PRPD image generation circuit, method, and partial discharge type identification method

Abstract

This invention discloses a PRPD image generation circuit, method, and partial discharge type identification method, belonging to the field of power equipment discharge detection technology. The PRPD image generation circuit of this invention includes: a power frequency generation unit for generating a power frequency synchronization signal; a counter for updating the phase number starting at the falling edge of the power frequency synchronization signal; a peak hold unit for acquiring partial discharge signals; a first digital-to-analog converter for updating the reference voltage; a comparator for comparing the partial discharge voltage of the partial discharge signal with the reference voltage; and a data generation unit for generating and outputting an image matrix. The PRPD image generation circuit of this invention does not require storing, processing, and transmitting large amounts of partial discharge data, generating PRPD images with amplitude and phase information using a low-complexity, low-data-volume method. The partial discharge type identification method of this invention identifies the partial discharge type based on the analysis of the partial discharge amplitude and phase of the PRPD image.

Description

Technical Field

[0001] This invention relates to the field of power equipment discharge detection technology, and in particular to a PRPD image generation circuit, method, and method for identifying partial discharge types. Background Technology

[0002] Partial discharge (PD) has different types, including tip discharge, particulate discharge, air gap discharge, and suspended discharge. PD classification is a major component of electrical insulation fault diagnosis in high-voltage equipment. With the development of artificial intelligence technology, deep learning has become an important method for partial discharge type identification. Compared with traditional waveform recognition methods, deep learning-based PD pattern recognition methods have advantages such as low complexity, high recognition accuracy, and high intelligence, as described in Chinese Patent Publication No. CN111814834B. In actual field PD classification and recognition based on deep learning models such as convolutional neural networks, PRPD (Phase Resolved Partial Discharge) images are first generated from partial discharge data collected on-site, and then online pattern recognition is performed using a deep learning model. The key information provided by the PRPD image for pattern recognition includes the phase and amplitude characteristics of the partial discharge pulse. By describing the relationship between the amplitude, phase, and number of discharges of the discharge signal, the PRPD image not only reflects the variation of the discharge signal amplitude with phase but also shows the distribution of the number of discharges in different phase intervals.

[0003] Traditional UHF PD receivers digitize the signal using an ADC (Analog-to-Digital Converter) after filtering, amplification, and detection, followed by relevant signal processing to extract amplitude information and record the phase information of the PD pulse. Unlike typical carrier-modulated wireless communication signals, PD signals are short-interval pulsed broadband baseband signals, with most of the time lacking useful information. With the development of AIoT (Artificial Intelligence of Things) technology, edge computing at sensing terminals is becoming a technological trend. Sensing terminals are typically data processing platforms with limited hardware resources, often based on microcontrollers. Traditional methods using ADCs for real-time sampling of partial discharge signals generate large amounts of data, much of which lacks useful information. Furthermore, high-speed ADCs consume significant power, posing a considerable challenge to IoT sensing terminals with limited hardware resources and typically battery-powered operation. Therefore, a low-data-volume, low-complexity method is needed to generate PRPD images without an ADC, enabling the extraction of PD pulse amplitude and phase feature information for low-cost PD type identification at IoT sensing terminals. Summary of the Invention

[0004] To address the difficulty of identifying partial discharge types at the sensing end of the Internet of Things (IoT), this invention provides a PRPD image generation circuit, a method, and a partial discharge type identification method. The PRPD image generation circuit of this invention does not require storing, processing, and transmitting large amounts of partial discharge data, generating PRPD images with amplitude and phase information using a low-complexity, low-data-volume method. Furthermore, the partial discharge type identification method of this invention identifies the partial discharge type by analyzing the amplitude and phase of the partial discharge based on the PRPD image.

[0005] The objective of this invention can be achieved through the following technical means:

[0006] A circuit for generating PRPD images, comprising:

[0007] The power frequency generation unit is used to generate power frequency synchronization signals;

[0008] A counter used to update the phase count starting at the falling edge of the power frequency synchronization signal;

[0009] The peak hold unit is used to acquire partial discharge signals, and its input is connected to the main circuit.

[0010] The first digital-to-analog converter is used to update the reference voltage;

[0011] The comparator is used to compare the partial discharge voltage U1 of the partial discharge signal with the reference voltage U2. The first input terminal of the comparator is connected to the first digital-to-analog converter, the second input terminal of the comparator is connected to the peak hold unit, and the output terminal is connected to the first digital-to-analog converter.

[0012] A data generation unit is used to generate and output an image matrix. The data generation unit is connected to a first digital-to-analog converter and a counter.

[0013] If U1 is greater than λU2, the comparator outputs a rising edge of the partial amplifier to the first digital-to-analog converter; otherwise, the comparator outputs a falling edge of the partial amplifier to both the first digital-to-analog converter and the counter. λ is the precision parameter.

