Method for calculating and identifying abnormality of multi-loop current unbalance degree of distribution box

Abstract

The application discloses a power distribution box multi-loop current unbalance degree calculation and abnormality discrimination method, relates to the field of power distribution box multi-loop online monitoring, and defines three sections of the head, the middle and the tail in the target loop busbar measurement groove and obtains corresponding differential magnetic field quantities; a space weight coefficient is introduced to perform weighted summation on the three-section differential magnetic field quantities to extract a longitudinal even component, and to perform opposite difference on the head-tail differential magnetic field quantities to extract a longitudinal odd component; the even component is used to reconstruct a real measurement current and calculate a multi-loop current unbalance degree; the unbalance degree and the odd component maximum value are searched respectively to lock candidate loops, and multi-dimensional cross-cause operation is performed in combination with double alarm threshold values. The application completely decouples the real current fluctuation and the false differential interference caused by local structure distortion at the data bottom layer, accurately discriminates real load imbalance and geometric pseudo-unbalance, and effectively eliminates false alarms under complex working conditions.

Description

Technical Field

[0001] This invention relates to the field of online monitoring of multiple circuits in distribution boxes, specifically to a method for calculating and identifying the current imbalance of multiple circuits in distribution boxes. Background Technology

[0002] In the evolution of power distribution systems towards digitalization and intelligence, high-density loop monitoring has become a core element in ensuring industrial production safety and refined power management. Especially in high-load power consumption scenarios such as semiconductor manufacturing, large computing centers, and precision industrial processing, real-time sensing of branch current status directly affects system operational safety and energy efficiency optimization. Currently, online monitoring methods based on the principle of electromagnetic induction are widely used. This method typically involves creating a narrow measuring slot in the measurement area of ​​the current-carrying busbar, causing the loop current to flow around the sides of the slot and generating a spatial magnetic field gradient, thereby indirectly obtaining the current intensity while maintaining electrical insulation strength. To suppress stray electromagnetic field interference in the environment, existing solutions often place a sensor array in the middle of the measuring slot, using the differential signal at this location to cancel far-end common-mode noise and calculate the loop current-carrying capacity.

[0003] The actual machining accuracy of the busbar measuring channel is affected by punching process deviations, edge burrs, or residual conductive debris, making it difficult to achieve the ideal geometric symmetry. If there are local residual conductive connection structures at the edge of the measuring channel, the topology of the current-carrying path in the local area of ​​the channel will be significantly deflected. This type of structural deviation causes the magnetic field distribution at the beginning and end of the measuring channel to exhibit an opposite shift, and this pattern is significantly different from the unidirectional distribution under current-driven conditions. Existing single-section monitoring logic can only acquire a one-dimensional signal at a specific location in the middle of the channel, and cannot analyze the distribution of the magnetic field in the longitudinal space, resulting in false signals caused by local structural defects being superimposed on the main current-carrying signal. When the power distribution system is in multi-circuit operation, the monitoring equipment has difficulty accurately determining the source of magnetic field fluctuations, and easily misinterprets minor structural deviations as branch load imbalances. Due to the lack of effective attribution identification methods, the early warning reliability and fault location accuracy of multi-circuit monitoring systems under complex operating conditions face severe challenges. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a method for calculating and identifying the current imbalance of multiple circuits in distribution boxes. This method solves the problem that existing single-section monitoring schemes cannot effectively isolate and distinguish between true load imbalance and geometric pseudo-imbalance caused by local structural defects, leading to incorrect fault attribution and low reliability of early warning under complex operating conditions.

[0005] To achieve the above objectives, the present invention provides the following technical solution: For each target circuit in the entire monitoring circuit in the distribution box, the first section, the middle section and the last section of the target circuit busbar measurement slot are defined. Measurement points are set up on both sides of each section to obtain the original magnetic field quantity. The original magnetic field quantity is subjected to a lateral spatial difference operation to generate the differential magnetic field quantity at the first, middle and last sections of the slot. A generalized spatial weighting coefficient with central dominance is introduced to perform weighted summation and normalization operations on the differential magnetic field quantities at the beginning, middle, and end of the slot to extract the longitudinal even components. An out-of-direction difference operation is performed on the differential magnetic field quantities at the beginning and end of the slot to extract the longitudinal odd components. Multiply the longitudinal dipole component by the preset current conversion factor to calculate the corrected actual measured current. Within the same time slice, summarize the corrected actual measured current of all monitoring loops to calculate the multi-loop average current. Place the minimum positive constant in the denominator and perform normalization deviation calculation by combining the corrected actual measured current and the multi-loop average current to calculate the multi-loop current imbalance. The maximum value of the multi-loop current imbalance is retrieved to lock the candidate loops of the actual load imbalance, and the maximum value of the absolute value of the longitudinal odd component is retrieved to lock the candidate loops of the structural pseudo-anomaly. Based on the imbalance alarm threshold and the structural distortion alarm threshold, cross-attribution operation is performed on the candidate loops of the actual load imbalance and the candidate loops of the structural pseudo-anomaly to generate the final state discrimination instruction.

