A circuit fault diagnosis method and system for an electric energy metering box

Abstract

This invention relates to the field of electrical parameter measurement and circuit fault diagnosis, and discloses a method and system for diagnosing circuit faults in an energy metering box. The method includes: synchronously acquiring high-frequency voltage and current sampling sequences; locating the fundamental zero-crossing point based on the voltage sequence and performing phase calibration on the current sequence; extracting sampling segments from the neighborhood of the zero-crossing point and dividing them into rising and falling edge sets according to the slope polarity; calculating the admittance sequence of sampling points in each set; identifying and removing sampling points in the noise dead zone; extracting the cumulative integral of the admittance difference at symmetrical positions to generate an asymmetric deviation; and determining that a circuit has a deterioration fault when the deviation exceeds a threshold. This invention utilizes the nonlinear barrier characteristics of connection points to achieve physical-level decoupling of external interference, improves the sensitivity of sensing hidden physical defects, ensures that faults are detected before they evolve into fires, and enhances the proactive defense capabilities of operation and maintenance.

Description

Technical Field

[0001] This invention relates to a method and system for diagnosing circuit faults in an electricity metering box, belonging to the field of electrical parameter measurement and circuit fault diagnosis technology. Background Technology

[0002] Currently, low-voltage power metering boxes, as end nodes of the distribution network, are responsible for distributing and metering electrical energy. The mainstream methods for monitoring the health status of circuits include monitoring the root mean square value of current, voltage deviation, and deploying temperature sensors. These methods have good universality in identifying circuit overload, short circuit, or obvious overheating faults. However, the connection terminals in the circuit undergo physical degradation under the alternating effects of environmental stress and thermal stress. The oxide film or micro gaps generated at the terminal interface exhibit complex nonlinear barrier characteristics. Affected by the high current breakdown effect in the peak range of power frequency voltage, such defects exhibit low resistance characteristics when the voltage amplitude is high, resulting in electrical parameter characteristics in the defective state being highly similar to those in the normal state. Traditional measurement methods, due to the use of full-cycle energy averaging calculations, cause transient distortions in the voltage zero-crossing neighborhood to be covered by the main power frequency energy, leaving the monitoring system in a signal silent state during the physical connection degradation evolution period.

[0003] To improve fault identification sensitivity, increasing the sampling frequency or introducing harmonic analysis are common improvement attempts. However, the current distortion generated by the nonlinear load on the user side and the electrical characteristics excited by the physical degradation of the contact surface overlap in the frequency domain. It is difficult to achieve physical separation between external load interference and internal connection defects by simply relying on frequency domain filtering. In addition to the limitations of hardware-level monitoring, the control logic faces challenges in stripping load interference. For example, Chinese invention patent application CN114509647A discloses a fault arc detection method and system. It obtains the coupling coefficient between the power frequency signal and the high-frequency component through load training and sets a detection threshold to identify the distortion signal. This statistical training algorithm relies on the completeness of the preset sample library. In the actual operation of the metering box, the physical degradation of the contact interface has randomness and dynamic evolution characteristics. The preset coupling characteristic curve is difficult to reproduce the intrinsic physical behavior of a specific oxide film at the moment of polarity reversal. Detached from the physical action mechanism and simply relying on the feature vector mapping statistical model, it is easy to miss the alarm or have insufficient resolution for identifying weak physical degradation signals when facing environmental noise fluctuations and nonlinear load coupling due to model mismatch.

[0004] Therefore, the technical problem to be solved by this invention is how to extract electrical parameter evolution measures that reflect the nonlinear potential barrier characteristics of the contact surface inside the metering box by utilizing the original sampling sequence of universal metering hardware and removing the interference of complex nonlinear loads on the user side. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A method for diagnosing circuit faults in an electricity metering box, comprising the following steps: Step 101: Simultaneously acquire the high-frequency voltage sampling sequence and high-frequency current sampling sequence of the load side of the power metering box, wherein the sampling frequency is not less than 100kHz. Step 102: Extract the fundamental zero-crossing point of the high-frequency voltage sampling sequence, calculate the phase deviation between the high-frequency voltage sampling sequence and the high-frequency current sampling sequence, and perform time-shift calibration on the high-frequency current sampling sequence based on the phase deviation to obtain the calibrated current sequence. Step 103: Using the fundamental zero-crossing point as the center, extract a local voltage sampling segment with a span of 500μs and a synchronous local current sampling segment. Step 104: Based on the slope polarity of the local voltage sampling segment, divide the intercepted sampling points into a set of rising edge zero crossings and a set of falling edge zero crossings. Step 105: Calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of rising edges to obtain the first admittance sequence; simultaneously calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of falling edges to obtain the second admittance sequence. Step 106: Identify noise dead zone sampling points in the local voltage sampling segment whose absolute values ​​are between 1.5V and 3.0V, and remove the corresponding sampling points in the first admittance sequence and the second admittance sequence; Step 107: Extract the admittance difference at the symmetrical phase angle position in the first admittance sequence and the second admittance sequence, calculate the cumulative integral of the admittance difference within the local voltage sampling segment span, and generate the asymmetric deviation. Step 108: When the asymmetric deviation exceeds the preset symmetric reference threshold, it is determined that there is a contact resistance deterioration fault in the power distribution circuit inside the power metering box, and an early warning signal is output.

