A method and system for arc fault traveling wave location in a string inverter

Abstract

The application discloses an arc fault traveling wave positioning method and system of a group string inverter, and relates to the technical field of power transmission and transformation monitoring. The arrival time of a traveling wave is detected by a sensor installed at the entrance of each group string, and a fault group string branch is identified. The traveling wave signal is separated from the main channel signal of the fault group string branch, and the wave front steepness of the initial traveling wave is calculated. According to a pre-labeled mapping curve, the wave front steepness is mapped to a fault distance, and a candidate reflected wave sequence meeting the fault distance is screened. The polarity, wave head energy attenuation factor and arrival time of each candidate reflected wave are reasonable, and the fault point first reflected wave, the opposite end reflected wave and the crosstalk reflected wave are distinguished. According to the initial wave head time, the first arrival time of the fault point first reflected wave and the second arrival time of the opposite end reflected wave, the fault point position is calculated. According to the embodiment of the application, the source of the reflected wave can be reliably identified under the condition of strong noise background and multi-branch crosstalk, and accurate fault point positioning is realized.

Description

Technical Field

[0001] This invention relates to the field of power transmission and transformation monitoring technology, and in particular to a method and system for locating arc faults in string inverters using traveling waves. Background Technology

[0002] In the current string photovoltaic inverter architecture, multiple photovoltaic strings are directly connected in parallel to the inverter's multi-channel maximum power point tracking input, without a separate DC combiner box. String inverters typically have a built-in arc fault detection function, but this function can only determine "whether an arc has occurred in a certain branch," and cannot pinpoint the specific location of the arc within the string.

[0003] The single-ended traveling wave method is a widely used fault location technique. It deploys a sensor at one end of the line and locates the fault by detecting the time difference between the initial traveling wave and the reflected wave from the fault point. However, when the existing single-ended traveling wave method is directly applied to photovoltaic string scenarios, the sequence of reflected waves received by the sensor contains a mixture of reflected waves from the fault point, the other end of the line, branch points, etc., making it difficult to accurately identify which reflected wave is the primary reflection wave from the fault point. Summary of the Invention

[0004] This invention aims to provide a method and system for locating arc faults in string inverters, which can reliably identify the source of reflected waves under conditions of strong noise background and multi-branch crosstalk, and achieve accurate fault location.

[0005] In a first aspect, embodiments of the present invention provide a method for locating traveling waves of an arc fault in a string inverter, comprising: The faulty string branch is identified based on the arrival time of the traveling wave detected by the sensors installed at the inlet of each string in the multi-branch string inverter. Separate the traveling wave signal from the main channel signal of the sensor on the faulty string branch, and calculate the wavefront steepness of the initial traveling wave based on the traveling wave signal; According to the pre-calibrated mapping curve, the wavefront steepness is mapped to the fault distance, and candidate reflected wave sequences that meet the fault distance are selected from the traveling wave signal. By considering the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence, the primary reflected wave at the fault point, the reflected wave at the opposite end, and the crosstalk reflected wave are distinguished. The location of the fault point is calculated based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end.

[0006] As an improvement to the above scheme, the step of identifying the faulty string branch based on the arrival time of the traveling wave detected by sensors installed at the inlet of each string in the multi-branch string inverter includes: The arrival times of traveling waves detected by each sensor are collected in chronological order by sensors installed at the inlet of each string of a multi-branch string inverter. If the arrival times of the first n traveling waves are consistent, then the faulty string branch is determined based on the current drop of the string branch corresponding to the arrival times of the first n traveling waves; where, ; Otherwise, the string branch corresponding to the arrival time of the first traveling wave will be identified as a faulty string branch.

[0007] As an improvement to the above scheme, the step of separating the traveling wave signal from the main channel signal of the sensor on the faulty group branch and calculating the wavefront steepness of the initial traveling wave based on the traveling wave signal includes: Collect background noise signals and obtain the main channel signals of the sensors on the faulty string branch; The background noise signal is separated from the main channel signal to obtain a traveling wave signal; The traveling wave signal is subjected to wavelet transform to obtain a wavelet modulus maxima sequence; Based on the wavelet modulus maxima sequence, the initial amplitude and rise time of the initial traveling wave are extracted, and the wavefront steepness of the initial traveling wave is calculated.

[0008] As an improvement to the above solution, separating the background noise signal from the main channel signal to obtain the traveling wave signal includes: The main channel signal and the background noise signal are input into an adaptive filter. By utilizing the coupling relationship between the background noise and the correlated noise in the main channel, the correlated noise component is adaptively subtracted from the main channel signal to obtain the residual signal after preliminary denoising. An improved fully adaptive noise set empirical mode decomposition or variational mode decomposition is adopted. By utilizing the difference in the distribution of the arc traveling wave signal and the residual noise in the time and frequency domain, the residual noise is removed from the residual signal, and a pure traveling wave signal is obtained.

[0009] As an improvement to the above scheme, the pre-calibration method of the mapping curve includes: Simulate an electric arc fault at a preset distance and record the wavefront steepness corresponding to each preset distance; The wavefront steepness and the preset distance are fitted and mapped to obtain the mapping curve between the wavefront steepness and the distance.

[0010] As an improvement to the above scheme, the step of mapping the wavefront steepness to a fault distance based on a pre-calibrated mapping curve, and filtering candidate reflected wave sequences from the traveling wave signal that satisfy the fault distance, includes: The fault distance is estimated based on the pre-calibrated mapping curve between wavefront steepness and distance and the wavefront steepness of the traveling wave signal. Based on the fault distance and nominal wave velocity, calculate the expected arrival time window of the primary reflected wave at the fault point; Searching for wavelet mode maxima within the expected arrival time window yields candidate reflected wave sequences.

