CN115061018A - High-voltage switch cabinet partial discharge monitoring method - Google Patents
High-voltage switch cabinet partial discharge monitoring method Download PDFInfo
- Publication number
- CN115061018A CN115061018A CN202210646834.4A CN202210646834A CN115061018A CN 115061018 A CN115061018 A CN 115061018A CN 202210646834 A CN202210646834 A CN 202210646834A CN 115061018 A CN115061018 A CN 115061018A
- Authority
- CN
- China
- Prior art keywords
- signal
- partial discharge
- point
- switch cabinet
- resonance
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Relating To Insulation (AREA)
Abstract
A partial discharge monitoring method for a high-voltage switch cabinet relates to the technical field of partial discharge on-line monitoring of the high-voltage switch cabinet, in particular to a partial discharge monitoring method for the high-voltage switch cabinet based on transient earth electric wave (TEV) and non-contact ultrasonic wave (AA) combined detection. The invention utilizes a double extreme value average threshold value method to carry out denoising pretreatment, realizes high resonance and low resonance decomposition through a resonance sparse algorithm, forms a denoised oscillation and periodic signal and a recombined signal of an impact signal rich in characteristic information, and inputs the recombined signal and an over-complete atom library constructed based on an over-complete redundancy function into a BP neural network data processing terminal for comparative analysis, thereby avoiding interference caused by errors or noise and extracting effective partial discharge pulse waveforms. By adjusting the response speed, the quality factor and the central frequency parameter, the requirements of partial discharge online monitoring and data analysis under the environment are met.
Description
Technical Field
The invention relates to the technical field of local discharge on-line monitoring of a high-voltage switch cabinet, in particular to a high-voltage switch cabinet local discharge monitoring method based on Transient Earth Voltage (TEV) and non-contact ultrasonic (AA) joint detection.
Background
Partial discharge is mainly the discharge of the internal insulation of transformers, switches, power cables and other high-voltage power equipment under the action of high voltage. It is one of the main causes of failure of high voltage power equipment.
The partial discharge detection is a common method for evaluating the insulation performance of high-voltage power equipment, and the extraction of a partial discharge pulse signal is taken as a key step of a pretreatment stage of a partial discharge measurement signal and is the basis for deep analysis of the partial discharge signal. At present, a main method for extracting the partial discharge pulse is a wavelet threshold denoising method, and the processing of the partial discharge pulse signal can achieve a certain effect, but the method also has problems.
Signals acquired by the online partial discharge detection method often contain a large amount of noise, so that the edge change of a single partial discharge pulse signal is not obvious, and the later partial discharge pulse signal characteristic analysis and identification difficulty is high due to the reasons of large calculation amount, obvious signal distortion, unclear pulse edge and the like in the prior art. The partial discharge monitoring method based on the bipolar value average threshold method and the resonance sparse decomposition algorithm has the advantages that the effect of extracting partial discharge signals by using a data processing terminal of the BP neural network is obvious, and the partial discharge monitoring method has the capacity of resisting interference and noise.
Disclosure of Invention
The invention aims to provide a high-voltage switch cabinet partial discharge monitoring method based on Transient Earth Voltage (TEV) and non-contact ultrasonic (AA) joint detection.
In order to achieve the purpose, the invention provides a partial discharge monitoring method based on a bipolar value average threshold method and a resonance sparse decomposition algorithm, which comprises the following steps of:
step 1: a sensor acquisition end of a partial discharge monitoring device is arranged in the switch cabinet, the acquisition end is connected to a data processing terminal, and a switch cabinet discharge signal x (t) is acquired through the sensor acquisition end of the monitoring device;
step 2: processing the sampled signal x (t) according to a bipolar average threshold method:
and step 3: adopting a threshold sliding window method, and according to the threshold Th and the window width M;
and 4, step 4: taking the signal between the point A and the point B as the once complete partial discharge waveform and storing the partial discharge waveform;
and 5: repeating the step 3 until reaching the end point of the partial discharge signal x (t), stopping moving, and obtaining the partial discharge signal x after denoising pretreatment 1 (t)；
Step 6: partial discharge signal x using FFT on Matlab 1 (t) transforming to time-frequency domain to obtain partial discharge signal x in time-frequency domain 2 (t)；
And 7: using over-complete redundancy functions
For signal x 2 (t) processing to construct an overcomplete atom library D ═ g γ }; alpha is a set of coefficients and is,
a matrix of M × N order basis functions, g, formed for the basis function vector γ Is an atom defined by the parameter set γ;
and 8: signal x 2 (t) performing resonance sparse decomposition by a high-pass filter and a low-pass filter respectively;
and step 9: decomposing the high resonance into a signal x G (t) and low resonance resolved signal x D And (t) inputting the signals into a BP neural network data processing terminal, comparing and analyzing the signals with a sparse over-complete dictionary library, and extracting effective partial discharge pulse signals.
