A method and system for rapid identification of arc fault in power distribution cabinet

Abstract

The application relates to the technical field of arc detection, and discloses a power distribution cabinet internal arc fault rapid identification method and system. The method comprises the following steps: collecting a dry circuit current signal, branch current signals, an arc light intensity signal and a sound pressure signal, performing harmonic decomposition on the branch current signals to obtain branch harmonic intensity indexes and a dry circuit fundamental wave effective value; performing summation on the branch harmonic intensity indexes and performing division on the dry circuit fundamental wave effective value to obtain a characteristic dilution index D; assigning dynamic weights to each fault indication value according to the size interval of the characteristic dilution index D and performing weighted summation to obtain a comprehensive confidence C; determining a confirmation counter threshold value according to a dilution change rate; when the number of windows in which the comprehensive confidence C continuously exceeds the decision threshold value reaches the confirmation counter threshold value, outputting a stable fault decision result, positioning a fault branch and outputting a tripping instruction. The application improves the accuracy and decision stability of the power distribution cabinet internal arc fault identification under complex load working conditions.

Description

Technical Field

[0001] This application relates to the field of arc detection technology, and in particular to a method and system for rapid identification of arc faults in a power distribution cabinet. Background Technology

[0002] As the core equipment of a power distribution system, the distribution cabinet is highly susceptible to arcing faults due to factors such as insulation aging, loose wiring, and overload short circuits. Existing technologies for identifying arcing faults within distribution cabinets primarily rely on single signal dimensions for judgment. These include methods based on the time-frequency characteristics of high-frequency current components, photoelectric detection methods based on sudden changes in arc intensity, and acoustic detection methods based on sound pressure signal spectrum analysis. These methods extract feature quantities from their respective signal dimensions, compare them with preset thresholds, output fault judgment results, and, upon triggering the judgment, drive the circuit breaker to perform a disconnection action.

[0003] However, existing technologies have significant shortcomings. In actual operation scenarios of distribution cabinets, multiple load branches operate in parallel. The harmonic currents generated by each branch during normal operation are superimposed on the main circuit, causing the time-domain characteristics of the fault arc, such as "zero rest" and amplitude spikes, in the main circuit current waveform to be diluted, and the frequency domain distinguishability to decrease significantly. As a result, the false negative rate of current signal-based identification methods increases dramatically in scenarios with multiple loads in parallel. Although introducing arc light and sound pressure signals can compensate for the deficiencies of current signals to some extent, existing multi-source fusion methods generally use fixed weights to weight and superimpose features of each dimension, without considering the dynamic changes in the representational capabilities of each signal dimension under different load conditions. This results in the distortion features of the current signal still participating in the decision-making at a fixed proportion in high-interference load scenarios, lowering the overall reliability of the fusion identification results. Summary of the Invention

[0004] This application provides a method and system for rapid identification of arc faults in power distribution cabinets, which solves the problems of high false negative rate caused by current feature dilution in multi-load parallel scenarios and decision jump error caused by dynamic weight switching, thereby improving the accuracy and decision stability of arc fault identification in power distribution cabinets under complex load conditions.

[0005] Firstly, this application provides a method for rapid identification of arc faults in a power distribution cabinet, the method comprising:

[0006] Step S1: Collect the main circuit current signal, branch circuit current signal, arc intensity signal and sound pressure signal of the distribution cabinet, and perform harmonic decomposition on the branch circuit current signal to obtain the harmonic intensity index of each branch circuit and the effective value of the fundamental wave of the main circuit.

[0007] Step S2: Sum the harmonic intensity indices of each branch and divide them by the effective value of the fundamental wave of the main road to obtain the characteristic dilution index D;

[0008] Step S3: Based on the size range of the feature dilution index D, assign corresponding dynamic weights to the fault indication values ​​of the arc intensity signal, the sound pressure signal, and the main circuit current signal, respectively, and sum the fault indication values ​​with the corresponding dynamic weights to obtain the comprehensive confidence level C.

[0009] Step S4: Calculate the difference of the characteristic dilution index D between adjacent periods to obtain the dilution change rate. Determine the confirmation counter threshold based on the dilution change rate. When the number of consecutive windows in which the comprehensive confidence C exceeds the decision threshold reaches the confirmation counter threshold, output a stable fault decision result. Locate the faulty branch based on the sorting of the high-frequency energy mutation of the current signals of each branch and output a trip command to the faulty branch.

[0010] Secondly, this application provides a rapid arc fault identification system for power distribution cabinets, the rapid arc fault identification system for power distribution cabinets comprising:

[0011] The acquisition module is used to acquire the main circuit current signal, branch circuit current signals, arc intensity signal and sound pressure signal of the distribution cabinet, and to perform harmonic decomposition on the branch circuit current signals to obtain the harmonic intensity index of each branch circuit and the effective value of the fundamental wave of the main circuit.

[0012] The summation module is used to sum the harmonic intensity indices of each branch and then divide them by the effective value of the fundamental wave of the main branch to obtain the characteristic dilution index D.

[0013] The allocation module is used to assign corresponding dynamic weights to the fault indication values ​​of the arc intensity signal, the sound pressure signal, and the main circuit current signal according to the size range of the feature dilution index D, and to sum the fault indication values ​​with the corresponding dynamic weights to obtain the comprehensive confidence level C.

[0014] The positioning module is used to calculate the difference of the characteristic dilution index D between adjacent periods to obtain the dilution change rate, determine the confirmation counter threshold based on the dilution change rate, and output a stable fault judgment result when the number of continuous windows in which the comprehensive confidence C exceeds the decision threshold reaches the confirmation counter threshold. It also locates the faulty branch based on the sorting of the high-frequency energy mutation of the current signals of each branch and outputs a trip command to the faulty branch.

