A photovoltaic array fault detection and location method based on time domain reflectometry

Abstract

The present application relates to the field of fault diagnosis of photovoltaic power generation system, and particularly to a photovoltaic array fault detection and positioning method based on time domain reflection method, which specifically comprises the following steps: injecting a step signal into the photovoltaic array; adjusting the sampling interval according to the size of the array to be measured to obtain the reflection waveform under normal state and the reflection waveform within the fault detection range; carrying out wavelet threshold denoising processing on the reflection signal of the fault array; calculating the signal amplitude; determining whether a fault occurs according to the preset fault detection threshold, and positioning the fault point according to the time delay estimation algorithm if a fault occurs. The present application detects faults according to the characteristic that the impedance value of the fault point changes when the photovoltaic array fails, positions the fault point by measuring the time difference between the incident signal and the reflection signal of the fault point, and can realize the fault diagnosis function by using an external detection device without adding or modifying electrical equipment of the photovoltaic power generation system, thus having good fault point detection and positioning effect, avoiding the problem of manually removing each component and detecting them one by one in the photovoltaic array fault detection process, and improving the operation and maintenance efficiency of the photovoltaic power station and ensuring the safety of maintenance personnel.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis of photovoltaic power generation systems, specifically a method for fault detection and location of photovoltaic arrays based on time-domain reflectometry. Background Technology

[0002] In recent years, my country's cumulative installed photovoltaic (PV) capacity has increased year by year, reaching 300.2 GW by the end of March 2022, of which centralized PV power generation accounted for 63.6%. Simultaneously, the National Development and Reform Commission (NDRC) released policies regarding feed-in tariffs for new energy sources, implementing grid parity for newly registered PV projects. Against the backdrop of a steadily increasing share of PV power generation, rapidly declining costs, and PV power generation entering a phase of grid parity without subsidies, focusing on key issues such as high labor costs and low levels of automation in the PV industry, further reducing industry operating costs, and promoting high-quality development of the PV power generation industry are important measures for my country to achieve its carbon peaking and carbon neutrality goals on schedule.

[0003] With the large-scale development of photovoltaic (PV) power generation and the reduction of PV subsidies, the importance of daily maintenance and fault diagnosis of PV power plants is becoming increasingly prominent. Influenced by the distribution patterns of solar energy resources and land management policies, PV power generation systems are often installed in harsh climates such as deserts, Gobi, and wastelands. During operation, they are susceptible to external environmental factors such as solar irradiance, temperature, humidity, dust, and hail, leading to various electrical faults caused by factors such as cell breakage, poor terminal contact, and cable damage and aging. These faults include grounding faults, short circuits between lines, open circuits, and arcing faults. These faults not only reduce power generation efficiency but can also lead to more destructive consequences, such as module damage and fires, severely degrading the power generation performance of the entire PV module series array or even preventing it from generating power. Therefore, corresponding fault detection and isolation technologies are essential to protect the PV system.

[0004] Photovoltaic arrays often employ overcurrent protection devices and ground fault detection circuit breakers to prevent faults. However, the unique operating characteristics of photovoltaic systems can cause these protection devices to malfunction. For example, under low irradiance conditions, the current of the faulty component is low, and the overcurrent protection device cannot detect the inter-line short circuit fault; ground fault detection circuit breakers may fail under low irradiance and high resistance faults; and the optimization of the power grid by maximum power point tracking (MPPT) technology can prevent circuit breakers from identifying circuit anomalies. Furthermore, photovoltaic arrays often use a series-parallel (SP) structure, meaning that when a fault occurs, these devices can only locate the faulty component in the series array, not the precise fault location. According to the relevant design standards in the "Design Code for Photovoltaic Power Plants" (GB50797-2012), the maximum number of photovoltaic modules in series in a photovoltaic power plant can reach 25 to 30, and they are often installed in harsh environments such as high mountains and deserts. Current fault diagnosis methods for photovoltaic arrays still primarily rely on manual, module-by-module inspection offline, resulting in long maintenance cycles and high risks. Therefore, establishing rapid and effective fault diagnosis and location technologies is of great significance for strengthening the daily maintenance and fault diagnosis of photovoltaic power plants and promoting the safe, stable and reliable operation of photovoltaic power generation systems. Summary of the Invention

[0005] To address the aforementioned problems, the present invention aims to provide a photovoltaic array fault detection and location method based on the time-domain reflectometry method. This method detects faults by utilizing the characteristic of impedance changes at fault points in the photovoltaic array, and has a good effect on fault location in photovoltaic arrays.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for photovoltaic array fault detection and location based on time-domain reflectometry includes the following steps:

[0008] Step 1, Connect the detection device: Connect an external detection device to the photovoltaic array under test to transmit and collect signals;

