A low-voltage cabinet branch circuit electric parameter anomaly detection method and system

Abstract

The application relates to the technical field of fault detection, and discloses a low-voltage cabinet branch circuit electric parameter abnormality detection method and system, wherein the method comprises the following steps: acquiring an arc light signal in a cabinet and each branch circuit current signal, and synchronously collecting a multi-source synchronous signal set; adopting adaptive band pass filtering and an arc light pulse generation mechanism to obtain a pretreatment signal set; extracting a circuit energy feature set through an arc light pulse density driven adaptive window length calculation mechanism; completing photoelectricity coherence analysis and forming a circuit discrimination feature set by using an arc light trigger transient change amount comparison mechanism; performing double threshold relative advantage determination according to the circuit discrimination feature set, and outputting a low-voltage cabinet branch circuit electric parameter abnormality determination result. By unifying multi-source time bases, suppressing external interference, and comparing transient responses of each circuit at an arc light trigger moment, the application can distinguish a fault circuit from adjacent circuit crosstalk, and improve the pertinence and anti-interference ability of abnormality identification under a multi-branch circuit light load working condition.

Description

Technical Field

[0001] This invention relates to the field of fault detection technology, and in particular to a method and system for detecting abnormal electrical parameters in a low-voltage switchgear branch circuit. Background Technology

[0002] Currently, in the detection of arc faults in branch circuits of low-voltage distribution cabinets, arc sensors and current transformers in each circuit are commonly used to synchronously acquire signals. Fault identification is achieved by monitoring the temporal correspondence between arc events and current surges. A typical processing path is: when the arc sensor detects a light pulse and the current change in a branch circuit exceeds a preset threshold, an arc fault is determined to have occurred in that circuit. This method can locate faults using significant changes in electrical parameters under conditions of independent single-circuit operation or large load current, and has certain engineering applicability. However, in light-load scenarios where multiple lighting or socket branch circuits operate in parallel and the current in each circuit is generally below 10A, the current characteristics generated by arc faults are extremely weak. Traditional fixed threshold methods struggle to accurately extract effective signals from noisy backgrounds, leading to a significant decrease in detection reliability.

[0003] For example, in a low-voltage switchgear operating multiple lighting or socket branch circuits simultaneously, the high-frequency current component excited by the arc under light load conditions has an amplitude of only milliamperes, which is easily submerged in background noise and crosstalk signals transmitted from adjacent circuits through electromagnetic coupling. Simultaneously, the complex physical structure within the switchgear causes reflection, scattering, and transmission delays in the optical signal from the arc sensor, blurring the time-stamp correlation between the optical pulse and the circuit current surge. When using traditional threshold judgment strategies, if the current surge threshold is set too low, transient processes such as normal load switching or operational overvoltage in adjacent circuits may be misjudged as arc faults in this circuit, leading to false tripping of non-faulty circuits; if the threshold is set too high, the truly weak arc signal will be ignored, resulting in missed detections, which, over time, will create a fire hazard. Existing technology cannot fully meet the dual requirements of accurate detection and non-false tripping of arc faults under light load conditions in multiple circuits.

[0004] Therefore, there is an urgent need for an electrical parameter anomaly detection method that can reliably identify arc faults in low-voltage switchgear where multiple branch circuits are under light load conditions and electromagnetic crosstalk from adjacent circuits exists. This method should deeply integrate the high-frequency components of the current in each circuit with the arc signal within the switchgear to accurately distinguish between arc faults in the current circuit and interference from adjacent circuits. This would effectively reduce the false alarm rate, prevent false tripping of non-faulty circuits, and improve the operational safety and power supply continuity of the low-voltage distribution system. Summary of the Invention

[0005] To address the aforementioned technical shortcomings, the purpose of this invention is to propose a method for detecting abnormal electrical parameters in branch circuits of low-voltage switchgear. This method aims to solve the technical problem that the existing single-circuit electrical parameter threshold method cannot distinguish between arc faults in the current circuit and interference from adjacent circuits in low-voltage switchgear where multiple branch circuits are under light load conditions and there is electromagnetic crosstalk between adjacent circuits.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a method for detecting abnormal electrical parameters of a low-voltage switchgear branch circuit.

[0007] The method for detecting abnormal electrical parameters in a low-voltage switchgear branch circuit includes: Step S10: Acquire the multi-source detection signal of the low-voltage switchgear. The multi-source detection signal of the low-voltage switchgear includes the arc light signal inside the switchgear and the current signal of each branch circuit. Based on the multi-source detection signal of the low-voltage switchgear, a synchronous acquisition mechanism is used to perform the multi-source signal acquisition task and output the multi-source synchronous signal set. Step S20: Based on the multi-source synchronous signal set, an adaptive bandpass filtering and arc pulse generation mechanism are used to perform signal preprocessing tasks, and a preprocessed signal set is output; Step S30: Based on the preprocessed signal set, perform the energy feature extraction task using an adaptive window length calculation mechanism driven by arc pulse density, and output the loop energy feature set; Step S40: Based on the circuit energy feature set, perform photoelectric coherence analysis using the arc-triggered transient change comparison mechanism, and output the circuit discrimination feature set; Step S50: Perform a dual-threshold relative advantage determination based on the circuit discrimination feature set, and output the abnormal electrical parameter determination result of the low-voltage switchgear branch circuit.

[0008] Preferably, step S10, which involves acquiring the multi-source detection signal of the low-voltage switchgear, performing a multi-source signal acquisition task based on the multi-source detection signal of the low-voltage switchgear using a synchronous acquisition mechanism, and outputting a multi-source synchronous signal set, specifically includes: Step S101: Set the sampling frequency of the arc light signal inside the cabinet and the current signal of each branch circuit to not less than 200kHz; Step S102: The arc light sensor is used to collect the arc light signal inside the cabinet, and the Rogowski coil current transformers set on each branch circuit are used to collect the current signal of each branch circuit. The same synchronous sampling clock is used to control the arc light sensor and each Rogowski coil current transformer to perform synchronous sampling. Step S103: Perform anti-aliasing low-pass filtering on the arc light signal inside the cabinet and the current signal of each branch circuit obtained by synchronous sampling, respectively. Set the cutoff frequency of the anti-aliasing low-pass filter to no more than half of the sampling frequency, and time-align the filtered arc light signal inside the cabinet and the current signal of each branch circuit according to the sampling time to obtain the multi-source synchronization signal set.

