Method for detecting leakage of explosion-proof distribution box

Abstract

This invention discloses a leakage current detection method for explosion-proof distribution boxes, specifically relating to the field of electrical safety technology in flammable and explosive environments. It addresses the problems of leakage signal attenuation and insufficient detection sensitivity caused by the eddy current effect of metal explosion-proof enclosures. The method acquires the original leakage current signal from the wires inside the box and constructs an eddy current compensation factor based on the enclosure material and thickness for gain compensation. Simultaneously, it collects the surface temperature distribution and three-dimensional magnetic field components of the enclosure, and determines the compensation failure state through collaborative analysis of the high-temperature region migration acceleration and magnetic field polarization angle offset. When compensation is effective, it calculates the area fraction of the spatial magnetic field curl on the enclosure surface; when its temporal gradient exceeds a preset threshold, it triggers a tripping action, overcoming the shielding effect of the enclosure eddy current on weak leakage current signals, and achieving high-sensitivity leakage current protection while maintaining explosion-proof integrity.

Description

Technical Field

[0001] This invention relates to the field of electrical safety technology in flammable and explosive environments, and more specifically, to a method for detecting leakage current in explosion-proof distribution boxes. Background Technology

[0002] In power systems operating in flammable and explosive environments, explosion-proof distribution boxes provide safe power distribution for downstream equipment (such as lighting and motors). As a key piece of equipment in power distribution, the reliability of leakage current detection in explosion-proof distribution boxes is directly related to production safety. According to requirements, explosion-proof distribution boxes must adopt a metal shell structure to achieve explosion-proof function. Currently, the industry generally uses zero-sequence current transformers to detect leakage current, which relies on the induced output on the secondary side of the transformer due to changes in the magnetic field of the conductor.

[0003] However, the metal explosion-proof enclosure generates eddy currents under an alternating magnetic field, causing distortion and attenuation of the leakage current detection signal. Due to the inherent metal material and structural thickness of the explosion-proof enclosure, the alternating leakage current magnetic field will induce eddy currents within the enclosure. The reverse magnetic field generated by these eddy currents significantly weakens the original magnetic field strength, resulting in a reduction in the signal amplitude output by the zero-sequence current transformer. This physical effect leads to a decrease in the detection sensitivity for small current leakage (especially below 30mA), making it difficult to meet the safety protection requirements of explosive atmospheres. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a leakage current detection method for explosion-proof distribution boxes to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] Leakage detection methods for explosion-proof distribution boxes include:

[0007] S1. Obtain the original leakage current signal of the wires inside the explosion-proof distribution box;

[0008] S2. Based on the material and thickness of the metal shell of the explosion-proof distribution box, construct an eddy current compensation factor to perform gain compensation on the original leakage signal and generate a compensated leakage signal.

[0009] S3. Collect the temperature distribution on the surface of the metal shell of the explosion-proof distribution box and obtain the three-dimensional magnetic field component corresponding to the compensation leakage signal;

[0010] S4. Extract the migration acceleration of the high-temperature region based on the temperature distribution, and calculate the magnetic field polarization angle offset based on the three-dimensional magnetic field components; when the product of the migration acceleration and the polarization angle offset exceeds the preset collaborative failure threshold, generate a compensation failure flag and start the calibration mode.

[0011] S5. When the compensation failure flag is not triggered, obtain the spatial magnetic field distribution of the compensation leakage signal and calculate the area fraction of the curl of the spatial magnetic field distribution on the shell surface.

[0012] S6. When the gradient of the area fraction change over time exceeds the preset vortex mutation threshold, a leakage fault command is output to the tripping actuator.

[0013] Furthermore, the original leakage current signal of the wires inside the explosion-proof distribution box is obtained, including:

[0014] Install the zero-sequence current transformer around the grounding neutral wire of the three-phase conductors inside the explosion-proof distribution box;

[0015] The signal conditioning circuit performs bandpass filtering on the induced signal output by the zero-sequence current transformer to filter out high-frequency electromagnetic interference components and low-frequency drift components. The passband of the bandpass filtering process covers the fundamental frequency and integer harmonic frequencies of the power frequency.

[0016] The filtered signal is converted from analog to digital to generate the original digitized leakage current signal.

[0017] Furthermore, an eddy current compensation factor is constructed based on the material and thickness of the metal casing of the explosion-proof distribution box to perform gain compensation on the original leakage current signal, generating a compensated leakage current signal, including:

[0018] The resistivity and permeability of the metal shell material of the explosion-proof distribution box can be obtained by querying the pre-stored material parameter mapping table.

[0019] The reference eddy current attenuation coefficient is calculated based on resistivity and permeability.

[0020] The compensation weight coefficient is determined based on the thickness of the metal shell of the explosion-proof distribution box, wherein the compensation weight coefficient increases non-linearly with the increase of thickness;

[0021] The eddy current compensation factor is generated by multiplying the reference eddy current attenuation coefficient by the compensation weight coefficient.

[0022] The original leakage signal is amplified by a gain factor corresponding to the eddy current compensation factor to generate a compensated leakage signal.

[0023] Furthermore, the material parameter mapping table includes resistivity and permeability data for carbon steel, stainless steel, and aluminum alloy.

[0024] Furthermore, the temperature distribution on the surface of the metal casing of the explosion-proof distribution box is collected, and the three-dimensional magnetic field components corresponding to the compensation leakage signal are obtained, including:

[0025] Infrared thermal radiation images of the surface of the metal shell of the explosion-proof distribution box are acquired at fixed time intervals using an infrared thermal imager to generate temperature distribution data.

[0026] Multiple triaxial magnetic field sensors are arranged at equal intervals on the surface of the metal shell of the explosion-proof distribution box to simultaneously acquire the X-axis magnetic field component, Y-axis magnetic field component and Z-axis magnetic field component corresponding to the compensation leakage signal.

[0027] The temperature distribution data collected at the same time are timestamped and stored in alignment with the three-dimensional magnetic field components.

[0028] Furthermore, the infrared thermal imager's field of view covers the entire surface of the explosion-proof distribution box's metal casing.

[0029] Furthermore, the migration acceleration in the high-temperature region is extracted based on the temperature distribution, and the magnetic field polarization angle offset is calculated based on the three-dimensional magnetic field components. When the product of the migration acceleration and the polarization angle offset exceeds a preset collaborative failure threshold, a compensation failure flag is generated and a calibration mode is initiated, including:

[0030] Identify continuous high-temperature regions in the temperature distribution of a continuous time series where the temperature exceeds a preset temperature threshold;

[0031] Calculate the displacement change of the centroid of the high-temperature region at adjacent time points;

[0032] Calculate the migration acceleration in the high-temperature region based on the displacement change and time interval;

[0033] Extract the X-axis and Y-axis magnetic field components from the three-dimensional magnetic field components;

[0034] The instantaneous magnetic field polarization angle is calculated based on the ratio of the X-axis magnetic field component to the Y-axis magnetic field component.

[0035] The standard deviation of the instantaneous magnetic field polarization angle within the preset time window is used as the magnetic field polarization angle offset.

[0036] When the product of migration acceleration and magnetic field polarization angle offset exceeds the preset collaborative failure threshold, a compensation failure flag is generated and calibration mode is initiated.

[0037] Furthermore, when the compensation failure flag is not triggered, the spatial magnetic field distribution of the compensation leakage signal is acquired, and the area fraction of the curl of the spatial magnetic field distribution on the shell surface is calculated, including:

[0038] The spatial magnetic field vector distribution corresponding to the compensation leakage signal is obtained by multiple triaxial magnetic field sensors deployed on the surface of the metal shell of the explosion-proof distribution box.