As a preferable scheme, the step 2 specifically comprises the following steps:
step 2-1, an extreme value is taken for the sampling signal, and primary enveloping maximum value distribution and primary enveloping minimum value distribution are obtained;
step 2-2: continuously taking a secondary extreme value from the distribution of the enveloped primary extreme values to obtain a secondary extreme value containing a local extreme point;
2-3, filtering out local extreme values representing noise disturbance and interference, and reserving an effective extreme value range;
step 2-4: and obtaining a threshold Th of the current time period by averaging the extreme values with a certain length according to the secondary maximum value distribution and the minimum value distribution.
As another preferable scheme, step 3 of the present invention specifically includes the following steps:
step 3-1: starting from the first point of the partial discharge signal x (t), gradually shifting x (t), and when the absolute values of the signal amplitudes in the window when reaching the point A are all larger than a threshold value Th, taking the point A as the starting point of the signal waveform and recording the point A;
step 3-2: and continuing to move the partial discharge signal x (t) until the absolute values of the signal amplitudes in the window are smaller than a threshold Th when the point B is reached, namely taking the point B as an end point of the signal waveform and recording the point B.
Secondly, step 7 of the present invention specifically comprises the following steps:
step 7-1: selecting and signal x from overcomplete atomic pool 2 (t) the atoms that are most relevant for the residual, the signal can be decomposed into two parts, the component on the best matching atom and the residual component:
x 2 (t)＝<x 2 (t)，gγ 0 >gγ 0 +R 1 x 2 (t)
step 7-2: and repeatedly decomposing the residual signal after the optimal matching for many times:
R n x 2 (t)＝<R n x 2 (t)，gγ n >gγ n +R n+1 x 2 (t)
and 7-3: after k times of decomposition, obtaining a sparse decomposition formula of the signal after the residual meets the requirement:
and 7-4: and processing partial discharge standard maps of different power generation tests by using a sparse decomposition formula, establishing a sparse overcomplete dictionary base, and inputting the sparse overcomplete dictionary base to a BP neural network data processing terminal.
In addition, step 8 of the present invention specifically includes the following steps:
step 8-1: the quality factor Q is adjusted so that,
alpha and beta are low-pass and high-pass scale factors, and r is redundancy;
step 8-2: when Q is 1, low resonance decomposition; when Q is 3, high resonance decomposition is carried out;
step 8-3: in the resonance sparse decomposition process, different decomposition layer numbers L correspond to different central frequencies f c The TEV sensor detects electromagnetic wave signals of 1 MHz-100 MHz, and the non-contact ultrasonic sensor detects ultrasonic wave signals with the center frequency of 20 kHz-200 kHz;
step 8-4:
f s is the frequency of the original sampled signal x (t);
and 8-5: obtaining the denoised oscillation and periodic signals x with high resonance decomposition G (t), and a low resonance resolved characteristic information-rich impact signal x D (t)。
The invention has the beneficial effects.
The invention relates to a partial discharge pulse monitoring method, which utilizes a double extremum mean threshold method to carry out denoising pretreatment, realizes high resonance and low resonance decomposition through a resonance sparse algorithm, forms a denoised oscillation and periodic signal and a recombined signal of an impact signal rich in characteristic information, and inputs the recombined signal and an over-complete atomic library constructed based on an over-complete redundant function into a BP neural network data processing terminal for comparative analysis, so that the interference caused by errors or noise can be avoided, an effective partial discharge pulse waveform is extracted, and the requirements of partial discharge online monitoring and data analysis under the environment are met by adjusting the response speed (corresponding to the step 3 in a specific implementation mode), the quality factor (corresponding to the step 8 in the specific implementation mode) and the central frequency parameter (corresponding to the step 8 in the specific implementation mode).
Drawings
The invention is further described with reference to the following figures and detailed description. The scope of the invention is not limited to the following expressions.
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2 is a waveform diagram of an original partial discharge sampling signal.