[0011] As an improvement to the above scheme, the step of distinguishing between primary reflection waves, counterpart reflection waves, and crosstalk reflection waves by considering the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflection wave in the candidate reflection wave sequence includes: Based on the initial wavefront amplitude and the reflected wave amplitude, calculate the wavefront energy attenuation factor of each candidate reflected wave in the candidate reflected wave sequence; If the arrival time of each candidate reflected wave is the same as the arrival time of the traveling wave of the faulty string branch, determine the reasonableness of the arrival time of the first reflected wave or the opposite reflected wave at the fault point of the candidate reflected wave. If the candidate reflected wave does not have a reasonable arrival time and the wavefront energy attenuation factor is greater than the expected value, then the candidate reflected wave is marked as a crosstalk reflected wave. If the candidate reflected wave has a reasonable arrival time, positive polarity, and the wavefront energy attenuation factor matches the first distance, then the candidate reflected wave is marked as the first reflected wave of the fault point; wherein, the first distance is the distance from the fault point to the sensor; If the candidate reflected wave has a reasonable arrival time, negative polarity, and the wavefront energy attenuation factor matches the second distance, then the candidate reflected wave is marked as the opposite-end reflected wave; wherein, the second distance is the sum of the distance from the fault point to the opposite end and the distance from the opposite end to the sensor.

[0012] As an improvement to the above scheme, after distinguishing between primary reflected waves, opposite-end reflected waves, and crosstalk reflected waves by considering the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence, the arc fault traveling wave localization method for string inverters further includes: If, at the same time, wavefronts with amplitudes similar to the crosstalk reflected wave appear on sensors in multiple non-faulty string branches, then the crosstalk reflected wave is marked as a verified crosstalk reflected wave.

[0013] As an improvement to the above scheme, the step of calculating the fault location based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end includes: The first arrival time of the primary reflected wave at the fault point and the second arrival time of the reflected wave at the opposite end are identified; Calculate the first time difference between the first arrival time and the initial wavefront time, and calculate the second time difference between the second arrival time and the first arrival time; Based on the total length of the fault string, the first time difference, and the second time difference, the distance from the fault point to the sensor on the branch of the fault string is calculated to obtain the location of the fault point.

[0014] Secondly, embodiments of the present invention also provide an arc fault traveling wave location system for a string inverter, comprising: The fault string branch identification module is used to identify fault string branches based on the arrival time of traveling waves detected by sensors installed at the entrance of each string of a multi-branch string inverter. The wavefront steepness calculation module is used to separate the traveling wave signal from the main channel signal of the sensor on the fault string branch, and calculate the wavefront steepness of the initial traveling wave based on the traveling wave signal. The candidate reflected wave sequence screening module is used to map the wavefront steepness to the fault distance according to the pre-calibrated mapping curve, and to screen candidate reflected wave sequences that meet the fault distance from the traveling wave signal. The reflected wave differentiation module is used to differentiate the primary reflected wave at the fault point, the reflected wave at the opposite end, and the crosstalk reflected wave by the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence. The fault location module is used to calculate the location of the fault point based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end.

[0015] Compared with existing technologies, the present invention discloses a traveling wave localization method and system for arc faults in string inverters. This method identifies the faulty string branch based on the arrival time of the traveling wave detected by sensors installed at the inlet of each string in a multi-branch string inverter. The traveling wave signal is separated from the main channel signal of the sensor on the faulty string branch, and the wavefront steepness of the initial traveling wave is calculated based on the traveling wave signal. The wavefront steepness is mapped to a fault distance according to a pre-calibrated mapping curve, and candidate reflected wave sequences that meet the fault distance are selected from the traveling wave signal. The polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence are used to distinguish between the primary reflected wave at the fault point, the reflected wave at the opposite end, and the crosstalk reflected wave. The fault location is calculated based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end. Using this invention, the source of reflected waves can be reliably identified under strong noise backgrounds and multi-branch crosstalk conditions, achieving accurate fault location. Attached Figure Description

[0016] Figure 1This is a flowchart illustrating the steps of a method for locating an arc fault traveling wave in a string inverter according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of an arc fault traveling wave location system for a string inverter provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and 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.

[0018] In the description and claims, it should be understood that the terms "first," "second," etc., used in the description and claims are only for the purpose of distinguishing the description of the same technical features, and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated, nor necessarily the order of description or chronological order. The terms are interchangeable where appropriate. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature.

[0019] A photovoltaic DC string system consists of multiple photovoltaic modules connected in series. The system voltage level is typically 900V-1200V, and the normal operating current of a single string is approximately 9.59A (the actual current fluctuates between 5A and 15A depending on sunlight). The connection points between the modules within the string and between the modules and the cables are potential locations for series arcing faults.

[0020] In current string photovoltaic (PV) inverter architectures, multiple PV strings are directly connected in parallel to the inverter's multiple MPPT (Maximum Power Point Tracking) inputs, without a separate DC combiner box. Although string inverters typically have built-in arc fault detection (AFCI) functionality, it can only determine whether an arc has occurred in a specific branch, and cannot meet the need for precise fault location. Maintenance personnel still need to check each long string containing dozens of modules one by one, which is inefficient.

[0021] When a series arc occurs, the fault point exhibits high impedance characteristics, and the circuit current drop is only a few percent to a dozen percent of the nominal current (approximately 0.5A-2A). However, the temperature at the center of the arc can reach thousands of degrees Celsius, making it extremely easy to ignite surrounding materials. Since a DC arc has no natural zero-crossing point, once formed, it will continue to burn until the circuit is broken, posing a serious safety hazard.

[0022] Currently, traveling wave fault location technologies in the field of photovoltaic DC arc locating mainly include the double-end traveling wave method and the single-end traveling wave method.

[0023] The dual-end traveling wave method involves placing traveling wave sensors at both ends of a DC line (such as the inverter side and the combiner box side), achieving time synchronization via GPS / BeiDou timing, and calculating the fault location using the time difference in arrival times at both ends of the traveling wave. However, string inverter architectures do not have a combiner box side, therefore it is not suitable for string inverter architectures.

[0024] The single-ended traveling wave method involves deploying a sensor at only one end of the line (such as the DC input of an inverter) and locating the fault by detecting the time difference between the initial traveling wave and the reflected wave at the fault point.

[0025] However, at a string current of 9.59A, the traveling wave signal generated by the series arc has an amplitude of only milliamperes. Inverter switching noise and MPPT high-frequency disturbances (with amplitudes reaching amperes) completely drown out the milliampere-level arc traveling wave signal, making it impossible for traditional filtering methods to effectively extract it. Furthermore, at a string current of 9.59A, the series arc only reduces the current by 0.5A-2A (approximately 4%-15%), preventing traditional overcurrent protection from triggering, and the arc may persist undetected for several hours.