[0009] Step 2, Signal transmission: The step signal emitted by the detection device is injected into the photovoltaic array under test through one of the interfaces of the T-type BNC tee connector;

[0010] Step 3, acquire waveforms under fault-free conditions: Adjust the sampling interval according to the size of the array under test, and acquire a set of photovoltaic array reflection signal waveforms under fault-free conditions;

[0011] Step 4, acquire waveforms under fault conditions: adjust the sampling interval according to the size of the array under test, and obtain the reflected signal waveform within the fault detection range through the signal acquisition device;

[0012] Step 5, Denoising: The wavelet threshold denoising algorithm is used to denoise the reflected signal waveforms in Step 3 and Step 4 respectively to obtain the denoised reflected signal waveforms.

[0013] Step 6, obtain the signal amplitude: After denoising, subtract the sampling sequence of the photovoltaic array reflected signal waveform under fault conditions from the sampling sequence of the photovoltaic array reflected signal waveform under fault-free conditions to obtain the signal amplitude;

[0014] Step 7, Fault Determination and Location: Set a preset fault detection threshold. Determine whether a fault has occurred based on the preset fault detection threshold and the signal amplitude obtained in Step 6. If a fault has occurred, calculate the fault distance and locate the fault point through the fault distance.

[0015] Furthermore, the detection device includes a signal generating device, a signal acquiring device, and a T-type BNC tee connector. The positive terminals of the signal generating device and the signal acquiring device are respectively connected to any two ports of the T-type BNC tee connector. The negative terminals of the signal acquiring device and the signal generating device are connected to the photovoltaic array ground wire or the photovoltaic array metal frame. The positive terminal of the photovoltaic array is connected to the remaining port of the T-type BNC tee connector. Preferably, this invention connects the photovoltaic array to the T-type BNC tee by introducing a "BNC-MC4" adapter interface.

[0016] The entire system connection process requires close coordination. This means ensuring continuous cable impedance and good conductivity, while also considering the impact of connectors to prevent signal attenuation and failure to obtain reflected signals due to impedance mismatch. Furthermore, impedance matching between the signal generator and signal acquisition device must be ensured to reduce interference. To guarantee smooth testing, the signal generator, signal acquisition device, T-type BNC connector, BNC-MC4 adapter, and photovoltaic module must be tightly connected. The interfaces of the signal generator, signal acquisition device, and BNC connector should be impedance matched, with input and output impedances of 50Ω. Since the BNC connector is a coaxial output interface with two signal outputs (positive and negative), while the MC4 interface is a single output, a simple BNC-MC4 adapter needs to be constructed. This involves splitting the BNC male connector lead into two wires (positive and negative). The positive wire is connected to the MC4 connector terminal, and the negative wire is connected to the photovoltaic module ground wire to complete the signal loop.

[0017] Furthermore, the signal generating device in step 1 is a square wave signal generator that generates a step signal with a rise time of less than 10ns, a voltage amplitude of less than 5V, and a frequency of less than 100KHz.

[0018] Furthermore, the signal sampling device in step 1 uses a sampling rate greater than 50 MSPS.

[0019] To enable fault detection in photovoltaic arrays, specific requirements are placed on the design of the signal generator and signal acquisition device. The time-domain reflectometry method primarily detects faults based on signal singularities; that is, fault detection is only achieved when a certain output frequency is reached at the signal abrupt change. By adopting the above technical solution, to achieve high-frequency signal output, the signal generator should be a square wave signal generator capable of producing low-voltage, high-speed step signals with a rise time of less than 10ns, a voltage amplitude of less than 5V, and a frequency of less than 100kHz.

[0020] Meanwhile, during the acquisition of high-frequency signals, if the sampling rate and resolution are too low, fault signals cannot be effectively captured, affecting fault detection. In actual use, the sampling rate of the signal acquisition device should be greater than 50 MSPS, and it should also have a certain resolution to clearly identify small-scale faults, such as cable damage or loose connections. Therefore, the signal sampling device should use a sampling rate greater than 50 MSPS and a resolution of 8 bits or more.

[0021] Furthermore, the noise reduction process in step 4 includes the following steps:

[0022] Step 41: Perform multi-level wavelet decomposition on the original signal using the DB4 wavelet basis, and extract the approximate coefficient vector A = [a1, a2, a3, ..., a...]. n Extract the detail coefficient vector D = [d1, d2, d3, ..., d] from each layer. n ];

[0023] Step 42: Denoise the detail coefficients containing noisy signals, removing at least the first 3 layers of detail coefficients;

[0024] Step 43: Perform wavelet reconstruction on the denoised wavelet coefficients, using the approximation coefficient a. n and detail coefficient d n The waveform of the reflected signal after denoising is obtained.

