Space charge measurement method and device based on cable terminal

Abstract

The invention relates to the technical field of charge measurement, and discloses a space charge measurement method and device based on a cable terminal, and the method comprises the following steps: building a space charge test platform based on an electro-acoustic pulse method; accurately matching the charge test time with the alternating current phase; applying a high-voltage pulse to the cable terminal sample by using the high-voltage pulse power supply so as to excite space charges in the cable terminal sample to vibrate and generate sound waves; capturing sound waves by using the portable space charge sensor, and converting the sound waves into electric signals; after amplification processing, a processor of the data acquisition and processing system executes optimization to obtain and output a charge waveform for eliminating environmental interference; and after wavelet threshold noise reduction processing and space charge positioning algorithm calculation are carried out, measurement of the space charge of the cable terminal is realized. According to the invention, through hardware structure optimization and algorithm innovation, on-site accurate measurement of the space charge of the cable terminal is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of charge measurement, more particularly, it relates to a method and device for measuring space charge based on cable terminal. BACKGROUND

[0002] With the modernization and economic development, the demand for electric energy in various industries continues to grow. Due to the advantages of not affecting the city appearance, less environmental interference, high reliability of power supply, and less land occupation, power cables have become the core equipment for high-power power transmission and distribution in urban power distribution networks, industrial parks and other scenarios. Among them, the cable terminal, as the connection node of the cable and other power equipment, is the weak link of the cable system insulation. During the long-term operation, space charges are easily accumulated in the insulation layer due to factors such as electric field distortion and local heating.

[0003] The continuous accumulation of space charges not only distorts the electric field strength inside the insulation, accelerates the aging of the insulation material, and in severe cases, can also cause insulation breakdown, leading to power outage accidents and threatening the safe and stable operation of the power system. Therefore, accurate measurement of space charges has become a key means to assess the insulation state of the cable terminal and predict potential faults. Currently, the PEA method (Piezoelectric Effect Analysis) is the mainstream technology for measuring space charge distribution in the laboratory. It stimulates the vibration of charges by high-voltage pulses to generate sound waves, and then captures the signals through sensors to realize the inversion of charge distribution. However, this technology still has significant limitations in field applications.

[0004] The existing space charge measurement technology and equipment based on the PEA method are too large in size, making it difficult to meet the demand for accurate measurement of cable terminals in the field. In addition, during the measurement process, there are many factors that interfere with the signal, and the processing capacity is insufficient. The signal is easily affected by electromagnetic noise, temperature and humidity fluctuations in the environment, power voltage fluctuations, and sensor adhesion deviations. However, the existing technology lacks targeted signal optimization methods, resulting in poor noise reduction effect and environmental interference that cannot be offset, further reducing the measurement accuracy. SUMMARY

[0005] To solve the above technical problems, the present application provides a method and device for measuring space charge based on cable terminal, which realizes the accurate measurement of space charge of cable terminal in the field through hardware structure optimization and algorithm innovation.

[0006] To achieve the above purpose, the technical scheme of the present application is as follows:

[0007] The method for measuring space charge based on cable terminal comprises the following steps:

[0008] S1, a space charge test platform is built based on the piezoelectric effect analysis method, which includes a high-voltage AC power supply, a high-voltage pulse power supply, an AC matching circuit, a portable space charge sensor, an amplifier, a data acquisition and processing system, and a resistance divider.

[0009] S2, precisely match the charge test time with the AC phase through the AC matching circuit;

[0010] S3, use the high-voltage pulse power supply to apply a high-voltage pulse to the cable terminal sample to excite the space charge inside the cable terminal sample to vibrate and generate acoustic waves;

[0011] S4, use the portable space charge sensor to capture the acoustic waves generated in step S3 and convert the acoustic wave signals into electrical signals;

[0012] S5, transmit the electrical signals obtained in step S4 to the amplifier for amplification processing, and then use the processor of the data acquisition and processing system to perform optimization to obtain an output charge waveform that eliminates environmental interference ;

[0013] S6, the data acquisition and processing system performs wavelet threshold denoising processing and space charge positioning algorithm calculation on the After wavelet threshold denoising processing and space charge positioning algorithm calculation, the space charge distribution curve inside the cable terminal sample is generated and displayed on the computer, and the measurement of the space charge of the cable terminal is realized.

[0014] As a preferred scheme of the present application, the electrical connection relationship of the hardware of the space charge test platform in S1 is as follows: the high-voltage AC power supply provides an AC electric field environment for the sample to be tested through the AC matching circuit; the AC matching circuit performs electrical signal matching to realize the synchronization of the test timing and the AC phase; the high-voltage pulse power supply applies a high-voltage pulse to the sample to be tested through the capacitor C to excite the space charge to vibrate; the space charge sensor, the amplifier, and the data acquisition and processing system are electrically connected in sequence to form a signal path of acoustic-electric conversion-signal amplification-data processing.

