A method for identifying and monitoring resistive capacitive residual current of an electrical device

Abstract

This invention relates to the field of power system insulation monitoring technology and discloses a method for identifying and monitoring resistive and capacitive residual current in electrical equipment. The method includes: synchronously acquiring a bus voltage reference signal and a total residual current signal; extracting the zero-crossing moment of the lockout voltage using the fundamental component; dividing the time-series orthogonal intervals and calculating the piecewise charge integral; comparing the integral differences between the positive and negative half-wave mirror symmetrical positions to determine the phase shift angle caused by the nonlinearity of the measurement link; shifting the integral boundary accordingly to maintain the physical phase orthogonality; and, based on the intrinsic independence of resistive and capacitive components, using a covariance optimization mechanism to remove projection interference and output resistive residual current data. This invention utilizes waveform symmetry to self-calibrate the sensor's dynamic phase shift, eliminating the projection penetration of capacitive components onto the resistive observation axis, ensuring that the identified data corresponds to the true physical loss properties of the insulation medium, and improving the accuracy of insulation degradation trend assessment.

Description

Technical Field

[0001] This invention belongs to the field of power system insulation monitoring technology, and particularly relates to a method for identifying and monitoring resistive and capacitive residual current in electrical equipment. Background Technology

[0002] Currently, in the operation and maintenance of power distribution systems and high-voltage electrical equipment, monitoring residual current to assess the insulation status of equipment is a common technical method. The usual practice is to use a residual current transformer to collect the total residual current signal in the circuit and determine the insulation health status based on whether the current value exceeds a preset threshold. This mode can provide basic data support for preventing electrical faults in a purely linear load environment. However, due to the widespread use of frequency conversion drive equipment and long-distance power cables in industrial sites, the power distribution system generates a huge distributed capacitance to ground. At this time, the amplitude of the capacitive component is usually tens of times that of the resistive leakage current. In order to avoid frequent false trips caused by the fluctuation of the capacitive component, the existing technology often adopts the strategy of raising the alarm threshold to maintain the continuity of power supply. This approach makes it difficult for the system to identify early minor insulation degradation trends, leaving insulation hazards in the monitoring blind zone for a long time.

[0003] Although the industry has attempted to remove resistive components by adding hardware filters or injecting probe signals, these methods are susceptible to harmonic interference under complex operating conditions. The fundamental problem lies in the fact that strong magnetic fields and ambient temperature drift in industrial environments cause dynamic and nonlinear phase shifts in the transformer core. This phase distortion, induced by the physical characteristics of the sensing hardware, physically displaces the original orthogonal reference frame, resulting in large capacitive components projecting onto the resistive observation axis and causing identification errors. While adding hardware filters or injecting probe signals can improve monitoring sensitivity, software-level synchronization and data processing algorithms still have limitations. For example, the method described in publication number CN118191663A... A Chinese invention patent application discloses a method for detecting the instantaneous value of residual current based on a dual synchronization mechanism. The dual synchronization mechanism achieves logical alignment of the sampling time axis of multiple modules and optimizes the current synthesis accuracy of dual power supply loops. However, under actual complex working conditions, the nonlinear magnetization of the transformer core and the large changes in ambient temperature induce dynamic physical phase shifts in the sensing link. The phase distortion dominated by physical laws causes a physical offset between the sampling window and the true phase of the voltage fundamental. The aforementioned scheme only focuses on the logical synchronization of multi-point sampling and lacks a self-compensation mechanism for nonlinear distortion at the sensor bottom layer. When applied to the monitoring of large capacitive background insulation, large numerical capacitive components inevitably project onto the resistive observation axis, masking the true physical loss.

[0004] Therefore, the technical problem to be solved by this invention is how to eliminate the interference of dynamic changes in the distributed capacitance of the power grid on the insulation status assessment, offset the projection crosstalk caused by the dynamic phase shift of the sensing hardware, and extract high-purity resistive leakage index to characterize the true health status of the equipment. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A method for identifying and monitoring resistive and capacitive residual current in electrical equipment, comprising: Step 101: Synchronously acquire the bus voltage reference signal and the total residual current signal, and extract the fundamental component of the bus voltage reference signal to obtain the voltage zero crossing time; Step 102: Divide the time into four time-series orthogonal intervals based on the voltage zero-crossing time, and calculate the piecewise charge integral value of the total residual current signal in each time-series orthogonal interval to generate a primary integral characteristic quantity that characterizes the aliasing state of resistive and capacitive components. Step 103: Compare the piecewise charge integral values ​​at symmetrical positions within adjacent half-wave periods, calculate the half-wave difference components of the two, and determine the phase shift angle caused by the nonlinearity of the magnetic circuit of the induction unit based on the preset proportional mapping relationship between the half-wave difference components and the phase shift angle. Step 104: Based on the dynamic physical phase offset angle, the integral time domain boundary of the time-series orthogonal interval is compensated and shifted in real time. By adjusting the sampling start point and the phase span of the sampling window, the integral time domain boundary after the shift correction is kept absolutely orthogonal to the fundamental physical phase of the bus voltage reference signal. Step 105: Within each integral time domain boundary after translation correction, based on the intrinsic independence of capacitive and resistive components, iterative optimization is performed by minimizing the covariance between the total residual current signal and the capacitive projection component, thereby removing the projection interference of the capacitive component on the resistive axis, and calculating and outputting the decoupled resistive residual current data.