[0006] Preferably, step 102 is further refined into the following sub-steps: Step 1021, identify the low-voltage waveform interval in the high-frequency voltage sampling sequence where the sampled value is lower than the preset voltage threshold; Step 1022, shield the sampling point data in the low-voltage waveform interval and determine the zero offset of the voltage transformer; Step 1023, use the waveform stable segment of the high-frequency voltage sampling sequence to synchronously align the calibration current sequence, and use a phase compensation algorithm based on the fundamental zero-crossing time difference to make the calibrated sampling data have a reference with physical symmetry.

[0007] Preferably, step 105 is further refined into the following sub-steps: Step 1051, calculate the transient admittance value corresponding to each sampling time of the local voltage sampling segment; Step 1052, map the transient admittance value to the corresponding transient voltage change rate to generate a dynamic admittance evolution curve centered on the fundamental zero-crossing point, wherein the dynamic admittance evolution curve is used to characterize the energy loss state of the metal connection point during the polarity reversal process.

[0008] Preferably, step 107 is further refined into the following sub-steps: Step 1071, obtain the first admittance feature component in the positive zero-crossing direction and the second admittance feature component in the negative zero-crossing direction; Step 1072, calculate the numerical deviation between the first admittance feature component and the second admittance feature component, and quantify the degree of dynamic admittance asymmetry excited in the positive and negative zero-crossing directions; Step 1073, perform an integral operation on the numerical deviation over the time scale to obtain the asymmetric deviation.

[0009] Preferably, after step 108, the following steps are also included: Step 109, retrieve the historical fault diagnosis records of the power metering box; Step 110, extract the evolution slope of the asymmetric deviation as the operating cycle changes; Step 111, when the evolution slope shows a monotonically increasing trend and the rate of increase exceeds the preset rate of change threshold, determine that the internal connection point of the power metering box is in the deterioration evolution period, and output a diagnosis signal.

[0010] Preferably, step 101 is further refined into the following sub-steps: step 1011, using a wideband current transformer to collect analog voltage signals and analog current signals from the load side of the power metering box; step 1012, using an analog-to-digital converter circuit to convert the analog voltage signals and analog current signals into high-frequency voltage sampling sequences and high-frequency current sampling sequences.

[0011] Preferably, step 107 further includes the following steps: step 1074, mapping the sampled data after phase deviation calibration to a normalized unit coordinate system; step 1075, calculating the center symmetry deviation value of the dynamic admittance evolution curve relative to the origin in the normalized unit coordinate system.

[0012] Preferably, step 108, determining that there is a contact resistance degradation fault in the power distribution circuit inside the power metering box, includes: step 1081, matching the asymmetric deviation with the fault feature fingerprints corresponding to different degrees of degradation in the preset database; step 1082, identifying the specific fault type in the power distribution circuit of the power metering box caused by loose terminals or oxidation of contacts.

[0013] Preferably, step 108 further includes the following steps: step 1083, determining the fault risk level based on the value of the asymmetric deviation; step 1084, outputting corresponding early warning instructions based on the fault risk level, and uploading the fault diagnosis results to the operation and maintenance management platform.

[0014] A circuit fault diagnosis system for an electricity metering box, comprising: The data acquisition module is used to synchronously acquire the high-frequency voltage sampling sequence and high-frequency current sampling sequence of the load side of the power metering box, wherein the sampling frequency is not less than 100kHz. The phase calibration module is used to extract the fundamental zero-crossing point of the high-frequency voltage sampling sequence, calculate the phase deviation between the high-frequency voltage sampling sequence and the high-frequency current sampling sequence, and perform time-shift calibration on the high-frequency current sampling sequence based on the phase deviation to obtain the calibration current sequence. The sample extraction module is used to extract a local voltage sampling segment with a span of 500μs, centered on the zero-crossing point of the fundamental frequency, as well as a synchronous local current sampling segment. The polarity classification module is used to divide the intercepted sampling points into rising edge zero-crossing set and falling edge zero-crossing set according to the slope polarity of the local voltage sampling segment. The admittance calculation module is used to calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of rising edges to obtain the first admittance sequence; and simultaneously calculates the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of falling edges to obtain the second admittance sequence. The noise filtering module is used to identify noise dead zone sampling points in the local voltage sampling segment with an absolute value between 1.5V and 3.0V, and to remove the corresponding sampling points in the first admittance sequence and the second admittance sequence. The asymmetric analysis module is used to extract the admittance difference at the symmetrical phase angle position in the first admittance sequence and the second admittance sequence, calculate the cumulative integral of the admittance difference within the local voltage sampling segment span, and generate an asymmetric deviation quantity that reflects the nonlinear barrier characteristics of the connection point. The fault determination module is used to determine that there is a contact resistance deterioration fault in the power distribution circuit inside the power metering box when the asymmetric deviation exceeds the preset symmetric reference threshold, and outputs a warning signal.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the fault diagnosis of the power metering box circuit, by locking the low-energy sampling window in the neighborhood of the zero-crossing voltage, the nonlinear barrier characteristics exhibited by the oxide film or micro gap at the contact connection are transformed from a macroscopic main energy shielding state into an observable discrete conductivity distortion. Traditional measurement methods, which rely on full-cycle energy averaging or steady-state impedance judgment, often only produce a response after the contact resistance has generated a significant thermal effect. This scheme utilizes the charge charging and discharging effect of the oxide film in the low-voltage range and the dielectric strength recovery process to extract microsecond-level transient signals reflecting the intrinsic deterioration state of the contacts at an earlier stage of physical fire risk. This measurement mechanism moves the observation dimension from macroscopic thermal accumulation to the microscopic physical barrier evolution level, ensuring that fault characteristics are accurately captured before they evolve into destructive physical phenomena.