[0006] Furthermore, the cross-sections at the beginning, middle, and end of the target circuit busbar measurement slot are defined, including: A three-dimensional local coordinate system is established with the center of the current inflow end edge of the target circuit busbar measuring channel as the origin, the positive current carrying direction of the busbar is defined, and the longitudinal length of the measuring channel is read. Along the positive current-carrying direction of the busbar, the section at the beginning of the measuring channel is set at one-sixth of the longitudinal length of the measuring channel, the section at the middle of the measuring channel is set at one-half of the longitudinal length of the measuring channel, and the section at the end of the measuring channel is set at five-sixths of the longitudinal length of the measuring channel.

[0007] Furthermore, the original magnetic field quantities are obtained, including: When the target circuit is in a power-off and no-load state, multiple no-load samples are continuously collected by sensors configured at the measuring points and the arithmetic average is performed to generate a static zero-point voltage, which is then stored in a non-volatile memory. During the formal operation phase, the real-time output voltage of the sensor at the measuring point is acquired synchronously, the difference between the real-time output voltage and the static zero-point voltage is calculated, and the difference is divided by the voltage of the sensor at the measuring point to the magnetic field sensitivity to generate the original magnetic field quantity.

[0008] Furthermore, weighted summation and normalization operations are performed on the differential magnetic field quantities at the beginning, middle, and end of the slot to extract the longitudinal even components, and an out-of-direction difference operation is performed on the differential magnetic field quantities at the beginning and end of the slot to extract the longitudinal odd components, including: According to the weight combination of 1:2:1, the generalized spatial weight coefficients are configured for the differential magnetic field quantities at the beginning, middle and end of the slot, respectively. The differential magnetic field quantities at the beginning, middle and end of the slot are weighted and summed. The weighted summation result is divided by the coefficient value of four to extract the longitudinal even component. Calculate the out-of-direction difference between the differential magnetic field quantity at the beginning of the slot and the differential magnetic field quantity at the end of the slot. Divide the out-of-direction difference by the coefficient value to perform amplitude normalization and extract the longitudinal odd component.

[0009] Furthermore, the preset current conversion factor is obtained, including: When the target circuit is in the factory stage or offline state, inject two known non-zero standard test currents into the target circuit in sequence, and obtain two longitudinal even component outputs corresponding to the two known non-zero standard test currents. Calculate the ratio of the difference between two known non-zero standard test currents to the difference between the two longitudinal even component outputs, calibrate based on the ratio to generate a preset current conversion factor, and write the preset current conversion factor into a non-volatile memory.

[0010] Furthermore, by placing the minimum positive constant in the denominator and performing a normalized deviation calculation on the corrected actual measured current and the multi-loop average current, the multi-loop current imbalance is calculated, including: Minimum positive constant based on the minimum non-zero calibration current configuration of the hardware system; The multi-loop current imbalance is calculated by performing a division operation with the absolute difference between the corrected actual measured current and the average current of the multi-loop as the numerator and the sum of the average current of the multi-loop and the minimum positive constant as the denominator.

[0011] Furthermore, the maxima of the multi-loop current imbalance are retrieved to identify candidate loops with actual load imbalance, and the maxima of the absolute values ​​of the longitudinal odd components are retrieved to identify candidate loops with structural pseudo-anomalies, including: If multiple loops have the same current imbalance value, select the loop with the largest corrected actual measured current value as the candidate loop for actual load imbalance; if the corrected actual measured current values ​​are still equal, select the loop with the smallest independent loop number as the candidate loop for actual load imbalance. If multiple loops have the same absolute value of longitudinal odd component, the loop with the largest sum of the absolute values ​​of the differential magnetic field at the beginning and end of the slot is selected as the candidate loop for structural pseudo-anomaly; if the sum of the absolute values ​​of the differential magnetic field at the beginning and end of the slot is still equal, the loop with the smallest independent loop number is selected as the candidate loop for structural pseudo-anomaly.

[0012] Furthermore, based on the imbalance alarm threshold and the structural distortion alarm threshold, a cross-attribution operation is performed on the candidate loops of actual load imbalance and the candidate loops of structural pseudo-anomalies to generate a final state discrimination instruction, including: If the current imbalance of the multi-circuit candidate circuit corresponding to the actual load imbalance is less than the imbalance alarm threshold, and the absolute value of the longitudinal odd component corresponding to the structural pseudo-anomaly candidate circuit is less than the structural distortion alarm threshold, the distribution box is determined to be in a global healthy balance state, and a normal operation command is output.

[0013] Furthermore, the multidimensional cross-anomaly attribution operation includes: If the actual load imbalance candidate circuit and the structural pseudo-anomaly candidate circuit are the same circuit, and the multi-circuit current imbalance degree corresponding to the actual load imbalance candidate circuit is greater than or equal to the imbalance degree alarm threshold, and the absolute value of the longitudinal odd component corresponding to the structural pseudo-anomaly candidate circuit is greater than or equal to the structural distortion alarm threshold, the actual load imbalance candidate circuit is determined to be a composite anomaly circuit. If the actual load imbalance candidate circuit and the structural pseudo-anomaly candidate circuit are the same circuit, and only one parameter in the absolute value of the multi-circuit current imbalance degree corresponding to the actual load imbalance candidate circuit and the vertical odd component corresponding to the structural pseudo-anomaly candidate circuit is greater than or equal to the corresponding alarm threshold, the actual load imbalance candidate circuit will be downgraded to a single abnormal circuit that exceeds the corresponding alarm threshold.