[0015] Thirdly, a device for rapid identification of arc faults in a power distribution cabinet is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the device for rapid identification of arc faults in a power distribution cabinet to execute the aforementioned method for rapid identification of arc faults in a power distribution cabinet.

[0016] Fourthly, a computer-readable storage medium is provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform the aforementioned method for rapid identification of arc faults in a power distribution cabinet.

[0017] The technical solution provided in this application synchronously acquires the main circuit current signal, branch circuit current signals, arc intensity signal, and sound pressure signal of the distribution cabinet. Harmonic decomposition is then performed on each branch circuit current signal to obtain the harmonic intensity index of each branch and the effective value of the main circuit fundamental wave. This allows the harmonic contribution of each branch to be quantified independently rather than being aggregated, fundamentally avoiding the problem of harmonic masking between branches when analyzing a single main circuit signal. The characteristic dilution index D is obtained by summing the harmonic intensity indices of each branch and dividing it by the effective value of the main circuit fundamental wave. This index transforms the current load topology of the distribution cabinet into a dimensionless quantity that can be updated in real time, enabling the system to have its own current... The ability to quantitatively perceive the reliability of signals is the core difference between this application and existing fixed-weight fusion schemes. Based on the size range of the feature dilution index D, corresponding dynamic weights are assigned to the fault indication values ​​of the arc intensity signal, sound pressure signal, and main circuit current signal. The comprehensive confidence level C is obtained by weighted summation of each fault indication value and the dynamic weight. The dynamic weight mechanism makes the two physical mechanisms of arc and sound pressure independent of the sensing dimension of the current signal. When the current feature is severely diluted, it can automatically obtain higher decision-making power. Thus, without increasing the number of sensors, active suppression of multi-load parallel interference is achieved only by adjusting the weight allocation logic.

[0018] The dilution rate of change is obtained by calculating the difference between the dilution index D of adjacent cycles, and this rate drives the dynamic setting of the confirmation counter threshold. The number of windows in which the comprehensive confidence level C continuously exceeds the decision threshold is compared with the confirmation counter threshold, and a stable fault decision result is output. This mechanism directly links the severity of load state switching with the strictness of decision confirmation. When the load is stable, confirmation is allowed to be completed in the shortest window. When the load is drastically switched, the confirmation window is automatically extended, so that the short-term jump in comprehensive confidence level caused by dynamic weight switching is effectively filtered out, and the output of the decision result is no longer directly affected by the weight switching rate. After completing the stable fault decision, the fault branch is further located by sorting the high-frequency energy mutation of the current signal of each branch and outputting a trip command to it. The identification result is accurately located to the specific circuit rather than just giving a general conclusion that there is an electric arc in the cabinet. This makes the linkage protection action have clear circuit directionality and avoids the expansion of the whole cabinet power outage. Overall, the characteristic dilution index D runs through the entire process from the summation calculation of the harmonic intensity index of each branch to the output of the stable fault judgment result. It is both the driving factor for dynamic weight allocation and the basis for the adjustment of the confirmation counter threshold. It connects the three links of signal credibility assessment, fusion judgment and judgment stability control with a single intermediate quantity, so that the technical characteristics form an internal causal logic chain rather than a simple stacking. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of an embodiment of the method for rapid identification of arc faults in a distribution cabinet in this application.

[0021] Figure 2 This is a schematic diagram illustrating the relationship between the absolute value of the dilution rate of change and the confirmation counter threshold in an embodiment of this application. Detailed Implementation

[0022] This application provides a method and system for rapid identification of arc faults in a power distribution cabinet. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0023] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the method for rapid identification of arc faults in power distribution cabinets in this application includes:

[0024] Step S1: Collect the main circuit current signal, branch circuit current signal, arc intensity signal and sound pressure signal of the distribution cabinet, and perform harmonic decomposition on the branch circuit current signal to obtain the harmonic intensity index of each branch circuit and the effective value of the fundamental wave of the main circuit.

[0025] Specifically, the main circuit current signal and the current signals of each branch circuit are acquired through current transformers with a transformation ratio set to 1000:1 and a sampling frequency set to 10kHz, corresponding to a sampling period of 0.1ms. The three signals are synchronously acquired triggered by a unified clock, with a time alignment error not exceeding 0.1ms. The harmonic intensity index of each branch circuit refers to the quantity obtained by extracting and summing the amplitudes of the 3rd, 5th, and 7th harmonics after performing a Fast Fourier Transform on the current signal of each branch circuit within one power frequency cycle (20ms, corresponding to 200 sampling points). The 3rd, 5th, and 7th harmonics are selected because the nonlinear loads (such as switching power supplies and frequency converters) in the distribution cabinet mainly generate these odd harmonics during normal operation, and the sum of these three harmonics can quantitatively reflect the contribution of the branch circuit to the superposition of harmonics in the main circuit current. The effective value of the main circuit fundamental frequency is obtained by taking the root mean square of the main circuit current signal after low-pass filtering within one power frequency cycle, and is used to characterize the fundamental frequency amplitude level of the main circuit current.

