A method for detecting resistive current of a surge arrester based on adaptive noise cancellation

Abstract

The application discloses a lightning arrester resistance current detection method based on adaptive noise cancellation, which comprises the following steps: compensating the phase of target leakage current and target voltage based on a PID algorithm to obtain compensated leakage current and voltage; de-noising the compensated leakage current and reserving signal characteristics based on a wavelet de-noising method to obtain de-noised leakage current and signal fault characteristics thereof; extracting the de-noised leakage current based on an adaptive noise cancellation method (ANC) and the compensated voltage to obtain a resistance current component of the de-noised leakage current. The resistance component of the leakage current is obtained by canceling the capacitive current component of the de-noised leakage current through the adaptive noise cancellation algorithm (ANC), and the variable step LMS strategy is combined to dynamically adapt the equivalent capacitance parameter of the lightning arrester. The method can improve the anti-interference ability of the measurement system, can separate the resistance component in real time, has good dynamic characteristics, and is suitable for lightning arrester resistance current detection in complex environments.

Description

Technical Field

[0001] This invention relates to the field of resistive current detection of surge arresters, and more specifically, to a method for detecting resistive current of surge arresters based on adaptive noise cancellation. Background Technology

[0002] In modern power systems, zinc oxide surge arresters (MOAs) are key devices for overvoltage protection, and the reliability of their insulation performance directly affects the safe operation of the entire power grid. The core material of an MOA is a zinc oxide (ZnO) nonlinear resistive element, which exhibits a high resistivity state during normal operation, allowing only microampere-level leakage current to pass through. Under overvoltage impact, its resistance drops sharply, rapidly dissipating energy to protect power equipment. However, with the development of ultra-high voltage (UHV) transmission technology, the operating environment of surge arresters has become more demanding, subjecting them to higher operating voltages, stronger electromagnetic interference, and more complex climatic conditions (such as humidity and pollution). During long-term operation, the nonlinear characteristics of MOAs may deteriorate due to moisture, aging, or mechanical damage, causing the resistive component of the leakage current to gradually increase. This increase in resistive current causes the temperature of the resistive element to rise, accelerating aging and even triggering thermal collapse accidents, seriously threatening the stable operation of the power grid. Therefore, accurate detection of the resistive component of the MOA leakage current is an important means of assessing its insulation status, predicting its remaining life, and preventing faults.

[0003] However, resistive current typically accounts for only 1% to 10% of the total leakage current and is easily overwhelmed by capacitive components (accounting for over 90%) and power grid harmonic noise, making it difficult to acquire resistive current with high precision. The equivalent capacitance of surge arresters drifts (±20%) with temperature and aging, but existing fixed-parameter models cannot track this in real time, leading to long-term distortion of monitoring results. Common-mode noise in complex electromagnetic environments (such as ground potential difference and switching transients) couples to the measurement circuit through the sensor, while traditional hardware filtering schemes can only suppress interference in specific frequency bands, with limited dynamic noise suppression effectiveness. Therefore, a new resistive current detection method with high precision, strong anti-interference capability, and low computational complexity is urgently needed to provide reliable technical support for online monitoring and fault early warning of surge arrester insulation status.

[0004] In existing technologies, the full-current detection method proposes a monitoring method based on the change of the full current amplitude. This method alarms by setting a current threshold and has the advantage of simple implementation, and was widely used in early power grids. However, with the increasing requirements for power system monitoring, this method has revealed serious defects. This method cannot distinguish between resistive and capacitive components, is prone to false alarms when changes in ambient humidity cause an increase in capacitive current, has low sensitivity, and can only detect when insulation is severely deteriorated (resistive current increases by more than 50%). It is also greatly affected by grid voltage fluctuations, with measurement errors reaching ±15%. The third harmonic method proposes a method to indirectly calculate the resistive current (reflecting the insulation state) by analyzing the third harmonic component in the leakage current. This method does not require voltage phase synchronization. Traditional capacitive current compensation methods require accurate measurement of the phase difference between voltage and current, while the third harmonic method only requires a current signal, reducing the impact of synchronization errors. Moreover, it is sensitive to early aging of surge arresters, with the third harmonic significantly increasing even when the valve plate is slightly deteriorated. However, this method is susceptible to high-order harmonic interference. The 5th and 7th harmonics present in the power grid may overlap with the 3rd harmonic frequency band, causing the measured value to be larger than the actual value and resulting in false alarms. The fundamental phase difference method calculates the resistive current component by measuring the phase difference between the voltage and the fundamental leakage current and using orthogonal decomposition. This algorithm has low complexity and requires less processing power, making it suitable for resource-constrained embedded systems. Theoretically, it has high accuracy. However, in practical applications, due to harmonic pollution from the power grid voltage, sensor phase shift errors, and additional phase shifts from the signal conditioning circuit, this method has extremely high requirements for synchronous sampling and hardware calibration, leading to difficulties in field implementation and insufficient long-term stability. Summary of the Invention

[0005] This invention provides a surge arrester resistive current detection method based on adaptive noise cancellation. The method uses an adaptive noise cancellation algorithm (ANC) to cancel the capacitive current component of the denoised leakage current to obtain the resistive component of the leakage current. Combined with a variable step size LMS strategy, the method dynamically adapts to the equivalent capacitance parameters of the surge arrester. This method not only improves the anti-interference capability of the measurement system, but also can separate the resistive component in real time, exhibiting good dynamic characteristics and making it suitable for surge arrester resistive current detection in complex environments.

