Photovoltaic DC arc detection method, device, medium and equipment

Abstract

The present invention discloses a photovoltaic DC arc detection method, apparatus, medium, and equipment, relating to the technical field of fault diagnosis in new energy power generation systems. The photovoltaic DC arc detection method includes: obtaining M first eigenvectors based on electrical characteristic quantities corresponding to M photovoltaic inverters extracted from a photovoltaic inverter data transmission network; determining a first reference vector associated with the M photovoltaic inverters based on the M first eigenvectors; calculating the difference between the M first eigenvectors and the first reference vector; comparing the difference with a first threshold to obtain M comparison results associated with the M photovoltaic inverters; and, based on the M comparison results, first determining whether any of the M photovoltaic inverters generates a DC arc. The present invention, through this solution, can improve the accuracy of determining whether a photovoltaic inverter generates a DC arc.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis of new energy power generation systems, and in particular to a photovoltaic direct current arc detection method, device, medium and equipment. Background Art

[0002] At present, solar photovoltaic systems have been widely used due to their advantages of no pollution to the environment, high efficiency of energy conversion and high space utilization.

[0003] However, over the long term, PV system wiring may age due to environmental weathering, potentially leading to DC arcing. Once a DC arc occurs, it is often difficult to extinguish naturally. The temperature generated by the DC arc is high enough to ignite surrounding materials, potentially causing PV system failures and significant economic losses. Therefore, detecting DC arcs in PV systems is crucial for preventing fires and protecting personnel and equipment. It is a key measure for improving the overall reliability and safety of PV systems.

[0004] Existing methods for photovoltaic DC arc detection fall into two main categories: one involves detecting physical phenomena such as light and heat generated by a DC arc. However, this method is susceptible to interference from the external environment, and the difficulty in predicting the location of the fault arc leads to inaccurate sensor installation and poor arc detection. The other method extracts the time-domain and frequency-domain characteristics of the current and voltage in the photovoltaic system and uses artificial intelligence algorithms to determine the occurrence of a DC arc. However, due to significant differences in fault signatures at different voltages and the influence of noise, this method often results in misjudgments and missed detections.

[0005] In summary, the existing technology has poor arc detection effect when detecting photovoltaic DC arcs. Summary of the Invention

[0006] Based on this, it is necessary to provide a photovoltaic DC arc detection method, device, medium and equipment to address the above technical problems.

[0007] The present invention adopts the following technical solutions:

[0008] In a first aspect, the present invention provides a photovoltaic DC arc detection method, the method comprising:

[0009] Based on electrical characteristic quantities corresponding to M photovoltaic inverters extracted from a photovoltaic inverter data transmission network, M first characteristic vectors are obtained;

[0010] Determining a first reference vector associated with the M photovoltaic inverters based on the M first eigenvectors;

[0011] Calculating the difference between the M first eigenvectors and the first reference vector;

[0012] Comparing the difference with a first threshold value to obtain M comparison results associated with the M photovoltaic inverters;

[0013] Based on the M comparison results, it is first determined whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc.

[0014] Furthermore, the method of obtaining M first feature vectors based on the electrical feature quantities corresponding to the M photovoltaic inverters extracted from the photovoltaic inverter data transmission network specifically includes:

[0015] Obtaining data types of electrical characteristic quantities corresponding to the M photovoltaic inverters;

[0016] Classifying the electrical characteristic quantities corresponding to the M photovoltaic inverters according to the data type to obtain M groups of standard electrical characteristic quantities;

[0017] The M groups of standard electrical characteristic quantities are respectively implanted into arrays according to the data types to obtain the M first characteristic vectors.

[0018] Furthermore, before obtaining the data types of the electrical characteristic quantities corresponding to the M photovoltaic inverters, the method further includes:

[0019] Collecting electrical characteristic quantities corresponding to the M photovoltaic inverters through a sliding window to obtain electrical characteristic quantities corresponding to the M photovoltaic inverters;

[0020] Based on the number of photovoltaic modules in the photovoltaic array connected to the M photovoltaic inverters and the series or parallel connection mode, the electrical characteristic quantities corresponding to the M photovoltaic inverters are standardized.

[0021] Furthermore, determining a first reference vector associated with the M photovoltaic inverters based on the M first eigenvectors specifically includes:

[0022] Determine any two first eigenvectors among the M first eigenvectors as initial cluster centroids, and determine the remaining first eigenvectors as first eigenvectors to be clustered;

[0023] Calculating the distance between each of the first eigenvectors to be clustered and the initial cluster centroid;

[0024] assigning the first eigenvector to be clustered to the initial cluster centroid according to the distance;

[0025] Calculating an average value of the initial cluster centroids, and updating the initial cluster centroids according to the average value of the initial cluster centroids;

[0026] When the number of times the initial cluster centroids are updated reaches a preset number, the dividing plane between the cluster centroids is determined as the first reference vector.

