Direct-current arc identification method of photovoltaic power station and related device

Abstract

According to the direct-current arc identification method of the photovoltaic power station and the related device, the target electric signal is obtained from the direct-current cabinet, the target electric signal is subjected to signal decomposition processing based on the noise auxiliary analysis method, and multiple modal components are obtained; and performing modal screening based on the plurality of modal components, the target electric signal and the screening condition to obtain a target modal component, performing feature extraction on the target modal component to obtain a feature vector, and performing analysis based on the feature vector and a DC arc identification model to obtain a DC arc identification result. According to the scheme, the direct-current cabinet electric signal at each to-be-analyzed moment is continuously monitored, so that the operation is simple, analysis is carried out based on methods such as noise assistance and modal screening, and the direct-current arc occurrence condition in the photovoltaic power station at the to-be-analyzed moment is finally obtained. Identification does not depend on the number and quality of experimental samples, the detection calculation amount of the direct-current arc is reduced, and the accuracy and timeliness of direct-current arc detection are improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a direct current arc identification method and related devices for a photovoltaic power station. Background Art

[0002] With the widespread application of new energy technologies, the photovoltaic industry has flourished. Photovoltaic power stations, as a power generation system that converts sunlight into electrical energy through solar panels, have attracted widespread attention and application due to their clean and pollution-free characteristics. There are multiple arrays and components in a photovoltaic power station, which are connected together by cables, connectors, etc.

[0003] However, during the long-term operation of a photovoltaic power station, the array and component connectors may age and fall off, the wires may be damaged, the connections may fail, and the cables may have poor contact, which may lead to DC fault arcs. Since DC arcs have no zero-crossing characteristics, the combustion temperature is high and it is not easy to extinguish naturally, which may cause serious fires. Therefore, it is necessary to monitor the DC side of the photovoltaic power station in real time.

[0004] At present, the detection of DC arcs in photovoltaic power stations still relies on manual operation and positioning, has delayed detection response time, high false alarm rate, and complex operation. In addition, the analysis and calculation process is computationally intensive and has high hardware requirements. Summary of the invention

[0005] The embodiments of the present application provide a DC arc identification method and related devices for a photovoltaic power station, which are beneficial to reducing the amount of DC arc detection calculations and improving the accuracy and timeliness of DC arc detection.

[0006] In a first aspect, an embodiment of the present application provides a DC arc identification method for a photovoltaic power station, which is applied to a server of a photovoltaic power generation system, wherein the photovoltaic power generation system also includes a DC cabinet, and the method includes:

[0007] Obtaining a target electrical signal from a DC cabinet, where the target electrical signal is an electrical signal obtained from the DC cabinet at a time to be analyzed, and the DC cabinet is used to aggregate electrical signals generated in a photovoltaic power station;

[0008] The target electrical signal is decomposed based on the noise-assisted analysis method to obtain multiple modal components, which are used to characterize the characteristics of the target signal in the frequency domain;

[0009] Performing modal screening based on multiple modal components, target electrical signals, and screening conditions to obtain target modal components, where the screening conditions are preset screening conditions for limiting the target modal components;

[0010] Feature extraction is performed on the target modal component to obtain a feature vector, and the target electrical signal and the feature vector have a corresponding relationship;

[0011] The DC arc identification result is obtained by analyzing the characteristic vector and the DC arc identification model. The DC arc identification model is a pre-trained model. The DC arc identification result is used to characterize the occurrence of DC arc in the photovoltaic power station at the time to be analyzed.

[0012] In a possible embodiment, before the step of acquiring the target electrical signal from the DC cabinet, the method further includes:

[0013] Acquire multiple arc fault electrical signals, each arc fault electrical signal is generated by an arc occurrence model, each arc occurrence model represents a different arc occurrence position, and the arc occurrence model is used to generate the arc fault electrical signal;

[0014] Model training is performed based on at least one arc fault electrical signal and a neural network model to obtain a DC arc identification model.

[0015] In a possible embodiment, model training is performed based on at least one arc fault electrical signal and a neural network model to obtain a DC arc identification model, including:

[0016] Performing signal decomposition processing on the arc fault electrical signal to obtain at least one training modal component, where the training modal component is a modal component used for model training;

[0017] Performing validity analysis based on at least one training modal component and at least one arc fault electrical signal to obtain a target training modal component;

[0018] Perform feature extraction on the target training modal component to obtain multiple training feature vectors;

[0019] A neural network model is constructed according to a plurality of training feature vectors and the neural network model is trained to obtain a DC arc identification model.

[0020] In a possible embodiment, the target electrical signal is subjected to signal decomposition processing based on a noise-assisted analysis method to obtain multiple modal components including:

[0021] Adding a plurality of noise signals to the target electrical signal in sequence to obtain a plurality of first electrical signals, wherein the noise signal is a preset signal;

[0022] Performing signal dissociation on each first electrical signal to obtain multiple frequency components, each first electrical signal corresponding to the multiple frequency components;

[0023] An average operation is performed based on multiple frequency components corresponding to each first electrical signal to obtain multiple modal components.

[0024] In a possible embodiment, the screening condition includes an effective threshold, and modal screening is performed based on multiple modal components, a target electrical signal, and the screening condition, and the target modal components obtained include:

[0025] Performing modal screening on each modal component based on multiple modal components and target electrical signals to obtain multiple validity parameters, where the validity parameters are used to characterize the degree of correlation between the modal component and the target electrical signal corresponding to the modal component;

[0026] The target modal component is obtained according to a plurality of validity parameters and a validity threshold, wherein the validity threshold is preset, wherein if it is detected that the validity parameter is greater than the validity threshold, the modal component is the target modal component.

[0027] In a possible embodiment, modal screening is performed on each modal component based on multiple modal components and target electrical signals to obtain multiple validity parameters, including:

[0028] The validity parameters are obtained by performing correlation analysis on multiple modal components according to the following formula:

[0029]

[0030] in, is the validity parameter, x(t) is the target electrical signal, c i (t) is the modal component, t is time, i is an integer greater than or equal to 1 and less than or equal to M, and M is the number of modal components.

