Method for detecting leakage current of 0.4kV branch line of transformer area

Abstract

The present application relates to low-voltage distribution network leakage detection technical field, disclose a kind of 0.4kV branch line leakage current fast detection method of table area, first acquisition table area total outgoing line leakage current as physical reference;Non-invasive current signal is collected by flexible Rogowski coil to each branch, the difference between total reference and the sum of measured branch leakage is used to estimate system cumulative drift, and the current branch leakage value is backtracked and corrected;Fault branch is detected by section and progression, and the leakage current step mutation is identified by moving window statistical process control, combined with load fluctuation characteristics to remove false mutation, and automatically lock physical fault point.The present application does not need power off, does not need communication network, single line detection efficiency is improved by more than 80%, suitable for emergency rapid investigation of no communication coverage table area, significantly improve distribution network operation efficiency and power supply reliability.

Description

Technical Field

[0001] This invention relates to the field of leakage current detection technology in low-voltage distribution networks, and more specifically, to a rapid detection method for leakage current in 0.4kV branch lines of a transformer substation. Background Technology

[0002] As the power supply terminal of the power system, the 0.4kV low-voltage distribution network directly serves end users, and its operational safety is directly related to personal safety and power supply reliability. Leakage faults are common and frequent faults in low-voltage distribution areas, which can not only cause electric shock injuries and deaths, but also electrical fires due to insulation aging. Furthermore, excessive leakage current can cause residual current circuit breakers to trip frequently, severely affecting users' normal power supply. Therefore, achieving rapid and accurate detection of leakage current in distribution area branch lines and fault location is one of the core requirements for the operation and maintenance of low-voltage distribution networks.

[0003] In existing leakage current detection solutions, maintenance personnel carry portable testing equipment to perform non-invasive testing on each branch line on-site, without power outages or reliance on communication networks. While this solution offers advantages in portability and flexibility, existing technology still suffers from the following technical challenges: Because portable devices are battery-powered, and ambient temperature changes during on-site testing cause the battery voltage to gradually decrease as it discharges, this leads to slow drift in the zero point and gain of the detection system. As maintenance personnel test each branch line individually, this system drift accumulates over time, resulting in systemic deviations in the initial leakage current detection values ​​for each branch. Consequently, the sum of branch leakage currents cannot match the overall leakage current baseline value, leading to misjudgments in fault branch screening. Therefore, there is an urgent need to provide a rapid leakage current detection method for 0.4kV branch lines in a distribution area to address these issues. Summary of the Invention

[0004] To address the aforementioned technical issues, this invention provides a rapid detection method for leakage current in 0.4kV branch lines of a transformer substation. It achieves timing drift compensation based on the overall outgoing line reference, and combines moving window SPC with load fluctuation rejection to automatically and accurately locate fault points. The method employs a flexible Rogowski coil for non-invasive detection, eliminating the need for power outages and significantly improving the detection efficiency of a single line.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] A rapid detection method for leakage current of 0.4kV branch lines in a transformer substation includes:

[0007] The three-phase and neutral current signals of the main outgoing line of the transformer area are collected in a non-invasive manner, and the reference value of the total leakage current of the transformer area is calculated.

[0008] The three-phase and neutral current signals of each branch line in the transformer area are collected in a non-intrusive manner, and the original leakage current value of each branch is calculated after signal processing.

[0009] The cumulative system drift is determined by the difference between the total leakage current reference value of the transformer area and the original leakage current value of the measured branch. The original leakage current value of the current branch is retrospectively corrected to obtain the accurate leakage current value of each branch. The faulty branch is identified by comparison and screening.

[0010] The identified faulty branch is subjected to segmented progressive detection to obtain the leakage current sequence at each detection point along the line. Statistical process control methods are used to identify step changes in leakage current. Combined with load fluctuation interference elimination rules, the physical fault point is located.

[0011] As a preferred embodiment of the present invention, the non-invasive method employs four independent flexible open-type Rogowski coils respectively encircling the outer walls of the three-phase live wire and neutral wire. The Rogowski coils have a wide range covering milliampere to ampere levels and are free from magnetic saturation characteristics. Before acquiring the current signal, the detection system is subjected to full-range static accuracy calibration. A mapping relationship between the coil output voltage and the measured current is established and stored at multiple characteristic current calibration points. During on-site measurement, the calibration coefficient is corrected in real time based on the current ambient temperature and battery voltage.

[0012] As a preferred embodiment of the present invention, the signal processing includes: digitally integrating the acquired differential current signal to obtain an instantaneous current signal, and automatically performing zero-point calibration to eliminate the initial integration error after each coil installation;

[0013] The instantaneous current signal is subjected to multi-stage digital filtering processing, using a low-pass filter to remove high-frequency noise and a power frequency notch filter to eliminate power frequency common-mode interference.

[0014] The fundamental phasor is extracted from the filtered time-domain current. Based on Kirchhoff's current law, the fundamental current phasors of the three-phase live wire and neutral wire are vector summed and the modulus is taken to obtain the effective value.

[0015] As a preferred embodiment of the present invention, determining the cumulative drift of the system includes:

[0016] The difference between the reference value of the total leakage current of the transformer area and the sum of the original leakage current values ​​of each branch measured in the previous step is calculated. The average drift of a single branch is determined by a first-order inertial recursive model. The recursive model uses the inter-branch drift learning coefficient to perform a weighted average of the difference.