[0006] Preferably, step 104 further includes the following sub-steps: step 1041, obtaining phase compensation weights characterizing changes in ambient temperature and nonlinear saturation characteristics of the magnetic core; step 1042, using the phase compensation weights to perform weighted calculations on the phase offset angles to generate a phase compensation time step; step 1043, using the phase compensation time step to synchronously shift the sampling start time of each time-series orthogonal interval.

[0007] Preferably, the process of stripping component interference in step 105 includes the following sub-steps: Step 1051, extracting the decoupled resistive sequence and capacitive sequence within a preset sliding window; Step 1052, calculating the correlation coefficient between the resistive sequence and the capacitive sequence, and determining whether the correlation coefficient reaches a preset independence threshold; Step 1053, if the correlation coefficient does not reach the independence threshold, adjusting the translation step size of the integral boundary according to the positive or negative polarity of the correlation coefficient until the correlation coefficient converges to a minimum value.

[0008] Preferably, before step 102, the method further includes the following steps: step 401, performing low-pass filtering on the acquired total residual current signal to extract the fundamental residual current component; step 402, performing amplitude normalization processing on the fundamental residual current component to generate a unit reference vector for projection decoupling.

[0009] Preferably, after step 105, the following steps are also included: Step 601, recording the resistive residual current data within the continuous monitoring period and fitting the time evolution curve of the insulation degradation trend; Step 602, calculating the slope of the time evolution curve and generating an insulation status warning command when the slope exceeds a preset change rate threshold of 15%.

[0010] Preferably, obtaining the voltage zero-crossing time in step 101 specifically involves: using a digital phase-locked loop algorithm to track the phase of the bus voltage reference signal and correcting the sampling frequency in real time, so that the deviation between the sampling frequency and the grid fundamental frequency is kept within 0.01Hz.

[0011] Preferably, the four time-series orthogonal intervals in step 102 correspond to the 0 to π / 2, π / 2 to π, π to 3π / 2, and 3π / 2 to 2π radian regions of the fundamental period of the bus voltage reference signal, respectively.

[0012] Preferably, the process of obtaining the nonlinear saturation characteristics of the magnetic core in step 1041 includes the following sub-steps: Step 10411, real-time acquisition of magnetic flux change rate data on the secondary side of the current transformer; Step 10412, based on the magnetic flux change rate data, a preset magnetization curve mapping table is retrieved to determine the phase lag correction amount corresponding to the current magnetic flux state, and the phase lag correction amount is included in the phase compensation weight.

[0013] Preferably, the method further includes the following steps: Step 1001, performing cluster calculation on the resistive residual current data within the historical period to establish a baseline distribution of insulation parameters under normal operating conditions of the electrical equipment; Step 1002, calculating the deviation of the resistive residual current data obtained by real-time decoupling from the baseline distribution of insulation parameters, and outputting the health rating result of the electrical equipment based on the deviation.

[0014] Compared with existing technologies, the resistive and capacitive residual current identification and monitoring method for electrical equipment of the present invention has the following advantages: 1. In the identification and monitoring of resistive and capacitive residual current in electrical equipment, by extracting the fundamental component of the phase voltage signal and locking the zero-crossing time of the voltage, four time-series orthogonal integration intervals are constructed. The total residual current signal is processed by piecewise definite integration in the time domain. This mechanism utilizes the phase difference between resistive leakage current and capacitive leakage current in the AC cycle to separate the resistive component mixed in the total current from the high-amplitude capacitive background. This decoupling method based on physical phase difference eliminates the interference of dynamic changes in the distributed capacitance of the power grid on insulation monitoring, so that the output resistive current component can directly correspond to the energy loss properties of the insulation medium, thereby improving the sensitivity of insulation degradation identification.

[0015] 2. To address the dynamic phase shift caused by nonlinear magnetization of the transformer core and ambient temperature drift, the method introduces boundary self-calibration logic based on the mirror symmetry of physical waveforms. The system generates a time step fine-tuning amount by comparing the differential balance state of the current integral in adjacent half-waves, and performs real-time translation correction on the time domain boundary of the integration interval. This closed-loop calibration process ensures that the integration window is always aligned with the true phase of the voltage fundamental wave, blocking the projection of the large capacitive vector into the resistive observation axis. Under the premise of utilizing the existing acquisition hardware of the system, the method eliminates the pseudo-resistive leakage increment caused by the phase distortion of the measurement link through adaptive adjustment at the data logic level, ensuring the physical authenticity of the health status assessment data. Attached Figure Description

[0016] Figure 1 This is the main flowchart of the resistive-capacitive residual current decoupling and dynamic phase shift compensation of the present invention; Figure 2 This is the correlation iterative optimization and insulation state evolution early warning diagram of the present invention. Detailed Implementation

[0017] 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.