[0039] Calculate the magnetic field curl vector at each spatial location point based on the spatial magnetic field vector distribution;

[0040] The surface of the metal shell of the explosion-proof distribution box is discretized into several surface units;

[0041] Calculate the product of the normal component of the curl vector of the magnetic field within each surface unit and the area of ​​the surface unit;

[0042] Summing the product of all surface units yields the area fraction of the curl of the spatial magnetic field distribution on the shell surface.

[0043] Furthermore, when the gradient of the area fraction change over time exceeds a preset vortex mutation threshold, a leakage fault command is output to the tripping actuator, including:

[0044] Obtain the area fraction of the curl of the spatial magnetic field distribution over a continuous time series on the shell surface;

[0045] Calculate the difference in area fractions at adjacent time points;

[0046] Divide the difference by the corresponding time interval to obtain the rate of change of the area fraction over time.

[0047] When the rate of change per unit time exceeds the preset vortex mutation threshold, a leakage fault command is generated.

[0048] The leakage fault command is transmitted to the tripping actuator to trigger the circuit breaker to trip.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. By constructing an eddy current compensation factor, dynamic gain compensation is performed on the leakage current signal to counteract the magnetic field attenuation caused by the eddy current effect of the metal shell. The compensation factor is calculated based on the physical characteristics of the shell material resistivity, permeability and thickness. It can accurately restore the signal for different explosion-proof enclosure structures and significantly improve the detection sensitivity of small leakage current. At the same time, multi-modal sensing of temperature distribution and three-dimensional magnetic field is adopted to establish a thermal-magnetic coupling analysis model. When the migration of local overheated areas of the shell and the abnormal polarization of the magnetic field produce a correlation effect, the compensation failure calibration mechanism is automatically triggered to ensure the long-term effectiveness of the compensation parameters under complex working conditions.

[0051] 2. By performing area integral calculation of the curl of the spatial magnetic field on the shell surface, the eddy current field distribution characteristics of leakage current are transformed into quantifiable indicators. This area integral gradient detection breaks through the limitations of traditional single signal amplitude judgment and is extremely sensitive to the changes in non-uniform eddy current field formed in the early stage of leakage. Combined with the preset eddy current mutation threshold, it achieves millisecond-level fault response. Under the premise of maintaining the explosion-proof structure, the leakage current detection signal chain is reconstructed from the physical field level, taking into account both the safety constraints of explosive environment and the reliability of protection. Attached Figure Description

[0052] Figure 1 This is a flowchart of the leakage current detection method for the explosion-proof distribution box of the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Example: Figure 1 The present invention provides a leakage current detection method for explosion-proof distribution boxes, comprising:

[0055] S1. Obtain the original leakage current signal of the wires inside the explosion-proof distribution box;

[0056] S2. Based on the material and thickness of the metal shell of the explosion-proof distribution box, construct an eddy current compensation factor to perform gain compensation on the original leakage signal and generate a compensated leakage signal.

[0057] S3. Collect the temperature distribution on the surface of the metal shell of the explosion-proof distribution box and obtain the three-dimensional magnetic field component corresponding to the compensation leakage signal;

[0058] S4. Extract the migration acceleration of the high-temperature region based on the temperature distribution, and calculate the magnetic field polarization angle offset based on the three-dimensional magnetic field components; when the product of the migration acceleration and the polarization angle offset exceeds the preset collaborative failure threshold, generate a compensation failure flag and start the calibration mode.

[0059] S5. When the compensation failure flag is not triggered, obtain the spatial magnetic field distribution of the compensation leakage signal and calculate the area fraction of the curl of the spatial magnetic field distribution on the shell surface.

[0060] S6. When the gradient of the area fraction change over time exceeds the preset vortex mutation threshold, a leakage fault command is output to the tripping actuator.

[0061] S1. Obtain the original leakage current signal of the wires inside the explosion-proof distribution box, specifically as follows:

[0062] When acquiring the original leakage current signal of the conductors inside the explosion-proof distribution box, the first step is to install the zero-sequence current transformer. The toroidal zero-sequence current transformer is wrapped around the grounding neutral wire of the three-phase conductors inside the explosion-proof distribution box. During installation, ensure a reasonable gap between the inner diameter of the transformer and the outer diameter of the conductors, for example, within the range of 3 to 5 mm, to avoid signal distortion caused by physical contact. The transformer core material is made of a high-permeability alloy, such as permalloy with an initial permeability greater than 15,000. The secondary winding is made of high-temperature resistant enameled wire with a cross-sectional area of ​​0.5 square millimeters or more. After installation, an air gap test is performed. The gap between the closed surfaces of the core is measured using a feeler gauge and controlled within 0.1 mm to ensure complete magnetic circuit closure.

[0063] The signal conditioning circuit uses a fourth-order active filter architecture to implement bandpass filtering for processing the transformer output signal. During filter design, based on regulations for power frequency grid harmonics, the lower passband limit is set to the fundamental frequency minus 5 Hz, and the upper passband limit is set to 9 times the fundamental frequency. For example, in a 50 Hz power grid system, the passband range is set to 45 Hz to 450 Hz, fully covering the 50 Hz fundamental frequency and its 3rd, 5th, and 7th harmonic components. Filter component parameters are determined through calculations using standard filter design equations: the feedback resistor is a metal film resistor with a temperature coefficient below 50 ppm / ℃, and the capacitor dielectric is polypropylene material with a dielectric loss angle below 0.1%. High-frequency interference filtering addresses transient interference generated by switching device operation, setting a stopband attenuation rate of no less than 24 dB / oct; low-frequency drift suppression addresses environmental temperature drift, providing attenuation greater than 40 dB below 10 Hz.

[0064] The analog-to-digital (A / D) conversion process employs a successive approximation converter, with signal level adaptation performed before conversion. A programmable gain amplifier adjusts the signal amplitude to 80% to 90% of the A / D converter's range; for example, when the converter's range is ±5V, the target signal peak is controlled between ±4V and ±4.5V. The sampling frequency is set to at least 2.5 times the highest frequency component of the signal, based on Shannon's sampling theorem; for example, a sampling rate of 1125Hz or higher is used for the 450Hz harmonic. A bandgap reference with an initial accuracy of 0.05% is selected as the conversion reference voltage source, and a π-type filter network consisting of a 10μF tantalum capacitor and a 100nF ceramic capacitor is installed at the power input. Dithering technology is introduced during the conversion process to suppress quantization errors; that is, a triangular wave disturbance signal with an amplitude of 0.5LSB is superimposed on the input signal.

[0065] Dynamic adaptation of the bandpass filter is achieved through programmable analog devices. When a change in the mains frequency is detected, such as switching from a 50 Hz to a 60 Hz system, the filter resistor network is adjusted via a digital potentiometer to synchronously switch the passband range to 55 Hz to 550 Hz. Filter performance verification employs a frequency sweep test method: a 10 mV sweep signal is input through a function generator, and the frequency is linearly swept from 1 Hz to 1 kHz, recording the output amplitude-frequency response curve. When the passband fluctuation exceeds ±1 dB or the stopband attenuation is less than 20 dB, a calibration program is triggered to automatically correct the component parameters.