[0017] The backtracking correction includes:

[0018] The original leakage current value of the current branch is compensated by the average drift amount determined by the preceding branch to obtain the accurate leakage current value; the average drift amount is iteratively updated as the number of measured branches increases.

[0019] As a preferred embodiment of the present invention, the comparison screening includes: sorting the precise leakage current values ​​of each branch according to their numerical values, or directly comparing them with a preset safety threshold, and identifying the branches with abnormally large leakage currents as faulty branches; when multiple branches exceed the threshold at the same time, the branch with the largest leakage value is dealt with first.

[0020] As a preferred embodiment of the present invention, the segmented progressive detection includes: starting from the beginning of the fault branch, sequentially deploying detection at the cable joint, T-junction branch point, and user meter box access point; if there is no actual leakage change in each section of the main line, then proceeding down to the individual household inlet line to check for leakage on the user side.

[0021] As a preferred embodiment of the present invention, the statistical process control method includes: performing fixed-width moving window statistics on the leakage current sequence of detection points along the line, calculating the moving mean and moving standard deviation of the leakage current within the window, constructing a dynamic control upper limit and a dynamic control lower limit based on the moving mean and moving standard deviation, wherein the dynamic control lower limit is forced to be non-negative;

[0022] The identification of step change in leakage current includes: when the difference in leakage current between adjacent detection points exceeds a preset multiple of the moving standard deviation, a change pending confirmation state is triggered; if the leakage current of multiple subsequent detection points exceeds the dynamic control upper limit, it is determined to be a real physical fault point, otherwise it is determined to be instantaneous electromagnetic interference.

[0023] As a preferred embodiment of the present invention, the load fluctuation interference elimination rule includes: when a sudden change pending confirmation state is triggered, if the three-phase load current corresponding to the detection point changes proportionally and the leakage current does not change accordingly, it is determined to be a false sudden change caused by the switching of three-phase balanced load; or if there is a single-phase or two-phase load current change rate exceeding a preset ratio and the change rate of the current of the remaining phases is less than a preset tolerance, and the difference between the change rate of the neutral current and the change rate of the current of that phase is less than a preset tolerance, it is determined to be a false sudden change caused by the switching of single-phase or two-phase load.

[0024] As a preferred embodiment of the present invention, it further includes anomaly classification and data closed-loop management based on the accurate leakage current value; the anomaly classification includes comparing the accurate leakage current value with a preset safety threshold, classifying it into a warning level and an alarm level, and performing on-site verification or emergency repair respectively; the data closed-loop management includes transmitting the fault handling results to refresh the leakage baseline data, and generating a leakage trend curve based on historical detection data, and including old lines with continuously rising leakage values ​​in the pre-inspection and maintenance list in advance.

[0025] A rapid detection system for leakage current of a 0.4kV branch line in a transformer substation, comprising:

[0026] The main outgoing line detection module is used to collect the three-phase and neutral current signals of the main outgoing line of the transformer area in a non-intrusive manner, and calculate the reference value of the total leakage current of the transformer area.

[0027] The branch detection module is used to collect the three-phase and neutral current signals of each branch line in the transformer area in a non-intrusive manner, and calculate the original leakage current value of each branch after signal processing.

[0028] The drift compensation and screening module is used to determine the cumulative system drift by using the difference between the total leakage current reference value of the transformer area and the original leakage current value of the measured branch, to backtrack and correct the original leakage current value of the current branch, to obtain the accurate leakage current value of each branch, and to identify the faulty branch by comparison and screening.

[0029] The segmented location module is used to perform segmented progressive detection on the fault branch determined by the drift compensation and screening module, obtain the leakage current sequence of each detection point along the line, identify the step change of leakage current using statistical process control method, and lock the physical fault point by combining the load fluctuation interference elimination rule.

[0030] The beneficial technical effects of this invention are:

[0031] This invention proposes a timing drift compensation algorithm based on the total outgoing line reference of the transformer area. Using the physical constraint of the total leakage current as a benchmark, it estimates the cumulative drift of the system in real time through a first-order inertial recursive model and backtracks to correct the original leakage values ​​of each branch. This algorithm effectively solves the problem of inconsistent detection values ​​caused by temperature drift and battery voltage drop during on-site testing, ensuring uniformity of leakage detection values ​​across branches and significantly improving the accuracy of fault branch screening. Simultaneously, combined with a real-time on-site temperature and pressure calibration mechanism, it further enhances the accuracy of single-point detection.

[0032] This invention employs a moving window statistical process control method, constructing dynamic control limits through the 3σ principle and combining rate and continuity verification criteria to automatically identify step changes in leakage current without requiring manual experience. This algorithm balances detection sensitivity and statistical stability, accurately identifying fault points over short distances and achieving algorithm-level automatic fault location, significantly reducing reliance on the experience of maintenance personnel.

[0033] The invention designs targeted load fluctuation interference elimination rules, which can effectively distinguish between genuine leakage current surges and false surges caused by user load switching. By utilizing the proportionality of three-phase current changes and the synchronicity of single-phase load changes, measurement interference caused by user load switching is eliminated, solving the false alarm problem caused by load fluctuations in traditional methods and further improving the accuracy of fault location.

[0034] Utilizing flexible Rogowski coils to encircle the three-phase live and neutral wires, this system allows for rapid deployment in complex circuit scenarios without power outages, wire stripping, or insulation damage. It improves single-line detection efficiency by over 80%, effectively reducing power outage frequency and ensuring stable power supply for users. Employing a purely portable detection mode, it is suitable for rapid emergency troubleshooting of sudden faults. Its flexible deployment, low cost, and lack of power outage requirements ensure uninterrupted power supply for users. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0036] Figure 2 This is a schematic diagram of the detection host in this invention.