[0014] Furthermore, the multidimensional cross-anomaly attribution operation also includes: If the candidate loop of actual load imbalance and the candidate loop of structural pseudo-anoma are not the same loop, perform independent attribution determination operation. If the current imbalance of the multi-loop current corresponding to the candidate loop of the actual load imbalance is greater than or equal to the imbalance alarm threshold, the candidate loop of the actual load imbalance is determined to be the actual load imbalance loop. If the absolute value of the longitudinal odd component corresponding to the candidate loop of structural pseudo-anomaly is greater than or equal to the structural distortion alarm threshold, the candidate loop of structural pseudo-anomaly is determined to be a pseudo-anomaly loop of end asymmetric structural distortion.

[0015] Compared with existing technologies, it has the following advantages: The proposed method for calculating and identifying current imbalance in multi-circuit distribution boxes expands the spatial analytical dimension of electromagnetic monitoring by constructing a multi-section topology covering the beginning, middle, and end of the measurement channel. This effectively addresses the limitation of existing technologies, which cannot distinguish between current intensity fluctuations and structural deviations due to the single sampling point. By utilizing the weighting and parity decomposition logic of the signals from each section, the method deeply decouples the structural distortion component with longitudinal odd symmetry from the longitudinal even symmetry component reflecting the main current flow state of the circuit. This method eliminates the need for additional complex compensation algorithms, effectively removing geometric pseudo-imbalance signals caused by manufacturing tolerances or residual conductive bridges from the underlying data level. This significantly improves the accuracy of reconstructing the true current flow state of the circuit and enhances monitoring robustness.

[0016] Building upon this foundation, a cross-attribution mechanism based on the imbalance sequence and the longitudinal odd component sequence is introduced, establishing a correlation model between load conditions and the busbar micro-topology. This mechanism, by locking the mapping relationship between load imbalance loops and structural anomaly loops, constructs a highly discriminative logical judgment matrix, completely eliminating false alarms and attribution confusion caused by unilateral structural abrupt changes. The multi-dimensional cross-validation method enables the system to perform classification and degradation processing for composite anomalies, single load anomalies, and purely structural pseudo-anomalies, significantly reducing the false alarm rate under complex operating conditions while ensuring the determinism of early warnings. The calibration and real-time judgment closed-loop logic established in this invention provides stable and reliable technical support for the refined operation and maintenance of multiple loops within the distribution cabinet. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0018] 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.

[0019] Please see Figure 1 This application provides a method for calculating and identifying the current imbalance of multiple circuits in a distribution box; The method specifically includes the following steps: Before executing the formal monitoring steps, the system performs initialization and parameter pre-calibration. Based on two known non-zero standard test currents of the target circuit during the factory test or in offline state, and the corresponding longitudinal even component outputs, the current conversion factor of the target circuit is calibrated by calculating the ratio of the difference between the two standard test currents to the difference of the corresponding longitudinal even component. And write it to non-volatile memory, current conversion factor The geometric electromagnetic coupling relationship between the sensor array and the busbar is encapsulated. Simultaneously, based on the minimum non-zero resolvable calibration current resolution of the hardware system, the system pre-calculates and configures a minimum positive constant c and an imbalance alarm threshold. and structural distortion alarm threshold This serves as the static baseline boundary for subsequent calculations.

[0020] Step 1: In the specific implementation of the method for calculating and identifying the current imbalance of multiple circuits in the distribution box, a three-section spatial magnetic field differential array is first constructed.

[0021] For the distribution box numbered i ( The target loop is defined using a three-dimensional local coordinate system with the center of the current-carrying inlet edge of the busbar measuring slot as the origin. The measuring slot is a narrow, elongated notch created within the rectangular busbar measuring area of ​​the measured loop, used to alter the local flow path and create a measurable spatial magnetic field gradient. The positive current-carrying direction of the busbar is defined as the x-axis, the direction parallel to the transverse direction of the measuring slot as the y-axis, and the direction perpendicular to the busbar surface as the z-axis. The longitudinal length L and transverse width W of the measuring slot are read from the system's pre-configured values. Three spatial sections are defined inside the measuring slot along the x-axis: the section H at the beginning of the slot is set at the x-axis coordinate... The cross-section M in the middle of the groove is set in the x-axis coordinate as The cross-section T at the tail end of the groove is set to the x-axis coordinate as The coordinate anchoring distribution ratio aims to avoid the nonlinear magnetic field distortion region caused by the sharp geometric corners of the slot and ensure that the sampling points avoid the non-uniform leakage magnetic field region caused by the end edge effect. The coordinate space orientation does not change with the installation posture of the busbar in the distribution box. Among them, the end edge effect is affected by the end asymmetric structural distortion. End asymmetric structural distortion refers to the undesigned anomalies in the local structure of the measurement slot, such as microscopic abrupt changes such as residual conductive bridges on one side of the slot edge. Such distortion will cause the current-carrying path to deflect asymmetrically at the end of the slot.