[0015] As a preferred scheme of the present application, in S1, the following signal transmission relationship exists:

[0016] The high-voltage pulse power supply outputs a high-voltage pulse electrical signal to excite the space charge inside the sample to vibrate, so that the charge vibration generates an acoustic wave signal; the space charge sensor receives the acoustic wave signal and converts it into a weak electrical signal, which is transmitted to the amplifier; the amplifier amplifies and enhances the weak electrical signal, and then transmits the strong electrical signal to the data acquisition and processing system; the data acquisition and processing system collects the amplified electrical signal, analyzes and operates it, generates a space charge distribution curve, and displays the result on the computer to realize the visualization of the charge distribution.

[0017] As a preferred scheme of the present application, the portable space charge sensor in S1 comprises a ring-shaped buckle and a detection piece, the detection piece comprises three parts of a measuring electrode, a piezoelectric film and a sensor, the measuring electrode is in the uppermost layer, the sensor is in the lowermost layer, the piezoelectric film is sandwiched between the two, the area of the piezoelectric film is the same as that of the inner aluminum plate of the sensor and only contacts with the inner aluminum plate; the measuring electrode and the sensor are fixed by screws, the contact surface of the measuring electrode is in an arc structure, and the sensor support structure is composed of four layers of materials nested together, the inner layer is polymethyl methacrylate material, the second layer from the inside to the outside is aluminum plate material, the third layer is polytetrafluoroethylene, and the outermost layer is aluminum plate material.

[0018] As a preferred scheme of the present application, the specific optimization process in S5 comprises:

[0019] S51, the processor runs a variable step size LMS noise suppression algorithm: define the input vector , wherein is the filter order, is the output of the PVDF sensor, and the pure charge waveform calibrated in the laboratory is taken as the expected output , the output , the error is calculated; the weight vector is updated according to (wherein is the initial step size, is the error signal variance, and is the input signal variance), and the iteration is performed until , so as to obtain the filtered signal ;

[0020] S52, the processor runs a fit degree adaptive compensation algorithm: the amplitude of is calculated, the deviation of the amplitude from the calibrated amplitude is calculated, the fit degree coefficient is fitted, and the signal is compensated to , so as to correct the influence of the loose fit of the PVDF sensor and the stress cone;

[0021] S53, the processor runs a multivariate LSTM environment compensation algorithm: an input matrix is constructed (wherein is a humidity parameter, is a temperature parameter, and is a power fluctuation parameter), a pre-trained LSTM model (trained with historical environment-charge data, with 64-256 hidden layer units) is called, and the cell state is updated according to (wherein is the cell state, In hidden state, These are the activation values ​​for the forget gate, input gate, and output gate, respectively. This is the weight matrix. For bias terms, (Hadamard product), outputting a charge waveform that eliminates environmental interference. .

[0022] As a preferred embodiment of the present invention, the filter order of the variable step-size LMS noise suppression algorithm described in S51 is... Dynamically adjust based on the signal bandwidth of the PVDF sensor: when the signal frequency is higher than 10kHz, take... When the signal frequency is below 10kHz, take ;

[0023] The calibration amplitude described in S52 The signal amplitude was obtained by performing multiple measurements on a standard sample with a known charge density under standard testing conditions (temperature 25℃±2℃, humidity 50%±5%), and taking the average value as the result. ;

[0024] The training process of the LSTM model described in S53 includes: collecting charge signal samples under different environmental parameters (temperature -10℃~60℃, humidity 20%~90%, power fluctuation ±5%), dividing the training set and validation set in a 7:3 ratio, iteratively training with the Adam optimizer (learning rate 0.001) until the validation set loss function is less than 1e-5, and saving the model parameters.

[0025] As a preferred embodiment of the present invention, the wavelet threshold noise reduction processing in S6 is used to eliminate signal interference caused by the miniaturization design of portable space charge sensors. The specific algorithm is as follows:

[0026] Define the original electrical signal as ,in For real signals, This is a noise signal. The standard deviation of noise. It is a time variable;

[0027] Perform N-level wavelet decomposition on the signal to obtain wavelet coefficients. ( The number of decomposition layers, (for coefficient index)

[0028] Adaptive threshold function is used to process wavelet coefficients:

[0029]

[0030] in Interlayer threshold, Let j be the standard deviation of the noise at layer j. The signal length;

[0031] Processed wavelet coefficients Reconstruct the signal to obtain the denoised signal. .

[0032] As a preferred embodiment of the present invention, the space charge localization algorithm described in S6 combines the arc-shaped structural features of the portable space charge sensor to calculate the charge position through the sound wave propagation time, as shown in the following formula:

[0033] Define the coordinates of the space charge position as (cylindrical coordinate system) Radial distance, For circumferential angle, (where is the axial distance), and the central angle of the arc-shaped detection element of the sensor is... The radius of curvature is ;

[0034] The speed of sound wave propagation in the cable termination medium is ,in For the Young's modulus of the medium, The density of the medium;

[0035] The time difference matrix of the sound waves generated by the charge reaching the sensor detection point is:

[0036]

[0037] in The circumferential angle corresponds to the nth sampling point of the mth detection unit of the sensor. This is the time difference of sound wave propagation at that point;

[0038] By solving the inverse problem of the time difference matrix using the least squares method, the three-dimensional coordinate distribution of space charge is obtained:

[0039]

[0040] in This refers to the number of sensor detection units. The number of sampling points for each unit, This is for the theoretical calculation of the time difference.