[0025] Furthermore, the signal amplitude calculation method in step 6 is as follows: N sub,i =A si -A Ni In the formula A si and A Ni These are the amplitudes of the reflected signals at the i-th sampling point of the fault-free system and the faulty system, respectively.

[0026] Furthermore, the preset fault detection threshold in step 7 is set to 5%-10% of the step signal voltage amplitude. When the signal amplitude obtained in step 6 is greater than the fault detection threshold, a fault is determined to have occurred.

[0027] Furthermore, the fault distance in step 7 can be calculated using the following method:

[0028] (1) The incident delay t0 is the time when the signal rises rapidly, the reflection delay t1 is the time when the reflected wave exceeds the threshold, and the propagation speed v of the signal in the photovoltaic array is the signal propagation speed v.

[0029] (2) Calculate the signal delay t after the signal passes through the photovoltaic array and is reflected back, where t = t1 - t0;

[0030] (3) Calculate the fault distance L at the fault point, which is obtained by a time delay estimation algorithm:

[0031]

[0032] Furthermore, the fault distance in step 7 can also be calculated using the following method:

[0033] For ease of detection, the fault distance is typically equated in the art to the time delay t of a single photovoltaic module. d ,now that In the formula L d The signal propagation distance corresponding to a single component;

[0034] By transforming v, we can obtain v. Substituting v into the time delay estimation algorithm, we get:

[0035] Since the fault distance is positively correlated with the number of components, the fault location can be determined as follows: In the formula, n represents the position of the nth component in the photovoltaic array.

[0036] Therefore, the fault distance described above can be calculated using the following steps:

[0037] (1) Calculate the total signal delay t: the time when the signal rises rapidly is the incident delay t0, the time when the reflected wave exceeds the threshold is the reflection delay t1, and the signal delay t when the signal passes through the photovoltaic array and is reflected back is calculated, where t = t1 - t0.

[0038] (2) Determine the time delay t corresponding to a single component. d Select a standard photovoltaic module, inject a step signal into the standard photovoltaic module, and measure the corresponding time delay through calibration. This time delay t corresponds to a single module. d ;

[0039] (3) Calculate the fault distance n: The fault distance n is calculated using the following formula, thereby determining the location of the fault point.

[0040]

[0041] The time-domain reflection method is mainly based on the directionality and reflectivity of wave propagation to locate faults. It mainly achieves accurate fault location by analyzing the output signal in the time domain.

[0042] In photovoltaic power generation systems, various transmission lines and power transmission media are widely used as core components for energy transport, including the main and auxiliary grid lines connecting photovoltaic cells and the connecting cables between photovoltaic modules. In location technology based on the time-domain reflectometry method, the basic approach for fault location and diagnosis relies on the traveling wave transmission theory in transmission lines. When the frequency of the transmitted signal increases, the propagation mode of the signal in the transmission line is no longer considered as that of an ordinary conductor, but rather as exhibiting distributed resistance, distributed inductance, distributed capacitance, and distributed conductance effects. High-frequency current flowing through the conductor surface generates a skin effect, reducing the effective conductive area of ​​the conductor and increasing the high-frequency resistance; high-frequency current generates high-frequency magnetism around the conductor; and due to the voltage between the conductors, a high-frequency electric field exists between them. Furthermore, because the medium surrounding the conductor is not ideally insulated, leakage current exists. When the geometric length of the transmission line is on the same order of magnitude as the wavelength of the electromagnetic wave, the phase lag, skin effect, phase shift caused by radiation effects, increased transmission impedance, and signal attenuation characteristics during wave transmission cannot be ignored. The signal changes with time and displacement during transmission, and this method applies this principle for detection.

[0043] According to traveling wave propagation theory, a wave on a transmission line is formed by the superposition of an incident wave and a reflected wave. When the impedance loads connected to both ends of the transmission line match the characteristic impedance of the transmission line, the incident signal will be completely transmitted from one end to the other without reflection or transmission. Conversely, when the traveling wave encounters impedance mismatch points such as fault points, termination joints, or loads, it will be reflected and transmitted. Time-domain reflectometry (TD-RS) often injects a low-voltage pulse signal into the photovoltaic system under test for fault diagnosis. When the line impedance at any point within the transmission line is matched, the incident signal will be completely transmitted from one end to the other without reflection or refraction. Conversely, when the traveling wave encounters impedance mismatch points such as short circuits, open circuit fault points, termination joints, or loads, it will be reflected and transmitted. Therefore, the transmission line wave state at a corresponding moment can be described by measuring the change in the reflection coefficient during signal propagation, enabling fault detection and location of the photovoltaic array.