[0026] When a series arc fault occurs in a string, the resulting traveling wave signal not only propagates and reflects within the faulty string but also shunts to other non-faulty MPPT branches at the DC bus node inside the inverter. The traveling wave is transmitted at impedance discontinuities, enters other branches, and returns to the detection end after reflection at their far ends, forming "inter-branch crosstalk reflection waves" mixed in with the reflected wave sequence. Existing single-ended traveling wave methods are only applicable to single-branch models and cannot distinguish between the reflected wave of the branch itself and the crosstalk reflection wave, leading to location failures or extremely large errors.

[0027] Based on the above considerations, this invention provides a method for locating traveling waves of arc faults in string inverters. Please refer to... Figure 1 In this embodiment, the arc fault traveling wave localization method for the string inverter is specifically executed through steps S1 to S5: S1. Identify the faulty string branch based on the arrival time of the traveling wave detected by the sensors installed at the inlet of each string in the multi-branch string inverter. S2. Separate the traveling wave signal from the main channel signal of the sensor on the faulty group branch, and calculate the wavefront steepness of the initial traveling wave based on the traveling wave signal; S3. Based on the pre-calibrated mapping curve, the wavefront steepness is mapped to the fault distance, and candidate reflected wave sequences that meet the fault distance are selected from the traveling wave signal. S4. By considering the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence, distinguish the primary reflected wave at the fault point, the reflected wave at the opposite end, and the crosstalk reflected wave. S5. Calculate the location of the fault point based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end.

[0028] It should be noted that each string input of a multi-branch string inverter, that is, each maximum power point tracking (MPPT) input of the inverter, connects the string to the inverter.

[0029] Preferably, a high-sensitivity sensor is used to capture traveling wave signals, and all sensors synchronously and continuously acquire the DC current signals of their respective branches.

[0030] For example, a high permeability microcrystalline magnetic core with an initial permeability ≥80000 is used. The sensor has 200-500 coil turns, a bandwidth of 10kHz-2MHz, a sampling rate ≥2MS / s, and a sensitivity better than 1mA. It can detect weak traveling wave signals at the milliampere level, and the triggering method is rising edge threshold triggering (the threshold is set to 2%-5% of the normal operating current, about 0.3A-0.7A).

[0031] When an arc fault occurs, a traveling wave propagates from the fault point along the faulty string to both ends. The traveling wave propagating towards the inverter will first reach the sensors connected to the faulty string, so the arrival time of the traveling wave detected by each sensor can identify the branch of the faulty string.

[0032] A traveling wave signal is superimposed on the main channel signal of the sensor on the faulty string branch. It is necessary to use high-pass filter or wavelet transform to separate the relatively weak traveling wave signal from the main channel signal.

[0033] After the initial traveling wave reaches the bus or fault point, it will be reflected / refracted. The fault distance determines the time delay range of the initial traveling wave to propagate to the fault point and return. The search range of reflected waves can be narrowed by using the mapping curve, thereby effectively eliminating a large number of irrelevant wavefronts.

[0034] In the candidate reflected wave sequence of the faulty string branch sensor, three types of wavefronts may appear: the primary reflected wave from the fault point, the reflected wave from the opposite end, and the crosstalk reflected wave. Among them, the primary reflected wave from the fault point and the reflected wave from the opposite end are the target wavefronts to be identified, while the crosstalk reflected wave from other non-faulty branches are the interference wavefronts.

[0035] In the traditional single-ended traveling wave method, the reflected wave sequence received by the sensor contains a mixture of reflected waves from the fault point, the other end of the line, and branch points. Furthermore, since the technical goal of the traditional single-ended traveling wave method is only to identify the faulty branch, it does not distinguish between these different reflected waves.

[0036] In this embodiment of the invention, based on the differences in characteristics between the primary reflected wave, the reflected wave at the opposite end, and the crosstalk reflected wave at the fault point, the real fault reflected wave and the interference wave are comprehensively distinguished from three dimensions: polarity, energy attenuation, and timing rationality, thereby improving the reliability of positioning.

[0037] The time difference between the initial wavefront and the first reflected wave from the fault corresponds to one time the distance to the fault point. The time difference between the initial wavefront and the reflected wave from the opposite end corresponds to a certain functional relationship between the total length and the distance to the fault point. By using the first arrival time of the first reflected wave from the fault point and the second arrival time of the reflected wave from the opposite end, the fault location can be calculated without relying on the wave velocity, thus avoiding the impact of the uncertain wave velocity on the accuracy of fault location calculation.

[0038] In the above scheme, by accurately distinguishing the reflected wave from the fault point of this branch, the reflected wave from the opposite end of this branch, and the crosstalk reflected wave from other branches, the problem of crosstalk reflected waves between branches mixed in the reflected wave sequence received by each branch sensor can be solved. Thus, under strong noise background and multi-branch crosstalk conditions, the source of reflected waves can be reliably identified, and the fault point can be accurately located.

[0039] As a preferred embodiment, step S1, identifying the faulty string branch based on the arrival time of the traveling wave detected by sensors installed at the inlet of each string in the multi-branch string inverter, includes: The arrival times of traveling waves detected by each sensor are collected in chronological order by sensors installed at the inlet of each string of a multi-branch string inverter. If the arrival times of the first n traveling waves are consistent, then the faulty string branch is determined based on the current drop of the string branch corresponding to the arrival times of the first n traveling waves; where, ; Otherwise, the string branch corresponding to the arrival time of the first traveling wave will be identified as a faulty string branch.

[0040] When an arc fault occurs, a traveling wave propagates from the fault point along the faulty string towards both ends. The traveling wave propagating towards the inverter will first reach the sensors connected to the faulty string. Other branch sensors may also detect the traveling wave transmitted through the common bus, but their arrival time is significantly later than the arrival time of the traveling wave in the faulty string branch (typically ≥1 second later). (and the amplitude is relatively small).