[0018] A method for identifying and monitoring resistive and capacitive residual current in electrical equipment, comprising: Step 101: Synchronously acquire the bus voltage reference signal and the total residual current signal, and extract the fundamental component of the bus voltage reference signal to obtain the voltage zero crossing time; Step 102: Divide the time into four time-series orthogonal intervals based on the voltage zero-crossing time, and calculate the piecewise charge integral value of the total residual current signal in each time-series orthogonal interval to generate a primary integral characteristic quantity that characterizes the aliasing state of resistive and capacitive components. Step 103: Compare the piecewise charge integral values ​​at symmetrical positions within adjacent half-wave periods, calculate the half-wave difference components of the two, and determine the phase shift angle caused by the nonlinearity of the magnetic circuit of the induction unit based on the preset proportional mapping relationship between the half-wave difference components and the phase shift angle. Step 104: Based on the dynamic physical phase offset angle, the integral time domain boundary of the time-series orthogonal interval is compensated and shifted in real time. By adjusting the sampling start point and the phase span of the sampling window, the integral time domain boundary after the shift correction is kept absolutely orthogonal to the fundamental physical phase of the bus voltage reference signal. Step 105: Within each integral time domain boundary after translation correction, based on the intrinsic independence of capacitive and resistive components, iterative optimization is performed by minimizing the covariance between the total residual current signal and the capacitive projection component, thereby removing the projection interference of the capacitive component on the resistive axis, and calculating and outputting the decoupled resistive residual current data.

[0019] Preferably, step 104 further includes the following sub-steps: step 1041, obtaining phase compensation weights characterizing changes in ambient temperature and nonlinear saturation characteristics of the magnetic core; step 1042, using the phase compensation weights to perform weighted calculations on the phase offset angles to generate a phase compensation time step; step 1043, using the phase compensation time step to synchronously shift the sampling start time of each time-series orthogonal interval.

[0020] Preferably, the process of stripping component interference in step 105 includes the following sub-steps: Step 1051, extracting the decoupled resistive sequence and capacitive sequence within a preset sliding window; Step 1052, calculating the correlation coefficient between the resistive sequence and the capacitive sequence, and determining whether the correlation coefficient reaches a preset independence threshold; Step 1053, if the correlation coefficient does not reach the independence threshold, adjusting the translation step size of the integral boundary according to the positive or negative polarity of the correlation coefficient until the correlation coefficient converges to a minimum value.

[0021] Preferably, before step 102, the method further includes the following steps: step 401, performing low-pass filtering on the acquired total residual current signal to extract the fundamental residual current component; step 402, performing amplitude normalization processing on the fundamental residual current component to generate a unit reference vector for projection decoupling.

[0022] Preferably, the calculation of the half-wave difference component in step 103 follows the formula below: ,in, For half-wave difference components, This represents the piecewise charge integral value within the first time-orthogonal interval of the positive half-cycle. It represents the piecewise charge integral value within the time-series orthogonal interval corresponding to the symmetrical position in the negative half-cycle.

[0023] Preferably, after step 105, the following steps are also included: Step 601, recording the resistive residual current data within the continuous monitoring period and fitting the time evolution curve of the insulation degradation trend; Step 602, calculating the slope of the time evolution curve and generating an insulation status warning command when the slope exceeds a preset change rate threshold of 15%.

[0024] Preferably, obtaining the voltage zero-crossing time in step 101 specifically involves: using a digital phase-locked loop algorithm to track the phase of the bus voltage reference signal and correcting the sampling frequency in real time, so that the deviation between the sampling frequency and the grid fundamental frequency is kept within 0.01Hz.

[0025] Preferably, the four time-series orthogonal intervals in step 102 correspond to the 0 to π / 2, π / 2 to π, π to 3π / 2, and 3π / 2 to 2π radian regions of the fundamental period of the bus voltage reference signal, respectively.

[0026] Preferably, the process of obtaining the nonlinear saturation characteristics of the magnetic core in step 1041 includes the following sub-steps: Step 10411, real-time acquisition of magnetic flux change rate data on the secondary side of the current transformer; Step 10412, based on the magnetic flux change rate data, a preset magnetization curve mapping table is retrieved to determine the phase lag correction amount corresponding to the current magnetic flux state, and the phase lag correction amount is included in the phase compensation weight.

[0027] Preferably, the method further includes the following steps: Step 1001, performing cluster calculation on the resistive residual current data within the historical period to establish a baseline distribution of insulation parameters under normal operating conditions of the electrical equipment; Step 1002, calculating the deviation of the resistive residual current data obtained by real-time decoupling from the baseline distribution of insulation parameters, and outputting the health rating result of the electrical equipment based on the deviation.