[0014] When the first digital-to-analog converter receives the rising edge of the partial discharge, the reference voltage U2 increases. When the first digital-to-analog converter and the counter receive the falling edge of the partial discharge, the first digital-to-analog converter outputs the current reference voltage U2, and the counter outputs the current phase number.

[0015] The data generation unit generates multiple arrays containing reference voltages and phase numbers. It normalizes the reference voltages and phase numbers of the arrays according to the number of rows and columns of the image matrix, and then generates the elements of the image matrix according to the number of rows and columns of each array.

[0016] In this invention, the output of the comparator is also connected to a peak hold unit. When the comparator generates a rising edge of the partial discharge, it outputs a high level to the peak hold unit; when the comparator generates a falling edge of the partial discharge, it outputs a low level to the peak hold unit.

[0017] When the comparator signal received by the peak hold unit remains low, the peak hold unit enters the envelope tracking state.

[0018] When the comparator signal received by the peak hold unit remains high or transitions from low to high, the peak hold unit enters the peak detection state.

[0019] When the comparator signal received by the peak hold unit transitions from a high level to a low level, the peak hold unit releases the current partial discharge signal.

[0020] A method for generating a PRPD image based on the aforementioned PRPD image generation circuit includes the following steps:

[0021] Step 1: The data generation unit creates an image matrix, the power frequency generation unit generates a power frequency synchronization signal, and the counter resets the phase count and starts counting after receiving the falling edge of the power frequency synchronization signal.

[0022] Step 2: The peak hold unit enters the envelope tracking state, the first digital-to-analog converter initializes the reference voltage U2, and the peak hold unit acquires the partial discharge signal;

[0023] Step 3: The comparator compares the partial discharge voltage U1 of the partial discharge signal with the reference voltage U2. If U1 is greater than λU2, proceed to step 4; otherwise, return to step 2. λ is the accuracy parameter.

[0024] Step 4: The peak hold unit enters the peak detection state, the comparator outputs a rising edge of partial discharge to the first digital-to-analog converter, and the reference voltage U2 increases;

[0025] Step 5: The peak hold unit re-acquires the partial discharge signal, and the comparator re-compares the partial discharge voltage U1 with the reference voltage U2. If U1 is greater than λU2, return to step 4; otherwise, proceed to step 6.

[0026] Step 6: The comparator outputs a falling edge of partial discharge to the first digital-to-analog converter and the counter. The first digital-to-analog converter outputs the current reference voltage U2, the counter outputs the current phase number, and the reference voltage U2 is reset.

[0027] Step 7: Repeat steps 2 to 6 to generate multiple arrays containing reference voltage and phase number. Normalize the reference voltage and phase number of the arrays according to the number of rows and columns of the image matrix. Then generate the elements of the image matrix according to the normalized values ​​of reference voltage and phase number, and output the image matrix.

[0028] In this invention, in step 4, the increment step size U of the reference voltage U2 is... step =0.001V to 0.005V, the image matrix is ​​a 32×32 two-dimensional matrix, the power frequency synchronization signal is a square wave signal with a period of 20ms, and the voltage of the power frequency synchronization signal is 3.3V.

[0029] In this invention, in step 4, the peak hold unit receives the reference voltage U from the second digital-to-analog converter. ref The partial discharge voltage U1 = U0 - U ref U0 is the current voltage of the partial discharge signal.

[0030] In this invention, in step 7, the image matrix is ​​an N x M matrix, based on the maximum reference voltage U. max Set the normalization ratio M / U max M is the number of columns in the image matrix, and the normalized value of the reference voltage U2 is Floor(U2M / U max )+1, Floor() is the floor function.

[0031] In this invention, the number K of arrays with a normalized value of n for the number of statistical phases and a normalized value of m for the reference voltage is... nm Element P in the image matrix nm =(1-0.02K nm / T)×255, where T is the detection duration, n≤N, m≤M.

[0032] A method for identifying partial discharge types based on the PRPD image generation method, comprising the following steps:

[0033] Step 100: Generate at least two sets of partial discharge signals of partial discharge types, implement the PRPD image generation method, and generate multiple sets of image matrices;

[0034] Step 200: Train a partial discharge classification model based on multiple sets of image matrices and their corresponding partial discharge types;

[0035] Step 300: Generate an image matrix according to the PRPD image generation method, and identify the partial discharge type of the image matrix according to the partial discharge classification model.

[0036] In this invention, in step 100, a partial discharge signal is generated based on the gas-insulated switchgear, and the partial discharge types include tip discharge, particle discharge, air gap discharge and suspension discharge.

[0037] In this invention, the partial discharge classification model includes multiple sets of alternating hourglass-shaped attention layers and transition layers. The hourglass-shaped attention layer is composed of multiple stacked hourglass-shaped attention units, each of which has a lightweight convolutional block and an attention block for adjusting the feature weights of the image matrix.