[0041] It should be noted that, theoretically, it is impossible for multiple traveling waves to arrive at different sensors simultaneously, because the path length of the wave from the fault point to different MPPT ports is different. However, we can assume an extreme case where multiple sensors detect traveling waves simultaneously. In this case, we can still determine which string branch is the faulty string branch by observing the current drop in each string branch. The principle is that the faulty string branch should have the largest current drop.

[0042] If there is a significant time difference between the arrival times of the traveling waves, the branch where the fault occurs will be the shortest path for the traveling waves to propagate along the string to the sensor. Therefore, the sensor in that branch will be the first to detect the traveling waves. The branch corresponding to the first arrival time is the branch with the shortest traveling wave propagation path, and logically, it must be the branch where the fault occurred.

[0043] In a preferred embodiment, the absolute times at which each sensor detects the traveling wave are compared. Since the arrival times of the traveling waves are all different, the string branch k that detects the traveling wave earliest is identified as the faulty string branch. Furthermore, the initial wavefront time is denoted as... .

[0044] The above scheme only requires comparing a small number of preceding arrival times, without the need for complex analysis of all branch data, and can handle both inconsistent and consistent time scenarios simultaneously. It covers the simultaneous arrival of traveling waves under multiple parallel branches and introduces current reduction to identify faulty string branches under consistent time conditions.

[0045] In a preferred embodiment, step S2 involves separating the traveling wave signal from the main channel signal of the sensor on the faulty string branch, calculating the wavefront steepness of the initial traveling wave based on the traveling wave signal, and performing this step through steps S21-S24. S21. Collect background noise signals and obtain the main channel signals of the sensors on the faulty string branch; S22. Separate the background noise signal from the main channel signal to obtain a traveling wave signal; S23. Perform wavelet transform on the traveling wave signal to detect the wavelet modulus maxima sequence; S24. Based on the wavelet modulus maxima sequence, extract the initial amplitude and rise time of the initial traveling wave, and calculate the wavefront steepness of the initial traveling wave.

[0046] The main channel signal is the raw electrical signal directly acquired by the sensor, which includes traveling wave signal, background noise signal and other interference components.

[0047] In a preferred embodiment of the present invention, background noise signals are acquired in real time through a noise reference channel. To ensure that no traveling wave signals are mixed into the background noise signals, the noise reference channel uses an independent coil that is not electrically coupled to the cable under test and is placed near the inverter.

[0048] For example, the background noise signal includes at least inverter switching noise and MPPT high-frequency scanning pulses.

[0049] Background noise and traveling wave signals are superimposed in the main channel. By acquiring the background noise signal, the background noise signal can be subtracted from the main channel signal, retaining only the traveling wave component caused by the fault, thus avoiding noise interference in subsequent calculations.

[0050] Traveling wave signals exhibit singularity, with wavelet transform modulus maxima displaying significant peaks at abrupt change points. Utilizing the multi-scale analysis capabilities of wavelet transform, abrupt change points in traveling wave signals can be accurately located, forming a sequence of modulus maxima. Furthermore, the initial oscillation point of the initial traveling wave can be automatically identified using this wavelet modulus maxima sequence.

[0051] In some preferred embodiments, the wavefront steepness of the initial traveling wave is expressed as: ; in, This is the initial amplitude value; The rising time.

[0052] In the above scheme, the influence of environmental and system-specific noise on the extraction of traveling wave features is eliminated through background noise separation preprocessing; furthermore, the initial traveling wave is accurately locked through wavelet transform in order to calculate the wavefront steepness that can characterize the propagation distance of the traveling wave.

[0053] Further, preferably, step S22, separating the background noise signal from the main channel signal to obtain a traveling wave signal, includes: The main channel signal and the background noise signal are input into an adaptive filter. By utilizing the coupling relationship between the background noise and the correlated noise in the main channel, the correlated noise component is adaptively subtracted from the main channel signal to obtain the residual signal after preliminary denoising. An improved fully adaptive noise set empirical mode decomposition or variational mode decomposition is adopted. By utilizing the difference in the distribution of the arc traveling wave signal and the residual noise in the time and frequency domain, the residual noise is removed from the residual signal, and a pure traveling wave signal is obtained.

[0054] The main channel signal contains traveling wave signals and noise, while the noise reference channel contains only noise and is correlated with some of the noise components in the main channel (co-source interference). The adaptive filter can dynamically model this coupling correlation and then adaptively subtract the correlated noise components from the main channel, retaining the traveling wave signal and uncorrelated noise.

[0055] In some preferred embodiments, the adaptive filter employs the least mean square (LMS) or recursive least squares (RLS) algorithm to subtract the correlated noise component from the main channel signal, thereby initially improving the signal-to-noise ratio.

[0056] Adaptive filters can only remove noise related to the reference channel, but there is still noise unrelated to the reference channel in the main channel, which is residual noise. By further separating the uncorrelated residual noise in the residual signal that has not been processed by the adaptive filter, a purer traveling wave signal can be obtained.

[0057] The traveling arc signal exhibits sparse spike pulses in the time domain and is concentrated in the frequency domain between 1MHz and 2MHz, while the residual noise exhibits regular periodic pulses. Based on this principle, for the residual signal after cancellation, improved fully adaptive noise ensemble empirical mode decomposition (ICEEMDAN) or variational mode decomposition (VMD) can be used to further separate the traveling arc signal from the residual noise.

[0058] The above scheme employs a two-stage denoising and separation method: first, adaptive filtering removes relevant noise, and then modal decomposition removes residual unrelated noise. Compared to direct high-pass filtering or fixed filters, this method better preserves the temporal morphology and abrupt change characteristics of the traveling wave.

[0059] As a preferred embodiment, the pre-calibration method for the mapping curve includes: Simulate an electric arc fault at a preset distance and record the wavefront steepness corresponding to each preset distance; The wavefront steepness and the preset distance are fitted and mapped to obtain the mapping curve between the wavefront steepness and the distance.

[0060] The wavefront steepness decreases with increasing distance, but this decrease is affected by factors such as line type, insulation medium, and frequency characteristics. By measuring the wavefront steepness at preset distances (such as 50m, 100m, 200m, and 500m), the mapping curve can be made applicable, avoiding errors caused by theoretical calculations.