[0026] Step S2: Sum the harmonic intensity indices of each branch and divide them by the effective value of the fundamental wave of the main road to obtain the characteristic dilution index D;

[0027] Specifically, the characteristic dilution index D is defined as the dimensionless ratio obtained by dividing the sum of the harmonic intensity indices of all branches (i.e., the total amount of harmonic superposition in the main circuit) by the effective value of the fundamental wave in the main circuit. The D value is updated every 20ms power frequency cycle. The physical meaning of the D value is as follows: when the number or power of nonlinear loads operating in parallel in the distribution cabinet increases, the harmonic currents generated by the normal operation of each branch are superimposed in the main circuit, increasing the proportion of harmonic components in the main circuit current waveform. At this time, the D value increases, indicating that the main circuit current signal has a reduced ability to characterize the characteristics of fault arcs. Conversely, when the load is mainly linear, the D value is smaller, and the reliability of the main circuit current signal is higher. The segmented thresholds of the D value are set at 0.15 and 0.35. Below 0.15 corresponds to a low-interference state dominated by linear loads, and above 0.35 corresponds to a high-interference state with multiple nonlinear loads in parallel. The state between the two thresholds is a transition state. The above segmentation is determined based on the measured statistical distribution of harmonics under typical load combinations in the distribution cabinet.

[0028] Step S3: Based on the size range of the characteristic dilution index D, assign corresponding dynamic weights to the fault indication values ​​of the arc intensity signal, sound pressure signal and main circuit current signal respectively, and sum the fault indication values ​​with the corresponding dynamic weights to obtain the comprehensive confidence level C.

[0029] Specifically, the fault indication values ​​of the arc intensity signal, sound pressure signal, and main circuit current signal are all binary quantities (with values ​​of 0 or 1), determined based on whether the arc intensity change, the short-time sound pressure energy change, and the high-frequency current energy change exceed their respective decision thresholds. The dynamic weights are selected in real time from three preset weighted sets based on the interval of the D value. All three weighted sets satisfy the normalization constraint that the sum of the three weights is 1. The comprehensive confidence level C is the weighted sum of the three fault indication values ​​and the corresponding dynamic weights, with a value range of 0 to 1. The decision threshold is set to 0.50. This value ensures that at least one of the arc or sound pressure signals must be triggered in a high-dilution state for C to exceed the threshold, avoiding false judgments triggered by a weak response from a single current signal.

[0030] Step S4: Calculate the difference between adjacent periodic characteristic dilution index D to obtain the dilution rate of change. Determine the confirmation counter threshold based on the dilution rate of change. When the number of consecutive windows in which the comprehensive confidence C exceeds the decision threshold reaches the confirmation counter threshold, output a stable fault decision result. Locate the faulty branch based on the sorting of high-frequency energy mutations in the current signals of each branch and output a trip command to the faulty branch.

[0031] Specifically, the dilution rate of change is defined as the difference between the characteristic dilution index D of two adjacent power frequency cycles, reflecting the severity of load state switching. The confirmation counter threshold is set in three levels based on the absolute value of the dilution rate of change: a threshold of 2 (corresponding to a 2ms confirmation window) when the absolute value is not greater than 0.05; a threshold of 5 (corresponding to 5ms) when the absolute value is between 0.05 and 0.15; and a threshold of 10 (corresponding to 10ms) when the absolute value is greater than 0.15. The three thresholds are set based on the following: when the load state is stable, the weight allocation is reliable, allowing for rapid confirmation; when the load is switched on or the motor starts, the D value fluctuates drastically, requiring a longer confirmation window to filter out false judgments caused by weight switching. Fault branch location is achieved by comparing the magnitude of high-frequency energy mutations in each branch, ranking them accordingly. The branch with the largest mutation is identified as the fault branch. When the maximum value does not exceed the location judgment threshold, the fault is determined to be located on the main busbar. Subsequently, a 50ms pulse width tripping command is output to the fault branch, driving the circuit breaker to complete the tripping action.

[0032] In one specific embodiment, step S1 includes:

[0033] Triggered by a unified clock signal, the main circuit current signal, branch circuit current signal, arc intensity signal and sound pressure signal of the power distribution cabinet are collected synchronously.

[0034] The arc light intensity signal is subjected to median filtering to obtain the arc light filtered signal; the sound pressure signal is subjected to bandpass filtering to obtain the sound pressure filtered signal.

[0035] The trunk current signal is low-pass filtered to obtain the trunk fundamental component; the trunk current signal is band-pass filtered to obtain the trunk high-frequency component; the current signals of each branch are band-pass filtered to obtain the high-frequency component of each branch.

[0036] Based on the high-frequency components of each branch, a fast Fourier transform is performed on the current signal of each branch to extract the amplitudes of the 3rd, 5th, and 7th harmonics of the current signal of each branch. The amplitudes of the 3rd, 5th, and 7th harmonics of the same branch are summed to obtain the harmonic intensity index of each branch. Based on the fundamental component of the main circuit, the root mean square of the current signal of the main circuit is taken within one power frequency cycle to obtain the effective value of the fundamental component of the main circuit.

[0037] Specifically, a unified clock signal is used for triggering, with a trigger period set to 0.1ms, corresponding to a sampling frequency of 10kHz. The three sensors complete sampling on the same trigger edge, with a time alignment error not exceeding 0.1ms. The median filter window length for the arc intensity signal is set to 5 sampling points to remove pulse noise generated by occasional interference from external light sources in the arc sensor, while preserving the abrupt change trend of the arc intensity, thus obtaining the arc-filtered signal. The bandpass filter for the sound pressure signal is set to 1kHz to 40kHz, covering the characteristic sound wave frequency range generated by arc discharge. Power frequency mechanical vibrations below 1kHz and ultrasonic noise above 40kHz are filtered out, resulting in the sound pressure filter signal. The low-pass filter cutoff frequency for the main circuit current signal is set to 5kHz, preserving the fundamental and lower harmonic components, thus obtaining the main circuit fundamental component. The bandpass filter for both the main circuit current signal and the branch circuit current signals is set to 2kHz to 40kHz. Within this frequency band, the high-frequency energy is low during normal load operation but significantly increases during fault arcing, resulting in the main circuit high-frequency component and the high-frequency components of each branch circuit, respectively.