[0037] Figure 3 This is a schematic diagram of the detection host and the flexible Rogowski coil in this invention. Detailed Implementation

[0038] In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0039] Combination Figures 1-3 The present invention provides the following embodiments:

[0040] Example 1:

[0041] A method for rapid detection of leakage current in a 0.4kV branch line of a transformer substation, comprising:

[0042] The three-phase and neutral current signals of the main outgoing line of the transformer area are collected in a non-invasive manner, and the reference value of the total leakage current of the transformer area is calculated.

[0043] The three-phase and neutral current signals of each branch line in the transformer area are collected in a non-intrusive manner, and the original leakage current value of each branch is calculated after signal processing.

[0044] The cumulative system drift is determined by the difference between the total leakage current reference value of the transformer area and the original leakage current value of the measured branch. The original leakage current value of the current branch is retrospectively corrected to obtain the accurate leakage current value of each branch. The faulty branch is identified by comparison and screening.

[0045] The identified faulty branch is subjected to segmented progressive detection to obtain the leakage current sequence at each detection point along the line. Statistical process control methods are used to identify step changes in leakage current. Combined with load fluctuation interference elimination rules, the physical fault point is located.

[0046] Furthermore, the non-invasive method employs four independent flexible open-type Rogowski coils, which are respectively arranged around the outer walls of the three-phase live wire and neutral wire. The Rogowski coils have a wide range of measurement from milliamperes to amperes and are free from magnetic saturation characteristics. Before acquiring the current signal, the detection system is subjected to full-range static accuracy calibration. The mapping relationship between the coil output voltage and the measured current is established and stored at multiple characteristic current calibration points. During on-site measurement, the calibration coefficient is corrected in real time based on the current ambient temperature and battery voltage.

[0047] The four independent coils can collect the amplitude and phase information of the current in each phase, providing a data foundation for subsequent vector calculations based on Kirchhoff's current law. This is different from the traditional single coil, which can only measure the total residual current and cannot distinguish the contribution of each phase. The flexible open structure does not require stripping or disconnecting the line, enabling true uninterrupted operation, and can be adapted to complex wiring scenarios such as T-connections and multi-strand winding.

[0048] Rogowski coils operate based on the principle of electromagnetic induction. Since they have no iron core, there is no magnetic saturation problem. They can simultaneously cover the measurement requirements of small current leakage of 1mA and large current load of 500A. Full-range static calibration can eliminate the inherent system errors of sensors and circuits. On-site temperature and pressure real-time correction can compensate for the drift of coil mutual inductance coefficient caused by changes in ambient temperature and the drift of circuit gain caused by battery voltage drop, ensuring the accuracy of single-point measurement.

[0049] Furthermore, the signal processing includes: digitally integrating the acquired differential current signal to obtain the instantaneous current signal, and automatically performing zero-point calibration to eliminate the initial integration error after each coil installation;

[0050] The instantaneous current signal is subjected to multi-stage digital filtering processing, using a low-pass filter to remove high-frequency noise and a power frequency notch filter to eliminate power frequency common-mode interference.

[0051] The fundamental phasor is extracted from the filtered time-domain current. Based on Kirchhoff's current law, the fundamental current phasors of the three-phase live wire and neutral wire are vector summed and the modulus is taken to obtain the effective value.

[0052] The Rogowski coil outputs a differential signal that is proportional to the rate of change of current. It must be integrated to be restored to the true instantaneous current signal. The integrating circuit has an inherent zero drift. Zero calibration is performed after each coil installation to eliminate the initial value error of the integration and avoid the accumulation of error over time.

[0053] The low-voltage distribution area contains a large number of high-frequency harmonics, electromagnetic radiation interference, and power frequency common-mode interference generated by power electronic equipment. Multi-stage filtering can effectively filter out these interference components and retain only the 50Hz power frequency fundamental wave effective signal. The main energy of leakage current is concentrated in the power frequency fundamental wave. Extracting the fundamental wave phasor for vector operation can maximize the accuracy of leakage current calculation, which is in line with the physical essence of Kirchhoff's current law.

[0054] Furthermore, determining the cumulative drift of the system includes:

[0055] The difference between the reference value of the total leakage current of the transformer area and the sum of the original leakage current values ​​of each branch measured in the previous step is calculated. The average drift of a single branch is determined by a first-order inertial recursive model. The recursive model uses the inter-branch drift learning coefficient to perform a weighted average of the difference.

[0056] The backtracking correction includes:

[0057] The original leakage current value of the current branch is compensated by the average drift amount determined by the preceding branch to obtain the accurate leakage current value; the average drift amount is iteratively updated as the number of measured branches increases.

[0058] According to Kirchhoff's current law, the total leakage current of the transformer area is equal to the vector sum of the leakage currents of all branches. This is a physical constraint that is not affected by the measurement system. During the line-by-line testing process, the system drift caused by temperature drift and battery voltage drop will accumulate over time, resulting in a difference between the sum of the original leakage values ​​of each branch and the total leakage reference value. This difference is the system cumulative drift.

[0059] The first-order inertial recursive model can smooth the cumulative drift and avoid drift estimation fluctuations caused by instantaneous interference. As the number of measured branches increases, the sample size of drift estimation increases, and the estimation accuracy will gradually improve. Using the average drift of the preceding branches to backtrack and correct the current branch can ensure that the leakage current value of all branches is under the same quantitative benchmark, thus solving the problem of quantitative consistency in cross-branch detection.