[0027] Furthermore, the first determining, based on the M comparison results, whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc specifically includes:

[0028] When a difference value corresponding to any one or more comparison results among the M comparison results is greater than the first threshold, determining for the first time that the photovoltaic inverter associated with any one or more comparison results has generated a DC arc;

[0029] When the difference corresponding to any one or more comparison results among the M comparison results is not greater than the first threshold, the photovoltaic inverter associated with any one or more comparison results is determined as not generating a DC arc.

[0030] Furthermore, after determining for the first time whether any photovoltaic inverter generates a DC arc among the M photovoltaic inverters based on the M comparison results, the method further includes:

[0031] issuing a first alarm signal through the photovoltaic inverter that is first determined to have generated a DC arc;

[0032] After the time when the photovoltaic inverter that is first determined to generate a DC arc sends out a first alarm signal reaches a preset time, adjusting the power parameters of the photovoltaic inverter that is first determined to generate a DC arc;

[0033] Calculating, based on the power parameter, a second eigenvector associated with the photovoltaic inverter that is first determined to generate a DC arc;

[0034] Based on the second eigenvector, calculating a second reference vector associated with the second eigenvector;

[0035] Based on the second reference vector and a preset threshold, it is determined whether the photovoltaic inverter that is first determined to have generated a DC arc has generated a DC arc.

[0036] Further, after determining whether the photovoltaic inverter that is first determined to generate a DC arc generates a DC arc based on the second reference vector and a preset threshold, the method further includes:

[0037] In the case where the photovoltaic inverter that is first determined to have generated a DC arc generates a DC arc, a second alarm signal is issued and an arc extinguishing operation is performed;

[0038] In the case that the photovoltaic inverter that is first determined to have generated a DC arc does not generate a DC arc, a misjudgment signal is issued.

[0039] In a second aspect, the present invention provides a photovoltaic DC arc detection device, comprising:

[0040] A first determining module is configured to obtain M first feature vectors based on electrical feature quantities corresponding to M photovoltaic inverters extracted from a photovoltaic inverter data transmission network;

[0041] A second determining module is configured to determine, based on the M first eigenvectors, a first reference vector associated with the M photovoltaic inverters;

[0042] A calculation module, configured to calculate the difference between the M first eigenvectors and the first reference vector;

[0043] a comparison module, configured to compare the difference with a first threshold value to obtain M comparison results associated with the M photovoltaic inverters;

[0044] The judgment module is used to: determine for the first time whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc based on the M comparison results.

[0045] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the photovoltaic DC arc detection method is implemented.

[0046] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the photovoltaic DC arc detection method is implemented when the processor executes the program.

[0047] The at least one technical solution adopted by the present invention can achieve the following beneficial effects:

[0048] The present invention first extracts electrical characteristics corresponding to M photovoltaic inverters from a photovoltaic inverter data transmission network to obtain M first eigenvectors. Based on the M first eigenvectors, a first reference vector associated with the M photovoltaic inverters is determined, and the difference between the M first eigenvectors and the first reference vector is calculated. The difference is then compared with a first threshold to obtain M comparison results associated with the M photovoltaic inverters. Finally, based on the M comparison results, a first determination is made as to whether any of the M photovoltaic inverters has generated a DC arc.

[0049] The present invention detects DC arcs based on the collected electrical characteristic quantities of each photovoltaic inverter, reducing the error of the existing technology in DC arc detection by deploying light sensors and heat sensors in the photovoltaic inverter data transmission network to detect physical phenomena such as light and heat generated by the DC arc, and avoiding the influence of the external environment on the detection of whether the photovoltaic inverter generates a DC arc.

[0050] Due to significant differences in the electrical characteristics of photovoltaic inverters under different operating environments, DC arc detection often results in misjudgments and omissions. The present invention detects DC arcs based on the electrical characteristics of multiple photovoltaic inverters collected under the same operating environment. This avoids errors in the electrical characteristics collected from photovoltaic inverters under different operating environments and improves the accuracy of determining whether DC arcs are generated by multiple photovoltaic inverters under the same operating environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0052] Figure 1 A schematic flow chart of a photovoltaic DC arc detection method provided by the present invention;

[0053] Figure 2 A schematic diagram of a photovoltaic DC arc detection device provided by the present invention;

[0054] Figure 3 A schematic diagram of a computer device for implementing a photovoltaic DC arc detection method provided by the present invention. DETAILED DESCRIPTION

[0055] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0056] The server mentioned in the present invention can be a server installed on a business platform, or a device such as a desktop computer or laptop computer capable of executing the solution of the present invention. For ease of explanation, the following description will only use the server as the execution entity. The following, combined with the accompanying drawings, details the technical solutions provided by various embodiments of the present invention.

[0057] Figure 1The following is a flow chart of a photovoltaic DC arc detection method according to the present invention, which specifically includes the following steps:

[0058] S10: Obtain M first feature vectors based on electrical feature quantities corresponding to M photovoltaic inverters extracted from the photovoltaic inverter data transmission network.

