Photovoltaic arc discharge detection method and system based on time-frequency domain analysis and wavelet transform

Abstract

The invention discloses a photovoltaic arcing detection method and system based on time-frequency domain analysis and wavelet transform, and relates to the technical field of photovoltaic power generation system safety detection.The photovoltaic arcing detection method comprises the steps that photovoltaic panel current data are collected in real time, and a high-frequency signal of the current data is calculated and compared with an alarm threshold value; performing wavelet decomposition on the current data, and recording a wavelet coefficient; and through wavelet detail coefficient calculation and time-domain signal calculation and comparison, if the time-domain signal and the frequency-domain signal both continuously exceed the alarm threshold, it is regarded that an arc fault is generated. According to the method, on the basis of a traditional time domain analysis method, a wavelet transform frequency domain analysis method and an FFT frequency domain analysis method are added, high-frequency signals of fault current are effectively processed, precision is improved, the misoperation rate is reduced, and response is faster.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation system safety detection, and in particular to a photovoltaic arc detection method and system based on time-frequency domain analysis and wavelet transform. Background Art

[0002] In photovoltaic power generation systems, over time, due to factors such as aging wiring and animal bites, poor contact between panel chips, between chips and guide frames, and between wiring and junction boxes can easily lead to series arcing. These low-voltage DC arcs lack zero crossings and are difficult to extinguish on their own, making them prone to fires. Because series arcs introduce nonlinear loads into the original circuit, they reduce the circuit current. Consequently, traditional low-voltage circuit breakers and fuses are unable to effectively detect and protect against arc faults.

[0003] Research on series arc fault detection has been extensive in areas such as aircraft, electric vehicles, and DC low-voltage distribution cabinets. However, in the photovoltaic power generation sector, traditional DC arc detection technology is no longer sufficient, as arc faults are significantly affected by factors such as sunlight intensity, temperature fluctuations, and converter noise. When an arc fault occurs, the voltage across the arc suddenly increases, while the current across the arc suddenly decreases. Since the arc fault location cannot be determined, fault detection can only be performed on the current in the line. When an arc fault occurs, current amplitude detection in the time domain, combined with frequency domain analysis and filtering and wavelet transform processing of the high-frequency noise generated by the arc, can further improve detection accuracy. Combining time and frequency domain analysis methods to optimize the judgment criteria can improve the accuracy and reliability of photovoltaic arc fault detection. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the existing DC arc detection method has poor anti-interference ability and low reliability, and how to improve the detection accuracy.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform, comprising real-time collection of photovoltaic panel current data, calculation of the high-frequency signal of the current data and comparison with the alarm threshold; wavelet decomposition of the current data and recording of the wavelet coefficients; after wavelet detail coefficient calculation and time domain signal calculation and comparison, if the time domain signal and the frequency domain signal both continuously exceed the alarm threshold, it is considered that an arc fault has occurred; the frequency domain signal includes signal characteristics obtained by performing frequency domain analysis on the current data, reflecting the energy distribution of the current data at different frequencies.

[0007] As a preferred solution of the photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform described in the present invention, the high-frequency signal of the current data is calculated, including performing FFT transformation on the collected current data, calculating the energy spectrum of the processed current data, dividing the frequency band by step size and summing them, and generating a characteristic frequency band energy spectrum matrix.

[0008] As a preferred solution of the photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform described in the present invention, the FFT transformation includes converting the time domain signal into a frequency domain signal, decomposing the signal frequency components to reveal the spectrum characteristics, analyzing the arc current energy spectrum distribution, and determining the characteristic frequency band of the fault arc.

[0009] As a preferred solution of the photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform described in the present invention, the characteristic frequency band includes a specific frequency interval where the energy in the current signal is significantly concentrated when a fault arc occurs. The arc current energy spectral density is analyzed by FFT, and the frequency range with maximum energy is determined as the characteristic frequency band to identify the high-frequency noise characteristics of the fault arc.