[0041] As a preferred embodiment of the present invention, the noise standard deviation described in S6 The calibration is obtained through calibration using the no-load signal of a portable space charge sensor, and the calibration formula is as follows: , where median is the median function, used to eliminate the influence of sensor inherent noise on threshold calculation.

[0042] A cable terminal space charge measurement device includes a high-voltage AC power supply, a high-voltage pulse power supply, an AC matching circuit, a portable space charge sensor, an amplifier, a data acquisition and processing system, and a resistive voltage divider. The input terminal of the resistive voltage divider is electrically connected to the output terminal of the high-voltage AC power supply, and its output terminal is electrically connected to the voltage monitoring input terminal of the data acquisition and processing system. The output terminal of the high-voltage AC power supply is also electrically connected to one end of the cable terminal sample, and the output terminal of the high-voltage pulse power supply is electrically connected to the other end of the cable terminal sample. The AC matching circuit consists of a capacitor C and a resistor R connected in series. Its input terminal is connected to the signal output terminal of the high-voltage AC power supply, and its output terminal is connected to the phase signal input terminal of the data acquisition and processing system. The portable space charge sensor is connected to the input terminal of the amplifier, and the output terminal of the amplifier is connected to the core signal input terminal of the data acquisition and processing system. The processor of the data acquisition and processing system is configured to execute a variable-step-size LMS noise suppression algorithm, a fit-adaptive compensation algorithm, a multivariable LSTM environmental compensation algorithm, wavelet threshold noise reduction processing, and a space charge localization algorithm to generate and display the internal space charge distribution curve of the cable terminal.

[0043] The beneficial technical effects of this invention are:

[0044] By designing a portable space charge sensor that includes a ring-shaped buckle and an arc-shaped detection element, the arc-shaped concave structure of the detection element is perfectly matched with the structure and size of the stress cone shell of the cable terminal, which can achieve a tight fit between the sensor and the terminal surface, replacing the single-point contact of the traditional flat electrode, ensuring complete capture of the acoustic signal, and significantly reducing measurement errors caused by structural mismatch.

[0045] The sensor features a lightweight design, making it easy to transport to cable operation sites. This solves the problem of laboratory equipment not being able to be deployed on-site and meets the measurement needs of complex scenarios such as underground cable wells and outdoor substations.

[0046] By employing a variable step size LMS noise suppression algorithm and wavelet threshold denoising, electromagnetic noise and inherent sensor noise are effectively eliminated. A multivariable LSTM environmental compensation algorithm is used to offset the interference of temperature, humidity, and power fluctuations on the signal in real time. An adaptive fit compensation algorithm is used to correct the effect of fit deviation between the sensor and the terminal, thereby improving signal quality in multiple dimensions and ensuring the accuracy of measurement results.

[0047] An innovative design incorporates a spatial charge localization algorithm based on the arc-shaped structure of a sensor. It establishes a sound wave propagation time difference matrix in cylindrical coordinates and solves the three-dimensional coordinates of the charge using the least squares method. This algorithm adapts to the arc-shaped medium morphology of cable terminals, solves the compatibility problem of traditional planar localization algorithms, and accurately inverts the radial, circumferential, and axial distribution of spatial charges inside the terminal, providing precise charge location information for insulation condition assessment.

[0048] This method enables accurate on-site measurement of space charge at cable terminals. The generated space charge distribution curve can intuitively reflect the charge accumulation state inside the terminal insulation, providing a reliable basis for predicting insulation aging and troubleshooting potential faults. It effectively reduces the risk of power outages caused by cable terminal breakdown and ensures the safe and stable operation of the power system. Attached Figure Description

[0049] Figure 1 This is a flowchart of the measurement method of the present invention;

[0050] Figure 2 This is a schematic diagram of the measurement system of the present invention;

[0051] Figure 3 This is a schematic diagram of the structure of the portable space charge sensor in this invention;

[0052] Figure 4 This is a schematic diagram of the waveform output of the AC matching circuit in this invention. Detailed Implementation

[0053] In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0054] Combination Figure 1 - Figure 4 The present invention provides the following embodiments:

[0055] The method for measuring space charge at cable terminals includes the following steps:

[0056] S1. A space charge testing platform is built based on the electroacoustic pulse method. The space charge testing platform includes a high-voltage AC power supply, a high-voltage pulse power supply, an AC matching circuit, a portable space charge sensor, an amplifier, a data acquisition and processing system, and a resistor divider.

[0057] S2. The charge test time is precisely matched with the AC phase through the AC matching circuit.