[0059] In this embodiment, the electrical characteristics of the M photovoltaic inverters can be extracted through a pre-established photovoltaic inverter data transmission network. The photovoltaic inverter data network can be established by, but is not limited to, utilizing the inverter's own communication functions or by using an external communication module. The communication methods used include, but are not limited to, optical fiber, 4G, Wireless Fidelity (Wi-Fi), and ZigBee (ZigBee technology is a short-range, low-power wireless communication technology).

[0060] S20: Determine a first reference vector associated with the M photovoltaic inverters based on the M first eigenvectors.

[0061] In this embodiment, the method for determining the first reference vector includes but is not limited to: calculating the average value of the first feature vectors of the M photovoltaic inverters in the first n frames at the time of arc detection, and using the calculation result as the first reference vector.

[0062] S30: Calculate the difference between the M first eigenvectors and the first reference vector.

[0063] In this embodiment, the differences between the M first eigenvectors and the first reference vector are calculated by the server.

[0064] S40: Compare the difference with a first threshold to obtain M comparison results associated with the M photovoltaic inverters.

[0065] In this embodiment, the server is used to compare the difference with the first threshold to obtain M comparison results associated with the M photovoltaic inverters.

[0066] Optionally, a method for calculating the first threshold includes but is not limited to: taking an average value X% of the M difference vectors as the first threshold, and a specific value of X can be determined based on experience, experiments, and the like.

[0067] S50: Based on the M comparison results, it is determined for the first time whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc.

[0068] In this embodiment, the first determination of whether the M photovoltaic inverters generate a DC arc is performed as follows: if the difference exceeds the first threshold, it is determined that the corresponding photovoltaic inverter is suspected of generating a DC arc.

[0069] based on Figure 1 A photovoltaic DC arc detection method is shown. First, electrical characteristic quantities corresponding to M photovoltaic inverters are extracted from a photovoltaic inverter data transmission network to obtain M first characteristic vectors. Based on these M first characteristic vectors, a first reference vector associated with the M photovoltaic inverters is determined, and the difference between the M first characteristic vectors and the first reference vector is calculated. This difference is then compared with a first threshold to obtain M comparison results associated with the M photovoltaic inverters. Finally, based on these M comparison results, a first determination is made as to whether any of the M photovoltaic inverters has generated a DC arc. By quantizing the electrical characteristic quantities, the present invention avoids the influence of the environment on DC arc detection and improves the accuracy of determining whether a photovoltaic inverter has generated a DC arc.

[0070] When applying the photovoltaic DC arc detection method provided by the present invention, it is not necessary to Figure 1 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0071] In addition, in one or more embodiments of the present invention, obtaining M first feature vectors based on the extracted electrical feature quantities of the M photovoltaic inverters specifically includes:

[0072] S101: Obtain data types of electrical characteristic quantities corresponding to the M photovoltaic inverters.

[0073] Optionally, before obtaining the data types of the electrical characteristic quantities corresponding to the M photovoltaic inverters, the electrical characteristic quantities corresponding to the M photovoltaic inverters are standardized. Methods for standardizing the electrical characteristic quantities corresponding to the M photovoltaic inverters include, but are not limited to, standardizing the electrical characteristic quantities by comprehensively considering the number of photovoltaic modules in the photovoltaic array connected to the inverter and the series and parallel configuration. Based on this method of standardizing the electrical characteristic quantities corresponding to the M photovoltaic inverters, the effects of different dimensions and combinations of photovoltaic modules on the electrical characteristic quantities can be eliminated.

[0074] S102: Classifying the electrical characteristic quantities corresponding to the M photovoltaic inverters according to the data types to obtain M groups of standard electrical characteristic quantities.

[0075] In this embodiment, the data types include but are not limited to: current variance value, current spectrum energy and energy entropy, and current wavelet energy entropy. Each set of electrical feature quantities includes the above data types.

[0076] S103: respectively embed the M groups of standard electrical characteristic quantities into an array according to the data type to obtain the M first characteristic vectors.

[0077] In this embodiment, the current variance value, current spectrum energy and energy entropy, and current wavelet energy entropy are put into an array in order as feature quantities, as the first feature vector in, represents the first eigenvector of the inverter numbered n at time t, t 1n_1 Represents the first component of the first eigenvector of inverter numbered n.

[0078] This embodiment standardizes the electrical characteristic quantities corresponding to the photovoltaic inverter, classifies the electrical characteristic quantities according to different data types, and then sequentially implants the classified electrical characteristic quantities into an array to obtain a first characteristic vector, thereby reducing the calculation error caused by multiple data types.

[0079] In addition, in one or more embodiments of the present invention, before obtaining the data types of the electrical characteristic quantities corresponding to the M photovoltaic inverters, the method further includes:

[0080] S1011: Collect electrical characteristic quantities corresponding to the M photovoltaic inverters through a sliding window to obtain electrical characteristic quantities corresponding to the M photovoltaic inverters:.