[0016] 2. Establish a load characteristic decoupling mechanism and reduce false alarm frequency. This method utilizes the dynamic conductance asymmetry measure of the zero-crossing neighborhood of the voltage rising and falling edges to construct a targeted identification logic for defects inside the metering box. There are a large number of nonlinear switching loads at the end of the power system, and the harmonic signals they generate have a high degree of central symmetry in the time domain topology. This scheme calculates the divergence deviation of the positive and negative zero-crossing feature sequences to reveal the polarity relaxation difference that only exists in the degradation process of the physical connection point of the metering box. This asymmetric mapping of the topological space, without relying on complex filtering hardware, physically separates the current distortion caused by external power loads from the power offset caused by internal intrinsic physical defects at the logic level. This processing method enables the system to maintain accurate identification of the internal connection status of the metering box even when facing complex power conditions.

[0017] 3. Compensating for sensor system errors and enhancing engineering applicability: This method eliminates measurement artifacts introduced by the inherent phase-frequency characteristic differences between voltage and current transformers in the high-frequency response range by embedding a phase background alignment link based on the fundamental zero-crossing difference before admittance calculation. In actual distribution network operation environments, the non-ideal frequency response of the transformers can cause random phase shifts in transient admittance calculations. This scheme uses steady-state waveform characteristics for adaptive calibration, combined with the elimination of sampling points in the low-voltage dead zone range, to shield the interference caused by the white noise of the power grid background and the measurement blind zone of the transformer on the extraction of weak conductivity features. This software-level compensation mechanism for hardware non-ideal characteristics improves the algorithm's compatibility with existing universal metering hardware and ensures the stability of high-precision fault monitoring capabilities under different operating conditions. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method for identifying nonlinear degradation of circuit connection points in an energy metering box according to the present invention. Figure 2 This is a logic diagram showing the switching logic of each functional module in the diagnostic system of the present invention.

[0019] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0021] A method for diagnosing circuit faults in an electricity metering box includes the following steps: Step 101: Simultaneously acquire the high-frequency voltage sampling sequence and high-frequency current sampling sequence of the load side of the power metering box, wherein the sampling frequency is not less than 100kHz. Step 102: Extract the fundamental zero-crossing point of the high-frequency voltage sampling sequence, calculate the phase deviation between the high-frequency voltage sampling sequence and the high-frequency current sampling sequence, and perform time-shift calibration on the high-frequency current sampling sequence based on the phase deviation to obtain the calibrated current sequence. Step 103: Using the fundamental zero-crossing point as the center, extract a local voltage sampling segment with a span of 500μs and a synchronous local current sampling segment. Step 104: Based on the slope polarity of the local voltage sampling segment, divide the intercepted sampling points into a set of rising edge zero crossings and a set of falling edge zero crossings. Step 105: Calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of rising edges to obtain the first admittance sequence; simultaneously calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of falling edges to obtain the second admittance sequence. Step 106: Identify noise dead zone sampling points in the local voltage sampling segment whose absolute values ​​are between 1.5V and 3.0V, and remove the corresponding sampling points in the first admittance sequence and the second admittance sequence; Step 107: Extract the admittance difference at the symmetrical phase angle position in the first admittance sequence and the second admittance sequence, calculate the cumulative integral of the admittance difference within the local voltage sampling segment span, and generate the asymmetric deviation. Step 108: When the asymmetric deviation exceeds the preset symmetric reference threshold, it is determined that there is a contact resistance deterioration fault in the power distribution circuit inside the power metering box, and an early warning signal is output.

[0022] Preferably, step 102 is further refined into the following sub-steps: Step 1021, identify the low-voltage waveform interval in the high-frequency voltage sampling sequence where the sampled value is lower than the preset voltage threshold; Step 1022, shield the sampling point data in the low-voltage waveform interval and determine the zero offset of the voltage transformer; Step 1023, use the waveform stable segment of the high-frequency voltage sampling sequence to synchronously align the calibration current sequence, and use a phase compensation algorithm based on the fundamental zero-crossing time difference to make the calibrated sampling data have a reference with physical symmetry.

[0023] Preferably, step 105 is further refined into the following sub-steps: Step 1051, calculate the transient admittance value corresponding to each sampling time of the local voltage sampling segment; Step 1052, map the transient admittance value to the corresponding transient voltage change rate to generate a dynamic admittance evolution curve centered on the fundamental zero-crossing point, wherein the dynamic admittance evolution curve is used to characterize the energy loss state of the metal connection point during the polarity reversal process.

[0024] Preferably, step 107 is further refined into the following sub-steps: Step 1071, obtain the first admittance feature component in the positive zero-crossing direction and the second admittance feature component in the negative zero-crossing direction; Step 1072, calculate the numerical deviation between the first admittance feature component and the second admittance feature component, and quantify the degree of dynamic admittance asymmetry excited in the positive and negative zero-crossing directions; Step 1073, perform an integral operation on the numerical deviation over the time scale to obtain the asymmetric deviation.

[0025] Preferably, after step 108, the following steps are also included: Step 109, retrieve the historical fault diagnosis records of the power metering box; Step 110, extract the evolution slope of the asymmetric deviation as the operating cycle changes; Step 111, when the evolution slope shows a monotonically increasing trend and the rate of increase exceeds the preset rate of change threshold, determine that the internal connection point of the power metering box is in the deterioration evolution period, and output a diagnosis signal.