[0010] As a preferred solution of the photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform described in the present invention, the alarm threshold includes key parameters for judging arc faults, which include time domain and frequency domain. In the time domain, the standard deviation of the high-frequency current signal is calculated and compared with the standard deviation of the current data under normal working conditions to set the time domain alarm threshold. In the frequency domain, the wavelet detail coefficient of the current data is analyzed by wavelet transform, and the wavelet detail coefficients of the fault arc and the normal current are compared to set the frequency domain alarm threshold. The time domain alarm threshold and the frequency domain alarm threshold are collectively referred to as the alarm threshold.

[0011] As a preferred solution of the photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform described in the present invention, the wavelet decomposition of the current data includes using the db4 wavelet basis function, calculating the approximate coefficients through a low-pass filter and the detail coefficients through a high-pass filter, performing multi-scale decomposition on the current data, and extracting the signal components of the characteristic frequency band of the current data.

[0012] As a preferred solution of the photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform described in the present invention, the wavelet detail coefficient calculation includes decomposing the current data into detail coefficients of different frequency bands through wavelet transform, selecting the corresponding wavelet detail coefficient as the target by matching the frequency range of the FFT characteristic band energy spectrum matrix, and establishing an association between the frequency domain signal and the number of wavelet decomposition layers.

[0013] As a preferred solution of the photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform described in the present invention, the domain signal calculation includes analyzing the fluctuation characteristics of the high-frequency component of the current waveform based on the collected current data, quantifying the discrete degree of the current data by using statistical methods, capturing abnormal fluctuations by comparing with the preset alarm threshold, and identifying the time domain characteristics of the fault arc.

[0014] As a preferred solution of the photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform described in the present invention, wherein: the time domain signal and the frequency domain signal both continuously appear to exceed the alarm threshold, including, based on dual comprehensive judgment of the time and frequency domains, calculating the standard deviation of the high-frequency signal in the time domain, and extracting the wavelet detail coefficient through wavelet transform in the frequency domain, when the standard deviation and the wavelet detail coefficient of the high-frequency signal both continuously appear to exceed the preset alarm threshold, an alarm action is triggered.

[0015] Another object of the present invention is to provide a photovoltaic arc detection system based on time-frequency domain analysis and wavelet transform, which can process current data by combining time domain analysis and frequency domain analysis to determine the characteristic value of the fault arc, thereby solving the problem of misjudgment and missed judgment caused by the lack of reliability of a single criterion in current photovoltaic arc detection technology.

[0016] As a preferred solution of the photovoltaic arc detection system based on time-frequency domain analysis and wavelet transform described in the present invention, it includes: a data acquisition module, a feature analysis module, and a fault judgment module. The data acquisition module is used to capture the arc current data of the photovoltaic system in real time, and provide high-precision original data for the time-frequency domain joint analysis of the fault arc. The sampling points are used to collect the current at the outlet end of the junction box at high speed, and fully capture the high-frequency noise characteristics when the arc occurs. The collected data is used for FFT transformation, and the energy spectrum of the arc current data is analyzed. The frequency interval with the largest energy is selected as the characteristic frequency band of the fault arc, and the wavelet detail coefficient adapted to the characteristic frequency band is determined to provide an input signal for the wavelet transform; the feature analysis module includes a time domain analysis module and a frequency domain analysis module. The time domain analysis module is used to extract the high-frequency time domain characteristics of the current signal and judge the time domain alarm threshold, so as to utilize The fault arc and current data under normal working conditions collected by the sampling equipment are used to calculate the standard deviation of the high-frequency time domain signal, which is compared with the standard deviation of the time domain signal of the current data under normal working conditions to determine the alarm threshold. The frequency domain analysis module is used to extract the high-frequency noise characteristics of the current signal and optimize the frequency domain judgment criteria of the fault arc. The adaptive db4 wavelet basis function and decomposition layer number are selected based on the characteristic frequency band obtained by FFT transformation, and the wavelet detail coefficient threshold is calculated. The time domain analysis and frequency domain analysis are repeated many times, and arc detection is performed based on the threshold of the comprehensive characteristic value; the fault judgment module is used to synchronously execute the time domain standard deviation calculation and the frequency domain wavelet detail coefficient extraction, and dynamically compare the calculation results with the preset alarm threshold. The time domain amplitude fluctuation and the high-frequency noise characteristics simultaneously trigger the alarm threshold to exceed the limit. When the time domain standard deviation and the corresponding frequency band detail coefficient exceed the limit in continuous monitoring, the arc fault is determined and the alarm is triggered.