[0058] S3. Using the cable terminal as the test sample, apply a high-voltage pulse to the cable terminal sample using the high-voltage pulse power supply to excite the space charge inside the cable terminal sample to vibrate and generate sound waves.

[0059] S4. Use the portable space charge sensor to capture the sound waves generated in step S3, and convert the sound wave signal into an electrical signal;

[0060] S5. The electrical signal obtained in step S4 is transmitted to the amplifier for amplification. Then, the processor of the data acquisition and processing system performs optimization to obtain an output charge waveform that eliminates environmental interference. ;

[0061] S6. The data acquisition and processing system performs the same steps as in step S5. After performing wavelet threshold denoising and space charge localization algorithm calculations in sequence, the space charge distribution curve inside the cable terminal sample is generated and displayed on the computer, thereby realizing the measurement of the space charge of the cable terminal.

[0062] Furthermore, the electrical connections of the various hardware components of the space charge testing platform described in S1 are as follows: a high-voltage AC power supply provides an AC electric field environment for the test sample via an AC matching circuit; the AC matching circuit performs the electrical signal matching function, synchronizing the test timing with the AC phase; a high-voltage pulse power supply applies a high-voltage pulse to the test sample through capacitor C, exciting the space charge vibration; the space charge sensor, amplifier, and data acquisition and processing system are sequentially electrically connected, forming a signal path of acoustic-electric conversion-signal amplification-data processing. This complete hardware path ensures synchronized test timing and AC phase, guarantees efficient excitation of charge vibration by the high-voltage pulse, and avoids attenuation of weak signals after acoustic-electric conversion, providing a stable original signal for subsequent processing and laying the foundation for subsequent measurements.

[0063] Furthermore, in S1, the following signal transmission relationships also exist:

[0064] A high-voltage pulse power supply outputs a high-voltage pulse electrical signal, which excites the vibration of space charges within the test sample, causing the charge vibration to generate an acoustic signal. A space charge sensor receives the acoustic signal, converts it into a weak electrical signal, and transmits it to an amplifier. The amplifier amplifies and enhances the weak electrical signal before transmitting the strong electrical signal to a data acquisition and processing system. The data acquisition and processing system acquires the amplified electrical signal, analyzes and processes it, generates a space charge distribution curve, and displays the results on a computer, thus visualizing the charge distribution. This process completes the conversion from charge vibration to acoustic wave to electrical signal to data to image, amplifies the weak electrical signal to meet sampling requirements, and ultimately visualizes the charge distribution, facilitating intuitive acquisition of insulation status information.

[0065] Furthermore, the portable space charge sensor described in S1 includes a ring-shaped clip and a detection element. The detection element comprises three parts: a measuring electrode, a piezoelectric film, and a sensor. The measuring electrode is on the top layer, the sensor is on the bottom layer, and the piezoelectric film is sandwiched between them. The area of ​​the piezoelectric film is the same as the area of ​​the inner aluminum plate of the sensor and it only contacts the inner aluminum plate. The measuring electrode and the sensor are fixed together by screws. The contact surface of the measuring electrode has an arc-shaped structure. The sensor support structure is composed of four nested layers of materials: the inner layer is polymethyl methacrylate, the second layer from the inside out is an aluminum plate, the third layer is polytetrafluoroethylene, and the outermost layer is an aluminum plate. The arc-shaped contact surface adapts to the terminal shape and reduces acoustic signal loss; the multi-layer support isolates electromagnetic interference and ensures signal purity; the lightweight design combined with the ring-shaped clip enables convenient on-site deployment.

[0066] Furthermore, the specific optimization process in S5 includes:

[0067] S51. Processor runs variable step size LMS noise suppression algorithm: Define input vector ,in Let the filter order be . The output electrical signal is from the PVDF sensor; the desired output is a pure charge waveform calibrated in the laboratory. Calculate the output ,error ;according to (in The initial step size, The variance of the error signal. Update the weight vector (for the variance of the input signal) Iterate to The filtered signal is obtained. ;

[0068] S52, Processor runs adaptive compensation algorithm for fit: Calculation amplitude With the calibration amplitude deviation Fit coefficient Signal compensation To correct the effect of poor adhesion between the PVDF sensor and the stress cone;

[0069] S53. Processor runs multivariable LSTM environment compensation algorithm: Constructing input matrix (in For humidity parameters, For temperature parameters, (For power fluctuation parameters), call the pre-trained LSTM model (trained with historical environment-charge data, hidden layer unit number 64~256), and press... , Update the cell state (where In cellular state, In hidden state, These are the activation values ​​for the forget gate, input gate, and output gate, respectively. This is the weight matrix. For bias terms, (Hadamard product), outputting a charge waveform that eliminates environmental interference. .