[0038] The advantages of implementing the PRPD image generation circuit, method, and partial discharge type identification method of the present invention are as follows:

[0039] 1. This invention greatly simplifies the overall circuit structure, can make full use of the built-in hardware circuit resources of the microcontroller, and adopts a single-chip partial discharge peak holding circuit scheme, making the system more stable, reliable and easy to implement.

[0040] 2. This invention realizes a PRPD image generation method with low data volume and low complexity, which can realize the peak and phase measurement of partial discharge pulse signal without ADC converter, thus reducing system requirements and costs.

[0041] 3. This invention combines PRPD image generation circuit and deep learning model to realize partial discharge pattern recognition on a microcontroller, which is suitable for application in power Internet of Things sensing terminals with limited resources. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of an ultra-high frequency partial discharge signal;

[0043] Figure 2 This is a block diagram of the PRPD image generation circuit of the present invention;

[0044] Figure 3 This is a schematic diagram of the PRPD image generation circuit of the present invention;

[0045] Figure 4 This is a circuit diagram of the peak hold unit of the present invention;

[0046] Figure 5 This is a voltage schematic diagram of the subtraction amplifier circuit of the present invention;

[0047] Figure 6 This is a communication principle diagram of the power frequency generation unit of the present invention;

[0048] Figure 7 This is a flowchart illustrating the operation of the counter of the present invention;

[0049] Figure 8 This is a schematic diagram of the PRPD image of the present invention;

[0050] Figure 9 A schematic diagram of a PRPD image generated on a PC based on a tip discharge signal acquired using existing technology;

[0051] Figure 10 A schematic diagram of a PRPD image generated on a PC based on particle discharge signals acquired using existing technology;

[0052] Figure 11 A schematic diagram of a PRPD image generated on a PC based on an air gap discharge signal acquired using existing technology;

[0053] Figure 12 A schematic diagram of a PRPD image generated on a PC based on a suspended discharge signal acquired using existing technology;

[0054] Figure 13 This is a schematic diagram of the PRPD image generated at the sensing end by the tip discharge signal acquired in this invention.

[0055] Figure 14 This is a schematic diagram of the PRPD image generated at the sensing end by the particle discharge signal collected in this invention;

[0056] Figure 15 This is a schematic diagram of the PRPD image generated at the sensing end by the air gap discharge signal acquired in this invention;

[0057] Figure 16 This is a schematic diagram of the PRPD image generated at the sensing end by the suspended discharge signal collected in this invention;

[0058] Figure 17 This is a flowchart of the PRPD image generation method according to the PRPD image generation circuit of the present invention;

[0059] Figure 18 This is a flowchart of the partial discharge type identification method of the present invention;

[0060] Figure 19 This is a schematic diagram of the partial discharge classification model of the present invention. Detailed Implementation

[0061] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0062] In existing technologies, partial discharge model identification is typically performed on a PC, utilizing the PC's data processing capabilities to analyze the raw partial discharge data. However, in the identification of ultra-high frequency (UHF) partial discharge, the raw data contains a large amount of invalid information. With the development of power IoT technology, it is necessary to implement UHF partial discharge mode identification at the sensing end using a microcontroller. To embed the identification algorithm into the microcontroller, lightweight data acquisition and model implementation are required. The PRPD image generation circuit, method, and partial discharge type identification method of this invention can fully utilize the built-in hardware circuit resources of the microcontroller to generate PRPD images with low data volume and low complexity, making it suitable for resource-constrained power IoT sensing terminal applications. Example 1

[0063] like Figures 1 to 16 The PRPD image generation circuit of the present invention, as shown, includes: a power frequency generation unit, a counter, a peak hold unit, a comparator, a first digital-to-analog converter (DAC), a data generation unit, and a second DAC. The present invention uses the ultra-low-power microcontroller STM32L476RG. This microcontroller has a main frequency of 80MHz and integrates a comparator COMP, DAC1, DAC2, and a timer, which serve as the comparator, first DAC, second DAC, and counter, respectively. The STM32L476RG's built-in COMP can respond within a time range of 55~80ns, and the setup time of DAC1 and DAC2 is 1.4us, enabling rapid detection and tracking of changes in the partial discharge signal. DAC1 and DAC2 can be configured with different bit depths; this invention uses 8 bits.