[0038] Based on the high-frequency components of each branch, a Fast Fourier Transform (FFT) is performed on the current signal of each branch within one power frequency cycle (20ms, corresponding to 200 sampling points) to extract the amplitudes of the 3rd (150Hz), 5th (250Hz), and 7th (350Hz) harmonics. The selection of these three odd harmonics is based on the fact that nonlinear loads such as switching power supplies and frequency converters in the distribution cabinet mainly inject the 3rd, 5th, and 7th odd harmonics into the main circuit during normal operation. The amplitudes of the 3rd harmonics of the same branch are directly summed to obtain the harmonic intensity index of that branch. This index quantitatively reflects the magnitude of the contribution of that branch to the superposition of harmonics in the main circuit. Based on the fundamental component of the main circuit, the root mean square (RMS) is taken within 200 sampling points in the same power frequency cycle to obtain the effective value of the fundamental component of the main circuit. The RMS operation eliminates the positive and negative symmetry of the fundamental component, ensuring that the effective value of the fundamental component of the main circuit is always positive and monotonically corresponds to the amplitude of the fundamental current of the main circuit.

[0039] In one specific embodiment, in step S2, the summation of the harmonic intensity indices of each branch is divided by the effective value of the fundamental wave of the main branch to obtain the characteristic dilution index D, including:

[0040] The harmonic intensity indices of each branch are summed one by one to obtain the total superposition of harmonics in the main road.

[0041] Based on the total amount of harmonic superposition in the main line and the effective value of the fundamental wave in the main line, the total amount of harmonic superposition in the main line is divided by the effective value of the fundamental wave in the main line to obtain the characteristic dilution index D; wherein, the characteristic dilution index D is updated once per power frequency cycle.

[0042] Specifically, the total harmonic superposition of the main circuit is a scalar value obtained by summing the harmonic intensity indices of all branches obtained in step S1. The summation object is all parallel branches in the distribution cabinet. The harmonic intensity index of each branch is the sum of the amplitudes of the 3rd, 5th, and 7th harmonics of that branch. The total harmonic superposition of the main circuit is the arithmetic sum of the harmonic intensity indices of all branches, reflecting the total harmonic level injected into the main circuit of the distribution cabinet at the current moment by the normal operation of all parallel loads.

[0043] The characteristic dilution index D is obtained by dividing the total harmonic superposition of the main circuit by the effective value of the fundamental frequency of the main circuit, and is a dimensionless ratio. When the number of parallel nonlinear loads or their power increases in the distribution cabinet, the cumulative harmonic current injected into the main circuit from each branch increases, the total harmonic superposition of the main circuit increases, while the effective value of the fundamental frequency of the main circuit changes relatively slowly, resulting in an increase in the D value. This indicates that the proportion of harmonic components in the main circuit current waveform increases, and the ability of the main circuit current signal to characterize the fault arc characteristics decreases. Conversely, when the load is mainly linear, the total harmonic superposition of the main circuit is small, the D value decreases, and the reliability of the main circuit current signal is higher. The characteristic dilution index D is updated once every power frequency cycle (20ms). The update cycle is consistent with the transform window length of the fast Fourier transform in step S1 to ensure that the D value and the harmonic intensity index of each branch are calculated within the same time window, avoiding distortion of the D value due to time misalignment.

[0044] In one specific embodiment, in step S3, based on the size range of the characteristic dilution index D, corresponding dynamic weights are assigned to the fault indication values ​​of the arc intensity signal, sound pressure signal, and main circuit current signal, respectively, including:

[0045] The maximum value of the arc light filtered signal is taken within a continuous sampling window, and the difference between the maximum values ​​of adjacent windows is obtained to obtain the arc light mutation amount; when the arc light mutation amount exceeds the arc light decision threshold, the arc light fault indicator value is set to 1, otherwise it is set to 0.

[0046] The short-time energy of the sound pressure filtered signal is calculated within a continuous sampling window to obtain the sound pressure energy value; when the sound pressure energy value exceeds the sound pressure decision threshold, the sound pressure fault indication value is set to 1, otherwise it is set to 0.

[0047] The energy mutation rate of the high-frequency component of the main circuit is calculated within a continuous sampling window to obtain the high-frequency energy mutation rate of the current. When the high-frequency energy mutation rate of the current exceeds the current judgment threshold, the current fault indication value is set to 1, otherwise it is set to 0.

[0048] Based on the size range of the feature dilution index D, when the feature dilution index D is not greater than the first threshold, the dynamic weights corresponding to the arc fault indication value, sound pressure fault indication value, and current fault indication value are respectively set as the first weighted reassembly; when the feature dilution index D is greater than the first threshold but not greater than the second threshold, the corresponding dynamic weights are set as the second weighted reassembly; when the feature dilution index D is greater than the second threshold, the corresponding dynamic weights are set as the third weighted reassembly; wherein, in the third weighted reassembly, the dynamic weight corresponding to the arc fault indication value is greater than the dynamic weight corresponding to the arc fault indication value in the first weighted reassembly, and the dynamic weight corresponding to the current fault indication value in the third weighted reassembly is less than the dynamic weight corresponding to the current fault indication value in the first weighted reassembly.