[0009] Preferably, step S20, which involves performing signal preprocessing based on the multi-source synchronization signal set using an adaptive bandpass filtering and arc pulse generation mechanism, and outputting the preprocessed signal set, specifically includes: Step S201: Perform zero-phase-shift bandpass filtering on the current signals of each branch circuit in the multi-source synchronization signal set, set the lower limit frequency of the initial passband to 10kHz, and set the upper limit frequency of the initial passband to 80kHz to obtain the high-frequency current components of each branch circuit. Step S202: Perform differential processing and absolute value processing on the arc light signals inside the cabinet in the multi-source synchronization signal set in sequence, determine the dynamic threshold of each sampling time according to the preset dynamic threshold update rule, and compare the absolute value processing result with the dynamic threshold of the corresponding sampling time; when the absolute value processing result changes from not greater than the dynamic threshold to greater than the dynamic threshold, determine the rising edge of the arc light pulse, and generate an arc light pulse sequence according to the sampling time corresponding to each rising edge of the arc light pulse; Step S203: According to a preset detection period, perform a Fast Fourier Transform on the current signals of each branch circuit before the zero-phase-shift bandpass filtering, and determine the noise baseline based on the median of the transformed spectral amplitude; divide the initial passband into multiple detection sub-bands. When the increment of the average spectral amplitude of any detection sub-band relative to the noise baseline is consistently not less than 15dB, and the increment of the spectral amplitude of any detection sub-band at the sampling time corresponding to the rising edge of the arc pulse is lower than a preset synchronization increment threshold, determine any detection sub-band as an external interference sub-band, and adjust the lower limit frequency of the initial passband to above the upper limit frequency of the external interference sub-band; combine the high-frequency current components of each branch circuit obtained after adjusting the bandpass frequency and the arc pulse sequence into the preprocessed signal set.

[0010] Preferably, step S202, the step of determining the dynamic threshold at each sampling time according to the preset dynamic threshold update rule, specifically includes: Acquire the arc light signal inside the cabinet within a 5ms time window prior to the current sampling time; Calculate the absolute value of the arc light signal inside the cabinet at each sampling point within the 5ms time window, and calculate the mean value of the absolute value of the arc light signal inside the cabinet at each sampling point to obtain the mean absolute value of the signal in the window. The mean absolute value of the window signal is multiplied by a preset threshold coefficient to obtain the dynamic threshold at the current sampling time, and the 5ms time window and the dynamic threshold are updated as the sampling time progresses.

[0011] Preferably, step S30, which involves performing energy feature extraction based on the preprocessed signal set using an adaptive window length calculation mechanism driven by arc pulse density, and outputting a loop energy feature set, specifically includes: Step S301: In the first Within a preset pulse density statistical time window where the sampling time ends, the number of sampling points corresponding to the rising edge of the arc pulse in the arc pulse sequence is counted, and the number of sampling points is calculated based on the number of sampling points and the duration of the preset pulse density statistical time window. Arc pulse density at each sampling time ; Step S302: Based on the arc pulse density Calculate the first Variable window length at each sampling time :

[0012] in, The sampling time number. For the first The variable window length at each sampling time. To preset the basic window length, The preset window length adjustment factor is used. It is the hyperbolic tangent function. The preset pulse density threshold is, and and They have the same dimensions; Step S303: Adjust the variable window length The ratio of the two samples is rounded down, and the larger of the rounded-down result and 1 is determined as the first value. Number of variable window sampling points at each sampling time Calculate the first according to the following formula The branch loop in the first Short-time energy value at each sampling time :

[0013] in, For the first The branch loop in the first The short-time energy value at each sampling time. The branch loop number, The sampling point number, For the first The number of variable window sampling points at each sampling time. For the first The branch loop in the first High-frequency current components at each sampling time; Arrange the short-time energy values ​​according to the sampling time to obtain the short-time energy sequence of each branch circuit, and combine the short-time energy sequence of each branch circuit, the arc pulse density, the arc pulse sequence and the high-frequency current component of each branch circuit into the circuit energy feature set.

[0014] Preferably, in step S40, the step of performing photoelectric coherence analysis based on the circuit energy feature set using an arc-triggered transient change comparison mechanism and outputting the circuit discrimination feature set specifically includes: Step S401: The first arc pulse in the arc pulse sequence... The sampling time corresponding to the rising edge of the arc pulse is determined as the nth arc pulse. The trigger moment is determined by extracting the high-frequency current components from each branch circuit as described above. A transient segment that starts at a trigger time and lasts for a preset transient analysis duration; Step S402: Determine the number of sampling intervals included in the transient segment based on the preset transient analysis duration and sampling time interval. Calculate the first according to the following formula The branch loop in the first Equivalent transient change at each trigger moment :

[0015] in, For the first The branch loop in the first The transient change equivalent at each trigger moment This is the rising edge number of the arc pulse. For the first The sampling point number corresponding to each trigger time. The number of sampling intervals contained in the transient segment. The sampling point number within the transient segment. For the first The branch loop in the first High-frequency current components at each sampling time The sampling time interval; Step S403: Place the first The ratio between the transient change equivalent of the first branch loop and the average transient change equivalent of the remaining branch loops is determined as the value of the first branch loop. The branch loop in the first Contrast of relative changes at each trigger moment And a preset zero-prevention constant is added to the average value of the transient change equivalent of the remaining branch loops, and the first branch loop is calculated according to the following formula. Transient coherence of a branch loop :

[0016] in, For the first The transient coherence of a branch loop, To preset the number of rising edges of the arc pulse within the judgment time window. For the first The branch loop in the first Contrast of relative changes at each trigger moment It is a natural logarithmic function, and the preset zero-prevention constant has the same dimensions as the transient change equivalent; The transient coherence of each branch loop and the short-time energy sequence of each branch loop are combined to form the loop discrimination feature set.