[0066] The signal link's self-calibration mechanism is executed every 24 hours. During calibration, the current transformer input is disconnected, and a standard sine wave with an amplitude of 100mV and a frequency equal to the power frequency is injected into the signal conditioning circuit input. The output signal is acquired, and the gain error and phase shift are calculated. When the gain deviation exceeds 1% or the phase difference is greater than 0.5 degrees, the calibration coefficients are updated and stored in non-volatile memory. The calibration data includes a frequency response compensation table for real-time correction of filter characteristic drift.

[0067] The digital processing of the raw leakage current signal employs a sliding window averaging filter algorithm. The window width is set to 20 power frequency cycles; for example, a 400-millisecond window corresponds to a 50 Hz system. Each data frame contains 1024 sampling points, and the frame header embeds a synchronization identifier 0xAA55 for data packet integrity verification. Data transmission uses Manchester encoding format and is conducted through an optically isolated serial interface. The communication protocol includes the device address code, data length, payload, and CRC16 checksum field.

[0068] The mechanical stability of the installation structure is ensured through vibration damping design. The transformer bracket is made of non-magnetic stainless steel, and a silicone rubber vibration damping pad with a Shore hardness of 60 is installed between the bracket and the distribution box housing. The signal transmission cable adopts a double-layer shielding structure: the inner layer is a braided copper mesh with a coverage of more than 85% to shield high-frequency interference, and the outer layer is an aluminum-plastic composite film to shield low-frequency magnetic fields, with an overall shielding effectiveness of 70dB@100MHz.

[0069] The current-carrying capacity of the grounding neutral conductor is verified through a temperature rise test. When the conductor carries the rated current, its surface temperature is monitored using an infrared thermometer to ensure the temperature rise does not exceed 25°C above the ambient temperature. The magnetic saturation characteristics of the transformer core are verified through a DC bias test: a DC current of 10% of the conductor's rated current is applied, and the change in the transformer's output AC gain is measured to be no more than 1%.

[0070] Clock synchronization for analog-to-digital conversion is achieved by using a phase-locked loop (PLL) to track the power frequency signal. A digital phase detector compares the phase difference between the sampling clock and the zero-crossing point of the power frequency, dynamically adjusting the voltage-controlled oscillator (VCO) frequency to control the sampling clock jitter within 50 picoseconds. Nonlinear error compensation for the conversion result uses a 16-point piecewise linear interpolation method, and the calibration coefficients are calibrated at the factory using a precision voltage reference source.

[0071] Real-time monitoring of key parameters for signal link fault diagnosis includes operational amplifier bias current, filter Q-value drift, and analog-to-digital converter lost codes. When the bias current change exceeds 10% of its initial value or three consecutive data packet verification failures occur, a system alarm is triggered, and the system switches to the redundant backup channel. In environments with strong electromagnetic interference, twisted-pair shielded cables are used for signal transmission, with a twist pitch of less than 20 mm to suppress common-mode interference.

[0072] S2. Based on the material and thickness of the metal casing of the explosion-proof distribution box, an eddy current compensation factor is constructed to compensate the gain of the original leakage current signal, generating a compensated leakage current signal. The specific implementation is as follows:

[0073] When constructing the eddy current compensation factor based on the material and thickness of the metal casing of the explosion-proof distribution box, the first step is to query the material parameter mapping table. This material parameter mapping table is pre-stored in non-volatile memory and contains the electromagnetic characteristic parameters of commonly used metal materials in explosion-proof distribution boxes. The construction process of the material parameter mapping table is as follows: for carbon steel materials, resistivity data is obtained through a four-probe resistance test; for stainless steel materials, permeability is measured using an alternating magnetic field induction method, for example, the typical relative permeability of 304 austenitic stainless steel is 1.02; for aluminum alloy materials, conductivity is measured using a conductivity meter and then converted to resistivity. Each material entry is associated with its standard heat treatment state, for example, cold-rolled state and annealed state have separate data records.

[0074] The calculation of the baseline eddy current attenuation coefficient is based on a physical model grounded in the principle of electromagnetic induction. The input parameters are resistivity ρ and permeability μ, obtained from a material parameter mapping table. The permeability needs to be converted to absolute permeability in standard units. The calculation process follows the fundamental principle of eddy current loss: the baseline eddy current attenuation coefficient is directly proportional to the square root of the angular frequency ω, directly proportional to the square root of the permeability μ, and inversely proportional to the square root of the resistivity ρ. The angular frequency ω is determined based on the actual power grid frequency; for example, in a 50 Hz system, the angular frequency is 2π × 50 rad / s. The calculation result is a dimensionless coefficient, typically ranging from 0.5 to 3.0.

[0075] The compensation weighting coefficient is determined based on the physical relationship between shell thickness and eddy current shielding effect. The magnetic field attenuation characteristics of metal plates with different thicknesses are calibrated experimentally: a series of test samples with thicknesses ranging from 1 mm to 15 mm are prepared. A standard alternating magnetic field is applied to one side of the sample, and the attenuation rate of the magnetic field strength on the other side is measured. Experimental data shows that the compensation weighting coefficient exhibits a non-linear increasing characteristic with increasing thickness; for example, for every 1 mm increase in thickness, the increment of the weighting coefficient increases by approximately 1.5 times. In practical applications, a segmented calculation strategy is adopted: a linear compensation model is used when the thickness is less than 6 mm, a quadratic function compensation model is used when the thickness is between 6 mm and 12 mm, and a constant upper limit compensation model is used when the thickness exceeds 12 mm. The thickness data is obtained from the values ​​marked on the design drawings of the explosion-proof distribution box, and the measurement accuracy is controlled within ±0.1 mm.

[0076] The eddy current compensation factor is generated through a digital arithmetic unit. The baseline eddy current attenuation coefficient and the compensation weight coefficient are multiplied by a floating-point multiplier, and the result is stored in a dedicated register. The calculation process includes a numerical range protection mechanism: when the product result is greater than 3.5, it is automatically limited to 3.5; when it is less than 0.5, it is set to 0.5. The calculation result retains three significant decimal places; for example, the product of the baseline coefficient 1.82 and the weight coefficient 1.75 is 3.185. The update mechanism uses an event-triggered mode: recalculation is initiated when a change in shell material or thickness parameter is detected; otherwise, the historical valid value is retained.

[0077] Gain compensation is implemented through a programmable analog gain circuit. After reading the eddy current compensation factor K, the gain of the amplifier circuit is set to the reciprocal of K; for example, when K is 2.0, the gain is set to 0.5. Dynamic range detection is performed before the original leakage current signal is input to the amplifier circuit: when the instantaneous peak value of the signal exceeds 90% of the full scale, a 1 / 2 voltage divider preprocessing circuit is activated. The compensated signal is output through a high-precision digital-to-analog converter, with the output voltage range calibrated to ±5V. The compensation accuracy verification method is as follows: injecting a 10mA standard power frequency current into the system, the amplitude error of the compensated signal is measured to be no more than ±1% of the nominal value.

[0078] The material parameter mapping table maintenance includes an automatic update function. When a new metal material type is added, the material identification information is entered through the human-machine interface, and the system automatically retrieves the default parameters from the built-in material database. Parameter verification is performed periodically: a standard material sample is attached to the shell surface, and the electromagnetic parameters are measured using an eddy current analyzer and compared with the values ​​stored in the mapping table. When the deviation exceeds 5%, the parameter correction process is initiated.