[0006] The technical solution adopted in this invention is:

[0007] A method for detecting resistive current in surge arresters based on adaptive noise cancellation, comprising:

[0008] S1. Based on the current sensor and the capacitive voltage divider, the leakage current and voltage of the surge arrester are collected and preprocessed to obtain the target leakage current and target voltage respectively;

[0009] S2. Based on the PID algorithm, phase compensation is performed on the target leakage current and target voltage to obtain the compensated leakage current and voltage.

[0010] S3. Based on the wavelet denoising method, the compensated leakage current is denoised and its signal features are preserved to obtain the denoised leakage current and its signal fault characteristics.

[0011] S4. Based on the adaptive noise cancellation method (ANC) and the compensated voltage, the denoised leakage current is extracted to obtain the resistive current component of the denoised leakage current.

[0012] Further, step S1 includes:

[0013] S11. Install the current sensor on the grounding wire of the surge arrester. Connect the common-mode choke coil in series at the output terminal of the current sensor. The current sensor outputs a differential signal. The common-mode choke coil suppresses the common-mode noise on the differential signal. Then, it is differentially amplified by the AD8421 instrumentation amplifier. Finally, it is filtered out by a second-order Butterworth low-pass filter with a cutoff frequency of 300Hz to further remove the high-frequency interference signal contained in the leakage current and obtain the target leakage current.

[0014] S12. Connect the capacitor voltage divider to the high-voltage terminal of the surge arrester. Connect the output terminal of the capacitor voltage divider to the common-mode choke coil. Obtain the voltage of the high-voltage terminal of the surge arrester through the capacitor voltage divider. Suppress common-mode interference through the common-mode choke coil. At the same time, form a high-frequency noise discharge path with the Y capacitor. Then, perform differential amplification through the INA188 instrumentation amplifier. Then, filter out high-frequency noise through a second-order Butterworth low-pass filter to obtain the target voltage.

[0015] Further, step S2 includes:

[0016] S21. Initialization phase: A standard sine wave test signal with an amplitude of 1Vpp and a frequency of 50Hz is injected through the DDS signal generator.

[0017] S22. Use a synchronous sampling method to measure the voltage and current channels, and obtain the phase difference between the target leakage current and the target voltage, which is recorded as the original phase difference;

[0018] S23. Based on the PID algorithm, the resistance of the digital potentiometer is dynamically adjusted to drive the all-pass filter to compensate for the original phase difference within a range of ±5°, thereby obtaining the compensated leakage current and voltage.

[0019] Further, step S3 includes:

[0020] S31. Using the Db4 wavelet basis in the wavelet denoising method, the compensated leakage current is decomposed into 5 levels, resulting in the following expression:

[0021]

[0022] In the formula, These are the approximation coefficients for the 5th layer, representing the low-frequency main signal, which includes power frequency components; These are the detail coefficients for layers 1 to 5, representing high-frequency noise and harmonics. The higher the layer number, the lower the frequency it represents. The leakage current after compensation by the surge arrester;

[0023] S32, Based on the first layer detail coefficients containing the most noise The noise standard deviation is calculated, and the noise threshold is determined using the following expression:

[0024]

[0025] In the formula, The standard deviation of noise; is the detail coefficient for the first layer; 0.6745 is the conversion factor between the absolute deviation of the median and the standard deviation under the Gaussian distribution;

[0026] S33. After determining the noise threshold, filter out the noise while retaining the compensated leakage current fault characteristics, and selectively retain or suppress wavelet coefficients.

[0027] 1) Calculate the threshold for each layer, as shown in the following expression:

[0028]

[0029] In the formula, The noise threshold for each layer; The length of the compensated leakage current signal; The wavelet decomposition levels are 1 to 5; The standard deviation of noise;

[0030] 2) Apply the improved soft thresholding function to the detail coefficients of each layer, as shown in the following expression:

[0031]

[0032] In the formula, The detail coefficient after noise reduction; These are the detail coefficients of each layer before noise reduction; The noise threshold for each layer; The standard deviation of noise;

[0033] 3) The signal is reconstructed through inverse transformation to obtain the denoised leakage current and its signal fault characteristics for subsequent processing;

[0034] The formula for the inverse transform is as follows:

[0035]

[0036] In the formula, The leakage current after reconstruction; These are the approximation coefficients for the 5th layer; These are the detail coefficients for layers 1 to 5 after denoising; This is the inverse transformation.