[0017] Preferably, step S50, which involves determining the relative dominance of the circuit based on the circuit discrimination feature set and outputting the result of abnormal electrical parameters of the low-voltage switchgear branch circuit, specifically includes: Step S501: Calculate the average short-time energy of each branch circuit within a preset judgment time window, and normalize the average short-time energy based on the preset background high-frequency energy benchmark value and the pre-calibrated reference energy corresponding to each branch circuit to obtain the normalized short-time energy value of each branch circuit; determine the fusion coefficient based on the preset load category corresponding to each branch circuit, and perform weighted fusion of the normalized short-time energy value and the transient coherence based on the fusion coefficient to obtain the comprehensive index of electrical parameter anomalies of each branch circuit; Step S502: Compare the comprehensive index of electrical parameter anomalies of each branch circuit with the preset start threshold. When the comprehensive index of electrical parameter anomalies of any branch circuit is greater than the preset start threshold, the branch circuit is determined as a branch circuit to be judged. The initial value of the preset start threshold is 0.25. The preset start threshold is updated according to the maximum comprehensive index of electrical parameter anomalies during the most recent 1 hour of fault-free operation before the current sampling time according to the preset update cycle. The updated preset start threshold is the larger of the value obtained by multiplying the maximum comprehensive index of electrical parameter anomalies by 0.8 and 0.2. Step S503: Compare the comprehensive index of electrical parameter anomalies of the branch circuit to be judged with the comprehensive index of electrical parameter anomalies of the other branch circuits; when the comprehensive index of electrical parameter anomalies of the branch circuit to be judged is not lower than the product of the maximum comprehensive index of electrical parameter anomalies of the other branch circuits and the preset advantage factor, and the maximum comprehensive index of electrical parameter anomalies of the other branch circuits is not greater than 1.5 times the average comprehensive index of electrical parameter anomalies of the other branch circuits, output the judgment result that the branch circuit to be judged has an electrical parameter anomaly.

[0018] This invention also provides a low-voltage switchgear branch circuit electrical parameter anomaly detection system, comprising: The signal acquisition module is used to acquire multi-source detection signals from the low-voltage switchgear. Based on the multi-source detection signals from the low-voltage switchgear, a synchronous acquisition mechanism is used to perform multi-source signal acquisition tasks and output a set of multi-source synchronous signals. The signal preprocessing module is used to perform signal preprocessing tasks based on the multi-source synchronous signal set using an adaptive bandpass filtering and arc pulse generation mechanism, and output the preprocessed signal set. The energy feature extraction module is used to perform energy feature extraction tasks based on the preprocessed signal set using an adaptive window length calculation mechanism driven by arc pulse density, and outputs a loop energy feature set. The coherence analysis module is used to perform photoelectric coherence analysis based on the circuit energy feature set using an arc-triggered transient change comparison mechanism, and outputs a circuit discrimination feature set. The anomaly determination module is used to perform dual-threshold relative advantage determination based on the circuit discrimination feature set and output the anomaly determination result of the electrical parameters of the low-voltage switchgear branch circuit.

[0019] The present invention also provides a low-voltage switchgear branch circuit electrical parameter anomaly detection device, the low-voltage switchgear branch circuit electrical parameter anomaly detection device comprising: a memory, a processor, and a low-voltage switchgear branch circuit electrical parameter anomaly detection program stored in the memory and executable on the processor, the low-voltage switchgear branch circuit electrical parameter anomaly detection program implementing the above method when executed by the processor.

[0020] The present invention also provides a computer program product, the computer program product including a low-voltage switchgear branch circuit electrical parameter anomaly detection program, which implements the above method when executed by a processor.

[0021] The beneficial effects of this invention are as follows: This invention establishes a multi-source synchronous signal set with a unified time base by acquiring the arc light signal and the current signal of each branch circuit using the same synchronous sampling clock, performing anti-aliasing low-pass filtering and time alignment; and then combining zero-phase-shift bandpass filtering, arc light dynamic threshold pulse generation and external interference sub-band identification to ensure that the high-frequency current components of each circuit correspond to the rising edge of the arc light in time, and reduce the impact of continuous narrowband electromagnetic interference and slow changes in background light on the preprocessing results, thus providing a consistent data basis for subsequent cross-circuit feature comparison.

[0022] By adjusting the short-time energy calculation window length based on the arc pulse density and converting the variable window length into the number of discrete sampling points, the high-frequency current components of each circuit are processed by square averaging on the same scale. This allows for the preservation of rapid energy changes when arc activity is dense and the increase of effective sampling points for statistical participation when arc activity is sparse. This balances the temporal resolution of the short-time energy sequence with the statistical stability of the background stage, and reduces the weakening of the intermittent light-load arc characteristics by the fixed window length.

[0023] By calculating the transient change equivalent of each circuit within the transient segment triggered by the rising edge of the arc pulse, and using the average transient change equivalent of the remaining circuits as a reference to form a relative change contrast, it is possible to distinguish the independent transient response of the faulty circuit from the common-mode crosstalk of the adjacent circuits under the same arc event. Then, the transient coherence is fused with the normalized short-time energy, and the start-up threshold and relative advantage judgment are executed in sequence, so that the abnormal results are simultaneously constrained by historical background, load category and cross-circuit differences, improving the pertinence of fault circuit identification under light load conditions of multi-branch circuits and reducing the possibility of misjudgment or missed judgment caused by a single threshold. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating the first embodiment of a method for detecting abnormal electrical parameters in a low-voltage switchgear branch circuit according to the present invention. Detailed Implementation

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0026] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Example 1: As Figure 1 The diagram shown is a flowchart of the first embodiment of the method for detecting abnormal electrical parameters in a branch circuit of a low-voltage switchgear according to the present invention. The first embodiment of the method for detecting abnormal electrical parameters in a branch circuit of a low-voltage switchgear according to the present invention is presented.

[0028] In the first embodiment, the method for detecting abnormal electrical parameters of a low-voltage switchgear branch circuit includes: Step S10: Acquire the multi-source detection signal of the low-voltage switchgear. The multi-source detection signal of the low-voltage switchgear includes the arc light signal inside the switchgear and the current signal of each branch circuit. Based on the multi-source detection signal of the low-voltage switchgear, a synchronous acquisition mechanism is used to perform the multi-source signal acquisition task and output the multi-source synchronous signal set. The multi-source detection signal of the low-voltage switchgear in step S10 refers to the arc light signal inside the switchgear and the current signal of each branch circuit obtained within the same low-voltage switchgear. The arc light signal inside the switchgear is collected by an arc light sensor, and the current signal of each branch circuit is collected by Rogowski coil current transformers respectively installed on the corresponding branch circuits. The arc light sensor and each Rogowski coil current transformer share the same synchronous sampling clock, and the sampling frequency is set to no less than 200kHz, so that the same number sampling points of different channels correspond to the same sampling time. Each signal after synchronous sampling is subjected to anti-aliasing low-pass filtering, and the cutoff frequency is set according to the actual sampling frequency and is no more than half of the sampling frequency; the filtering result is then time-aligned according to the sampling time. For channels where sampling is lost, the time is marked as an invalid sampling point and skipped in subsequent calculations. When the consecutive loss reaches a preset number, an abnormal sampling status is output, and the interpolation result is not used in the arc light triggering determination. The multi-source synchronous signal set formed by the above processing refers to the arc light signal sequence inside the switchgear and the current signal sequence of each branch circuit with a unified time base, channel identifier, and validity mark. This signal set serves as the input for step S20.