[0079] On-site verification of the casing thickness was conducted using an ultrasonic thickness gauge. Nine evenly distributed measurement points were selected on the surface of the distribution box, including four corner positions, four edge positions, and one center position. The final thickness was calculated as the arithmetic mean of the measurements taken at each point. A thickness consistency detection mechanism was implemented: when the difference between the maximum and minimum measured values ​​exceeded 10% of the nominal thickness, abnormal deformation of the casing was identified, triggering a manual inspection process. The allowable deviation between the measurement data and the design drawings was controlled within ±5%.

[0080] The temperature compensation mechanism is implemented through a dual-signal path design. The main signal path performs conventional compensation processing, while the auxiliary monitoring path collects shell temperature data in real time. When the temperature change exceeds ±15℃, the compensation factor is adjusted according to a preset temperature-compensation coefficient relationship table. For example, for every 10℃ increase in temperature, the compensation factor for carbon steel increases by 0.02. The relationship table data comes from environmental simulation tests: the temperature drift characteristics of each material are measured in 10℃ increments within a temperature range of -25℃ to +85℃.

[0081] The signal path anomaly monitoring system employs a three-level protection mechanism: Level 1 monitors the amplifier output DC offset, initiating automatic zeroing when it exceeds 10% of full scale; Level 2 analyzes the signal spectrum components, classifying an abnormal state when the fundamental frequency energy percentage falls below 80%; Level 3 compares the signal energy ratio before and after compensation, triggering a system reset when it exceeds the theoretical range of 1.5 to 2.5. All abnormal events are recorded in a time-stamped operation log for subsequent fault analysis.

[0082] The online verification of the compensation effect employs a reference sensor comparison method. A reference current transformer and a main current transformer are installed in parallel on the grounding neutral line, with the reference transformer positioned outside the enclosure to avoid shielding. During system operation, the amplitude ratio of the two signals is calculated in real time. If this ratio deviates from the theoretical compensation factor by more than 5% for more than 5 seconds, a compensation failure state is automatically marked, triggering subsequent processing. The reference transformer is installed at least 200 mm away from the enclosure surface to avoid electromagnetic coupling interference.

[0083] S3. Collect the temperature distribution on the surface of the metal shell of the explosion-proof distribution box and obtain the three-dimensional magnetic field component corresponding to the compensation leakage signal. The specific implementation is as follows:

[0084] When collecting temperature distribution data and acquiring three-dimensional magnetic field components from the surface of the metal casing of an explosion-proof distribution box, the installation position and parameters of the infrared thermal imager are first configured. The infrared thermal imager is fixedly installed on a bracket 1.5 to 2.5 meters away from the surface of the metal casing of the explosion-proof distribution box. Geometric optics calculations ensure that the field of view completely covers the maximum projected area of ​​the casing. For example, for a distribution box with dimensions of 1 meter × 0.8 meters, a thermal imager lens with a horizontal field of view of not less than 60 degrees is selected. The operating parameters of the thermal imager are set as follows: thermal sensitivity of 0.05℃ at 30℃, spectral response range of 8μm to 14μm, and image resolution of 640×480 pixels. Data acquisition uses a fixed time interval trigger mode, for example, acquiring one frame of infrared thermal radiation image every 200 milliseconds. The process of converting each frame of image into temperature matrix data includes the conversion calculation from radiation value to temperature. The material emissivity parameter in the conversion formula is preset according to the casing material: 0.85 for carbon steel, 0.60 for stainless steel, and 0.10 for aluminum alloy. Background temperature compensation is corrected using real-time measurements from an ambient temperature sensor.

[0085] The triaxial magnetic field sensors are deployed using a gridded layout with equal spacing. The minimum number of sensors is calculated based on the surface area of ​​the explosion-proof distribution box's metal casing: spacing should not exceed 150 mm, for example, at least 48 sensor points per square meter. Each sensor point is fixed to the casing surface using an explosion-proof mounting base, with a non-magnetic insulating pad between the base and the casing to eliminate the influence of mechanical stress. The sensor power supply uses intrinsically safe circuitry, with the operating voltage limited to below 12VDC. Synchronous acquisition of the triaxial magnetic field sensors is achieved through a hardware trigger signal: the main controller sends a synchronization pulse signal, and all sensors synchronously start sampling within 5 microseconds of receiving the pulse. Each sensor outputs three-dimensional magnetic field data corresponding to the compensated leakage current signal, including measurements in three orthogonal directions: X-axis magnetic field component, Y-axis magnetic field component, and Z-axis magnetic field component. The measurement range covers ±50 mT, with a resolution of 0.1 μT.

[0086] The timestamp-aligned storage mechanism is driven by a high-precision clock source. The main controller has a built-in temperature-compensated crystal oscillator with a clock accuracy better than one part per million. Each data packet is marked with a 64-bit timestamp: the infrared thermal imager records a timestamp at the moment the exposure is completed, and the three-axis magnetic field sensor records a timestamp at the moment the sampling window closes. During storage processing, a time-aligned queue is established: based on the timestamp of the earliest arriving data packet, all temperature distribution data and three-dimensional magnetic field component data arriving within 200 milliseconds are grouped into the same data frame. The data frame storage format is defined as follows: the frame header is marked with a 4-byte synchronization word, the temperature data area contains 307,200 floating-point numbers (corresponding to a 640×480 matrix), the magnetic field data area stores the three-axis component values ​​of each point in the order of sensor number, and a 32-bit cyclic redundancy check code is appended to the frame tail.

[0087] The infrared thermal imager's field-of-view calibration is performed with laser positioning assistance. In calibration mode, a 650nm red laser projects markers at the four corners of the housing, and the thermal imager's pan-tilt unit is adjusted so that all four markers are located 10 pixels inside the image edge. The thermal imager's focus control employs an autofocus algorithm: a high-contrast area on the housing surface (e.g., the nameplate location) is selected, and the motorized focusing lens is driven by a contrast evaluation function. The focus is locked when the function value reaches its peak. During continuous operation, an autofocus check is performed every 24 hours, triggering a refocusing process when image sharpness decreases by more than 10%.

[0088] The field calibration of the triaxial magnetic field sensor is assisted by a standard magnetic field source. During calibration, a Helmholtz coil is set up near the distribution box, and a standard current of 10mA is passed through it to generate a uniform magnetic field of 1mT. Each sensor is rotated sequentially to three orthogonal directions, and the deviation coefficient between the output value of each axis and the theoretical value is recorded. The calibration data is stored in the sensor's local memory, and subsequent measurement data is multiplied by a correction matrix in real time. For example, when the X-axis sensitivity deviation is +1.2%, the original measured value is multiplied by a correction factor of 0.988.

[0089] Data storage integrity is ensured using a dual-buffering mechanism. Two 512MB storage blocks are allocated in memory: when block 1 receives data, block 2 performs a solid-state drive (SSD) write operation. Industrial-grade SSDs are used as the storage medium, with wear leveling algorithms extending their lifespan. Storage directories are named according to time, for example, the format 20240805-120000 indicates a data collection task that started at 12:00 on August 5, 2024. Every 24 hours, data is automatically packaged, compressed, and transferred to the network storage server.

[0090] Environmental interference suppression measures include multi-level protection. A compressed air dust removal system is deployed for the thermal imager lens, spraying clean airflow for 0.2 seconds every 60 seconds to prevent dust accumulation. The magnetic field sensor signal transmission uses twisted-pair shielded cable with a twist pitch of less than 20 mm, and the shielding layer is grounded at a single point on the controller. An electromagnetic shielding fence is installed 3 meters around the distribution box, using a continuous grounded shielding body constructed from 1 mm thick galvanized steel plate.

