High-altitude partial discharge detector and partial discharge analysis method thereof

Abstract

The invention discloses a high-altitude partial discharge detector and a partial discharge analysis method thereof.The detector comprises a double-hemisphere metal cover, an electromagnetic pulse sensor, an acoustic array sensor, a signal processing unit and a power source, and the electromagnetic pulse sensor and the acoustic array sensor are arranged in the double-hemisphere metal cover; the electromagnetic pulse sensor is connected with the acoustic array sensor, the signal processing unit is respectively connected with the electromagnetic pulse sensor and the acoustic array sensor, and the power supply is respectively connected with the electromagnetic pulse sensor, the acoustic array sensor and the signal processing unit; the analysis method comprises the following steps: installing a detector and collecting a signal, calculating a signal difference value and detecting an instantaneous condition, extracting a high-frequency component and analyzing a frequency spectrum, calculating the intensity of partial discharge and analyzing a fault type, and outputting a result and alarming. According to the invention, through combination of electromagnetic and acoustic signals, filtering and data fusion processing and accurate spectrum analysis, powerful support is provided for state monitoring and fault diagnosis of high-altitude power equipment.

Description

Technical Field

[0001] The invention belongs to the technical field of partial discharge detection, and in particular relates to a high-altitude partial discharge detector and a partial discharge analysis method thereof. Background Art

[0002] Partial discharge is an important status indicator to measure the monitoring status of high-voltage equipment in operation. Partial discharge is usually detected by telemetry and direct measurement. Inevitably, there is a technical pain point of noise filtering. In terms of how to improve the signal-to-noise ratio, researchers and developers have made various optimizations in terms of sensors, whole machine filters, software algorithms, etc. However, in recent years, with the continuous improvement of high-altitude operation technology, local discharge detection at high altitude has become a new application method. Its unique advantages are: compared with ground telemetry, the probe is closer to the high-altitude high-voltage equipment, and the theoretical signal-to-noise ratio is better. However, through application and demonstration, it is difficult to distinguish partial discharge signals from high-altitude power equipment due to large current, high magnetic field strength, and strong stray electric field. Whether it is electrical method or acoustic method, the electric field, electromagnetic, and electroacoustic noise are very large, and it is difficult to achieve the ideal effect. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a high-altitude partial discharge detector and a partial discharge analysis method thereof which have high detection accuracy, strong anti-interference capability and can improve the accuracy of partial discharge analysis.

[0004] The technical solution of the present invention is:

[0005] A high-altitude partial discharge detector comprises: a double-hemispherical metal cover, an electromagnetic pulse sensor, an acoustic array sensor, a signal processing unit and a power supply. The double-hemispherical metal cover is composed of two convex surfaces of the hemispherical metal covers connected back to back, and the openings of the two hemispherical metal covers are opposite. One end of the electromagnetic pulse sensor is connected to the center position of the concave surface of the hemispherical metal cover. The electromagnetic pulse sensor is composed of a plurality of coils or capacitors connected. The electromagnetic pulse sensor comprises at least three channels, and the probe of each channel is arranged along the spherical concentric axis direction of the hemispherical metal cover. The acoustic array sensor is a hollow annular structure and comprises a plurality of acoustic sensors. The other end of the electromagnetic pulse sensor is connected to the center position of the acoustic array sensor. The signal processing unit is respectively connected to the electromagnetic pulse sensor and the acoustic array sensor. The power supply is respectively connected to the electromagnetic pulse sensor, the acoustic array sensor and the signal processing unit.

[0006] Furthermore, the double-hemispherical metal cover serves as a signal ground terminal GND of the electromagnetic pulse sensor and the acoustic array sensor.

[0007] Furthermore, the probes of the channels of the electromagnetic pulse sensor are arranged at different distances in the hemispherical metal cover.

[0008] Furthermore, the acoustic sensor is composed of a plurality of acoustic microphone arrays connected together.

[0009] Furthermore, the signal processing unit comprises: a filtering module, a data fusion module and a partial discharge analysis module, and the filtering module, the data fusion module and the partial discharge analysis module are sequentially connected to form the signal processing unit.

[0010] Furthermore, the power supply includes: a rechargeable battery and a solar power supply module, and the rechargeable battery and the solar power supply module are connected.

[0011] A partial discharge analysis method for a high-altitude partial discharge detector comprises the following steps:

[0012] Step 1: Detector installation and signal acquisition, including the following steps:

[0013] S1. Install detectors: Install at least three electromagnetic pulse detectors inside the hemispherical metal cover. The three detectors are located on the same axis but at different distances from the center of the sphere. The corresponding electric field strength E 1 , the electric field strength E corresponding to detector 2 2 , the corresponding electric field strength E of detector 3 3 ;

[0014] S2. Electric signal acquisition: The electric field strength signals collected by the three detectors are respectively recorded as E 1 (t), E 2 (t) and E 3 (t), the expression of the electric field strength signal collected by the three detectors is:

[0015]

[0016] Where Q(t) represents the instantaneous charge of the discharge source at time t, ∈ 0 is the dielectric constant in free space, d 1 Indicates the distance between detector 1 and the discharge source, d 2 Indicates the distance between detector 2 and the discharge source, d 3 Indicates the distance between detector 3 and the discharge source;