[0031] In a possible embodiment, extracting features from the target modal component to obtain a feature vector includes:

[0032] Performing statistical calculation on each target modal component to obtain a first statistic and a second statistic, wherein the first statistic includes a modal mean and the second statistic includes a modal standard deviation;

[0033] Performing distribution statistical analysis on each target modal component according to the first statistic and the second statistic to obtain a kurtosis value, wherein each target modal component corresponds to a kurtosis value;

[0034] The eigenvectors of the target modal components are obtained based on all kurtosis values.

[0035] In a second aspect, an embodiment of the present application provides a DC arc identification device for a photovoltaic power station, which is applied to a server of a photovoltaic power generation system, wherein the photovoltaic power generation system also includes a DC cabinet, and the device includes:

[0036] An acquisition module is used to acquire a target electrical signal from a DC cabinet, where the target electrical signal is acquired from the DC cabinet, and the DC cabinet is used to aggregate photovoltaic electrical signals generated in a photovoltaic power station;

[0037] A decomposition module is used to perform signal decomposition processing on the target electrical signal based on the noise-assisted analysis method to obtain multiple modal components, and the modal components are used to characterize the characteristics of the target signal in the frequency domain;

[0038] A screening module, used for performing modal screening based on multiple modal components, target electrical signals and screening conditions to obtain target modal components, wherein the screening conditions are preset screening conditions for limiting the target modal components;

[0039] An extraction module is used to extract features of target modal components to obtain feature vectors, and the target electrical signal and the feature vectors have a corresponding relationship;

[0040] The identification module is used to obtain a DC arc identification result by analyzing based on the characteristic vector and the DC arc identification model. The DC arc identification model is a pre-trained model. The DC arc identification result is used to characterize the occurrence of the DC arc.

[0041] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the program includes instructions for executing some or all of the steps described in the first aspect of the embodiment of the present application.

[0042] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium on which a computer program is stored. The computer storage medium stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, the processor executes some or all of the steps described in the first aspect.

[0043] In a fifth aspect, an embodiment of the present application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps described in the first aspect of the embodiment of the present application. The computer program product may be a software installation package.

[0044] By implementing the embodiments of the present application, a target electrical signal is obtained from a DC cabinet, the target electrical signal is an electrical signal obtained from the DC cabinet at the time to be analyzed, and the DC cabinet is used to summarize the electrical signals generated in the photovoltaic power station; the target electrical signal is subjected to signal decomposition processing based on a noise-assisted analysis method to obtain a plurality of modal components, and the modal components are used to characterize the characteristics of the target signal in the frequency domain; modal screening is performed based on the plurality of modal components, the target electrical signal, and screening conditions to obtain the target modal component, and the screening condition is a preset screening condition for limiting the target modal component; feature extraction is performed on the target modal component to obtain a feature vector, and the target electrical signal and the feature vector have a corresponding relationship; based on the feature vector and DC arc identification The DC arc identification model is used to analyze and obtain the DC arc identification result. The DC arc identification model is a pre-trained model. The DC arc identification result is used to characterize the occurrence of DC arc in the photovoltaic power station at the time to be analyzed. Compared with the existing strategy of collecting photovoltaic side data at multiple time points for arc monitoring, this scheme continuously monitors the DC cabinet electrical signal at each time to be analyzed, which makes the operation simple, and based on this, noise assistance and mode screening methods are added for analysis, and finally the occurrence of DC arc in the photovoltaic power station at the time to be analyzed is obtained, which realizes that the identification does not depend on the quantity and quality of experimental samples, reduces the detection calculation amount of DC arc, and improves the accuracy and timeliness of DC arc detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the background technology, the drawings required for use in the embodiments of the present invention or the background technology will be described below.

[0046] Figure 1 It is a schematic diagram of the architecture of a photovoltaic power generation system provided in an embodiment of the present application;

[0047] Figure 2 It is a flow chart of a DC arc identification method for a photovoltaic power station provided in an embodiment of the present application;

[0048] Figure 3 It is a flow chart of signal decomposition processing of a DC arc identification method for a photovoltaic power station provided by an embodiment of the present application;

[0049] Figure 4 It is a scene schematic diagram of a DC arc identification method for a photovoltaic power station provided in an embodiment of the present application;

[0050] Figure 5 is a flowchart illustrating a DC arc identification method for a photovoltaic power station proposed in an embodiment of the present application;

[0051] Figure 6 It is a structural schematic diagram of a DC arc identification device for a photovoltaic power station proposed in an embodiment of the present application;

[0052] Figure 7 It is a structural schematic diagram of an electronic device proposed in an embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0054] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or electronic device that includes a series of steps or units is not limited to the listed steps or units, but in an optional example also includes steps or units that are not listed, or in an optional example also includes other steps or units inherent to these processes, methods, products, or electronic devices.

[0055] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0056] The DC arc identification method and related device of the photovoltaic power station provided in the embodiment of the present application can be applied to Figure 1 In the photovoltaic power generation system shown, please refer to Figure 1 , Figure 1 1 is a schematic diagram of the architecture of a photovoltaic power generation system provided in an embodiment of the present application. The photovoltaic power generation system 100 includes a server 110 and an acquisition terminal 120 . The acquisition terminal 120 can perform data transmission with the server 110 .

[0057] The server 110 refers to a remote computer used to process a large number of computing tasks and store data. In this solution, a program for DC arc identification of a photovoltaic power station is deployed on the server 110 to identify the occurrence of DC arcs in the photovoltaic power station. The server 110 can also be used to collect data during the use of the model, so as to facilitate the subsequent optimization of the program for DC arc identification of the photovoltaic power station.

[0058] The acquisition terminal 120 is a device for acquiring the target electrical signal of the photovoltaic power station, which may be a DC cabinet or other device for managing the DC power generated by the photovoltaic module, which is not limited here. The acquisition terminal 120 is connected to the photovoltaic module, and the photovoltaic modules may also be connected to each other. The DC arc may occur between the photovoltaic module, the photovoltaic module and / or the acquisition terminal 120. The photovoltaic module is a device for converting solar energy into electrical energy.

[0059] In a possible embodiment, Figure 1 As shown, the photovoltaic power generation system 100 further includes a user terminal, and the user terminal can communicate with the server 110 .