[0070] This optimization process achieves multi-dimensional improvement in signal quality through layered interference suppression and precise deviation correction. Specifically, the S51's variable step-size LMS algorithm is based on adaptive filtering principles, dynamically adjusting the step size by calculating the variance between the error signal and the input signal in real time. The step size decreases as the error decreases, avoiding convergence oscillations or slow convergence caused by a fixed step size. It can effectively filter out electromagnetic noise in the sensor output signal, such as on-site power frequency interference and equipment electromagnetic radiation, while retaining the true charge signal components.

[0071] The S52's fit compensation algorithm is based on the mechanism that amplitude deviation is positively correlated with fit tightness. By comparing the amplitude of the filtered signal with the calibrated amplitude under standard conditions, it calculates the fit deviation and fits the compensation coefficient to correct the signal amplitude, thereby offsetting the signal attenuation caused by the poor fit between the sensor and the terminal and restoring the amplitude characteristics of the real charge signal.

[0072] S53's multivariate LSTM algorithm utilizes the temporal feature extraction capabilities of deep learning. It takes environmental parameters such as temperature and humidity, power fluctuations, and the compensated signal as inputs. Through the LSTM gating mechanism, it learns in real time and cancels the nonlinear interference of environmental factors on the signal, and finally outputs a charge waveform that is close to the real state, providing a high-quality signal for subsequent noise reduction and positioning.

[0073] Furthermore, the filter order of the variable step-size LMS noise suppression algorithm described in S51... Dynamically adjust based on the signal bandwidth of the PVDF sensor: when the signal frequency is higher than 10kHz, take... When the signal frequency is below 10kHz, take ;

[0074] The calibration amplitude described in S52 The signal amplitude was obtained by performing multiple measurements on a standard sample with a known charge density under standard testing conditions (temperature 25℃±2℃, humidity 50%±5%), and taking the average value as the result. ;

[0075] The training process of the LSTM model described in S53 includes: collecting charge signal samples under different environmental parameters (temperature -10℃~60℃, humidity 20%~90%, power fluctuation ±5%), dividing the training set and validation set in a 7:3 ratio, and iteratively training using the Adam optimizer (learning rate 0.001) until the validation set loss function is less than 0.001. The model parameters are saved. Wavelet decomposition is used to distinguish between real signals and sensor miniaturization noise, and adaptive thresholding is used to remove interference, thereby improving the signal-to-noise ratio and providing a clean signal for positioning.

[0076] Furthermore, the wavelet threshold noise reduction process described in S6 is used to eliminate signal interference caused by the miniaturization design of portable space charge sensors. The specific algorithm is as follows:

[0077] Define the original electrical signal as ,in For real signals, This is a noise signal. The standard deviation of noise. It is a time variable;

[0078] Perform N-level wavelet decomposition on the signal to obtain wavelet coefficients. ( The number of decomposition layers, (for coefficient index)

[0079] Adaptive threshold function is used to process wavelet coefficients:

[0080]

[0081] in Interlayer threshold, Let j be the standard deviation of the noise at layer j. The signal length;

[0082] Processed wavelet coefficients Reconstruct the signal to obtain the denoised signal. .

[0083] To specifically eliminate interference caused by sensor miniaturization, wavelet decomposition can decompose the original signal into wavelet coefficients of different scales. The wavelet coefficients of the true charge signal have larger amplitudes, while the wavelet coefficients of the inherent noise introduced by the simplified structure of the miniaturized sensor have smaller amplitudes. An adaptive threshold function is used to mitigate interference from inter-layer thresholds. Filter the coefficients: those with amplitudes greater than or equal to The coefficients are shrunk, and the amplitude is smaller than The coefficients are set to avoid signal distortion or noise residue caused by traditional fixed thresholds; finally, the processed coefficients are restored to the time domain signal through wavelet reconstruction, so as to retain the real charge signal, remove miniaturized interference, further improve the signal-to-noise ratio, and provide a more accurate signal input for subsequent space charge positioning.

[0084] Furthermore, the space charge localization algorithm described in S6 combines the arc-shaped structural features of the portable space charge sensor to calculate the charge position through sound wave propagation time, as shown in the following formula:

[0085] Define the coordinates of the space charge position as (cylindrical coordinate system) Radial distance, For circumferential angle, (where is the axial distance), and the central angle of the arc-shaped detection element of the sensor is... The radius of curvature is ;

[0086] The speed of sound wave propagation in the cable termination medium is ,in For the Young's modulus of the medium, The density of the medium;

[0087] The time difference matrix of the sound waves generated by the charge reaching the sensor detection point is:

[0088]

[0089] in The circumferential angle corresponds to the nth sampling point of the mth detection unit of the sensor. This is the time difference of sound wave propagation at that point;

[0090] By solving the inverse problem of the time difference matrix using the least squares method, the three-dimensional coordinate distribution of space charge is obtained:

[0091]

[0092] in This refers to the number of sensor detection units. The number of sampling points for each unit, This is for the theoretical calculation of the time difference.

