power point, I min is the minimum value in the current array, I max is the maximum value in the current array, W is the width of the IV curve image; The calculation formula of the coordinate of the virtual maximum power point on the horizontal axis of the IV curve image is as follows: ； in, ; The calculation formula of the coordinate of the virtual maximum power point on the vertical axis of the IV curve image is as follows: ； in, is the short circuit current.

2. The method according to claim 1, characterized in that The step of obtaining the IV data of the photovoltaic strings includes: Acquire historical fault IV data of the photovoltaic string under different fault types and different fault degrees; The IV data of the photovoltaic string when in normal use is obtained.

3. The method according to claim 1, characterized in that The preprocessing of the IV data, converting the preprocessed IV data into an IV curve image and preparing a data set comprises: Cleaning the IV data to remove erroneous data and obtain IV data that meets the conditions; The IV data meeting the conditions are drawn into an IV curve, and then normalized or resized or the IV curve coordinate axis is not displayed to produce an IV curve image; The IV curve images are classified according to different fault types and fault degrees and a data set is prepared.

4. The method according to claim 1, characterized in that: The step of training the data set to obtain a photovoltaic fault type identification model comprises: The photovoltaic fault type recognition model is obtained by training the data set based on a convolutional neural network model.

5. The method according to claim 2, characterized in that: The fault type includes: at least one of: shielding, hot spot, glass breakage, open circuit, short circuit and aging; The fault degree includes at least one of a minor fault, a medium fault and a major fault.

6. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the photovoltaic string fault diagnosis method as described in any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the fault diagnosis method for a photovoltaic string as described in any one of claims 1 to 5 is implemented.

Images