[0017] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform.

[0018] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform.

[0019] Beneficial effects of the present invention: The photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform provided by the present invention collects arc current data through a high-frequency sampling device, performs FFT transformation on the fault arc, accurately captures the high-frequency noise characteristics of the arc, and provides a reliable data basis for subsequent frequency domain analysis. The current data is processed by combining time domain analysis and frequency domain analysis to determine the characteristic value of the fault arc. The accuracy of fault arc monitoring is significantly improved through dual judgment criteria, and environmental noise interference is reduced. The alarm threshold is set according to the characteristic value of the fault arc. The judgment result is set through the dynamic alarm threshold to determine the arc fault and trigger the alarm. Multiple judgments are made continuously to avoid the false alarm rate caused by instantaneous interference, and improve the reliability and stability of detection. The present invention achieves better results in detection accuracy, anti-interference ability and response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 This is an overall flow chart of a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform provided in the first embodiment of the present invention.

[0022] Figure 2 A photovoltaic fault arc current simulation diagram of a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform provided in the second embodiment of the present invention.

[0023] Figure 3 A simulation diagram of the fault arc current after filtering by a high-pass filter in a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform provided in the second embodiment of the present invention.

[0024] Figure 4 This is an FFT analysis diagram within 100kHz of a fault arc of a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform provided by the second embodiment of the present invention.

[0025] Figure 5 A fault arc simulation model diagram based on a photovoltaic inverter is provided for a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to the second embodiment of the present invention.

[0026] Figure 6 This is a current waveform diagram of a fault arc after passing through a 1kHz high-pass filter in a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform provided in the second embodiment of the present invention.

[0027] Figure 7 A waveform diagram of a fault arc current signal after wavelet transformation in a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transformation provided in a second embodiment of the present invention.

[0028] Figure 8 An overall schematic diagram of a photovoltaic arc detection system based on time-frequency domain analysis and wavelet transform is provided for the third embodiment of the present invention. DETAILED DESCRIPTION

[0029] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0030] Example 1, with reference to Figure 1 , as one embodiment of the present invention, provides a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform, comprising:

[0031] S1: Collect the photovoltaic panel current data in real time, calculate the high-frequency signal of the current data and compare it with the alarm threshold.

[0032] Furthermore, calculating the high-frequency signal of the current data includes performing FFT transformation on the collected current data, calculating the energy spectrum of the processed current data, dividing the frequency bands by step size and summing them to generate a characteristic frequency band energy spectrum matrix.

[0033] The energy spectrum matrix is ​​expressed as:

[0034]

[0035] Among them, [F1, F2, F3......, Fm] is is the characteristic frequency band of the step size, Fs is the sampling frequency, N is the number of sampling data points, S[k] is the unilateral power spectrum, k is the index value of the frequency domain component, and m is the coefficient.

[0036] It should be noted that the FFT transformation includes converting the time domain signal into the frequency domain signal, decomposing the signal frequency components to reveal the spectrum characteristics, analyzing the arc current energy spectrum distribution, and determining the characteristic frequency band of the fault arc.

[0037] The FFT transform is expressed as:

[0038]

[0039] Among them, X[k] is the complex value of the kth frequency domain component in the frequency domain, N is the number of sampled data, j is the imaginary unit, x[n] is the nth point of the sampled current, k is the index value of the frequency domain component, and n is the index value of the sampled current. The FFT transform is used to calculate the energy spectral density of the current signal to accurately identify the characteristic frequency band of the fault arc, providing a basis for the selection of the number of wavelet transform decomposition layers.

[0040] It should also be noted that the characteristic frequency band includes a specific frequency range where the energy in the current signal is significantly concentrated when a fault arc occurs. By analyzing the arc current energy spectrum density through FFT, the frequency range with the maximum energy is determined as the characteristic frequency band to identify the high-frequency noise characteristics of the fault arc.