[0028] Example 1: In a continuously operating large-scale industrial microgrid power distribution system, the system includes long-distance high-voltage cables and high-power frequency converter drive equipment. The operating environment faces diurnal temperature fluctuations and periodic heavy load impact conditions. Under such conditions, the inherent parasitic capacitance of long-distance cables and frequency converter equipment generates high-amplitude capacitive residual currents, which are tens of times larger than resistive residual currents. At the same time, strong magnetic field interference and drastic changes in ambient temperature cause nonlinear magnetization of the current transformer core, resulting in dynamic and nonlinear physical phase shifts in the measurement link. This dynamic phase distortion causes the static phase synchronization mechanism to fail, resulting in the capacitive vector being projected onto the resistive observation axis and causing projection interference. The high-amplitude capacitive current noise and distortion error mask the small resistive increments that characterize the insulation degradation state, leading to technical problems such as false alarms in insulation status warnings and low sensitivity.

[0029] The monitoring system synchronously acquires the bus voltage reference signal and the total residual current signal. It extracts the fundamental component of the bus voltage reference signal to obtain the voltage zero-crossing moment. Based on this zero-crossing moment, it divides the system into four orthogonal time intervals. Based on the intrinsic physical property that the definite integral of the pure sinusoidal fundamental signal is zero within a complete physical cycle, and considering the monotonically accumulating error generated by the DC bias component during time integration, the system initiates a zero-point drift suppression procedure before calculating the piecewise charge integral value. Starting from the voltage zero-crossing moment, it extracts the arithmetic mean of the total residual current signal within the complete grid fundamental cycle time window. Using the DC bias reference as the basis, the system subtracts the DC bias reference point by point from the total residual current signal sequence to generate a zero-bias residual current signal. Subsequent integration operations are all based on the zero-bias residual current signal, and the piecewise charge integral value of the total residual current signal in each time-orthogonal interval is calculated to generate a primary integral characteristic quantity representing the aliasing state of resistive and capacitive components. Based on the mirror-symmetric physical property of the capacitive leakage current in the AC half-wave, the system compares the piecewise charge integral values ​​at symmetrical positions in adjacent half-wave periods to calculate the half-wave difference component. The specific calculation formula is as follows: ,in, For half-wave difference components, This represents the piecewise charge integral value within the first time-orthogonal interval of the positive half-cycle. This represents the piecewise charge integral value within the time-series orthogonal interval corresponding to the symmetrical position in the negative half-cycle. In this physical calculation process, due to the continuous impact of high-amplitude capacitive currents, which are tens of times greater than the resistive component, in the industrial environment, the dynamic hysteresis loop of the high-permeability crystal core of the current transformer is forcibly pushed into the nonlinear operating region during the alternating positive and negative magnetization process. This surface magnetic circuit saturation causes an asymmetrical physical distortion in the rising edge response slope and falling edge attenuation slope of the induced current on the secondary side. Therefore, even with a pure sine wave input that eliminates DC bias, the output waveform after coupling through the transformer has lost its odd harmonic symmetry. Consequently, the positive and negative half-waves, which originally had a theoretical integral difference of zero, generate residual accumulated charge in a specific interval. The actual physical charge amount of this accumulation breaks the theoretical paradox, making the reverse quantitative characterization of the nonlinear distortion and phase drift of the core using the half-wave difference component a directly extractable engineering physical feature.

[0030] The system calculates the phase shift angle caused by the nonlinearity of the magnetic circuit of the induction unit based on a preset proportional mapping relationship between the half-wave differential component and the phase shift angle. Then, based on this physical phase shift angle, it shifts the integral time-domain boundaries of the orthogonal time-series intervals. By adjusting the sampling start point and the phase span of the sampling window, the system ensures that the shifted and corrected integral time-domain boundaries remain orthogonal to the fundamental physical phase of the bus voltage reference signal. Within each shifted and corrected integral time-domain boundary, the system extracts resistive and capacitive sequences within a sliding window covering multiple grid start-up and shutdown cycles. Based on the physical properties that the resistive residual current exhibits slow time-varying characteristics and the capacitive residual current exhibits high-frequency fluctuation characteristics, the system calculates the correlation coefficient between the resistive and capacitive sequences. When the correlation coefficient is greater than or equal to a preset independence threshold, the system adjusts the translation step size of the integral boundary according to the positive or negative polarity of the correlation coefficient and performs an iterative calculation process until the correlation coefficient is less than the independence threshold. This process is used to remove the projection component of the capacitive component onto the resistive axis. The system outputs the decoupled resistive residual current data and records this data within a continuous monitoring period to fit the time evolution curve of the insulation degradation trend. When the slope of the calculated time evolution curve is greater than or equal to a preset rate of change threshold of 15%, the system generates an insulation status early warning command. The output resistive characteristic index is used to characterize the physical loss of the insulating medium, constituting the insulation degradation trend assessment data under the fault prediction and health management architecture.