[0064] The power frequency generation unit is used to generate a power frequency synchronization signal. Specifically, the power frequency generation unit converts 220V, 50Hz power frequency AC power into a 3.3V, 50Hz power frequency synchronization signal (square wave signal) to achieve precise phase synchronization. The power frequency generation unit uses zero-crossing detection technology to extract the phase information of the power frequency voltage, and through an optocoupler isolation circuit, realizes the conversion from 220V power frequency AC power to a 3.3V square wave signal. The power frequency generation unit ensures that the partial discharge signal collected by this invention is in phase with the power frequency voltage of the power equipment, providing a zero-starting reference point. In actual substations and other locations, the 220V power frequency power socket is usually some distance away from the partial discharge detection point. Therefore, a wireless transceiver circuit (wireless transmitting circuit, wireless receiving circuit) is needed to transmit the power frequency synchronization signal generated by the power frequency generation unit to the detection point, such as... Figure 6 As shown. The wireless transmitting and receiving circuits use MAX1472 and MAX1470, which have low power consumption characteristics and simple and easy-to-implement modulation and demodulation methods, making them suitable for use in the embedded devices of this invention.

[0065] The counter updates the phase count starting at the falling edge of the power frequency synchronization signal. The power frequency synchronization signal provides a reference point for the counter, which begins counting at the falling edge of the power frequency synchronization signal. The counting period of the counter in this invention is one power frequency cycle of 20ms, which corresponds to the horizontal coordinate (phase direction) range of the PRPD image. Each falling edge of the power frequency synchronization signal represents the 0° phase position, and the count value (CNT) corresponds to the specific value of the horizontal coordinate of the partial discharge signal in the image. This invention can set the prescaler and counting period of the timer (counter) according to the microcontroller's main frequency, so that the counter's counting range is 1 to N. For example, when the counting range is 20, the counting interval is 1ms. The function of the prescaler is to divide the clock frequency into a lower counting frequency to meet the requirements of the power frequency cycle. By reasonably setting the prescaler and period, the counting frequency of the counter is made an integer multiple of the power frequency, thereby achieving accurate phase acquisition.

[0066] The peak hold unit is used to acquire partial discharge signals, and its input is connected to the main circuit. The main circuit refers to the current circuit of the power equipment that needs to measure partial discharge, such as high-voltage power systems and mains power systems. Specifically, the peak hold unit is implemented by the HMC1120 chip, where pin 8, Mode_Sel, is used to select the operating mode. When Mode_Sel is high, it enters peak detection mode, capturing and holding the peak value of the signal; when it is low, it releases the saved peak value and enters envelope tracking mode. When pin 6, PH_Cap, is directly grounded without an external capacitor, the peak value can be held for approximately 150µs, giving the microcontroller sufficient time to complete the partial discharge signal peak measurement and acquisition. With an external capacitor, the peak hold time can be even longer. Pin 7, ETOUT, outputs a range of 0.95V-1.65V, where 0.95V is a fixed DC component, and the dynamic range is 0-0.7V.

[0067] To ensure the accuracy of subsequent signal processing, the peak hold unit includes a subtraction amplifier circuit to remove the fixed DC component and amplify the output peak signal. The subtraction amplifier circuit is implemented using a COS8092 chip, which integrates two operational amplifiers: one designed as the subtraction amplifier circuit, and the other as an emitter follower to isolate the peak hold circuit from the microcontroller. (Refer to...) Figure 5 The input voltage of the subtraction amplifier circuit is the reference voltage of the second digital-to-analog converter (DAC2). The amplification factor of the operational amplifier can be set according to actual needs; in this embodiment, the amplified output voltage is approximately 1.5V. The component parameters of the peak hold circuit are as follows: C1=C2=C4=1nF, C3=10nF, C5=0.1nF, C6=C7=100nF, R1=50Ω, R2=604Ω, R3=R6=5kΩ, R4=R5=15kΩ.

[0068] The first digital-to-analog converter (DAC) is used to update the reference voltage. When the first DAC receives a rising edge of the partial discharge, the reference voltage U2 increments. When the first DAC and the counter receive a falling edge of the partial discharge, the first DAC outputs the current reference voltage U2, and the counter outputs the current phase number.

[0069] The comparator compares the partial discharge voltage U1 of the partial discharge signal with the reference voltage U2. The first input of the comparator is connected to the first digital-to-analog converter (DAC), the second input is connected to the peak hold unit, and the output is connected to the DAC. If U1 is greater than λU2, the comparator outputs a rising edge of the partial discharge signal to the DAC; otherwise, the comparator outputs a falling edge of the partial discharge signal to both the DAC and the counter.

[0070] The comparator's output is also connected to a peak hold unit. When the comparator generates a rising edge of the partial discharge signal, it outputs a high level to the peak hold unit; when it generates a falling edge, it outputs a low level. When the comparator signal received by the peak hold unit remains low, the peak hold unit enters envelope tracking mode. When the comparator signal received by the peak hold unit remains high or transitions from low to high, the peak hold unit enters peak detection mode. When the comparator signal received by the peak hold unit transitions from high to low, the peak hold unit releases the current partial discharge signal.