[0017] Step 2, calculation of signal difference and instantaneous condition detection, includes the following steps:

[0018] S3, calculating the electric field strength difference, the calculation formula of the electric field strength difference is:

[0019] S 1 (t) = E 1 (t)-E 2 (t)

[0020] S2 (t) = E 2 (t)-E 3 (t)

[0021] S 3 (t) = E 1 (t)-E 3 (t),

[0022] Among them, E 1 (t) represents the electric field strength signal collected by the detector at time t, E 2 (t) represents the electric field strength signal collected by the detector 2 at time t, E 3 (t) represents the electric field strength signal collected by the detector 3 at time t, S 1 (t) represents the difference in electric field strength between detector 1 and detector 2 at time t, S 2 (t) represents the difference in electric field strength between detector 2 and detector 3 at time t, S 3 (t) represents the difference in electric field intensity between detector 1 and detector 3 at time t;

[0023] S4. Detecting unstable conditions: Compare S 1 (t) and S 2 (t) is used to detect the instantaneous electric field intensity fluctuation. When there is partial discharge, the electric field is unstable, that is:

[0024] ΔS(t)=S 1 (t)-S 2 (t)≠0,

[0025] Among them, ΔS(t) represents the fluctuation of electric field intensity;

[0026] Step 3, high frequency component extraction and spectrum analysis, including the following steps:

[0027] S5. Extract high frequency components: 1 (t), S 2 (t) or ΔS(t) is high-pass filtered and the high-frequency part of the signal is extracted using wavelet transform, that is:

[0028] S 1，high (t) = HighPass(S 1 (t))

[0029] S 2，high (t) = HighPass(S 2 (t))

[0030] ΔS high (t) = HighPass(ΔS(t)),

[0031] Among them, S 1，high (t) represents the signal S 1 (t) The high-frequency part obtained after high-pass filtering, S 2，high (t) represents the signal S 2 (t) The high-frequency part obtained after high-pass filtering, ΔS high (t) represents the high-frequency part obtained after high-pass filtering of ΔS(t);

[0032] S6. Spectrum analysis: Perform spectrum analysis on the extracted high-frequency components to obtain the spectrum characteristics of the partial discharge signal, and use Fourier transform to obtain the spectrum representation of the signal;

[0033] Step 4: Calculation of the intensity of partial discharge and analysis of the fault type, including the following steps:

[0034] S7. Calculate the intensity of partial discharge: Calculate the intensity of partial discharge according to the amplitude and energy characteristics of the high-frequency component. The higher the energy, the greater the intensity of partial discharge. The calculation formula of the intensity of partial discharge is:

[0035] Energy=∫|F(ΔS high (f))| 2 df,

[0036] Energy represents the total energy value obtained after spectrum analysis. The greater the energy, the greater the local discharge intensity. high (t) represents the high-frequency part obtained by high-pass filtering the electric field intensity fluctuation, F(ΔS high (f)) represents the change of ΔS high (t) The result after Fourier transformation;

[0037] S8. Fault type analysis: Classify the partial discharge type according to the frequency characteristics of the high-frequency components and the waveform of the discharge signal;

[0038] Step 5: Result output and alarm system, including the following steps:

[0039] S9, result output: output the calculated discharge intensity and analyzed fault type to the monitoring system;

[0040] S10, alarm system: set a threshold for discharge intensity. Once the discharge intensity is detected to exceed the set threshold, an alarm is triggered to remind maintenance personnel to check and handle it.

[0041] Furthermore, the specific steps of extracting the high frequency part of the signal using wavelet transform in step S5 are:

[0042] 1. Obtaining time domain signal: From the time domain signal S 1 (t), S2 (t) or ΔS(t) to obtain the discrete sequence S 1 [n], S 2 [n] or ΔS[n], the sampling expression of the time domain signal is:

[0043] S 1 [n] = S 1 (n·Δt), n=0, 1, 2,..., N-1,

[0044] S 2 [n] = S 2 (n·Δt), n=0, 1, 2,..., N-1,

[0045] ΔS[n]=ΔS(n·Δt), n=0, 1, 2,..., N-1,

[0046]

[0047] f s ≥2·f max ,

[0048]

[0049] Among them, Δt represents the time interval, N represents the number of sampling points, and f s represents the sampling frequency, f max represents the highest frequency of the signal, Δf represents the frequency resolution, and n represents the position of the sampling point in the sequence;

[0050] 2. Perform wavelet decomposition on the signal: Use discrete wavelet transform to decompose the signal into components of different scales. The expression of the wavelet decomposition is:

[0051] S 1 [n] → {A J , D J , D J-1 , ..., D 1},

[0052]

[0053] Among them, J represents the number of decomposition levels, A J represents the approximation coefficient of the Jth layer, D J represents the detail coefficient of the Jth layer, S j [k] represents the input signal of the jth layer, g[k] represents the low-pass filter coefficient, h[k] represents the high-pass filter coefficient, A j [n] represents the approximation coefficient of the jth layer, D j [n] represents the detail coefficient of the jth layer;