[0026] Preferably, step 101 is further refined into the following sub-steps: step 1011, using a wideband current transformer to collect analog voltage signals and analog current signals from the load side of the power metering box; step 1012, using an analog-to-digital converter circuit to convert the analog voltage signals and analog current signals into high-frequency voltage sampling sequences and high-frequency current sampling sequences.

[0027] Preferably, step 107 further includes the following steps: step 1074, mapping the sampled data after phase deviation calibration to a normalized unit coordinate system; step 1075, calculating the center symmetry deviation value of the dynamic admittance evolution curve relative to the origin in the normalized unit coordinate system.

[0028] Preferably, step 108, determining that there is a contact resistance degradation fault in the power distribution circuit inside the power metering box, includes: step 1081, matching the asymmetric deviation with the fault feature fingerprints corresponding to different degrees of degradation in the preset database; step 1082, identifying the specific fault type in the power distribution circuit of the power metering box caused by loose terminals or oxidation of contacts.

[0029] Preferably, step 108 further includes the following steps: step 1083, determining the fault risk level based on the value of the asymmetric deviation; step 1084, outputting corresponding early warning instructions based on the fault risk level, and uploading the fault diagnosis results to the operation and maintenance management platform.

[0030] A circuit fault diagnosis system for an electricity metering box, comprising: The data acquisition module is used to synchronously acquire the high-frequency voltage sampling sequence and high-frequency current sampling sequence of the load side of the power metering box, wherein the sampling frequency is not less than 100kHz. The phase calibration module is used to extract the fundamental zero-crossing point of the high-frequency voltage sampling sequence, calculate the phase deviation between the high-frequency voltage sampling sequence and the high-frequency current sampling sequence, and perform time-shift calibration on the high-frequency current sampling sequence based on the phase deviation to obtain the calibration current sequence. The sample extraction module is used to extract a local voltage sampling segment with a span of 500μs, centered on the zero-crossing point of the fundamental frequency, as well as a synchronous local current sampling segment. The polarity classification module is used to divide the intercepted sampling points into rising edge zero-crossing set and falling edge zero-crossing set according to the slope polarity of the local voltage sampling segment. The admittance calculation module is used to calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of rising edges to obtain the first admittance sequence; and simultaneously calculates the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of falling edges to obtain the second admittance sequence. The noise filtering module is used to identify noise dead zone sampling points in the local voltage sampling segment with an absolute value between 1.5V and 3.0V, and to remove the corresponding sampling points in the first admittance sequence and the second admittance sequence. The asymmetric analysis module is used to extract the admittance difference at the symmetrical phase angle position in the first admittance sequence and the second admittance sequence, calculate the cumulative integral of the admittance difference within the local voltage sampling segment span, and generate an asymmetric deviation quantity that reflects the nonlinear barrier characteristics of the connection point. The fault determination module is used to determine that there is a contact resistance deterioration fault in the power distribution circuit inside the power metering box when the asymmetric deviation exceeds the preset symmetric reference threshold, and outputs a warning signal.

[0031] Example 1: In a low-voltage power metering box circuit operating in an industrial park, a nonlinear oxide film forms at the incoming terminal connection due to environmental thermal stress. This oxide film breaks down in the high-amplitude range of the power frequency voltage, exhibiting a low impedance state, resulting in transient distortion in the neighborhood of the voltage zero-crossing point. This invention employs a high-frequency sampling unit to simultaneously acquire high-frequency voltage and current sampling sequences from the load side, with the sampling frequency set to 100kHz. The processor extracts the fundamental zero-crossing point of the high-frequency voltage sampling sequence and calculates the phase deviation between the high-frequency voltage and current sampling sequences. Time-shift calibration is performed on the high-frequency current sampling sequence based on the phase deviation to generate a calibration current sequence. A local voltage sampling segment with a time span of 500 μs and a synchronous local current sampling segment are extracted, centered on the fundamental zero-crossing point. The processor divides the extracted sampling points into rising edge zero-crossing sets and falling edge zero-crossing sets based on the slope polarity of the local voltage sampling segments. The instantaneous ratio of the current amplitude to the voltage amplitude corresponding to each sampling point in the rising edge zero-crossing set and the falling edge zero-crossing set are calculated respectively, generating a first admittance sequence and a second admittance sequence. During this calculation process, the processor identifies and removes local current sampling segments. Sampling points with absolute voltage values ​​between 1.5V and 3.0V in the voltage sampling segment correspond to a preset noise dead zone. Specifically, the phase deviation between the high-frequency voltage and current calculated here essentially refers to the hardware delay constant fixed by the difference in the background of the high-frequency response characteristics of the voltage and current transformers under the initial calibration state without external effective electrical load. When performing time-shift calibration on subsequent high-frequency current sampling sequences, the system only performs inverse compensation operation on this background hardware delay constant, thereby ensuring that the physical phase deviation of the real reactive power generated by the external nonlinear load is completely correct. This avoids the distortion of underlying electrical parameters caused by forcibly aligning the load fundamental current and voltage zero-crossing points. In the power frequency AC system, the 500μs time span before and after the fundamental zero-crossing point physically covers the characteristic breakdown critical region of the metal-oxide-metal interface of the connection terminal. Due to the inherent Schottky barrier effect of the oxide film interface, when the transient voltage amplitude is lower than the above-mentioned dead zone boundary, the barrier exhibits a high-damping blocking state, and the admittance fluctuation is dominated by the system's background white noise. When the transient voltage crosses the dead zone boundary, the interface charge undergoes nonlinear tunneling and microscopic polarity relaxation, which excites dynamic admittance distortion in a specific direction.