[0060] Among them, the user terminal refers to the device used by the user, and the DC arc identification result obtained by the server 110 analysis can be transmitted to the user terminal. The device may include a tablet computer, a PDA, an in-vehicle electronic device, a server, a laptop computer, a mobile Internet electronic device (MID, Mobile Internet Devices) or a wearable electronic device (such as a smart watch, a Bluetooth headset), a smart phone (such as an Android phone, an iOS phone, a Windows Phone phone, etc.), etc. The above are only examples, not exhaustive, including but not limited to the above electronic devices. The server 110 can send the DC arc identification result of the photovoltaic power station to the user terminal for the user to view.

[0061] Based on this, the present application provides a DC arc identification method and related devices for a photovoltaic power station, and the present application is described in detail below in conjunction with the accompanying drawings.

[0062] See also Figure 2 , Figure 2 1 is a flow chart of a method for identifying a DC arc in a photovoltaic power station provided in an embodiment of the present application. The method is applied to a server of a photovoltaic power generation system, and the photovoltaic power generation system also includes a DC cabinet. Figure 2 As shown, the method comprises the following steps:

[0063] S210, obtaining a target electrical signal from a DC cabinet, where the target electrical signal is an electrical signal obtained from the DC cabinet at a time to be analyzed, and the DC cabinet is used to aggregate electrical signals generated in a photovoltaic power station.

[0064] The DC cabinet obtains an electrical signal from the photovoltaic module of the photovoltaic power station, the target electrical signal may be an electrical signal obtained from the DC cabinet at the time to be analyzed, the electrical signal may be a current signal, and the electrical signal may be a total electrical signal, which may be used to analyze whether a DC arc exists at multiple locations where a DC arc may occur in the entire photovoltaic module at the time to be analyzed. The above-mentioned locations where a DC arc may occur may be obtained through experiments or expert experience, or may be obtained by other methods, which are not limited here.

[0065] The time to be analyzed may be every continuous moment, and analysis is performed at every moment, or may be a specified moment, such as performing analysis once every unit interval time. The specific time is determined according to actual conditions and is not limited here.

[0066] For example, if it is set to perform an analysis once per unit time interval, and the unit time interval is 1 second, the DC cabinet will send the electrical signal collected at that moment as the target electrical signal to the server every second, and the server will obtain and analyze it. If it is set to perform an analysis once per unit time period, the moment to be analyzed refers to the unit moment in the unit time period. For example, if the unit time period is T, the target electrical signal corresponding to the unit moment in the unit time period is x(t). The above T can contain multiple ts or one t, which is not limited here.

[0067] S220, performing signal decomposition processing on the target electrical signal based on the noise-assisted analysis method to obtain multiple modal components, where the modal components are used to characterize the characteristics of the target signal in the frequency domain.

[0068] Among them, the noise-assisted analysis method can be to inject multiple groups of noise into the target electrical signal to obtain multiple modal classifications corresponding to the target electrical signal, and the above modal components are used to characterize the frequency domain characteristics of the target electrical signal at the acquisition time, that is, the time to be analyzed. The above injected noise can be a Gaussian white noise signal, or other noise signals, such as white noise signals, adaptive white noise signals, and auxiliary white noise signals, etc., which are not limited here.

[0069] In a possible embodiment, the target electrical signal is subjected to signal decomposition processing based on a noise-assisted analysis method to obtain multiple modal components, including: adding multiple noise signals to the target electrical signal in sequence to obtain multiple first electrical signals, where the noise signal is a preset signal; performing signal decomposition on each first electrical signal to obtain multiple frequency components, each first electrical signal corresponds to multiple frequency components; and performing averaging operations on the multiple frequency components corresponding to each first electrical signal to obtain multiple modal components.

[0070] The signal decomposition process may be an empirical mode decomposition (EMD), and the intrinsic mode function components are intrinsic mode functions (IMFs) that meet specific conditions extracted in the EMD process, which represent the local characteristics of the original signal, i.e., the target electrical signal, at different time scales. The process of performing signal decomposition processing on the target electrical signal based on the noise-assisted analysis method to obtain multiple modal components is a process based on ensemble empirical mode decomposition (EEMD).

[0071] Among them, multiple noise signals are added to the target electrical signal to obtain multiple first electrical signals, a single first electrical signal is a composite signal of the target electrical signal and a single noise signal, one target electrical signal corresponds to multiple first electrical signals, and the noise signal can be a Gaussian white noise signal, or other noise signals, such as white noise signals, adaptive white noise signals, and auxiliary white noise signals, etc., which are not limited here. After obtaining multiple first electrical signals, the multiple first electrical signals are subjected to signal dissociation to obtain multiple frequency components. The frequency component is the intrinsic mode function component of the first electrical signal, which reflects the local characteristics of the target electrical signal on different time scales. Then, the multiple frequency components are averaged to obtain multiple modal components.

[0072] For details, please refer to Figure 3 , Figure 3 1 is a flow chart of signal decomposition processing of a DC arc identification method for a photovoltaic power station provided by an embodiment of the present application. Figure 3 As shown, if the target electrical signal is x(t), N groups of noise signals are added to the target electrical signal to obtain the first electrical signal x j (t), including x 1 (t), x 2 (t), …, x j (t), j is an integer greater than or equal to 1 and less than or equal to N, then the first electrical signal x j (t) Perform signal dissociation. The signal dissociation process can be EMD to obtain multiple frequency components C ij (t), including C i1 (t), C i2 (t),…,C ij (t), where each first electrical signal x j (t) corresponds to M frequency components C ij (t), i is an integer greater than or equal to 1 and less than or equal to M. For each first electrical signal x j (t) corresponds to the i-th frequency component C ij(t) calculate the average value and obtain i modal components C i (t), where

[0073] It can be seen that in this embodiment, white noise is added to the target electrical signal and the signal is decomposed to obtain multiple modal components. The target electrical signal is decomposed into multiple modal components. The multiple modal components can characterize the characteristics of the frequency domain and time domain at different positions, reducing the problem of modal aliasing, which is beneficial to improving the accuracy of the target electrical signal decomposition, and further beneficial to improving the accuracy of DC arc identification.

[0074] S230, performing modal screening based on multiple modal components, the target electrical signal and a screening condition to obtain a target modal component, where the screening condition is a preset screening condition for limiting the target modal component.

[0075] Among them, the target modal component is a modal component that meets the screening conditions, and the screening conditions are used to limit the screening conditions of the target modal component. The specific screening index needs to be obtained based on multiple modal components and target electrical signals.