[0054] 3. Extract high-frequency components: The detail coefficients after wavelet decomposition correspond to the high-frequency part of the signal;

[0055] 4. Reconstruct high-frequency signal: Use inverse wavelet transform to reconstruct the high-frequency detail coefficients back to the time domain to obtain the time domain signal of the high-frequency part. The expression of the time domain signal of the high-frequency part is:

[0056] S 1，high [n] = IDWT(D j ),

[0057]

[0058] Among them, D j [n] and D j Both represent the detail coefficients of the jth layer, S j [k] represents the input signal of the jth layer, h[k] represents the high-pass filter coefficient, S 1,high [n] represents the extracted signal S 1 The high frequency part of [n];

[0059] Steps 2 to 4 also apply to S 2 [n] and ΔS[n], and get S 2,high [n] and ΔS high [n].

[0060] Furthermore, the spectrum analysis in step S6 includes the following steps:

[0061] 1. Obtaining a high-frequency signal: Obtaining a high-frequency signal S through step S5 1,high [n], S 2,high [n] and ΔS high [n], the above sequence contains N sampling points, and the sampling frequency is f s ;

[0062] Second, perform fast Fourier transform on the high-frequency signal to convert the time domain signal into a frequency domain signal. The expression of the fast Fourier transform is:

[0063]

[0064] Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔS high )[k] represents the high frequency signal ΔS high[n] represents the frequency domain after fast Fourier transform, k represents the frequency index, and n represents the sampling point index of the time series;

[0065] 3. Calculate frequency resolution and frequency axis: The expression of frequency resolution is:

[0066]

[0067] Where N represents the number of sampling points, f s represents the sampling frequency, Δf represents the frequency resolution;

[0068] The expression of the frequency axis is:

[0069] f k =k·Δf,k=0,1,2,...,N / 2,

[0070] Among them, f k represents the actual frequency corresponding to the frequency index k, Δf represents the frequency resolution, and N represents the number of sampling points;

[0071] Fourth, calculate the amplitude spectrum, take the absolute value of the complex structure of the frequency domain signal, and obtain the amplitude of each frequency component:

[0072]

[0073] Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔS high )[k] represents the high frequency signal ΔS high [n] represents the frequency domain after fast Fourier transform, k represents the frequency index, represents the real part of the frequency domain signal, represents the imaginary part of the frequency domain signal, N represents the number of sampling points, and S[n] represents the discrete sequence of the input signal;

[0074] 5. Calculate the power spectrum density. The calculation formula of the power spectrum density is:

[0075]

[0076] Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔShigh )[k] represents the high frequency signal ΔS high [n] represents the frequency domain after fast Fourier transform, k represents the frequency index;

[0077] 6. Analyze spectrum characteristics, including:

[0078] Spectrum feature extraction, the expression of the spectrum feature extraction is:

[0079]

[0080] Δf band =f upper -f lower ,

[0081] Among them, f peak Indicates the peak frequency in the spectrum, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, f band Indicates the frequency bandwidth, f upper Indicates the upper frequency where the power spectrum density is greater than the threshold, f lower Indicates the lower frequency where the power spectrum density is less than the threshold;

[0082] Discharge intensity estimation, calculate the total energy of the signal in the high frequency range, and estimate the total energy of the partial discharge signal by integrating the spectrum or accumulating the power spectrum density. The calculation formula for the total energy of the signal in the high frequency range is:

[0083]

[0084] Among them, Energy high Indicates the total energy of the signal in the high frequency range, PSD (S 1,high )[k] represents the high frequency signal S 1,high [n] Power spectral density at frequency index k, Δf represents the frequency resolution, k cutoff Indicates the frequency index corresponding to the cutoff frequency of the high-pass filter, and N represents the number of sampling points;

[0085] 7. Result analysis and output: According to the spectrum analysis results, determine the characteristics of the partial discharge signal, classify these characteristics, and determine the type and severity of the partial discharge.

[0086] Beneficial effects of the present invention:

[0087] 1. Improve the detection accuracy of partial discharge: The present invention adopts a combination of electromagnetic pulse sensor and acoustic array sensor, which can simultaneously detect the electromagnetic signal and acoustic signal generated by partial discharge, and enhance the ability to capture partial discharge signals; electromagnetic signal can quickly respond to the instantaneous change of partial discharge, while acoustic signal can enhance the detection effect in the case of large discharge. The combination of the two signal sources improves the sensitivity and detection accuracy of partial discharge events;

[0088] 2. Enhanced anti-interference capability: The double hemispherical metal covers of the present invention are connected back to back to form a closed structure. This design not only provides mechanical protection for the sensor, but also effectively shields external electromagnetic interference and reduces the impact of environmental noise on the sensor. The metal cover serves as the common signal ground (GND) of the electromagnetic pulse sensor and the acoustic array sensor, which can ensure that the sensor signals work at the same reference potential, which helps to reduce common-mode interference during signal transmission.