Fig. 3 is a graph of extracted partial discharge waveforms.
Detailed Description
The following detailed description of embodiments of the present invention will be made with reference to the accompanying drawings and examples.
The partial discharge monitoring method based on the bipolar value average threshold method and the resonance sparse decomposition algorithm comprises the following steps of:
step 1: a sensor acquisition end of a partial discharge monitoring device is arranged in the switch cabinet, the acquisition end is connected to a data processing terminal, and a switch cabinet discharge signal x (t) is acquired through the sensor acquisition end of the monitoring device;
step 2: as shown in fig. 2, the sampled signal x (t) is processed according to a bipolar average threshold method:
2-1, taking an extreme value of the sampled signal to obtain primary enveloping maximum value distribution and primary enveloping minimum value distribution;
step 2-2: continuously taking a secondary extreme value from the distribution of the enveloped primary extreme values to obtain a secondary extreme value containing a local extreme point;
2-3, filtering out local extreme values representing noise disturbance and interference, and reserving an effective extreme value range;
step 2-4: obtaining a threshold Th of the current time period by averaging extrema of a certain length according to the distribution of quadratic maxima and minima (the step considers that the average values are different in different lengths and can adjust the length according to requirements to adjust the threshold, so that compared with the existing threshold method, the step can dynamically adjust the threshold and can dynamically adapt to various interferences such as spurs, square waves and the like);
and step 3: adopting a threshold sliding window method, and according to the threshold Th and the window width M;
step 3-1: starting from the first point of the partial discharge signal x (t), gradually shifting x (t), and when the absolute values of the signal amplitudes in the window when reaching the point A are all larger than a threshold value Th, taking the point A as the starting point of the signal waveform and recording the point A;
step 3-2: continuing to move the partial discharge signal x (t) until the absolute values of the signal amplitudes in the window are smaller than a threshold Th when the point B is reached, namely taking the point B as an end point of the signal waveform, and recording the point B;
and 4, step 4: taking the signal between the point A and the point B as the once complete partial discharge waveform and storing the partial discharge waveform;
and 5: repeating the step 3 until the partial discharge signal x (t) reaches the end point, stopping moving, and obtaining the partial discharge signal x after denoising pretreatment 1 (t)；
Step 6: partial discharge signal x is analyzed on Matlab using FFT (fast Fourier analysis) 1 (t) transforming to time-frequency domain to obtain partial discharge signal x in time-frequency domain 2 (t)；
And 7: using over-complete redundancy functions
For signal x 2 (t) processing to construct an overcomplete atom library D ═ g γ }; (a is a set of coefficients,
a matrix of M × N order basis functions formed for the basis function vector, g γ being an atom defined by the set of parameters γ)
Step 7-1: selecting and signal x from overcomplete atomic pool 2 (t) the most residual related atom, the signal can be decomposed into two parts, the component on the best matching atom and the residual component:
x 2 (t)＝<x 2 (t)，g γ0 >g γ0 +R 1 x 2 (t)
step 7-2: and repeatedly decomposing the residual signal after the optimal matching for many times:
R n x 2 (t)＝<R n x 2 (t)，gγn>g γn +R n+1 x 2 (t)
and 7-3: after k times of decomposition, obtaining a sparse decomposition formula of the signal after the residual meets the requirement:
and 7-4: processing partial discharge standard maps of different power generation tests by using a sparse decomposition mode, establishing a sparse overcomplete dictionary base, and inputting the sparse overcomplete dictionary base to a BP neural network data processing terminal;
and 8: signal x 2 (t) performing resonance sparse decomposition by a high-pass filter and a low-pass filter respectively;
step 8-1: the quality factor Q is adjusted so that,
(α and β are low-pass and high-pass scale factors, r is redundancy);
step 8-2: when Q is 1, low resonance decomposition is carried out; when Q is 3, high resonance decomposition is carried out;
step 8-3: in the resonance sparse decomposition process, different decomposition layer numbers L correspond to different central frequencies f c The TEV sensor detects electromagnetic wave signals of 1 MHz-100 MHz, and the non-contact ultrasonic sensor detects ultrasonic wave signals with the center frequency of 20 kHz-200 kHz;
and 8-4:
(f s for the frequency of the original sampled signal x (t)
And 8-5: obtaining the denoised oscillation and periodic signals x with high resonance decomposition G (t), and a low resonance resolved characteristic information-rich impact signal x D (t)；
And step 9: decomposing the high resonance into a signal x G (t) and low resonance resolved signal x D (t) inputting the signals into a BP neural network data processing terminal, comparing and analyzing the signals with a sparse over-complete dictionary library, and extracting effective partial discharge pulse signals, as shown in fig. 3.