[0032] The processor extracts the admittance difference at symmetrical phase angle positions in the first and second admittance sequences. By calculating the cumulative integral of the admittance difference within a local voltage sampling segment, it generates an asymmetric deviation characteristic of the connection point's nonlinearity. The processor then applies the formula... Solve for the asymmetric deviation S, where, This represents the transient admittance value at the i-th discrete sampling point in the first admittance sequence. This represents the transient admittance value at the corresponding centrosymmetric temporal position in the second admittance sequence, and N represents the total number of sampling points within the truncation segment. The time interval constant represents the interval between adjacent sampling points. Furthermore, the processor divides the instantaneous voltage value of each sampling point within the phase-calibrated voltage segment by the absolute peak value of the segment voltage, and divides the synchronously acquired instantaneous current value by the absolute peak value of the current. These values ​​are mapped to a normalized unit coordinate system to construct a dynamic admittance evolution curve. The processor calculates the cumulative Euclidean geometric distance between each discrete coordinate point on the curve and the ideal symmetrical point extending backward from the origin. It then extracts the central symmetry deviation value of the dynamic admittance evolution curve relative to the origin. When the asymmetry deviation exceeds a preset symmetry reference threshold, it determines that there is a nonlinear degradation fault in the contact resistance of the power distribution circuit inside the energy metering box and outputs a warning signal. At this time, the processor extracts the generated asymmetry deviation and the aforementioned central symmetry deviation value to construct a two-dimensional feature vector, which is then input into the fault identification module. To eliminate the randomness of data at a single point and meet the time evolution prerequisite requirements of the dynamic sequence comparison algorithm, the processor performs a pre-processing step before inputting the two-dimensional feature vector into the fault identification module for subsequent judgment. It extracts historical two-dimensional feature vector data from several consecutive power frequency cycles, including the current cycle, and extracts and arranges them sequentially according to time. It assembles the originally isolated static feature points into a two-dimensional variable-length data sequence with time-series evolution trajectory characteristics. The module calculates the dynamic time warping (DTW) spatial distance between the current two-dimensional feature vector and the feature vectors of fingerprint templates of each degradation level in the memory's preset database. If the calculated shortest warping path distance is less than the set classification tolerance limit, it locks and outputs the specific fault type caused by terminal loosening or contact oxidation that matches the corresponding fingerprint template. This method utilizes the dynamic conductance asymmetric evolution of the zero-crossing neighborhood of the voltage rising and falling edges to advance the observation point from the overall heat accumulation stage to the local physical barrier evolution stage. It utilizes the polar relaxation difference generated by the oxide film in the low voltage range to achieve physical-level decoupling between the charge offset caused by internal connection point defects and the central symmetric current distortion caused by external nonlinear loads.

[0033] Example 2: The test environment included an adjustable AC power supply and a load circuit containing a precision shunt. The load circuit was connected to terminal blocks with varying degrees of degradation. A multi-channel synchronous acquisition card was used to simultaneously acquire voltage and current sampling sequences. The card had a resolution of 16 bits and a maximum sampling rate of 200kHz. Gaussian white noise with a signal-to-noise ratio of 20dB and a 1% 50Hz third harmonic were superimposed on the original signal to simulate the electromagnetic interference environment of an industrial setting. The sampling frequency... The settings need to balance the accuracy of capturing transient distortions during oxide film breakdown with the computational load of the processor. Since the charge migration process during oxide film penetration lasts in the 50μs neighborhood, to ensure that at least 5 sampling points are obtained within this process, the parameters are adjusted according to the signal reconstruction requirements. The kHz setting was 100kHz. The experiment was divided into an experimental group and a control group. The control group used an impedance determination method based on the full-cycle effective value calculation. The healthy terminals were measured under interference conditions. The average value of the first admittance sequence corresponding to the zero-crossing set of the rising edge was 1.25S, and the average value of the second admittance sequence corresponding to the zero-crossing set of the falling edge was 1.24S. The calculated result of the generated asymmetric deviation was 0.08. The terminal sample was replaced with a degraded terminal with a 0.5μm thick oxide film. Due to the low impedance performance of the oxide film in the high voltage amplitude range, the effective voltage measured by the control group was 220.5V, the effective current was 17.6A, and the full-cycle equivalent impedance was 12.5284Ω. Compared with 12.5000Ω in the healthy state, there was only a 0.22% numerical deviation, which did not reach the 5% warning threshold.

[0034] The experimental group used the method of this invention to process the signal of the degraded terminal. After correcting the transformer phase deviation of 1.2ms through phase calibration, the fundamental zero-crossing point was accurately located and sampling points in the noise dead zone of 1.5V to 3.0V were eliminated. The measured first admittance sequence of the degraded sample showed a value of 0.85s during the rising phase due to charge accumulation, and the second admittance sequence showed a value of 1.42s during the falling phase due to polarity reversal hysteresis. This difference in polarity relaxation caused the admittance difference to jump at a symmetrical position, and the accumulated asymmetric deviation reached 8.52, which produced a numerical gain compared to the healthy state. To verify the rationality of the truncation span setting, an out-of-range control group was established and the truncation span was adjusted to 2ms. Because the sampling segment contained too much linear conductance data, nonlinear conductance data was also affected. The admittance difference caused by the voltage barrier is diluted during the integration process, the asymmetric bias is reduced to 1.15, and the signal-to-noise ratio is reduced by 86.5%. Therefore, 500 μs is determined as the working window for capturing local physical degradation signals. In addition, the test on the control group with missing features shows that if the admittance difference extraction step at the symmetrical phase angle position is omitted and only the single-sided zero-crossing waveform analysis is performed, the system cannot distinguish between the harmonic distortion of the order of 0.12 caused by the external nonlinear load and the signal of the order of 8.52 caused by oxide film defects. The above data confirm that the method of the present invention utilizes the dynamic admittance evolution characteristics of the voltage zero-crossing neighborhood to achieve decoupling between internal physical defects and external load interference through symmetrical cancellation, providing data support for the identification of the early stage of nonlinear degradation of contact resistance.

