Electric pulse partial discharge detection method and system

Abstract

The invention discloses an electric pulse partial discharge detection method and system. The method comprises the following steps: firstly, carrying out system wiring and parameter initialization, matching RLC type detection impedance, and setting parameters such as a sampling rate and a filtering gear; standard pulses are injected through the calibration pulse generator to complete system calibration, the signal transmission proportion is determined, and the high-voltage power supply is started to collect partial discharge signals. Power frequency is positioned and eliminated through a frequency domain blanking algorithm, and a partial discharge characteristic frequency band is reserved through a frequency domain windowing algorithm, and out-of-band noise is shielded. The system is composed of a computer, a partial discharge host, an RLC type detection impedance, a calibration pulse generator and signal processing system software. The partial discharge host integrates a signal input unit, a conditioning unit, an A / D conversion unit and a central control unit to complete signal purification and digitization. And the RLC type detection impedance adopts a multi-model design. Through the synergistic effect of the frequency domain algorithm, the detection signal-to-noise ratio is improved, and a scientific basis can be provided for insulation state evaluation and operation and maintenance decision making of power equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of partial discharge detection of power equipment, and in particular to an electric pulse partial discharge detection method and system. BACKGROUND

[0002] Partial discharge is a key indicator for evaluating the insulation state of power equipment such as transformers, transformers, bushings, etc. Timely and accurate detection of partial discharge can effectively prevent equipment insulation breakdown failure and ensure safe operation of the power system. At present, the pulse current method is the mainstream technology for partial discharge detection, but its existing implementation scheme still has many limitations in interference suppression, system performance, operation adaptability and data management, etc., and it is difficult to meet the precise detection needs in complex scenarios. The main problems are:

[0003] (1) In terms of interference suppression, there are generally 50Hz power frequency and its harmonics, high-frequency electromagnetic radiation, mechanical noise and other interference signals in the detection environment of power equipment, and the existing technology lacks targeted frequency domain processing means. Traditional methods rely more on high-pass and low-pass filtering, which can only filter out interference in a fixed frequency band, and cannot accurately suppress specific interference such as power frequency harmonics and narrowband radio interference. At the same time, there is a lack of selective retention mechanism for the partial discharge characteristic frequency band, resulting in that the 40kHz-800kHz effective partial discharge signal is easily submerged by out-of-band noise, the signal-to-noise ratio is low, and it is difficult to capture weak discharge signals such as <0.1pC.

[0004] (2) The existing detection impedance is designed with fixed parameters, and the tuning capacitance range is narrow, which cannot cover the loop resonance capacitance requirements of different test objects, such as 25pF-400pF for oil-immersed transformers and 400pF-6000pF for small transformers, resulting in that the detection loop resonance frequency deviates from the measurement frequency band, the signal conversion efficiency is low, and the sensitivity is insufficient.

[0005] (3) The existing system calibration needs to manually adjust the gain and filter position repeatedly, and needs to be recalibrated after switching different detection channels or frequency bands, without forming a standardized process. The selection of calibration pulse parameters such as calibration capacitance and voltage depends on the experience of the operator. At the same time, the flexibility of the wiring mode is insufficient, and the traditional wiring cannot simultaneously measure the partial discharge signal and the test voltage, and when the test power supply and the instrument power supply are not in the same phase, there is a lack of reliable external zero mark adaptation scheme, the phase reference deviation is large, and the phase distribution characteristics of the partial discharge pulse cannot be accurately analyzed.

[0006] In summary, the existing partial discharge detection technology has deficiencies in interference suppression, system adaptability, calibration convenience, and needs an electric pulse partial discharge detection method and system with efficient frequency domain interference suppression, wide adaptability, standardized calibration and perfect data management. SUMMARY

[0007] In view of the above problems of the prior art, the present application proposes a technical scheme of electric pulse partial discharge detection by combining frequency domain signal processing technology, modular hardware design and intelligent software.

[0008] S1: Assembling the detection system, selecting a matching type of RLC detection impedance according to the capacitance of the test product, and setting the sampling rate, filter position and zero mark mode;

[0009] S2: Selecting a detection circuit wiring mode according to the detection requirement, wherein the wiring mode includes a wiring mode for simultaneously measuring partial discharge and test voltage and a wiring mode for only measuring partial discharge;

[0010] S3: System calibration, injecting a standard calibration pulse into the detection circuit through the calibration pulse generator, matching the gain and filter position, and recording the signal transmission ratio;

[0011] S4: Starting the high-voltage power supply to apply voltage to the test product, collecting partial discharge signals and performing pretreatment;

[0012] S5: Executing frequency domain blanking algorithm and frequency domain windowing algorithm processing on the pretreated signals to eliminate power frequency and harmonic interference and retain partial discharge characteristic frequency band signals;

[0013] S6: Extracting characteristic parameters of the processed signals and performing stability verification; the characteristic parameters include apparent discharge quantity and phase distribution;

[0014] S7: Outputting and storing the detection results and generating a detection report; the detection report includes a waveform diagram, a data table and an analysis conclusion.

[0015] Preferably, in S1, the tuning capacitance range of the RLC detection impedance needs to cover the detection circuit resonance capacitance Ct, which is calculated by the formula Ct=Cx×Ck / (Cx+Ck), wherein Cx is the test product capacitance and Ck is the coupling capacitance; the zero mark mode includes an internal zero mark mode and an external zero mark mode.