[0031] Example 2: In a physical test platform simulating insulation degradation of high-voltage variable frequency drive equipment, a variable frequency load with a rated power of 500kW and a 10kV cross-linked polyethylene cable with a length of 2km were configured to construct the measurement error source of the test platform. Gaussian white noise with a signal-to-noise ratio of 20dB was injected into the current transformer measurement link, along with an external power frequency harmonic interference signal with an amplitude of 15% of the fundamental component. A dynamic ambient temperature disturbance from 25℃ to 85℃ was applied at a constant heating rate using a temperature control box, inducing nonlinear drift in the permeability of the current transformer core and generating dynamic phase distortion. Under this superimposed disturbance condition, the actual resistive leakage current reference value was set to 10.5mA. The total capacitive residual current induced during the frequency converter start-up and shutdown process was... The residual current amplitude is 480.0 mA. The time span of the sliding data extraction window is selected based on the mapping relationship between the real-time data acquisition and the data processing load. When the rate of change of the ambient temperature at the measurement link is greater than the predetermined temperature change threshold, the dynamic phase drift of the current transformer exhibits high-frequency abrupt change characteristics. To avoid phase calibration lag, the time span of the sliding data extraction window tends to the lower limit of the value. The rate of change of the ambient temperature is set to a temperature rise of 5℃ / min. Based on the physical mapping rule of the negative correlation between the rate of change of the ambient temperature and the phase shift response time, the time span of the sliding data extraction window is determined to be 512 grid fundamental frequency cycles, taking into account both the phase shift capture resolution and the data processing load of the computing system.

[0032] Comparative sample group 1, comparative sample group 2, and the sample group of this invention were established. Comparative sample group 1 adopted a static zero-crossing phase synchronization mechanism to shield the integral boundary shift and covariance optimization. Comparative sample group 2 enabled the integral boundary shift and set the independence threshold of the correlation coefficient to 0.50 (out-of-range index). The sample group of this invention enabled the integral boundary shift and covariance optimization and set the independence threshold of the correlation coefficient to 0.05. When the ambient temperature rises to 80℃, the current transformer generates a physical phase shift. The sample group of this invention synchronously acquires the bus voltage reference signal and the total residual current signal, and calculates the piecewise charge integral value in the first time-series orthogonal interval within the positive half-cycle. Given a temperature of 45.2 μC, calculate the piecewise charge integral value within the time-series orthogonal interval at the symmetrical position during the negative half-cycle. The temperature is 38.6 μC, according to the mathematical formula. Calculate the half-wave difference components, where, For half-wave difference components, This represents the piecewise charge integral value within the first time-orthogonal interval of the positive half-cycle. The piecewise charge integral values ​​within the time-series orthogonal intervals at symmetrical positions within the negative half-cycle are used to calculate the half-wave difference component. The value is 6.6 μC; based on this half-wave difference component, the phase shift angle caused by the nonlinearity of the magnetic circuit of the induction unit is determined to be 3.15°, and the time domain boundaries of each integration are shifted by 3.15°.

[0033] After shifting the integral boundary, the sample group of this invention extracts resistive and capacitive sequences within a sliding window of 512 power grid cycles. The initial calculation shows that the correlation coefficient between the resistive and capacitive sequences is 0.42. The specific underlying logic of the above sequence extraction operation is as follows: The monitoring system treats each fundamental power grid cycle within the sliding window as an independent discrete sampling frame. Within each frame, the system locks the zero-point coordinates after shift correction, directly extracts the piecewise charge integral values ​​within the absolute phase intervals of the corresponding bus voltage reference signal from 0 to π / 2 and from π to 3π / 2, and sums them to generate a single resistive data point for that frame. At the same time, it extracts the piecewise charge integral values ​​within the corresponding phase intervals from π / 2 to π and from 3π / 2 to 2π and sums them to generate... The system generates a single capacitive data point for the frame. It then arranges and assembles the scattered points generated within the 512 sampling frames sequentially along the time axis. From the original aliased total current signal, it physically and discretely separates resistive and capacitive data arrays with independent dimensions that can be directly used for mathematical correlation functions. Based on the positive polarity of the correlation coefficient, the system adjusts the translation of the integration boundary in 0.1° increments. After seven iterations, the correlation coefficient converges to 0.04, which is less than the independence threshold of 0.05. The decoupled resistive residual current data output by the sample group of this invention is 10.8 mA. In contrast, sample group one without dynamic projection interference and aliased capacitive components on the resistive observation axis outputs 42... The resistance residual current of the first sample group was 0.5mA. In contrast, the independence threshold of the second sample group exceeded the optimal range, and the iterative calculation terminated when the correlation coefficient dropped to 0.48, failing to filter out projection interference. The output resistive residual current data was 18.3mA. Neither static phase shift compensation nor an out-of-range iterative threshold could suppress the nonlinear penetration of the capacitive component. The physical coupling effect of integral boundary shifting and covariance optimization eliminated the masking effect of the underlying distortion of the measurement link on the insulation degradation characteristics. The resistance residual current data of the first sample group increased from 15.2mA to 65.7mA, showing an exponential growth deviation. In the range where the ambient temperature change rate is below 8℃ / min, the sample group of this invention... The output resistive residual current data is stably distributed in the range of 10.6mA to 10.9mA, showing stability against temperature disturbances and core distortion. When the temperature change rate exceeds the physical response limit of 8℃ / min, the transformer core enters a deep magnetic saturation state, and the half-wave differential component loses its linear mapping basis with the phase shift angle. The variance of the resistive data output by the sample group of this invention increases. This performance inflection point verifies that the boundary setting of the sliding window parameter and the independence threshold conforms to the objective physical limitations of electromagnetic materials. The resistive residual current data output by the sample group of this invention filters out the phase shift interference at the bottom layer of the system, characterizes the physical loss of the insulating medium, and constitutes the data basis for assessing the insulation degradation trend under the fault prediction and health management architecture.

