Method for verifying effectiveness of direct-current arc current signal measurement

Abstract

A DC arc current signal measurement validity verification method mainly comprises the following steps: collecting a DC arc current signal which does not contain inverter switching noise and / or is aliasing with the inverter switching noise, and carrying out analog-to-digital conversion on the DC arc current signal; time windows are set, a moving average method is adopted to calculate the current moving average value of the direct-current arc current signal with the influence of the initial transient effect removed in each time window, the current moving average value serves as the normal distribution mu under each window, and the normal distribution standard deviation sigma is further calculated; setting a dynamic threshold value as k sigma; setting an upper boundary of a threshold range to be a sum obtained by adding the current moving average value mu to the dynamic threshold value k sigma in each time window and setting a lower boundary to be a difference obtained by subtracting the dynamic threshold value k sigma from the current moving average value mu in each time window by using positive and negative deviations of the current moving average value mu and the dynamic threshold value k sigma; and marking the point corresponding to the instantaneous current value exceeding the threshold boundary as a potential fault point.

Description

Technical Field

[0001] The present invention relates to the technical field of power transmission and transformation equipment state measurement, and in particular to a method for verifying the effectiveness of direct current arc current signal measurement. Background Art

[0002] In DC systems, DC discharges can occur for various reasons, influenced by environmental and operating conditions. Arc faults in DC systems are a serious problem. Once a DC discharge develops into an arc, it is difficult to extinguish because the current does not naturally cross zero.

[0003] Numerous manufacturers have begun developing DC arc detection devices, but validating their effectiveness still faces a series of challenges. These validation methods fail to consider the actual operating conditions of DC systems and fail to fully account for the interference of switching noise amplitude and frequency on the detection device, resulting in inaccurate testing. Therefore, a more comprehensive, integrated, and practical validation approach is necessary to ensure greater accuracy and reliability.

[0004] The information disclosed in this Background section is only for enhancement of understanding of the background of the invention and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0005] To address the shortcomings and drawbacks of the existing technologies, a method for verifying the effectiveness of DC arc current signal measurements is provided. This method simulates DC system inverter switching noise by modifying the output signal characteristics of a signal generator. The method is simple, easy to implement, and inexpensive. Furthermore, the noise signal can be easily modified, resulting in highly accurate verification results.

[0006] The purpose of the present invention is achieved through the following technical solutions.

[0007] A method for verifying the effectiveness of DC arc current signal measurement includes the following steps:

[0008] Collecting a DC arc current signal that does not contain inverter switching noise and / or is aliased with inverter switching noise, and performing analog-to-digital conversion on the signal;

[0009] Preprocessing the DC arc current signal after analog-to-digital conversion to obtain a DC arc current signal with the influence of initial transient effects removed;

[0010] Based on the premise that signal fluctuations obey the normal distribution, μ is defined as the mean value and σ is the standard deviation. According to the 3σ principle, 68.27% of the data fall within the range of μ±σ, 95.45% of the data fall within the range of μ±2σ, and 99.73% of the data fall within the range of μ±3σ. Therefore, the dynamic threshold range is defined as μ±kσ, where k is the threshold factor, and k takes the value of 2 or 3.

[0011] Setting a time window, using a moving average method to calculate the current moving average of the DC arc current signal within each time window after removing the influence of the initial transient effect, and using this as μ of the normal distribution under each window, and further calculating the standard deviation σ of the normal distribution; setting a dynamic threshold value as kσ;

[0012] Using the positive and negative deviations of the current moving average μ and the dynamic threshold kσ, the upper boundary of the threshold range is set as the sum of the current moving average μ in each time window plus the dynamic threshold kσ, and the lower boundary is set as the difference between the current moving average μ in each time window and the dynamic threshold kσ.

[0013] For points corresponding to instantaneous current values ​​exceeding the threshold boundary, they are marked as potential fault points.

[0014] Preferably,

[0015] The threshold factor k is 2.

[0016] Preferably, the method further comprises the following steps:

[0017] Based on a minimum fault duration corresponding to the inverter preset from multiple experiments, transient interference areas with a duration less than the minimum fault duration are filtered out, and other areas are retained as potential valid fault areas;

[0018] The total number of all potential fault points in the potential effective fault area is counted and compared with the fault point number threshold preset based on multiple experiments under the minimum fault duration. If the total number of points exceeds the fault point number threshold, it is determined that a fault exists; otherwise, it is determined that there is no fault.