[0037] Further, step S4 includes:

[0038] S41. Based on the compensated voltage, the voltage derivative at the non-sampling boundary point is calculated using the fifth-order central difference method. The calculation formula is as follows:

[0039]

[0040] In the formula, This is the voltage derivative at the non-sampling boundary point at the current moment; The sampling interval; This is the voltage sample value; The signal on the time axis One sampling point;

[0041] S42. Based on the compensated voltage, the voltage derivative at the sampling boundary point is calculated using the third-order finite difference method. The calculation formula is as follows:

[0042]

[0043] In the formula, Voltage derivative at sampling boundary points; The sampling interval; ~ These are voltage sample values ​​at the boundary points;

[0044] S43. Based on the equivalent capacitance value at the current moment and the voltage derivative at the corresponding moment, the capacitive current component of the denoised leakage current is calculated. The calculation formula is as follows:

[0045]

[0046] In the formula, This represents the capacitive current component of the leakage current after noise reduction. This is the equivalent capacitance value at the current moment; This is the voltage derivative at the non-sampling boundary point at the current moment;

[0047] S44. Based on the adaptive noise cancellation method, a signal with the same magnitude but opposite phase to the calculated capacitive current component is used to cancel the capacitive current component of the denoised leakage current. The resistive current component of the denoised leakage current is then calculated from the capacitive current component of the denoised leakage current. The calculation formula is as follows:

[0048]

[0049] In the formula, This represents the resistive current component of the leakage current after noise reduction. The leakage current after noise reduction; This represents the capacitive current component of the leakage current after noise reduction.

[0050] Furthermore, the equivalent capacitance value is dynamically updated using the variable step size least mean square (LMS) algorithm, and the specific process is as follows:

[0051] 1) Obtaining temperature data of surge arrester: Install a temperature sensor near the leakage current sampling point of surge arrester, and after anti-electromagnetic interference processing, send it to ADC for sampling, and obtain smooth temperature data by averaging multiple sampling points.

[0052] 2) Obtain the initial equivalent capacitance value of the surge arrester. And define temperature reference values ​​based on temperature data. The temperature reference is the typical operating temperature of the surge arrester when it is in a healthy state and the environmental conditions are stable. To collect the average temperature over a period of time during the initial operation of the surge arrester;

[0053] 3) Perform parameter initialization settings: Set the algorithm step size parameter: baseline step size Maximum step size minimum step size ; Set adjustment coefficient: current amplitude weight Temperature coefficient ;

[0054] 4) Calculate the error signal at the current time:

[0055]

[0056] In the formula, This is an error signal; This is the instantaneous value of the total leakage current of the surge arrester after hardware filtering and phase compensation. This is the equivalent capacitance value; This is the voltage derivative at the non-sampling boundary point at the current moment; This is the voltage sample value;

[0057] 5) Based on the step size parameter and adjustment coefficient, the step size is dynamically adjusted to obtain the real-time changing step size. :

[0058]

[0059] In the formula, The step size changes in real time; Used as the reference step size; Weighted by current amplitude; This is an error signal; Temperature coefficient; For a certain moment Temperature sampling value; The temperature reference value is typically 25°C; the step size for real-time variation is... At the maximum step size with minimum step size The values ​​are taken between these ranges to avoid oscillations when the step size is too large or slow convergence when the step size is too small.

[0060] 6) The updated equivalent capacitance value is obtained by using the variable step size least mean square (LMS) algorithm. The calculation formula is as follows:

[0061]

[0062] In the formula, This is the updated equivalent capacitance value; The equivalent capacitance value before the update; The step size changes in real time; This is an error signal; This is the voltage sample value;

[0063] When the error signal Or voltage sample value As the temperature increases, the capacitance correction amount increases; Deviation from temperature reference value Real-time changing step size It will automatically shrink to suppress temperature drift, thereby achieving a more accurate estimate of the equivalent capacitance value.

[0064] Furthermore, it also includes: constraining the boundaries of the equivalent capacitance value to prevent the capacitance correction from diverging due to interference, with the following constraint conditions:

[0065]

[0066] In the formula, For the updated equivalent capacitance value; This is the initial equivalent capacitance value.

[0067] Compared with the prior art, the present invention has the following advantages:

[0068] 1) This invention uses a high-precision current sensor to convert current into a voltage signal and integrates a common-mode noise suppression circuit on the secondary side of the sensor to suppress ground loop interference; it also uses a capacitor voltage divider and a voltage follower to ensure that the voltage signal and the current signal are triggered synchronously with a phase deviation of less than 0.1 degrees; on this basis, a second-order Butterworth low-pass filter is introduced to suppress high-frequency noise and other high-frequency interference that are aliased onto the low-frequency signal during synchronous ADC sampling, thereby reducing the interference of high-frequency noise on the algorithm calculation and improving the accuracy of capacitive current calculation.