[0041] The arc current energy spectrum density is expressed as:

[0042] S[k]=|X[k]| 2

[0043] Among them, S[k] is the unilateral power spectrum, X[k] is the frequency domain signal after FFT transformation, and k is the index value of the frequency domain component. Calculating the energy spectrum density can quickly locate the characteristic frequency band of the fault arc, provide a basis for the subsequent selection of the number of decomposition layers of wavelet transform, and avoid the deviation of manual experience judgment.

[0044] It should also be noted that the alarm threshold includes key parameters for judging arc faults, which include time domain and frequency domain. In the time domain, the standard deviation of the high-frequency current signal is calculated and compared with the standard deviation of the current data under normal working conditions to set the time domain alarm threshold. In the frequency domain, the wavelet detail coefficient of the current data is analyzed by wavelet transform, and the wavelet detail coefficients of the fault arc and the normal current are compared to set the frequency domain alarm threshold. The time domain alarm threshold and the frequency domain alarm threshold are collectively referred to as the alarm threshold.

[0045] It should also be noted that by collecting arc current signals and using FFT transformation, accurate analysis of the fault arc spectrum characteristics can be achieved, and the current energy spectrum density can be calculated. This can solve the limitations of traditional manual experience in judging characteristic frequency bands, avoid calculation deviations, and provide a basis for wavelet transformation.

[0046] S2: Perform wavelet decomposition on the current data and record the wavelet coefficients.

[0047] Furthermore, the wavelet decomposition of the current data includes using the db4 wavelet basis function, calculating the approximate coefficients through a low-pass filter and calculating the detail coefficients through a high-pass filter, performing multi-scale decomposition on the current data, and extracting the signal components of the characteristic frequency band of the current data.

[0048] It should be noted that the wavelet transform is expressed as:

[0049]

[0050]

[0051] Among them, a k is the approximate coefficient, h[m] is the low-pass filter coefficient, x[2k-m] is the sampling data, d k is the wavelet detail coefficient, h[0]-h[4] is the value of h[m], g[m] is the high-pass filter coefficient, F s is the sampling frequency.

[0052] It should also be noted that the alarm threshold of frequency domain analysis is expressed as:

[0053] T dk =2·(|d k -d' k |) / 3+d' k

[0054] Among them, T dk is the frequency domain alarm threshold, d k is the wavelet detail coefficient, d' k is the wavelet detail coefficient of normal current data.

[0055] It should also be noted that by combining time domain analysis and frequency domain analysis, high-precision detection of fault arcs is achieved. The wavelet detail coefficients corresponding to the characteristic frequency bands are extracted using wavelet transform, and the current data is compared through dynamic alarm thresholds to solve the problem that a single detection method is susceptible to noise interference and has a high false alarm rate, which significantly improves the accuracy and anti-interference ability of arc fault detection.

[0056] S3: After wavelet detail coefficient calculation and time domain signal calculation and comparison, if both the time domain signal and the frequency domain signal continuously exceed the alarm threshold, it is considered that an arc fault has occurred.

[0057] Furthermore, the calculation of wavelet detail coefficients includes decomposing the current data into detail coefficients of different frequency bands through wavelet transform, selecting the corresponding wavelet detail coefficients as the target by matching the frequency range of the FFT characteristic band energy spectrum matrix, and establishing the association between the frequency domain signal and the number of wavelet decomposition layers.

[0058] It should be noted that the time domain signal calculation includes analyzing the fluctuation characteristics of the high-frequency components of the current waveform based on the collected current data, quantifying the discreteness of the current data using statistical methods, capturing abnormal fluctuations by comparing with the preset alarm threshold, and identifying the time domain characteristics of the fault arc.

[0059] It should also be noted that the continuous occurrence of time domain signals and frequency domain signals exceeding the alarm threshold includes, based on dual comprehensive judgment of time and frequency domains, calculating the standard deviation of the high-frequency signal in the time domain, and extracting the wavelet detail coefficient through wavelet transform in the frequency domain. When the standard deviation and wavelet detail coefficient of the high-frequency signal continuously exceed the preset alarm threshold, an alarm action is triggered.