[0016] Preferably, in S2, the wiring mode for simultaneously measuring partial discharge and test voltage is that the upper end of the coupling capacitance C1 is connected to the high-voltage end of the test product, and the lower end is connected to the input end of the RLC detection impedance; the upper end of the coupling capacitance C2 is connected to the partial discharge host zero mark input interface, and the lower end is commonly grounded with the output end of the detection impedance; the wiring mode for only measuring partial discharge is that the upper end of the coupling capacitance C1 is connected to the high-voltage end of the test product, and the lower end is connected to the input end of the detection impedance, and the output end of the detection impedance is directly grounded.

[0017] Preferably, the S3 further comprises selecting a calibration capacitance Cq according to the test product capacitance Cx and the coupling capacitance Ck; the calibration capacitance Cq satisfies 10 pF≤Cq<0.1×(Cx+Ck); a standard calibration charge Q is calculated by the formula Q=Uq×Cq, wherein Uq is a calibration pulse voltage; during the calibration, the amplifier gain is adjusted to make the pulse amplitude reach 80% of the screen full scale.

[0018] Preferably, in the S5, the frequency domain blanking algorithm comprises:

[0019] S511: performing fast Fourier transform on the collected time domain signal array with a length of n to obtain complex frequency spectrum data with a length of 2n including real and imaginary parts, including n positive frequencies and n negative frequencies;

[0020] S512: setting a blanking target frequency point and a blanking bandwidth, determining a blanking range, calculating the actual frequency corresponding to each positive frequency, and if it falls within the blanking range, setting the corresponding complex frequency spectrum real and imaginary parts to zero; if harmonic interference needs to be eliminated, setting a frequency multiplication as the target frequency point and repeating the operation;

[0021] S513: performing inverse fast Fourier transform on the blanked complex frequency spectrum to restore the processed frequency spectrum to a time domain signal array.

[0022] Preferably, in the S5, the frequency domain windowing algorithm comprises:

[0023] S521: performing fast Fourier transform on the collected time domain signal array with a length of n to obtain complex frequency spectrum data with a length of 2n including real and imaginary parts, including n positive frequencies and n negative frequencies;

[0024] S522: setting an effective frequency band according to the high-frequency characteristics of the partial discharge signal, calculating the actual frequency corresponding to each positive frequency, and if it falls outside the effective frequency band, setting the complex frequency spectrum real and imaginary parts of the corresponding frequency point to zero;

[0025] S523: performing IFFT on the windowed complex frequency spectrum to restore it to a time domain signal array.

[0026] Preferably, in the S6, the apparent discharge quantity is calculated by comparing the processed partial discharge signal with the standard calibration charge; the phase distribution analysis takes the zero reference signal as the benchmark, judges the discharge type according to the pulse phase distribution characteristics; and the stability verification measures whether the apparent discharge quantity fluctuation within a period of time is less than a threshold value.

[0027] The second object of the present application provides an electric pulse partial discharge detection system, comprising a computer, a partial discharge host, an RLC type detection impedance, a calibration pulse generator and a signal processing system software; the computer is in communication connection with the partial discharge host, and the RLC type detection impedance is in electrical connection with the partial discharge host; the calibration pulse generator is used for injecting a standard pulse into a detection loop; the signal processing system software runs on the computer and realizes signal processing and analysis functions; the signal processing system software comprises a calibration module, a signal acquisition module, a frequency domain processing module, a feature analysis module and a data management module; the frequency domain processing module comprises a frequency domain blanking algorithm and a frequency domain windowing algorithm.

[0028] Preferably, the partial discharge host comprises a signal input unit, a signal conditioning unit, an AD conversion unit, a central control unit, a synchronization unit and a power supply system; the signal input unit comprises a double-channel input interface and an external zero mark input interface; the signal conditioning unit comprises a filtering module, an impedance conversion module, a program-controlled amplification module and a hardware filtering module; and the AD conversion unit comprises a high-speed AD converter and a buffer.

[0029] Preferably, the RLC type detection impedance adopts an RLC parallel structure, comprises a balanced impedance type and a general impedance type, and the tuning capacitance range covers 25pF-6000pF, which is suitable for different capacitive specifications of test objects; an equivalent circuit thereof is composed of an inductor Lm, a tuning capacitor Cm and a damping resistor Rm in parallel, and the detection loop resonance frequency falls within a 10kHz-500kHz measurement frequency band.

[0030] The present application has the following beneficial effects:

[0031] (1) The present application solves the problems of incomplete interference filtering and easy loss of effective signals in traditional detection techniques through the cooperative application of the patented frequency domain windowing algorithm and the frequency domain blanking algorithm. The frequency domain blanking algorithm can accurately locate and eliminate the common power frequency 50Hz and its harmonic interference in the power environment. The frequency domain windowing algorithm constructs a dedicated frequency domain window for the characteristic frequency band of the partial discharge signal, and only retains the effective signals in this interval. The two algorithms are used in cooperation, which can improve the signal-to-noise ratio of detection.

[0032] (2) The present application adopts an RLC type detection impedance, covers different test objects through multiple models of tuning capacitors, such as the loop resonance requirements of oil-immersed transformers, small transformers and bushings, etc., so that the detection loop resonance frequency stably falls within the measurement frequency band, and the signal conversion efficiency is improved. The calibration link adopts a standardized calibration pulse generator, and the calibration process can automatically match the gain and filter position through software, without the need for manual repeated debugging, so that the calibration error is low, and the accuracy of the apparent discharge quantity measurement is ensured.

[0033] (3) The application can adapt to various detection scenes and test product types through flexible design of wiring mode, zero mark mode and channel configuration.