[0044] Compared with existing technologies, the photovoltaic array fault detection and location method of the present invention has the following advantages:

[0045] This invention detects faults based on the characteristic that the impedance value of the fault point changes when a photovoltaic array malfunctions. It locates the fault point by measuring the time difference between the incident signal and the reflected signal at the fault point. The fault diagnosis function can be realized using an external detection device without adding or modifying electrical equipment in the photovoltaic power generation system. It has a good fault point detection and location effect, avoiding the problem of manually disassembling and inspecting each component one by one during the photovoltaic array fault detection process, thus improving the operation and maintenance efficiency of photovoltaic power plants and ensuring the safety of maintenance personnel. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the fault detection and location method of the present invention;

[0047] Figure 2 This is a block diagram showing the connection between the detection device and the photovoltaic array required for fault detection and location in this invention;

[0048] Figure 3 This is an initial reflection waveform diagram of the reflected signal within the fault detection range obtained by the present invention.

[0049] Figure 4 This is a waveform diagram of the reflected signal after the noise reduction process of this invention. Detailed Implementation

[0050] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention.

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments and accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Example 1

[0053] like Figures 1-2 As shown, this invention provides a photovoltaic array fault detection and location method based on time-domain reflectometry, comprising the following steps:

[0054] Step 1, Connect the detection device: Connect an external detection device to the photovoltaic array under test to transmit and collect signals. This detection device includes a signal acquisition device, a signal generator, a T-type BNC tee connector, and a BNC-MC4 adapter. Specifically, the positive terminals of the signal generator and signal acquisition device are connected to any two ports of the T-type BNC tee connector, respectively. The negative terminals of the signal acquisition and signal generator are connected to the photovoltaic array ground wire, or, if no ground wire is available, to the photovoltaic array metal frame. The positive terminal of the photovoltaic array is connected to the remaining port of the T-type BNC tee connector. This invention connects the photovoltaic array to the T-type BNC tee by introducing a "BNC-MC4" adapter. Specifically, the lead wire of the BNC male connector is split into two wires, positive and negative. The positive wire is connected to the MC4 connector terminal, and the negative wire is connected to the photovoltaic module ground wire to complete the signal loop.

[0055] Preferably, in this invention, a square wave signal generator that generates a low-voltage, high-speed step signal with a rise time of less than 10 ns, a voltage amplitude of less than 5V, and a frequency of less than 100KHz is selected as the signal generating device, and a signal sampling device with a sampling rate of greater than 50MSPS is selected.

[0056] It should be noted that, for the signal generating device and signal sampling device mentioned in this step, any arbitrary waveform function generator, DDS signal generator, self-made signal generating hardware circuit board, etc., that can generate a square wave signal that meets the detection requirements are all considered signal generating devices, and any oscilloscope, ADC analog-to-digital converter, self-made analog signal sampling hardware circuit board, etc., that meets the signal sampling requirements are all considered signal sampling devices.

[0057] Step 2, Signal transmission: The step signal emitted by the signal generator is injected into the photovoltaic array under test through one of the interfaces of the T-type BNC tee connector to start detecting whether the photovoltaic array is faulty and collecting relevant data;

[0058] Step 3: First, the present invention needs to collect a set of waveforms under fault-free conditions. The sampling interval is adjusted according to the size of the array under test. After adjusting the sampling interval, a set of photovoltaic array reflection signal waveforms under fault-free conditions are collected.

[0059] Step 4, continue to collect waveforms under fault conditions: adjust the sampling interval according to the size of the array under test, and obtain the reflected signal waveform within the fault detection range through the signal acquisition device;

[0060] In this invention, the sampling interval adjustment method is as follows: the propagation time t of the incident signal through a single photovoltaic module is generally about 100 ns. Since the incident signal is reflected back in the photovoltaic array, the minimum sampling interval should be set to the number of modules in the photovoltaic array under test, n*2t0, and the maximum sampling interval should not exceed n*3t0, so as to ensure that the signal waveform has sufficient resolution.

[0061] Step 5, Denoising: The wavelet threshold denoising algorithm is used to denoise the reflected signal waveforms in Step 3 and Step 4 respectively to obtain the denoised reflected signal waveforms; the specific processing steps are as follows: (1) The original signal is decomposed into multiple wavelet layers using the DB4 wavelet basis, and the approximate coefficient vectors of each layer are extracted A = [a1, a2, a3, ..., a n Extract the detail coefficient vector D = [d1, d2, d3, ..., d] from each layer. n (2) Denoise the detail coefficients of the noisy signal. To achieve a suitable signal waveform smoothness, at least the first three layers of detail coefficients should be removed to shield the high-frequency part of the signal. (3) Perform wavelet reconstruction on the denoised wavelet coefficients and use the approximation coefficient a. n and detail coefficient d n The waveform of the reflected signal after denoising is obtained.