[0060] Furthermore, the comparison and screening includes: sorting the precise leakage current values ​​of each branch according to their numerical values, or directly comparing them with a preset safety threshold, and identifying branches with abnormally large leakage currents as faulty branches; when multiple branches exceed the threshold at the same time, the branch with the largest leakage current value is dealt with first.

[0061] The severity of leakage faults is positively correlated with the magnitude of leakage current. The larger the leakage current, the higher the risk of electric shock, fire, and tripping. By sorting the values ​​or comparing thresholds, branches with potential faults can be quickly screened out. Prioritizing the handling of the branch with the largest leakage value can eliminate the highest safety risk, improve operation and maintenance efficiency, and avoid confusion in handling multiple fault scenarios.

[0062] Furthermore, the segmented progressive detection includes: starting from the beginning of the faulty branch, sequentially deploying detection at the cable joint, T-junction branch point, and user meter box access point; if there is no actual leakage change in any section of the main line, then proceeding down to the individual household inlet line to check for leakage on the user side.

[0063] Cable joints, T-junctions, and user meter box access points are the weakest points in the insulation of low-voltage lines and are also high-risk areas for leakage faults. Segment-by-segment testing at these critical points can significantly reduce the workload and improve fault location efficiency. Leakage current flows from the power source to the fault point. Testing segment by segment along the line from the beginning to the end will reveal a step change in leakage current at the fault point. If no such change occurs in any segment of the main line, it indicates that the fault is not on the main line but located on the user side of a branch. Therefore, it is necessary to investigate down to the individual household inlet line to cover all possible fault scenarios.

[0064] Furthermore, the statistical process control method includes: performing fixed-width moving window statistics on the leakage current sequence of detection points along the line, calculating the moving mean and moving standard deviation of the leakage current within the window, constructing a dynamic control upper limit and a dynamic control lower limit based on the moving mean and moving standard deviation, wherein the dynamic control lower limit is forced to be non-negative;

[0065] Under normal circumstances, the fluctuation of leakage current measurement follows a normal distribution law. Using moving window statistics can adapt to the leakage baseline level of different lines. Since leakage current is an effective value, it is physically impossible for it to be negative. Therefore, the lower limit is forcibly controlled to be non-negative to avoid invalid criteria.

[0066] The identification of step change in leakage current includes: when the difference in leakage current between adjacent detection points exceeds a preset multiple of the moving standard deviation, a change pending confirmation state is triggered; if the leakage current of multiple subsequent detection points exceeds the dynamic control upper limit, it is determined to be a real physical fault point, otherwise it is determined to be instantaneous electromagnetic interference.

[0067] A real leakage fault will cause a continuous step change in the leakage current, while transient electromagnetic interference will only cause a single point of abnormality. The rate criterion can quickly capture the characteristics of the sudden change, and the verification of multiple consecutive detection points exceeding the limit can effectively filter out transient interference, so as to accurately distinguish between real faults and interference, without relying on the manual experience of maintenance personnel.

[0068] Furthermore, the load fluctuation interference elimination rules include: when a sudden change pending confirmation state is triggered, if the three-phase load current corresponding to the detection point changes proportionally and the leakage current does not change accordingly, it is determined to be a false sudden change caused by the switching of three-phase balanced loads; or if there is a single-phase or two-phase load current change rate exceeding a preset proportion and the change rate of the current of the remaining phases is less than a preset tolerance, and the difference between the change rate of the neutral current and the change rate of the current of that phase is less than a preset tolerance, it is determined to be a false sudden change caused by the switching of a single-phase or two-phase loads.

[0069] Ideally, leakage current calculations based on Kirchhoff's current law are unaffected by load fluctuations. However, in actual measurements, phase errors in the Rogowski coil and asynchronous ADC sampling can lead to false jumps in leakage current during load switching. When a three-phase balanced load is switched, the three-phase currents change proportionally, and the neutral current remains essentially constant. When a single-phase or two-phase load is switched, the corresponding phase current changes significantly, while the currents of the other phases remain essentially constant, and the change in neutral current is approximately equal to the change in the phase current. By utilizing these typical current characteristics of load switching, it is possible to effectively distinguish between genuine leakage current jumps and false jumps caused by load fluctuations, further reducing the false alarm rate.

[0070] Furthermore, it also includes anomaly classification and data closed-loop management based on the accurate leakage current value; the anomaly classification includes comparing the accurate leakage current value with a preset safety threshold, classifying it into a warning level and an alarm level, and performing on-site verification or emergency repair respectively.

[0071] Different levels of leakage current correspond to different safety risks. Tiered handling can reasonably allocate operation and maintenance resources, carry out emergency repairs on high-risk faults, and conduct regular checks on low-risk hazards, thereby improving operation and maintenance efficiency.

[0072] The data closed-loop management includes transmitting fault handling results to refresh leakage baseline data, generating leakage trend curves based on historical detection data, and including old lines with continuously rising leakage values ​​in the pre-inspection and maintenance list in advance.

[0073] After the fault is handled, the leakage baseline data is refreshed, which can avoid misjudging the normal baseline as abnormal during subsequent testing; leakage trend analysis based on historical data can identify hidden dangers of line insulation aging in advance, realize the transformation from passive emergency repair to proactive preventive operation and maintenance, extend the service life of the line, and reduce the failure rate.