[0034] The application simplifies the detection process, reduces the operation difficulty, effectively reduces the dependence on the experience of the operator, improves the field detection efficiency, and provides a scientific basis for the condition-based maintenance of power equipment. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 A partial discharge detection method step diagram for the embodiment of the application;

[0036] Figure 2 A partial discharge detection system structure schematic diagram for the embodiment of the application;

[0037] Figure 3 A partial discharge detection system host hardware block diagram for the embodiment of the application;

[0038] Figure 4 A partial discharge detection system detection impedance principle schematic diagram for the embodiment of the application;

[0039] Figure 5 A partial discharge detection system detection loop equivalent circuit schematic diagram for the embodiment of the application;

[0040] Figure 6 A wiring schematic diagram for measuring partial discharge and test voltage simultaneously in the method of the embodiment of the application;

[0041] Figure 7 A wiring schematic diagram for measuring only partial discharge in the method of the embodiment of the application. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0043] Embodiment one

[0044] The embodiment is based on the pulse current method, integrates RLC type detection impedance, frequency domain signal processing technology and system calibration mechanism, and realizes accurate detection of partial discharge signals of power equipment such as transformers, transformers and bushings, such as Figure 1 As shown in the figure, the method comprises the following steps:

[0045] S1: System assembly and parameter initialization. First, complete the detection system hardware assembly, the system core is composed of a computer, a partial discharge host, an RLC type detection impedance, a calibration pulse generator and a signal processing system software, and each component connection needs to ensure that the signal transmission has no loss. Among them, the RLC type detection impedance needs to be matched according to the capacitance of the test product, such as testing oil-immersed mutual inductor to select the detection impedance of the tuning capacitance 25pF-400pF, testing small transformer to select the detection impedance of 400pF-6000pF, its equivalent circuit is composed of inductance Lm, capacitance Cm and resistance Rm in parallel, and the tuning capacitance range needs to be confirmed in advance to cover the detection loop resonance capacitance Ct, Ct=Cx×Ck / (Cx+Ck), Cx is the capacitance of the test product, Ck is the coupling capacitance, to ensure the detection sensitivity.

[0046] Subsequently, the system parameter initialization is carried out, and the sampling rate can be set to 50MHz in the signal processing software to match the performance of the 12-bit high-speed A / D converter, and to ensure the capture of high-frequency partial discharge pulses. Turn on the 7KHz high-pass filter to filter out the power frequency low frequency interference basis. The hardware filter gear is preset according to the common partial discharge signal characteristic frequency band, for example, temporarily set to "high frequency 200KHz, low frequency 40KHz". Select the zero mark mode, if the test power supply and the instrument power supply are in phase, use the internal zero mark as the phase reference; if they are not in phase, connect the test power supply to the external zero mark input interface through the voltage divider according to the required voltage, to ensure the phase synchronization accuracy.

[0047] S2: Detection loop wiring. According to the detection requirements, one of the two wiring methods is selected, and the wiring process needs to ensure that the high-voltage end is insulated and grounded reliably.

[0048] (1) Simultaneous measurement of partial discharge and test voltage wiring. As shown in Figure 6 , according to the principle of coupling capacitor voltage divider, the upper end of coupling capacitor C1 is connected to the high-voltage end of the test product, and the lower end of C1 is connected to the input end of the RLC type detection impedance; the upper end of coupling capacitor C2 is connected to the "zero mark input" interface of the partial discharge host, and the lower end of C2 is connected to the output end of the detection impedance together. This wiring can synchronously obtain the test voltage signal and the partial discharge signal through the voltage divider characteristic, and the actual high-voltage value can be displayed by multiplying the voltage signal by the voltage divider ratio.

[0049] (2) Only partial discharge measurement wiring. As shown in Figure 7 , the coupling capacitor voltage divider loop is simplified, only C1 is kept as the coupling capacitor, the upper end of C1 is connected to the high-voltage end of the test product, and the lower end is connected to the input end of the detection impedance, the output end of the detection impedance is directly grounded, and the zero mark input interface does not need to be connected to C2, only the internal zero mark or the external zero mark is used to provide the phase reference alone.

[0050] After wiring, check the circuit to ensure that the high-voltage wire is short and thick to avoid corona, and that the test circuit has the smallest surrounding area to reduce interference and introduction. All metal parts such as the booster transformer shell and the detection impedance shell are firmly grounded, and there is no floating potential.

[0051] S3: System calibration. The purpose of calibration is to determine the signal transmission ratio of the detection circuit to ensure the accuracy of apparent discharge measurement. The operation is as follows:

[0052] S31: Turn off the high-voltage power supply of the test object, and place the output end of the calibration pulse generator close to the high-voltage end of the test object. This can shorten the output connection and reduce signal attenuation. According to the capacitance Cx of the test object and the coupling capacitance Ck, select the calibration capacitance Cq, which needs to satisfy 10pF≤Cq<0.1×(Cx+Ck). Set the calibration pulse voltage grade, such as 5.0V. At this time, the standard calibration charge Q=Uq×Cq. For example, when Cq=10pF and Uq=5.0V, Q=50pC.

[0053] S32: Enter the software "calibration interface", select the current detection channel such as CH-A, start the calibration pulse generator, and the working indicator light is on. Observe the pulse waveform in the interface. Adjust the amplifier gain and waveform amplification factor. The gain has four grades: -20dB, 0dB, 20dB, and 40dB, which can be selected to make the pulse amplitude reach 80% of the screen full scale, taking into account the precision and margin. If there are interference pulses covering the calibration pulse, select the area containing only the calibration pulse through the "window" function, and shield the interference.

[0054] S33: Click "Start calibration". The software automatically records the gain, filter grade, and calibration charge corresponding relationship of the current channel. After calibration, the interface displays "calibrated" and the calibration time. The calibration parameters will be used as the basis for subsequent measurements and cannot be changed arbitrarily.

[0055] S4: Partial discharge signal acquisition.