[0034] Example 3: In industrial microgrid power distribution scenarios facing temperature fluctuations, the magnetic core of the current transformer undergoes nonlinear magnetization with temperature changes, inducing dynamic physical phase drift in the measurement link. Constant compensation mechanisms are insufficient to eliminate this dynamic projection interference. The monitoring system collects ambient temperature data from the current transformer and calculates the quotient of the ambient temperature difference between adjacent sampling periods and the sampling time interval to generate the temperature change rate. The monitoring system calculates the phase compensation weight based on the ratio of the temperature change rate to a preset reference temperature change rate. The specific calculation formula is as follows: ,in, For phase compensation weights, The rate of change of ambient temperature. The physical basis for this overall thermodynamic proportional parameter, which is a preset reference temperature change rate, to directly reflect and compensate for the nonlinear saturation characteristics of the surface electromagnetics is that a sharp rise or fall in temperature at the industrial site will induce a severe transient thermodynamic temperature gradient between the external hard encapsulation layer and the internal high-permeability nanocrystalline magnetic core of the current transformer. Due to the rigidity difference in the thermal expansion coefficients of the materials, the mechanical thermal stress converted from this gradient will forcibly squeeze the boundary of the internal surface magnetic domains through the material piezomagnetic effect, causing the response time of the magnetic core to reach saturation magnetic flux density under high-frequency and high-current impact to be nonlinearly delayed. Therefore, the minute-level overall temperature change rate measured outside the system is extracted as a priori compensation weight, which can quantitatively offset the millisecond-level grid-induced phase distortion error caused by the deformation of the surface magnetic domains in a feedforward manner in the physical path.

[0035] The monitoring system synchronously acquires the bus voltage reference signal and the total residual current signal. It compares the piecewise charge integral values ​​at symmetrical positions within adjacent half-wave cycles to calculate the half-wave difference component. Based on this half-wave difference component, the monitoring system calculates the base phase offset angle. The product of the base phase offset angle and the aforementioned phase compensation weight generates the phase compensation time step. The monitoring system uses this phase compensation time step to shift the sampling start time of the orthogonal time-series interval. After shifting the integration boundary, the monitoring system synchronously extracts the resistive and capacitive sequences within the sliding window and calculates their correlation coefficient. In this covariance optimization and correlation calculation process, the capacitive sequence (i.e., the capacitive projected component) is not an unknown black-box variable that the system needs to solve in advance. Instead, the monitoring system directly uses the translation-corrected integration boundary under the current trial step size as the forced measurement coordinate system. The transient deterministic value is obtained by performing forced arithmetic integration on the real-time and unchanging total residual current physical signal within the capacitive orthogonal time window corresponding to the coordinate system. Each boundary translation generates a set of completely known and deterministic transient capacitive and transient resistive data. These two deterministic data sources are fed into the covariance calculation formula as basic inputs, thereby reducing the deadlock analytical problem of nonlinear equations to a unidirectional positive trial-and-error iterative process of finding the minimum covariance value in a bounded phase space. The monitoring system determines whether the correlation coefficient is less than the preset independence threshold. When the correlation coefficient is greater than or equal to the independence threshold, the monitoring system starts the optimization iterative operation. The monitoring system extracts the absolute value of the correlation coefficient and calculates the product of the basic translation step size and the absolute value to generate the dynamic adjustment step size. The specific calculation formula is as follows: ,in, To dynamically adjust the step size, To preset the basic translation step size, is the correlation coefficient between resistive and capacitive sequences.