[0089] 3. Improve the effectiveness of signal processing: The electromagnetic pulse sensor of the present invention is composed of at least three channels, and the probes are arranged along the concentric axis direction of the sphere, and each channel is at a different distance, so that the difference in electric field strength can be analyzed; by analyzing the difference in electric field strength between the detectors, the abnormal changes caused by partial discharge can be further identified, and the accuracy of the detection signal can be improved; in the signal processing unit, a high-pass filter is used for processing, which can effectively remove low-frequency noise and only retain high-frequency components related to partial discharge events;

[0090] 4. Improve the accuracy of partial discharge analysis: Fast Fourier transform and power spectrum density calculation of high-frequency components can analyze the frequency characteristics of partial discharge signals and evaluate the intensity of discharge based on the energy of the high-frequency part; by analyzing the signal energy in different frequency ranges, different types of partial discharge phenomena can be distinguished, such as corona discharge, flashover discharge, etc., thereby providing more detailed diagnostic information for equipment maintenance;

[0091] 5. Improve the adaptability of high-altitude operations: The design of the double-hemispherical metal cover of the present invention enables the detector to have better resistance to wind pressure and mechanical impact in the high-altitude environment, ensuring that the equipment can operate stably for a long time in the harsh high-altitude environment; the rechargeable battery and solar power supply module ensure that the detector can work continuously at high altitude without relying on external power supply, and is suitable for long-term unattended high-altitude equipment monitoring;

[0092] 6. Real-time monitoring and alarm: Through the collaborative work of the filtering module, data fusion module and partial discharge analysis module, the detector can process high-altitude partial discharge signals in real time. Once abnormal discharge intensity or spectrum characteristics are found, the alarm is immediately triggered to remind maintenance personnel to conduct inspections; through the synchronous monitoring of electromagnetic and acoustic signals, combined with the results of real-time analysis, it can quickly respond to changes in the status of the equipment, and

[0093] Discover and locate the fault source in time to avoid long-term damage to the equipment caused by partial discharge problems;

[0094] 7. Improve the stability and reliability of discharge signal detection: The present invention uses electromagnetic pulse sensors with multiple channels. The probes detect signals at different distances, which can more comprehensively capture the changes in electric field strength during the discharge process, help smooth out local noise fluctuations, and improve measurement stability. The electric field difference relationship between detectors is used for analysis to quantify abnormal fluctuations in the electric field, thereby improving the reliability and stability of discharge signal detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 It is a structural schematic diagram of a high-altitude partial discharge detector of the present invention.

[0096] Figure 2 The present invention is a schematic structural diagram of an acoustic array sensor of a high-altitude partial discharge detector.

[0097] In the figure: 1- double hemispherical metal cover, 2- electromagnetic pulse sensor, 3- acoustic array sensor. DETAILED DESCRIPTION

[0098] like Figure 1 As shown, a high-altitude partial discharge detector includes: a double-hemispherical metal cover 1, an electromagnetic pulse sensor 2, an acoustic array sensor 3, a signal processing unit and a power supply. The double-hemispherical metal cover 1 is composed of two convex surfaces of the hemispherical metal covers connected back to back, and the openings of the two hemispherical metal covers are opposite. One end of the electromagnetic pulse sensor 2 is connected to the center position of the concave surface of the hemispherical metal cover. The electromagnetic pulse sensor 2 is composed of a plurality of coils or capacitors connected. The electromagnetic pulse sensor 3 includes at least three channels, and the probe of each channel is arranged along the spherical concentric axis direction of the hemispherical metal cover. The acoustic array sensor 3 is a hollow annular structure and includes a plurality of acoustic sensors. The other end of the electromagnetic pulse sensor 2 is connected to the center position of the acoustic array sensor 3. The signal processing unit is respectively connected to the electromagnetic pulse sensor 2 and the acoustic array sensor 3, and the power supply is respectively connected to the electromagnetic pulse sensor 2, the acoustic array sensor 3 and the signal processing unit.

[0099] Preferably, the double-hemispherical metal cover 1 serves as a signal ground terminal GND of the electromagnetic pulse sensor 2 and the acoustic array sensor 3 .

[0100] Preferably, the probes of the channels of the electromagnetic pulse sensor 2 are arranged at different distances in the hemispherical metal cover.

[0101] Preferably, the acoustic sensor is composed of a plurality of acoustic microphone arrays connected together.

[0102] Preferably, the signal processing unit comprises: a filtering module, a data fusion module and a partial discharge analysis module, and the filtering module, the data fusion module and the partial discharge analysis module are sequentially connected to form the signal processing unit.

[0103] Preferably, the power source comprises: a rechargeable battery and a solar power supply module, and the rechargeable battery and the solar power supply module are connected.