[0035] Example 3: This example combines Figures 1 to 2 This document describes a method and system for diagnosing circuit faults in an electricity metering box. Figure 1As shown, the logic flow from top to bottom executes each step sequentially. Step 101 executes the acquisition of high-frequency voltage and current sampling sequences on the load side of the metering box. Step 102 executes the extraction of the fundamental zero-crossing point, calculates the phase difference, and calibrates the current sequence with time shift. Step 103 executes the extraction of local voltage and current segments centered on the fundamental zero-crossing point. Step 104 executes the division of sampling points into rising and falling edge sets according to slope polarity. Step 105 executes the calculation of the current-voltage ratio of the two sets to obtain the first and second admittance sequences. Step 106 executes the identification and removal of noise dead zone sampling points from the two admittance sequences. Step 107 executes the integration of the symmetrical phase angle admittance difference to generate an asymmetrical deviation. Step 108 executes the determination of contact resistance degradation and outputs a warning if the deviation exceeds the threshold.

[0036] like Figure 2 As shown, each operational stage forms a closed-loop flow structure. The system is in a data synchronous acquisition state. When the condition of frequency not lower than 100kHz is met, the state flow is triggered to enter the phase alignment calibration state. By extracting the fundamental zero-crossing point, it enters the zero-neighborhood interception state. After performing the action of intercepting a 500μs segment, it flows to the polarity admittance calculation state. Then, based on the process of calculating the current-voltage amplitude ratio, it enters the noise dead zone elimination state. After that, after identifying the voltage noise dead zone, it enters the asymmetric evolution judgment state. Finally, when the deviation exceeds the threshold, it returns to the data synchronous acquisition state.

[0037] Example 4: In a power distribution system where the electricity metering box is connected to a distributed photovoltaic power source, due to inverter operation and grid load fluctuations, the fundamental frequency of the circuit shifts within the range of 49.8Hz to 50.2Hz. This dynamic frequency shift results in a 500... The timing misalignment caused by the truncation of the voltage span and the physical zero-crossing phase introduces spurious components from phase positioning errors into the calculated asymmetric deviation, leading to false alarms in the diagnostic system under harmonic backgrounds. After acquiring the high-frequency voltage sampling sequence, the processor initiates a frequency adaptive compensation program based on sample counting, counting the total number of samples N between three adjacent power frequency cycles, and then applying the formula... Calculate the real-time quasi-period duration, where, The quasi-period duration is calculated in real time, where N is the total number of sample points. The sampling frequency is set to 100kHz. The processor uses the Lagrange interpolation algorithm to resample the samples in the neighborhood of the zero-crossing point, locking the index value of the fundamental zero-crossing point in the discrete sequence within the sub-sampling period accuracy range. This ensures that the extracted 500μs local voltage sampling segment is centered on the physical zero-crossing point. During the initialization phase, the processor reads the pre-stored calibration factor from the memory. The high-frequency current sampling sequence is subjected to reverse shift calibration to eliminate the fixed time delay deviation introduced by the data transmission link, so that the voltage and current sampling points are physically aligned in the microsecond dimension.

[0038] When the system operates within the loop current range of 2A to 5A without drastic changes, the processor continuously collects 50 fundamental frequency cycles of asymmetric deviation samples and calculates the standard deviation. The symmetrical benchmark threshold for determining the presence of a nonlinear degradation fault in the contact resistance of the power distribution circuit inside the power metering box is set as follows: Where H is the symmetric reference threshold. The standard deviation is the initial standard deviation obtained in an industrial environment with 20 dB Gaussian white noise. The value is 0.15, and the generated symmetric reference threshold H is 0.75. When there is a 0.5μm oxide film fault in the circuit, the measured asymmetric deviation changes from the background level of 0.12 to 8.52. This value exceeds the symmetric reference threshold H, and the processor outputs a warning signal accordingly. This process will determine the threshold and associate it with the environmental background noise. The frequency adaptive reconstruction method is used to eliminate the dilution of characteristic signals caused by changes in the power grid operating state, thereby realizing the extraction of local physical degradation features.

[0039] Example 5: In the commissioning scenario after the installation of a new energy metering box, the system starts a calibration program for the polarity of the current transformer and the transmission link delay, and the processor obtains the sampling frequency. Sample a 100kHz voltage sequence and identify the phase shift introduced by the analog front-end circuitry. The time delay constant is determined by calculating the time index deviation of the sampling points. Where D is the time delay constant, This is the phase offset. The rated frequency of the power grid is 50Hz. The processor stores the time delay constant as the initial compensation factor in the memory, monitors the common-mode interference level in the circuit, and determines the lower limit of the voltage judgment dead zone based on the envelope peak value of the background noise. When the zero-crossing jitter deviation of the voltage signal waveform is within the preset error range, it confirms that the current circuit meets the benchmark conditions for extracting nonlinear degradation features.