[0061] During calibration, only a limited number of preset distances can be selected. However, actual faults may occur at any distance. By fitting and establishing a continuous mapping relationship, the steepness at uncalibrated distances can be interpolated or reasonably extrapolated to achieve continuous distance positioning.

[0062] In some preferred embodiments, the wavefront steepness and the preset distance are fitted by exponential decay or a quadratic polynomial.

[0063] In the above scheme, considering that the cable length and aging degree of different strings are different, and the traveling wave propagation speed varies between branches, affecting the positioning accuracy, a pre-calibration method of mapping curve is provided, which can map distance to wavefront steepness without the need for complex physical modeling.

[0064] In a preferred embodiment, step S3, mapping the wavefront steepness to a fault distance according to a pre-calibrated mapping curve, and filtering candidate reflected wave sequences that satisfy the fault distance from the traveling wave signal, includes: The fault distance is estimated based on the pre-calibrated mapping curve between wavefront steepness and distance and the wavefront steepness of the traveling wave signal. Based on the fault distance and nominal wave velocity, calculate the expected arrival time window of the primary reflected wave at the fault point; Searching for wavelet mode maxima within the expected arrival time window yields candidate reflected wave sequences.

[0065] The wavefront steepness monotonically decreases with increasing propagation distance, and the mapping curve provides a quantitative relationship for this decrease. By substituting the wavefront steepness of the traveling wave signal into the mapping curve, the corresponding fault distance can be solved in reverse, serving as a spatial constraint for subsequent reflected wave searches.

[0066] The propagation path of the primary reflected wave from the fault point is "sensor-fault point-sensor", and the total distance is twice the estimated fault distance. Based on the wave speed, this distance can be converted into round-trip time.

[0067] In some preferred embodiments, the expected arrival time window is represented as: ; in, This is the initial wavefront time; For the estimated fault distance; The nominal wave velocity is taken as 1.8 × 10⁸ m / s. The window width is used to account for errors. For example, if the error is ±50 μm, the corresponding window width is approximately ±0.56 μs.

[0068] By utilizing the detected abrupt change points in the wavelet modulus maxima sequence, candidate reflected waves within the expected arrival time window can be quickly extracted, providing a highly relevant candidate reflected wave sequence for subsequent steps.

[0069] In the above scheme, the wavefront steepness is first used to perform coarse ranging, and then the coarse ranging result is converted into a time window for local search. This avoids full-time-domain reflection wave matching of the entire traveling wave signal. By using the expected arrival time window, a large number of temporally unreasonable reflection waves are eliminated.

[0070] As a preferred implementation, step S4, distinguishing between primary reflection waves at the fault point, reflection waves at the opposite end, and crosstalk reflection waves based on the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflection wave in the candidate reflection wave sequence, includes: Based on the initial wavefront amplitude and the reflected wave amplitude, calculate the wavefront energy attenuation factor of each candidate reflected wave in the candidate reflected wave sequence; If the arrival time of each candidate reflected wave is the same as the arrival time of the traveling wave of the faulty string branch, determine the reasonableness of the arrival time of the first reflected wave or the opposite reflected wave at the fault point of the candidate reflected wave. If the candidate reflected wave does not have a reasonable arrival time and the wavefront energy attenuation factor is greater than the expected value, then the candidate reflected wave is marked as a crosstalk reflected wave. If the candidate reflected wave has a reasonable arrival time, positive polarity, and the wavefront energy attenuation factor matches the first distance, then the candidate reflected wave is marked as the first reflected wave of the fault point; wherein, the first distance is the distance from the fault point to the sensor; If the candidate reflected wave has a reasonable arrival time, negative polarity, and the wavefront energy attenuation factor matches the second distance, then the candidate reflected wave is marked as the opposite-end reflected wave; wherein, the second distance is the sum of the distance from the fault point to the opposite end and the distance from the opposite end to the sensor.

[0071] In some preferred embodiments, the wavefront energy attenuation factor is expressed as: ; in, This represents the amplitude of the reflected wave. This is the initial wavefront amplitude.

[0072] For the reflected wave from the fault point in this branch, its propagation path is fault point - sensor, and the distance is... Wavehead energy attenuation factor For the reflected wave at the opposite end of this branch, the propagation path is fault point - opposite end - sensor, distance... (Approximate) Wavehead energy attenuation factor For crosstalk reflections between branches, the propagation path is: fault point - MPPT terminal of the faulty branch - common bus - entrance of the non-faulty branch (sensor location) - far end of the non-faulty branch - reflection back to the sensor. The path length is much longer than the reflection path of this branch and involves transmission loss; therefore, its energy attenuation factor is... It is significantly greater than the expected value of the reflected wave from the same branch at the same distance.

[0073] Arrival time of crosstalk reflected waves between branches and The intervals usually do not meet the expected window of this branch (i.e., they will not fall within the calculated expected arrival time window), and often manifest as the reflected waves of multiple branches being discretely distributed in time.

[0074] In some preferred embodiments, determining the reasonableness of the arrival time of the primary reflection wave or the opposite-end reflection wave of the candidate reflected wave if the arrival time of each candidate reflected wave is equal to the arrival time of the traveling wave of the faulty string branch includes: If the arrival time of the candidate reflected wave falls within the expected arrival time window, then it is determined that the first reflected wave or the opposite reflected wave at the fault point of the candidate reflected wave has a reasonable arrival time. If the arrival time of the candidate reflected wave does not fall within the expected arrival time window, it is determined that the first reflected wave or the opposite reflected wave at the fault point of the candidate reflected wave does not have a reasonable arrival time.

[0075] The inverter's DC-side input capacitor exhibits near-short-circuit characteristics in the traveling wave frequency band (1MHz-2MHz), with a reflection coefficient... Therefore, the traveling wave undergoes negative reflection at the inverter end. The polarity of the reflected wave at the opposite end of this branch (reflected back from the far end of the string, then reflected again at the inverter end before reaching the sensor) is negative.

[0076] The series arc fault point exhibits high impedance (arc resistance can reach several ohms to tens of ohms, far exceeding the characteristic impedance of the cable, which is approximately 50-100Ω), and a high reflection coefficient. Therefore, the polarity of the primary reflected wave (reflected directly from the fault point without being reflected by the inverter) at the fault point of this branch is positive.