[0049] Specifically, the sampling window length is uniformly set to 1ms, corresponding to 10 sampling points (sampling frequency 10kHz). The arc flash mutation is the difference between the maximum value of the arc flash filtered signal within the current 1ms window and the maximum value of the previous window. The arc flash decision threshold is set to 500 lux. This value is determined based on the upper limit of arc flash intensity fluctuation caused by normal switching operations in the distribution cabinet. The arc flash mutation during normal operation usually does not exceed 300 lux. Setting the threshold to 500 lux can effectively eliminate interference from normal operation while retaining the sensitivity of fault arc flash response. The sound pressure energy value is the sum of the squares of the 10 sampling points of the sound pressure filtered signal within the current 1ms window. The sound pressure decision threshold is set to 0.5Pa. 2 This value is determined based on the upper limit of sound pressure energy generated by normal operations such as motor starting and circuit breaker opening and closing within the distribution cabinet; the current high-frequency energy mutation is the difference between the sum of the squares of the 10 sampling points of the main circuit high-frequency component within the current 1ms window and the sum of the squares of the previous window, and the current decision threshold is set to 0.02A. 2 This value is determined based on the upper limit of the high-frequency component energy fluctuation of the main circuit when the load in the distribution cabinet is normally switched on and off. The fault indication values ​​of the above three signals are all binary quantities, set to 1 when the corresponding judgment threshold is exceeded, and set to 0 otherwise.

[0050] The specific values ​​for the three weighted reassemblies are as follows: The first weighted reassembly corresponds to a low dilution state where the feature dilution index D is no greater than 0.15. In this state, the reliability of the main circuit current signal is high, and the dynamic weights for the arc fault indication value, sound pressure fault indication value, and current fault indication value are set to 0.25, 0.15, and 0.60, respectively. The second weighted reassembly corresponds to a medium dilution state where D is greater than 0.15 but no greater than 0.35, and the dynamic weights for the three channels are adjusted to 0.40, 0.25, and 0.35, respectively. The third weighted reassembly corresponds to a high dilution state where D is greater than 0.35. In this state, the main circuit current signal is severely distorted due to the superposition of multiple load harmonics, and the dynamic weights for the three channels are adjusted to 0.55, 0.35, and 0.10, respectively. All three weighted reassemblies satisfy the normalization constraint that the sum of the three weights is 1. The first threshold is set to 0.15, and the second threshold is set to 0.35. The boundary between the two thresholds is determined based on the measured statistical distribution of harmonics under typical load combinations of the distribution cabinet: when the D value is less than 0.15, the main circuit current is dominated by the fundamental wave, and the degree of harmonic superposition is low; when the D value is greater than 0.35, the superposition of nonlinear load harmonics has severely distorted the characteristics of the main circuit current, and the identification authority needs to be transferred to the arc light and sound pressure signals.

[0051] In one specific embodiment, step S4, calculating the difference in dilution index D between adjacent periods to obtain the dilution rate of change, includes:

[0052] The rate of change of dilution is obtained by subtracting the current period's characteristic dilution index D from the previous period's characteristic dilution index D.

[0053] When the absolute value of the dilution rate change is not greater than the first rate threshold, the confirmation counter threshold is set to the first count value; when the absolute value of the dilution rate change is greater than the first rate threshold but not greater than the second rate threshold, the confirmation counter threshold is set to the second count value; when the absolute value of the dilution rate change is greater than the second rate threshold, the confirmation counter threshold is set to the third count value; wherein, the third count value is greater than the second count value, and the second count value is greater than the first count value.

[0054] Specifically, the rate of change of dilution is defined as the difference between the characteristic dilution index D calculated at the end of the current power frequency cycle (20ms) and the characteristic dilution index D calculated at the end of the previous power frequency cycle. It is a signed scalar that reflects the switching speed of the load state between two adjacent power frequency cycles. When a high-power nonlinear load is suddenly connected in the distribution cabinet, the harmonic intensity index of each branch increases rapidly within one power frequency cycle, causing the characteristic dilution index D to produce a large positive jump between adjacent cycles, and the absolute value of the rate of change of dilution increases accordingly. When the load state is stable, the characteristic dilution index D changes slowly between adjacent cycles, and the absolute value of the rate of change of dilution is close to zero.

[0055] The first rate threshold is set to 0.05, and the second rate threshold is set to 0.15. These two thresholds divide the absolute value of the dilution rate of change into three intervals, corresponding to three scenarios: stable load status, load switching, and drastic load switching. The first count value is set to 2, corresponding to two 1ms detection windows (i.e., 2ms confirmation time), suitable for scenarios where the load status is stable, dynamic weight allocation is reliable, and rapid fault confirmation is allowed. The second count value is set to 5, corresponding to a 5ms confirmation time, suitable for scenarios where load switching occurs and weight allocation is in a transitional state. The third count value is set to 10, corresponding to a 10ms confirmation time, suitable for strong interference scenarios such as high-power load switching or motor starting. In this case, the feature dilution index D fluctuates drastically, causing the dynamic weight to switch rapidly between adjacent cycles. It is necessary to extend the confirmation window to filter out short-term jumps in the overall confidence level C caused by weight switching, and avoid outputting a stable fault judgment result that is misjudged. The increasing relationship of the three count values ​​ensures that the more unstable the load status, the more continuous windows are required for confirmation, strictly corresponding to the increasing interval of the absolute value of the dilution rate of change.