[0022] After defining the spatial cross-section, spatial measuring points are arranged to construct a data acquisition array. Measuring points are symmetrically arranged on the left and right sides of the cross-section H at the beginning of the trench, the cross-section M in the middle of the trench, and the cross-section T at the end of the trench. The measuring points are independently numbered as k. Measuring points 1 and 2 are located on the left and right sides of the front section H along the negative y-axis and positive y-axis, respectively. Measuring points 3 and 4 are located on the left and right sides of the middle section M along the negative y-axis and positive y-axis, respectively. Measuring points 5 and 6 are located on the left and right sides of the rear section T along the negative y-axis and positive y-axis, respectively. The plane containing all measuring points is parallel to the busbar plate surface, and the normal installation height from the busbar plate surface is uniformly set to the calibration distance d. The lateral geometric gap between the magnetic field sampling point and the corresponding outer edge of the measuring tank is set as parameter g, and the value range of parameter g is configured to be between 0.1W and 0.5W. In a specific embodiment, parameter g is set to a value of The parameters g and calibrated distance d are set to balance the limitations of sensor insulation installation gaps and maximize the magnetic field gradient coupling gain.

[0023] In a preferred embodiment, a distribution cabinet containing three circuits is used as an example. The system pre-configured measurement slot for circuit 1 has a longitudinal length L of 60 mm and a transverse width W of 10 mm. The calibration distance d is set to 2 mm, and the transverse geometric clearance g is set to 2.5 mm. The system synchronously acquires six original magnetic field quantities for circuit 1 at the beginning, middle, and both sides of the end of the slot. to The values ​​are 0.40 millitrile, 1.60 millitrile, 0.50 millitrile, 1.70 millitrile, 0.40 millitrile, and 1.60 millitrile, respectively. A transverse spatial subtraction operation is performed to calculate the three differential magnetic field quantities of loop 1. , as well as All are 1.20 millitriles. Similarly, the three differential magnetic field quantities of the abnormal loop 2 were collected and calculated as follows: 1.40 milliseconds For 1.50 milliseconds and It is 1.60 milliseconds.

[0024] After establishing the acquisition array, the original magnetic field quantity at each measuring point is obtained after zero-point correction. With the target circuit in a power-off and no-load state, a plurality of no-load samples are continuously collected from the sensor at measuring point k, and the arithmetic mean is calculated as the static zero-point voltage of the corresponding sensor. Stored in non-volatile memory. During the formal operation phase, the real-time output voltage of the sensor at measurement point k is synchronously acquired. Combined voltage to magnetic field sensitivity Calculate the original magnetic field quantity. The calculation formula is as follows: The subtraction operation between the real-time output voltage and the static zero-point voltage eliminates the static bias of the sensor hardware and the ambient geomagnetic background, thus improving sensitivity. Dimensional conversion was completed. Sampling at all six measurement points was driven by the same trigger source to ensure phase consistency of data acquisition.

[0025] After obtaining the original magnetic field quantity, based on the symmetrical pairing relationship of the measuring points on the left and right sides of the same cross section, a transverse spatial subtraction operation is performed to generate the differential magnetic field quantity at the head of the slot. Differential magnetic field quantity in the middle of the slot and the differential magnetic field quantity at the tail end of the slot The calculation formula is as follows: After the main current-carrying current is shunted by the measuring tank, it generates magnetic field vectors in opposite directions on the left and right sides of the cross-section. The stray interference field inside the distribution box exhibits common-mode characteristics within the transverse geometric gap g. The difference operation effectively suppresses the background common-mode field and multiplies the differential-mode signal of the target circuit. The sensors on the left and right sides of the same cross-section are of the same model and have consistent parameters. All sensor sensitive axes are kept parallel and aligned with the direction of the main component of the magnetic field generated by the current shunting in the measuring tank.

[0026] Step 2: Read the differential magnetic field quantity at the beginning of the target loop slot output in Step 1. Differential magnetic field quantity in the middle of the slot and the differential magnetic field quantity at the tail end of the slot The three differential magnetic field quantities of the target loop are input into the arithmetic unit to perform longitudinal parity decomposition operation and extract the longitudinal even component. With vertical odd components Among them, the longitudinal even component represents the main differential magnetic field components that are distributed in the same direction along the longitudinal direction of the measuring slot, and is used to reflect the main current flow state of the loop; the longitudinal odd component represents the characteristic quantity of the degree of magnetic field offset between the beginning and end of the measuring slot, and is used to reflect the degree of end structure distortion.