[0019] Preferably, the method further comprises the following steps:

[0020] Dynamically adjust the noise signal to simulate the different switching frequencies and waveform characteristics of different inverters. Simultaneously change the output frequency and waveform type of the signal generator, and the gain of the power amplifier to simulate differences in noise intensity under different load conditions, bus voltages, or inverter power levels.

[0021] Compared with the prior art, the present invention has the following beneficial effects: the present invention is economical and low-cost; the previous method requires constant replacement of the inverter to adjust the switching noise characteristics; the present method can simulate the noise characteristics by changing the output signal of the signal generator.

[0022] The above description is only an overview of the technical solution of the present invention. In order to make the technical means of the present invention clearer and easier to understand, so that those skilled in the art can implement it according to the contents of the description, and in order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are illustrated below. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Various other advantages and benefits of the present invention will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiments below. The accompanying drawings are intended only to illustrate preferred embodiments and are not to be construed as limiting the present invention. It should be understood that the drawings described below are merely examples of the present invention, and that those skilled in the art will be able to derive other drawings from these drawings without inventive effort. Throughout the drawings, identical reference numerals are used to denote identical components.

[0024] In the attached figure:

[0025] Figure 1 1 is a flow chart of a method for verifying the effectiveness of DC arc current signal measurement disclosed in the present invention in one embodiment;

[0026] Figure 2 1 is a schematic diagram of the overall structure of a device used to perform the method in another embodiment, wherein the device mainly includes three circuits, namely a switching noise simulation circuit, an arc circuit, and a sensing circuit;

[0027] Figure 3 The figure shows a simulation signal without inverter switching noise. When there is no inverter switching noise, the arc current waveform is clean, smooth, and has clear edges, which can more realistically reflect the current characteristics of the DC arc itself.

[0028] Figure 4 Figure 1 is a schematic diagram of a simulated signal with inverter switching noise. When inverter switching noise is present, the arc current waveform is distorted, with high-frequency oscillations, glitches, and spikes superimposed.

[0029] Figure 5 This is a schematic diagram of the simulated inverter switching noise signal. The simulated inverter switching noise has a frequency of 12kHz and a peak-to-peak value of 200mA.

[0030] Figure 6 Figure 1 is a schematic diagram of the actual system inverter switching noise signal. The waveform of the actual inverter switching noise signal is consistent with that of the simulated inverter switching noise signal.

[0031] Figure 7Schematic diagram of detection of current signal with aliased inverter switching noise of 20kHz frequency, 500mA peak-to-peak and 32kHz frequency, 600mA peak-to-peak.

[0032] The present invention will be further explained below with reference to the accompanying drawings and embodiments. DETAILED DESCRIPTION

[0033] Specific embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although specific embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0034] It should be noted that certain words are used in the specification and claims to refer to specific components. Those skilled in the art should understand that technicians may use different nouns to refer to the same component. This specification and claims do not use the difference in nouns as a way to distinguish components, but use the difference in the functions of the components as the criterion for distinction. As mentioned throughout the specification and claims, "including" or "comprising" is an open term, so it should be interpreted as "including but not limited to". The subsequent description of the specification is a preferred embodiment of the present invention, but the description is based on the general principles of the specification and is not intended to limit the scope of the invention. The scope of protection of the present invention shall be as defined in the attached claims.

[0035] To facilitate understanding of the embodiments of the present invention, several specific embodiments will be further explained below with reference to the accompanying drawings. However, the accompanying drawings do not limit the embodiments of the present invention.

[0036] For better understanding, Figures 1 to 7 As shown, the present invention discloses a method for verifying the effectiveness of DC arc current signal measurement, comprising the following steps:

[0037] Collecting a DC arc current signal that does not contain inverter switching noise and / or is aliased with inverter switching noise (e.g., inverter switching noise with a frequency of 12 kHz and a peak-to-peak value of 200 mA), and performing analog-to-digital conversion on the signal;

[0038] The DC arc current signal after analog-to-digital conversion is preprocessed to obtain a DC arc current signal that removes the influence of initial transient effects; for example, the signal within the initial 5-50ms is ignored, or a sliding window averaging method is applied to eliminate the starting impact;

[0039] Based on the premise that signal fluctuations obey the normal distribution, μ is defined as the mean value and σ is the standard deviation. According to the 3σ principle, 68.27% of the data fall within the range of μ±σ, 95.45% of the data fall within the range of μ±2σ, and 99.73% of the data fall within the range of μ±3σ. Therefore, the dynamic threshold range is defined as μ±kσ, where k is the threshold factor, and k takes the value of 2 or 3.