[0036] Based on the geometric mapping property that the sign of the cross-covariance algebra remains consistent with the direction of the phase deflection angle when two sets of orthogonal sinusoidal signals deviate in phase, the monitoring system reads the sign bit of the correlation coefficient to determine the translation direction. If the sign bit is positive, it is determined that the capacitive component has generated a leading projection onto the resistive observation axis, thus determining the forward translation direction. If the sign bit is negative, it is determined that the capacitive component has generated a lagging projection, thus determining the forward translation direction. The monitoring system then applies a dynamic adjustment step size to translate the integral boundary again based on the above translation direction. Based on the updated integral time domain boundary, the resistive and capacitive sequences are re-extracted and the correlation coefficient is recalculated. The optimization iterative operation continues until the correlation coefficient is less than the independence threshold. The resistive sequence at the end of the optimization iterative operation is output as the decoupled resistive residual current data. This calculation process avoids the oscillation and divergence phenomenon caused by a constant step size, compensates for the dynamic phase shift interference induced by external environmental temperature fluctuations, and the output resistive characteristic index characterizes the physical loss properties of the insulating medium, constituting the data basis for assessing the insulation degradation status under the fault prediction and health management framework.

[0037] Example 4: In a power distribution network deployment scenario with cables of unknown material, the monitoring system initiates a pre-calibration process under no-load conditions with the load de-energized. The monitoring system simultaneously acquires the bus voltage reference signal and the total residual current signal, extracts the piecewise charge integral values ​​at symmetrical positions within adjacent half-wave cycles at different sweep frequencies to calculate the reference half-wave difference component, and reads the physical phase offset angle recorded by the phase meter at the corresponding frequency. It then fits the reference half-wave difference component with the physical phase offset angle to extract the scatter plot slope, which is determined as the coefficient of a preset proportional mapping relationship. The specific calculation formula is as follows: ,in, For mapping coefficients, This is the physical phase offset angle. Using the reference half-wave differential component, the monitoring system writes this coefficient into the underlying register. While performing the above-mentioned no-load hardware coefficient calibration process, the monitoring system simultaneously initiates an offline exhaustive calibration operation targeting the magnetization characteristics of a specific material. It injects a multi-band high-voltage simulated current with known precise phase into the primary side of the current transformer under test in a stepped manner with a preset step size. Using a high-frequency digital oscilloscope and a current probe, it synchronously records the actual magnetic flux change rate of the secondary side induced output and aligns and compares to extract the corresponding waveform initial phase hysteresis correction value. The tens of thousands of sets of discrete comparison points are input into the host computer for two-dimensional polynomial surface fitting and smoothing. After removing random measurement noise, the data is solidified into a static two-dimensional data index table structure. This structure is permanently burned into the read-only memory unit of the device's microprocessor as the aforementioned preset magnetization curve mapping table, thereby providing reliable historical measured underlying data support for real-time fast table lookup after the device is deployed and put into operation.

[0038] After the load reaches the rated power, the monitoring system initiates the independence threshold calibration process. The monitoring system extracts the resistive and capacitive sequences within the initial integration time domain boundary and calculates the square root of the initial variance product. The monitoring system increments the integration boundary shift and calculates the absolute value of the covariance of the resistive and capacitive sequences after the shift. The monitoring system calculates the ratio sequence of the absolute value of the covariance to the square root of the initial variance product under each shift, identifies the noise floor convergence interval in the ratio sequence where the values ​​tend to fluctuate steadily, extracts the arithmetic mean of the ratios within the noise floor convergence interval, adds a 5% measurement margin to generate the independence threshold of the correlation coefficient, and the output independence threshold constitutes the initialization benchmark adapted to the current power grid topology.

[0039] Example 5: In the insulation aging simulation test scenario before equipment deployment, the monitoring system executes an offline optimization procedure for a preset rate of change threshold to control the false alarm and missed alarm phenomena of the early warning system under electromagnetic interference environment. The monitoring system quantifies the early warning accuracy as an objective function and configures it as a weighted arithmetic sum of the false alarm rate and the missed alarm rate. The monitoring system collects the total residual current signal and the bus voltage reference signal of the standard cross-linked polyethylene cable during the entire life cycle of the accelerated thermal aging test, and continuously outputs the decoupled resistive residual current historical data by using integral boundary translation and covariance optimization calculation.

[0040] The monitoring system uses the least squares method to process historical resistive residual current data, fits the time evolution curve of insulation degradation trend, and extracts the slope of change. The monitoring system sets the traversal search interval of the preset change rate threshold to 5% to 30%, and sets the search step size to 1%. At each traversal node, the monitoring system compares the extracted change slope with the current node threshold to generate test warning results. It compares the test warning results with the cable physical breakdown records to calculate the false alarm rate and missed alarm rate at the current node. Based on the calculated false alarm rate and missed alarm rate, it solves for the corresponding objective function value. The monitoring system extracts the threshold coordinates corresponding to the minimum point of the objective function in all traversal nodes, determines the threshold coordinates as the preset change rate threshold of 15%, and writes them into the system read-only memory. The output preset change rate threshold constitutes the quantitative benchmark for judging the health status of the insulation medium during long-term operation.