[0027] Differential magnetic field quantity at the beginning of the slot Differential magnetic field quantity in the middle of the slot Differential magnetic field quantity at the tail end of the slot Perform weighted summation and normalization operations to calculate the longitudinal even components. Introducing a generalized spatial weighting coefficient that favors the central region. , and The weighting logic satisfies the following conditions: corresponding to the cross-sections at the beginning, middle, and end of the channel, respectively. and This ensures that the differential signal in the middle of the slot dominates the computation. In a specific embodiment, the generalized spatial weighting coefficients are configured to... and The symmetrical proportional relationship is used, that is, a weighted average is calculated using a weighted combination of 1:2:1 and a coefficient value of 4. The calculation formula is as follows: The weighted summation and normalization operations are mathematically equivalent to discretizing the spatial convolution of the longitudinal magnetic field distribution in the measuring tank. The middle section of the tank is close to the main flow region and experiences minimal leakage magnetic disturbance from the edges. The weighted operation, through the differential magnetic field quantity in the middle of the tank, is used to determine the differential magnetic field quantity... By assigning the highest weight, asymmetric electromagnetic disturbances at the measurement point are suppressed in both the time and spatial domains, and the principal differential magnetic field components that maintain a unidirectional distribution along the longitudinal direction of the measurement slot are extracted. The longitudinal even component is also extracted. It effectively isolates interference signals caused by end structure asymmetry, providing a stable numerical characterization for the subsequent reconstruction loop current.

[0028] Differential magnetic field quantity at the beginning of the slot Differential magnetic field quantity at the tail end of the slot Perform out-of-direction interpolation to calculate the longitudinal odd components. In one specific embodiment, the difference between the differential magnetic field quantities at the beginning and end of the slot is divided by a coefficient of 2 for amplitude normalization. The calculation formula is as follows: In an ideally symmetrical measuring tank structure, the differential magnetic field quantities at the beginning and end of the tank are driven by the same main current, and their magnitudes and vector directions are generally consistent, with the difference between the beginning and end approaching zero. When the measuring tank has longitudinal asymmetric current-guiding distortion (such as microscopic abrupt structures like residual conductive bridges on one side of the tank edge), the local change in current-guiding topology forces the current to undergo asymmetric path deflection at the inflow and outflow ends, resulting in a decrease in the differential magnetic field quantity at the beginning of the tank. Differential magnetic field quantity at the tail end of the slot This results in localized magnetic field biases in opposite directions. The opposite-direction difference operation uses subtraction logic to cancel out the same-direction main current components contained at both ends, causing the opposite bias difference to be superimposed and multiplied. The longitudinal odd component is extracted. The invisible end-structure asymmetric anomalies are transformed into quantifiable numerical features with high signal-to-noise ratio, thereby achieving feature decoupling between structural anomaly signals and main carrier signals at the data layer.

[0029] In a preferred embodiment, for loop 1, , as well as Substituting the data of 1.20 milliseconds into the longitudinal parity decomposition formula, the longitudinal even component of loop 1 is calculated. It is 1.20 millitrile, with longitudinal odd components. The value of 0 millitrile indicates that the measuring slot structure of loop 1 has ideal symmetry. For the abnormal loop 2, [the value will be...]. milliseconds Hammer and Substituting Hamt into the formula, the longitudinal even component of loop 2 is calculated. It is 1.50 millitrile, with a longitudinal odd component. It is -0.10 millitrile.

[0030] After completing the vertical parity decomposition operation, the vertical even components will be... With vertical odd components Each data point is written to a separate storage unit for independent use in subsequent real current reconstruction and composite anomaly attribution steps.

[0031] Step 3: Read the vertical even component cached within the storage unit. Retrieve the preset current conversion factor in the non-volatile memory. , the longitudinal even component Current conversion factor The arithmetic unit is loaded synchronously to perform scale-mapped multiplication operations and calculate the corrected actual measured current. The calculation formula is as follows: In the formula, the corrected actual measured current This refers to the characteristic current value used to accurately reflect the current-carrying capacity of the circuit after stripping away the distortion interference from the asymmetrical structure at the end. Current conversion factor. The electromagnetic coupling transfer function between the target loop rectangular busbar and the sensor array is encapsulated. The scaling mapping multiplication operation characterizes the longitudinal even component of the longitudinally oriented main magnetic field of the measurement slot. Linear mapping to current dimensions. Calculation of the corrected actual measured current. The differential interference signal caused by the asymmetric structural distortion at the end was removed at the algebraic level, and the high-fidelity loop current carrying capacity independent of microstructural defects was reconstructed.

[0032] Extract the corrected actual measured current of all N circuits in the distribution box within the same time slice. The corrected actual measured current of all N loops The arithmetic mean operation is performed by loading the arithmetic unit to calculate the average current of the multi-loop circuit. The calculation formula is as follows: Arithmetic averaging aggregates the true current flow status of all loops within the current monitoring system at a macroscopic level, smoothing out transient load fluctuations in a single loop. It calculates the average current across multiple loops. A dynamic comparison benchmark for the global load distribution of the distribution box under the current operating conditions was established, providing an adaptive reference base for measuring local imbalances.