[0040] When the circuit load structure changes or the ambient electromagnetic noise intensity fluctuates dynamically, the processor initiates a boundary update program. Within the acquisition window where the operating current of the power metering box is between 2A and 10A and the waveform characteristics meet a healthy distribution, the processor continuously calculates the asymmetric deviation over 50 fundamental cycles and extracts the statistical variance of the dataset. The symmetric baseline threshold H is updated according to the preset statistical model. Where H is the symmetric reference threshold and k is the scaling factor. To statistically measure the variance, a dynamic threshold compensation mechanism determined by environmental noise is introduced to anchor the system's sensing sensitivity to the fundamental characteristics of the current background electromagnetic environment. This enables the power metering box circuit to enter a monitoring state for nonlinear degradation faults in contact resistance. The preset statistical model here is essentially a probability density estimation function based on the Gaussian distribution of the white noise in the system's underlying physical environment. By importing the discrete deviation set within the aforementioned continuous acquisition period into this function for normal fitting and filtering, transient pulse distortion stray points exceeding three times the standard deviation boundary are filtered out. The second-order central moment is then calculated on the converged subset of data to map out the statistical variance parameter that purely characterizes the steady-state energy width of the noise floor. This provides a reliable underlying support operator for formula calculations, eliminating extreme value interference.

[0041] Example 6: In the initialization scenario for deploying an industrial load-side energy metering box, the processor uses a fourth-order Lagrange interpolation algorithm to reconstruct the discrete sampling points in the neighborhood of the fundamental zero-crossing point, based on a preset sub-sampling period displacement. The physical coordinates of the fundamental zero-crossing point on the time axis are determined. The basis functions of the fourth-order Lagrange interpolation algorithm consist of the voltage amplitudes of four consecutive high-frequency sampling points, controlling 500... The fitting residual between the center position of the local voltage sampling segment and the physical zero crossing point is less than 1. The processor determines the target false alarm probability. Determine the value of the proportionality coefficient k, and calculate the relationship using the inverse function of the standard normal distribution: With the false alarm probability limited to Under the specified operating conditions, the calibration value of parameter k is set to 5.2, where k is the proportionality coefficient. The probability of a false positive is the target. It is the inverse function of the standard normal distribution.

[0042] When the system acquires the ambient background noise caused by the nonlinear leakage current of the sensor, the processor collects the residual voltage sequence on the load side while the circuit breaker is open. It calculates the peak-to-peak distribution of the residual voltage sequence and uses three times this value as the boundary value of the noise dead zone. This yields the distribution range of the simulated front-end thermal noise and electromagnetic induction noise in the 1.5V to 3.0V range. Based on this, the processor locks and eliminates the exclusion interval in the computational logic to remove the initial value interference of the background noise on the calculation results of the first and second admittance sequences. During real-time operation, the processor monitors the envelope consistency of the first and second admittance sequences. When the fluctuation rate of the admittance difference exceeds 20% over three consecutive cycles due to external non-transient pulse interference, the system suspends asymmetric... The deviation is calculated and enters an online self-calibration state, ensuring that the sensing logic of the power metering box circuit is within the effective response bandwidth for the physical barrier signal. This 1.5V to 3.0V dead zone absolute physical range is derived from the derivation conclusion of the non-ideal lead parasitic capacitance model of the mainstream 0.5-class wideband current transformer in the industry under room temperature conditions, reflecting the physical upper limit of the noise floor that can be excited by the leakage current of the general analog front end. The system uses this theoretical electrical parameter base range with universal engineering applicability as the initial value mapping coordinate, and performs adaptive proportional reduction alignment with the actual three times residual peak-to-peak value measured on site, ensuring that the dead zone setting has both the rigorous derivation support of the laws of general electronic components and can adapt to the background differences of aging equipment in different industrial sites.

[0043] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0044] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for diagnosing circuit faults in an electricity metering box, characterized in that, Includes the following steps: Step 101: Simultaneously acquire the high-frequency voltage sampling sequence and high-frequency current sampling sequence of the load side of the power metering box, wherein the sampling frequency is not less than 100kHz. Step 102: Extract the fundamental zero-crossing point of the high-frequency voltage sampling sequence, calculate the phase deviation between the high-frequency voltage sampling sequence and the high-frequency current sampling sequence, and perform time-shift calibration on the high-frequency current sampling sequence based on the phase deviation to obtain the calibrated current sequence. Step 103: Using the fundamental zero-crossing point as the center, extract a local voltage sampling segment with a span of 500μs and a synchronous local current sampling segment. Step 104: Based on the slope polarity of the local voltage sampling segment, divide the intercepted sampling points into a set of rising edge zero crossings and a set of falling edge zero crossings. Step 105: Calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of rising edges to obtain the first admittance sequence; simultaneously calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of falling edges to obtain the second admittance sequence. Step 106: Identify noise dead zone sampling points in the local voltage sampling segment whose absolute values ​​are between 1.5V and 3.0V, and remove the corresponding sampling points in the first admittance sequence and the second admittance sequence; Step 107: Extract the admittance difference at the symmetrical phase angle position in the first admittance sequence and the second admittance sequence, calculate the cumulative integral of the admittance difference within the local voltage sampling segment span, and generate the asymmetric deviation. Step 108: When the asymmetric deviation exceeds the preset symmetric reference threshold, it is determined that there is a contact resistance deterioration fault in the power distribution circuit inside the power metering box, and an early warning signal is output.