[0060] The standard deviation is expressed as:

[0061]

[0062] T σ =2·(|σ-σ'|) / 3+σ'

[0063] Where, σ is the standard deviation of the fault current, N is the number of sampling data, x(n) is the sampling data, μ is the average value of the current data amplitude, T σ is the time domain alarm threshold, and σ' is the normal current standard deviation.

[0064] Whether the time domain standard deviation exceeds the time domain alarm threshold is expressed as:

[0065]

[0066] Among them, S t Whether the time domain standard deviation exceeds the time domain alarm threshold, exceeding the time domain alarm threshold is represented by 1, and not exceeding the time domain alarm threshold is represented by 0, σ1 is the current variance.

[0067] Whether the frequency domain wavelet detail coefficient exceeds the frequency domain alarm threshold is expressed as:

[0068]

[0069] Among them, S f Whether the frequency domain wavelet detail coefficient exceeds the frequency domain alarm threshold, exceeding the frequency domain alarm threshold is represented by 1, and not exceeding the frequency domain alarm threshold is represented by 0, d k is the wavelet detail coefficient, T dk is the frequency domain alarm threshold.

[0070] It should also be noted that by combining dynamic alarm threshold setting with dual criteria, high-reliability detection of fault current can be achieved, solving the problem that traditional fixed alarm thresholds are easily affected by environmental noise and have a high false alarm rate. The alarm threshold can be adaptively adjusted to effectively distinguish between normal operating conditions and fault arcs, significantly improving the robustness and accuracy of fault arc detection and reducing false alarm and missed alarm rates.

[0071] Example 2, reference Figure 2-Figure 7, which is an embodiment of the present invention, provides a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0072] First, establish a photovoltaic arc detection simulation, such as Figure 2 As shown in the current simulation diagram of the photovoltaic power generation device, when the fault current occurs, the current in the circuit will suddenly decrease, accompanied by a series of high-frequency noise signals, such as Figure 3 The simulated waveform of the fault arc current after filtering by a 1kHz high-pass filter can be observed. It can be observed that a fault arc signal is generated in the circuit after 0.3s, the current suddenly decreases, and after 0.2s, it begins to burn stably, accompanied by high-frequency random noise. After analysis and calculation, the standard deviation of the high-frequency time domain signal of the fault arc is 0.1449. Figure 4 This is an FFT analysis of the fault arc current signal in the stable combustion stage. After calculation, it can be determined that the characteristic frequency band of the photovoltaic fault arc is between 20kHz and 65kHz. The wavelet detail coefficient d3 is extracted as the basis for fault arc judgment. The wavelet transform is decomposed into 3 layers. After performing wavelet transform on the fault arc, it is known that the standard deviation of d3 is 0.2582. Then, after analyzing and calculating the normal current data, it is known that the standard deviation of the high-frequency time domain signal of the normal current signal is 0.1180. After performing wavelet transform on the normal current signal, the standard deviation of the detail coefficient d3 is 0.00893. The wavelet detail coefficient, that is, the frequency domain alarm threshold T, is determined respectively. dk =0.17517, high-frequency time domain threshold is also the time domain alarm threshold T σ =0.13593.

[0073] Secondly, a fault arc detection simulation is established to simulate the occurrence of series arc in the photovoltaic inverter device, such as Figure 5 The figure below shows the simulation model of a photovoltaic inverter. The front stage consists of a photovoltaic cell, a boost circuit, and an MPPT controller, while the back stage consists of a three-phase full-bridge inverter circuit, an inverter control module, and a three-phase AC source. The Cassie arc model is connected in series to the DC bus to simulate a series arc. The simulation duration is set to 1s, and the arcing test is started at 0.3s. The test results are shown in Figure 2. Figure 6 . Figure 6 The arc current simulation waveform after filtering is shown in Figure 2. The current standard deviation of the high-frequency signal is 0.13905, which is greater than the time domain alarm threshold T. σ Then the arc current waveform data is subjected to wavelet transform, and the result is as follows: Figure 7 As shown. By calculation, we can know that the standard deviation of detail coefficients d1 = 0.1315, d2 = 0.3159, d3 = 0.3293, and we can know that d3>T dkAccording to steps 13) to 18), when the time domain signal and the frequency domain signal are both higher than the alarm threshold for 3 to 5 times in a row, an arc alarm signal is issued.