[0104] A partial discharge analysis method for a high-altitude partial discharge detector comprises the following steps:

[0105] Step 1: Detector installation and signal acquisition, including the following steps:

[0106] S1. Install detectors: Install at least three electromagnetic pulse detectors inside the hemispherical metal cover. The three detectors are located on the same axis but at different distances from the center of the sphere. The corresponding electric field strength E 1 , the electric field strength E corresponding to detector 2 2 , the corresponding electric field strength E of detector 3 3 ;

[0107] S2. Electric signal acquisition: The electric field strength signals collected by the three detectors are respectively recorded as E 1 (t), E 2 (t) and E 3 (t), the expression of the electric field strength signal collected by the three detectors is:

[0108]

[0109] Where Q(t) represents the instantaneous charge of the discharge source at time t, ∈ 0 is the dielectric constant in free space, d 1 Indicates the distance between detector 1 and the discharge source, d 2 Indicates the distance between detector 2 and the discharge source, d 3 Indicates the distance between detector 3 and the discharge source;

[0110] Step 2, calculation of signal difference and instantaneous condition detection, includes the following steps:

[0111] S3, calculating the electric field strength difference, the calculation formula of the electric field strength difference is:

[0112] S 1 (t) = E 1 (t)-E 2 (t)

[0113] S 2 (t) = E 2 (t)-E 3 (t)

[0114] S 3 (t) = E 1 (t)-E 3 (t),

[0115] Among them, E 1 (t) represents the electric field strength signal collected by the detector at time t, E 2 (t) represents the electric field strength signal collected by the detector 2 at time t, E 3 (t) represents the electric field strength signal collected by the detector 3 at time t, S 1 (t) represents the difference in electric field strength between detector 1 and detector 2 at time t, S 2 (t) represents the difference in electric field strength between detector 2 and detector 3 at time t, S 3 (t) represents the difference in electric field strength between detector 1 and detector 3 at time t;

[0116] S4. Detecting unstable conditions: Compare S 1 (t) and S 2 (t) is used to detect the instantaneous electric field intensity fluctuation. When there is partial discharge, the electric field is unstable, that is:

[0117] ΔS(t)=S 1 (t)-S 2 (t)≠0,

[0118] Among them, ΔS(t) represents the fluctuation of electric field intensity;

[0119] Step 3, high frequency component extraction and spectrum analysis, including the following steps:

[0120] S5. Extract high frequency components: 1 (t), S 2 (t) or ΔS(t) is high-pass filtered and the high-frequency part of the signal is extracted using wavelet transform, that is:

[0121] S 1，high (t) = HighPass(S 1 (t))

[0122] S 2，high (t) = HighPass(S 2 (t))

[0123] ΔS high (t) = HighPass(ΔS(t)),

[0124] Among them, S 1，high (t) represents the signal S 1(t) The high-frequency part obtained after high-pass filtering, S 2，high (t) represents the signal S 2 (t) The high-frequency part obtained after high-pass filtering, ΔS high (t) represents the high-frequency part obtained after high-pass filtering of ΔS(t);

[0125] S6. Spectrum analysis: Perform spectrum analysis on the extracted high-frequency components to obtain the spectrum characteristics of the partial discharge signal, and use Fourier transform to obtain the spectrum representation of the signal;

[0126] Step 4: Calculation of the intensity of partial discharge and analysis of the fault type, including the following steps:

[0127] S7. Calculate the intensity of partial discharge: Calculate the intensity of partial discharge according to the amplitude and energy characteristics of the high-frequency component. The higher the energy, the greater the intensity of partial discharge. The calculation formula of the intensity of partial discharge is:

[0128] Energy=∫|F(ΔS high (f))| 2 df,

[0129] Energy represents the total energy value obtained after spectrum analysis. The greater the energy, the greater the local discharge intensity. high (t) represents the high-frequency part obtained by high-pass filtering the electric field intensity fluctuation, F(ΔS high (f)) represents the change of ΔS high (t) The result after Fourier transformation;

[0130] S8. Fault type analysis: Classify the partial discharge type according to the frequency characteristics of the high-frequency components and the waveform of the discharge signal;

[0131] Step 5: Result output and alarm system, including the following steps:

[0132] S9, result output: output the calculated discharge intensity and analyzed fault type to the monitoring system;

[0133] S10, alarm system: set a threshold for discharge intensity. Once the discharge intensity is detected to exceed the set threshold, an alarm is triggered to remind maintenance personnel to check and handle it.

[0134] Preferably, the specific steps of extracting the high-frequency part of the signal using wavelet transform in step S5 are:

[0135] 1. Obtaining time domain signal: From the time domain signal S 1 (t), S 2 (t) or ΔS(t) to obtain the discrete sequence S 1 [n], S2 [n] or ΔS[n], the sampling expression of the time domain signal is:

[0136] S 1 [n] = S 1 (n·Δt), n=0, 1, 2,..., N-1,

[0137] S 2 [n] = S 2 (n·Δt), n=0, 1, 2,..., N-1,

[0138] ΔS[n]=ΔS(n·Δt), n=0, 1, 2,..., N-1,

[0139]

[0140] f s ≥2·f max ,

[0141]

[0142] Among them, Δt represents the time interval, N represents the number of sampling points, and f s represents the sampling frequency, f max represents the highest frequency of the signal, Δf represents the frequency resolution, and n represents the position of the sampling point in the sequence;

[0143] 2. Perform wavelet decomposition on the signal: Use discrete wavelet transform to decompose the signal into components of different scales. The expression of the wavelet decomposition is:

[0144] S 1 [n] → {A J , D J , D J-1 , ..., D 1},

[0145]

[0146] Among them, J represents the number of decomposition levels, A J represents the approximation coefficient of the Jth layer, D J represents the detail coefficient of the Jth layer, S j [k] represents the input signal of the jth layer, g[k] represents the low-pass filter coefficient, h[k] represents the high-pass filter coefficient, A j [n] represents the approximation coefficient of the jth layer, D j [n] represents the detail coefficient of the jth layer;