[0041] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1. A method for identifying and monitoring resistive-capacitive residual current in electrical equipment, characterized in that, include: Step 101: Synchronously acquire the bus voltage reference signal and the total residual current signal, and extract the fundamental component of the bus voltage reference signal to obtain the voltage zero crossing time; Step 102: Divide the time into four time-series orthogonal intervals based on the voltage zero-crossing time, and calculate the piecewise charge integral value of the total residual current signal in each time-series orthogonal interval to generate a primary integral characteristic quantity that characterizes the aliasing state of resistive and capacitive components. Step 103: Compare the piecewise charge integral values ​​at symmetrical positions within adjacent half-wave periods, calculate the half-wave difference components of the two, and determine the phase shift angle caused by the nonlinearity of the magnetic circuit of the induction unit based on the preset proportional mapping relationship between the half-wave difference components and the phase shift angle. Step 104: Based on the dynamic physical phase offset angle, the integral time domain boundary of the time-series orthogonal interval is compensated and shifted in real time. By adjusting the sampling start point and the phase span of the sampling window, the integral time domain boundary after the shift correction is kept absolutely orthogonal to the fundamental physical phase of the bus voltage reference signal. Step 105: Within each integral time domain boundary after translation correction, based on the intrinsic independence of capacitive and resistive components, iterative optimization is performed by minimizing the covariance between the total residual current signal and the capacitive projection component, thereby removing the projection interference of the capacitive component on the resistive axis, and calculating and outputting the decoupled resistive residual current data.

2. The method for identifying and monitoring resistive-capacitive residual current in electrical equipment according to claim 1, characterized in that, Step 104 further includes the following sub-steps: Step 1041, obtaining phase compensation weights characterizing changes in ambient temperature and nonlinear saturation features of the magnetic core; Step 1042, using the phase compensation weights to perform weighted calculations on the phase offset angles to generate a phase compensation time step; Step 1043, using the phase compensation time step to synchronously shift the sampling start time of each time-series orthogonal interval.

3. The method for identifying and monitoring resistive-capacitive residual current in electrical equipment according to claim 1, characterized in that, The process of removing component interference in step 105 includes the following sub-steps: Step 1051, extract the decoupled resistive sequence and capacitive sequence within a preset sliding window; Step 1052, calculate the correlation coefficient between the resistive sequence and the capacitive sequence, and determine whether the correlation coefficient reaches the preset independence threshold; Step 1053, if the correlation coefficient does not reach the independence threshold, adjust the translation step size of the integral boundary according to the positive or negative polarity of the correlation coefficient until the correlation coefficient converges to a minimum value.

4. The method for identifying and monitoring resistive-capacitive residual current in electrical equipment according to claim 1, characterized in that, Before step 102, the following steps are also included: Step 401, low-pass filtering is performed on the acquired total residual current signal to extract the fundamental residual current component; Step 402, amplitude normalization processing is performed on the fundamental residual current component to generate a unit reference vector for projection decoupling.

5. The method for identifying and monitoring resistive-capacitive residual current in electrical equipment according to claim 1, characterized in that, Step 105 is followed by the following steps: Step 601, record the resistive residual current data within the continuous monitoring period and fit the time evolution curve of the insulation degradation trend; Step 602, calculate the slope of the time evolution curve and generate an insulation status warning command when the slope exceeds a preset change rate threshold of 15%.

6. The method for identifying and monitoring resistive-capacitive residual current in electrical equipment according to claim 1, characterized in that, The specific steps for obtaining the zero-crossing time of the voltage in step 101 are as follows: the phase of the bus voltage reference signal is tracked using a digital phase-locked loop algorithm, and the sampling frequency is corrected in real time to keep the deviation between the sampling frequency and the fundamental frequency of the power grid within 0.01Hz.

7. The method for identifying and monitoring resistive-capacitive residual current in electrical equipment according to claim 1, characterized in that, In step 102, the four time-series orthogonal intervals correspond to the 0 to π / 2, π / 2 to π, π to 3π / 2 and 3π / 2 to 2π radian regions of the fundamental period of the bus voltage reference signal, respectively.

8. The method for identifying and monitoring resistive-capacitive residual current in electrical equipment according to claim 2, characterized in that, The process of obtaining the nonlinear saturation characteristics of the magnetic core in step 1041 includes the following sub-steps: Step 10411, real-time acquisition of magnetic flux change rate data on the secondary side of the current transformer; Step 10412, based on the magnetic flux change rate data, searching the preset magnetization curve mapping table, determining the phase lag correction amount corresponding to the current magnetic flux state, and including the phase lag correction amount in the phase compensation weight.

9. The method for identifying and monitoring resistive-capacitive residual current in electrical equipment according to claim 1, characterized in that, It also includes the following steps: Step 1001: Perform clustering calculations on the resistive residual current data within the historical period to establish a baseline distribution of insulation parameters under normal operating conditions of the electrical equipment; Step 1002: Calculate the deviation of the resistive residual current data obtained from real-time decoupling from the baseline distribution of insulation parameters, and output the health rating result of the electrical equipment based on the deviation.

Images