[0062] It should be noted that the wavelet thresholding denoising algorithm can be implemented using various software such as MATLAB and C++. The specific number of detail coefficients to be removed can be selected according to the specific waveform characteristics, but at least the first 3 layers of detail coefficients must be removed.

[0063] Step 6, Obtain the signal amplitude: After denoising, subtract the sampling sequence of the photovoltaic array reflected signal waveform under fault conditions from the sampling sequence of the photovoltaic array reflected signal waveform under fault-free conditions to obtain the signal amplitude N. sub,i Its calculation expression is N sub,i =A si -A Ni In the formula A si and A Ni These are the amplitudes of the reflected signals at the i-th sampling point of the fault-free system and the faulty system, respectively.

[0064] Step 7, Fault Judgment and Location: Preset a fault detection threshold, which is set to 5%-10% of the step signal voltage amplitude. Determine whether a fault has occurred based on the preset fault detection threshold and the signal amplitude obtained in Step 6. A fault occurs when the signal amplitude is greater than the fault detection threshold. Further calculate the fault distance L. After calculating the fault distance, the location of the fault point can be determined, and the fault point can be located and repaired.

[0065] In this embodiment, the fault distance needs to be calculated by obtaining the signal delay t when the signal passes through the photovoltaic array and is reflected back, and the signal propagation speed v in the photovoltaic array. The specific steps are as follows: (1) Obtain the incident delay t0 when the signal generates a rapid rise by calibration measurement, and obtain the reflection delay t1 when the reflected wave exceeds the threshold critical point and the signal propagation speed v in the photovoltaic array; (2) Calculate the signal delay t when the signal passes through the photovoltaic array and is reflected back, where t = t1 - t0; (3) Calculate the fault distance L of the fault point by using a time delay estimation algorithm. The specific calculation formula is as follows: This allows us to pinpoint the location of the fault.

[0066] Example 2

[0067] like Figures 1-2 As shown, this invention provides a photovoltaic array fault detection and location method based on time-domain reflectometry, comprising the following steps:

[0068] Step 1, Connect the detection device: Connect an external detection device to the photovoltaic array under test to transmit and collect signals. This detection device includes a signal acquisition device, a signal generator, a T-type BNC tee connector, and a BNC-MC4 adapter. Specifically, the positive terminals of the signal generator and signal acquisition device are connected to any two ports of the T-type BNC tee connector, respectively. The negative terminals of the signal acquisition and signal generator are connected to the photovoltaic array ground wire, or, if no ground wire is available, to the photovoltaic array metal frame. The positive terminal of the photovoltaic array is connected to the remaining port of the T-type BNC tee connector. This invention connects the photovoltaic array to the T-type BNC tee by introducing a "BNC-MC4" adapter interface. Specifically, the lead wire of the BNC male connector is split into two wires, positive and negative. The positive wire is connected to the MC4 connector terminal, and the negative wire is connected to the photovoltaic module ground wire to complete the signal loop.

[0069] Preferably, in this invention, a square wave signal generator that generates a low-voltage, high-speed step signal with a rise time of less than 10 ns, a voltage amplitude of less than 5V, and a frequency of less than 100KHz is selected as the signal generating device, and a signal sampling device with a sampling rate of greater than 50MSPS is selected.

[0070] It should be noted that, for the signal generating device and signal sampling device mentioned in this step, any arbitrary waveform function generator, DDS signal generator, self-made signal generating hardware circuit board, etc., that can generate a square wave signal that meets the detection requirements are all considered signal generating devices, and any oscilloscope, ADC analog-to-digital converter, self-made analog signal sampling hardware circuit board, etc., that meets the signal sampling requirements are all considered signal sampling devices.

[0071] Step 2, Signal transmission: The step signal emitted by the signal generator is injected into the photovoltaic array under test through one of the interfaces of the T-type BNC tee connector to start detecting whether the photovoltaic array is faulty and collecting relevant data;

[0072] Step 3: First, the present invention needs to collect a set of waveforms under fault-free conditions. The sampling interval is adjusted according to the size of the array under test. After adjusting the sampling interval, a set of photovoltaic array reflection signal waveforms under fault-free conditions are collected.

[0073] Step 4, continue to collect waveforms under fault conditions: adjust the sampling interval according to the size of the array under test, and obtain the reflected signal waveform within the fault detection range through the signal acquisition device;

[0074] Step 5, Denoising: The wavelet threshold denoising algorithm is used to denoise the reflected signal waveforms in Step 3 and Step 4 respectively to obtain the denoised reflected signal waveforms; the specific processing steps are as follows: (1) The original signal is decomposed into multiple wavelet layers using the DB4 wavelet basis, and the approximate coefficient vectors of each layer are extracted A = [a1, a2, a3, ..., a n Extract the detail coefficient vector D = [d1, d2, d3, ..., d] from each layer. n (2) Denoise the detail coefficients of the noisy signal. To achieve a suitable signal waveform smoothness, at least the first three layers of detail coefficients should be removed to shield the high-frequency part of the signal. (3) Perform wavelet reconstruction on the denoised wavelet coefficients and use the approximation coefficient a. n and detail coefficient d n The waveform of the reflected signal after denoising is obtained.