[0056] S41: Remove the calibration pulse generator, start the high-voltage power supply, and gradually apply voltage to the test object according to the test standard, from 0 to the rated test voltage, with a voltage rise rate ≤1kV / s. At the same time, start the signal acquisition function.

[0057] S42: The partial discharge pulse generated by the test object is coupled to the RLC type detection impedance through the coupling capacitance C1. The impedance converts the pulse current signal into a voltage signal. This signal is first filtered by a 7KHz high-pass filter to filter out 50Hz power frequency and lower frequency interference, and then matched to the host input impedance through impedance transformation. Subsequently, it enters the preamplifier, which amplifies according to the gain set during calibration, and then passes through hardware filtering to maintain the calibrated high frequency of 200KHz and low frequency of 40KHz grade, further purifying the signal.

[0058] S43: The amplified filtered analog signal is sent to a 12-bit 50MHz high-speed A / D converter, and after conversion to a digital signal, it is stored in the host 3M cache unit; at the same time, the zero marker signal (internal / external zero marker) is synchronously collected, used to mark the corresponding test voltage phase angle of each partial discharge pulse, and the collection duration is set according to the detection scene, usually 10-30 minutes in the field, and more than 1 hour in the laboratory, covering intermittent discharge.

[0059] S5: Frequency domain signal processing. Perform frequency domain processing on the cached digital signal, and sequentially remove specific interference through frequency domain blanking algorithm and retain valid signals through frequency domain windowing algorithm, including:

[0060] S51: Frequency domain blanking algorithm, used to locate and eliminate the spectral components of known interference frequency points, such as power frequency 50Hz and its 3rd and 5th harmonics, including the following steps:

[0061] (1) Time domain to frequency domain: Call the FFT module in the signal processing software to perform fast Fourier transform (FFT) on the collected time domain signal array to obtain complex spectrum data containing real and imaginary parts, with the real part corresponding to the signal amplitude and the imaginary part corresponding to the phase. Let the time domain signal length be n, and after FFT, a 2n-length array is generated, with the first n corresponding to positive frequency and the last n corresponding to negative frequency, and only the positive frequency part is actually analyzed.

[0062] (2) Interference frequency point positioning and blanking: Set the blanking target frequency point to 50Hz, and the blanking bandwidth to 4Hz, i.e. covering the range of 50Hz±2Hz, to ensure that the power frequency and its small frequency deviation are included. Traverse each frequency point k in the positive frequency interval, with the frequency calculation formula being freq=k×sampling rate / n, k=0 to n-1, and if freq falls within the interval "50Hz-2Hz" to "50Hz+2Hz", set the real and imaginary parts of the complex spectrum corresponding to this frequency point to 0 to achieve interference component zeroing. If harmonic interference needs to be eliminated, repeat this step, setting the target frequency point to 150Hz, 250Hz, etc., and the bandwidth to 4Hz.

[0063] (3) Frequency domain to time domain: Perform inverse fast Fourier transform (IFFT) on the blanked complex spectrum to restore the processed spectrum to a time domain signal array, which has removed the power frequency and harmonic interference and only retains the high-frequency partial discharge signal and a small amount of out-of-band noise.

[0064] S52: Frequency domain windowing algorithm, which retains the partial discharge characteristic frequency band by constructing a frequency domain window to filter out the characteristic frequency band of the partial discharge signal and shield out-of-band noise, including the following steps:

[0065] (1) Second FFT conversion: Perform FFT again on the time domain signal after frequency domain blanking processing to generate a new complex spectrum, ensuring accurate reflection of the frequency distribution of the remaining signal.

[0066] (2) Window function construction and spectrum screening: According to the high frequency characteristics of partial discharge signals, the effective frequency band is usually 40 kHz-800 kHz, and the lowest frequency fLow=40 kHz and the highest frequency fHigh=800 kHz of the reserved frequency band are set. Traverse each frequency point k in the positive frequency interval, calculate freq=k*sample rate / n, if freq<40 kHz or freq>800 kHz, set the real and imaginary parts of the corresponding frequency point to 0. If freq is in the interval of 40 kHz-800 kHz, the original spectrum data is reserved, that is, the effective frequency band is reserved by "windowing". The window function here is a rectangular window, which has the characteristics of clear boundary and no transition band, ensuring the complete reservation of partial discharge signals.

[0067] (3) Secondary IFFT conversion: Perform IFFT on the windowed complex spectrum to restore it to a time-domain signal array. At this time, the signal has removed power frequency interference and out-of-band noise, and only retains the pure partial discharge pulse signal, which can be used for subsequent feature analysis.

[0068] S6: Partial discharge signal feature extraction and analysis.

[0069] S61: Calculation of apparent discharge quantity. The software compares the partial discharge signal after frequency domain processing with the standard charge of 50pC calibrated in step three, and automatically calculates the apparent discharge quantity according to the proportional relationship between the signal amplitude and the calibration pulse amplitude. If the "windowing" function is turned on, such as selecting the 0-90° phase interval of the test voltage, the discharge quantity corresponding to the maximum pulse amplitude in the window is taken as the effective discharge quantity. If the window is not opened, the discharge quantity corresponding to the maximum pulse amplitude in the full cycle 360° is taken.

[0070] S62: Phase distribution analysis. Based on the zero reference signal, the phase distribution of partial discharge pulses in one cycle of test voltage is counted. If the pulses are concentrated near 270° phase, it is determined as corona discharge. If the pulses are scattered in the 0-90° and 180-270° phase intervals, it is determined as internal discharge of the test object, such as insulation defect discharge. At the same time, the rising edge of the pulse less than 60 ns, the falling edge greater than 100 μs and the pulse width are extracted to assist in judging the defect type. For example, the pulse width of bubble discharge is relatively narrow, and the pulse width of surface discharge is relatively wide.