[0147] 3. Extract high-frequency components: The detail coefficients after wavelet decomposition correspond to the high-frequency part of the signal;

[0148] 4. Reconstruct high-frequency signal: Use inverse wavelet transform to reconstruct the high-frequency detail coefficients back to the time domain to obtain the time domain signal of the high-frequency part. The expression of the time domain signal of the high-frequency part is:

[0149] S 1，high [n] = IDWT(D j ),

[0150]

[0151] Among them, D j [n] and D j Both represent the detail coefficients of the jth layer, S j [k] represents the input signal of the jth layer, h[k] represents the high-pass filter coefficient, S 1,high [n] represents the extracted signal S 1 The high frequency part of [n];

[0152] Steps 2 to 4 also apply to S 2 [n] and ΔS[n], and get S 2,high [n] and ΔS high [n].

[0153] Preferably, the spectrum analysis in step S6 comprises the following steps:

[0154] 1. Obtaining a high-frequency signal: Obtaining a high-frequency signal S through step S5 1,high [n], S 2,high [n] and ΔS high

[0155] [n], the above sequence contains N sampling points, and the sampling frequency is f s ;

[0156] Second, perform fast Fourier transform on the high-frequency signal to convert the time domain signal into a frequency domain signal. The expression of the fast Fourier transform is:

[0157]

[0158] Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔS high )[k] represents the high frequency signal ΔS high [n] represents the frequency domain after fast Fourier transform, k represents the frequency index, and n represents the sampling point index of the time series;

[0159] 3. Calculate frequency resolution and frequency axis: The expression of frequency resolution is:

[0160]

[0161] Where N represents the number of sampling points, f s represents the sampling frequency, Δf represents the frequency resolution;

[0162] The expression of the frequency axis is:

[0163] f k =k·Δf,k=0,1,2,...,N / 2,

[0164] Among them, f k represents the actual frequency corresponding to the frequency index k, Δf represents the frequency resolution, and N represents the number of sampling points;

[0165] Fourth, calculate the amplitude spectrum, take the absolute value of the complex structure of the frequency domain signal, and obtain the amplitude of each frequency component:

[0166]

[0167] Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔS high )[k] represents the high frequency signal ΔS high [n] represents the frequency domain after fast Fourier transform, k represents the frequency index, represents the real part of the frequency domain signal, represents the imaginary part of the frequency domain signal, N represents the number of sampling points, and S[n] represents the discrete sequence of the input signal;

[0168] 5. Calculate the power spectrum density. The calculation formula of the power spectrum density is:

[0169]

[0170] Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔS high )[k] represents the high frequency signal ΔS high[n] represents the frequency domain after fast Fourier transform, k represents the frequency index;

[0171] 6. Analyze spectrum characteristics, including:

[0172] Spectrum feature extraction, the expression of the spectrum feature extraction is:

[0173]

[0174] Δf band =f upper -f lower ,

[0175] Among them, f peak Indicates the peak frequency in the spectrum, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, f band Indicates the frequency bandwidth, f upper Indicates the upper frequency where the power spectrum density is greater than the threshold, f lower Indicates the lower frequency where the power spectrum density is less than the threshold;

[0176] Discharge intensity estimation, calculate the total energy of the signal in the high frequency range, and estimate the total energy of the partial discharge signal by integrating the spectrum or accumulating the power spectrum density. The calculation formula for the total energy of the signal in the high frequency range is:

[0177]

[0178] Among them, Energy high Indicates the total energy of the signal in the high frequency range, PSD (S 1,high )[k] represents the high frequency signal S 1,high [n] Power spectral density at frequency index k, Δf represents the frequency resolution, k cutoff Indicates the frequency index corresponding to the cutoff frequency of the high-pass filter, and N represents the number of sampling points;

[0179] 7. Result analysis and output: According to the spectrum analysis results, determine the characteristics of the partial discharge signal, classify these characteristics, and determine the type and severity of the partial discharge.

[0180] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A high altitude partial discharge detector, characterized in that: include: A double hemispherical metal cover (1), an electromagnetic pulse sensor (2), an acoustic array sensor (3), a signal processing unit and a power supply. The double hemispherical metal cover (1) is composed of two convex surfaces of the hemispherical metal covers connected back to back, the openings of the two hemispherical metal covers are opposite, one end of the electromagnetic pulse sensor (2) is connected to the center position of the concave surface of the hemispherical metal cover, the electromagnetic pulse sensor (2) is composed of a plurality of coils or capacitors connected, the electromagnetic pulse sensor (2) includes at least three channels, and the probe of each channel is arranged along the spherical concentric axis direction of the hemispherical metal cover. The acoustic array sensor (3) is a hollow annular structure and includes a plurality of acoustic sensors. The other end of the electromagnetic pulse sensor (2) is connected to the center position of the acoustic array sensor (3), the signal processing unit is respectively connected to the electromagnetic pulse sensor (2) and the acoustic array sensor (3), and the power supply is respectively connected to the electromagnetic pulse sensor (2), the acoustic array sensor (3) and the signal processing unit.