[0075] It should be noted that the denoising algorithm can be implemented using various software such as MATLAB and C++. The specific number of detail coefficients to be removed can be selected according to the specific waveform characteristics, but at least the first 3 layers of detail coefficients must be removed.

[0076] Step 6, Obtain the signal amplitude: After denoising, subtract the sampling sequence of the photovoltaic array reflected signal waveform under fault conditions from the sampling sequence of the photovoltaic array reflected signal waveform under fault-free conditions to obtain the signal amplitude N. sub,i Its calculation expression is N sub,i =A si -A Ni In the formula A si and A Ni These are the amplitudes of the reflected signals at the i-th sampling point of the fault-free system and the faulty system, respectively.

[0077] Step 7: Preset a fault detection threshold. The fault detection threshold is set to 5%-10% of the step signal voltage amplitude. Based on the preset fault detection threshold and the signal amplitude obtained in Step 6, it is determined whether a fault has occurred. When the signal amplitude is greater than the fault detection threshold, a fault occurs. Further calculate the fault distance L. After calculating the fault distance, the location of the fault point can be determined, and the fault point can be located and repaired.

[0078] In this embodiment, the second method is used to calculate the fault distance. For ease of detection, the fault distance is often equivalent to the time delay t of a single photovoltaic module. d ,now that In the formula L d Let the signal propagation distance corresponding to a single component be the formula. Substituting the formula into the time delay estimation algorithm formula, we get: Since the fault distance is positively correlated with the number of components, the fault location can be determined as follows: In the formula, n represents the position of the nth component in the photovoltaic array.

[0079] Specifically, the propagation time of a signal in a single photovoltaic module is determined through calibration measurement. This involves selecting a standard photovoltaic module, injecting the signal into it, measuring the signal delay, and determining the time delay t corresponding to a single module. d The fault location is achieved by using the signal delay t obtained from the detection and the arrangement pattern of the photovoltaic array.

[0080] Example 3

[0081] This embodiment provides a specific operation of the photovoltaic array fault detection and location method based on the time-domain reflectometry of the present invention: An external detection device is connected to the photovoltaic array under test. The detection device includes a signal acquisition device, a signal generation device, a T-type BNC tee connector, and a BNC-MC4 adapter. The specific connection relationship is as follows: the positive terminals of the signal generation device and the signal acquisition device are respectively connected to any two interfaces of the T-type BNC tee connector. The negative terminals of the signal acquisition device and the signal generation device are connected to the ground wire of the photovoltaic array, or to the metal frame of the photovoltaic array if there is no ground wire. The positive terminal of the signal generation device is connected to the BNC male connector on the "BNC-MC4" adapter through the interface of the T-type BNC tee connector. The lead wire of the BNC male connector is divided into two wires, positive and negative. The positive wire is connected to the terminal of the MC4 connector, and the negative wire is connected to the ground wire of the photovoltaic module to complete the signal loop. In other words, the positive terminal of the photovoltaic array is connected to the remaining interface of the T-type BNC tee connector through the "BNC-MC4" adapter interface. In this embodiment, the signal generating device can be any one of a square wave signal waveform function generator, a DDS signal generator, or a self-made signal generating hardware circuit board, and the signal sampling device can be any one of an oscilloscope, an ADC analog-to-digital converter, or a self-made analog signal sampling hardware circuit board.

[0082] After connecting the detection device and powering it on, the step signal generated by the signal generator is injected into the photovoltaic array under test through one of the interfaces of the T-type BNC three-way connector to start detecting whether the photovoltaic array has a fault and collecting relevant data.

[0083] A set of waveforms under fault-free conditions is collected. The sampling interval is adjusted according to the size of the array under test. After the sampling interval is adjusted, a set of photovoltaic array reflection signal waveforms under fault-free conditions are collected for calculating the signal amplitude.

[0084] Acquiring waveforms under fault conditions: Adjust the sampling interval according to the size of the array under test, and obtain the reflected signal waveform within the fault detection range through the signal acquisition device, i.e., the initial reflected waveform diagram, such as... Figure 3 As shown in the diagram, the fault detection type is photovoltaic array open circuit fault detection. A total of 8 open circuit fault points are set, namely: the connection interruption between the photovoltaic DC cable and the photovoltaic array; and 7 connection interruptions between each component. The components closest to the detection equipment are numbered starting from 1. The interruption fault between components 1 and 2 is the first group of detection, and the interruption fault between components 7 and 8 is the last group of detection.