[0071] S63: Stability verification: continuously monitor for 10-15 minutes, and observe the fluctuation range of apparent discharge quantity. If the fluctuation is ≤±5%, it is determined that the discharge is stable and the data is valid; if the fluctuation is too large, check the loop ground or filter position to eliminate interference and reacquire the data.

[0072] S7: Result output and data storage.

[0073] S71: Real-time display. The software interface displays the results in multiple modes, including an elliptical window showing the pulse phase-amplitude distribution, a sinusoidal window showing the synchronization relationship between the pulse and the test voltage, and a spectrum window showing the frequency distribution of the frequency domain processed signal. The local area can be enlarged for detailed observation through the zoom function.

[0074] S72: Data recording. Key parameters are recorded, including test voltage value calculated according to the divider ratio, apparent discharge peak and average value, discharge phase distribution statistics, and detection time, environmental temperature and humidity, detection impedance model, and filter position.

[0075] S73: Report generation and storage. The detection report is generated in a preset format, including waveform graphs, data tables, and analysis conclusions. The original collected time domain signal and frequency domain spectrum, processed signal data, and report are stored on the hard disk, supporting subsequent historical data query, waveform playback, and multi-batch detection result comparison, providing a basis for the insulation state evaluation of the test product.

[0076] After the detection is completed, the test voltage is gradually reduced to zero, the high-voltage power supply is turned off, the detection circuit is removed, and the equipment is arranged, completing the entire detection process.

[0077] The method of this embodiment uses the above steps, and through the synergistic effect of frequency domain algorithm, the signal-to-noise ratio is improved, the 0.1 pC level weak discharge signal can be captured, the apparent discharge measurement is accurate, and it is suitable for laboratory precision measurement and on-site detection in multiple scenes, and is suitable for various types of power equipment partial discharge detection.

[0078] Example Two

[0079] The electric pulse partial discharge detection system of this embodiment is designed based on the pulse current method, integrates hardware acquisition, signal conditioning, frequency domain processing, and software analysis functions, and can realize accurate acquisition, interference suppression, and feature analysis of power equipment partial discharge signals, and is suitable for laboratory precision measurement and on-site detection scenes. The system is composed of a computer, a partial discharge host, an RLC type detection impedance, a calibration pulse generator, and a signal processing system software, as shown in Figure 2 The technical details and working principles of each part are as follows:

[0080] I. Computer. It undertakes the functions of instruction issuance, data reception, result display, and storage management, and is the carrier of user operation and system control. Its functions include:

[0081] (1) Man-machine interaction control: Run the signal processing system software to provide a visual operation interface, support users to complete system parameter configuration such as sampling rate, filter gear, enter the calibration interface to adjust the gain and pulse parameters for calibration operation, and detect start / stop instruction input. The interface includes calibration interface, partial discharge home page, spectrum analysis, history record and other modules, which can switch to view data of different detection links in real time, including calibration pulse waveform, partial discharge phase distribution, frequency domain filter effect, etc.

[0082] (2) Data receiving and processing: Connect with the partial discharge host through the RJ45 network interface or dedicated communication line, receive the digitized partial discharge signal and synchronous zero mark signal transmitted by the host. Preliminary preprocessing is performed on the received data, such as data buffering and format conversion, to provide data basis for subsequent frequency domain processing and feature analysis.

[0083] (3) Result display and storage: Display the detection results in multiple forms, including elliptical window, sine window, and spectrum window, support waveform scaling and local windowing to view details. Automatically store the time domain signal, frequency domain spectrum of the original collected data, the pure signal data after processing with blanking / windowing, and the detection report including discharge quantity, phase statistics, and environmental parameters. The storage format is compatible with common document types, facilitating subsequent historical data query and multi-batch result comparison.

[0084] (4) Remote collaboration and expansion: Some models support wireless communication function, which can upload detection data to cloud server in real time for multi-terminal viewing. At the same time, it is compatible with external display to meet the needs of on-site multi-person collaborative analysis.

[0085] II. Partial discharge host, as shown in Figure 3 , is responsible for receiving partial discharge signals transmitted by detection impedance, completing signal conditioning, A / D conversion, synchronization control and data temporary storage, and is the key link connecting hardware collection and software analysis. The host integrates signal input unit, signal conditioning unit, A / D conversion unit, central control unit, synchronization unit and power system, and each unit works together to realize efficient signal processing.

[0086] (1) Signal input unit. The signal input unit is the entrance of partial discharge signal and synchronization signal, including A and B dual-channel input interface and external zero mark input interface, supporting single-channel independent detection or multi-channel synchronous acquisition to meet the needs of simultaneous detection of multiple test objects.

[0087] Channel characteristics: Each input channel is independent of each other, equipped with a dedicated signal interface, which can connect the output end of RLC type detection impedance. The channel input impedance is matched with the detection impedance, usually high impedance, to avoid attenuation or distortion during signal transmission.

[0088] Zero interface: support internal and external zero switching, external zero input range is ac 10-220v, 30hz-300hz, can access the test voltage signal after voltage reduction by voltage divider, provide accurate phase reference for partial discharge pulse. If the test power supply and instrument power supply are different phases, phase synchronization is realized through external zero.

[0089] (2) Signal conditioning unit. Responsible for purifying and amplifying the input raw partial discharge signal, eliminating low-frequency interference and adjusting the signal amplitude to the range suitable for A / D conversion, including high-pass filter module, impedance conversion module, programmable amplification module and hardware filter module.