[0093] Speed ​​of sound waves in the cable termination medium Young's modulus of the medium With density Since the time difference is a fixed value, the sound wave propagation time is directly proportional to the straight-line distance from the charge to the sensor detection point; the time difference matrix incorporates the central angle of the sensor's arc-shaped structure. radius of curvature Circumferential angle of the detection point A mapping relationship between the three-dimensional coordinates of the charge and the time difference of multiple detection points was constructed, which is adapted to the arc-shaped medium morphology of the cable terminal and solves the positioning deviation caused by the traditional planar positioning algorithm ignoring the arc structure of the terminal. The least squares method solves the inverse problem of the time difference matrix (i.e., deduce the coordinates from the time difference) by minimizing the sum of the squares of the actual measurement time difference and the theoretical calculation time difference. This can effectively suppress the influence of measurement noise on the positioning results and finally obtain the accurate distribution of charge in the radial, circumferential and axial directions, providing accurate location information for assessing the insulation aging degree at different locations of the cable terminal.

[0094] Furthermore, the noise standard deviation described in S6 The calibration is obtained through calibration using the no-load signal of a portable space charge sensor, and the calibration formula is as follows: ,in This is the median function, used to eliminate the influence of inherent sensor noise on threshold calculation. The standard deviation of noise is calibrated using an unloaded signal. The median function resists extreme interference, ensuring accurate wavelet denoising thresholds and avoiding signal distortion or residual noise.

[0095] A cable terminal space charge measurement device includes a high-voltage AC power supply, a high-voltage pulse power supply, an AC matching circuit, a portable space charge sensor, an amplifier, a data acquisition and processing system, and a resistive voltage divider. The input terminal of the resistive voltage divider is electrically connected to the output terminal of the high-voltage AC power supply, and its output terminal is electrically connected to the voltage monitoring input terminal of the data acquisition and processing system. The output terminal of the high-voltage AC power supply is also electrically connected to one end of the cable terminal sample, and the output terminal of the high-voltage pulse power supply is electrically connected to the other end of the cable terminal sample. The AC matching circuit consists of a capacitor C and a resistor R connected in series. Its input terminal is connected to the signal output terminal of the high-voltage AC power supply, and its output terminal is connected to the phase signal input terminal of the data acquisition and processing system. The portable space charge sensor is connected to the input terminal of the amplifier, and the output terminal of the amplifier is connected to the core signal input terminal of the data acquisition and processing system. The processor of the data acquisition and processing system is configured to execute a variable-step-size LMS noise suppression algorithm, a fit-adaptive compensation algorithm, a multivariable LSTM environmental compensation algorithm, wavelet threshold noise reduction processing, and a space charge localization algorithm to generate and display the internal space charge distribution curve of the cable terminal.

[0096] The output of the high-voltage AC power supply is connected to one end of the cable terminal via an AC matching circuit, and the other end is connected to a high-voltage pulse power supply. The input of the resistor divider is connected in parallel to the high-voltage AC power supply, and the output is connected to the voltage monitoring port of the data acquisition system. The portable sensor is fixed to the middle of the terminal stress cone by a ring buckle. The sensor output is connected to the amplifier input, and the amplifier output is connected to the core signal port of the data acquisition system.

[0097] After power is applied, when the high-voltage pulse power supply outputs a 5kV pulse, the amplifier can detect a weak electrical signal, indicating that the sound-to-electric conversion-signal amplification path is normal.

[0098] The high-voltage AC power supply sinusoidal signal is converted into a power frequency square wave through an AC matching circuit. After phase correction, the data acquisition system is triggered to ensure that the high-voltage pulse is accurately applied at the zero-crossing point of the AC voltage, thereby synchronizing the test time with the AC phase and avoiding signal distortion caused by phase misalignment.

[0099] Set the high-voltage pulse power supply output parameters: amplitude 5kV, frequency 1600Hz. Apply the pulse to the terminal through a 0.01μF capacitor. The pulsed electric field excites the space charge inside the terminal to generate mechanical vibration. The vibration energy is transferred through the insulating medium, forming an acoustic signal.

[0100] The portable sensor's arc-shaped measuring electrode is attached to the terminal surface, and the sound wave acts on the piezoelectric film in the middle of the sensor, converting the mechanical energy of the sound wave into a weak electrical signal through the piezoelectric effect.

[0101] The sensor output signal frequency is 8kHz, and the filter order is L=10; the desired output is the XLPE pure charge waveform calibrated in the laboratory. Initial step size Iteratively update the weight vector to the error variance The filtered signal is obtained by filtering out the power frequency electromagnetic noise at the site. ;

[0102] calculate Amplitude =1.8V and rated amplitude =2V deviation =|1.8-2| / 2=0.1, Fit coefficient =1 - 0.8 × 0.1 = 0.92; The compensated signal =1.8 / 0.92≈1.96V, correcting for signal attenuation caused by a slight gap between the sensor and the terminal;

[0103] On-site temperature and humidity recorder display =32℃ =65%, power fluctuation monitor shows =+2%, construct the input matrix = Use a pre-trained LSTM model with 256 hidden layers, and train the Adam optimizer until the validation set loss is < By using a gating mechanism to cancel out nonlinear interference from temperature, humidity, and power fluctuations, the final output is a charge waveform that eliminates environmental interference. .