[0069] 2) The phase compensation circuit is used to compensate for the inherent phase shift of the sensor, the phase shift of the signal conditioning circuit, and the sampling time deviation of the ADC, so as to ensure that the phase difference between the sampling voltage and the sampling current only reflects the true state of the surge arrester.

[0070] 3) Before extracting the resistive component of the surge arrester leakage current, residual noise is further removed by wavelet denoising to suppress non-stationary noise while retaining the transient characteristics of the leakage current.

[0071] 4) By combining the adaptive noise cancellation algorithm (ANC) with the variable step size LMS strategy, the resistive and capacitive components of the leakage current in the voltage and current signals after noise reduction and phase compensation are separated in real time, and the equivalent capacitance parameters of the surge arrester are dynamically adapted. This method can not only improve the anti-interference capability of the measurement system, but also separate the resistive component in real time. It has good dynamic characteristics and is suitable for high-precision monitoring of the resistive current of surge arresters in complex environments. Attached Figure Description

[0072] The present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0073] Figure 1 This is a schematic flowchart of a surge arrester resistive current detection method based on adaptive noise cancellation according to the present invention.

[0074] Figure 2 This is a flowchart of the leakage current and voltage acquisition and preprocessing process of the surge arrester in this invention;

[0075] Figure 3 This is a flowchart of the phase compensation process in this invention;

[0076] Figure 4 This is a comparison image of the original signal before wavelet denoising processing in this invention;

[0077] Figure 5 This is a comparison diagram of the reconstructed signals obtained based on the wavelet denoising method in this invention;

[0078] Figure 6 This is a waveform diagram of the total leakage current, capacitive current component, and resistive current component in this invention. Detailed Implementation

[0079] Example

[0080] like Figure 1 As shown, a method for detecting resistive current in a surge arrester based on adaptive noise cancellation includes:

[0081] S1. Based on the current sensor and the capacitive voltage divider, the leakage current and voltage of the surge arrester are collected and preprocessed to obtain the target leakage current and target voltage respectively;

[0082] Specifically, such as Figure 2 As shown in Figure S11, the current sensor is mounted on the grounding wire of the surge arrester. A common-mode choke coil is connected in series on the output terminal of the current sensor. The current sensor outputs a differential signal. The common-mode choke coil suppresses the common-mode noise on the differential signal. Then, it is differentially amplified by an AD8421 instrumentation amplifier. Finally, it is filtered out by a second-order Butterworth low-pass filter with a cutoff frequency of 300Hz to further remove the high-frequency interference signal contained in the leakage current and obtain the target leakage current. The current sensor is a closed-loop Hall sensor.

[0083] S12. Connect the capacitor voltage divider to the high-voltage terminal of the surge arrester. Connect the output terminal of the capacitor voltage divider to the common-mode choke coil. Obtain the voltage of the high-voltage terminal of the surge arrester through the capacitor voltage divider. Suppress common-mode interference through the common-mode choke coil. At the same time, form a high-frequency noise discharge path with the Y capacitor. Then, perform differential amplification through the INA188 instrumentation amplifier. Then, filter out high-frequency noise through a second-order Butterworth low-pass filter to obtain the target voltage.

[0084] S2. Based on the PID algorithm, phase compensation is performed on the target leakage current and target voltage to obtain the compensated leakage current and voltage.

[0085] Specifically, in S21, the initialization phase, a standard sine wave test signal with an amplitude of 1Vpp and a frequency of 50Hz is injected through the DDS signal generator;

[0086] S22. Use a synchronous sampling method to measure the voltage and current channels, and obtain the phase difference between the target leakage current and the target voltage, which is recorded as the original phase difference;

[0087] S23. Based on the PID algorithm, the resistance of the digital potentiometer is dynamically adjusted to drive the all-pass filter to compensate for the original phase difference within a range of ±5°, thereby obtaining the compensated leakage current and voltage.

[0088] Phase compensation flowchart as follows Figure 3As shown, the core hardware platform is the MAX7420 configurable all-pass filter chip. This chip integrates a second-order active all-pass network and adopts an operational amplifier feedback topology to provide continuously adjustable phase delay while maintaining a constant signal amplitude and within the range of 0° to -180°. The phase adjustment is performed using an AD5171 high-precision digital potentiometer, which has a digitally programmable resistor with 256 tap positions. It receives control commands through a three-wire SPI interface and maintains stable adjustment performance even in harsh environments.

[0089] Note: Due to inherent phase shifts in the target leakage current and target voltage at the current sensor and capacitive voltage divider, phase shifts generated when the signal passes through the signal conditioning circuit, and time deviations caused by ADC sampling, it is necessary to achieve accurate phase compensation through hardware circuitry and digital control in order to ensure that the phase difference between the target leakage current and target voltage can accurately reflect the true state of the surge arrester.

[0090] S3. Based on the wavelet denoising method, the compensated leakage current is denoised and its signal features are preserved to obtain the denoised leakage current and its signal fault characteristics.