[0091] Real-time diagnostics of equipment status. Thermal imager status monitoring includes: activating thermoelectric cooling when the lens temperature exceeds 50°C, and triggering gain compensation when the average image brightness falls below 50 gray levels. Magnetic field sensor monitoring includes: generating an alarm when the sensor temperature exceeds 85°C, and marking an anomaly when the effective value of the signal noise exceeds 5% of the range. All diagnostic data is logged to a separate log file once per minute.

[0092] The spatiotemporal synchronization accuracy was verified using an oscilloscope test method. A high-impedance probe was connected in parallel to the exposure signal line of the infrared thermal imager and the trigger line of the magnetic field sensor, and the time difference between the two was measured using a digital oscilloscope. When the synchronization error exceeded 100 microseconds, the clock compensation parameters of the main controller were adjusted. Long-term stability tests showed that the maximum synchronization error did not exceed 250 microseconds within a temperature range of -20℃ to +65℃.

[0093] The data acquisition anomaly handling process includes a three-level response: Level 1 anomaly is the loss of a single frame of data, which is automatically filled by inserting the previous frame of data; Level 2 anomaly is the loss of five consecutive frames, which triggers a device restart; Level 3 anomaly is a hardware failure, which switches to a redundant backup device. The switching process ensures data continuity, with a maximum interruption time of no more than 500 milliseconds.

[0094] S4. Extract the migration acceleration of the high-temperature region based on the temperature distribution, and calculate the magnetic field polarization angle offset based on the three-dimensional magnetic field components; when the product of the migration acceleration and the polarization angle offset exceeds the preset collaborative failure threshold, a compensation failure flag is generated and the calibration mode is started. The specific implementation is as follows:

[0095] When extracting migration acceleration of high-temperature regions based on temperature distribution, the first step is to identify continuous high-temperature regions. From the continuous time-series temperature distribution data, image processing algorithms identify connected regions where the temperature exceeds a preset temperature threshold. The preset temperature threshold is determined based on the safe operating temperature of the metal casing material of the explosion-proof distribution box; for example, the threshold is set to 80℃ for carbon steel. The identification process includes three steps: First, median filtering is applied to the temperature matrix to eliminate noise, with a filtering window size of 5×5 pixels. Second, temperature threshold binarization is applied, marking pixels above the threshold as 1 and the rest as 0. Third, morphological closing operations are performed to connect adjacent high-temperature pixels, forming continuous high-temperature regions. The coordinates of the contour boundary points of each region are recorded; regions with an area less than 100 square millimeters are considered noise points and filtered out.

[0096] When calculating the displacement change of the centroid in the high-temperature region, a spatial centroid algorithm is used. For the high-temperature region identified in frame t, the centroid coordinates are calculated by a weighted average: the X-coordinate equals the sum of the X-coordinate values ​​of all pixels in the region multiplied by their corresponding temperature values, divided by the sum of the temperature values; the Y-coordinate is calculated similarly. The centroid displacement change at adjacent time points is calculated using the Euclidean distance formula: the displacement change equals the spatial straight-line distance between the centroid coordinates of frame t and frame t-1. A fixed sampling period value is used for the time interval, for example, 200 milliseconds. The calculation process includes an outlier removal mechanism: when the displacement change exceeds three times the average of the previous five measurements, it is considered a positioning error and interpolation processing is initiated.

[0097] The migration acceleration in the high-temperature region is calculated based on a classical kinematic model. The input parameters are the displacement change Δs and the time interval Δt. The acceleration a is calculated using the formula a = 2Δs / (Δt)². The calculation results are expressed in meters per second squared, with values ​​typically ranging from 0.01 m / s² to 5 m / s². The calculation process includes direction determination: the migration direction angle is determined based on the trend of the centroid coordinate changes; for example, the positive X-axis is 0 degrees, increasing to 360 degrees counterclockwise. The acceleration vector is decomposed into X and Y components and stored for subsequent directional correlation analysis.

[0098] During the extraction and processing of 3D magnetic field component data, data separation is performed for each triaxial magnetic field sensor point. Each sampling point data packet contains three independent fields: the X-axis magnetic field component is stored as a 32-bit floating-point number, the Y-axis magnetic field component is stored as a 32-bit floating-point number, and the Z-axis magnetic field component is stored as a 32-bit floating-point number. The extraction operation retains the X-axis and Y-axis magnetic field component fields, while the Z-axis magnetic field component is moved to a buffer for later use. The extraction process includes validity verification: when any magnetic field component exceeds the range of ±50mT, the data at that point is marked as invalid, and interpolation compensation using adjacent points is employed.

[0099] The instantaneous magnetic field polarization angle is calculated using a planar magnetic field vector analysis method. For each valid sensor point, its X-axis magnetic field component Bx and Y-axis magnetic field component By are read, and the polarization angle θ is calculated using the arctangent function: θ = arctan(By / Bx). The calculation result is expressed in radians, covering a range from -π to +π. Quadrant correction is automatically adjusted based on the signs of Bx and By: when Bx < 0 and By > 0, θ = θ + π; when Bx < 0 and By < 0, θ = θ - π. The calculation result retains four decimal places of precision; for example, when Bx = 1.2 mT and By = 0.8 mT, θ ≈ 0.5880 radians.

[0100] The statistical processing of magnetic field polarization angle offset is performed using a fixed time window. The preset time window length is set according to the system response requirements, for example, 1 second corresponding to 5 sampling periods. Within each window period, the instantaneous magnetic field polarization angle sequence of all sensor points is collected, and its standard deviation σ is calculated. The standard deviation is calculated using an unbiased estimation formula: first, the arithmetic mean θavg of all angle values ​​within the window is calculated; then, the sum of squared deviations of each angle value θi from θavg is calculated; finally, the square root is taken after dividing by the number of samples minus 1. The calculated result σ is the magnetic field polarization angle offset, with the unit being radians, consistent with the polarization angle, and a typical value range between 0.01 rad and 0.5 rad.

[0101] The collaborative failure determination process involves multiplication and threshold comparison. The migration acceleration *a* and magnetic field polarization angle offset *σ* in the current high-temperature region are read, and the product value *P* = *a* × *σ* is calculated. The preset collaborative failure threshold is determined based on experimental calibration data; for example, a typical value is set to 0.25 m / rad·s². The comparison process employs a hysteresis comparison strategy: when *P* first exceeds the threshold, a warning state is marked; if it exceeds the threshold for three consecutive cycles, it is considered a valid event. When the determination condition is met, two operations are performed: a compensation failure flag (binary 1) is written to the system status register, and a start pulse signal is sent to the calibration control unit.

[0102] The calibration mode startup process consists of three stages. The first stage freezes the current data processing thread and saves intermediate calculation results to non-volatile memory. The second stage calls the S2 compensation factor reconstruction program: it re-queries the material parameter mapping table, corrects the resistivity parameter based on the current temperature, remeasures the shell thickness, updates the compensation weight coefficients, and generates a new eddy current compensation factor. The third stage resumes data acquisition; the first 10 sampling cycles use a mixed transition strategy of old and new compensation factors to avoid signal jumps. The entire calibration process is completed within 500 milliseconds.

[0103] A temperature compensation mechanism for high-temperature zone identification adjusts the threshold in real time. The ambient temperature (Tenv) is monitored by a temperature sensor. When Tenv changes by more than ±5°C, the preset temperature threshold is adjusted proportionally: for example, for every 10°C increase in ambient temperature, the threshold for carbon steel is lowered by 2°C. The correction coefficient is stored in a temperature-threshold relationship table, which is calibrated through material temperature rise tests: measuring the gradient of safe contact temperature changes on different material surfaces within an ambient temperature range of 30°C to 70°C.