[0076] The screening condition may be a preset threshold or may be compliance with a certain condition, which is not limited here.

[0077] In a possible embodiment, the screening condition includes an effective threshold, and modal screening is performed based on multiple modal components, a target electrical signal, and the screening condition to obtain the target modal component, including: performing modal screening on each modal component based on multiple modal components and the target electrical signal to obtain multiple effectiveness parameters, the effectiveness parameters are used to characterize the degree of association between the modal component and the target electrical signal corresponding to the modal component; obtaining the target modal component based on multiple effectiveness parameters and the effective threshold, the effective threshold is preset, wherein, if it is detected that the effectiveness parameter is greater than the effective threshold, the modal component is the target modal component.

[0078] The specific screening index is the validity parameter obtained by modal screening based on multiple modal components and the target electrical signal, the screening condition is the effective threshold, and the target modal component is obtained based on the validity parameter and the effective threshold. Specifically, if it is detected that the validity parameter is greater than the effective threshold, it indicates that the modal component corresponding to the validity parameter is the target modal component.

[0079] The process of obtaining the validity parameter by the above-mentioned modal screening can be to perform a correlation calculation, specifically, it can be a Pearson correlation coefficient calculation, a Spearman rank correlation coefficient calculation, a distance correlation coefficient calculation, and other calculation methods that can calculate the correlation of several values, which are not specifically limited here. The validity parameter obtained after the modal screening can characterize the degree of correlation between the modal component and the original quantity, that is, the target electrical signal, and the screening condition is used to limit the degree of correlation between the modal component.

[0080] It can be seen that in this embodiment, it is proposed to screen the modal components based on the validity parameters of the target electrical signal and the modal components, and obtain the modal components whose correlation degree with the target electrical signal is greater than the valid threshold as the target modal components for subsequent use, and screen out those modal components that can best reflect the arc characteristics in the target electrical signal, which is beneficial to improving the accuracy of DC arc identification.

[0081] In a possible embodiment, performing modal screening on each modal component based on the multiple modal components and the target electrical signal to obtain multiple validity parameters includes: performing correlation analysis on the multiple modal components according to the following formula to obtain the validity parameters:

[0082]

[0083] in, is the validity parameter, x(t) is the target electrical signal, c i (t) is the modal component, t is time, i is an integer greater than or equal to 1 and less than or equal to M, and M is the number of modal components.

[0084] The validity parameter is obtained by performing correlation analysis according to the above correlation formula. The validity parameter can be a correlation parameter used to characterize the magnitude of the correlation between the modal component and the target electrical signal.

[0085] After the validity parameter is obtained by performing correlation analysis according to the above correlation formula, the validity threshold can be set to a number between 0 and 1, which is not limited here. For example, the validity threshold can be 0.7. Then the c i (t) is the target modal component; if This means that the modal component does not meet the screening criteria and is discarded as a useless modal component.

[0086] It can be seen that in this embodiment, a specific correlation calculation formula is proposed to perform modal screening, and those modal components that can best reflect the arc characteristics in the target electrical signal are screened out, which is beneficial to improving the accuracy of DC arc identification.

[0087] S240, extracting features of the target modal component to obtain a feature vector, and the target electrical signal and the feature vector have a corresponding relationship.

[0088] Among them, the process of extracting features from the target modal component may include calculations such as kurtosis coefficient calculation, skewness calculation, fourth-order moment calculation or kurtosis calculation, and the method may be selected as needed, and the specific method is not limited here. By performing the above calculations on the target modal component, a characteristic component corresponding to the target modal component is obtained, and the characteristic components corresponding to all target modal components corresponding to the target electrical signal are taken as characteristic components to obtain a characteristic vector corresponding to the target electrical signal.

[0089] In a possible embodiment, extracting features from a target modal component to obtain a feature vector includes: performing statistical calculations on each target modal component to obtain a first statistic and a second statistic, the first statistic including a modal mean, and the second statistic including a modal standard deviation; performing distribution statistical analysis on each target modal component according to the first statistic and the second statistic to obtain a kurtosis value, each target modal component corresponding to a kurtosis value; and obtaining a feature vector of the target modal component based on all kurtosis values.

[0090] The statistical calculation may include performing mean calculation and standard deviation calculation, and the first statistic and the second statistic may be calculated by the following formula:

[0091]

[0092] in, is the first statistic, s is the second statistic, c i (t) is the target modal component, T is the sequence length of the target modal component, and the sequence length of the above-mentioned target modal component may refer to the numerical quantity of the target modal component, that is, corresponding to the description in the above-mentioned step S210, T may include multiple ts or one t, and the quantity is not limited.

[0093] It should be noted that if T includes multiple ts, it means that there are multiple target electrical signals. After performing steps S210 to S230 on each target electrical signal, when the statistics are calculated in step S240, the statistics may be for multiple target electrical signals, that is, the obtained feature vector result is for the entire T. If T includes one t, it means that there is only one target electrical signal, and the feature vector corresponds to the target electrical signal.

[0094] Among them, after obtaining the first statistic and the second statistic, the kurtosis value calculation can be selected to perform characteristic component calculation, that is, the distribution statistical analysis of each target modal component is performed according to the first statistic and the second statistic to obtain the kurtosis value, each target modal component corresponds to a kurtosis value, and the kurtosis value is used as the characteristic component to construct a characteristic vector, and each target electrical signal corresponds to a characteristic vector.

[0095] Specifically, the kurt value kurt(c i (t)):

[0096]

[0097] For example, i is an integer greater than or equal to 1 and less than or equal to M. Assuming that there are K target modal components screened out, K kurtosis values ​​are calculated, and the K kurtosis values ​​are used as feature components to form a feature vector, and a K-dimensional feature vector K=[k 1 ,k 2 ,…,k k ], the feature vector corresponds to a target electrical signal.

[0098] It should be noted that if the sample follows a normal distribution, its kurtosis value is 0; if the kurtosis value is greater than 0, it means that the distribution is steeper than the normal distribution, and if the kurtosis value is less than 0, it means that the distribution is flatter than the normal distribution. If the kurtosis value of a modal component is particularly large, it means that the data of the modal component fluctuates most violently, and the probability of containing fault characteristics is also greater.