[0074] Example 2:

[0075] This invention also provides a rapid detection system for leakage current of a 0.4kV branch line in a transformer substation, comprising:

[0076] The main outgoing line detection module is used to collect the three-phase and neutral current signals of the main outgoing line of the transformer area in a non-intrusive manner, and calculate the reference value of the total leakage current of the transformer area.

[0077] The branch detection module is used to collect the three-phase and neutral current signals of each branch line in the transformer area in a non-intrusive manner, and calculate the original leakage current value of each branch after signal processing.

[0078] The drift compensation and screening module is used to determine the cumulative system drift by using the difference between the total leakage current reference value of the transformer area and the original leakage current value of the measured branch, to backtrack and correct the original leakage current value of the current branch, to obtain the accurate leakage current value of each branch, and to identify the faulty branch by comparison and screening.

[0079] The segmented location module is used to perform segmented progressive detection on the fault branch determined by the drift compensation and screening module, obtain the leakage current sequence of each detection point along the line, identify the step change of leakage current using statistical process control method, and lock the physical fault point by combining the load fluctuation interference elimination rule.

[0080] Application example:

[0081] A rapid detection method for leakage current of 0.4kV branch lines in a transformer substation is proposed. This method's core logic includes on-site mobile detection, establishment of a total outgoing line benchmark for the substation, time-sequence drift compensation for each branch, rapid screening of each branch, and precise local closed-loop management via segmented progressive SPC. Maintenance personnel, carrying a portable detection host, first measure the leakage current of the total outgoing line of the substation as a physical benchmark. Then, they non-invasively collect the differential signals of the three-phase and neutral currents of each branch on-site, and obtain the original leakage current value through integration and vector calculation. The difference between the total leakage current and the sum of the branch leakage currents is used to estimate the system's cumulative drift, and the measurement results are retrospectively corrected. After identifying the faulty branch through sorting and screening, segmented progressive detection is performed. A combination of moving window SPC control limits and rate-continuous verification criteria is used to automatically identify leakage current step changes, supplemented by load fluctuation exclusion rules to eliminate user switching interference, accurately pinpointing the physical fault point. The specific steps are as follows:

[0082] Step S1: Detection system setup and full-range static calibration

[0083] 1.1 Hardware System Preparation

[0084] The detection system consists of a portable leakage current detection unit and a flexible open-type Rogowski coil. The unit integrates a main control unit, signal integration circuit, 16-bit ADC acquisition module, internal temperature sensor, battery voltage monitoring module, LCD display module, and a fast-charging lithium battery. The fast-charging lithium battery can be fully charged in half an hour, and the offline battery life is ≥2 hours. The Rogowski coil has a range of 1mA-500A and a nominal mutual inductance coefficient. It has no magnetic saturation characteristics and its flexible structure is suitable for complex field scenarios such as cable T-junctions, narrow wiring, and overhead winding lines.

[0085] 1.2 Full-range static accuracy calibration

[0086] Under standard power frequency current source and constant temperature Stable power supply voltage Under these conditions, the entire machine was jointly calibrated at 7 key current nodes, the mapping relationship between coil output voltage and measured current was established, and the data was stored in the host's on-chip Flash.

[0087] The calibration points are: 1mA, 10mA, 100mA, 1A, 10A, 100A, 500A; the calibration technical specifications are: overall current measurement error ≤ ±1%, fundamental current phase measurement error ≤ 0.5°; the calibration coefficient array is: ,in , For the first Standard current values ​​at each calibration point This corresponds to the output voltage value of the coil.

[0088] 1.3 Temperature and pressure drift coefficient factory calibration

[0089] Calibrate the host unit's temperature drift coefficient in a constant temperature chamber and adjustable power supply environment. With battery voltage drift coefficient It is then cured to Flash. During on-site measurement, the calibration coefficient is directly corrected by temperature and pressure, and the calculation formula is as follows:

[0090]

[0091] In the formula Current temperature Current battery voltage Next Correction factor for each calibration point; The temperature drift coefficient was calibrated through testing on 100 prototype units. The voltage drift coefficient was calibrated through testing on 100 prototype units.

[0092] 1.4 System Parameter Initialization

[0093] According to the low-voltage power distribution design specifications, graded safety thresholds are set: among which the first-level warning threshold is... Level 2 alarm threshold Initialize the cumulative drift estimate .

[0094] Step S2: Measurement of the reference line for the main outgoing line of the transformer area

[0095] Maintenance personnel carried the testing unit to the main outgoing line of the transformer substation. Under the condition of no power outage, they wrapped four flexible Rogowski coils around the outer walls of the live wires and neutral wires of phases A, B, and C of the main outgoing line, respectively, and performed a complete signal acquisition and leakage current calculation process to obtain the original effective value of the leakage current of the main outgoing line of the transformer substation. .

[0096] This value serves as the physical reference for subsequent branch-by-branch drift compensation. Its physical meaning is: the vector sum of leakage currents of all branch lines within the transformer area, including the total drift of the measurement system.

[0097] Step S3: Single-branch signal acquisition and leakage current calculation

[0098] 3.1 Non-invasive synchronous acquisition of current signals

[0099] Maintenance personnel carried the host to the On the test branch line, without power interruption or wire stripping, four Rogowski coils were wrapped around the outer walls of the live and neutral wires of phases A, B, and C of the branch. Portable testing mode was activated, with the ADC sampling frequency set to 2kHz, to acquire the differential current signals of each phase in real time. Rogowski coil output:

[0100]

[0101] in, The equivalent mutual inductance coefficient after temperature and pressure correction is derived from the current calibration coefficient. We can deduce this by working backwards; The output voltage is the instantaneous induced voltage of the coil; This represents the instantaneous current in the conductor being measured.