[0085] In this embodiment, 8 sets of open circuit faults are used as samples. In actual testing, multiple types of faults such as short circuits and open circuits can be detected. At the same time, the specific component testing type and quantity are not limited in actual testing. Due to the differences in measurement environment and component parameters, the maximum detection range is subject to the on-site testing results.

[0086] Figure 3 The initial reflection waveform shown can only roughly estimate the fault point and has a large detection error. Therefore, it is necessary to eliminate the influence of noise in the signal and improve the fault detection accuracy. The specific denoising process is as follows: use the wavelet threshold denoising algorithm to denoise the reflection signal waveforms in steps 3 and 4 respectively to obtain the denoised reflection signal waveform; the specific processing steps are as follows: (1) use the DB4 wavelet basis to perform multi-level wavelet decomposition on the original signal and extract the approximate coefficient vector A = [a1, a2, a3, ..., a n Extract the detail coefficient vector D = [d1, d2, d3, ..., d] from each layer. n (2) Denoise the detail coefficients of the noisy signal. To achieve a suitable signal waveform smoothness, at least the first three layers of detail coefficients should be removed to shield the high-frequency part of the signal. (3) Perform wavelet reconstruction on the denoised wavelet coefficients and use the approximation coefficient a. n and detail coefficient d n The denoised reflected signal waveform is obtained. This operation eliminates multiple reflections caused by impedance inconsistencies within the photovoltaic array itself. Step E then determines whether a fault has occurred and locates the fault point.

[0087] It should be noted that the denoising algorithm can be implemented using various software such as MATLAB and C++. The specific denoising threshold setting range is selected according to the specific waveform characteristics. However, in order to achieve a suitable signal smoothness, at least the first 3 layers of detail coefficients should be removed to shield the high-frequency part of the signal.

[0088] After noise reduction, the signal amplitude is obtained. The signal amplitude N is obtained by subtracting the sampling sequence of the photovoltaic array reflected signal waveform under fault conditions from the sampling sequence under fault-free conditions. sub,i Its calculation expression is N sub,i =A si -A Ni In the formula A si and A Ni These represent the amplitudes of the reflected signals at the i-th sampling point of the fault-free system and the faulty system, respectively. The above formula is applied under the condition that the sampling standard for the photovoltaic array reflected waveforms is the same in both faulty and healthy states.

[0089] A preset fault detection threshold is set, which is an amplitude threshold. This invention sets 5%-10% of the step signal voltage amplitude emitted by the signal generator as the fault detection threshold. The fault is determined based on the preset fault detection threshold and the signal amplitude. When the signal amplitude is greater than the fault detection threshold, a fault occurs. The fault distance L is then calculated. After calculating the fault distance, the location of the fault point can be determined, and the fault point can be located and repaired.

[0090] The time it takes for the reflected wave to exceed the threshold is called the reflection delay, and the time it takes for the signal to generate a fast rising edge is called the incident delay. Subtracting the incident delay from the reflection delay gives the signal delay as it passes through the photovoltaic array and is reflected back.

[0091] like Figure 4 As shown, the reflection waveform after noise reduction has obvious amplitude variation characteristics. The fault-free section of the photovoltaic array is represented by an amplitude of around 0 in the image. When a fault occurs, the signal amplitude drops rapidly and exceeds the detection threshold to determine that a fault has occurred.

[0092] Specifically, the time at which the signal generates a fast rising edge (1) is the incident delay, and the time at which the signal intersects with the threshold is the reflection delay, such as the breakpoint (2) between the DC cable and the photovoltaic module, or the interruption fault (3) between modules 1 and 2. The dashed line (4) represents the detection threshold. The signal delay is used to determine the fault distance through a delay estimation algorithm, which is expressed as follows: In the formula, v is the propagation speed of the signal in the photovoltaic array, and t is the signal delay.

[0093] Specifically, the propagation speed v of the signal in the photovoltaic array is determined by calibration measurement. A standard photovoltaic module and a photovoltaic DC cable of a given length, such as a 10m long photovoltaic DC cable, are selected. The signal is injected into the photovoltaic module and the time delay of the signal within it is measured. The location of any point is achieved based on the detected signal time delay and the arrangement of the photovoltaic array.