[0104] Anomaly handling for polarization angle calculations is implemented to address specific scenarios. When both Bx and By are less than 0.1 mT, the magnetic field strength is deemed too low, and the effective value from the previous cycle is used as a substitute. When the sensor point failure rate exceeds 30%, a sparse data reconstruction algorithm is activated: a spatial magnetic field distribution model is established based on the effective point data, and the failure point data is calculated through interpolation. The reconstruction algorithm employs an inverse distance weighting method, with a weighting exponent set to 2.0 and a search radius set to 150 mm.

[0105] The dynamic adjustment of the collaborative failure threshold is based on historical operational data. The system analyzes the distribution of P-values ​​in the most recent 100 determined events. When 90% of the P-values ​​in events determined to be failures are concentrated in the range of 1.2 to 1.5 times the threshold, the threshold is automatically increased by 5%. A safety boundary is set during the adjustment process: the upper limit of the threshold does not exceed 150% of the initial value, and the lower limit is not lower than 50%. Each adjustment generates an operation log, including a timestamp, the original threshold, the new threshold, and the basis for the adjustment.

[0106] Performance verification of the calibration mode is performed immediately after completion. A standard test signal is injected into the system, the compensated signal parameters are measured, and the results are compared with the data before calibration. Calibration is considered successful when the following conditions are met: signal gain change is less than 5%, phase shift is less than 1 degree, and harmonic distortion rate is reduced by more than 10%. If verification fails, a secondary calibration process is triggered: switching to the backup sensor network to reacquire data and repeating the complete calibration steps.

[0107] Spatiotemporal data correlation is ensured through a unified clock source. The timestamp for temperature distribution data is based on the end time of the thermal imager exposure, while the timestamp for magnetic field data is based on the center time of the sampling window; both are aligned using a time calibration curve. The calibration curve is determined during initial system installation: a standard timescale generator is used to send synchronization pulses, the response delay of each device is recorded, and a time offset compensation table between devices is established. During operation, the time offset is remeasured every 24 hours, and the compensation table data is updated.

[0108] The application scope of the compensation failure flag covers subsequent processing procedures. When the flag is in position, the system performs three linked operations: suspends the S5 curl integral calculation task, forwards the original S3 magnetic field data to the fault diagnosis port, and displays a red alarm indicator on the human-machine interface. The flag remains in this state until calibration is completed and verification is passed, at which point it is automatically reset or manually cleared by maintenance personnel.

[0109] S5. When the compensation failure flag is not triggered, obtain the spatial magnetic field distribution of the compensation leakage signal, and calculate the area fraction of the curl of the spatial magnetic field distribution on the shell surface. Specifically, the implementation is as follows:

[0110] When performing spatial magnetic field distribution processing without triggering the compensation failure flag, the spatial magnetic field vector distribution corresponding to the compensation leakage signal is first acquired. Data is simultaneously collected by multiple triaxial magnetic field sensors evenly spaced on the surface of the explosion-proof distribution box's metal casing. The sensor spacing is set to a range of 100 mm to 150 mm based on the casing's radius of curvature. Each sensor output contains magnetic field measurements in three orthogonal directions: the X-axis magnetic field component records the magnetic field strength along the length of the distribution box, the Y-axis magnetic field component records the component along the width, and the Z-axis magnetic field component records the component perpendicular to the casing surface. A fifth-order low-pass filter is used during the acquisition process to suppress high-frequency noise, with a cutoff frequency set to 500 Hz. A data validity verification mechanism monitors in real time: when the output fluctuation of a single sensor exceeds three times the average of the three adjacent sensors, that point is marked as invalid, and interpolation compensation for adjacent points is initiated.

[0111] When calculating the curl vector of the magnetic field at various spatial locations based on the spatial magnetic field vector distribution, the central difference approximation method is used. For any point P on the shell surface, six adjacent sensor points are selected to form a calculation unit. The X component of the magnetic field curl vector is calculated as follows: it equals the partial derivative of the difference in the Z-axis magnetic field components between adjacent points in the Y direction minus the partial derivative of the difference in the Y-axis magnetic field components in the Z direction. Since all points are located on the shell surface, the Z-direction partial derivative is simplified to the calculation of the height difference between adjacent points. The size of the calculation unit is controlled within 50 mm × 50 mm to ensure that the spatial resolution meets the curl calculation requirements. The calculation results are stored in a three-dimensional vector format, with each component retaining four significant digits, and the unit is amperes per square meter.

[0112] When discretizing the surface of the metal casing of the explosion-proof distribution box into several surface elements, an unstructured mesh generation algorithm is used. Based on the 3D CAD model of the casing, surface triangulation is performed, with the maximum side length of the triangular element limited to 0.8 times the sensor spacing, for example, 120 mm. The element vertices are precisely matched with the positions of the magnetic field sensors. When the sensor distribution does not meet the vertex requirements, virtual nodes are inserted in areas with drastic changes in surface curvature. Discretization quality verification indicators: the interior angle of the element is controlled between 30 and 120 degrees, and the area ratio of adjacent elements does not exceed 2.0. Finally, a surface element topology table is generated, recording the vertex coordinates, adjacent element indices, and normal vector direction of each element.

[0113] When calculating the normal component of the curl vector of the magnetic field within each surface element, a vector projection operation is performed. First, the normal direction of the surface element is determined: two edge vectors are calculated using the coordinates of the three vertices of the element, then the normal vector is obtained through a vector cross product, and finally normalized to a unit vector. The normal component is calculated as a dot product of the curl vector and the unit normal vector. The dot product result is a scalar value; the sign indicates the relationship between the curl direction and the normal: positive values ​​point outwards, and negative values ​​point inwards. The calculation result is multiplied by the surface element area: the area value is obtained using Heron's formula based on the three side lengths of a triangular element, retaining a precision of one square millimeter. The product result represents the contribution of that element to the surface integral value.

[0114] The summation of the product of all surface elements is achieved using a double-precision floating-point accumulator. The accumulation process employs an order-independent algorithm: all elements are first partitioned according to their spatial location, each partition is accumulated independently, and then the sums are made up. The summation algorithm includes an error control mechanism: if the accumulated value exceeds 1 × 10⁻⁶, an error will occur. 6 The data is automatically converted to scientific notation for storage to avoid loss of precision. The final area fraction of the curl of the spatial magnetic field distribution on the shell surface is a scalar value with dimensions in amperes-meters, typically ranging from 0.01 A·m to 10 A·m. The summation result is stored in a circular buffer for subsequent time series analysis.

[0115] Temperature compensation for magnetic field curl calculation is achieved through a real-time correction algorithm. Each sensor point is associated with temperature monitoring data; when the sensor temperature changes by more than ±10℃, the original magnetic field reading is corrected according to a preset temperature-sensitivity curve. The compensation curve is established through laboratory calibration: in a temperature chamber ranging from -25℃ to +85℃, the sensor's response deviation to a standard magnetic field is measured at each temperature point in 10℃ increments. For example, at +70℃, the sensor near the carbon steel housing requires an additional 1.2% gain compensation.