[0099] It can be seen that in this embodiment, the kurtosis calculation is selected for feature vector extraction, and the obtained kurtosis value is used as the characteristic component to construct the feature vector. The kurtosis value can characterize the intensity of data fluctuation. The kurtosis can be used as a feature for fault identification. The features characterizing the existence of arcs in the target electrical signal can be extracted, which is beneficial to improve the accuracy of DC arc identification.

[0100] S250, analyzing based on the feature vector and the DC arc identification model to obtain a DC arc identification result, where the DC arc identification model is a pre-trained model, and the DC arc identification result is used to characterize the occurrence of DC arc in the photovoltaic power station at the time to be analyzed.

[0101] Among them, the DC arc identification results include normal or specific arc occurrence positions. For example, the DC arc identification results include: normal, position 1, position 2, position 3, position 4, position 5. The above position 1, position 2, position 3, position 4, position 5 are preset and trained positions where arcs may occur.

[0102] For examples, see Figure 4 , Figure 4 is a schematic diagram of a scenario of a DC arc identification method for a photovoltaic power station provided in an embodiment of the present application, such as Figure 4As shown, the electrical signal generated by the photovoltaic module is transmitted to the DC cabinet through the busbar. The server obtains the target electrical signal from the DC cabinet, and performs data processing to finally obtain a feature vector, which is input as the input of the DC arc identification model. After the DC arc identification model is analyzed, the DC arc identification result is obtained. The result may be normal, position 1, position 2, position 3, position 4, position 5. The above positions 1, position 2, position 3, position 4, and position 5 represent Figure 4 A DC arc occurs at a position where a DC arc may occur, and a normal position indicates that no DC arc occurs at the above positions 1, 2, 3, 4, and 5. Please refer to Table 1 below for details. Table 1 is used to illustrate the relationship between the input feature vector and the output result and the relationship between the input and output of the DC arc identification model:

[0103] Kurtosis 1 Kurtosis 2 Kurtosis 3 Kurtosis 4 …… Kurtosis K DC arc identification results 1 9.10 8.00 4.00 1.00 …… 10.00 Position 1 2 2.00 8.54 5.00 5.30 …… 8.56 Location 5 3 7.00 7.00 9.00 3.00 …… 3.00 Location 3 4 6.00 1.00 7.00 4.00 …… 3.00 Location 3 5 4.00 1.00 8.00 9.30 …… 5.00 Position 4 6 7.00 6.00 3.00 3.00 …… 3.00 Position 1 7 7.54 3.00 6.00 4.55 …… 9.00 Position 4 8 7.00 8.00 6.65 5.00 …… 8.00 Location 3 9 10.00 7.00 2.00 1.00 …… 9.00 Position 4 10 4.00 9.98 4.00 5.00 …… 4.00 Location 5 11 3.09 3.11 2.36 2.10 …… 1.98 normal 12 2.65 2.00 3.00 2.00 …… 1.00 normal 13 2.98 2.40 2.00 1.00 …… 3.00 normal …… …… …… …… …… …… …… ……

[0104] Table 1

[0105] The DC arc identification model is a pre-trained model, and the training process of the model may be to use the training data obtained in the same process as step S210 to step S240 to perform model training.

[0106] In a possible embodiment, before the step of obtaining the target electrical signal from the DC cabinet, the method also includes: obtaining multiple arc fault electrical signals, each arc fault electrical signal is generated by an arc occurrence model, each arc occurrence model represents a different arc occurrence location, and the arc occurrence model is used to generate the arc fault electrical signal; performing model training based on at least one arc fault electrical signal and a neural network model to obtain a DC arc identification model.

[0107] The arc occurrence model is a pre-built model for simulating photovoltaic modules in a real photovoltaic power station. Multiple arc occurrence sub-models can be constructed according to different DC arc occurrence locations, and multiple arc occurrence sub-models together constitute the arc occurrence model. The arc fault electrical signal is generated by an arc occurrence sub-model, which is an electrical signal that characterizes the occurrence of a DC arc at a location in the photovoltaic module. The multiple arc fault electrical signals may include multiple electrical signals generated by an arc occurrence sub-module multiple times, or multiple electrical signals generated by multiple arc occurrence sub-modules, which may be multiple times or a single time, and are not limited here.

[0108] Specifically, model training is performed based on at least one arc fault electrical signal and a neural network model to obtain a DC arc identification model, including: performing signal decomposition processing based on the arc fault electrical signal to obtain at least one training modal component, where the training modal component is a modal component used for model training; performing validity analysis based on at least one training modal component and at least one arc fault electrical signal to obtain a target training modal component; performing feature extraction on the target training modal component to obtain a plurality of training feature vectors; constructing a neural network model based on the plurality of training feature vectors and training the neural network model to obtain a DC arc identification model.

[0109] Among them, the method for obtaining the above-mentioned training modal component corresponds to the process of obtaining the modal component described in step S220, the method for obtaining the above-mentioned target training modal component corresponds to the process of obtaining the target modal component described in step S230, and the method for obtaining the above-mentioned training feature vector corresponds to the process of obtaining the feature vector described in step S240, which will not be repeated here.

[0110] It can be seen that in this embodiment, the initial training data, i.e., the arc fault electric signal, is obtained through the arc occurrence model, and data under various arc occurrence conditions can be obtained, and the arc occurrence model is flexible and changeable. The DC arc occurrence condition can be obtained by directly processing the target electric signal to obtain the feature vector and inputting it into the trained arc occurrence model, thereby achieving simple operation, reducing the amount of DC arc detection calculation, improving the accuracy and calculation speed of DC arc detection, and improving the accuracy and timeliness of DC arc detection.

[0111] In a possible embodiment, a neural network model is constructed according to a plurality of training feature vectors and the neural network model is trained to obtain a DC arc identification model, including:

[0112] 51. Input the training feature vector as an input into an input layer of a neural network model, wherein the neural network model includes an input layer, a hidden layer, and an output layer;

[0113] 52. Setting model parameters, the model parameters include a function center and a function width of a neural network model; and determining an implicit output value of an implicit node based on the function center and the function width and a training feature vector, wherein the hidden layer has a plurality of hidden nodes, and each of the hidden nodes outputs an implicit output value; and determining an output weight of an output layer based on a stochastic gradient descent method, wherein the output weight is used to limit a proportion of each implicit output value, and each implicit output value corresponds to an output weight;

[0114] 53. Obtain output results based on the above model parameters and neural network model;

[0115] 54. Iterative training is performed based on the output results, model parameters, neural network model and target results to obtain a DC arc identification model, wherein the target result is an arc occurrence condition set by the arc occurrence model corresponding to the training feature vector.