[0071] Specifically, the comparator sets an initial output threshold based on actual needs to filter out some noise. In the absence of a partial discharge signal, the partial discharge voltage U1 ≤ λU2, and the comparator output remains low. At this time, the threshold setting isolates the noise signal. λ is a precision parameter, typically ranging from 1 to 1.001. When a partial discharge signal is acquired, its voltage exceeds the threshold THRH (in this embodiment, THRH = 0.1V). At this point, U1 > λU2 (the partial discharge voltage is slightly greater than the reference voltage), and the output level of the first digital-to-analog converter jumps from low to high, generating a rising edge of the partial discharge signal. Then, the first digital-to-analog converter updates the reference voltage U2, increasing U2 by a fixed step size U each time. step During the output adjustment process of the first digital-to-analog converter (DAC), the comparator maintains a high-level output, and the high-level Mode_Sel puts the circuit in peak hold mode. When the partial discharge voltage U1 ≤ λU2 (the reference voltage U2 is slightly greater than the partial discharge voltage U1), the comparator output switches from high to low, generating a partial discharge falling edge. The first DAC outputs a reference voltage U2, which is close to the partial discharge voltage U1, i.e., the amplitude of the partial discharge signal. Simultaneously, the first DAC resets to its initial threshold, and the low-level Mode_Sel puts the peak hold unit into envelope tracking mode, ready to start acquiring the next partial discharge signal.

[0072] This invention uses a comparator to monitor signal changes in real time, and in conjunction with the adjustment of the first digital-to-analog converter, ensures the capture of the amplitude of the partial discharge signal. It then gradually adjusts the reference voltage U2 to approximate the amplitude, thus completing the acquisition and processing of the partial discharge amplitude. The accuracy of the amplitude measurement depends on the set step size U. step U step The larger the step size, the worse the accuracy, and the shorter the time it takes for the reference voltage U2 to reach slightly higher than the partial discharge voltage U1, and vice versa. In addition, the smaller the step size, the lower the accuracy parameter λ is, so as to improve the detection accuracy simultaneously.

[0073] The data generation unit is used to generate and output an image matrix. The data generation unit is connected to the first digital-to-analog converter and the counter. The data generation unit generates multiple arrays containing reference voltages and phase numbers. It normalizes the reference voltages and phase numbers of the arrays according to the number of rows and columns of the image matrix, and then generates the elements of the image matrix according to the corresponding number of rows and columns of each array.

[0074] Reference Figure 8 PRPD images are typically composed of an array of multiple power frequency cycles superimposed. The pixel value of a PRPD image is determined by the number of times a partial discharge signal appears at the corresponding phase and amplitude; for example, the maximum number of occurrences within 1 second is 50. In power equipment, different types of partial discharges exhibit different distribution characteristics with varying power frequency phases, and these distribution characteristics determine that different types of partial discharges have different PRPD images. Example 2

[0075] like Figures 8 to 17 As shown, a PRPD image generation method based on the PRPD image generation circuit of the present invention includes the following steps.

[0076] Step 1: The data generation unit creates an image matrix, and the power frequency generation unit generates a power frequency synchronization signal. Upon receiving the falling edge of the power frequency synchronization signal, the counter resets the phase count and begins counting. The number of rows and columns of the image matrix is ​​preset according to the microcontroller's capacity. Assume an M×N PRPD image needs to be acquired, and the image matrix is ​​an M×N matrix. The phase corresponds to the horizontal coordinate of the PRPD image; therefore, the counter (TIM) uses counts from 1 to N to represent different phase intervals within one power frequency cycle (20ms). Larger values ​​of M and N result in higher system accuracy, while smaller values ​​of M and N result in smaller system storage requirements.

[0077] Step 2: The peak hold unit enters the envelope tracking state, the first digital-to-analog converter initializes the reference voltage U2, and the peak hold unit acquires the partial discharge signal. The initialized reference voltage U2 is the threshold THRH of Example 1.

[0078] Step 3: The comparator compares the partial discharge voltage U1 of the partial discharge signal with the reference voltage U2. If U1 is greater than λU2, proceed to step 4; otherwise, return to step 2, where λ is the accuracy parameter. This embodiment determines whether to enter peak detection mode by detecting the Q level of the comparator. The comparator's interrupt function is set to edge-triggered mode. When the Q level changes, the comparator's interrupt function is triggered, and different processing flows are entered depending on the type of change (rising edge or falling edge of the partial discharge). If the Q level remains low, it indicates that the partial discharge voltage has not exceeded the reference voltage, and the program will continue to detect the Q level state, returning to step 2 until a valid change signal is detected.