[0040] For example, according to actual engineering practice, a threshold factor k of 2 is applicable to most cases, so it is preferably 2;

[0041] Setting a time window, using a moving average method to calculate the current moving average of the DC arc current signal within each time window after removing the influence of the initial transient effect, and using this as μ of the normal distribution under each window, and further calculating the standard deviation σ of the normal distribution; setting a dynamic threshold value as kσ;

[0042] Using the positive and negative deviations of the current moving average μ and the dynamic threshold kσ, the upper boundary of the threshold range is set as the sum of the current moving average μ in each time window plus the dynamic threshold kσ, and the lower boundary is set as the difference between the current moving average μ in each time window and the dynamic threshold kσ.

[0043] For points corresponding to instantaneous current values ​​exceeding the threshold boundary, they are marked as potential fault points.

[0044] In another embodiment, to further improve reliability,

[0045] Based on a minimum fault duration (e.g., 1ms) corresponding to the inverter preset from multiple experiments, transient interference areas with a duration less than the minimum fault duration are filtered out, and other areas are retained as potential valid fault areas;

[0046] The total number of all potential fault points in the potential effective fault area is counted and compared with the fault point number threshold preset based on multiple experiments under the minimum fault duration. If the total number of points exceeds the fault point number threshold, it is determined that a fault exists; otherwise, it is determined that there is no fault.

[0047] For example, the DC arc current signal measurement validity verification method disclosed in the present invention is used to detect two DC arc current signals, one with 20kHz frequency and 500mA peak-to-peak inverter switching noise aliased, and the other with 32kHz frequency and 600mA peak-to-peak inverter switching noise aliased. The results are as follows: Figure 7As shown. Both signals actually exhibit faults and are detected as faulty. In practical applications, there are numerous inverter models, each generating different inverter switching noise waveforms, frequencies, and peak-to-peak values. Therefore, when applying the present invention to a wider range of inverters, it is necessary to experimentally obtain different inverter switching noise interference conditions to determine the aforementioned time window, minimum fault duration, threshold factor, and other factors, tailored to each inverter.

[0048] In another embodiment, the following steps are also included:

[0049] Dynamically adjust the noise signal to simulate the different switching frequencies and waveform characteristics of different inverters. Simultaneously change the output frequency and waveform type of the signal generator, and the gain of the power amplifier to simulate differences in noise intensity under different load conditions, bus voltages, or inverter power levels.

[0050] For example, the switching frequency and peak-to-peak noise of inverter models A and B are shown in Table 1: .

[0051] It should be noted that when using a Type A inverter, the signal generator's output frequency was set to 20kHz, the waveform type to square wave, and the power amplifier gain was changed from 1.5 to 3.0 to simulate power variations of a 1.0-3.6kW non-inductive resistive load, a fixed 193.5V bus voltage, and varying noise intensities of 0.5A to 1.0A from a low-power string to a high-power string. When using a Type B inverter, the signal generator's output frequency was set to 32kHz, the waveform type to square wave, and the power amplifier gain was changed from 2.2 to 2.5 to simulate capacitive load resonance characteristics, bus voltage fluctuations of 193.5-450V, and varying noise intensities of 0.6A to 0.7A from a 3.6kW power level.

[0052] In another embodiment,

[0053] 1) Based on the characteristics of the actual inverter switching noise signal, a signal generator is used to generate the inverter switching noise signal and the DC arc current signal when a fault occurs. The inverter switching noise waveform, peak-to-peak value, and frequency parameters are determined based on the characteristics of the actual inverter switching noise signal. The arc generator's stepper motor is controlled to move the electrode rod at a set speed of 1 mm / s. The electrode rod is adjusted to move to 1 mm to generate a stable and repeatable DC arc current signal when a fault occurs.