[0102] 3.2 Signal Integration and Zero-Point Calibration

[0103] The host computer reconstructs the instantaneous current using a combination of hardware integrating circuitry and software digital integrating algorithm. Each time the coil is installed and before data acquisition begins, the host computer automatically performs a 100ms zero-point calibration to eliminate initial integration values. The error is calculated using the following formula:

[0104]

[0105] 3.3 Digital Filtering Processing

[0106] A 64th-order FIR low-pass filter is used to filter out high-frequency noise; a 50Hz power frequency notch filter is used to eliminate power frequency common-mode interference and fully preserve the fundamental effective current component.

[0107] 3.4 Fundamental Current Phasor Extraction

[0108] Perform a Fast Fourier Transform (FFT) on the four filtered time-domain currents to extract the 50Hz fundamental amplitude and initial phase, obtaining the fundamental current phasors for each phase:

[0109]

[0110] 3.5 Calculation of Leakage Current Vector

[0111] Current direction convention: three-phase current The positive direction is defined as the flow from the power source to the load; neutral current. The positive direction is defined as the flow returning from the load to the power source.

[0112] Applying Kirchhoff's Current Law (KCL) to the load node: the sum of the currents flowing into the node equals the sum of the currents flowing out of the node, that is:

[0113]

[0114] Rearranging terms yields the leakage current phasor:

[0115]

[0116] The actual effective value of the leakage current is the magnitude of the phasor:

[0117]

[0118] 3.6 Local Terminal Display

[0119] The main unit's LCD screen displays the real-time RMS values ​​and phase values ​​of the three-phase current, neutral current, and leakage current; the original detection data of this branch is stored locally, including the branch number, Phase, timestamp, internal temperature Battery voltage .

[0120] Step S4: Cross-branch time series drift self-compensation and branch-by-branch rapid screening

[0121] 4.1 Calculation of original leakage current value for a single branch

[0122] No. The original effective value of leakage current obtained by the calculation in step S3 for each branch is denoted as: .

[0123] 4.2 Cumulative Drift Estimation

[0124] Using the total outgoing line benchmark value of the transformer area Before calculation Cumulative system drift of branch:

[0125]

[0126] The average drift of a single branch is estimated using a first-order inertial recursive model:

[0127]

[0128] In the formula: To complete the first The estimated average drift of a single branch after a branching pattern; For the first Average drift estimate after branching, initial value ; The inter-branch drift learning coefficient, with values ​​ranging from [value missing]. Verified through 50 field tests in various transformer substations, this value can track slow drift caused by temperature drift and battery voltage drop, and suppress drift estimation fluctuations caused by instantaneous interference.

[0129] 4.3 Current Branch Leakage Current Retrospective Compensation

[0130] For the first The original leakage current value of each branch is adjusted for drift to obtain the compensated leakage current value:

[0131]

[0132] In the formula, For the first The accurate effective value of leakage current of each branch after time-drift compensation; : No. The cumulative average drift estimate after the completion of the branch is used as the first branch. The pre-compensation amount for each branch.

[0133] 4.4 Rapid screening of each branch

[0134] The maintenance personnel sequentially performed steps S2 to 4.3 on each branch under test within the transformer area to obtain the compensated leakage current value for each branch. The host performs comparison screening using locally stored data:

[0135] Each branch Sort by numerical value, or compare directly with a preset threshold;

[0136] Will The branch is identified as a candidate set of faulty branches;

[0137] If multiple branches exceed the threshold simultaneously, the branch with the largest leakage current value will be dealt with first.

[0138] This process replaces the traditional method of shutting down each branch one by one to troubleshoot, enabling rapid initial location of the faulty branch on-site.

[0139] Step S5: Precise fault branch segmentation based on SPC

[0140] 5.1 Segmented Progressive Detection

[0141] Maintenance personnel, carrying the main unit, arrive at the faulty branch site identified in step S4. Starting from the beginning of the branch, they repeat steps 3.1 to 3.5 sequentially at high-leakage locations such as cable joints, T-junctions, and user meter box access points to obtain the effective value sequence of leakage current at each detection point along the line. . These are the serial numbers of the inspection points along the route. This represents the total number of segments in the faulty branch.

[0142] 5.2 Calculation of moving window statistics

[0143] A moving window statistical process control is performed on the leakage current sequence, with a fixed window width. The first detection point. For the first... Each testing point ( ),calculate:

[0144] Moving average:

[0145] Moving standard deviation:

[0146] in For the first Average leakage current within the moving window; For the first Standard deviation of leakage current within the moving window; To adjust the width of the movable window, take the following value: It balances the sensitivity of mutation detection with statistical stability.

[0147] 5.3 SPC Dynamic Control Limit Construction

[0148] Based on the statistics from the previous detection point, calculate the... Upper and lower limits of dynamic control of points: , .

[0149] in, For the first Point controls the upper limit; For the first The lower limit of the point control is forced to be non-negative, and the leakage current is an effective value, which cannot be negative physically. To control the limit factor, the value is selected. It adopts the industry-standard 3σ principle, which can cover 99.73% of normal measurement fluctuations, with a false alarm rate of less than 0.3%.