2. A high altitude partial discharge detector according to claim 1, characterized in that: The double-hemispherical metal cover (1) serves as a signal ground terminal GND of the electromagnetic pulse sensor (2) and the acoustic array sensor (3).

3. A high altitude partial discharge detector according to claim 1, characterized in that: The probes of the channels of the electromagnetic pulse sensor (2) are arranged at different distances in the hemispherical metal cover.

4. A high altitude partial discharge detector according to claim 1, characterized in that: The acoustic sensor is composed of a plurality of acoustic microphone arrays connected together.

5. A high altitude partial discharge detector according to claim 1, characterized in that: The signal processing unit comprises: a filtering module, a data fusion module and a partial discharge analysis module, and the filtering module, the data fusion module and the partial discharge analysis module are connected in sequence to form the signal processing unit.

6. A high altitude partial discharge detector according to claim 1, characterized in that: The power source comprises: a rechargeable battery and a solar power supply module, and the rechargeable battery and the solar power supply module are connected.

7. A partial discharge analysis method for a high altitude partial discharge detector according to claim 1, characterized in that: The following steps are involved: Step 1: Detector installation and signal acquisition, including the following steps: S1. Install detectors: install at least three electromagnetic pulse detectors inside the hemispherical metal cover. The three detectors are located on the same axis but at different distances from the center of the sphere. Detector one corresponds to electric field strength E1, detector two corresponds to electric field strength E2, and detector three corresponds to electric field strength E3; S2. Electric signal acquisition: The electric field strength signals acquired by the three detectors are respectively recorded as E1(t), E2(t) and E3(t). The expressions of the electric field strength signals acquired by the three detectors are: Where Q(t) represents the instantaneous charge of the discharge source at time t, ∈0 represents the dielectric constant in free space, d1 represents the distance between detector 1 and the discharge source, d2 represents the distance between detector 2 and the discharge source, and d3 represents the distance between detector 3 and the discharge source; Step 2, calculation of signal difference and instantaneous condition detection, includes the following steps: S3, calculating the electric field strength difference, the calculation formula of the electric field strength difference is: S1(t)=E1(t)-E2(t) S2(t)=E2(t)-E3(t) S3(t)=E1(t)-E3(t), Wherein, E1(t) represents the electric field strength signal collected by detector 1 at time t, E2(t) represents the electric field strength signal collected by detector 2 at time t, E3(t) represents the electric field strength signal collected by detector 3 at time t, S1(t) represents the difference in electric field strength between detector 1 and detector 2 at time t, S2(t) represents the difference in electric field strength between detector 2 and detector 3 at time t, and S3(t) represents the difference in electric field strength between detector 1 and detector 3 at time t; S4, detection of unstable conditions: Compare S1(t) and S2(t) to detect instantaneous electric field intensity fluctuations. When partial discharge exists, the electric field is unstable, that is: ΔS(t)=S1(t)-S2(t)≠0, Among them, ΔS(t) represents the fluctuation of electric field intensity; Step 3, high frequency component extraction and spectrum analysis, including the following steps: S5, extract high-frequency components: perform high-pass filtering on S1(t), S2(t) or ΔS(t), and use wavelet transform to extract the high-frequency part of the signal, that is: S 1，high (t)=HighPass(S1(t)) S 2，high (t)=HighPass(S2(t)) ΔS high (t)=HighPass(ΔS(t)), Among them, S 1，high (t) represents the high-frequency part obtained by high-pass filtering the signal S1(t), S 2，high (t) represents the high-frequency part obtained by high-pass filtering the signal S2(t), ΔS high (t) represents the high-frequency part obtained after high-pass filtering of ΔS(t); S6. Spectrum analysis: Perform spectrum analysis on the extracted high-frequency components to obtain the spectrum characteristics of the partial discharge signal, and use Fourier transform to obtain the spectrum representation of the signal; Step 4: Calculation of the intensity of partial discharge and analysis of the fault type, including the following steps: S7. Calculate the intensity of partial discharge: Calculate the intensity of partial discharge according to the amplitude and energy characteristics of the high-frequency component. The higher the energy, the greater the intensity of partial discharge. The calculation formula of the intensity of partial discharge is: Energy=∫|F(ΔS high (f))| 2 df, Energy represents the total energy value obtained after spectrum analysis. The greater the energy, the greater the local discharge intensity. high (t) represents the high-frequency part obtained by high-pass filtering the electric field intensity fluctuation, F(ΔS high (f)) represents the change of ΔS high (t) The result after Fourier transformation; S8. Fault type analysis: Classify the partial discharge type according to the frequency characteristics of the high-frequency components and the waveform of the discharge signal; Step 5: Result output and alarm system, including the following steps: S9, result output: output the calculated discharge intensity and analyzed fault type to the monitoring system; S10, alarm system: set a threshold for discharge intensity. Once the discharge intensity is detected to exceed the set threshold, an alarm is triggered to remind maintenance personnel to check and handle it.