[0079] Step 4: The peak hold unit enters the peak detection state. The comparator outputs a rising edge of partial discharge to the first digital-to-analog converter, and the reference voltage U2 increases. If the Q level jumps from low to high (rising edge of partial discharge), it indicates that the amplitude of the partial discharge signal exceeds the analog value output by the first digital-to-analog converter, and the program will enter the peak finding process (peak detection state). In this stage, the present invention follows a preset step size U... step The reference voltage is gradually increased, and the updated reference voltage is output through a first digital-to-analog converter. By gradually increasing the analog value output by the first digital-to-analog converter, this invention can approximate the peak value of the partial discharge signal. This process is iterative; after each increase in the reference voltage, the Q level is re-detected to determine whether the peak value of the partial discharge signal has been reached. This is a peak value finding and approximation stage for the partial discharge signal.

[0080] The increment step size U of the reference voltage U2 step =0.001V to 0.005V, the image matrix can be a 32×32 two-dimensional matrix. In a preferred embodiment, the peak hold unit receives the reference voltage U from the second digital-to-analog converter. ref The partial discharge voltage U1 = U0 - U ref U0 is the current voltage of the partial discharge signal. The reference voltage U... ref =0.95V, the reference voltage serves as noise reduction and is used in conjunction with the subtraction amplifier circuit. To ensure data comparability, the reference voltage can also be processed using the same subtraction method after the partial discharge voltage is subtracted.

[0081] Step 5: The peak hold unit re-acquires the partial discharge signal, and the comparator re-compares the partial discharge voltage U1 with the reference voltage U2. If U1 is greater than λU2, return to step 4; otherwise, proceed to step 6. Specifically, if the Q level jumps from high to low (falling edge of partial discharge), it indicates that the analog value output by the first digital-to-analog converter has gradually increased to a value higher than the amplitude of the partial discharge signal, and the peak value is saved, proceeding to step 6.

[0082] Step 6: The comparator outputs a falling edge of the partial discharge signal to the first digital-to-analog converter (DAC) and the counter. The DAC outputs the current reference voltage U2, the counter outputs the current phase number, and the reference voltage U2 is reset. This invention saves the current reference voltage to an array to record the peak value of the partial discharge signal. After saving the peak data, this invention resets the reference voltage to the threshold THRH and releases the partial discharge peak value through the Mode_Sel terminal to enter the envelope tracking state and restart the detection of the next partial discharge signal.

[0083] Step 7: Repeat steps 2 to 6 to generate multiple arrays containing reference voltages and phase numbers. Normalize the reference voltage (partial discharge voltage amplitude) and phase number of the arrays according to the number of rows and columns of the image matrix. Then, generate the elements of the image matrix based on the normalized reference voltage and phase number values, and output the image matrix. The number of repetitions depends on the number of arrays required. For example, if the number of power frequency cycles within 1 second is 50, 50 arrays can be acquired. The arrays acquired in this invention undergo amplitude normalization and pixel normalization. Amplitude normalization is used to unify partial discharge signals with different amplitudes into the same range, avoiding image quality issues caused by excessive amplitude differences. Pixel normalization converts the processed data into an image format suitable for pattern recognition.

[0084] Reference Figure 8 The image matrix is ​​an N x M matrix, based on the maximum phase number W. max Set the normalization ratio N / W max N is the number of rows in the image matrix, and the normalized value of the phase number W is n = Floor(WN / W max +1, based on the maximum reference voltage U max Set the normalization ratio M / U max M is the number of columns in the image matrix, and the normalized value of the reference voltage U2 is m = Floor(U2M / U max )+1, Floor() is the floor function. The number K of arrays with a normalized phase number of n and a normalized reference voltage of m. nm Element P in the image matrix nm =(1-0.02K nm / T)×255, where T is the detection duration. n≤N, m≤M.

[0085] The image matrix corresponds to the PRPD image, and the element values ​​of the image matrix are the pixel values ​​of the corresponding pixels. The smaller the pixel value, the more partial discharges occur. The PRPD image is an array of data collected over multiple power frequency cycles, superimposed according to amplitude and phase. If the PRPD image size is M×N=10×10, both amplitude and phase are divided into 10 equal parts, with each phase interval being 36°. The minimum pixel value of the PRPD image is 0, indicating that all partial discharge signals appear in the corresponding phase and amplitude. The maximum value is 255, indicating that no partial discharge signals appear in either phase or amplitude. Example 3

[0086] like Figures 8 to 19 The present invention provides a method for identifying the type of partial discharge based on the PRPD image generation method, comprising the following steps.

[0087] Step 100: Generate partial discharge signals of at least two types of partial discharge, implement the PRPD image generation method, and generate multiple image matrices. Typically, laboratory GIS (Gas Insulated Switchgear) partial discharge test platforms have four built-in partial discharge models, which generate corresponding partial discharge signals based on gas-insulated switchgear. The partial discharge types include tip discharge, particle discharge, air gap discharge, and suspension discharge.