[0104] right A 5-level wavelet decomposition was performed, using the db4 wavelet basis. The noise standard deviation was calibrated using the sensor's no-load signal: the median absolute value of the 3rd-level wavelet coefficients under no-load conditions was 0.03V. ≈0.044V; interlayer threshold =0.044×√(2ln1024)≈0.202V. After shrinking coefficients with amplitudes ≥0.202V and setting coefficients <0.202V to zero, a clean signal with a signal-to-noise ratio improved to 35dB is obtained after reconstruction. .

[0105] Calculating the velocity of sound waves: Young's modulus of XLPE medium =1.2GPa, density =920kg / m³, according to the formula We get v≈1140m / s;

[0106] Time difference matrix: The sensor contains 8 detection units. =8; 32 points are sampled per unit. =32; No. =3 detection units =Circumferential angle of 15 sampling points =30°, measured time difference =45μs;

[0107] 3D positioning: By solving the inverse time difference problem using the least squares method, the charge coordinates are obtained: radial... =5mm, circumferential =30°, axial =200mm, corresponding to a peak charge density of 1.5μC / m³.

[0108] The computer displays the charge distribution curve inside the cable terminal, clearly showing that "there is local charge accumulation at 200mm axially and 5mm radially", with a charge density of 1.2-1.5μC / m³ in the accumulation area.

[0109] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for measuring space charge at cable terminals, characterized in that, Includes the following steps: S1. A space charge testing platform is built based on the electroacoustic pulse method. The space charge testing platform includes a high-voltage AC power supply, a high-voltage pulse power supply, an AC matching circuit, a portable space charge sensor, an amplifier, a data acquisition and processing system, and a resistor divider. S2. The charge test time is precisely matched with the AC phase through the AC matching circuit. S3. Using the cable terminal as the test sample, apply a high-voltage pulse to the cable terminal sample using the high-voltage pulse power supply to excite the space charge inside the cable terminal sample to vibrate and generate sound waves. S4. Use the portable space charge sensor to capture the sound waves generated in step S3, and convert the sound wave signal into an electrical signal; S5. The electrical signal obtained in step S4 is transmitted to the amplifier for amplification. Then, the processor of the data acquisition and processing system performs optimization to obtain an output charge waveform that eliminates environmental interference. ; S6. The data acquisition and processing system performs the same steps as in step S5. After performing wavelet threshold denoising and space charge localization algorithm calculations in sequence, the space charge distribution curve inside the cable terminal sample is generated and displayed on the computer, thereby realizing the measurement of the space charge of the cable terminal.

2. The method for measuring space charge at cable terminals according to claim 1, characterized in that, The electrical connections of the hardware components of the space charge testing platform described in S1 are as follows: a high-voltage AC power supply provides an AC electric field environment for the test sample via an AC matching circuit; the AC matching circuit performs the function of matching electrical signals to achieve synchronization of the test timing with the AC phase; a high-voltage pulse power supply applies a high-voltage pulse to the test sample through capacitor C to excite the space charge to vibrate; the space charge sensor, amplifier, and data acquisition and processing system are electrically connected in sequence to form a signal path of sound-to-electric conversion-signal amplification-data processing.

3. The method for measuring space charge at cable terminals according to claim 2, characterized in that, In S1, the following signal transmission relationships also exist: The high-voltage pulse power supply outputs a high-voltage pulse electrical signal, which excites the vibration of space charges within the test sample, causing the charge vibration to generate an acoustic signal. The space charge sensor receives the acoustic signal, converts it into a weak electrical signal, and transmits it to the amplifier. The amplifier amplifies and enhances the weak electrical signal, and then transmits the strong electrical signal to the data acquisition and processing system. The data acquisition and processing system acquires amplified electrical signals, analyzes and processes them to generate space charge distribution curves, and displays the results on a computer, thus visualizing the charge distribution.

4. The method for measuring space charge at cable terminals according to claim 1, characterized in that, The portable space charge sensor described in S1 includes a ring-shaped buckle and a detection element. The detection element consists of three parts: a measuring electrode, a piezoelectric film, and a sensor. The measuring electrode is on the top layer, the sensor is on the bottom layer, and the piezoelectric film is sandwiched between the two. The area of ​​the piezoelectric film is the same as the area of ​​the inner aluminum plate of the sensor and it only contacts the inner aluminum plate. The measuring electrode and the sensor are fixed together by screws. The contact surface of the measuring electrode has an arc-shaped structure. The sensor support structure is composed of four nested materials: the inner layer is polymethyl methacrylate, the second layer from the inside out is aluminum plate, the third layer is polytetrafluoroethylene, and the outermost layer is aluminum plate.