[0081] In this embodiment, a possible solution is to use a 10ms time sliding window to collect the electrical characteristics of the photovoltaic inverter. In this way, one frame of electrical quantity data can be obtained every 10ms. The time sliding window can also be changed to other time lengths according to actual needs. In this example, a sliding window is used to collect the current of the photovoltaic inverter. When the sliding window size is set to 10ms and the sampling frequency is 200kHz, each sampling group has a total of 2000 data points, which are recorded as {I n}, n is from 1 to 2000.

[0082] S1012: Based on the number of photovoltaic modules in the photovoltaic array connected to the M photovoltaic inverters and the series or parallel connection mode, perform standardization processing on the electrical characteristic quantities corresponding to the M photovoltaic inverters.

[0083] In this embodiment, electrical characteristics include, but are not limited to, time-domain, frequency-domain, and time-frequency-domain characteristics of the DC-side current and voltage of the photovoltaic inverter. Time-domain characteristics include, but are not limited to, effective values, peak values, and mean values. Frequency-domain characteristics include, but are not limited to, spectrum energy and spectrum energy entropy. Time-frequency characteristics include, but are not limited to, wavelet coefficients and wavelet entropy.

[0084] This embodiment collects data using a sliding window approach, allowing for different sliding window widths to be set as needed, thereby increasing the richness of the sampled data. Furthermore, this embodiment takes into account the number of PV modules in the PV array and their series or parallel configuration when normalizing electrical characteristic quantities, thus eliminating the impact of different dimensions and PV module combinations on the electrical characteristic quantities.

[0085] Furthermore, in one or more embodiments of the present invention, determining, based on the M first eigenvectors, first reference vectors associated with the M photovoltaic inverters specifically includes:

[0086] S201: Determine any two first eigenvectors among the M first eigenvectors as initial cluster centroids, and determine the remaining first eigenvectors as first eigenvectors to be clustered.

[0087] S202: Calculate the distance between each of the first eigenvectors to be clustered and the initial cluster centroid.

[0088] S203: Allocate the first feature vector to be clustered to the initial cluster centroid according to the distance.

[0089] S204: Calculate the average value of the initial cluster centroids, and update the initial cluster centroids according to the average value of the initial cluster centroids.

[0090] S205: When the number of times the initial cluster centroids are updated reaches a preset number, a dividing plane between the cluster centroids is determined as the first reference vector.

[0091] In this embodiment, a K-clustering algorithm is used based on multi-machine data. First, the number of cluster centers is determined, which in this method is set to 2. Using M first eigenvectors as input data, two first eigenvectors are randomly selected as initial cluster centroids. The distance between each data point and each cluster centroid is calculated, and the data point is assigned to the closest cluster. The average value of the data points in each cluster is calculated, and the cluster centroid is updated. This process is repeated until the maximum number of iterations is reached. The segmentation planes between different clusters are used as the first reference vectors.

[0092] Optionally, in this embodiment, another method for calculating the first reference vector is to calculate the average value of the first eigenvectors of the first M frames at the arc detection moment of M photovoltaic inverters, and use it as the first reference vector, where M≥1.

[0093] In addition, in one or more embodiments of the present invention, determining whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc specifically includes:

[0094] S501: When a difference corresponding to any one or more comparison results among the M comparison results is greater than a first threshold, determining for the first time that a photovoltaic inverter associated with the one or more comparison results has generated a DC arc.

[0095] In this embodiment, the first determination that a DC arc is generated refers to determination that a DC arc is suspected to be generated.

[0096] S502: If a difference value corresponding to any one or more comparison results among the M comparison results is not greater than the first threshold, determine that the photovoltaic inverter associated with any one or more comparison results does not generate a DC arc.

[0097] In this embodiment, the calculation method of the first threshold includes but is not limited to: taking the average value X% of the M difference vectors as the first threshold, and the specific value of X can be determined based on experience, experiments, etc.

[0098] Alternatively, another method for determining whether a PV inverter generates a DC arc is as follows:

[0099] From the M photovoltaic inverters, N photovoltaic inverters that need to be compared with the first eigenvector are determined to obtain an inverter group, where N is not less than two.

[0100] In this embodiment, a possible photovoltaic inverter grouping solution is to group photovoltaic inverters according to the similarity of photovoltaic array operating environments, and group N photovoltaic inverters with relatively similar operating environments into one group to obtain the inverter group.

[0101] In the inverter group, each photovoltaic inverter multicasts its associated first eigenvector to other photovoltaic inverters, so that any photovoltaic inverter in the inverter group obtains N first eigenvectors associated with different photovoltaic inverters.

[0102] In this embodiment, each photovoltaic inverter need not copy multiple first eigenvectors to send; instead, it only needs to send one associated first eigenvector. Multicast network devices (network devices running a multicast routing protocol) forward or copy multicast data as needed, enabling on-demand copying and sending. Furthermore, the first eigenvector is only sent to the photovoltaic inverters that have joined the inverter group; photovoltaic inverters that do not require it will not receive it. This solution improves network resource utilization and the efficiency with which each photovoltaic inverter receives the first eigenvectors of other photovoltaic inverters.