[0090] High-pass filter module: built-in 7khz high-pass filter, can filter out power frequency 50hz and lower frequency interference signals such as power grid low-frequency noise, and preliminarily purify partial discharge high-frequency pulse signals.

[0091] Impedance conversion module: impedance matching is realized through transformer or special chip, high impedance signal output by detection impedance is converted into low impedance signal, reducing signal transmission loss.

[0092] Programmable amplification module: provides-20db, 0db, 20db, 40db multi-gain adjustment, supports software remote control switching, can adjust the amplification according to the strength of partial discharge signal, such as selecting 40db gain for weak signal and-20db gain for strong signal, at the same time, it has ±20db gain fine tuning function, ensuring that the signal amplitude is stable in the best interval of A / D conversion;

[0093] Hardware filter module: contains multiple frequency upper limit 100k, 200k, 300k, 800khz, and multiple frequency lower limit low-pass 10k, 20k, 40k, 80khz filter, can be arbitrarily combined to form different bandwidths, such as 40-200khz bandwidth, to further filter out out-of-band interference.

[0094] (3) A / D conversion unit. Responsible for converting the conditioned analog partial discharge signal into digital signal, providing digitized data for subsequent frequency domain processing, the core components are 12-bit high-speed A / D converter and 3m byte buffer. 12-bit A / D converter: sampling rate up to 50mhz, can accurately capture the high-frequency characteristics of partial discharge pulse, such as pulse with rise time less than 60ns, conversion accuracy meets the accuracy requirements of gb7354-2018 for partial discharge measurement, linearity error <5%±1pC. 3m byte buffer: each channel is equipped with independent buffer, can temporarily store the continuously collected digital signal, avoid data loss, at the same time, provide buffer for batch data reading of central control unit, ensure smooth data transmission.

[0095] (4) Central control unit. Composed of high-performance CPU and peripheral circuits, responsible for coordinating the work of each unit of the partial discharge host, realizing signal flow control and data interaction.

[0096] Instruction execution: receive the control instructions issued by the computer, such as sampling rate setting, filter position switching, calibration start, and drive the corresponding unit to execute the operation.

[0097] Data scheduling: control the sampling timing of the A / D conversion unit, read the converted digital signal to the cache in batches, and then transmit it to the computer through the computer interface.

[0098] State monitoring: real-time monitoring of the working state of each unit in the host, such as power voltage, A / D conversion, if abnormal, timely feedback alarm information to the computer.

[0099] (5) Synchronization unit. Provides a phase reference to ensure the phase synchronization of the partial discharge signal and the test voltage, and supports both internal and external zero mark modes.

[0100] Internal zero mark mode: when the test power supply and the instrument power supply are in phase, the 50Hz standard synchronization signal is generated by the internal crystal oscillator of the host, which is used as the phase reference of the partial discharge pulse.

[0101] External zero mark mode: when the test power supply and the instrument power supply are out of phase, the test voltage signal is received by the external zero mark input interface through the voltage divider, and the synchronization signal is extracted through isolation conversion and signal processing. After the effective value conversion, it is sent to the A / D conversion unit to realize the accurate phase alignment of the partial discharge signal and the test voltage, and provides the basis for subsequent phase analysis.

[0102] (6) Power supply system. Provides stable power supply for each unit of the partial discharge host, supports AC 220V±10% input, and outputs various DC voltages such as ±5V, ±12V through internal voltage stabilization and filtering circuit to meet the power supply needs of different units. At the same time, it has overvoltage and overcurrent protection function, when the input voltage is abnormal or the internal circuit is overloaded, it automatically cuts off the power supply to protect the safety of the host hardware.

[0103] Three, RLC type detection impedance. It is a partial discharge signal conversion element, responsible for converting the partial discharge pulse current generated by the test product into a measurable voltage signal. Its structure and parameters directly affect the detection sensitivity, and it is suitable for different capacitance test products such as transformers, transformers, and bushings, as shown in Figure 4 、 Figure 5 .

[0104] (1) Structure and equivalent circuit. The detection impedance adopts RLC parallel structure, and the equivalent circuit is composed of inductance Lm, capacitance Cm and resistance Rm in parallel; Cm is a tuning capacitor, which can be adjusted by a knob; Lm is a fixed inductance; Rm is a damping resistance to suppress the overvoltage during resonance. In practical applications, the tuning capacitor range needs to be selected according to the resonant capacitance Ct of the detection loop, Ct = Cx × Ck / (Cx + Ck), Cx is the capacitance of the test product, Ck is the coupling capacitance, so that the loop resonant frequency falls within the system measurement frequency band, 10 kHz-500 kHz, to ensure that the amplitude of the converted voltage signal is maximum and the detection sensitivity is improved.

[0105] (2) Type and parameter. The detection impedance is divided into balanced impedance and ordinary impedance, a total of 12 models. The tuning capacitor range and maximum current effective value of different models are different, and they are suitable for different test products.

[0106] Ordinary impedance: only has signal conversion function, the input end is connected with coupling capacitor Ck, and the output end is connected with the input channel of the partial discharge host, which is suitable for laboratory environment with less interference.

[0107] Balanced impedance: has signal conversion and anti-interference functions, supports balanced measurement loop, one input end is connected with the test product side signal, and the other input end is connected with the reference signal, which can suppress common mode interference. At the same time, it can be used as ordinary impedance, that is, single input and the other input end is connected with ground. Or cooperate with coupling capacitor voltage divider to realize synchronous measurement of partial discharge and test voltage, the positive and negative input ends are connected with the lower end of C1 and the upper end of C2 respectively.