[0054] In this example, in order to verify the anti-interference and accuracy of arc measurement of a 5kW photovoltaic inverter under full load conditions, based on the typical switching frequency range of the photovoltaic inverter being 10kHz-30kHz and the measured noise current data having a peak-to-peak value of 1% to 5% of the system rated current, the signal generator is adjusted to output a triangle wave with a peak-to-peak value of 200mA and a frequency of 12kHz. Based on signal compatibility, actual working condition coverage and safety margins, the power amplifier gain can be adjusted in the range of 0.1 times to 10 times. For example, the power amplifier gain is set to 6 times, and the stepper motor of the arc generator is controlled to move the electrode rod at a set moving speed of 1mm / s, and the electrode rod is adjusted to move to 1mm to generate a stable and repeatable DC arc current signal during a fault. According to the parameters, the signal generator is adjusted to output a simulated inverter switching noise signal, such as Figure 5 The actual system inverter switching noise signal is as shown in Figure 6 shown.

[0055] 2) Start the oscilloscope acquisition function and record the DC arc current signal waveform mixed with inverter switching noise output by the sensor (for example, after converting voltage to current through a non-inductive resistive load and then passing through a current sensor);

[0056] For example, the wires of the arc circuit and the switching noise simulation circuit are simultaneously passed through the through-type current sensor, the sampling frequency of the oscilloscope is set to 100MS / s and the vertical / horizontal scale is set appropriately, the oscilloscope acquisition function is started, and the DC arc current signal waveform output by the sensor and aliased with the inverter switching noise is recorded, such as Figure 2 shown.

[0057] 3) Verify the measurement validity: Input the DC arc current signal mixed with inverter switching noise and the DC arc current signal without inverter switching noise into the arc fault detection algorithm respectively, and obtain two judgment results corresponding to the two DC arc current signals. Compare the two judgment results respectively to verify the validity of the DC arc current signal measurement.

[0058] In another embodiment, see Figure 2 The present invention discloses a device for performing a method for verifying the effectiveness of DC arc current signal measurement, comprising:

[0059] An arc circuit for generating a DC arc current signal, wherein the arc circuit includes a DC power supply forming a first series circuit, a non-inductive resistive load R1, and an arc generator; the arc generator includes an adjustable gap, a stepper motor, and two electrode rods, wherein the two electrode rods include a fixed electrode rod and a movable electrode rod; the stepper motor is controlled to move the movable electrode rod so as to generate an arc signal between the two electrode rods; wherein each time the movable electrode rod is moved, the adjustable gap between the movable electrode rod and the fixed electrode rod changes accordingly;

[0060] a switching noise simulation circuit for simulating high-frequency switching noise generated by an inverter, the switching noise simulation circuit comprising a power amplifier and a non-inductive resistive load R2 forming a second series circuit, and a signal generator connected to the power amplifier, wherein the signal generator outputs an inverter switching noise signal of a set waveform, frequency, and peak-to-peak value, which is amplified by the power amplifier and then loaded into the non-inductive resistive load; wherein the second series circuit and the first series circuit do not form a conductive circuit and are independent of each other;

[0061] The sensing circuit is used to collect the DC arc current signal mixed with inverter switching noise and record it with an oscilloscope. Exemplarily, the sensing circuit includes a current sensor and an oscilloscope connected to each other. The current sensor non-contactly senses the mixed current between the first series circuit of the arc circuit and the second series circuit of the switching noise simulation circuit, i.e., the DC arc current signal mixed with inverter switching noise.

[0062] It should be noted that

[0063] According to UL1699B, a DC arc generator consists of an adjustable gap, a stepper motor, and an electrode. In this example, the electrode is made of copper, the gap is 1 mm, and the stepper motor's speed is 1 mm / s. The arc generator uses a stepper motor to control the movement of the electrode, generating a DC arc and providing an arc signal for validating the detection algorithm.

[0064] The purpose of using a non-inductive resistive load is to avoid introducing additional inductive impedance at high frequencies and ensure a purely resistive load, thereby realistically simulating the behavior of high-frequency noise on a non-inductive path.

[0065] The signal generator and power amplifier can jointly simulate the high-frequency switching noise signals generated by different types of inverters.

[0066] In another embodiment,

[0067] The switching noise simulation circuit also includes a through-type current sensor and an oscilloscope. The through-type current sensor passes through the wires of the arc circuit and the switching noise simulation circuit at the same time, so that the sensing circuit can collect the DC arc current signal mixed with the inverter switching noise signal.

[0068] In this embodiment, the arc circuit and the switching noise circuit are electrically independent circuits, and there is no direct series or parallel connection between them. The sensing circuit realizes magnetic coupling between the arc circuit and the switching noise simulation circuit through the physical structure of the through-type current sensor. Placing the two key current paths, the arc circuit and the switching noise simulation circuit, simultaneously within the measurement range of the through-type current sensor is the key to the through-type current sensor output reflecting the superposition of the current signals of the two circuits.