[0091] Specifically, in S31, the Db4 wavelet basis in the wavelet denoising method is used to perform a 5-level decomposition on the compensated leakage current, resulting in the following expression:

[0092]

[0093] In the formula, These are the approximation coefficients for the 5th layer, representing the low-frequency main signal, which includes power frequency components; These are the detail coefficients for layers 1 to 5, representing high-frequency noise and harmonics. The higher the layer number, the lower the frequency it represents. The leakage current after compensation by the surge arrester;

[0094] S32, Based on the first layer detail coefficients containing the most noise The noise standard deviation is calculated, and the noise threshold is determined using the following expression:

[0095]

[0096] In the formula, The standard deviation of noise; is the detail coefficient for the first layer; 0.6745 is the conversion factor between the absolute deviation of the median and the standard deviation under the Gaussian distribution;

[0097] S33. After determining the noise threshold, filter out the noise while retaining the compensated leakage current fault characteristics, and selectively retain or suppress wavelet coefficients.

[0098] 1) Calculate the threshold for each layer, as shown in the following expression:

[0099]

[0100] In the formula, The noise threshold for each layer; The length of the compensated leakage current signal; The wavelet decomposition levels are 1 to 5; The standard deviation of noise;

[0101] 2) Apply the improved soft thresholding function to the detail coefficients of each layer, as shown in the following expression:

[0102]

[0103] In the formula, The detail coefficient after noise reduction; These are the detail coefficients of each layer before noise reduction; The noise threshold for each layer; The standard deviation of noise;

[0104] 3) The signal is reconstructed through inverse transformation to obtain the denoised leakage current and its signal fault characteristics for subsequent processing;

[0105] The formula for the inverse transform is as follows:

[0106]

[0107] In the formula, The leakage current after reconstruction; These are the approximation coefficients for the 5th layer; These are the detail coefficients for layers 1 to 5 after denoising; This is an inverse transformation;

[0108] Note: Hardware filtering often struggles to suppress non-stationary noise such as pulse interference. To preserve the transient characteristics of the compensated leakage current, wavelet denoising is used to further remove residual noise before extracting its resistive current component, while retaining the fault characteristics of the leakage current signal. Since the tight support and regularity of Db4 are suitable for capturing the transient characteristics of the compensated leakage current and have a high waveform matching degree with the power frequency signal, the Db4 wavelet basis in wavelet denoising is selected to perform multi-level decomposition of the compensated leakage current. To determine the noise threshold and avoid signal distortion due to excessive denoising, the noise standard deviation is calculated for the first-level detail coefficients containing the most noise. The noise follows a Gaussian distribution, and the median absolute deviation (MAD) is insensitive to pulse interference in the signal. This method is more efficient than directly calculating the standard deviation.

[0109] To verify the ability of wavelet denoising to remove noise from the compensated leakage current while preserving its signal fault characteristics, the signal waveforms before and after applying wavelet denoising were compared. Specific results are as follows: Figure 4 and Figure 5 As shown. By Figure 5 It can be seen that after the sampling signal is processed by wavelet denoising, the noise of the compensated leakage current is further filtered out, while the signal fault characteristics of the compensated leakage current are preserved.

[0110] S4. Based on the adaptive noise cancellation method (ANC) and the compensated voltage, the denoised leakage current is extracted to obtain the resistive current component of the denoised leakage current.

[0111] Specifically, S41, based on the compensated voltage, the voltage derivative at the non-sampling boundary point is calculated using the fifth-order central difference method, and the calculation formula is as follows:

[0112]

[0113] In the formula, This is the voltage derivative at the non-sampling boundary point at the current moment; The sampling interval; This is the voltage sample value; The signal on the time axis One sampling point;

[0114] S42. Based on the compensated voltage, the voltage derivative at the sampling boundary point is calculated using the third-order finite difference method. The calculation formula is as follows:

[0115]

[0116] In the formula, The voltage derivative at the sampling boundary point; The sampling interval; ~ These are voltage sample values ​​at the boundary points;

[0117] S43. Based on the equivalent capacitance value at the current moment and the voltage derivative at the corresponding moment, the capacitive current component of the denoised leakage current is calculated. The calculation formula is as follows:

[0118]

[0119] In the formula, This represents the capacitive current component of the leakage current after noise reduction. This is the equivalent capacitance value at the current moment; This is the voltage derivative at the non-sampling boundary point at the current moment;

[0120] S44. Based on the adaptive noise cancellation method, a signal with the same magnitude but opposite phase to the calculated capacitive current component is used to cancel the capacitive current component of the denoised leakage current. The resistive current component of the denoised leakage current is then calculated from the capacitive current component of the denoised leakage current. The calculation formula is as follows:

[0121]

[0122] In the formula, This represents the resistive current component of the leakage current after noise reduction. The leakage current after noise reduction; This represents the capacitive current component of the leakage current after noise reduction.