[0116] The dynamic optimization of surface discretization automatically adjusts according to curvature changes. The system calculates the Gaussian curvature distribution of the shell surface in real time and automatically refines the mesh in areas with a curvature radius of less than 200 mm: triangular elements are split into four sub-elements, and the normal vector of the new vertex is taken as the average of the normal vectors of the adjacent faces. The maximum recursion depth of the refinement operation is three levels to ensure that the total number of elements does not exceed five times the initial number. The mesh optimization trigger condition is: when the angle between the normal vectors of adjacent elements exceeds 15 degrees, it is identified as a high curvature region.

[0117] Special rules are set for boundary treatment in the calculation of normal components. For surface elements at the edge of the distribution box, a semi-element correction method is used: virtual mirror elements are added outside the shell boundary line, and the magnetic field curl vector of the mirror element is taken as the boundary symmetry value. For open areas (such as cable inlets and outlets), a closed loop is generated along the hole boundary. The width of the loop is equal to the standard element size, and the area inside the loop is not included in the integration calculation.

[0118] Error control for area fraction calculation includes triple verification. Level 1 verification: Calculate the average angle between the normal vectors of all surface elements and the Z-axis; if it exceeds 60 degrees, coordinate system correction is triggered. Level 2 verification: Calculate the standard deviation of each element's contribution; if it exceeds five times the average, mark the element as abnormal. Level 3 verification: Compare the current integral value with the rate of change from the previous period; if it exceeds 50%, a recalculation is initiated. All verification anomalies are logged in detail for analysis.

[0119] Parallel acceleration of data processing is achieved through task partitioning. The shell surface is divided into eight sector regions, each assigned an independent computation thread. Thread synchronization mechanism: A synchronization barrier is set during the integration and summation phase, waiting for all regions to complete their sub-integrations before performing global accumulation. Dynamic degradation is initiated when computational resources are insufficient: only 50% of the key sensor points are selected for simplified calculations, with the key points selected based on the top 50% of magnetic field strength.

[0120] Active cancellation technology is employed to suppress environmental electromagnetic interference. Reference magnetic field sensors are deployed around the distribution box to monitor the ambient background magnetic field. The background magnetic field component is subtracted from the raw data before calculating the curl; the background magnetic field is updated every 200 milliseconds. For power frequency interference, an adaptive notch filter is applied, with the notch center frequency tracking the actual frequency of the power grid, and a bandwidth set to ±0.5 Hz.

[0121] The unit conversion of the integration result is adjusted according to the application scenario. When used for leakage current detection, it retains the ampere-meter unit; when used for the display interface, it is converted to milliweber-meter with a conversion factor of multiplied by 1000. The displayed value is rounded to three decimal places, and the range is automatically switched and a dimensional suffix is ​​added when the display exceeds the range.

[0122] Real-time performance of the computation process is ensured through time slicing management. The entire S5 process is limited to a maximum completion time of 150 milliseconds, with 20 milliseconds for magnetic field acquisition, 50 milliseconds for curl calculation, 30 milliseconds for discretization, and 50 milliseconds for integration and summation. Timeout handling mechanism: When the actual time exceeds the limit, the mesh resolution is automatically reduced and recalculated, down to a minimum of 40% of the initial cell count. Each degradation operation generates a warning-priority runtime log.

[0123] The storage format for area fractions includes multidimensional metadata. The main data area records the integral value as a floating-point number; the extended data area includes: calculation timestamp, number of sensors involved in the calculation, proportion of invalid points, total number of grid cells, and version number of the temperature compensation coefficient. Data packets are stored in a timestamped circular queue with a queue depth of 1000 entries, covering historical data older than 200 seconds.

[0124] S6. When the gradient of the area fraction change over time exceeds the preset vortex mutation threshold, a leakage fault command is output to the tripping actuator. The specific implementation is as follows:

[0125] When performing leakage fault diagnosis, the area fraction of the curl of the spatial magnetic field distribution over a continuous time series on the shell surface is first obtained. The area fraction data sequence from the most recent sampling periods is extracted from the data buffer output in step S5. The number of sampling periods is set to a range of 20 to 50 periods based on system response requirements. The continuity of timestamps is verified during data extraction; if an abnormal time interval is detected, linear interpolation is used to supplement data points. The extracted data is stored in an array structure, containing timestamps accurate to milliseconds and corresponding area fraction values, with the area fraction unit maintained in ampere-meters.

[0126] When calculating the difference in area fractions between adjacent time points, the data sequence is processed in chronological order. For the area fraction at time point i, the difference is obtained by subtracting the area fraction at time point i-1. A positive difference indicates an increase in the area fraction, while a negative difference indicates a decrease. The time interval is obtained by subtracting the timestamps of two time points, accurate to 0.1 milliseconds. The calculation process includes an anomaly filtering mechanism: when the time interval exceeds the allowable deviation range of the standard sampling period, the calculation result is discarded. If three consecutive calculations are invalid, the data source diagnostic program is activated.

[0127] The rate of change of area fraction over time is obtained through division. The difference in area fraction between adjacent time points is divided by the corresponding time interval, and the result is expressed in amperes per second (amperes per meter per second). To suppress measurement noise, the original calculation result is smoothed: the arithmetic mean of five consecutive calculated values ​​is taken as the final rate of change. A division-by-zero error protection mechanism ensures a minimum time interval constraint and avoids mathematical calculation errors.

[0128] The calibration of the preset eddy current mutation threshold is based on an electromagnetic shielding characteristic experiment of the metal casing of an explosion-proof distribution box. The experimental procedure involves preparing casing samples of typical materials, injecting simulated leakage current with stepped variations into the internal conductors, and recording the correlation between the rate of change of the area fraction and the leakage current. The threshold is set to 80% of the minimum rate of change that causes a sudden change in the magnetic field distribution pattern; different materials correspond to different threshold levels. Calibration data is stored in non-volatile memory. During field operation, the corresponding threshold can be retrieved based on the actual casing material, allowing operators to fine-tune the threshold within a limited range to adapt to special operating conditions.

[0129] The leakage fault command is generated using comparison logic with a confirmation mechanism. A warning state is recorded when the rate of change per unit time first exceeds a preset vortex mutation threshold; if the rate of change exceeds the threshold for three consecutive sampling periods, a high-level valid leakage fault command signal is generated. The state is automatically locked after command generation until manually reset by maintenance personnel. Simultaneously, data records for a period before and after the triggering time are saved for subsequent event analysis.

[0130] Electrical isolation measures are employed when the leakage fault command is transmitted to the tripping actuator. The control signal is converted from a low-voltage level to a drive level and connected to the circuit breaker control circuit via a shielded cable. The transmission process includes a signal enhancement stage to ensure reliable long-distance transmission. The execution sequence that triggers the tripping action includes three stages: sending a start signal, status detection, and result feedback. Each stage has clear time constraints and status verification requirements.

[0131] The threshold dynamic adjustment mechanism periodically optimizes the set value. It statistically analyzes the characteristic data of historical triggering events, automatically adjusting the threshold when the rate of change in actual leakage events falls within a specific range, while simultaneously monitoring the false trigger rate for reverse correction. Adjustment operations set safety boundary limits, and each change generates a detailed operation record.

[0132] Environmental adaptability processing for the rate of change per unit time includes temperature compensation. The ambient temperature of the distribution box and sensor temperatures are monitored in real time, and a compensation coefficient is applied to the rate of change when the temperature exceeds the standard range. The compensation parameters are determined through environmental simulation tests, and independent correction relationships are established for different temperature points.