[0103] Each photovoltaic inverter calculates a difference between the N first eigenvectors associated with different photovoltaic inverters and the first reference vector.

[0104] In this embodiment, each inverter in the group calculates the difference between the N first eigenvectors it receives and the first reference vector. If the difference between the first eigenvector and the first reference vector of a certain inverter exceeds the first threshold, it is determined that the inverter is suspected of generating a DC arc, and the corresponding inverter sends a first alarm signal.

[0105] One possible solution is to calculate the difference between each of the N first eigenvectors and the first reference vector to obtain N difference vectors. X% of the average of the N difference vectors is used as a first threshold. If the difference exceeds the first threshold, it is determined that a DC arc has occurred in the corresponding inverter, and the inverter issues a first alarm signal. The specific value of X can be determined based on experience, experimentation, and other methods.

[0106] For example, in this example, N is 3, and the three first eigenvectors and the first reference vector at time t are calculated. The difference between , get 3 difference vectors, namely 120% of the average value of the three difference vectors is used as the first threshold. When the difference exceeds 120% of the average value of the three difference vectors, it is determined that the corresponding inverter is suspected of generating a DC arc, and the corresponding inverter sends a first alarm signal.

[0107] It is determined whether an arc occurs in the photovoltaic inverter according to the difference.

[0108] In this embodiment, each photovoltaic inverter in the group calculates the difference between the N first eigenvectors it receives and the first reference vector. If the difference between the first eigenvector and the first reference vector of a photovoltaic inverter exceeds a first threshold, it is determined that the photovoltaic inverter is suspected of generating a DC arc.

[0109] Through the above-mentioned solution, the present invention can narrow the scope of arc detection on the photovoltaic inverter and improve the efficiency of detecting whether the photovoltaic inverter generates an arc.

[0110] Furthermore, in one or more embodiments of the present invention, the first determining, based on the M comparison results, whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc further includes:

[0111] S601: Sending a first alarm signal via the photovoltaic inverter that is first determined to have generated a DC arc.

[0112] In this embodiment, the types of the alarm signal include but are not limited to: sound signals and optical signals.

[0113] S602: After the time when the photovoltaic inverter that is first determined to generate a DC arc sends out a first alarm signal reaches a preset time, adjust the power parameters of the photovoltaic inverter that is first determined to generate a DC arc.

[0114] S603: Calculate, based on the power parameter, a second eigenvector associated with the photovoltaic inverter that is first determined to generate a DC arc.

[0115] S604: Based on the second eigenvector and a preset threshold, determine whether the photovoltaic inverter that is first determined to generate a DC arc generates a DC arc.

[0116] In this embodiment, Solution 1: At time t, upon receipt of the first alarm signal, the DC voltage of the photovoltaic inverter first determined to have generated a DC arc is adjusted to cause the DC voltage to change by ΔV. The photovoltaic inverter stores the first eigenvectors of the frame before and after time t and records them as the second eigenvector. Regarding the selection of the first eigenvector, one possible solution is to select the variance of the voltage and current sampling values ​​as the first eigenvector and include it in the second eigenvector. Based on the change in the voltage and current variance before and after the DC arc occurs, a second threshold is appropriately set to determine whether the photovoltaic inverter first determined to have generated a DC arc has generated a DC arc. This second threshold can be determined based on the voltage and current variance when a DC arc occurs in the photovoltaic system, or based on a second threshold value received or stored by the photovoltaic inverter.

[0117] For example, in this example, for the current variance value, 120% of the current variance value of the frame before the DC voltage adjustment moment can be selected as its second threshold; for the current spectrum energy and spectrum energy entropy, 130% of the average value of the current spectrum energy entropy of the previous two frames can be selected as its second threshold; for the wavelet energy entropy, 120% of the current spectrum energy entropy of the previous five frames can be selected as its second threshold.

[0118] The current variance calculation formula is as follows:

[0119]

[0120] Among them, σ 2 is the current variance value, n is the total number of current sampling points, I i is each current value sampled, and μ is the average value of the sampled current values.

[0121] Solution 2: At time t, upon receiving the first alarm signal, the DC voltage of the photovoltaic inverter first identified as generating a DC arc is adjusted, resulting in a DC voltage change of ΔV. The photovoltaic inverter stores the first eigenvectors of m frames before and p frames after time t, and records them as second eigenvectors, where m is not less than 2 and p is not less than 1. Regarding the selection of the first eigenvector, one possible solution is to select the current spectral energy entropy as the first eigenvector and place it within the second eigenvector. The spectral energy entropy of the current m frames before and m frames after the DC voltage adjustment is calculated. Based on the change in the current spectral energy entropy before and after the DC arc occurs, a second threshold is appropriately set to determine whether the photovoltaic inverter first identified as generating a DC arc has generated a DC arc. This second threshold can be determined based on the spectral energy entropy of the current when the DC arc occurs in the photovoltaic system, or based on a second threshold value received or stored by the photovoltaic inverter.