[0069] In another embodiment,

[0070] The through-hole current sensor includes a Hall coil or a Rogowski coil.

[0071] In another embodiment,

[0072] The set waveform includes one or more combinations of sine wave, square wave or triangle wave.

[0073] In another embodiment,

[0074] The gain of the power amplifier is adjustable to change the amplitude of the noise signal.

[0075] In another embodiment,

[0076] The electrode rod is made of copper.

[0077] In another embodiment,

[0078] The gap between the two electrode rods is 0.5 to 2 mm, and the moving speed of the stepper motor is 0.1 to 2 mm / s.

[0079] In another embodiment,

[0080] The frequency range of the switching noise signal is 1kHz to 100kHz, and the peak-to-peak range is 100mA to 1A.

[0081] In another embodiment,

[0082] The sampling frequency of the sensor circuit is 100MS / s.

[0083] In another embodiment,

[0084] The method or the device is applied to photovoltaic systems, electric vehicle charging systems and energy storage systems.

[0085] In another embodiment,

[0086] The output terminal of the signal generator is connected to the input terminal of the power amplifier through a coaxial cable, and the output terminal and the ground terminal of the power amplifier are connected to both ends of the non-inductive resistance load through wires to form a closed loop of the noise signal.

[0087] In another embodiment,

[0088] In the arc circuit, the positive terminal of the DC power supply is connected to one end of the non-inductive resistive load via a wire, and the other end is connected to the end of the movable electrode rod of the arc generator via a wire. The end of the fixed electrode rod of the arc generator is connected to the negative terminal of the DC power supply via a wire, thereby forming a complete DC current path (i.e., the first series circuit mentioned above). By controlling the adjustable gap between the front ends of the movable electrode rod and the fixed electrode rod, a DC arc is generated during the movement of the movable electrode rod.

[0089] In another embodiment,

[0090] A through-hole current sensor is used as a current sensor and is connected in series with the oscilloscope.

[0091] Therefore, the DC arc current signal aliased with the inverter switching noise is collected by the current sensor and displayed on the oscilloscope.

[0092] In another embodiment,

[0093] Generate the inverter switching noise signal and DC arc current signal. The inverter switching noise waveform, peak-to-peak value, and frequency parameters are determined. In this example, the peak-to-peak value of the noise waveform is determined to be 200mA and the frequency is determined to be 12kHz. According to the parameters, the signal generator is adjusted to output the simulated inverter switching noise signal, such as Figure 5 The actual system inverter switching noise signal is as shown in Figure 6 As shown. Controlling the electrode movement generates a DC arc current signal. See further Figure 7 , which shows that for DC arc current signals aliased with 20kHz noise and DC arc current signals aliased with 32kHz noise, the method disclosed by the present invention can effectively identify faults.

[0094] Furthermore, this invention simulates real-world arc fault scenarios by precisely controlling the electrode separation process to generate a stable, repeatable DC arc current signal, providing realistic test input for the detection algorithm. The arc generator is designed in accordance with UL1699B, ensuring industry-recognized and standardized experimental results. The stepper motor drive system facilitates adjustment and control, making parameters such as arc generation timing, intensity, and duration controllable, facilitating the construction of a structured test data set. A signal generator outputs a noise signal with a set waveform, frequency, and amplitude; a power amplifier amplifies the noise signal and applies it to a non-inductive resistive load. The signal generator frequency (e.g., 1kHz to 100kHz), peak-to-peak value (e.g., 100mA to 1A), and power amplifier gain are adjustable. This method offers a low-cost alternative to inverter switching noise sources: Traditional methods require replacing different inverter models to obtain different noise characteristics, which is costly and complex. This method, using a signal generator combined with a power amplifier, can flexibly simulate a variety of inverter switching noise characteristics.