[0116] The neural network model may be a Gaussian radial basis function neural network model, or other neural network models having an input layer, a hidden layer, and an output layer, which is not limited here.

[0117] The hidden layer has multiple hidden nodes, the number of hidden nodes may be Q, and the number of hidden nodes Q may be determined according to the following formula: Q = log 2 m, m is the number of input layer nodes, and the number of input layer nodes is determined according to the dimension of the training feature vector, and specifically may be equal to the dimension of the training feature vector.

[0118] Among them, the Q hidden nodes from the input layer to the hidden layer are activated with Gaussian radial basis functions so that the input layer is mapped to the hidden layer, and each hidden node of the hidden layer corresponds to a Gaussian radial basis function. The implicit output value of the hidden node of the hidden layer can be set according to the following formula:

[0119]

[0120] X is the input training feature vector, ||Xc q || is the Euclidean norm, c q is the center of the Gaussian radial basis function corresponding to the qth hidden node, d q is the width of the Gaussian radial basis function corresponding to the qth hidden node. q and d q The determination can be determined by K-means clustering algorithm q , and then calculate the width d based on the standard deviation of the data distribution q , can also be determined by other methods, which are not limited here. The output weights of the above output layer can be calculated according to the stochastic gradient descent method to obtain a Q-dimensional weight: W = [ω 1 ,ω 2 ,…,ω q ,…,ω Q ] T , where ω q is the connection weight between the neurons in the hidden layer and the output layer, which is used to map the implicit output value of the hidden layer to the predicted result of the output layer, that is, the output result. The output result of the above output is determined according to the following formula: t =ω 1 h 1 +ω 1 h 2 +…+ωQ h Q ,p t is the probability value of DC arc generation output by the neural network model at time t.

[0121] Among them, in the iterative training step based on the output results, model parameters, neural network model and target results, the output results are adjusted by continuously adjusting the model parameters, so that the output results and the target results meet the preset training completion requirements, which are not limited here, and a DC arc identification model with higher accuracy is obtained.

[0122] It can be seen that in this embodiment, by constructing a radial basis vector neural network model and performing training iterations to finally obtain a neural network model whose input feature vector can obtain arc identification results, the amount of calculation is reduced, and the identification process has low hardware requirements, thereby achieving the recognition independent of the quantity and quality of experimental samples, reducing the amount of calculation for DC arc detection, and improving the accuracy and timeliness of DC arc detection.

[0123] For examples, see Figure 5 , Figure 5 FIG. 1 is a flowchart illustrating a method for identifying a DC arc in a photovoltaic power station according to an embodiment of the present application. Figure 5 As shown, the server obtains the target electrical signal from the DC cabinet, performs signal decomposition processing on the target electrical signal based on noise-assisted analysis, and performs modal screening based on screening conditions to obtain the target modal component, and performs feature extraction on the target modal component to obtain a K-dimensional feature vector; the feature components k1, k2, ..., kk of the K-dimensional feature vector are input into the DC arc identification model, and the output value pt of the output layer is obtained through analysis of the input layer, hidden layer, and output layer, and finally the DC arc identification result is obtained.

[0124] It can be seen that in this embodiment, by implementing the embodiment of the present application, a target electrical signal is obtained from a DC cabinet, the target electrical signal is an electrical signal at the time to be analyzed obtained from the DC cabinet, and the DC cabinet is used to summarize the electrical signals generated in the photovoltaic power station; the target electrical signal is subjected to signal decomposition processing based on the noise-assisted analysis method to obtain multiple modal components, and the modal components are used to characterize the characteristics of the target signal in the frequency domain; modal screening is performed based on multiple modal components, the target electrical signal, and screening conditions to obtain the target modal component, and the screening condition is a preset screening condition for limiting the target modal component; feature extraction is performed on the target modal component to obtain a feature vector, and the target electrical signal and the feature vector have a corresponding relationship; based on the feature vector and The DC arc identification model is used for analysis to obtain the DC arc identification result. The DC arc identification model is a pre-trained model. The DC arc identification result is used to characterize the occurrence of DC arc in the photovoltaic power station at the time to be analyzed. Compared with the existing strategy of collecting photovoltaic side data at multiple time points for arc monitoring, this scheme continuously monitors the DC cabinet electrical signal at each time to be analyzed, which makes the operation simple, and based on this, noise assistance and modal screening methods are added for analysis, and finally the occurrence of DC arc in the photovoltaic power station at the time to be analyzed is obtained, which realizes that the identification does not depend on the quantity and quality of experimental samples, reduces the detection calculation amount of DC arc, and improves the accuracy and timeliness of DC arc detection.

[0125] See also Figure 6 , Figure 6 Schematic diagram of the structure of a DC arc identification device for a photovoltaic power station proposed in an embodiment of the present application, which is applied to a server of a photovoltaic power generation system, wherein the photovoltaic power generation system also includes a DC cabinet, such as Figure 6 As shown, the DC arc identification device 600 of the photovoltaic power station includes: an acquisition module 610, a decomposition module 620, a screening module 630, an extraction module 640 and an identification module 650, wherein:

[0126] An acquisition module 610 is used to acquire a target electrical signal from a DC cabinet, where the target electrical signal is acquired from a DC cabinet, and the DC cabinet is used to aggregate photovoltaic electrical signals generated in a photovoltaic power station;

[0127] A decomposition module 620 is used to perform signal decomposition processing on the target electrical signal based on the noise-assisted analysis method to obtain multiple modal components, and the modal components are used to characterize the characteristics of the target signal in the frequency domain;

[0128] A screening module 630 is used to perform modal screening based on multiple modal components, a target electrical signal and a screening condition to obtain a target modal component, where the screening condition is a preset screening condition for limiting the target modal component;

[0129] An extraction module 640 is used to extract features from the target modal component to obtain a feature vector, and the target electrical signal and the feature vector have a corresponding relationship;

[0130] The identification module 650 is used to obtain a DC arc identification result by analyzing based on the characteristic vector and the DC arc identification model. The DC arc identification model is a pre-trained model. The DC arc identification result is used to characterize the occurrence of the DC arc.