[0122] In this embodiment, the current variance value, current spectrum energy and energy entropy, and current wavelet energy entropy are put into an array in order as feature quantities, as the first feature vector in, represents the first eigenvector of the inverter numbered n at time t, t 1n_1 Represents the first component of the first eigenvector of inverter numbered n.

[0123] For example, in this example, the first feature vector of the ten frames before the arc detection time t is taken Calculate the average value of the first eigenvector of the first ten frames of the inverter in the group in, The average as the first reference vector.

[0124] Solution 3: At the moment t when the first alarm signal is received, the DC voltage of the photovoltaic inverter that is first determined to have a DC arc is adjusted multiple times, so that it changes in sequence by ΔV1, ΔV2, ..., ΔV according to the time window shift. n , that is, each time window changes ΔV1, ΔV2, ..., ΔV n, where n is the number of DC voltage adjustments, and n is no less than 2. The photovoltaic inverter stores the first eigenvectors of the previous and subsequent frames for each DC voltage adjustment and records them as second eigenvectors, for a total of n+1 second eigenvectors. Regarding the selection of the first eigenvector, one possible approach is to select wavelet energy entropy as the first eigenvector and include it in the second eigenvector. Based on the differences in the frequency spectrum during different stages of the DC arc (before the DC arc occurs, during the DC arc occurs, and during the DC arc burn), a second threshold is appropriately set to determine whether the photovoltaic inverter initially determined to have generated a DC arc has generated a DC arc. This second threshold can be determined based on the frequency spectrum of the DC arc at different stages in the photovoltaic system, or based on a second threshold received or stored by the photovoltaic inverter. Furthermore, different second threshold components are set for different second eigenvectors. The second threshold components are arranged in an array in the order of the corresponding second eigenvectors to form the second threshold. In this example, based on the number of photovoltaic modules in the photovoltaic array connected to the inverter and their series and parallel configuration, the electrical quantity is first converted to the electrical quantity of the photovoltaic array corresponding to a specific number and series and parallel configuration. Then perform Fourier transform and calculate the spectrum energy and spectrum energy entropy. Put the standardized feature quantities into the array in a certain order to obtain the first feature vector. Among them, the spectrum energy Ef is the current value I n Obtained by Fourier transform.

[0125] to {I n The spectrum amplitude after Fourier transform is recorded as {F k}, k is a positive integer and k Not more than 2000, where the spectrum energy E f The calculation formula is as follows:

[0126]

[0127] Where k2 is the upper frequency of the frequency band The corresponding frequency band number, K1 is the lower limit frequency F of the frequency band k 1 Corresponding frequency band number, F k is the amplitude of the kth point of the spectrum of the sampling group.

[0128] If the sampling frequency is 200kHz, the corresponding signal frequency upper limit is 100kHz. The frequency bands can be set to 1Hz-10kHz, 10kHz-20kHz, 20kHz-30kHz, 30kHz-40kHz, 40kHz-50kHz, 50kHz-60kHz, 60kHz-70kHz, 70kHz-80kHz, 80kHz-90kHz, and 90kHz-100kHz, respectively, and the spectrum energy of the 10 frequency bands is used as 10 feature information.

[0129] Spectral energy entropy H f The definition of is shown in formula (3):

[0130]

[0131] Among them, p fk The calculation method is shown in formula (4):

[0132]

[0133] Where n is the total number of spectrum points, P fk is the proportion of the kth point in the total energy of the spectrum, E fsum The total energy of a sampling group between 1Hz and 100kHz. Spectral energy entropy can reflect the complexity and randomness of a signal. The greater the spectral energy entropy, the more variations the signal spectrum contains.

[0134] The wavelet energy entropy calculation method is as follows:

[0135] Current sampling signal I n After wavelet decomposition, the coefficient of the j-th subband is c j (k),j=1,2,…,N;k=1,2,…,K。 N is the number of subbands, K is the number of coefficients, then the energy of the j-th subband is as shown in formula (5):

[0136]

[0137] Perform m-layer tree wavelet decomposition on the current signal, let the vector E=(E1,E2,…,E 2m ) is 2 m The energy of the sub-band is E sum is the energy E of each sub-band j The sum of p j =E j / E sum , then the calculation method of wavelet energy entropy H is as shown in formula (6):

[0138]

[0139] Solution 4: At time t, upon receiving the first alarm signal, the DC voltage of all PV inverters in the inverter group is adjusted by a change of ΔV. Each PV inverter stores the first eigenvector of the frame following the DC voltage adjustment at time t and records it as the second eigenvector. Communication technology is used to share the second eigenvector data within the group. By comparing the second eigenvectors received by the PV inverters, it is determined whether the PV inverter first identified as generating a DC arc has generated a DC arc. Specifically, one possible solution is to calculate the average value of the second eigenvectors of the PV inverters in the inverter group one frame after the DC voltage adjustment, use this as the second reference vector, calculate the difference between each second eigenvector and the second reference vector, and determine that the PV inverter first identified as generating a DC arc that exceeds a second threshold is the inverter generating the DC arc. The second threshold can be determined based on the second threshold received or stored by the PV inverter.