[0133] Multiple technical measures are employed to ensure the reliability of signal transmission. The transmission line adopts an anti-interference design, adds a data verification mechanism, and sets up a timeout monitoring function. When the main channel fails, it automatically switches to the backup transmission path to ensure that commands are delivered reliably.

[0134] Monitoring the execution status of the tripping operation is achieved through current characteristic analysis. The trip coil current waveform parameters are detected, including rise time, peak level, and duration. When abnormal waveform characteristics are detected, a maintenance prompt message is generated.

[0135] Historical data management employs a cyclical storage strategy. Change rate data is stored at a fixed frequency, ensuring that data within the most recent time range remains searchable. Data related to fault events is permanently stored and supports multi-condition retrieval and analysis.

[0136] Real-time status display provides visual monitoring. The human-machine interface dynamically displays the current rate of change, threshold level, historical extreme values, and actuator status. Color-coded indicators show how close the threshold is to being reached, and the refresh rate meets the operator's observation needs.

[0137] Data anomaly recovery employs predictive compensation technology. When consecutive data loss exceeds a set threshold, missing values ​​are predicted and supplemented based on data trend changes; the prediction results are used solely for display purposes.

[0138] The electrical characteristics of the tripping actuator meet safety standards. It employs a retaining relay structure, and the contact capacity meets high-voltage breaking requirements. The relay control circuit is equipped with electromagnetic shielding to prevent interference.

[0139] The event log records the complete operation process. Each time a leakage fault command is triggered, a structured log containing precise time, key parameters, environmental conditions, and execution results is recorded. The log file is archived chronologically to meet long-term traceability requirements.

[0140] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0141] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0142] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0143] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0144] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0145] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0146] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0147] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0148] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0149] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A leakage current detection method for explosion-proof distribution boxes, characterized in that, include: S1. Obtain the original leakage current signal of the wires inside the explosion-proof distribution box; S2. Based on the material and thickness of the metal casing of the explosion-proof distribution box, an eddy current compensation factor is constructed to compensate the gain of the original leakage current signal, generating a compensated leakage current signal, including: The resistivity and permeability of the metal shell material of the explosion-proof distribution box can be obtained by querying the pre-stored material parameter mapping table. The reference eddy current attenuation coefficient is calculated based on resistivity and permeability. The compensation weight coefficient is determined based on the thickness of the metal shell of the explosion-proof distribution box, wherein the compensation weight coefficient increases non-linearly with the increase of thickness; The eddy current compensation factor is generated by multiplying the reference eddy current attenuation coefficient by the compensation weight coefficient. The original leakage signal is amplified by a gain factor corresponding to the eddy current compensation factor to generate a compensated leakage signal. S3. Collect the temperature distribution on the surface of the metal shell of the explosion-proof distribution box and obtain the three-dimensional magnetic field component corresponding to the compensation leakage signal; S4. Extract the migration acceleration of the high-temperature region based on the temperature distribution, and calculate the magnetic field polarization angle offset based on the three-dimensional magnetic field components; when the product of the migration acceleration and the polarization angle offset exceeds the preset collaborative failure threshold, generate a compensation failure flag and start the calibration mode. S5. When the compensation failure flag is not triggered, acquire the spatial magnetic field distribution of the compensation leakage signal, and calculate the area fraction of the curl of the spatial magnetic field distribution on the shell surface, including: The spatial magnetic field vector distribution corresponding to the compensation leakage signal is obtained by multiple triaxial magnetic field sensors deployed on the surface of the metal shell of the explosion-proof distribution box. Calculate the magnetic field curl vector at each spatial location point based on the spatial magnetic field vector distribution; The surface of the metal shell of the explosion-proof distribution box is discretized into several surface units; Calculate the product of the normal component of the curl vector of the magnetic field within each surface unit and the area of ​​the surface unit; Summing the product of all surface units yields the area fraction of the curl of the spatial magnetic field distribution on the shell surface. S6. When the gradient of the area fraction change over time exceeds the preset vortex mutation threshold, a leakage fault command is output to the tripping actuator.

2. The leakage current detection method for explosion-proof distribution boxes according to claim 1, characterized in that, Obtain the original leakage current signal of the wires inside the explosion-proof distribution box, including: Install the zero-sequence current transformer around the grounding neutral wire of the three-phase conductors inside the explosion-proof distribution box; The signal conditioning circuit performs bandpass filtering on the induced signal output by the zero-sequence current transformer to filter out high-frequency electromagnetic interference components and low-frequency drift components. The passband of the bandpass filtering process covers the fundamental frequency and integer harmonic frequencies of the power frequency. The filtered signal is converted from analog to digital to generate the original digitized leakage current signal.

3. The leakage current detection method for explosion-proof distribution boxes according to claim 1, characterized in that, The material parameter mapping table includes resistivity and permeability data for carbon steel, stainless steel, and aluminum alloy.

4. The leakage current detection method for explosion-proof distribution boxes according to claim 1, characterized in that, The temperature distribution on the surface of the metal casing of the explosion-proof distribution box is collected, and the three-dimensional magnetic field components corresponding to the compensation leakage signal are obtained, including: Infrared thermal radiation images of the surface of the metal shell of the explosion-proof distribution box are acquired at fixed time intervals using an infrared thermal imager to generate temperature distribution data. Multiple triaxial magnetic field sensors are arranged at equal intervals on the surface of the metal shell of the explosion-proof distribution box to simultaneously acquire the X-axis magnetic field component, Y-axis magnetic field component and Z-axis magnetic field component corresponding to the compensation leakage signal. The temperature distribution data collected at the same time are timestamped and stored in alignment with the three-dimensional magnetic field components.

5. The leakage current detection method for explosion-proof distribution boxes according to claim 4, characterized in that, The infrared thermal imager's field of view covers the entire surface of the metal casing of the explosion-proof distribution box.

6. The leakage current detection method for explosion-proof distribution boxes according to claim 1, characterized in that, The migration acceleration of the high-temperature region is extracted based on the temperature distribution, and the magnetic field polarization angle offset is calculated based on the three-dimensional magnetic field components. When the product of migration acceleration and polarization angle offset exceeds a preset collaborative failure threshold, a compensation failure flag is generated and calibration mode is initiated, including: Identify continuous high-temperature regions in the temperature distribution of a continuous time series where the temperature exceeds a preset temperature threshold; Calculate the displacement change of the centroid of the high-temperature region at adjacent time points; Calculate the migration acceleration in the high-temperature region based on the displacement change and time interval; Extract the X-axis and Y-axis magnetic field components from the three-dimensional magnetic field components; The instantaneous magnetic field polarization angle is calculated based on the ratio of the X-axis magnetic field component to the Y-axis magnetic field component. The standard deviation of the instantaneous magnetic field polarization angle within the preset time window is used as the magnetic field polarization angle offset. When the product of migration acceleration and magnetic field polarization angle offset exceeds the preset collaborative failure threshold, a compensation failure flag is generated and calibration mode is initiated.

7. The leakage current detection method for explosion-proof distribution boxes according to claim 1, characterized in that, When the gradient of the area fraction change over time exceeds the preset vortex mutation threshold, a leakage fault command is output to the tripping actuator, including: Obtain the area fraction of the curl of the spatial magnetic field distribution over a continuous time series on the shell surface; Calculate the difference in area fractions at adjacent time points; Divide the difference by the corresponding time interval to obtain the rate of change of the area fraction over time. When the rate of change per unit time exceeds the preset vortex mutation threshold, a leakage fault command is generated. The leakage fault command is transmitted to the tripping actuator to trigger the circuit breaker to trip.

Images