[0056] Figure 2 This is a schematic diagram illustrating the relationship between the absolute value of the dilution rate of change and the confirmation counter threshold in an embodiment of this application. Figure 2 The figure illustrates the segmented stepwise relationship between the absolute value of the dilution rate change |ΔD| and the confirmation counter threshold and corresponding confirmation duration. The horizontal axis represents the absolute value of the difference between the characteristic dilution index D of adjacent power frequency cycles, the left vertical axis represents the confirmation counter threshold (unit: windows), and the right vertical axis represents the corresponding confirmation duration (unit: ms). The solid line in the figure represents the stepwise change curve of the confirmation counter threshold, and the dashed line represents the stepwise change curve of the corresponding confirmation duration. The two curves have the same value and the same shape. The two vertical dotted lines correspond to the first rate threshold of 0.05 and the second rate threshold of 0.15, respectively. The horizontal axis is divided into three segments: low rate interval, medium rate interval, and high rate interval. The confirmation counter threshold takes the first count value of 2, the second count value of 5, and the third count value of 10 in the three intervals, respectively, with corresponding confirmation durations of 2ms, 5ms, and 10ms.

[0057] In one specific embodiment, in step S4, when the number of persistent windows where the overall confidence level C exceeds the decision threshold reaches the acknowledgment counter threshold, a stable fault decision result is output, including:

[0058] Within each detection window, the overall confidence level C is compared with the decision threshold. When the overall confidence level C exceeds the decision threshold, the confirmation counter is incremented by 1; when the overall confidence level C does not exceed the decision threshold, the confirmation counter is reset to zero.

[0059] When the accumulated value of the acknowledgment counter reaches the acknowledgment counter threshold, a stable fault judgment result is output.

[0060] Specifically, the detection window duration is set to 1ms, consistent with the sampling window length for feature extraction of each signal in step S3. This ensures that the overall confidence level C within each detection window strictly corresponds to the arc flash mutation, sound pressure energy value, and current high-frequency energy mutation extracted within the same window. The decision threshold is set to 0.50. The basis for this setting is as follows: In the high dilution state (third weighting, arc flash dynamic weight is 0.55), when only the arc flash fault indication value is set to 1, the overall confidence level C is 0.55, which exceeds 0.50. This ensures that in multi-load parallel scenarios with severely distorted current signals, a single arc flash trigger can initiate the confirmation count, avoiding missed detections due to current channel failure. Simultaneously, in the low dilution state (first weighting, current dynamic weight is 0.60), when only the current fault indication value is set to 1, the overall confidence level C is 0.60, which also exceeds 0.50, ensuring rapid response capability when the current signal is reliable.

[0061] The confirmation counter performs a comparison at the end of each 1ms detection window: when the overall confidence level C is strictly greater than the decision threshold of 0.50, the confirmation counter increments by 1 based on the current value; when the overall confidence level C is not greater than the decision threshold of 0.50, the confirmation counter immediately resets to zero and restarts counting. The zeroing mechanism ensures that only the number of windows that continuously exceed the decision threshold is counted as valid confirmations, excluding the case where the overall confidence level C occasionally exceeds the threshold due to short-term signal fluctuations. When the accumulated value of the confirmation counter reaches the confirmation counter threshold (one of the first count value 2, the second count value 5, or the third count value 10) calculated in step S4, a stable fault decision result is output. The stable fault decision result is a binary quantity, with a value of 1 indicating that an arc fault has occurred in the distribution cabinet, triggering the subsequent fault branch location and tripping command output process.

[0062] In one specific embodiment, step S4, locating the faulty branch based on the order of high-frequency energy mutations in the current signals of each branch, and outputting a tripping command to the faulty branch, includes:

[0063] Based on the stable fault judgment result, the energy difference between adjacent windows is calculated for the high-frequency components of each branch within the preset time window before the confirmation time, and the high-frequency energy change of each branch current is obtained.

[0064] Arrange the high-frequency energy mutation of the current in each branch in descending order, and take the branch with the maximum value as the faulty branch; when the maximum value exceeds the location judgment threshold, the fault is determined to occur in the faulty branch; when the high-frequency energy mutation of the current in each branch does not exceed the location judgment threshold, the fault is determined to occur in the main busbar.

[0065] Output trip command to the faulty branch, and record the fault timestamp, faulty branch number, overall confidence level C, characteristic dilution index D, and fault indication value of each branch.

[0066] Specifically, the confirmation time is defined as the moment when the cumulative value of the confirmation counter first reaches the confirmation counter threshold, i.e., the moment when the stable fault judgment result is set from 0 to 1. The preset time window is set to 10ms before the confirmation time, corresponding to 100 sampling points (sampling frequency 10kHz). Within this window, the difference between the sum of squares of the high-frequency components of each branch obtained in step S1 and the sum of squares of the sums ... 2 This value is determined based on the upper limit of the high-frequency component energy fluctuation of each branch when the load is switched on and off normally in the distribution cabinet. The sudden change in high-frequency energy of the branch caused by normal switching usually does not exceed 0.03A. 2 The location decision threshold is set to 0.05A. 2 It can effectively distinguish between high-frequency energy mutations caused by fault arcs and normal switching interference.

[0067] After sorting the high-frequency energy mutations of each branch current in descending order, the branch number corresponding to the maximum value is taken as the fault branch number; when this maximum value is strictly greater than the location judgment threshold of 0.05A... 2 When the fault occurs in the faulty branch, it is determined that the high-frequency energy change of the current in all branches does not exceed 0.05A. 2 When the fault location is confirmed, it indicates that the fault point is located on the main busbar upstream of each branch point, and the fault is determined to have occurred on the main busbar. After the location is completed, within 3ms after the stable fault judgment result is set to 1, a 50ms pulse width tripping command is output to the circuit breaker coil corresponding to the faulty branch, driving the circuit breaker to complete the tripping action; the fault timestamp, faulty branch number, comprehensive confidence level C, characteristic dilution index D, and arc fault indication value, sound pressure fault indication value, and current fault indication value are recorded simultaneously and written to the local storage module for maintenance personnel to review later.