[0095] Furthermore, the present invention precisely controls noise parameters, enabling independent adjustment of the noise waveform, frequency, and amplitude to meet the robustness verification requirements for detection algorithms under different operating conditions. This improves test efficiency and flexibility, allowing for quick switching of noise types without physically replacing equipment, shortening test cycles and increasing verification efficiency. Current signals are acquired using a through-hole current sensor (such as a Rogowski coil or Hall effect coil). The arc loop and noise simulation loop conductors are simultaneously passed through the sensor's center. An oscilloscope is used to record the mixed signal waveform. Signal aliasing acquisition is achieved by synchronously acquiring the arc signal and switching noise signal, simulating the current environment in which both coexist in a real system and enhancing test authenticity. Non-invasive measurement: The through-hole sensor does not alter the circuit structure, avoiding the introduction of additional interference and ensuring measurement accuracy. High compatibility: It is compatible with various arc detection algorithm input interfaces, facilitating algorithm integration and verification. The noise amplitude is varied by varying the signal generator's output frequency and waveform and adjusting the power amplifier's gain. Multiple mixed signals are acquired under different noise conditions to comprehensively evaluate algorithm performance. Multiple noise parameter combinations are used to simulate different real-world operating environments and verify the algorithm's stability and accuracy under complex electromagnetic interference. Enhance the credibility of verification results, avoid accidental errors caused by verification under single noise conditions, and improve the scientific nature and reliability of algorithm evaluation. Support algorithm optimization iteration, provide rich test data support for subsequent improvement of detection algorithms, and guide parameter adjustment and model optimization. Input the collected mixed signal into the arc detection algorithm to be verified; record the arc judgment results output by the arc fault detection algorithm; compare the actual arc state with the algorithm judgment, and calculate indicators such as accuracy. Quantitatively evaluate the detection effect. By comparing the actual arc state with the algorithm recognition results, objectively evaluate the algorithm's detection accuracy, false positive rate, missed judgment rate and other key indicators. The closed-loop verification mechanism forms a closed-loop process of "signal input-algorithm processing-result comparison", which facilitates automated testing and batch verification. Support algorithm selection and deployment decisions: Provide data basis for selecting the optimal detection algorithm in engineering applications, and promote the implementation of algorithms in actual systems.

[0096] The basic principles of the present application have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in this application are merely illustrative and not restrictive, and it should not be assumed that these advantages, strengths, and effects are required of each embodiment of this application. In addition, the specific details disclosed above are merely illustrative and facilitating understanding, and are not restrictive. The above details do not limit this application to necessarily being implemented using the above specific details.

[0097] The above description has been provided for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for verifying the effectiveness of DC arc current signal measurement, characterized in that: The steps include: Collecting a DC arc current signal that does not contain inverter switching noise and / or is aliased with inverter switching noise, and performing analog-to-digital conversion on the signal; Preprocessing the DC arc current signal after analog-to-digital conversion to obtain a DC arc current signal with the influence of initial transient effects removed; Based on the premise that signal fluctuations obey the normal distribution, μ is defined as the mean value and σ is the standard deviation. According to the 3σ principle, 68.27% of the data fall within the range of μ±σ, 95.45% of the data fall within the range of μ±2σ, and 99.73% of the data fall within the range of μ±3σ. Therefore, the dynamic threshold range is defined as μ±kσ, where k is the threshold factor, and k takes the value of 2 or 3. Setting a time window, using a moving average method to calculate the current moving average of the DC arc current signal within each time window after removing the influence of the initial transient effect, and using this as μ of the normal distribution under each window, and further calculating the standard deviation σ of the normal distribution; setting a dynamic threshold value as kσ; Using the positive and negative deviations of the current moving average μ and the dynamic threshold kσ, the upper boundary of the threshold range is set as the sum of the current moving average μ in each time window plus the dynamic threshold kσ, and the lower boundary is set as the difference between the current moving average μ in each time window and the dynamic threshold kσ. For points corresponding to instantaneous current values ​​exceeding the threshold boundary, they are marked as potential fault points.

2. The method according to claim 1, characterized in that The threshold factor k is 2.

3. The method according to claim 1, characterized in that The following steps are also included: Based on a minimum fault duration corresponding to the inverter preset from multiple experiments, transient interference areas with a duration less than the minimum fault duration are filtered out, and other areas are retained as potential valid fault areas; The total number of all potential fault points in the potential effective fault area is counted and compared with the fault point number threshold preset based on multiple experiments under the minimum fault duration. If the total number of points exceeds the fault point number threshold, it is determined that a fault exists; otherwise, it is determined that there is no fault.

4. The method according to claim 1, wherein The following steps are also included: Dynamically adjust the noise signal to simulate the different switching frequencies and waveform characteristics of different inverters. Simultaneously change the output frequency and waveform type of the signal generator, and the gain of the power amplifier to simulate differences in noise intensity under different load conditions, bus voltages, or inverter power levels.

Images