[0150] 5.4 Automatic Identification Criteria for Step Changes

[0151] This invention employs a two-level combined criterion of rate triggering and continuous limit-crossing verification to accurately distinguish between real leakage current step changes and transient electromagnetic interference. The execution logic is as follows:

[0152] First, when the absolute value of the leakage current difference between the k-th detection point and the previous detection point exceeds a preset multiple of the standard deviation of the previous window movement, the system enters a sudden change pending confirmation state.

[0153]

[0154] in, This is the mutation rate threshold coefficient, with a value of [value missing]. Verified through 80 sets of artificial fault injection experiments, this value achieves the optimal balance between sensitivity and false alarm rate.

[0155] Subsequently, it was verified whether the leakage current at two consecutive detection points continuously exceeded the dynamic control limit after the pending confirmation state was triggered: .

[0156] Final determination: If the above consecutive limit-breaking conditions are met, the host will automatically trigger an audio-visual prompt, and the determination will be made for the [number]th [limit]. The first detection point and the first The line interval between each detection point is considered a real physical fault point; if this condition is not met, it is determined to be transient electromagnetic interference and no alarm is triggered.

[0157] 5.5 Load Fluctuation Disturbance Removal Rules

[0158] When to apply the rule: It is only executed after the mutation pending confirmation state is triggered, and is used to distinguish between real leakage current mutations and false load switching mutations caused by measurement system errors.

[0159] Ideally, leakage current calculations based on Kirchhoff's current law are unaffected by load fluctuations. However, in actual measurements, Rogowski coil phase errors and asynchronous sampling by multiple ADC channels can cause spurious jumps in leakage current during load switching. This rule identifies typical current characteristics during load switching to eliminate such spurious jumps, as follows:

[0160] If a sudden change pending confirmation state is triggered, and any of the following load switching characteristics are met simultaneously, it is determined to be a false change in the measurement system caused by load switching, and is not considered a leakage fault point:

[0161] Three-phase balanced load switching characteristics

[0162] The three-phase currents change proportionally and significantly, and the leakage current does not have a corresponding step change:

[0163]

[0164] in, , which represents the change in the total three-phase load; The load change rate threshold is set to a value of [value]. ; The tolerance for proportional variation of three-phase current is given by the following value: .

[0165] Single-phase or two-phase load switching characteristics

[0166] One or two phases of current change significantly on their own, while the currents of the other phases remain relatively stable, and the neutral current changes synchronously accordingly:

[0167] There exists a certain phase satisfy Furthermore, the change rates of the currents in the other two phases are both less than 10%, and the absolute value of the difference between the change rate of the neutral current and the change rate of the phase current is less than 10%.

[0168] Physical meaning of the rule: The ideal KCL vector sum is unaffected by load fluctuations, but in actual measurements, spurious abrupt changes may be introduced due to Rogowski coil phase errors and ADC sampling asynchrony. This rule is used to suppress such spurious abrupt changes in measurement systems, rather than to correct fundamental errors in the KCL principle.

[0169] 5.6 Troubleshooting leakage current on the user side

[0170] If each section of the branch line and main line is determined by SPC to have no real leakage current change, further inspection of the user's incoming line under that branch line will be carried out on a household-by-household basis to locate the user-side leakage current fault caused by aging household appliances or damaged lines.

[0171] Step S6: Anomaly Classification and Closed-Loop Management

[0172] 6.1 Local handling of leakage current levels

[0173] Level 1 Warning: The host computer triggers a yellow audible and visual alert and automatically records the potential hazard log. Maintenance personnel must complete the repair within 24 hours.

[0174] Level 2 Emergency Alert: The host triggers a red audible and visual alert, automatically records the fault log, immediately arranges emergency repairs, and, if necessary, takes measures such as power outages to power outages on local branches to eliminate the fault.

[0175] 6.2 Closed-loop management of handling results

[0176] After the on-site troubleshooting is completed, the maintenance personnel enter the cause of the fault, the repair method, and the material replacement information through the host human-machine interface; the host automatically refreshes the leakage baseline data of the branch line to complete the local closed loop.

[0177] 6.3 Preventive Maintenance Data Support

[0178] The host computer stores historical detection data of each branch locally for a long time, and supports local query and statistics by time, branch line, and leakage level. Maintenance personnel regularly export data to identify old lines with leakage values ​​that are rising month by month, and add them to the pre-inspection list in advance, realizing the transformation from fault repair to condition prediction maintenance.

[0179] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A rapid detection method for leakage current of a 0.4kV branch line in a transformer substation, characterized in that, include: The three-phase and neutral current signals of the main outgoing line of the transformer area are collected in a non-invasive manner, and the reference value of the total leakage current of the transformer area is calculated. The three-phase and neutral current signals of each branch line in the transformer area are collected in a non-intrusive manner, and the original leakage current value of each branch is calculated after signal processing. The cumulative system drift is determined by the difference between the total leakage current reference value of the transformer area and the original leakage current value of the measured branch. The original leakage current value of the current branch is retrospectively corrected to obtain the accurate leakage current value of each branch. The faulty branch is identified by comparison and screening. The identified faulty branch is subjected to segmented progressive detection to obtain the leakage current sequence at each detection point along the line. Statistical process control methods are used to identify step changes in leakage current. Combined with load fluctuation interference elimination rules, the physical fault point is located.