[0074] Example 3, reference Figure 8 , which is an embodiment of the present invention, provides a photovoltaic arc detection system based on time-frequency domain analysis and wavelet transform, including a data acquisition module 100, a feature analysis module 200, and a fault judgment module 300.

[0075] Among them, S4: data acquisition module 100 is used to capture the arc current data of the photovoltaic system in real time, provide high-precision original data for the time-frequency domain joint analysis of the fault arc, and perform high-speed acquisition of the current at the outlet end of the junction box by the sampling points, completely capturing the high-frequency noise characteristics when the arc occurs. The collected data is used for FFT transformation, analyzing the energy spectrum of the arc current data, selecting the frequency interval with the largest energy as the characteristic frequency band of the fault arc, determining the wavelet detail coefficient that is compatible with the characteristic frequency band, and providing the input signal for the wavelet transform.

[0076] Among them, S5: the feature analysis module 200 includes a time domain analysis module 201 and a frequency domain analysis module 202. The time domain analysis module 201 is used to extract the high-frequency time domain features of the current signal and judge the time domain alarm threshold. The fault arc collected by the sampling device and the current data under normal working conditions are used to calculate the standard deviation of the high-frequency time domain signal, and compare it with the standard deviation of the current data time domain signal under normal working conditions to determine the alarm threshold. The frequency domain analysis module 202 is used to extract the high-frequency noise features of the current signal and optimize the frequency domain judgment criteria of the fault arc. The adaptive db4 wavelet basis function and decomposition layer number are selected based on the characteristic frequency band obtained by FFT transformation, the wavelet detail coefficient threshold is calculated, the time domain analysis and frequency domain analysis are repeated many times, and the threshold of the comprehensive characteristic value is used for arc detection.

[0077] Among them, S6: The fault judgment module 300 is used to synchronously execute the time domain standard deviation calculation and the frequency domain wavelet detail coefficient extraction, and dynamically compare the calculation results with the preset alarm threshold. The time domain amplitude fluctuation and the high-frequency noise characteristics simultaneously trigger the alarm threshold to exceed the limit. When the time domain standard deviation and the corresponding frequency band detail coefficient are both exceeded in continuous monitoring, the arc fault is determined and the alarm is triggered.

Claims

1. A photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform, characterized in that: include: Collect photovoltaic panel current data in real time, calculate the high-frequency signal of the current data and compare it with the alarm threshold; Perform wavelet decomposition on the current data and record the wavelet coefficients; After wavelet detail coefficient calculation and time domain signal calculation and comparison, if both the time domain signal and the frequency domain signal continuously exceed the alarm threshold, it is considered that an arc fault has occurred; The frequency domain signal includes signal characteristics obtained by performing frequency domain analysis on the current data, reflecting the energy distribution of the current data at different frequencies.

2. The photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to claim 1, characterized in that: The high frequency signal for calculating current data includes: Perform FFT transformation on the collected current data, calculate the energy spectrum of the processed current data, divide the frequency band by step size and sum it. Generates the eigenband energy spectrum matrix.

3. The photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to claim 2, characterized in that: The FFT transformation includes: By converting the time domain signal into the frequency domain signal, decomposing the signal frequency components to reveal the spectrum characteristics, analyzing the arc current energy spectrum distribution, and determining the characteristic frequency band of the fault arc.

4. A photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to claim 2 or 3, characterized in that: The characteristic frequency bands include: When a fault arc occurs, the energy in the current signal is significantly concentrated in a specific frequency range. The arc current energy spectral density is analyzed by FFT, and the frequency range with the maximum energy is determined as the characteristic frequency band to identify the high-frequency noise characteristics of the fault arc.