[0108] Common model selection example: when testing oil-immersed current transformer, select No. 2 impedance, tuning capacitor 25pF-400pF, and the maximum current of unbalanced loop is 60mA. When testing small transformer, select No. 4 impedance, tuning capacitor 400pF-6000pF, and the maximum current of unbalanced loop is 250mA. The specific selection can refer to the capacitance of the test product and the resonant requirement of the detection loop.

[0109] (3) Working principle. When the test product generates partial discharge, the pulse current flows into the detection impedance through the coupling capacitor Ck. Since the detection impedance and Cx, Ck form a resonant loop, the resonant frequency is determined by Lm and Ct, and the pulse current excites resonance in the RLC parallel circuit, generating a high induced voltage across Rm. This voltage signal is the carrier of the partial discharge signal, which is transmitted to the signal input unit of the partial discharge host through the output end, completing the current-to-voltage conversion.

[0110] Four, calibration pulse generator. It is a system precision calibration element, which is used to generate standard partial discharge pulse signal, determine the signal transmission ratio of the detection loop, and ensure the accuracy of apparent discharge quantity measurement. Its performance meets the requirements of IEC 60270:2025 and GB7354-2018 standards.

[0111] (1) Structure and technical parameters. The generator is composed of a pulse generation unit, a calibration capacitor unit, a power supply unit, and an output interface. The core technical parameters are as follows:

[0112] Pulse characteristics: The output standard pulse rising edge is <60ns, and the falling edge is >100μs, which meets the high-frequency transient characteristics of partial discharge pulses. The pulse repetition frequency is fixed at 1000Hz to ensure the stability of the calibration signal.

[0113] Voltage and capacitance range: The pulse voltage range is divided into 0.5V, 1.0V, 2.0V, and 5.0V, and the calibration capacitor range is divided into 10pF and 100pF. The standard calibration charge can be generated by combining 5pC (0.5V x 10pF), 10pC (1.0V x 10pF), 100pC (1.0V x 100pF), and 500pC (5.0V x 100pF). Q=Uq x Cq, Uq is the pulse voltage, and Cq is the calibration capacitor.

[0114] Output and protection: The output resistance is <100Ω, which reduces signal transmission attenuation. The output port is protected by a metal ceramic discharge tube to prevent high voltage from damaging the internal circuit. The power supply unit uses a built-in 9V battery for power supply, with a timing shutdown function. It automatically shuts down after a few minutes of idling, extending the battery life.

[0115] (2) Working process. During calibration, the generator output end is close to the high voltage end of the test product, the output line is shortened, and the signal loss is reduced. Press and hold the power key for 3 seconds to start the generator. The working indicator light is on. Select the target voltage and capacitance range by pressing the selection button. The generator's internal pulse generation unit generates a high-frequency pulse corresponding to the voltage. After coupling through the calibration capacitor unit, the standard calibration pulse is output to the test product. The partial discharge host collects the standard pulse and adjusts the gain and filter range through software to make the pulse amplitude reach 80% of the screen full scale. Record the signal transmission ratio at this time to complete the system calibration. After calibration, the generator should be turned off and removed to avoid damaging the equipment during pressure testing.

[0116] Five, signal processing system software, running on a computer, integrating signal acquisition, frequency domain processing, feature analysis, and data management functions, based on patented frequency domain windowing and blanking algorithms to achieve interference suppression, and output accurate detection results. The software contains five core modules:

[0117] (1) Calibration module, responsible for completing system calibration to ensure measurement accuracy, including:

[0118] 1. Calibration parameter configuration: supports selecting detection channels such as CH-A and CH-B, calibration pulse range 5pC-1000pC, and frequency upper and lower limit combination filter range.

[0119] 2. Pulse adjustment and confirmation: real-time display of calibration pulse waveform, support adjustment of gain and amplification, make pulse amplitude stable. Provide "window" function to shield interference pulse, only keep standard pulse. After clicking "start calibration", automatically record the gain, filter and calibration charge corresponding relationship of the current channel, display "calibrated" state and calibration time, calibration parameters as subsequent measurement reference.

[0120] (2) Signal acquisition module. Control the local discharge host to complete the acquisition of local discharge signal and synchronization signal, the functions include:

[0121] 1. Acquisition parameter setting: set the sampling rate, default 50MHz, acquisition time 10-60 minutes, synchronization mode of internal zero or external zero;

[0122] 2. Real-time data receiving: receive the digitized local discharge signal and zero signal transmitted by the host through the communication interface, temporarily stored in the local cache, ensure continuous data without loss.

[0123] 3. Acquisition state monitoring: real-time display of acquisition progress, signal amplitude, and whether overload, if signal abnormality occurs, such as amplitude exceeding the range, automatically prompt to adjust gain or filter position.

[0124] (3) Frequency domain processing module. Based on frequency domain blanking and window algorithm, realize interference suppression and effective signal extraction.

[0125] (4) Feature analysis module, extract local discharge signal features, judge the insulation state of the test product, the functions include apparent discharge quantity calculation, phase distribution analysis, stability verification.

[0126] (5) Data management module, realize the display, storage and query of detection data.

[0127] The system of the embodiment realizes the whole process closed loop of signal acquisition, conditioning, processing, analysis and output through the cooperative work of the above several component parts, has the characteristics of strong anti-interference ability, high detection precision and convenient operation, and can meet the partial discharge detection needs of different power equipment.