2. The method for rapid detection of leakage current in a 0.4kV branch line of a transformer substation according to claim 1, characterized in that, The non-invasive method employs four independent flexible open-type Rogowski coils, which are respectively arranged around the outer walls of the three-phase live wire and neutral wire. The Rogowski coils have a wide range of measurement from milliamperes to amperes and are free from magnetic saturation characteristics. Before acquiring the current signal, the detection system is subjected to full-range static accuracy calibration. The mapping relationship between the coil output voltage and the measured current is established and stored at multiple characteristic current calibration points. During on-site measurement, the calibration coefficient is corrected in real time according to the current ambient temperature and battery voltage.

3. The rapid detection method for leakage current of a 0.4kV branch line in a transformer substation according to claim 2, characterized in that, The signal processing includes: digitally integrating the acquired differential current signal to obtain the instantaneous current signal, and automatically performing zero-point calibration to eliminate the initial integration error after each coil installation; The instantaneous current signal is subjected to multi-stage digital filtering processing, using a low-pass filter to remove high-frequency noise and a power frequency notch filter to eliminate power frequency common-mode interference. The fundamental phasor is extracted from the filtered time-domain current. Based on Kirchhoff's current law, the fundamental current phasors of the three-phase live wire and neutral wire are vector summed and the modulus is taken to obtain the effective value.

4. The rapid detection method for leakage current of a 0.4kV branch line in a transformer substation according to claim 1, characterized in that, The determination of the system's cumulative drift includes: The difference between the reference value of the total leakage current of the transformer area and the sum of the original leakage current values ​​of each branch measured in the previous step is calculated. The average drift of a single branch is determined by a first-order inertial recursive model. The recursive model uses the inter-branch drift learning coefficient to perform a weighted average of the difference. The backtracking correction includes: The original leakage current value of the current branch is compensated by the average drift amount determined by the preceding branch to obtain the accurate leakage current value; the average drift amount is iteratively updated as the number of measured branches increases.

5. The method for rapid detection of leakage current in a 0.4kV branch line of a transformer substation according to claim 1, characterized in that, The comparison and screening includes: sorting the precise leakage current values ​​of each branch according to their numerical values, or directly comparing them with a preset safety threshold, and identifying branches with abnormally large leakage currents as faulty branches; when multiple branches exceed the threshold at the same time, the branch with the largest leakage current value is dealt with first.

6. The rapid detection method for leakage current of a 0.4kV branch line in a transformer substation according to claim 1, characterized in that, The segmented progressive detection includes: starting from the beginning of the faulty branch, sequentially deploying and detecting at the cable joint, T-junction branch point, and user meter box access point; if there is no actual leakage change in any section of the main line, then proceeding down to the individual household inlet line to check for leakage on the user side.

7. The rapid detection method for leakage current of a 0.4kV branch line in a transformer substation according to claim 1, characterized in that, The statistical process control method includes: performing fixed-width moving window statistics on the leakage current sequence of detection points along the line, calculating the moving mean and moving standard deviation of the leakage current within the window, constructing a dynamic control upper limit and a dynamic control lower limit based on the moving mean and moving standard deviation, wherein the dynamic control lower limit is forced to be non-negative; The identification of step change in leakage current includes: when the difference in leakage current between adjacent detection points exceeds a preset multiple of the moving standard deviation, a change pending confirmation state is triggered; if the leakage current of multiple subsequent detection points exceeds the dynamic control upper limit, it is determined to be a real physical fault point, otherwise it is determined to be instantaneous electromagnetic interference.

8. The method for rapid detection of leakage current in a 0.4kV branch line of a transformer substation according to claim 1, characterized in that, The load fluctuation interference elimination rules include: when a sudden change pending confirmation state is triggered, if the three-phase load current corresponding to the detection point changes proportionally and the leakage current does not change accordingly, it is determined to be a false sudden change caused by the switching of three-phase balanced loads; or if there is a single-phase or two-phase load current change rate exceeding a preset ratio and the change rate of the current of the other phases is less than a preset tolerance, and the difference between the change rate of the neutral current and the change rate of the current of that phase is less than a preset tolerance, it is determined to be a false sudden change caused by the switching of a single-phase or two-phase loads.

9. The method for rapid detection of leakage current in a 0.4kV branch line of a transformer substation according to claim 1, characterized in that, It also includes anomaly classification and data closed-loop management based on the accurate leakage current value; the anomaly classification includes comparing the accurate leakage current value with the preset safety threshold, classifying it into early warning level and alarm level, and performing on-site verification or emergency repair respectively; the data closed-loop management includes transmitting the fault handling results to refresh the leakage baseline data, and generating a leakage trend curve based on historical detection data, and including old lines with continuously rising leakage values ​​in the pre-inspection list in advance.

10. A rapid detection system for leakage current of a 0.4kV branch line in a transformer substation, characterized in that, include: The main outgoing line detection module is used to collect the three-phase and neutral current signals of the main outgoing line of the transformer area in a non-intrusive manner, and calculate the reference value of the total leakage current of the transformer area. The branch detection module is used to collect the three-phase and neutral current signals of each branch line in the transformer area in a non-intrusive manner, and calculate the original leakage current value of each branch after signal processing. The drift compensation and screening module is used to determine the cumulative system drift by using the difference between the total leakage current reference value of the transformer area and the original leakage current value of the measured branch, to backtrack and correct the original leakage current value of the current branch, to obtain the accurate leakage current value of each branch, and to identify the faulty branch by comparison and screening. The segmented location module is used to perform segmented progressive detection on the fault branch determined by the drift compensation and screening module, obtain the leakage current sequence of each detection point along the line, identify the step change of leakage current using statistical process control method, and lock the physical fault point by combining the load fluctuation interference elimination rule.

Images