[0029] A unified synchronous sampling clock ensures that arc light changes and current changes in each branch circuit can directly correspond at the sampling time, while anti-aliasing low-pass filtering constrains signal components exceeding the identifiable frequency band before analog-to-digital conversion. The resulting multi-source synchronous signal set preserves the rapid changes in the initial stage of arc discharge while preventing time base offsets from different channels from being mistaken for circuit response differences. In step S20, when zero-phase-shift bandpass filtering is applied to the current signal and pulses are generated based on the rising edge of the arc light, the sequential relationship between the current component and the arc light event can be determined on the same sampling sequence, providing a verifiable time anchor point for subsequent extraction of transient segments.

[0030] Traditional multi-loop monitoring devices often use polling or independent channel clocks, which may lead to a misalignment of several sampling periods between the arc channel and the current channel. When the arc rise process is short, this misalignment can cause the transient current of the actual faulty loop to fall outside the arc trigger window. This implementation first unifies the clock, then performs anti-aliasing filtering and time alignment separately, ensuring that subsequent comparisons of each loop are based on the same physical time, rather than approximating asynchronous waveforms with software timestamps. This processing is particularly suitable for scenarios with a large number of branch loops and small currents in each loop, reducing discrimination bias introduced by sampling asynchrony and spectral folding.

[0031] In a low-voltage switchgear with 24 socket branch circuits, an in-cabinet arc sensor and 24 Rogowski coil current transformers can be connected to the same acquisition unit and operate synchronously at a sampling frequency of 250 kHz. When an arc is generated at the terminal of a branch circuit, the arc channel shows an upward change within a few microseconds, and the current channel of the corresponding branch circuit subsequently shows a high-frequency oscillation lasting approximately 150 microseconds. Adjacent circuits only record coupling fluctuations with small amplitudes. Since each channel uses the same sampling time, the acquisition unit can completely write the arc rise change and the current change of each circuit into the same multi-source synchronous signal set, and step S20 does not require re-estimating the inter-channel time delay.

[0032] Step S20: Based on the multi-source synchronous signal set, an adaptive bandpass filtering and arc pulse generation mechanism are used to perform signal preprocessing tasks, and a preprocessed signal set is output; The preprocessed signal set in step S20 refers to the combination of the high-frequency current components of each branch circuit after frequency band selection and the arc pulse sequence generated based on the arc signal inside the cabinet. For the current signal of each branch circuit, zero-phase-shift bandpass filtering is first used, with the initial lower limit of the passband set to 10kHz and the upper limit set to 80kHz to avoid changing the corresponding time of current change and arc rise edge due to filter phase delay. For the arc signal inside the cabinet, differential and absolute value processing are performed sequentially; the dynamic threshold of the current sampling time is obtained by multiplying the mean of the absolute values ​​of the arc signals at each sampling point within the previous 5ms time window by a preset threshold coefficient. When the absolute value processing result changes from not greater than the dynamic threshold to greater than the dynamic threshold, the corresponding sampling time is registered as the arc pulse rise edge, and an arc pulse sequence is generated accordingly. Furthermore, according to a preset detection cycle, a Fast Fourier Transform is performed on the current signals of each loop before bandpass filtering, using the median spectral amplitude as the noise baseline. The initial passband is divided into detection sub-bands. When the average spectral amplitude increment of a detection sub-band relative to the noise baseline is consistently not less than 15dB, and the spectral amplitude increment of that sub-band at the rising edge of the arc pulse is lower than a preset synchronization increment threshold, it is identified as an external interference sub-band, and the lower passband limit is adjusted to above the upper limit frequency of that interference sub-band. Peaks lasting only one detection cycle are not adjusted in the passband. For cases where the upper and lower limits coincide or are reversed after adjustment, the initial passband is restored, and preprocessing anomalies are recorded.

[0033] Zero-phase-shift filtering retains high-frequency current changes without introducing new time delays, while the dynamic threshold updates with the recent amplitude level of the arc signal, ensuring that the arc pulse sequence primarily marks sudden upward changes rather than following slow background light drift. The identification of external interference subbands uses both continuous amplitude increments and arc timing synchronization increments, thus distinguishing long-term, non-changing narrowband interference from arc-related broadband transients. The preprocessed signal set output in step S20 provides short-time energy calculation targets for step S30 and also preserves the arc rising edge and high-frequency current components of each loop for step S40, avoiding repeated data extraction from different versions of the signal in subsequent steps.

[0034] Fixed passband solutions cannot adapt to time-varying narrowband interference caused by equipment such as frequency converters and electronic ballasts in low-voltage switchgear, and fixed arcing thresholds are easily affected by ambient illuminance and sensor baseline changes. This implementation does not simply raise or lower a threshold; instead, it first uses short-time statistics of the arcing signal to form a dynamic threshold, and then identifies external interference sub-bands based on the continuous increment of the spectral amplitude relative to the noise baseline and arcing synchronization. Only frequency bands that simultaneously satisfy continuous enhancement and lack arcing synchronization abrupt changes trigger passband adjustment, thereby suppressing stable interference while preserving high-frequency current information synchronized with the arcing event.

[0035] In a lighting distribution cabinet, an electronic ballast can generate continuous narrowband noise around 42kHz, causing background components to appear in the passband of each branch circuit within the initial 10kHz to 80kHz range. After the system completes periodic spectrum detection, if the average amplitude of the 42kHz detection sub-band continuously increases by more than 15dB relative to the noise baseline, and the sub-band does not reach the synchronization increment threshold at the rising edge of the arc pulse, it is marked as an external interference sub-band, and the lower passband limit is adjusted to above the upper passband frequency. After adjustment, the continuous narrowband noise no longer enters the high-frequency current components of each circuit; the circuit where the arc occurs still retains the wideband variation synchronized with the arc pulse, and the two together constitute the preprocessed signal set for subsequent steps.