[0088] Step 200: Train a partial discharge (PRPD) classification model based on multiple sets of image matrices and their corresponding partial discharge types, and deploy the PRPD classification model to a microcontroller. The PRPD classification model of this invention is a lightweight deep learning model suitable for classifying PRPD images. The PRPD classification model includes multiple sets of alternately connected hourglass-shaped attention layers (Sandglass-SE layers) and transition layers. The hourglass-shaped attention layer is composed of multiple stacked hourglass-shaped attention units (Sandglass-SE blocks), each with lightweight convolutional blocks and attention blocks for adjusting the feature weights of the image matrix. In this patent application, the attention blocks may consist of convolutional operations.

[0089] Reference Figure 19The main body of the partial-effect classification model in this embodiment consists of alternating Sandglass-SE layers and transition layers. Convolutional and max-pooling layers are added between the input and the main body, followed by global pooling and fully connected layers. The Sandglass-SE layer is composed of multiple stacked Sandglass-SE blocks, employing lightweight convolutional blocks and optimized depthwise separable convolutions to reduce the number of model parameters. The SE attention mechanism adaptively adjusts the weights of each feature channel during feature map recovery, enhancing attention to important features and improving feature representation capabilities to more accurately capture important feature information. This lightweight model, deployed on the STM32L476RG microcontroller in Embodiment 1, occupies only 73.36KB of Flash and 41.09KB of RAM.

[0090] Step 300: Generate an image matrix according to the PRPD image generation method, and identify the partial discharge type of the image matrix according to the partial discharge classification model. The identification result of the partial discharge type can be sent to a host computer or transmitted to other sensing terminals via a wireless communication module.

[0091] Figures 9 to 12 The images of different partial discharge types (PRPDs) generated on a PC using traditional high-speed ADC and digital processor technology correspond to a 32×32 image matrix. Figures 13 to 16 The PRPD images of different partial discharge types generated by the microcontroller-based partial discharge peak and phase acquisition method of this invention are represented by a 32×32 image matrix. Figures 9 to 12 and Figures 13 to 16 The PRPD images show similar distribution characteristics. Among them, tip discharges are mainly concentrated in the first quadrant of the image, showing obvious phase distribution characteristics; particle discharges show a more dispersed characteristic, with no obvious phase distribution characteristics and a large number of discharges; air gap discharges mainly appear in the first and third quadrants of the image, and their discharge characteristics overlap with those of tip discharges; suspension discharges are also distributed in the first and third quadrants, but they are mainly concentrated in the upper middle area of ​​the image.

[0092] By deploying the trained lightweight partial discharge classification model onto a microcontroller, pattern recognition can be performed on the generated PRPD images to determine the type of partial discharge, enabling on-site monitoring of insulation faults in power equipment. The table below compares the results of partial discharge type identification on a PC (existing technology) and on a microcontroller (this invention). On the PC, the classification and identification uses PRPD images generated by traditional high-speed ADC and digital processor technology, trained on the PC. On the microcontroller, the classification and identification uses the partial discharge type identification method of the PRPD image generation method described in this invention. The table shows that the classification accuracy of this invention deployed on a microcontroller is close to that on a PC, verifying the effectiveness of this invention's low-complexity method for obtaining partial discharge feature data to generate PRPD images, and also demonstrating the portability of the lightweight partial discharge classification model on a microcontroller embedded platform.

[0093] PRPD image generation method Accuracy Recall rate F1 score Accuracy Existing technology / PC version 96.97% 96.88% 96.90% 96.88% This invention / microcontroller 93.22% 93.00% 93.05% 93.00%

[0094] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A circuit for generating PRPD images, characterized in that, include: The power frequency generation unit is used to generate power frequency synchronization signals; A counter used to update the phase count starting at the falling edge of the power frequency synchronization signal; The peak hold unit is used to acquire partial discharge signals, and its input is connected to the main circuit. The first digital-to-analog converter is used to update the reference voltage; The comparator is used to compare the partial discharge voltage U1 of the partial discharge signal with the reference voltage U2. The first input terminal of the comparator is connected to the first digital-to-analog converter, the second input terminal of the comparator is connected to the peak hold unit, and the output terminal is connected to the first digital-to-analog converter. A data generation unit is used to generate and output an image matrix. The data generation unit is connected to a first digital-to-analog converter and a counter. If U1 is greater than λU2, the comparator outputs a rising edge of the partial amplifier to the first digital-to-analog converter; otherwise, the comparator outputs a falling edge of the partial amplifier to both the first digital-to-analog converter and the counter. λ is the precision parameter. When the first digital-to-analog converter receives the rising edge of the partial discharge, the reference voltage U2 increases. When the first digital-to-analog converter and the counter receive the falling edge of the partial discharge, the first digital-to-analog converter outputs the current reference voltage U2, and the counter outputs the current phase number. The data generation unit generates multiple arrays containing reference voltages and phase numbers, normalizes the reference voltages and phase numbers of the arrays according to the number of rows and columns of the image matrix, and then generates the elements of the image matrix according to the corresponding number of rows and columns of each array. The comparator's output is also connected to a peak hold unit. When the comparator generates a rising edge of the partial discharge, it outputs a high level to the peak hold unit; when it generates a falling edge of the partial discharge, it outputs a low level to the peak hold unit. When the comparator signal received by the peak hold unit remains low, the peak hold unit enters the envelope tracking state. When the comparator signal received by the peak hold unit remains high or transitions from low to high, the peak hold unit enters the peak detection state. When the comparator signal received by the peak hold unit transitions from a high level to a low level, the peak hold unit releases the current partial discharge signal.