[0033] Read the system's pre-configured minimum positive constant c, and combine it with the target loop correction to accurately measure the current. With multi-loop average current Perform normalized deviation calculation to calculate the multi-loop current imbalance. The multi-loop current imbalance refers to the normalized deviation of the actual measured current of the target loop from the average current level of all monitored loops after correction. The minimum positive constant c maps to the underlying noise baseline of the hardware system, and the value of the minimum positive constant c is preferably configured to be between 0.1% and 5% of the system's preset minimum non-zero calibration current. In a specific embodiment, the minimum positive constant c is set to 1% of the system's preset minimum non-zero calibration current. The calculation formula is as follows: Normalized deviation calculation to correct for the actual measured current Deviation from multi-loop average current The absolute difference measures the degree of local load imbalance and is measured by averaging the current across multiple loops. The base is normalized by incorporating a minimal positive constant c. This minimal positive constant c, as a static bias term, participates in the denominator addition operation, smoothing out quantization truncation errors at the numerical level of the algorithm and preventing calculation breakdowns during division by zero under conditions of full-scale unloaded circuitry or low noise levels. Multi-loop current imbalance is then calculated. A relative balance quantification index was generated after filtering out structural pseudo-differential interference, providing a highly robust unbiased data parameter for subsequent location of real load anomalies.

[0034] In a preferred embodiment, the system's preset current conversion factor is read. All are 100 amps per milliwatt. The corrected actual measured currents for loops 1, 2, and 3 are calculated. , , The values ​​are 120 amps, 150 amps, and 120 amps respectively. The average current across the multiple circuits is calculated. The ampere is 130 amps. The minimum positive constant c is read as 1.3 amps. Normalized deviation calculation is performed to calculate the imbalance of loop 1. The unbalance of loop 2 is approximately 0.076. It is approximately 0.152.

[0035] After completing the normalized deviation calculation, the multi-loop current imbalance will be... Write it to the storage unit for independent use in subsequent exception attribution steps.

[0036] Step 4: Read the multi-loop current imbalance in the cache within the storage unit. With vertical odd components Perform multidimensional cross-attribution of anomalies and output the final state discrimination instruction.

[0037] Extracting multi-loop current imbalance The maximum value of is used to pinpoint the candidate loop number p for the actual load imbalance. The calculation formula is as follows: Maximum value search operation will improve the multi-loop current imbalance. As a sorted sequence, extract the loop index with the largest value. If there are multiple loops, the multi-loop current imbalance... The values ​​are exactly the same; select the corrected actual measured current. The circuit number with the largest value is designated as number p; if the values ​​are still equal, the circuit with the smallest independent number i is selected as number p. The maximum value search operation identifies the circuit node that deviates most severely from the global load distribution of the distribution box under the current operating conditions. Establishing the number p of the candidate circuit for the actual load imbalance isolates the actual load anomaly from the dynamic background of multiple circuits, providing a reliable data anchor for current anomaly determination.

[0038] Extracting vertical odd components The maximum absolute value determines the index q of the candidate loop for the structural pseudo-anomaly. The calculation formula is as follows: The asymmetric structural distortion at the end, located in the longitudinal position difference of the measuring groove, will cause longitudinal odd components. The polarity is reversed, and the absolute value operation algebraically shields the influence of polarity on the distortion intensity. The maximum value retrieval operation extracts the absolute value of the vertical odd components. The index of the largest loop. If multiple loops exist, the absolute value of their vertical odd components... Completely equal, select the differential magnetic field quantity at the beginning of the slot. Absolute value and differential magnetic field quantity at the tail end of the slot The loop number with the largest sum of absolute values ​​is designated as loop number q; if the sums are still equal, the loop with the smallest independent loop number i is selected as loop number q. The maximum value retrieval operation captures the loop nodes with the strongest anisotropic bias at the beginning and end of the measurement slot. Establishing the loop number q as a candidate for structural pseudo-anomalies provides a definite criterion for identifying end-asymmetric structural distortions.

[0039] Read the system's pre-configured imbalance alarm threshold With structural distortion alarm threshold Compare the multi-loop current imbalance corresponding to the number p of the candidate loop with the actual load imbalance. and imbalance alarm threshold And compare the absolute values ​​of the longitudinal odd components corresponding to the number q of the candidate circuit for structural pseudo-anomalies. With structural distortion alarm threshold .

[0040] If the multi-loop current imbalance Less than the imbalance alarm threshold And the absolute value of the vertical odd component Less than the structural distortion alarm threshold The system determines that the distribution box is in a globally healthy and balanced state and outputs a normal operation command. Imbalance alarm threshold. With structural distortion alarm threshold It isolates the sensor's underlying thermal noise and disordered extreme values ​​caused by weak load fluctuations, cutting off the false alarm path under healthy operating conditions.

[0041] If the multi-loop current imbalance Greater than or equal to the imbalance alarm threshold , or the absolute value of the vertical odd component Greater than or equal to the structural distortion alarm threshold Anomaly detection instructions are generated based on the cross-mapping relationship between numbers p and q: If the number p of the candidate loop for actual load imbalance is equal to the number q of the candidate loop for structural pseudo-anomaly, perform a threshold double check operation: if and only if the multi-loop current imbalance... Greater than or equal to the imbalance alarm threshold And the absolute value of the vertical odd component Greater than or equal to the structural distortion alarm threshold When the circuit corresponding to number p is determined to be a composite abnormal circuit, a judgment command is generated indicating that the circuit corresponding to number p exhibits a true load imbalance with the superimposed effect of end asymmetric structural distortion; if the multi-circuit current imbalance... absolute value of longitudinal odd components If only one parameter is greater than or equal to the corresponding alarm threshold, the circuit corresponding to number p will be downgraded to a single abnormal circuit that exceeds the corresponding alarm threshold.