[0140] Solution 5: At time t when the first alarm signal is received, the DC voltage of all PV inverters in the group is adjusted multiple times so that it changes in sequence by ΔV1, ΔV2, …, ΔV according to the time window shift. n , where n is the number of times the DC voltage is adjusted, and n is not less than 2. The photovoltaic inverter stores the first eigenvector of one frame after each DC voltage adjustment and records it as the second eigenvector. Communication technology is used to realize the sharing of the second eigenvector data within the inverter group. By comparing the second eigenvectors received by the photovoltaic inverter, it is realized to determine whether the photovoltaic inverter that was first determined to have generated a DC arc has generated a DC arc. Specifically, a possible solution is to calculate the average value of the second eigenvector of one frame after each DC voltage adjustment of the photovoltaic inverter in the inverter group, use it as the second reference vector, calculate the difference between each second eigenvector and the second reference vector, and determine the photovoltaic inverter that exceeds the second threshold as an inverter that has generated a DC arc. The second threshold here can be determined based on the second threshold obtained by the photovoltaic inverter through reception or storage.

[0141] Through the above-mentioned embodiments, the present invention compares the first eigenvectors corresponding to different photovoltaic inverters and then compares the data of the photovoltaic inverters before and after the DC voltage adjustment, thereby improving the accuracy of detecting whether the photovoltaic inverter generates a DC arc and reducing the misjudgment rate.

[0142] Furthermore, in one or more embodiments of the present invention, after determining, based on the second reference vector and a preset threshold, whether the photovoltaic inverter that is first determined to have generated a DC arc has generated a DC arc, the method further includes:

[0143] S6041: In the case where the photovoltaic inverter that is first determined to have generated a DC arc generates a DC arc, a second alarm signal is issued and an arc extinguishing operation is performed.

[0144] In this embodiment, the second alarm signal refers to an alarm signal for confirming the photovoltaic inverter that is first determined to have generated a DC arc as the photovoltaic inverter that has generated a DC arc.

[0145] S6042: When the photovoltaic inverter that is first determined to generate a DC arc does not generate a DC arc, a misjudgment signal is issued.

[0146] In this embodiment, the misjudgment signal refers to a signal that confirms that the photovoltaic inverter that is initially determined to generate a DC arc is a photovoltaic inverter that does not generate a DC arc.

[0147] Through the above technical solution, this embodiment can further confirm that a photovoltaic inverter suspected of generating a DC arc is a photovoltaic inverter that generates a DC arc or does not generate a DC arc, thereby improving the accuracy of DC arc detection on the photovoltaic inverter.

[0148] The above is a photovoltaic DC arc detection method provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding photovoltaic DC arc detection device, such as Figure 2 shown.

[0149] Figure 2 A schematic diagram of a photovoltaic DC arc detection device provided by the present invention includes:

[0150] A first determining module is configured to obtain M first feature vectors based on electrical feature quantities corresponding to M photovoltaic inverters extracted from a photovoltaic inverter data transmission network;

[0151] A second determining module is configured to determine, based on the M first eigenvectors, a first reference vector associated with the M photovoltaic inverters;

[0152] A calculation module, configured to calculate the difference between the M first eigenvectors and the first reference vector;

[0153] a comparison module, configured to compare the difference with a first threshold value to obtain M comparison results associated with the M photovoltaic inverters;

[0154] The judgment module is used to: determine for the first time whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc based on the M comparison results.

[0155] The specific definitions of a photovoltaic DC arc detection device can be found in the definitions of a photovoltaic DC arc detection method described above and will not be repeated here. Each module in the photovoltaic DC arc detection device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device memory in software form, allowing the processor to call and execute the corresponding operations of each module.

[0156] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the Figure 1 A photovoltaic DC arc detection method is provided.

[0157] The present invention also provides Figure 3 The structural diagram of the computer equipment shown in FIG. Figure 3 As shown, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the Figure 1 A photovoltaic DC arc detection method is provided.

[0158] Those skilled in the art will appreciate that all or part of the processes in the embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the various methods described. Among them, any reference to memory, storage, database or other media used in the various embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0159] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A photovoltaic DC arc detection method, characterized in that: include: Based on electrical characteristic quantities corresponding to M photovoltaic inverters extracted from a photovoltaic inverter data transmission network, M first characteristic vectors are obtained; Determining a first reference vector associated with the M photovoltaic inverters based on the M first eigenvectors; Calculating the difference between the M first eigenvectors and the first reference vector; Comparing the difference with a first threshold value to obtain M comparison results associated with the M photovoltaic inverters; determining, based on the M comparison results, whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc; The determining, based on the M first eigenvectors, of first reference vectors associated with the M photovoltaic inverters specifically includes: Determine any two first eigenvectors among the M first eigenvectors as initial cluster centroids, and determine the remaining first eigenvectors as first eigenvectors to be clustered; Calculating the distance between each of the first eigenvectors to be clustered and the initial cluster centroid; assigning the first eigenvector to be clustered to the initial cluster centroid according to the distance; Calculating an average value of the initial cluster centroids, and updating the initial cluster centroids according to the average value of the initial cluster centroids; When the number of times the initial cluster centroids are updated reaches a preset number, the dividing plane between the cluster centroids is determined as the first reference vector.