2. A method for generating a PRPD image using the PRPD image generation circuit according to claim 1, characterized in that, Includes the following steps: Step 1: The data generation unit creates an image matrix, the power frequency generation unit generates a power frequency synchronization signal, and the counter resets the phase count and starts counting after receiving the falling edge of the power frequency synchronization signal. Step 2: The peak hold unit enters the envelope tracking state, the first digital-to-analog converter initializes the reference voltage U2, and the peak hold unit acquires the partial discharge signal; Step 3: The comparator compares the partial discharge voltage U1 of the partial discharge signal with the reference voltage U2. If U1 is greater than λU2, proceed to step 4; otherwise, return to step 2. λ is the accuracy parameter. Step 4: The peak hold unit enters the peak detection state, the comparator outputs a rising edge of partial discharge to the first digital-to-analog converter, and the reference voltage U2 increases; Step 5: The peak hold unit re-acquires the partial discharge signal, and the comparator re-compares the partial discharge voltage U1 with the reference voltage U2. If U1 is greater than U2, return to step 4; otherwise, proceed to step 6. Step 6: The comparator outputs a falling edge of partial discharge to the first digital-to-analog converter and the counter. The first digital-to-analog converter outputs the current reference voltage U2, the counter outputs the current phase number, and the reference voltage U2 is reset. Step 7: Repeat steps 2 to 6 to generate multiple arrays containing reference voltage and phase number. Normalize the reference voltage and phase number of the arrays according to the number of rows and columns of the image matrix. Then generate the elements of the image matrix according to the normalized values ​​of reference voltage and phase number, and output the image matrix.

3. The PRPD image generation method according to claim 2, characterized in that, In step 4, the increment step size U of the reference voltage U2 is... step =0.001V to 0.005V, the image matrix is ​​a 32×32 two-dimensional matrix, the power frequency synchronization signal is a square wave signal with a period of 20ms, and the voltage of the power frequency synchronization signal is 3.3V.

4. The PRPD image generation method according to claim 2, characterized in that, In step 4, the peak hold unit receives the reference voltage U from the second digital-to-analog converter. ref The partial discharge voltage U1 = U0 - U ref U0 is the current voltage of the partial discharge signal.

5. The PRPD image generation method according to claim 2, characterized in that, In step 7, the image matrix is ​​an N x M matrix, based on the maximum reference voltage U. max Set the normalization ratio M / U max M is the number of columns in the image matrix, and the normalized value of the reference voltage U2 is m = Floor(U2M / U max )+1, Floor() is the floor function.

6. The PRPD image generation method according to claim 5, characterized in that, The number K of arrays where the normalized value of the number of phases is n and the normalized value of the reference voltage is m. nm Element P in the image matrix nm =(1-0.02K nm / T)×255, where T is the detection duration, n≤N, m≤M.

7. A method for identifying partial discharge types in the PRPD image generation method according to claim 2, characterized in that, Includes the following steps: Step 100: Generate at least two sets of partial discharge signals of partial discharge types, implement the PRPD image generation method, and generate multiple sets of image matrices; Step 200: Train a partial discharge classification model based on multiple sets of image matrices and their corresponding partial discharge types; Step 300: Generate an image matrix according to the PRPD image generation method, and identify the partial discharge type of the image matrix according to the partial discharge classification model.

8. The method for identifying partial discharge types according to claim 7, characterized in that, In step 100, a partial discharge signal is generated based on the gas-insulated switchgear. The partial discharge types include tip discharge, particle discharge, air gap discharge, and suspension discharge.

9. The method for identifying partial discharge types according to claim 7, characterized in that, The partial discharge classification model includes multiple sets of alternating hourglass-shaped attention layers and transition layers. The hourglass-shaped attention layer is composed of multiple stacked hourglass-shaped attention units, each with a lightweight convolutional block and an attention block for adjusting the feature weights of the image matrix.

Images