[0068] The above describes the method for rapid identification of arc faults in distribution cabinets in the embodiments of this application. The following describes the system for rapid identification of arc faults in distribution cabinets in the embodiments of this application. One embodiment of the system for rapid identification of arc faults in distribution cabinets in the embodiments of this application includes:

[0069] The acquisition module is used to acquire the main circuit current signal, branch circuit current signals, arc intensity signal and sound pressure signal of the distribution cabinet, and to perform harmonic decomposition on the branch circuit current signals to obtain the harmonic intensity index of each branch circuit and the effective value of the fundamental wave of the main circuit.

[0070] The summation module is used to sum the harmonic intensity indices of each branch and then divide them by the effective value of the fundamental wave of the main branch to obtain the characteristic dilution index D.

[0071] The allocation module is used to assign corresponding dynamic weights to the fault indication values ​​of the arc intensity signal, the sound pressure signal, and the main circuit current signal according to the size range of the feature dilution index D, and to sum the fault indication values ​​with the corresponding dynamic weights to obtain the comprehensive confidence level C.

[0072] The positioning module is used to calculate the difference of the characteristic dilution index D between adjacent periods to obtain the dilution change rate, determine the confirmation counter threshold based on the dilution change rate, and output a stable fault judgment result when the number of continuous windows in which the comprehensive confidence C exceeds the decision threshold reaches the confirmation counter threshold. It also locates the faulty branch based on the sorting of the high-frequency energy mutation of the current signals of each branch and outputs a trip command to the faulty branch.

[0073] This invention also provides a rapid arc fault identification device for power distribution cabinets. This device can be a server and includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor, designed as a computer, provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores the data corresponding to this embodiment. The network interface of the device is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0074] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the method for rapid identification of arc faults in the distribution cabinet.

[0075] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0076] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a rapid arc fault identification device in a power distribution cabinet (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0077] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for rapid identification of arc faults in a power distribution cabinet, characterized in that, The method includes: Step S1: Collect the main circuit current signal, branch circuit current signal, arc intensity signal and sound pressure signal of the distribution cabinet, and perform harmonic decomposition on the branch circuit current signal to obtain the harmonic intensity index of each branch circuit and the effective value of the fundamental wave of the main circuit. Step S2: Sum the harmonic intensity indices of each branch and divide them by the effective value of the fundamental wave of the main road to obtain the characteristic dilution index D; Step S3: Based on the size range of the feature dilution index D, assign corresponding dynamic weights to the fault indication values ​​of the arc intensity signal, the sound pressure signal, and the main circuit current signal, respectively, and sum the fault indication values ​​with the corresponding dynamic weights to obtain the comprehensive confidence level C. Step S4: Calculate the difference of the characteristic dilution index D between adjacent periods to obtain the dilution change rate. Determine the confirmation counter threshold based on the dilution change rate. When the number of consecutive windows in which the comprehensive confidence C exceeds the decision threshold reaches the confirmation counter threshold, output a stable fault decision result. Locate the faulty branch based on the sorting of the high-frequency energy mutation of the current signals of each branch and output a trip command to the faulty branch.

2. The method for rapid identification of arc faults in a power distribution cabinet according to claim 1, characterized in that, Step S1 includes: Triggered by a unified clock signal, the main circuit current signal, branch circuit current signal, arc intensity signal and sound pressure signal of the power distribution cabinet are collected synchronously. The arc light intensity signal is subjected to median filtering to obtain an arc light filtered signal; the sound pressure signal is subjected to bandpass filtering to obtain a sound pressure filtered signal. The trunk current signal is low-pass filtered to obtain the trunk fundamental component; the trunk current signal is band-pass filtered to obtain the trunk high-frequency component; the current signals of each branch are band-pass filtered to obtain the high-frequency component of each branch. Based on the high-frequency components of each branch, a fast Fourier transform is performed on the current signal of each branch to extract the amplitudes of the 3rd, 5th, and 7th harmonics of each branch current signal. The amplitudes of the 3rd, 5th, and 7th harmonics of the same branch are summed to obtain the harmonic intensity index of each branch. Based on the fundamental component of the main circuit, the root mean square of the current signal of the main circuit is taken within one power frequency cycle to obtain the effective value of the fundamental component of the main circuit.

3. The method for rapid identification of arc faults in a power distribution cabinet according to claim 2, characterized in that, In step S2, the summation of the harmonic intensity indices of each branch is divided by the effective value of the fundamental wave of the main branch to obtain the characteristic dilution index D, including: The harmonic intensity indices of each branch are summed one by one to obtain the total superposition of harmonics in the main road. Based on the total amount of harmonic superposition in the main line and the effective value of the fundamental wave in the main line, the total amount of harmonic superposition in the main line is divided by the effective value of the fundamental wave in the main line to obtain the characteristic dilution index D; wherein, the characteristic dilution index D is updated once per power frequency cycle.