[0036] Step S30: Based on the preprocessed signal set, perform the energy feature extraction task using an adaptive window length calculation mechanism driven by arc pulse density, and output the loop energy feature set; The loop energy feature set in step S30 refers to the combination of short-time energy sequences, arc pulse densities, arc pulse sequences, and high-frequency current components of each branch loop. Within a preset pulse density statistical time window ending at the m-th sampling time, the number of sampling points corresponding to the rising edge of the arc pulse is counted, and this number is divided by the duration of the statistical time window to obtain the arc pulse density at the m-th sampling time. The textual calculation rule for the variable window length is as follows: first, calculate the ratio of the current arc pulse density to the preset pulse density threshold; then, perform a hyperbolic tangent operation on this ratio; multiply the result by the preset window length adjustment factor; then subtract the product from 1; and finally, multiply by the preset base window length. The arc pulse density and the threshold use the same dimensions, and the window length adjustment factor is set within the range that keeps the calculation result positive; if the calculation result is less than one sampling time interval, it is treated as one sampling time interval. Divide the variable window length by the sampling time interval and round down, then compare it with 1 and take the larger value to obtain the number of variable window sampling points at the current time. For the k-th branch loop, obtain the corresponding number of high-frequency current sampling values ​​from the current sampling point backward, square them respectively, and take the arithmetic mean to form the short-time energy value of the loop at the current moment; arrange the short-time energy values ​​according to the sampling time to obtain the short-time energy sequence.

[0037] The arc pulse density directly determines the length of the energy statistical window. A shorter window is used when arc activity is dense to preserve the time difference between consecutive pulses; when arc activity is sparse, a window length close to the base window is used to increase the number of sampling points participating in the averaging. Each branch loop at the same moment shares the window length determined by the arc pulse density, thus ensuring that the short-time energy of each loop has a consistent statistical scale, allowing for lateral comparisons in steps S40 and S50. Simultaneously, the arc pulse sequence and high-frequency current components are retained in the loop energy feature set, enabling subsequent transient change calculations to return to the original high-frequency sampling points, rather than replacing the transient waveform with already averaged energy values.

[0038] When using a fixed short window, energy spikes are easily generated during the arc interval due to insufficient effective samples; when using a fixed long window, energy changes caused by short-duration arcs are over-averaged. This implementation uses a density adjustment window obtained statistically from the same arc pulse sequence and explicitly converts the window length into an integer number of sampling points to avoid inconsistencies in boundaries caused by directly indexing discrete signals with time length. When the starting position of the statistical window is earlier than the starting point of the current data segment, only existing effective sampling points are used and the actual window length is adjusted synchronously; when the number of effective sampling points is less than the preset minimum number, the short-duration energy value at that moment is not output, thereby avoiding missing data being treated as zero values ​​in the averaging.

[0039] For example, the pulse density statistical time window can be set to 10ms, the basic window length to 0.5ms, the pulse density threshold to 1 pulse per millisecond, and the window length adjustment factor to 0.8. When the arc pulse density at a certain moment is 1.5 pulses per millisecond, the variable window length obtained according to the above rules is approximately 0.14ms. At a sampling frequency of 250kHz, after converting this window length into an integer number of sampling points, the squared average of the corresponding number of high-frequency current sampling values ​​is calculated. During the dense arc phase, the faulty circuit forms short-time energy peaks corresponding to the rising edges of each arc, while the non-faulty circuit maintains a lower energy level under the same window length. After the arc activity stops, the variable window length gradually returns to the basic window length, and the short-time energy sequence subsequently recovers to a relatively stable background state.

[0040] Step S40: Based on the circuit energy feature set, perform photoelectric coherence analysis using the arc-triggered transient change comparison mechanism, and output the circuit discrimination feature set; The loop discrimination feature set in step S40 refers to the combination of the transient coherence of each branch loop and the short-time energy sequence of each branch loop. The sampling time corresponding to the j-th rising edge in the arc pulse sequence is taken as the j-th trigger time. A transient segment, starting from the trigger time and lasting for a preset transient analysis duration, is extracted from the high-frequency current component of each branch loop. The preset transient analysis duration is divided by the sampling time interval to obtain the number of sampling intervals contained in the transient segment; if less than one sampling interval, it is treated as one sampling interval; segments exceeding the end of the current valid data are deferred until the data is complete for calculation. The textual calculation rule for the transient change equivalent is as follows: starting from the first sampling point after the trigger time, the difference between the current high-frequency current sample value and the previous sample value is calculated point by point, and then the difference is divided by the sampling time interval to obtain the discrete change rate; the current high-frequency current sample value, the discrete change rate, and the sampling time interval are multiplied together and accumulated over the entire transient segment, and the result is taken as the transient change equivalent of the loop at the trigger time. The relative change contrast is obtained by dividing the transient change equivalent of this loop by the sum of the average of the transient change equivalents of the other branch loops and the preset zero-prevention constant. The zero-prevention constant and the transient change equivalent use the same dimensions. Finally, the relative change contrast corresponding to each arc pulse within the preset decision time window is first incremented by 1 and then the natural logarithm is taken. The arithmetic mean of the processing results of all arc pulses is then taken to obtain the transient coherence of this loop. If there is no valid arc pulse within the decision time window, the transient coherence is not updated, and the status flag of the previous valid window is retained for identification by step S50.

[0041] Using the rising edge of the arc as the trigger moment restricts cross-loop comparisons to a short time segment following the same arc event, reducing the impact of load switching and random noise during non-arc periods on the comparison results. The transient change equivalent utilizes both the current amplitude of the high-frequency current and the rate of change of adjacent samples. The rapid response caused by the arc in this loop is retained in the accumulated result, while the fluctuations of adjacent loops, which are only slowly coupled with the electromagnetic field within the cabinet, are relatively small. Using the average transient change equivalent of the remaining loops as a real-time reference, disturbances common to all loops can be incorporated into the same comparison benchmark. Natural logarithmic processing and cross-event arithmetic averaging limit the dominance of single extreme values ​​on transient coherence. The resulting loop discrimination feature set retains both energy intensity and arc synchronization changes for joint determination in step S50.

[0042] Full-time cross-correlation or single-loop RMS comparisons include a large number of steady-state sampling points unrelated to the arc flash in the statistics. When common-mode crosstalk exists in multi-branch loops, multiple loops may simultaneously exhibit similar correlation levels. This implementation does not perform a general correlation for the entire running period, but instead extracts an equal-length segment after each arc flash rising edge and compares the transient change equivalent of this loop with the average value of the remaining loops. This relative quantity, with other loops in the same cabinet as a reference, can reflect whether the faulty loop has an independent and prominent transient response when an arc flash event occurs. At the same time, it avoids the loss of numerical stability of the ratio by handling the case where the average value of the remaining loops is close to zero through the zero constant protection.

[0043] In a low-voltage switchgear with 18 branch circuits, three valid arc pulses were recorded within a preset decision time window. After extracting equal-length transient segments from each circuit for each pulse, the transient change equivalent of circuit 3 at the first trigger moment was 0.81, while the average value of the remaining 17 circuits was 0.043. After adding a preset zero-prevention constant, the relative change contrast of circuit 3 was approximately 18.8. The other two trigger moments were calculated using the same steps, without using the arc amplitude as an additional weight. After adding 1 to the natural logarithm of each of the three relative change contrasts and taking the arithmetic mean, circuit 3 obtained a higher transient coherence than the other circuits. This result, along with the short-time energy sequences of each circuit, was written into the circuit discrimination feature set, allowing step S50 to continue checking the energy conditions and relative dominance conditions.