2. The method for diagnosing circuit faults in an energy metering box according to claim 1, characterized in that, Step 102 is further refined into the following sub-steps: Step 1021, identify the low-voltage waveform interval in the high-frequency voltage sampling sequence where the sampled value is lower than the preset voltage threshold; Step 1022, shield the sampling point data in the low-voltage waveform interval and determine the zero offset of the voltage transformer; Step 1023, use the waveform stable segment of the high-frequency voltage sampling sequence to synchronously align the calibration current sequence, and use a phase compensation algorithm based on the fundamental zero-crossing time difference to make the calibrated sampling data have a reference with physical symmetry.

3. The method for diagnosing circuit faults in an energy metering box according to claim 1, characterized in that, Step 105 is further refined into the following sub-steps: Step 1051, calculate the transient admittance value corresponding to each sampling time of the local voltage sampling segment; Step 1052, map the transient admittance value to the corresponding transient voltage change rate to generate a dynamic admittance evolution curve centered on the fundamental zero-crossing point, wherein the dynamic admittance evolution curve is used to characterize the energy loss state of the metal connection point during the polarity reversal process.

4. The method for diagnosing circuit faults in an energy metering box according to claim 1, characterized in that, Step 107 is further refined into the following sub-steps: Step 1071, obtain the first admittance characteristic component in the positive zero-crossing direction and the second admittance characteristic component in the negative zero-crossing direction; Step 1072, calculate the numerical deviation between the first admittance characteristic component and the second admittance characteristic component, and quantify the degree of dynamic admittance asymmetry generated in the positive and negative zero-crossing directions; Step 1073, perform an integral operation on the numerical deviation over the time scale to obtain the asymmetric deviation.

5. The method for diagnosing circuit faults in an energy metering box according to claim 1, characterized in that, Step 108 is followed by the following steps: Step 109, retrieve the historical fault diagnosis records of the power metering box; Step 110, extract the evolution slope of the asymmetric deviation as the operating cycle changes; Step 111, when the evolution slope shows a monotonically increasing trend and the rate of increase exceeds the preset rate of change threshold, determine that the internal connection point of the power metering box is in the deterioration evolution period and output a diagnosis signal.

6. The method for diagnosing circuit faults in an energy metering box according to claim 1, characterized in that, Step 101 is further refined into the following sub-steps: Step 1011, using a wideband current transformer to collect analog voltage signals and analog current signals from the load side of the power metering box; Step 1012, using an analog-to-digital converter circuit to convert the analog voltage signals and analog current signals into high-frequency voltage sampling sequences and high-frequency current sampling sequences.

7. The method for diagnosing circuit faults in an energy metering box according to claim 1, characterized in that, Step 107 further includes the following steps: Step 1074, mapping the sampled data after phase deviation calibration to a normalized unit coordinate system; Step 1075, calculating the center symmetry deviation value of the dynamic admittance evolution curve relative to the origin in the normalized unit coordinate system.

8. The method for diagnosing circuit faults in an energy metering box according to claim 1, characterized in that, Step 108 determines that there is a contact resistance deterioration fault in the power distribution circuit inside the power metering box, including: Step 1081, matching the asymmetric deviation with the fault feature fingerprints corresponding to different degrees of deterioration in the preset database; Step 1082, identifying the specific fault type in the power distribution circuit of the power metering box caused by loose terminals or oxidation of contacts.

9. The method for diagnosing circuit faults in an energy metering box according to claim 1, characterized in that, Step 108 also includes the following steps: Step 1083, determine the fault risk level based on the value of the asymmetric deviation; Step 1084, output the corresponding early warning command based on the fault risk level, and upload the fault diagnosis result to the operation and maintenance management platform.

10. A circuit fault diagnosis system for an electricity metering box, used to implement the circuit fault diagnosis method for an electricity metering box as described in claim 1, characterized in that, include: The data acquisition module is used to synchronously acquire the high-frequency voltage sampling sequence and high-frequency current sampling sequence of the load side of the power metering box, wherein the sampling frequency is not less than 100kHz. The phase calibration module is used to extract the fundamental zero-crossing point of the high-frequency voltage sampling sequence, calculate the phase deviation between the high-frequency voltage sampling sequence and the high-frequency current sampling sequence, and perform time-shift calibration on the high-frequency current sampling sequence based on the phase deviation to obtain the calibration current sequence. The sample extraction module is used to extract a local voltage sampling segment with a span of 500μs, centered on the zero-crossing point of the fundamental frequency, as well as a synchronous local current sampling segment. The polarity classification module is used to divide the intercepted sampling points into rising edge zero-crossing set and falling edge zero-crossing set according to the slope polarity of the local voltage sampling segment. The admittance calculation module is used to calculate the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of rising edges to obtain the first admittance sequence; and simultaneously calculates the current-to-voltage amplitude ratio at each sampling point in the zero-crossing set of falling edges to obtain the second admittance sequence. The noise filtering module is used to identify noise dead zone sampling points in the local voltage sampling segment with an absolute value between 1.5V and 3.0V, and to remove the corresponding sampling points in the first admittance sequence and the second admittance sequence. The asymmetric analysis module is used to extract the admittance difference at the symmetrical phase angle position in the first admittance sequence and the second admittance sequence, calculate the cumulative integral of the admittance difference within the local voltage sampling segment span, and generate an asymmetric deviation quantity that reflects the nonlinear barrier characteristics of the connection point. The fault determination module is used to determine that there is a contact resistance deterioration fault in the power distribution circuit inside the power metering box when the asymmetric deviation exceeds the preset symmetric reference threshold, and outputs a warning signal.

Images