[0077] The above scheme distinguishes between primary reflected waves, opposite-end reflected waves, and crosstalk reflected waves at the fault point from three characteristic dimensions: polarity, wavefront energy attenuation factor, and arrival time rationality. This can solve the problem of crosstalk reflected waves between branches mixed in the reflected wave sequence received by each branch sensor.

[0078] In a preferred embodiment, after performing step S4, the arc fault traveling wave localization method for the string inverter further includes: If, at the same time, wavefronts with amplitudes similar to the crosstalk reflected wave appear on sensors in multiple non-faulty string branches, then the crosstalk reflected wave is marked as a verified crosstalk reflected wave.

[0079] In this embodiment of the invention, for a wavefront suspected to be a crosstalk reflection wave, it is checked whether it appears simultaneously on other non-faulty branch sensors. If wavefronts with similar amplitudes are detected on multiple non-faulty branch sensors at the same time, it is further confirmed that it is a crosstalk wave generated by the coupling of a common bus.

[0080] For example, the polarity is positive and the energy decays with The matched wavefront is marked as the first reflected wave from the fault point in this branch, and the arrival time is recorded as... The polarity is negative and the energy decay is related to the total length of the fault string. The matched wavefront is marked as the reflected wave from the opposite end of this branch, and its arrival time is denoted as . Wavefronts that meet the characteristics of crosstalk (excessive energy attenuation, arrival time not meeting the window of this branch, simultaneous occurrence in multiple branches, etc.) are marked as inter-branch crosstalk reflection waves and excluded from subsequent positioning calculations.

[0081] In the above scheme, crosstalk reflected waves were verified by using sensors on non-faulty string branches, ensuring accurate screening of crosstalk reflected waves.

[0082] As a preferred embodiment, step S5, calculating the location of the fault point based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end, includes: The first arrival time of the primary reflected wave at the fault point and the second arrival time of the reflected wave at the opposite end are identified; Calculate the first time difference between the first arrival time and the initial wavefront time, and calculate the second time difference between the second arrival time and the first arrival time; Based on the total length of the fault string, the first time difference, and the second time difference, the distance from the fault point to the sensor on the branch of the fault string is calculated to obtain the location of the fault point.

[0083] In some preferred embodiments, the distance from the fault point to the sensor on the fault string branch is expressed as: ; in, This represents the total length of the fault sequence; This is the initial wavefront time; The first arrival time; This is the second arrival time.

[0084] Preferably, the total length of the fault string is a known value, which can be calculated based on the number of photovoltaic modules and the spacing between the modules in that branch.

[0085] In some cases, if the second arrival time cannot be identified due to excessive energy attenuation of the reflected wave at the other end, and only the first arrival time of the first reflected wave at the fault point is identified, this embodiment of the invention also sets up a fallback strategy for this purpose. In this case, the distance from the fault point to the sensor on the fault string branch is calculated by substituting the wave velocity to obtain the location of the fault point.

[0086] Preferably, the wave velocity substitution method is expressed as: ; in, To correct wave velocity online in real time.

[0087] It should be noted that the online real-time correction wave velocity is obtained by periodically injecting test pulses into the faulty branch and measuring the round-trip time of a known length segment, and is automatically calibrated every 24 hours.

[0088] If the first arrival time cannot be identified, the positioning fails and an alarm "Precise positioning unavailable" is issued, retaining only the identification information of the faulty branch.

[0089] When both the first and second arrival times are successfully identified, a high-confidence result (confidence ≥ 90%) is output. If only the first arrival time is identified, a medium-confidence result is output and marked "suboptimal positioning". If the first arrival time is not identified, no positioning result is output, an alarm "precise positioning unavailable" is issued, and only the fault branch identification information is retained.

[0090] Furthermore, after calculating the location of the fault point, based on the known spacing between photovoltaic modules, d can be converted into the module number where the fault point is located, and presented to the operation and maintenance personnel in an intuitive form such as "near the 3rd module", so that the operation and maintenance personnel can quickly locate the fault.

[0091] In the above scheme, by using the first arrival time of the primary reflected wave at the fault point and the second arrival time of the reflected wave at the opposite end, a location equation that eliminates the traveling wave propagation velocity can be constructed. Therefore, its calculation results are not affected by changes in wave velocity due to environmental factors, aging, etc., which can improve the robustness of fault location.

[0092] MATLAB / Simulink simulations have verified that, under a typical three-channel MPPT string inverter architecture (fault string length 80m, nominal current 9.59A, arc resistance 0.5Ω, signal-to-noise ratio as low as 0dB), the relative error in fault location using this method is less than 3%, and the false rejection rate of crosstalk reflections from non-faulty branches is less than 5%. Compared to the existing single-ended traveling wave method, this method can still reliably identify the source of reflected waves in environments with strong noise and multi-branch crosstalk, significantly improving positioning accuracy and engineering practicality.

[0093] The arc fault traveling wave localization method for string inverters provided in this invention can accurately distinguish between reflected waves from the fault point in the current branch, reflected waves from the opposite end of the current branch, and crosstalk reflected waves from other branches by using features. This can solve the problem of crosstalk reflected waves mixed in the reflected wave sequence received by each branch sensor, and thus reliably identify the source of reflected waves under strong noise background and multi-branch crosstalk conditions, thereby achieving accurate fault point localization.

[0094] This invention provides an arc fault traveling wave location system for a string inverter. Please refer to [link to relevant documentation]. Figure 2The arc fault traveling wave localization system for the string inverter includes a fault string branch identification module 11, a wavefront steepness calculation module 12, a candidate reflected wave sequence screening module 13, a reflected wave differentiation module 14, and a fault point localization module 15, wherein: The fault string branch identification module 11 is used to identify the fault string branch based on the arrival time of the traveling wave detected by the sensor installed at the entrance of each string of the multi-branch string inverter. The wavefront steepness calculation module 12 is used to separate the traveling wave signal from the main channel signal of the sensor on the fault string branch and calculate the wavefront steepness of the initial traveling wave based on the traveling wave signal. The candidate reflected wave sequence screening module 13 is used to map the wavefront steepness to the fault distance according to the pre-calibrated mapping curve, and to screen candidate reflected wave sequences that meet the fault distance from the traveling wave signal. The reflected wave differentiation module 14 is used to differentiate the primary reflected wave at the fault point, the reflected wave at the opposite end, and the crosstalk reflected wave by the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence. The fault location module 15 is used to calculate the location of the fault point based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end.