8. The partial discharge analysis method of a high altitude partial discharge detector according to claim 7, characterized in that: The specific steps of using wavelet transform to extract the high frequency part of the signal in step S5 are:

1. Obtaining a time domain signal: Obtaining a discrete sequence S1[n], S2[n] or ΔS[n] from a time domain signal S1(t), S2(t) or ΔS(t), wherein the sampling expression of the time domain signal is: S1[n]=S1(n·Δt), n=0,1,2,...,N-1, S2[n]=S2(n·Δt), n=0, 1, 2,..., N-1, ΔS[n]=ΔS(n·Δt), n=0,1,2,...,N-1, f s ≥2·f max , Among them, Δt represents the time interval, N represents the number of sampling points, and f s represents the sampling frequency, f max represents the highest frequency of the signal, Δf represents the frequency resolution, and n represents the position of the sampling point in the sequence; 2. Perform wavelet decomposition on the signal: Use discrete wavelet transform to decompose the signal into components of different scales. The expression of the wavelet decomposition is: S1[n]→{A J ,D J ,D J-1 ,...,D1}, Among them, J represents the number of decomposition levels, A J represents the approximation coefficient of the Jth layer, D J represents the detail coefficient of the Jth layer, S j [k] represents the input signal of the jth layer, g[k] represents the low-pass filter coefficient, h[k] represents the high-pass filter coefficient, A j [n] represents the approximation coefficient of the jth layer, D j [n] represents the detail coefficient of the jth layer; 3. Extract high-frequency components: The detail coefficients after wavelet decomposition correspond to the high-frequency part of the signal; 4. Reconstruct high-frequency signal: Use inverse wavelet transform to reconstruct the high-frequency detail coefficients back to the time domain to obtain the time domain signal of the high-frequency part. The expression of the time domain signal of the high-frequency part is: S 1，high [n]=IDWT(D j ), Among them, D j [n] and D j Both represent the detail coefficients of the jth layer, S j [k] represents the input signal of the jth layer, h[k] represents the high-pass filter coefficient, S 1,high [n] represents the high frequency part of the extracted signal S1[n]; Steps 2 to 4 are also applicable to S2[n] and ΔS[n], and S is obtained 2,high [n] and ΔS high [n].

9. The partial discharge analysis method of a high altitude partial discharge detector according to claim 7, characterized in that: The spectrum analysis in step S6 comprises the following steps:

1. Obtaining a high-frequency signal: Obtaining a high-frequency signal S through step S5 1,high [n], S 2,high [n] and ΔS high [n], the above sequence contains N sampling points, and the sampling frequency is f s ; Second, perform fast Fourier transform on the high-frequency signal to convert the time domain signal into a frequency domain signal. The expression of the fast Fourier transform is: Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔS high )[k] represents the high frequency signal ΔS high [n] represents the frequency domain after fast Fourier transform, k represents the frequency index, and n represents the sampling point index of the time series; 3. Calculate frequency resolution and frequency axis: The expression of frequency resolution is: Where N represents the number of sampling points, f s represents the sampling frequency, Δf represents the frequency resolution; The expression of the frequency axis is: f k =k·Δf,k=0,1,2,...,N / 2, Among them, f k represents the actual frequency corresponding to the frequency index k, Δf represents the frequency resolution, and N represents the number of sampling points; Fourth, calculate the amplitude spectrum, take the absolute value of the complex structure of the frequency domain signal, and obtain the amplitude of each frequency component: Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔS high )[k] represents the high frequency signal ΔS high [n] represents the frequency domain after fast Fourier transform, k represents the frequency index, represents the real part of the frequency domain signal, represents the imaginary part of the frequency domain signal, N represents the number of sampling points, and S[n] represents the discrete sequence of the input signal; 5. Calculate the power spectrum density. The calculation formula of the power spectrum density is: Among them, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, F(S 2,high )[k] represents the high frequency signal S 2,high [n] Frequency domain representation after fast Fourier transform, F(ΔS high )[k] represents the high frequency signal ΔS high [n] represents the frequency domain after fast Fourier transform, k represents the frequency index; 6. Analyze spectrum characteristics, including: Spectrum feature extraction, the expression of the spectrum feature extraction is: Δf band =f upper -f lower , Among them, f peak Indicates the peak frequency in the spectrum, F(S 1,high )[k] represents the high frequency signal S 1,high [n] Frequency domain representation after fast Fourier transform, f band Indicates the frequency bandwidth, f upper Indicates the upper frequency where the power spectrum density is greater than the threshold, f lower Indicates the lower frequency where the power spectrum density is less than the threshold; Discharge intensity estimation, calculate the total energy of the signal in the high frequency range, and estimate the total energy of the partial discharge signal by integrating the spectrum or accumulating the power spectrum density. The calculation formula for the total energy of the signal in the high frequency range is: Among them, Energy high Indicates the total energy of the signal in the high frequency range, PSD (S 1,high )[k] represents the high frequency signal S 1,high [n] Power spectral density at frequency index k, Δf represents the frequency resolution, k cutoff Indicates the frequency index corresponding to the cutoff frequency of the high-pass filter, and N represents the number of sampling points; 7. Result analysis and output: According to the spectrum analysis results, determine the characteristics of the partial discharge signal, classify these characteristics, and determine the type and severity of the partial discharge.

Images