[0044] Step S50: Perform a dual-threshold relative advantage determination based on the circuit discrimination feature set, and output the abnormal electrical parameter determination result of the low-voltage switchgear branch circuit.

[0045] The abnormal electrical parameter determination result of the low-voltage switchgear branch circuit in step S50 refers to the branch circuit to be determined and its determination status that meet the start-up threshold and relative advantage conditions. First, the arithmetic mean of the short-time energy value of each branch circuit within the preset determination time window is taken, and normalized by combining the preset background high-frequency energy reference value and the pre-calibrated reference energy of the circuit: first, the background level represented by the background high-frequency energy reference is deducted, and then the normalized short-time energy value is obtained by converting according to the scale corresponding to the reference energy. The background high-frequency energy reference is determined by fault-free operation data, and the reference energy is obtained by calibration under known operating conditions; when the calibration data is insufficient, the factory reference value is used with an attached status mark indicating that the field calibration has not been completed. According to the preset load category of the branch circuit, a fusion coefficient is obtained, and the normalized short-time energy value and the transient coherence obtained in step S40 are weighted and fused to form a comprehensive index of abnormal electrical parameters. Circuits with a comprehensive index greater than a preset start threshold are identified as branch circuits to be judged. The initial start threshold is 0.25, and the maximum comprehensive index during the most recent hour of fault-free operation is read according to a preset update cycle. 0.8 times and 0.2 times this maximum comprehensive index are compared, and the larger of the two is used as the updated start threshold. Then, the circuit to be judged is compared with the other circuits: the comprehensive index of the circuit to be judged should not be lower than the product of the maximum comprehensive index of the other circuits and the preset dominance factor, and the maximum comprehensive index of the other circuits should not be greater than 1.5 times the average comprehensive index of the other circuits. If both conditions are met simultaneously, the result of an abnormal electrical parameter in the circuit to be judged is output. When there are multiple circuits to be judged, the same judgment is performed on each circuit, and the condition judgment is not replaced by the order of circuit numbers.

[0046] Normalization allows short-time energy under different current transformer ranges, background noise levels, and load levels to be included in the same judgment scale. The fusion coefficient corresponding to the load category is used to balance the contributions of energy characteristics and transient coherence characteristics. The start-up threshold is updated periodically based on the most recent fault-free operation data and can be adjusted according to changes in the long-term background level in the cabinet, but with a lower limit of 0.2 to avoid the threshold losing its constraint when the background is too low. The relative advantage condition further requires that the circuit to be judged has sufficient difference from other circuits and limits the maximum value of other circuits from being excessively prominent relative to the average value, thus allowing the cases of single-circuit anomalies and multiple circuits being disturbed to be handled separately.

[0047] A single absolute threshold scheme only judges whether the current or energy of a certain circuit exceeds the limit. It is easy to miss the fault when the arc characteristics are weak under light load. Lowering the threshold makes it easy to judge the switching of adjacent circuits or common-mode interference as faults. This implementation first uses the historical background of the circuit to form a normalized energy, and then fuses it with the transient coherence of the arc triggering. Subsequently, the start-up threshold and the relative advantage across circuits are checked in turn. Historical benchmark, cross-cabinet lateral reference and arc synchronous response constrain the abnormal intensity, inter-circuit differences and event correlation, respectively, so that the judgment result can be traced back to the specific data in the circuit energy feature set and the circuit discrimination feature set, rather than being directly given by a fixed threshold.

[0048] For example, a commercial building's low-voltage switchgear has 22 lighting and socket branch circuits. During the most recent hour of trouble-free operation, its maximum composite index was 0.16. According to the update rules, 0.8 times this value is less than 0.2, therefore the updated start-up threshold is 0.2. After a loose connection occurs in circuit 5, its normalized short-time energy value is 0.48 and its transient coherence is 0.35. If the fusion coefficient corresponding to its preset load category is 0.6, the weighted composite index is 0.428, which is higher than the start-up threshold. The maximum composite index of the remaining circuits is 0.10, and the average composite index is 0.07. The maximum value does not exceed 1.5 times the average value. When the preset dominance factor is 2.5, the composite index of circuit 5 is not lower than the product of the maximum value and the dominance factor of the remaining circuits. When both conditions are met, the system outputs a judgment result indicating that circuit 5 has an abnormal electrical parameter in a low-voltage switchgear branch circuit.

[0049] Example 2: Furthermore, the present invention provides a low-voltage switchgear branch circuit electrical parameter anomaly detection system, which employs a low-voltage switchgear branch circuit electrical parameter anomaly detection method described in the above embodiments, and can solve the technical problem of low-voltage switchgear branch circuit electrical parameter anomaly detection. The beneficial effects of the low-voltage switchgear branch circuit electrical parameter anomaly detection system provided by the present invention are the same as those of the low-voltage switchgear branch circuit electrical parameter anomaly detection method described in the above embodiments, and other technical features of the low-voltage switchgear branch circuit electrical parameter anomaly detection system are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0050] Example 3: This invention provides a low-voltage switchgear branch circuit electrical parameter anomaly detection device. The device includes at least one processor and a memory communicatively connected to the processor. The memory stores instructions executable by the processor, which are then executed to enable the processor to perform the low-voltage switchgear branch circuit electrical parameter anomaly detection method described in Example 1. The low-voltage switchgear branch circuit electrical parameter anomaly detection device in this invention can include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and vehicle terminals (e.g., vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. This low-voltage switchgear branch circuit electrical parameter anomaly detection device is merely an example and should not limit the functionality or scope of the invention. A low-voltage switchgear branch circuit electrical parameter anomaly detection device may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in a read-only memory or a program loaded from a storage device into a random access memory. The random access memory also stores various programs and data required for the operation of the low-voltage switchgear branch circuit electrical parameter anomaly detection device. The processing unit, read-only memory, and random access memory are interconnected via a bus. An I / O interface is also connected to the bus. Typically, the following systems can be connected to the I / O interface: input devices including touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices including liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices including magnetic tapes, hard disks, etc.; and communication devices. The communication device allows the low-voltage switchgear branch circuit electrical parameter anomaly detection device to communicate wirelessly or wiredly with other devices to exchange data. Although a low-voltage switchgear branch circuit electrical parameter anomaly detection device with various systems has been described, it should be understood that it is not required to implement or possess all the systems described. It can be implemented alternatively or with more or fewer systems.