[0094] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for fault detection and location of photovoltaic arrays based on time-domain reflectometry, characterized in that: The method includes the following steps: Step 1, Connect the detection device: Connect an external detection device to the photovoltaic array under test to transmit and collect signals; Step 2, Signal transmission: The step signal emitted by the detection device is injected into the photovoltaic array under test through one of the interfaces of the T-type BNC tee connector; Step 3, acquire waveforms under fault-free conditions: Adjust the sampling interval according to the size of the array under test, and acquire a set of photovoltaic array reflection signal waveforms under fault-free conditions; Step 4, acquire waveforms under fault conditions: adjust the sampling interval according to the size of the array under test, and obtain the reflected signal waveform within the fault detection range through the signal acquisition device; Step 5, Denoising: The wavelet threshold denoising algorithm is used to denoise the reflected signal waveforms in Step 3 and Step 4 respectively to obtain the denoised reflected signal waveforms. Step 6, obtain the signal amplitude: After denoising, subtract the sampling sequence of the photovoltaic array reflected signal waveform under fault conditions from the sampling sequence of the photovoltaic array reflected signal waveform under fault-free conditions to obtain the signal amplitude; Step 7, Fault Determination and Location: Set a preset fault detection threshold. Determine whether a fault has occurred based on the preset fault detection threshold and the signal amplitude obtained in Step 6. If a fault has occurred, calculate the fault distance and locate the fault point through the fault distance.

2. The photovoltaic array fault detection and location method based on time-domain reflectometry according to claim 1, characterized in that: The detection device includes a signal acquisition device, a signal generation device, and a T-type BNC tee connector. The positive terminals of the signal generation device and the signal acquisition device are respectively connected to any two interfaces of the T-type BNC tee connector. The positive terminal of the photovoltaic array is connected to the remaining interface of the T-type BNC tee connector. The negative terminals of the signal acquisition device and the signal generation device are connected to the photovoltaic array ground wire or the photovoltaic array metal frame, and the negative terminals share a common ground.

3. The photovoltaic array fault detection and location method based on time-domain reflectometry according to claim 2, characterized in that: The signal generating device is a square wave signal generator that generates a step signal with a rise time of less than 10ns, a voltage amplitude of less than 5V, and a frequency of less than 100KHz.

4. The photovoltaic array fault detection and location method based on time-domain reflectometry according to claim 2, characterized in that: The signal acquisition device uses a sampling rate greater than 50 MSPS.

5. The photovoltaic array fault detection and location method based on time-domain reflectometry according to claim 1, characterized in that: The noise reduction process in step 4 includes the following steps: Step 41: Perform multi-level wavelet decomposition on the original signal using the DB4 wavelet basis, and extract the approximate coefficient vector A=[a1,a2,a3,……,a…] for each level. n Extract the detail coefficient vector D=[d1,d2,d3,……,d] from each layer. n ]; Step 42: Denoise the detail coefficients containing noisy signals, removing at least the first 3 layers of detail coefficients; Step 43: Perform wavelet reconstruction on the denoised wavelet coefficients, using the approximation coefficient a. n and detail coefficient d n The waveform of the reflected signal after denoising is obtained.

6. The photovoltaic array fault detection and location method based on time-domain reflectometry according to claim 1, characterized in that: The method for calculating the signal amplitude in step 6 is as follows: N sub,i =A si -A Ni In the formula A si and A Ni These are the amplitudes of the reflected signals at the i-th sampling point of the fault-free system and the faulty system, respectively.

7. The photovoltaic array fault detection and location method based on time-domain reflectometry according to claim 1, characterized in that: In step 7, the preset fault detection threshold is set to 5%-10% of the step signal voltage amplitude. When the signal amplitude obtained in step 6 is greater than the fault detection threshold, a fault is determined to have occurred.

8. The photovoltaic array fault detection and location method based on time-domain reflectometry according to claim 1, characterized in that: The fault distance in step 7 is calculated using the following method: (1) The time when the signal rises rapidly is the incident delay. t 0 The time it takes for the reflected wave to exceed the threshold is called the reflection delay. t 1 and the speed of signal propagation in the photovoltaic array v ; (2) The signal delay after passing through the photovoltaic array and being reflected back is calculated. t ,in, t= t 1 - t 0 ; (3) Calculate the fault distance of the fault point L The fault distance L is calculated using the following formula: .

9. A photovoltaic array fault detection and location method based on time-domain reflectometry according to claim 1, characterized in that: The fault distance in step 7 is calculated using the following method: (1) Calculate the total signal delay t The time it takes for the signal to rise rapidly is the incident delay. t 0 The time it takes for the reflected wave to exceed the threshold is called the reflection delay. t 1 The signal delay after passing through the photovoltaic array and being reflected back was calculated. t， in t= t 1 - t 0 ; (2) Determine the time delay corresponding to a single component t d Select a standard photovoltaic module, inject a step signal into the standard photovoltaic module, and measure the corresponding time delay through calibration. This time delay corresponds to a single module. t d ; (3) Calculate the fault distance n The fault distance n is calculated using the following formula to determine the location of the fault point. .

Images