[0123] like Figure 6 The figure shows the total leakage current of the surge arrester and the capacitive and resistive current components extracted using the Adaptive Noise Cancellation (ANC) algorithm; this demonstrates that the ANC algorithm can effectively extract the resistive current component from the leakage current.

[0124] Explanation: The capacitive current component of a surge arrester is proportional to the voltage derivative. To improve the accuracy of the differential calculation of the capacitive current component, a fifth-order central difference method is used at non-sampling boundary points, and a third-order difference method is used at sampling boundary points to calculate the voltage derivative. Because the voltage signal cannot obtain a complete data window at sampling boundary points, such as during the power-on startup phase of the detection system or after sampling interruption and reconnection, there is insufficient data for fifth-order central difference. If the missing data is simply set to zero, a sudden change in the calculated value will be introduced at the boundary point, increasing the derivative error and resulting in inaccurate capacitive current calculations. Third-order difference, on the other hand, does not require historical data and can achieve a smaller calculation error at the boundary compared to fifth-order central difference. Therefore, third-order difference processing is used at sampling boundary points. Based on the capacitance characteristics of the surge arrester, the capacitive current component is obtained by multiplying the current equivalent capacitance value by the calculated voltage derivative. Then, an adaptive noise cancellation algorithm is used to cancel the capacitive current component, leaving only the resistive current component in the surge arrester leakage current, thereby achieving effective extraction and detection of the resistive current component in the leakage current.

[0125] As a further embodiment, the equivalent capacitance value is dynamically updated using a variable step size least mean square (LMS) algorithm, the specific process of which is as follows:

[0126] 1) Obtaining temperature data of surge arrester: Install a temperature sensor near the leakage current sampling point of surge arrester, and after anti-electromagnetic interference processing, send it to ADC for sampling, and obtain smooth temperature data by averaging multiple sampling points.

[0127] 2) Obtain the initial equivalent capacitance value of the surge arrester. And define temperature reference values ​​based on temperature data. The temperature reference is the typical operating temperature of the surge arrester when it is in a healthy state and the environmental conditions are stable. To collect the average temperature over a period of time during the initial operation of the surge arrester;

[0128] 3) Perform parameter initialization settings: Set the algorithm step size parameter: baseline step size Maximum step size minimum step size ; Set adjustment coefficient: current amplitude weight Temperature coefficient ;

[0129] 4) Calculate the error signal at the current time:

[0130]

[0131] In the formula, This is an error signal; This is the instantaneous value of the total leakage current of the surge arrester after hardware filtering and phase compensation. This is the equivalent capacitance value; This is the voltage derivative at the non-sampling boundary point at the current moment; This is the voltage sample value;

[0132] 5) Based on the step size parameter and adjustment coefficient, the step size is dynamically adjusted to obtain the real-time changing step size. :

[0133]

[0134] In the formula, The step size changes in real time; Used as the reference step size; Weighted by current amplitude; This is an error signal; Temperature coefficient; For a certain moment Temperature sampling value; The temperature reference value is typically 25°C; the step size for real-time variation is... At the maximum step size with minimum step size The values ​​are taken between these ranges to avoid oscillations when the step size is too large or slow convergence when the step size is too small.

[0135] 6) The updated equivalent capacitance value is obtained by using the variable step size least mean square (LMS) algorithm. The calculation formula is as follows:

[0136]

[0137] In the formula, This is the updated equivalent capacitance value; The equivalent capacitance value before the update; The step size changes in real time; This is an error signal; This is the voltage sample value;

[0138] Note: When the error signal Or voltage sample value As the temperature increases, the capacitance correction amount increases; Deviation from temperature reference value Real-time changing step size It will automatically shrink to suppress temperature drift, thereby achieving a more accurate estimate of the equivalent capacitance value.

[0139] As a further aspect of this embodiment, it also includes: constraining the boundary of the equivalent capacitance value to prevent the capacitance correction from diverging due to interference, wherein the constraint conditions are as follows:

[0140]

[0141] In the formula, For the updated equivalent capacitance value; This is the initial equivalent capacitance value.

[0142] Working principle of the invention:

[0143] 1) By integrating common-mode chokes on the secondary side of the current and voltage sensors, common-mode interference introduced by the grounding loop is effectively eliminated; high-frequency noise is suppressed by a second-order Butterworth filter with a front-end cutoff frequency of 300Hz.

[0144] 2) Employing the MAX7420 configurable all-pass filter chip and AD5171 high-precision digital potentiometer, during system initialization, a standard sine wave test signal is injected to measure the original phase difference of the voltage and current channels. A PID algorithm is used to adjust the resistance of the digital potentiometer, achieving precise compensation for the group delay introduced by the sensor and signal chain. The signal is decomposed into five levels using the Db4 wavelet basis, and the noise threshold is calculated based on the median absolute deviation (MAD). An improved soft thresholding function is used to selectively retain or suppress wavelet coefficients, and finally, the denoised signal is reconstructed. The combination of hardware filtering and wavelet denoising suppresses steady-state noise while preserving fault transient characteristics, improving the signal-to-noise ratio. The introduction of wavelet denoising solves the problem of insufficient suppression of non-stationary noise by traditional hardware filtering.