[0042] In a preferred embodiment, the system reads a pre-configured imbalance alarm threshold. The structural distortion alarm threshold is 0.10. The imbalance is 0.05 millitrile. A search revealed that the imbalance of loop 2... The value is 0.152, which is the maximum value across the entire field and exceeds the threshold. Therefore, the candidate loop number for the actual load imbalance is locked at p=2. Simultaneously, the absolute value of the longitudinal odd component of loop 2... The value is 0.10 millitrile, which is the maximum value across the entire field and exceeds the threshold. Therefore, the candidate loop number for the pseudo-anomaly of the locked structure is q=2. Since p=q=2, and and All exceeded the corresponding alarm thresholds, and the system finally output the judgment instruction: Loop 2 is a composite abnormal loop with real load imbalance and superimposed end asymmetric structural distortion.

[0043] If the candidate loop number p of the actual load imbalance is not equal to the candidate loop number q of the structural pseudo-anomaly, perform an independent attribution determination operation. (This applies to multi-loop current imbalance.) Greater than or equal to the imbalance alarm threshold Under the condition that the circuit corresponding to number p is determined to be a true unbalanced load circuit, the interference of structural distortion on the global imbalance of the circuit corresponding to number p is eliminated; in the absolute value of the longitudinal odd component Greater than or equal to the structural distortion alarm threshold Under the given conditions, the circuit corresponding to number q is determined to be a pseudo-abnormal circuit with an asymmetric end structure distortion, and it is confirmed that the circuit corresponding to number q exhibits the strongest end structure distortion and does not dominate the overall load imbalance of the distribution box.

[0044] The cross-mapping comparison and alarm threshold interception mechanism overcomes the algorithmic flaw of single-dimensional static thresholds, which are prone to causing false alarms. The generated final-state discrimination instruction achieves precise isolation and classification triggering of real circuit imbalance signals and geometric pseudo-differential signals at the data level. This cross-attribution operation writes the final-state discrimination instruction into the storage unit, completing the anomaly attribution.

[0045] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for calculating and identifying the current imbalance of multiple circuits in a distribution box, characterized in that, include: For each target circuit in the entire monitoring circuit in the distribution box, the first section, the middle section and the last section of the target circuit busbar measurement slot are defined. Measurement points are set up on both sides of each section to obtain the original magnetic field quantity. The original magnetic field quantity is subjected to a lateral spatial difference operation to generate the differential magnetic field quantity at the first, middle and last sections of the slot. A generalized spatial weighting coefficient with central dominance is introduced to perform weighted summation and normalization operations on the differential magnetic field quantities at the beginning, middle, and end of the slot to extract the longitudinal even components. An out-of-direction difference operation is performed on the differential magnetic field quantities at the beginning and end of the slot to extract the longitudinal odd components. Multiply the longitudinal dipole component by the preset current conversion factor to calculate the corrected actual measured current. Within the same time slice, summarize the corrected actual measured current of all monitoring loops to calculate the multi-loop average current. Place the minimum positive constant in the denominator and perform normalization deviation calculation by combining the corrected actual measured current and the multi-loop average current to calculate the multi-loop current imbalance. The maximum value of the multi-loop current imbalance is retrieved to lock the candidate loops of the actual load imbalance, and the maximum value of the absolute value of the longitudinal odd component is retrieved to lock the candidate loops of the structural pseudo-anomaly. Based on the imbalance alarm threshold and the structural distortion alarm threshold, cross-attribution operation is performed on the candidate loops of the actual load imbalance and the candidate loops of the structural pseudo-anomaly to generate the final state discrimination instruction.

2. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 1, characterized in that, Define the cross-sections at the beginning, middle, and end of the target circuit busbar within the measurement slot, including: A three-dimensional local coordinate system is established with the center of the current inflow end edge of the target circuit busbar measuring channel as the origin, the positive current carrying direction of the busbar is defined, and the longitudinal length of the measuring channel is read. Along the positive current-carrying direction of the busbar, the section at the beginning of the measuring channel is set at one-sixth of the longitudinal length of the measuring channel, the section at the middle of the measuring channel is set at one-half of the longitudinal length of the measuring channel, and the section at the end of the measuring channel is set at five-sixths of the longitudinal length of the measuring channel.

3. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 2, characterized in that, To obtain the raw magnetic field quantities, including: When the target circuit is in a power-off and no-load state, multiple no-load samples are continuously collected by sensors configured at the measuring points and the arithmetic average is performed to generate a static zero-point voltage, which is then stored in a non-volatile memory. During the formal operation phase, the real-time output voltage of the sensor at the measuring point is acquired synchronously, the difference between the real-time output voltage and the static zero-point voltage is calculated, and the difference is divided by the voltage of the sensor at the measuring point to the magnetic field sensitivity to generate the original magnetic field quantity.

4. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 1, characterized in that, Weighted summation and normalization operations are performed on the differential magnetic field quantities at the beginning, middle, and end of the slot to extract the longitudinal even components. An out-of-direction difference operation is performed on the differential magnetic field quantities at the beginning and end of the slot to extract the longitudinal odd components. This includes: According to the weight combination of 1:2:1, the generalized spatial weight coefficients are configured for the differential magnetic field quantities at the beginning, middle and end of the slot, respectively. The differential magnetic field quantities at the beginning, middle and end of the slot are weighted and summed. The weighted summation result is divided by the coefficient value of four to extract the longitudinal even component. Calculate the out-of-direction difference between the differential magnetic field quantity at the beginning of the slot and the differential magnetic field quantity at the end of the slot. Divide the out-of-direction difference by the coefficient value to perform amplitude normalization and extract the longitudinal odd component.

5. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 1, characterized in that, Obtain the preset current conversion factor, including: When the target circuit is in the factory stage or offline state, inject two known non-zero standard test currents into the target circuit in sequence, and obtain two longitudinal even component outputs corresponding to the two known non-zero standard test currents. Calculate the ratio of the difference between two known non-zero standard test currents to the difference between the two longitudinal even component outputs, calibrate based on the ratio to generate a preset current conversion factor, and write the preset current conversion factor into a non-volatile memory.

6. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 1, characterized in that, By placing the minimum positive constant in the denominator and performing a normalized deviation calculation on the corrected actual measured current and the multi-loop average current, the multi-loop current imbalance is calculated, including: Minimum positive constant based on the minimum non-zero calibration current configuration of the hardware system; The multi-loop current imbalance is calculated by performing a division operation with the absolute difference between the corrected actual measured current and the average current of the multi-loop as the numerator and the sum of the average current of the multi-loop and the minimum positive constant as the denominator.

7. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 1, characterized in that, The search for maxima of multi-loop current imbalance to pinpoint candidate loops with actual load imbalance, and the search for maxima of the absolute values ​​of longitudinal odd components to pinpoint candidate loops with structural pseudo-anomalies, include: If multiple loops have the same current imbalance value, select the loop with the largest corrected actual measured current value as the candidate loop for actual load imbalance; if the corrected actual measured current values ​​are still equal, select the loop with the smallest independent loop number as the candidate loop for actual load imbalance. If multiple loops have the same absolute value of longitudinal odd component, the loop with the largest sum of the absolute values ​​of the differential magnetic field at the beginning and end of the slot is selected as the candidate loop for structural pseudo-anomaly; if the sum of the absolute values ​​of the differential magnetic field at the beginning and end of the slot is still equal, the loop with the smallest independent loop number is selected as the candidate loop for structural pseudo-anomaly.

8. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 1, characterized in that, Based on the imbalance alarm threshold and the structural distortion alarm threshold, a cross-attribution operation is performed on the candidate loops of actual load imbalance and the candidate loops of structural pseudo-anomalies to generate a final state discrimination instruction, including: If the multi-circuit current imbalance degree corresponding to the candidate circuit of the actual load imbalance is less than the imbalance degree alarm threshold, and the absolute value of the longitudinal odd component corresponding to the candidate circuit of the structural pseudo-anomaly is less than the structural distortion alarm threshold, the distribution box is determined to be in a global healthy balance state, and a normal operation command is output.

9. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 8, characterized in that, Multidimensional cross-anomaly attribution computation includes: If the actual load imbalance candidate circuit and the structural pseudo-anomaly candidate circuit are the same circuit, and the multi-circuit current imbalance degree corresponding to the actual load imbalance candidate circuit is greater than or equal to the imbalance degree alarm threshold, and the absolute value of the longitudinal odd component corresponding to the structural pseudo-anomaly candidate circuit is greater than or equal to the structural distortion alarm threshold, the actual load imbalance candidate circuit is determined to be a composite anomaly circuit. If the actual load imbalance candidate circuit and the structural pseudo-anomaly candidate circuit are the same circuit, and only one parameter in the absolute value of the multi-circuit current imbalance degree corresponding to the actual load imbalance candidate circuit and the vertical odd component corresponding to the structural pseudo-anomaly candidate circuit is greater than or equal to the corresponding alarm threshold, the actual load imbalance candidate circuit will be downgraded to a single abnormal circuit that exceeds the corresponding alarm threshold.

10. The method for calculating and identifying the current imbalance of multiple circuits in a distribution box according to claim 8, characterized in that, Multidimensional cross-anomaly attribution computation also includes: If the candidate loop of actual load imbalance and the candidate loop of structural pseudo-anoma are not the same loop, perform independent attribution determination operation. If the current imbalance of the multi-loop current corresponding to the candidate loop of the actual load imbalance is greater than or equal to the imbalance alarm threshold, the candidate loop of the actual load imbalance is determined to be the actual load imbalance loop. If the absolute value of the longitudinal odd component corresponding to the candidate loop of structural pseudo-anomaly is greater than or equal to the structural distortion alarm threshold, the candidate loop of structural pseudo-anomaly is determined to be a pseudo-anomaly loop of end asymmetric structural distortion.

Images