5. The method for measuring space charge at cable terminals according to claim 1, characterized in that, The specific optimization process in S5 includes: S51. Processor runs variable step size LMS noise suppression algorithm: Define input vector ,in Let the filter order be . The output electrical signal is from the PVDF sensor; the desired output is a pure charge waveform calibrated in the laboratory. Calculate the output ,error ;according to (in The initial step size, The variance of the error signal. Update the weight vector (for the variance of the input signal) Iterate to The filtered signal is obtained. ; S52, Processor runs adaptive compensation algorithm for fit: Calculation amplitude With the calibration amplitude deviation Fit coefficient Signal compensation To correct the effect of poor adhesion between the PVDF sensor and the stress cone; S53. Processor runs multivariable LSTM environment compensation algorithm: Constructing input matrix ,in For humidity parameters, For temperature parameters, For power fluctuation parameters, call the pre-trained LSTM model, according to... , Update the cell state (where In cellular state, In hidden state, These are the activation values ​​for the forget gate, input gate, and output gate, respectively. This is the weight matrix. For bias terms, (Hadamard product), outputting a charge waveform that eliminates environmental interference. .

6. The method for measuring space charge at cable terminals according to claim 5, characterized in that, The filter order of the variable step-size LMS noise suppression algorithm described in S51 Dynamically adjust based on the signal bandwidth of the PVDF sensor: when the signal frequency is higher than 10kHz, take... When the signal frequency is below 10kHz, take ; The calibration amplitude described in S52 The signal amplitude was obtained by performing multiple measurements on a standard sample with a known charge density under standard testing conditions (temperature 25℃±2℃, humidity 50%±5%), and taking the average value as the result. ; The training process of the LSTM model described in S53 includes: collecting charge signal samples under different environmental parameters, dividing the training set and validation set in a 7:3 ratio, iteratively training with the Adam optimizer until the validation set loss function is less than 1e-5, and saving the model parameters.

7. The method for measuring space charge at cable terminals according to claim 1, characterized in that, The wavelet thresholding noise reduction process described in S6 is used to eliminate signal interference caused by the miniaturization design of portable space charge sensors. The specific algorithm is as follows: Define the original electrical signal as ,in For real signals, This is a noise signal. The standard deviation of noise. It is a time variable; Perform N-level wavelet decomposition on the signal to obtain wavelet coefficients. ( The number of decomposition layers, (for coefficient index) Adaptive threshold function is used to process wavelet coefficients: in Interlayer threshold, Let j be the standard deviation of the noise at layer j. The signal length; Processed wavelet coefficients Reconstruct the signal to obtain the denoised signal. .

8. The method for measuring space charge at cable terminals according to claim 1, characterized in that, The space charge localization algorithm described in S6 combines the arc-shaped structural features of the portable space charge sensor to calculate the charge location through sound wave propagation time. The specific formula is as follows: Define the coordinates of the space charge position as , Radial distance, For circumferential angle, The axial distance is given by the arc-shaped detection element of the sensor, and the central angle of the arc is given by the arc-shaped detection element. The radius of curvature is ; The speed of sound wave propagation in the cable termination medium is ,in For the Young's modulus of the medium, The density of the medium; The time difference matrix of the sound waves generated by the charge reaching the sensor detection point is: in The circumferential angle corresponds to the nth sampling point of the mth detection unit of the sensor. This is the time difference of sound wave propagation at that point; By solving the inverse problem of the time difference matrix using the least squares method, the three-dimensional coordinate distribution of space charge is obtained: in This refers to the number of sensor detection units. The number of sampling points for each unit, This is for the theoretical calculation of the time difference.

9. The method for measuring space charge at cable terminals according to claim 7, characterized in that, The noise standard deviation described in S6 The calibration is obtained through calibration using the no-load signal of a portable space charge sensor, and the calibration formula is as follows: , where median is the median function, used to eliminate the influence of sensor inherent noise on threshold calculation.

10. A cable terminal space charge measuring device, characterized in that, The system includes a high-voltage AC power supply, a high-voltage pulse power supply, an AC matching circuit, a portable space charge sensor, an amplifier, a data acquisition and processing system, and a resistive voltage divider. The input of the resistive voltage divider is electrically connected to the output of the high-voltage AC power supply, and its output is electrically connected to the voltage monitoring input of the data acquisition and processing system. The output of the high-voltage AC power supply is also electrically connected to one end of the cable terminal sample, and the output of the high-voltage pulse power supply is electrically connected to the other end of the cable terminal sample. The AC matching circuit consists of a capacitor C and a resistor R connected in series. Its input is connected to the signal output of the high-voltage AC power supply, and its output is connected to the phase signal input of the data acquisition and processing system. The portable space charge sensor is connected to the input of the amplifier, and the output of the amplifier is connected to the core signal input of the data acquisition and processing system. The processor of the data acquisition and processing system is configured to execute variable-step-size LMS noise suppression algorithm, adaptive fit compensation algorithm, multivariable LSTM environment compensation algorithm, wavelet threshold noise reduction processing, and space charge localization algorithm to generate and display the internal space charge distribution curve of the cable terminal.

Images