[0095] In a preferred embodiment, the fault string branch identification module 11 is specifically used for: The arrival times of traveling waves detected by each sensor are collected in chronological order by sensors installed at the inlet of each string of a multi-branch string inverter. If the arrival times of the first n traveling waves are consistent, then the faulty string branch is determined based on the current drop of the string branch corresponding to the arrival times of the first n traveling waves; where, ; Otherwise, the string branch corresponding to the arrival time of the first traveling wave will be identified as a faulty string branch.

[0096] In a preferred embodiment, the wavefront steepness calculation module 12 is specifically used for: Collect background noise signals and obtain the main channel signals of the sensors on the faulty string branch; The background noise signal is separated from the main channel signal to obtain a traveling wave signal; The traveling wave signal is subjected to wavelet transform to obtain a wavelet modulus maxima sequence; Based on the wavelet modulus maxima sequence, the initial amplitude and rise time of the initial traveling wave are extracted, and the wavefront steepness of the initial traveling wave is calculated.

[0097] Further, preferably, separating the background noise signal from the main channel signal to obtain the traveling wave signal includes: The main channel signal and the background noise signal are input into an adaptive filter. By utilizing the coupling relationship between the background noise and the correlated noise in the main channel, the correlated noise component is adaptively subtracted from the main channel signal to obtain the residual signal after preliminary denoising. An improved fully adaptive noise set empirical mode decomposition or variational mode decomposition is adopted. By utilizing the difference in the distribution of the arc traveling wave signal and the residual noise in the time and frequency domain, the residual noise is removed from the residual signal, and a pure traveling wave signal is obtained.

[0098] As a preferred embodiment, the pre-calibration method for the mapping curve includes: Simulate an electric arc fault at a preset distance and record the wavefront steepness corresponding to each preset distance; The wavefront steepness and the preset distance are fitted and mapped to obtain the mapping curve between the wavefront steepness and the distance.

[0099] In a preferred embodiment, the candidate reflected wave sequence screening module 13 is specifically used for: The fault distance is estimated based on the pre-calibrated mapping curve between wavefront steepness and distance and the wavefront steepness of the traveling wave signal. Based on the fault distance and nominal wave velocity, calculate the expected arrival time window of the primary reflected wave at the fault point; Searching for wavelet mode maxima within the expected arrival time window yields candidate reflected wave sequences.

[0100] In a preferred embodiment, the reflected wave differentiation module 14 is specifically used for: Based on the initial wavefront amplitude and the reflected wave amplitude, calculate the wavefront energy attenuation factor of each candidate reflected wave in the candidate reflected wave sequence; If the arrival time of each candidate reflected wave is the same as the arrival time of the traveling wave of the faulty string branch, determine the reasonableness of the arrival time of the first reflected wave or the opposite reflected wave at the fault point of the candidate reflected wave. If the candidate reflected wave does not have a reasonable arrival time and the wavefront energy attenuation factor is greater than the expected value, then the candidate reflected wave is marked as a crosstalk reflected wave. If the candidate reflected wave has a reasonable arrival time, positive polarity, and the wavefront energy attenuation factor matches the first distance, then the candidate reflected wave is marked as the first reflected wave of the fault point; wherein, the first distance is the distance from the fault point to the sensor; If the candidate reflected wave has a reasonable arrival time, negative polarity, and the wavefront energy attenuation factor matches the second distance, then the candidate reflected wave is marked as the opposite-end reflected wave; wherein, the second distance is the sum of the distance from the fault point to the opposite end and the distance from the opposite end to the sensor.

[0101] In a preferred embodiment, the arc fault traveling wave location system of the string inverter further includes a crosstalk reflection wave verification module, which responds to the completion command of the reflection wave differentiation module 14. The crosstalk reflection wave verification module is used for: If, at the same time, wavefronts with amplitudes similar to the crosstalk reflected wave appear on sensors in multiple non-faulty string branches, then the crosstalk reflected wave is marked as a verified crosstalk reflected wave.

[0102] In a preferred embodiment, the fault location module 15 is specifically used for: The first arrival time of the primary reflected wave at the fault point and the second arrival time of the reflected wave at the opposite end are identified; Calculate the first time difference between the first arrival time and the initial wavefront time, and calculate the second time difference between the second arrival time and the first arrival time; Based on the total length of the fault string, the first time difference, and the second time difference, the distance from the fault point to the sensor on the branch of the fault string is calculated to obtain the location of the fault point.

[0103] The arc fault traveling wave localization system for a string inverter provided in this invention can accurately distinguish between reflected waves from the fault point in the current branch, reflected waves from the opposite end of the current branch, and crosstalk reflected waves from other branches by using features. This solves the problem of crosstalk reflected waves mixed in the reflected wave sequence received by each branch sensor, and thus reliably identifies the source of reflected waves under strong noise background and multi-branch crosstalk conditions, achieving accurate fault point localization.

[0104] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0105] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for locating arc fault traveling wave in a string inverter, characterized in that, include: The faulty string branch is identified based on the arrival time of the traveling wave detected by the sensors installed at the inlet of each string in the multi-branch string inverter. Separate the traveling wave signal from the main channel signal of the sensor on the faulty string branch, and calculate the wavefront steepness of the initial traveling wave based on the traveling wave signal; According to the pre-calibrated mapping curve, the wavefront steepness is mapped to the fault distance, and candidate reflected wave sequences that meet the fault distance are selected from the traveling wave signal. By considering the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence, the primary reflected wave at the fault point, the reflected wave at the opposite end, and the crosstalk reflected wave are distinguished. The location of the fault point is calculated based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end.