[0128] The above is only the preferred embodiment of the present application, it should be pointed out that for ordinary skilled in the art, without departing from the technical principles of the present application, can make several improvements and refinements, these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A method of electrical pulse partial discharge detection, characterized by, The method comprises the following steps: S1: Assemble the detection system, select the RLC type detection impedance according to the capacitance of the test product, and set the sampling rate, filter position and zero mark mode; S2: Select the detection circuit wiring mode according to the detection requirement, the wiring mode includes the wiring mode of measuring partial discharge and test voltage at the same time and the wiring mode of measuring partial discharge only; S3: System calibration, inject standard calibration pulse into the detection circuit through the calibration pulse generator, match the gain and filter position, and record the signal transmission ratio; S4: Start the high-voltage power supply to apply voltage to the test product, collect the partial discharge signal and perform pretreatment; S5: Perform frequency domain blanking algorithm and frequency domain windowing algorithm on the pretreated signal to eliminate power frequency and harmonic interference and retain partial discharge characteristic frequency band signal; S6: Extract the characteristic parameters of the processed signal and verify the stability; the characteristic parameters include apparent discharge quantity and phase distribution; S7: Output and store the detection result and generate a detection report; the detection report includes waveform diagram, data table and analysis conclusion.

2. The method of claim 1, wherein, In S1, the tuning capacitance range of the RLC type detection impedance needs to cover the detection circuit resonance capacitance Ct, which is calculated by the formula Ct=Cx×Ck / (Cx+Ck), wherein Cx is the capacitance of the test product and Ck is the coupling capacitance; the zero mark mode includes internal zero mark mode and external zero mark mode.

3. The method of claim 1, wherein, In S2, the wiring mode of measuring partial discharge and test voltage at the same time is that the upper end of the coupling capacitance C1 is connected to the high-voltage end of the test product, and the lower end is connected to the input end of the RLC type detection impedance; the upper end of the coupling capacitance C2 is connected to the partial discharge host zero mark input interface, and the lower end is commonly grounded with the output end of the detection impedance; the wiring mode of measuring partial discharge only is that the upper end of the coupling capacitance C1 is connected to the high-voltage end of the test product, the lower end is connected to the input end of the detection impedance, and the output end of the detection impedance is directly grounded.

4. The method of claim 1, wherein, S3 also includes selecting calibration capacitance Cq according to the capacitance Cx of the test product and the coupling capacitance Ck; the calibration capacitance Cq satisfies 10pF≤Cq<0.1×(Cx+Ck); the standard calibration charge Q is calculated by the formula Q=Uq×Cq, wherein Uq is the calibration pulse voltage; During calibration, adjust the amplifier gain to make the pulse amplitude reach 80% of the screen full scale.

5. The method of claim 1, wherein, In S5, the frequency domain blanking algorithm includes: S511: Perform fast Fourier transform on the collected time domain signal array with a length of n to obtain complex frequency spectrum data with a length of 2n including real part and imaginary part, including n positive frequencies and n negative frequencies; S512: Set the blanking target frequency point and the blanking bandwidth, determine the blanking range, calculate the actual frequency corresponding to each positive frequency, and if it falls within the blanking range, set the corresponding complex frequency spectrum real part and imaginary part to zero; if harmonic interference needs to be eliminated, set the frequency multiplication as the target frequency point and repeat the operation; S513: Perform inverse fast Fourier transform on the blanked complex frequency spectrum to restore the processed frequency spectrum to a time domain signal array.

6. The method of claim 1, wherein, In S5, the frequency domain windowing algorithm includes: S521: Perform fast Fourier transform on the collected time domain signal array with a length of n to obtain complex frequency spectrum data with a length of 2n including real part and imaginary part, including n positive frequencies and n negative frequencies; S522: According to the high-frequency characteristics of the partial discharge signal, set the effective frequency band, calculate the actual frequency corresponding to each positive frequency, and if it falls outside the effective frequency band, set the real part and imaginary part of the complex spectrum corresponding to the frequency point to zero; S523: Perform IFFT on the windowed complex spectrum to restore the time-domain signal array.

7. The method of claim 1, wherein, In the S6, the apparent discharge quantity is calculated by comparing the processed partial discharge signal with the standard calibration charge; the phase distribution analysis takes the zero reference signal as the benchmark, and the discharge type is determined according to the pulse phase distribution characteristics; The stability verification measures whether the apparent discharge quantity fluctuation within a period of time is less than a threshold value.

8. An electrical pulse partial discharge detection system characterized by, The computer, the partial discharge host, the RLC detection impedance, the calibration pulse generator and the signal processing system software; the computer and the partial discharge host are in communication connection, and the RLC detection impedance and the partial discharge host are in electrical connection; the calibration pulse generator is used to inject standard pulses into the detection loop; the signal processing system software runs on the computer and realizes the functions of signal processing and analysis; the signal processing system software includes a calibration module, a signal acquisition module, a frequency domain processing module, a feature analysis module and a data management module; the frequency domain processing module includes a frequency domain blanking algorithm and a frequency domain windowing algorithm.

9. The electrical pulse partial discharge detection system of claim 8, wherein, The partial discharge host includes a signal input unit, a signal conditioning unit, an AD conversion unit, a central control unit, a synchronization unit and a power supply system; the signal input unit includes a double-channel input interface and an external zero reference input interface; the signal conditioning unit includes a filter module, an impedance conversion module, a programmable amplification module and a hardware filter module; the AD conversion unit includes a high-speed AD converter and a buffer.

10. The electrical pulse partial discharge detection system of claim 8, wherein, The RLC detection impedance adopts an RLC parallel structure, including a balanced impedance type and a general impedance type, and the tuning capacitor range covers 25pF-6000pF, which is suitable for different capacitor specifications of the test object; the equivalent circuit thereof is composed of an inductor Lm, a tuning capacitor Cm and a damping resistor Rm in parallel, and the detection loop resonance frequency falls within the 10kHz-500kHz measurement frequency band.

Images