5. The photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to claim 1, characterized in that: The alarm thresholds include: The key parameters for judging arc faults include time domain and frequency domain. In the time domain, the standard deviation of the high-frequency current signal is calculated and compared with the standard deviation of the current data under normal working conditions to set the time domain alarm threshold. In the frequency domain, the wavelet detail coefficient of the current data is analyzed by wavelet transform, and the wavelet detail coefficients of the fault arc and the normal current are compared to set the frequency domain alarm threshold. The time domain alarm threshold and the frequency domain alarm threshold are collectively referred to as the alarm threshold.

6. The photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to claim 1, characterized in that: The wavelet decomposition of the current data includes: The db4 wavelet basis function is used to calculate the approximate coefficients through a low-pass filter and the detail coefficients through a high-pass filter. The current data is decomposed into multiple scales to extract the signal components of the characteristic frequency band of the current data.

7. The photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to claim 1, characterized in that: The wavelet detail coefficient calculation includes: The current data is decomposed into detail coefficients of different frequency bands through wavelet transform. By matching the frequency range of the FFT characteristic band energy spectrum matrix, the corresponding wavelet detail coefficients are selected as the target, and the association between the frequency domain signal and the wavelet decomposition layer number is established.

8. The photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to claim 1, characterized in that: The time domain signal calculation includes: Based on the collected current data, the fluctuation characteristics of the high-frequency components of the current waveform are analyzed, and the discrete degree of the current data is quantified using statistical methods. The abnormal fluctuations are captured by comparing with the preset alarm threshold and the time domain characteristics of the fault arc are identified.

9. The photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to claim 1, characterized in that: The time domain signal and the frequency domain signal both continuously exceed the alarm threshold, including: Based on dual comprehensive judgment in the time and frequency domains, the standard deviation of the high-frequency signal is calculated in the time domain, and the wavelet detail coefficient is extracted through wavelet transform in the frequency domain. When the standard deviation and wavelet detail coefficient of the high-frequency signal continuously exceed the preset alarm threshold, the alarm action is triggered.

10. A photovoltaic arc detection system based on time-frequency domain analysis and wavelet transform, characterized by: It includes a data acquisition module (100), a feature analysis module (200), and a fault determination module (300); The data acquisition module (100) is used to capture the arc current data of the photovoltaic system in real time, provide high-precision original data for the time-frequency domain joint analysis of the fault arc, perform high-speed acquisition of the current at the outlet end of the combiner box at the sampling points, and completely capture the high-frequency noise characteristics when the arc occurs. The acquired data is used for FFT transformation, and the energy spectrum of the arc current data is analyzed. The frequency interval with the largest energy is selected as the characteristic frequency band of the fault arc, and the wavelet detail coefficient adapted to the characteristic frequency band is determined to provide an input signal for the wavelet transformation. The characteristic analysis module (200) comprises a time domain analysis module (201) and a frequency domain analysis module (202). The time domain analysis module (201) is used to extract the high-frequency time domain characteristics of the current signal and determine the time domain alarm threshold value. The standard deviation of the high-frequency time domain signal is calculated using the current data of the fault arc and the current data under normal working conditions collected by the sampling device. The standard deviation is compared with the standard deviation of the time domain signal of the current data under normal working conditions to determine the alarm threshold value. The frequency domain analysis module (202) is used to extract the high-frequency noise characteristics of the current signal and optimize the frequency domain judgment criterion of the fault arc. The adaptive db4 wavelet basis function and decomposition layer number are selected based on the characteristic frequency band obtained by FFT transformation, and the wavelet detail coefficient threshold value is calculated. The time domain analysis and frequency domain analysis are repeated multiple times, and arc detection is performed based on the threshold value of the comprehensive characteristic value. The fault determination module (300) is used to synchronously perform time domain standard deviation calculation and frequency domain wavelet detail coefficient extraction, dynamically compare the calculation result with a preset alarm threshold, and simultaneously trigger the alarm threshold to exceed the limit when the time domain amplitude fluctuation and the high-frequency noise characteristics both exceed the limit during continuous monitoring. When both the time domain standard deviation and the corresponding frequency band detail coefficient exceed the limit, an arcing fault is determined and an alarm is triggered.

11. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to any one of claims 1 to 9 are implemented.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a photovoltaic arc detection method based on time-frequency domain analysis and wavelet transform according to any one of claims 1 to 9 are implemented.

Images