4. The method for rapid identification of arc faults in a power distribution cabinet according to claim 3, characterized in that, In step S3, based on the size range of the characteristic dilution index D, corresponding dynamic weights are assigned to the fault indication values ​​of the arc intensity signal, the sound pressure signal, and the main circuit current signal, including: The maximum value of the arc light filtered signal is taken within a continuous sampling window, and the difference between the maximum values ​​of adjacent windows is obtained to obtain the arc light mutation amount; when the arc light mutation amount exceeds the arc light decision threshold, the arc light fault indication value is set to 1, otherwise it is set to 0. The short-time energy of the sound pressure filtered signal is calculated within a continuous sampling window to obtain the sound pressure energy value; when the sound pressure energy value exceeds the sound pressure decision threshold, the sound pressure fault indication value is set to 1, otherwise it is set to 0. The energy mutation amount of the high-frequency component of the main circuit is calculated within a continuous sampling window to obtain the high-frequency energy mutation amount of the current; when the high-frequency energy mutation amount of the current exceeds the current decision threshold, the current fault indication value is set to 1, otherwise it is set to 0. Based on the size range of the feature dilution index D, when the feature dilution index D is not greater than a first threshold, the dynamic weights corresponding to the arc fault indication value, sound pressure fault indication value, and current fault indication value are respectively set as a first weighted reassembly; when the feature dilution index D is greater than the first threshold and not greater than a second threshold, the corresponding dynamic weights are set as a second weighted reassembly; when the feature dilution index D is greater than the second threshold, the corresponding dynamic weights are set as a third weighted reassembly; wherein, in the third weighted reassembly, the dynamic weight corresponding to the arc fault indication value is greater than the dynamic weight corresponding to the arc fault indication value in the first weighted reassembly, and the dynamic weight corresponding to the current fault indication value in the third weighted reassembly is less than the dynamic weight corresponding to the current fault indication value in the first weighted reassembly.

5. The method for rapid identification of arc faults in a power distribution cabinet according to claim 4, characterized in that, In step S4, the difference between the characteristic dilution index D in adjacent periods is calculated to obtain the dilution rate of change, including: The difference between the characteristic dilution index D of the current period and the characteristic dilution index D of the previous period is used to obtain the dilution rate of change. When the absolute value of the dilution rate change is not greater than a first rate threshold, the confirmation counter threshold is set to a first count value; when the absolute value of the dilution rate change is greater than the first rate threshold but not greater than a second rate threshold, the confirmation counter threshold is set to a second count value; when the absolute value of the dilution rate change is greater than the second rate threshold, the confirmation counter threshold is set to a third count value; wherein the third count value is greater than the second count value, and the second count value is greater than the first count value.

6. The method for rapid identification of arc faults in a power distribution cabinet according to claim 5, characterized in that, In step S4, when the number of consecutive windows in which the overall confidence level C exceeds the decision threshold reaches the confirmation counter threshold, a stable fault decision result is output, including: Within each detection window, the overall confidence level C is compared with the decision threshold. When the overall confidence level C exceeds the decision threshold, the confirmation counter is incremented by 1; when the overall confidence level C does not exceed the decision threshold, the confirmation counter is reset to zero. When the cumulative value of the confirmation counter reaches the confirmation counter threshold, a stable fault judgment result is output.

7. The method for rapid identification of arc faults in a power distribution cabinet according to claim 6, characterized in that, In step S4, the faulty branch is located based on the order of high-frequency energy mutations in the current signals of each branch, and a tripping command is output to the faulty branch, including: Based on the stability fault judgment result, the energy difference between adjacent windows is calculated for the high-frequency components of each branch within a preset time window before the confirmation time, and the high-frequency energy change of each branch current is obtained. The high-frequency energy mutations of the current in each branch are arranged in descending order, and the branch with the maximum value is taken as the faulty branch. When the maximum value exceeds the location decision threshold, the fault is determined to occur in the faulty branch. When the high-frequency energy mutations of the current in each branch of all branches do not exceed the location decision threshold, the fault is determined to occur in the main busbar. Output a trip command to the faulty branch and record the fault timestamp, faulty branch number, overall confidence level C, feature dilution index D, and fault indication value for each branch.

8. A rapid identification system for arc faults in a power distribution cabinet, characterized in that, For implementing the method for rapid identification of arc faults in a distribution cabinet as described in any one of claims 1-7, the rapid identification system for arc faults in a distribution cabinet comprises: The acquisition module is used to acquire the main circuit current signal, branch circuit current signals, arc intensity signal and sound pressure signal of the distribution cabinet, and to perform harmonic decomposition on the branch circuit current signals to obtain the harmonic intensity index of each branch circuit and the effective value of the fundamental wave of the main circuit. The summation module is used to sum the harmonic intensity indices of each branch and then divide them by the effective value of the fundamental wave of the main branch to obtain the characteristic dilution index D. The allocation module is used to assign corresponding dynamic weights to the fault indication values ​​of the arc intensity signal, the sound pressure signal, and the main circuit current signal according to the size range of the feature dilution index D, and to sum the fault indication values ​​with the corresponding dynamic weights to obtain the comprehensive confidence level C. The positioning module is used to calculate the difference of the characteristic dilution index D between adjacent periods to obtain the dilution change rate, determine the confirmation counter threshold based on the dilution change rate, and output a stable fault judgment result when the number of continuous windows in which the comprehensive confidence C exceeds the decision threshold reaches the confirmation counter threshold. It also locates the faulty branch based on the sorting of the high-frequency energy mutation of the current signals of each branch and outputs a trip command to the faulty branch.

9. A device for rapid identification of arc faults in a power distribution cabinet, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the method for rapid identification of arc faults in the distribution cabinet as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the method for rapid identification of arc faults in the distribution cabinet as described in any one of claims 1 to 7.

Images