[0051] Example 4: This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the low-voltage switchgear branch circuit electrical parameter anomaly detection method described above. The computer program product provided by this invention can solve the technical problem of detecting abnormal electrical parameters in a low-voltage switchgear branch circuit. Compared with the prior art, the beneficial effects of the computer program product provided by this invention are the same as those of the low-voltage switchgear branch circuit electrical parameter anomaly detection method provided in the above embodiments, and will not be repeated here.

[0052] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a read-only memory. When the computer program is executed by a processing device, it performs the functions defined in the methods of the embodiments disclosed in this invention.

[0053] It should be understood that the various parts disclosed in this invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0054] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the present invention and its equivalents, the present invention also intends to include these modifications and variations.

Claims

1. A method for detecting abnormal electrical parameters in a low-voltage switchgear branch circuit, characterized in that, The method includes: Step S10: Acquire the multi-source detection signal of the low-voltage switchgear. The multi-source detection signal of the low-voltage switchgear includes the arc light signal inside the switchgear and the current signal of each branch circuit. Based on the multi-source detection signal of the low-voltage switchgear, a synchronous acquisition mechanism is used to perform the multi-source signal acquisition task and output the multi-source synchronous signal set. Step S20: Based on the multi-source synchronous signal set, an adaptive bandpass filtering and arc pulse generation mechanism are used to perform signal preprocessing tasks, and a preprocessed signal set is output; Specifically, step S20, which involves performing signal preprocessing based on the multi-source synchronization signal set using an adaptive bandpass filtering and arc pulse generation mechanism, and outputting the preprocessed signal set, includes: Step S201: Perform zero-phase-shift bandpass filtering on the current signals of each branch circuit in the multi-source synchronization signal set, set the lower limit frequency of the initial passband to 10kHz, and set the upper limit frequency of the initial passband to 80kHz to obtain the high-frequency current components of each branch circuit. Step S202: Perform differential processing and absolute value processing on the arc light signals inside the cabinet in the multi-source synchronization signal set in sequence, determine the dynamic threshold of each sampling time according to the preset dynamic threshold update rule, and compare the absolute value processing result with the dynamic threshold of the corresponding sampling time; when the absolute value processing result changes from not greater than the dynamic threshold to greater than the dynamic threshold, determine the rising edge of the arc light pulse, and generate an arc light pulse sequence according to the sampling time corresponding to each rising edge of the arc light pulse; Step S203: According to a preset detection period, perform a Fast Fourier Transform on the current signals of each branch circuit before the zero-phase-shift bandpass filtering, and determine the noise baseline based on the median of the resulting spectral amplitude; divide the initial passband into multiple detection sub-bands. When the increment of the average spectral amplitude of any detection sub-band relative to the noise baseline is consistently not less than 15dB, and the increment of the spectral amplitude of any detection sub-band at the sampling time corresponding to the rising edge of the arc pulse is lower than a preset synchronization increment threshold, determine any detection sub-band as an external interference sub-band, and adjust the lower limit frequency of the initial passband to above the upper limit frequency of the external interference sub-band; combine the high-frequency current components of each branch circuit obtained after adjusting the bandpass frequency with the arc pulse sequence to form the preprocessed signal set; Step S30: Based on the preprocessed signal set, perform the energy feature extraction task using an adaptive window length calculation mechanism driven by arc pulse density, and output the loop energy feature set; Specifically, step S30, which involves performing energy feature extraction based on the preprocessed signal set using an adaptive window length calculation mechanism driven by arc pulse density, and outputting a loop energy feature set, includes: Step S301: In the first Within a preset pulse density statistical time window where the sampling time ends, the number of sampling points corresponding to the rising edge of the arc pulse in the arc pulse sequence is counted, and the number of sampling points is calculated based on the number of sampling points and the duration of the preset pulse density statistical time window. Arc pulse density at each sampling time ; Step S302: Based on the arc pulse density Calculate the first Variable window length at each sampling time : in, The sampling time number. For the first The variable window length at each sampling time. To preset the basic window length, To preset the window length adjustment factor, It is the hyperbolic tangent function. The preset pulse density threshold is, and and They have the same dimensions; Step S303: Adjust the variable window length The ratio of the two samples is rounded down, and the larger of the rounded-down result and 1 is determined as the first value. Number of variable window sampling points at each sampling time Calculate the first according to the following formula The branch loop in the first Short-time energy value at each sampling time : in, For the first The branch loop in the first The short-time energy value at each sampling time. The branch loop number, The sampling point number, For the first The number of variable window sampling points at each sampling time. For the first The branch loop in the first High-frequency current components at each sampling time; Arrange the short-time energy values ​​according to the sampling time to obtain the short-time energy sequence of each branch circuit, and combine the short-time energy sequence of each branch circuit, the arc pulse density, the arc pulse sequence and the high-frequency current component of each branch circuit into the circuit energy feature set; Step S40: Based on the circuit energy feature set, perform photoelectric coherence analysis using the arc-triggered transient change comparison mechanism, and output the circuit discrimination feature set; Specifically, step S40, which involves performing photoelectric coherence analysis based on the circuit energy feature set using an arc-triggered transient change comparison mechanism and outputting the circuit discrimination feature set, includes: Step S401: The first arc pulse in the arc pulse sequence... The sampling time corresponding to the rising edge of the arc pulse is determined as the nth arc pulse. The trigger moment is determined by extracting the high-frequency current components from each branch circuit as described above. A transient segment that starts at a trigger time and lasts for a preset transient analysis duration; Step S402: Determine the number of sampling intervals included in the transient segment based on the preset transient analysis duration and sampling time interval. Calculate the first according to the following formula The branch loop in the first Equivalent transient change at each trigger moment : in, For the first The branch loop in the first The transient change equivalent at each trigger moment This is the rising edge number of the arc pulse. For the first The sampling point number corresponding to each trigger time. The number of sampling intervals contained in the transient segment. The sampling point number within the transient segment. For the first The branch loop in the first High-frequency current components at each sampling time The sampling time interval; Step S403: Place the first The ratio between the transient change equivalent of the first branch loop and the average transient change equivalent of the remaining branch loops is determined as the value of the first branch loop. The branch loop in the first Contrast of relative changes at each trigger moment And a preset zero-prevention constant is added to the average value of the transient change equivalent of the remaining branch loops, and the first branch loop is calculated according to the following formula. Transient coherence of a branch loop : in, For the first The transient coherence of a branch loop, To preset the number of rising edges of the arc pulse within the judgment time window. For the first The branch loop in the first Contrast of relative changes at each trigger moment It is a natural logarithmic function, and the preset zero-prevention constant has the same dimensions as the transient change equivalent; The transient coherence of each branch loop and the short-time energy sequence of each branch loop are combined to form the loop discrimination feature set; Step S50: Perform a dual-threshold relative advantage determination based on the circuit discrimination feature set, and output the abnormal electrical parameter determination result of the low-voltage switchgear branch circuit.