[0145] 4) The Adaptive Noise Cancellation (ANC) algorithm is used to separate the resistive and capacitive components of the leakage current in real time; the voltage derivative of the compensated voltage is accurately calculated using the fifth-order central difference method, thereby predicting the capacitive current component; the equivalent capacitance parameters of the surge arrester are dynamically updated using the variable step-size LMS algorithm to adapt to capacitance drift caused by factors such as temperature changes and aging; the anti-interference capability is improved compared with the traditional hardware filtering scheme, and the problem of slow algorithm convergence or oscillation is avoided by dynamically adjusting the step size, ensuring the real-time performance and accuracy of the parameters.

[0146] This invention utilizes an adaptive noise cancellation algorithm (ANC) to automatically cancel the effects of power grid harmonics. When the 3rd / 5th harmonic content reaches 20%, the resistive current detection fluctuation is <10μA, improving the anti-interference performance by 10 times compared to a pure hardware filtering scheme. Combined with a variable step size minimum mean square (LMS) strategy, the equivalent capacitance parameters of the surge arrester are updated online. Even if the capacitance value drifts by ±20% due to temperature or aging, the resistive current calculation error remains <5%, solving the failure problem caused by capacitance changes in traditional fixed parameter methods. Compared to the 100-200ms response time typically required by traditional monitoring methods, the dynamic response speed is controlled within 50ms through optimized design, enabling more timely capture of rapid changes in the surge arrester's insulation state, such as instantaneous moisture absorption after a lightning strike.

[0147] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the principles and essence of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for detecting resistive current in a surge arrester based on adaptive noise cancellation, characterized in that, include: S1. Based on the current sensor and the capacitive voltage divider, the leakage current and voltage of the surge arrester are collected and preprocessed to obtain the target leakage current and target voltage respectively; S2. Based on the PID algorithm, phase compensation is performed on the target leakage current and target voltage to obtain the compensated leakage current and voltage. S3. Based on the wavelet denoising method, the compensated leakage current is denoised and its signal features are preserved to obtain the denoised leakage current and its signal fault characteristics. S4. Based on the adaptive noise cancellation method (ANC) and the compensated voltage, the denoised leakage current is extracted to obtain the resistive current component of the denoised leakage current. Step S4 includes: S41. Based on the compensated voltage, the voltage derivative at the non-sampling boundary point is calculated using the fifth-order central difference method. The calculation formula is as follows: In the formula, This is the voltage derivative at the non-sampling boundary point at the current moment; The sampling interval; This is the voltage sample value; The signal on the time axis One sampling point; S42. Based on the compensated voltage, the voltage derivative at the sampling boundary point is calculated using the third-order finite difference method. The calculation formula is as follows: In the formula, The voltage derivative at the sampling boundary point; The sampling interval; ~ These are voltage sample values ​​at the boundary points; S43. Based on the equivalent capacitance value at the current moment and the voltage derivative at the corresponding moment, the capacitive current component of the denoised leakage current is calculated. The calculation formula is as follows: In the formula, This represents the capacitive current component of the leakage current after noise reduction. This is the equivalent capacitance value at the current moment; This is the voltage derivative at the non-sampling boundary point at the current moment; S44. Based on the adaptive noise cancellation method, a signal with the same magnitude but opposite phase to the calculated capacitive current component is used to cancel the capacitive current component of the denoised leakage current. The resistive current component of the denoised leakage current is then calculated from the capacitive current component of the denoised leakage current. The calculation formula is as follows: In the formula, This represents the resistive current component of the leakage current after noise reduction. The leakage current after noise reduction; This represents the capacitive current component of the leakage current after noise reduction. The equivalent capacitance value is dynamically updated using the variable step size least mean square (LMS) algorithm, and the specific process is as follows: 1) Obtaining temperature data of surge arrester: Install a temperature sensor near the leakage current sampling point of surge arrester, and after anti-electromagnetic interference processing, send it to ADC for sampling, and obtain smooth temperature data by averaging multiple sampling points. 2) Obtain the initial equivalent capacitance value of the surge arrester. And define temperature reference values ​​based on temperature data. The temperature reference is the typical operating temperature of the surge arrester when it is in a healthy state and the environmental conditions are stable. To collect the average temperature over a period of time during the initial operation of the surge arrester; 3) Perform parameter initialization settings: Set the algorithm step size parameter: baseline step size Maximum step size minimum step size ; Set adjustment coefficient: current amplitude weight Temperature coefficient ; 4) Calculate the error signal at the current time: In the formula, This is an error signal; This is the instantaneous value of the total leakage current of the surge arrester after hardware filtering and phase compensation. This is the equivalent capacitance value; This is the voltage derivative at the non-sampling boundary point at the current moment; This is the voltage sample value; 5) Based on the step size parameter and adjustment coefficient, the step size is dynamically adjusted to obtain the real-time changing step size. : In the formula, The step size changes in real time; Used as the reference step size; Weighted by current amplitude; This is an error signal; Temperature coefficient; For a certain moment Temperature sampling value; The temperature reference value is typically 25°C; the step size for real-time variation is... At the maximum step size with minimum step size The values ​​are taken between these ranges to avoid oscillations when the step size is too large or slow convergence when the step size is too small. 6) The updated equivalent capacitance value is obtained by using the variable step size least mean square (LMS) algorithm. The calculation formula is as follows: In the formula, This is the updated equivalent capacitance value; The equivalent capacitance value before the update; The step size changes in real time; This is an error signal; This is the voltage sample value; When the error signal Or voltage sample value As the temperature increases, the capacitance correction amount increases; Deviation from temperature reference value Real-time changing step size It will automatically shrink to suppress temperature drift, thereby achieving a more accurate estimate of the equivalent capacitance value.