2. The arc fault traveling wave localization method for a string inverter as described in claim 1, characterized in that, The method of identifying faulty string branches based on the arrival time of traveling waves detected by sensors installed at the inlet of each string in a multi-branch string inverter includes: The arrival times of traveling waves detected by each sensor are collected in chronological order by sensors installed at the inlet of each string of a multi-branch string inverter. If the arrival times of the first n traveling waves are consistent, then the faulty string branch is determined based on the current drop of the string branch corresponding to the arrival times of the first n traveling waves; where, ; Otherwise, the string branch corresponding to the arrival time of the first traveling wave will be identified as a faulty string branch.

3. The arc fault traveling wave localization method for a string inverter as described in claim 1, characterized in that, The step of separating the traveling wave signal from the main channel signal of the sensor on the faulty string branch and calculating the wavefront steepness of the initial traveling wave based on the traveling wave signal includes: Collect background noise signals and obtain the main channel signals of the sensors on the faulty string branch; The background noise signal is separated from the main channel signal to obtain a traveling wave signal; The traveling wave signal is subjected to wavelet transform to obtain a wavelet modulus maxima sequence; Based on the wavelet modulus maxima sequence, the initial amplitude and rise time of the initial traveling wave are extracted, and the wavefront steepness of the initial traveling wave is calculated.

4. The arc fault traveling wave localization method for a string inverter as described in claim 3, characterized in that, The step of separating the background noise signal from the main channel signal to obtain the traveling wave signal includes: The main channel signal and the background noise signal are input into an adaptive filter. By utilizing the coupling relationship between the background noise and the correlated noise in the main channel, the correlated noise component is adaptively subtracted from the main channel signal to obtain the residual signal after preliminary denoising. An improved fully adaptive noise set empirical mode decomposition or variational mode decomposition is adopted. By utilizing the difference in the distribution of the arc traveling wave signal and the residual noise in the time and frequency domain, the residual noise is removed from the residual signal, and a pure traveling wave signal is obtained.

5. The arc fault traveling wave localization method for a string inverter as described in claim 1, characterized in that, The pre-calibration method for the mapping curve includes: Simulate an electric arc fault at a preset distance and record the wavefront steepness corresponding to each preset distance; The wavefront steepness and the preset distance are fitted and mapped to obtain the mapping curve between the wavefront steepness and the distance.

6. The arc fault traveling wave localization method for a string inverter as described in claim 1 or 5, characterized in that, The step of mapping the wavefront steepness to a fault distance based on a pre-calibrated mapping curve, and then filtering candidate reflected wave sequences from the traveling wave signal that satisfy the fault distance, includes: The fault distance is estimated based on the pre-calibrated mapping curve between wavefront steepness and distance and the wavefront steepness of the traveling wave signal. Based on the fault distance and nominal wave velocity, calculate the expected arrival time window of the primary reflected wave at the fault point; Searching for wavelet mode maxima within the expected arrival time window yields candidate reflected wave sequences.

7. The arc fault traveling wave localization method for a string inverter as described in claim 1, characterized in that, The process of distinguishing between primary reflections at the fault point, reflections at the opposite end, and crosstalk reflections by considering the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflection in the candidate reflection sequence includes: Based on the initial wavefront amplitude and the reflected wave amplitude, calculate the wavefront energy attenuation factor of each candidate reflected wave in the candidate reflected wave sequence; If the arrival time of each candidate reflected wave is the same as the arrival time of the traveling wave of the faulty string branch, determine the reasonableness of the arrival time of the first reflected wave or the opposite reflected wave at the fault point of the candidate reflected wave. If the candidate reflected wave does not have a reasonable arrival time and the wavefront energy attenuation factor is greater than the expected value, then the candidate reflected wave is marked as a crosstalk reflected wave. If the candidate reflected wave has a reasonable arrival time, positive polarity, and the wavefront energy attenuation factor matches the first distance, then the candidate reflected wave is marked as the first reflected wave of the fault point; wherein, the first distance is the distance from the fault point to the sensor; If the candidate reflected wave has a reasonable arrival time, negative polarity, and the wavefront energy attenuation factor matches the second distance, then the candidate reflected wave is marked as the opposite-end reflected wave; wherein, the second distance is the sum of the distance from the fault point to the opposite end and the distance from the opposite end to the sensor.

8. The arc fault traveling wave localization method for a string inverter as described in claim 1, characterized in that, After distinguishing between primary reflected waves, opposite-end reflected waves, and crosstalk reflected waves by considering the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence, the arc fault traveling wave localization method for string inverters further includes: If, at the same time, wavefronts with amplitudes similar to the crosstalk reflected wave appear on sensors in multiple non-faulty string branches, then the crosstalk reflected wave is marked as a verified crosstalk reflected wave.

9. The arc fault traveling wave localization method for a string inverter as described in claim 1, characterized in that, The step of calculating the location of the fault point based on the initial wavefront time, the first arrival time of the primary reflected wave from the fault point, and the second arrival time of the reflected wave from the opposite end includes: The first arrival time of the primary reflected wave at the fault point and the second arrival time of the reflected wave at the opposite end are identified; Calculate the first time difference between the first arrival time and the initial wavefront time, and calculate the second time difference between the second arrival time and the first arrival time; Based on the total length of the fault string, the first time difference, and the second time difference, the distance from the fault point to the sensor on the branch of the fault string is calculated to obtain the location of the fault point.

10. A traveling wave location system for arc faults in a string inverter, characterized in that, include: The fault string branch identification module is used to identify fault string branches based on the arrival time of traveling waves detected by sensors installed at the entrance of each string of a multi-branch string inverter. The wavefront steepness calculation module is used to separate the traveling wave signal from the main channel signal of the sensor on the fault string branch, and calculate the wavefront steepness of the initial traveling wave based on the traveling wave signal. The candidate reflected wave sequence screening module is used to map the wavefront steepness to the fault distance according to the pre-calibrated mapping curve, and to screen candidate reflected wave sequences that meet the fault distance from the traveling wave signal. The reflected wave differentiation module is used to differentiate the primary reflected wave at the fault point, the reflected wave at the opposite end, and the crosstalk reflected wave by the polarity, wavefront energy attenuation factor, and arrival time rationality of each candidate reflected wave in the candidate reflected wave sequence. The fault location module is used to calculate the location of the fault point based on the initial wavefront time, the first arrival time of the primary reflected wave at the fault point, and the second arrival time of the reflected wave at the opposite end.

Images