2. The method for detecting abnormal electrical parameters in a low-voltage switchgear branch circuit as described in claim 1, characterized in that, Step S10, which involves acquiring the multi-source detection signal of the low-voltage switchgear, performing a multi-source signal acquisition task based on the multi-source detection signal of the low-voltage switchgear using a synchronous acquisition mechanism, and outputting a multi-source synchronous signal set, specifically includes: Step S101: Set the sampling frequency of the arc light signal inside the cabinet and the current signal of each branch circuit to not less than 200kHz; Step S102: The arc light sensor is used to collect the arc light signal inside the cabinet, and the Rogowski coil current transformers set on each branch circuit are used to collect the current signal of each branch circuit. The same synchronous sampling clock is used to control the arc light sensor and each Rogowski coil current transformer to perform synchronous sampling. Step S103: Perform anti-aliasing low-pass filtering on the arc light signal inside the cabinet and the current signal of each branch circuit obtained by synchronous sampling, respectively. Set the cutoff frequency of the anti-aliasing low-pass filter to no more than half of the sampling frequency, and time-align the filtered arc light signal inside the cabinet and the current signal of each branch circuit according to the sampling time to obtain the multi-source synchronization signal set.

3. The method for detecting abnormal electrical parameters in a low-voltage switchgear branch circuit as described in claim 1, characterized in that, Step S202, the step of determining the dynamic threshold at each sampling time according to the preset dynamic threshold update rule, specifically includes: Acquire the arc light signal inside the cabinet within a 5ms time window prior to the current sampling time; Calculate the absolute value of the arc light signal inside the cabinet at each sampling point within the 5ms time window, and calculate the mean value of the absolute value of the arc light signal inside the cabinet at each sampling point to obtain the mean absolute value of the signal in the window. The mean absolute value of the window signal is multiplied by a preset threshold coefficient to obtain the dynamic threshold at the current sampling time, and the 5ms time window and the dynamic threshold are updated as the sampling time progresses.

4. The method for detecting abnormal electrical parameters in a low-voltage switchgear branch circuit as described in claim 1, characterized in that, Step S50, which involves determining the relative dominance of the circuit based on the circuit discrimination feature set and outputting the abnormal electrical parameter determination result of the low-voltage switchgear branch circuit, specifically includes: Step S501: Calculate the average short-time energy of each branch circuit within a preset judgment time window, and normalize the average short-time energy based on the preset background high-frequency energy benchmark value and the pre-calibrated reference energy corresponding to each branch circuit to obtain the normalized short-time energy value of each branch circuit; determine the fusion coefficient based on the preset load category corresponding to each branch circuit, and perform weighted fusion of the normalized short-time energy value and the transient coherence based on the fusion coefficient to obtain the comprehensive index of electrical parameter anomalies of each branch circuit; Step S502: Compare the comprehensive index of electrical parameter anomalies of each branch circuit with the preset start threshold. When the comprehensive index of electrical parameter anomalies of any branch circuit is greater than the preset start threshold, the branch circuit is determined as a branch circuit to be judged. The initial value of the preset start threshold is 0.

25. The preset start threshold is updated according to the maximum comprehensive index of electrical parameter anomalies during the most recent 1 hour of fault-free operation before the current sampling time according to the preset update cycle. The updated preset start threshold is the larger of the value obtained by multiplying the maximum comprehensive index of electrical parameter anomalies by 0.8 and 0.

2. Step S503: Compare the comprehensive index of electrical parameter anomalies of the branch circuit to be judged with the comprehensive index of electrical parameter anomalies of the other branch circuits; when the comprehensive index of electrical parameter anomalies of the branch circuit to be judged is not lower than the product of the maximum comprehensive index of electrical parameter anomalies of the other branch circuits and the preset advantage factor, and the maximum comprehensive index of electrical parameter anomalies of the other branch circuits is not greater than 1.5 times the average comprehensive index of electrical parameter anomalies of the other branch circuits, output the judgment result that the branch circuit to be judged has an electrical parameter anomaly.

5. A low-voltage switchgear branch circuit electrical parameter anomaly detection system, applied to the low-voltage switchgear branch circuit electrical parameter anomaly detection method according to any one of claims 1 to 4, characterized in that, The low-voltage switchgear branch circuit electrical parameter anomaly detection system includes: The signal acquisition module is used to acquire multi-source detection signals from the low-voltage switchgear. Based on the multi-source detection signals from the low-voltage switchgear, a synchronous acquisition mechanism is used to perform multi-source signal acquisition tasks and output a set of multi-source synchronous signals. The signal preprocessing module is used to perform signal preprocessing tasks based on the multi-source synchronous signal set using an adaptive bandpass filtering and arc pulse generation mechanism, and output the preprocessed signal set. The energy feature extraction module is used to perform energy feature extraction tasks based on the preprocessed signal set using an adaptive window length calculation mechanism driven by arc pulse density, and outputs a loop energy feature set. The coherence analysis module is used to perform photoelectric coherence analysis based on the circuit energy feature set using an arc-triggered transient change comparison mechanism, and outputs a circuit discrimination feature set. The anomaly determination module is used to perform dual-threshold relative advantage determination based on the circuit discrimination feature set and output the anomaly determination result of the electrical parameters of the low-voltage switchgear branch circuit.

6. A device for detecting abnormal electrical parameters in a low-voltage switchgear branch circuit, characterized in that, The low-voltage switchgear branch circuit electrical parameter anomaly detection device includes: a memory, a processor, and a low-voltage switchgear branch circuit electrical parameter anomaly detection program stored in the memory and executable on the processor. When the low-voltage switchgear branch circuit electrical parameter anomaly detection program is executed by the processor, it implements a low-voltage switchgear branch circuit electrical parameter anomaly detection method according to any one of claims 1 to 4.

7. A computer program product, characterized in that, The computer program product includes a low-voltage switchgear branch circuit electrical parameter anomaly detection program, which, when executed by a processor, implements a low-voltage switchgear branch circuit electrical parameter anomaly detection method according to any one of claims 1 to 4.

Images