[0131] In a possible implementation, the decomposition module 620 performs signal decomposition processing on the target electrical signal based on the noise-assisted analysis method to obtain multiple modal components, specifically for:

[0132] Adding a plurality of noise signals to the target electrical signal in sequence to obtain a plurality of first electrical signals, wherein the noise signal is a preset signal;

[0133] Performing signal dissociation on each first electrical signal to obtain multiple frequency components, each first electrical signal corresponding to the multiple frequency components;

[0134] An average operation is performed based on multiple frequency components corresponding to each first electrical signal to obtain multiple modal components.

[0135] In a possible implementation, the screening module 630 performs modal screening based on multiple modal components, target electrical signals, and screening conditions to obtain target modal components, specifically for:

[0136] Performing modal screening on each modal component based on multiple modal components and target electrical signals to obtain multiple validity parameters, where the validity parameters are used to characterize the degree of correlation between the modal component and the target electrical signal corresponding to the modal component;

[0137] The target modal component is obtained according to a plurality of validity parameters and a validity threshold, wherein the validity threshold is preset, wherein if it is detected that the validity parameter is greater than the validity threshold, the modal component is the target modal component.

[0138] In a possible implementation, the screening module 630 performs modal screening on each modal component based on the multiple modal components and the target electrical signal to obtain multiple validity parameters, specifically for:

[0139] The validity parameters are obtained by performing correlation analysis on multiple modal components according to the following formula:

[0140]

[0141] in, is the validity parameter, x(t) is the target electrical signal, c i(t) is the modal component, t is time, i is an integer greater than or equal to 1 and less than or equal to M, and M is the number of modal components.

[0142] In a possible implementation, the extraction module 640 is specifically used to extract features from the target modal component to obtain a feature vector:

[0143] Performing statistical calculation on each target modal component to obtain a first statistic and a second statistic, wherein the first statistic includes a modal mean and the second statistic includes a modal standard deviation;

[0144] Performing distribution statistical analysis on each target modal component according to the first statistic and the second statistic to obtain a kurtosis value, wherein each target modal component corresponds to a kurtosis value;

[0145] The eigenvectors of the target modal components are obtained based on all kurtosis values.

[0146] It is worth pointing out that the specific functional implementation of the DC arc identification device of the photovoltaic power station can be found in the above Figure 2 The description of the DC arc identification method for the photovoltaic system shown, for example, the acquisition module 610 is used to implement the relevant content of executing S210. The various units or modules in the DC arc identification device 600 of the photovoltaic power station can be respectively or completely combined into one or several other units or modules to constitute, or one (some) of the units or modules can be further divided into multiple functionally smaller units or modules to constitute, which can achieve the same operation without affecting the realization of the technical effects of the embodiments of the present invention. The above-mentioned units or modules are divided based on logical functions. In actual applications, the functions of one unit (or module) are implemented by multiple units (or modules), or the functions of multiple units (or modules) are implemented by one unit (or module).

[0147] It can be seen that the DC arc identification device of the photovoltaic power station described in the embodiment of the present application obtains the target electrical signal from the DC cabinet, the target electrical signal is the electrical signal obtained from the DC cabinet at the time to be analyzed, and the DC cabinet is used to summarize the electrical signals generated in the photovoltaic power station; the target electrical signal is subjected to signal decomposition processing based on the noise-assisted analysis method to obtain multiple modal components, and the modal components are used to characterize the characteristics of the target signal in the frequency domain; modal screening is performed based on the multiple modal components, the target electrical signal and the screening conditions to obtain the target modal component, and the screening conditions are preset screening conditions for limiting the target modal component; feature extraction is performed on the target modal component to obtain a feature vector, and the target electrical signal and the feature vector have a corresponding relationship; based on the feature The eigenvector and the DC arc identification model are analyzed to obtain the DC arc identification result. The DC arc identification model is a pre-trained model. The DC arc identification result is used to characterize the occurrence of DC arc in the photovoltaic power station at the time to be analyzed. Compared with the existing strategy of collecting photovoltaic side data at multiple time points for arc monitoring, this scheme continuously monitors the DC cabinet electrical signal at each time to be analyzed, which makes the operation simple, and based on this, noise assistance and modal screening methods are added for analysis, and finally the occurrence of DC arc in the photovoltaic power station at the time to be analyzed is obtained, which realizes that the identification does not depend on the quantity and quality of experimental samples, reduces the detection calculation amount of DC arc, and improves the accuracy and timeliness of DC arc detection.

[0148] See also Figure 7 , Figure 7 It is a structural diagram of an electronic device proposed in an embodiment of the present application. As shown in the figure, the electronic device 700 includes a processor 710, a memory 720, a communication interface 730 and one or more programs 721. The one or more programs 721 are stored in the memory 720 and are configured to be executed by the processor 710.

[0149] The processor 710, the memory 720, and the communication interface 730 are connected to each other and perform communication work with each other;

[0150] The memory 720 may be a volatile memory such as a dynamic random access memory DRAM, or a non-volatile memory such as a mechanical hard disk. The memory 720 is used to store a set of executable program codes, and the processor 710 is used to call one or more programs 721 stored in the memory 720, and may execute part or all of the steps of any DC arc identification method for a photovoltaic system as described in the above embodiments.

[0151] Among them, the electronic device 700 may include smart phones (such as Android phones, iOS phones, Windows Phone phones, etc.), tablet computers, PDAs, driving recorders, vehicle-mounted electronic devices, servers, laptops, mobile Internet electronic devices (MID, Mobile Internet Devices) or wearable electronic devices (such as smart watches, Bluetooth headsets), etc. The above are only examples and not exhaustive, including but not limited to the above electronic devices.

[0152] An embodiment of the present application also provides a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute part or all of the steps of any method described in the above method embodiments, and the above computer includes an electronic device.

[0153] The embodiment of the present application also provides a computer program product, the computer program product includes a non-transitory computer-readable storage medium storing a computer program, the computer program is operable to cause a computer to execute some or all of the steps of any method described in the method embodiment. The computer program product may be a software installation package, and the computer includes an electronic device.