2. The surge arrester resistive current detection method based on adaptive noise cancellation according to claim 1, characterized in that, Step S1 includes: S11. Install the current sensor on the grounding wire of the surge arrester. Connect the common-mode choke coil in series at the output terminal of the current sensor. The current sensor outputs a differential signal. The common-mode choke coil suppresses the common-mode noise on the differential signal. Then, it is differentially amplified by the AD8421 instrumentation amplifier. Finally, it is filtered out by a second-order Butterworth low-pass filter with a cutoff frequency of 300Hz to further remove the high-frequency interference signal contained in the leakage current and obtain the target leakage current. S12. Connect the capacitor voltage divider to the high-voltage terminal of the surge arrester. Connect the output terminal of the capacitor voltage divider to the common-mode choke coil. Obtain the voltage of the high-voltage terminal of the surge arrester through the capacitor voltage divider. Suppress common-mode interference through the common-mode choke coil. At the same time, form a high-frequency noise discharge path with the Y capacitor. Then, perform differential amplification through the INA188 instrumentation amplifier. Then, filter out high-frequency noise through a second-order Butterworth low-pass filter to obtain the target voltage.

3. The surge arrester resistive current detection method based on adaptive noise cancellation according to claim 1, characterized in that, Step S2 includes: S21. Initialization phase: A standard sine wave test signal with an amplitude of 1Vpp and a frequency of 50Hz is injected through the DDS signal generator. S22. Use a synchronous sampling method to measure the voltage and current channels, and obtain the phase difference between the target leakage current and the target voltage, which is recorded as the original phase difference; S23. Based on the PID algorithm, the resistance of the digital potentiometer is dynamically adjusted to drive the all-pass filter to compensate for the original phase difference within a range of ±5°, thereby obtaining the compensated leakage current and voltage.

4. The surge arrester resistive current detection method based on adaptive noise cancellation according to claim 1, characterized in that, Step S3 includes: S31. Using the Db4 wavelet basis in the wavelet denoising method, the compensated leakage current is decomposed into 5 levels, resulting in the following expression: In the formula, These are the approximation coefficients for the 5th layer, representing the low-frequency main signal, which includes power frequency components; These are the detail coefficients for layers 1 to 5, representing high-frequency noise and harmonics. The higher the layer number, the lower the frequency it represents. The leakage current after compensation by the surge arrester; S32, Based on the first layer detail coefficients containing the most noise The noise standard deviation is calculated, and the noise threshold is determined using the following expression: In the formula, The standard deviation of noise; is the detail coefficient for the first layer; 0.6745 is the conversion factor between the absolute deviation of the median and the standard deviation under the Gaussian distribution; S33. After determining the noise threshold, filter out the noise while retaining the compensated leakage current fault characteristics, and selectively retain or suppress wavelet coefficients. 1) Calculate the threshold for each layer, as shown in the following expression: In the formula, The noise threshold for each layer; The length of the compensated leakage current signal; The wavelet decomposition levels are 1 to 5; The standard deviation of noise; 2) Apply the improved soft thresholding function to the detail coefficients of each layer, as shown in the following expression: In the formula, The detail coefficient after noise reduction; These are the detail coefficients of each layer before noise reduction; The noise threshold for each layer; The standard deviation of noise; 3) The signal is reconstructed through inverse transformation to obtain the denoised leakage current and its signal fault characteristics for subsequent processing; The formula for the inverse transform is as follows: In the formula, The leakage current after reconstruction; These are the approximation coefficients for the 5th layer; These are the detail coefficients for layers 1 to 5 after denoising; This is the inverse transformation.

5. The surge arrester resistive current detection method based on adaptive noise cancellation according to claim 1, characterized in that, It also includes: constraining the boundaries of the equivalent capacitance value to prevent the capacitance correction from diverging due to interference, with the following constraint conditions: In the formula, For the updated equivalent capacitance value; This is the initial equivalent capacitance value.

Images