2. The photovoltaic DC arc detection method according to claim 1, wherein: The method of obtaining M first feature vectors based on electrical feature quantities corresponding to M photovoltaic inverters extracted from the photovoltaic inverter data transmission network specifically includes: Obtaining data types of electrical characteristic quantities corresponding to the M photovoltaic inverters; Classifying the electrical characteristic quantities corresponding to the M photovoltaic inverters according to the data type to obtain M groups of standard electrical characteristic quantities; The M groups of standard electrical characteristic quantities are respectively implanted into arrays according to the data types to obtain the M first characteristic vectors.

3. The photovoltaic DC arc detection method according to claim 2, wherein: Before obtaining the data types of the electrical characteristic quantities corresponding to the M photovoltaic inverters, the method further includes: Collecting electrical characteristic quantities corresponding to the M photovoltaic inverters through a sliding window to obtain electrical characteristic quantities corresponding to the M photovoltaic inverters; Based on the number of photovoltaic modules in the photovoltaic array connected to the M photovoltaic inverters and the series or parallel connection mode, the electrical characteristic quantities corresponding to the M photovoltaic inverters are standardized.

4. The photovoltaic DC arc detection method according to claim 1, wherein: The first step of determining, based on the M comparison results, whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc specifically includes: When a difference value corresponding to any one or more comparison results among the M comparison results is greater than the first threshold, determining for the first time that the photovoltaic inverter associated with any one or more comparison results has generated a DC arc; When the difference corresponding to any one or more comparison results among the M comparison results is not greater than the first threshold, the photovoltaic inverter associated with any one or more comparison results is determined as not generating a DC arc.

5. The photovoltaic DC arc detection method according to claim 1, wherein: After determining for the first time whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc based on the M comparison results, the method further includes: When there is a photovoltaic inverter among the M photovoltaic inverters that is determined to have generated a DC arc for the first time, issuing a first alarm signal through the photovoltaic inverter that is determined to have generated a DC arc for the first time; After the time when the photovoltaic inverter that is first determined to generate a DC arc sends out a first alarm signal reaches a preset time, adjusting the power parameters of the photovoltaic inverter that is first determined to generate a DC arc; Calculating, based on the power parameter, a second eigenvector associated with the photovoltaic inverter that is first determined to generate a DC arc; Based on the second eigenvector, calculating a second reference vector associated with the second eigenvector; Whether the photovoltaic inverter that is first determined to have generated a DC arc generates a DC arc is determined based on the second reference vector and a preset threshold.

6. The photovoltaic DC arc detection method according to claim 5, wherein: After determining whether the photovoltaic inverter generating the DC arc generates the DC arc according to the second reference vector and the preset threshold, the method further includes: In the case where the photovoltaic inverter that is first determined to have generated a DC arc generates a DC arc, a second alarm signal is issued and an arc extinguishing operation is performed; In the case that the photovoltaic inverter that is first determined to have generated a DC arc does not generate a DC arc, a misjudgment signal is issued.

7. A photovoltaic DC arc detection device, characterized in that: include: A first determining module is configured to obtain M first feature vectors based on electrical feature quantities corresponding to M photovoltaic inverters extracted from a photovoltaic inverter data transmission network; A second determining module is configured to determine, based on the M first eigenvectors, a first reference vector associated with the M photovoltaic inverters; A calculation module, configured to calculate the difference between the M first eigenvectors and the first reference vector; a comparison module, configured to compare the difference with a first threshold value to obtain M comparison results associated with the M photovoltaic inverters; A judgment module is configured to: determine for the first time whether any photovoltaic inverter among the M photovoltaic inverters generates a DC arc based on the M comparison results; The second determining module is specifically configured to: Determine any two first eigenvectors among the M first eigenvectors as initial cluster centroids, and determine the remaining first eigenvectors as first eigenvectors to be clustered; Calculating the distance between each of the first eigenvectors to be clustered and the initial cluster centroid; assigning the first eigenvector to be clustered to the initial cluster centroid according to the distance; Calculating an average value of the initial cluster centroids, and updating the initial cluster centroids according to the average value of the initial cluster centroids; When the number of times the initial cluster centroids are updated reaches a preset number, the dividing plane between the cluster centroids is determined as the first reference vector.

8. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and when the computer program is executed by a processor, the photovoltaic DC arc detection method according to any one of claims 1 to 6 is implemented.

9. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the photovoltaic DC arc detection method according to any one of claims 1 to 6 is implemented.

Images