[0154] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0155] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0156] In the several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above-mentioned units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0157] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0158] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0159] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer electronic device (which can be a personal computer, electronic device or network electronic device, etc.) to perform all or part of the steps of the above-mentioned methods of each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0160] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, and the memory can include: a flash drive, a read-only memory (English: Read-Only Memory, abbreviated as: ROM), a random access memory (English: Random Access Memory, abbreviated as: RAM), a magnetic disk or an optical disk, etc.

[0161] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for identifying a DC arc in a photovoltaic power station, characterized in that: A server applied to a photovoltaic power generation system, wherein the photovoltaic power generation system further comprises a DC cabinet, and the method comprises: Acquire a target electrical signal from the DC cabinet, wherein the target electrical signal is an electrical signal acquired from the DC cabinet at a time to be analyzed, and the DC cabinet is used to aggregate electrical signals generated in the photovoltaic power station; Performing signal decomposition processing on the target electrical signal based on a noise-assisted analysis method to obtain a plurality of modal components, wherein the modal components are used to characterize the characteristics of the target signal in the frequency domain; Performing modal screening based on the multiple modal components, the target electrical signal and a screening condition to obtain a target modal component, wherein the screening condition is a preset screening condition for limiting the target modal component; Extracting features of the target modal component to obtain a feature vector, wherein the target electrical signal and the feature vector have a corresponding relationship; A DC arc identification result is obtained by analyzing the characteristic vector and the DC arc identification model, wherein the DC arc identification model is a pre-trained model, and the DC arc identification result is used to characterize the occurrence of DC arc in the photovoltaic power station at the time to be analyzed.

2. The method according to claim 1, characterized in that Before the step of acquiring the target electrical signal from the DC cabinet, the method further includes: Acquire a plurality of arc fault electrical signals, each of the arc fault electrical signals being generated by an arc occurrence model, each of the arc occurrence models representing a different arc occurrence location, and the arc occurrence model being used to generate the arc fault electrical signal; Model training is performed based on the at least one arc fault electrical signal and a neural network model to obtain the DC arc identification model.

3. The method according to claim 2, characterized in that The performing model training based on the at least one arc fault electrical signal and the neural network model to obtain the DC arc identification model comprises: Performing the signal decomposition processing according to the arc fault electrical signal to obtain at least one training modal component, wherein the training modal component is a modal component used for model training; Performing validity analysis based on the at least one training modal component and the at least one arc fault electrical signal to obtain a target training modal component; Performing feature extraction on the target training modal component to obtain a plurality of training feature vectors; A neural network model is constructed according to the multiple training feature vectors and the neural network model is trained to obtain the DC arc identification model.

4. The method according to claim 1, characterized in that: The step of performing signal decomposition processing on the target electrical signal based on a noise-assisted analysis method to obtain multiple modal components includes: Adding a plurality of noise signals to the target electrical signal in sequence to obtain a plurality of first electrical signals, wherein the noise signal is a preset signal; Performing signal dissociation on each of the first electrical signals to obtain a plurality of frequency components, each of the first electrical signals corresponding to the plurality of frequency components; The multiple modal components are obtained by performing an averaging operation on the multiple frequency components corresponding to each of the first electrical signals.

5. The method according to claim 1, characterized in that The screening condition includes a valid threshold, and the modal screening is performed based on the multiple modal components, the target electrical signal and the screening condition to obtain the target modal component including: Performing modal screening on each of the modal components based on the multiple modal components and the target electrical signal to obtain a plurality of validity parameters, wherein the validity parameters are used to characterize the degree of association between the modal component and the target electrical signal corresponding to the modal component; The target modal component is obtained according to the multiple validity parameters and the validity threshold, and the validity threshold is preset, wherein if it is detected that the validity parameter is greater than the validity threshold, the modal component is the target modal component.

6. The method according to claim 5, characterized in that The performing modal screening on each of the modal components based on the multiple modal components and the target electrical signal to obtain multiple validity parameters includes: The validity parameter is obtained by performing correlation analysis on the multiple modal components according to the following formula: Among them, ρ xci is the validity parameter, x(t) is the target electrical signal, c i (t) is the modal component, t is time, i is an integer greater than or equal to 1 and less than or equal to M, and M is the number of the modal components.

7. The method according to claim 1, characterized in that The step of extracting features from the target modal component to obtain a feature vector comprises: Performing statistical calculation on each of the target modal components to obtain a first statistic and a second statistic, wherein the first statistic includes a modal mean, and the second statistic includes a modal standard deviation; Performing distribution statistical analysis on each of the target modal components according to the first statistic and the second statistic to obtain a kurtosis value, each of the target modal components corresponding to one kurtosis value; A eigenvector of the target modal component is obtained based on all the kurtosis values.

8. A DC arc identification device for a photovoltaic power station, characterized in that: A server applied to a photovoltaic power generation system, wherein the photovoltaic power generation system further comprises a DC cabinet, and the device comprises: An acquisition module, used for acquiring a target electrical signal from the DC cabinet, wherein the target electrical signal is acquired from the DC cabinet, and the DC cabinet is used for aggregating photovoltaic electrical signals generated in the photovoltaic power station; A decomposition module, used for performing signal decomposition processing on the target electrical signal based on a noise-assisted analysis method to obtain a plurality of modal components, wherein the modal components are used to characterize the characteristics of the target signal in the frequency domain; a screening module, configured to perform modal screening based on the multiple modal components, the target electrical signal and a screening condition to obtain a target modal component, wherein the screening condition is a preset screening condition for limiting the target modal component; An extraction module, used for performing feature extraction on the target modal component to obtain a feature vector, wherein the target electrical signal and the feature vector have a corresponding relationship; The identification module is used to analyze and obtain a DC arc identification result based on the characteristic vector and the DC arc identification model, wherein the DC arc identification model is a pre-trained model, and the DC arc identification result is used to characterize the occurrence of the DC arc.

9. A computer-readable storage medium, characterized in that: A direct current arc identification program for a photovoltaic system is stored, including execution instructions. When a processor of an electronic device executes the execution instructions, the processor executes the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: It comprises a processor and a memory storing execution instructions, wherein the memory stores one or more programs; when the processor executes the execution instructions stored in the memory, the processor executes the method according to any one of claims 1 to 7.

Images