Compared with a partial discharge signal extraction method, the processing method in the steps 7, 8 and 9 of the application carries out denoising processing before sparse decomposition, so that the interference of noise on the extraction effect can be effectively reduced; and constructing a sparse overcomplete atom library based on different discharge test standard maps, comparing the sparse overcomplete atom library with high and low resonance decomposition signals, and more effectively identifying and extracting accurate partial discharge signals.
According to the partial discharge monitoring method, denoising pretreatment is carried out by using a double extreme value average threshold method, high resonance and low resonance decomposition is realized through a resonance sparse algorithm, denoised oscillation and periodic signals and impact signals rich in characteristic information are formed, and the denoised oscillation and periodic signals and the impact signals are input to a BP neural network data processing terminal together with an over-complete atomic library constructed based on an over-complete redundant function for comparative analysis, so that interference caused by errors or noise can be avoided, effective partial discharge pulse waveforms are extracted, and the requirements of partial discharge online monitoring and data analysis under the environment are met by adjusting parameters such as response speed, quality factors and central frequency.
It should be understood that the detailed description of the present invention is only for illustrating the present invention and is not limited by the technical solutions described in the embodiments of the present invention, and those skilled in the art should understand that the present invention can be modified or substituted equally to achieve the same technical effects; as long as the use requirements are met, the method is within the protection scope of the invention.
Claims (5)
1. A partial discharge monitoring method for a high-voltage switch cabinet is characterized by comprising the following steps:
step 1: a sensor acquisition end of a partial discharge monitoring device is arranged in the switch cabinet, the acquisition end is connected to a data processing terminal, and a switch cabinet discharge signal x (t) is acquired through the sensor acquisition end of the monitoring device;
step 2: processing the sampled signal x (t) according to a bipolar average threshold method:
and step 3: adopting a threshold sliding window method, and according to the threshold Th and the window width M;
and 4, step 4: taking the signal between the point A and the point B as a once complete partial discharge waveform and storing the partial discharge waveform;
and 5: repeating the step 3 until the partial discharge signal x (t) reaches the end point, stopping moving, and obtaining the partial discharge signal x after denoising pretreatment 1 (t)；
Step 6: partial discharge signal x using FFT on Matlab 1 (t) transforming to time-frequency domain to obtain partial discharge signal x in time-frequency domain 2 (t)；
And 7: using over-complete redundancy functions
For signal x 2 (t) processing to construct an overcomplete atom library D ═ g γ }; alpha is a set of coefficients and is,
a matrix of M × N order basis functions, g, formed for the basis function vector γ Is an atom defined by the parameter set γ;
and 8: signal x 2 (t) performing resonance sparse decomposition by a high-pass filter and a low-pass filter respectively;
and step 9: decomposing the high resonance into a signal x G (t) and low resonance resolved signal x D And (t) inputting the signals into a BP neural network data processing terminal, comparing and analyzing the signals with a sparse over-complete dictionary library, and extracting effective partial discharge pulse signals.
2. The method for monitoring the partial discharge of the high-voltage switch cabinet according to claim 1, wherein the step 2 specifically comprises the following steps:
step 2-1, an extreme value is taken for the sampling signal, and primary enveloping maximum value distribution and primary enveloping minimum value distribution are obtained;
step 2-2: continuously taking a secondary extreme value from the distribution of the enveloping primary extreme values to obtain a secondary extreme value containing a local extreme value point;
2-3, filtering local extreme values representing noise disturbance and interference, and reserving an effective extreme value range;
step 2-4: and obtaining a threshold Th of the current time period by averaging the extreme values with a certain length according to the secondary maximum value distribution and the minimum value distribution.
3. The method for monitoring the partial discharge of the high-voltage switch cabinet according to claim 1, wherein the step 3 specifically comprises the following steps:
step 3-1: starting from the first point of the partial discharge signal x (t), gradually shifting x (t), and when the absolute values of the signal amplitudes in the window when reaching the point A are all larger than a threshold value Th, taking the point A as the starting point of the signal waveform and recording the point A;
step 3-2: and continuing to move the partial discharge signal x (t) until the absolute values of the signal amplitudes in the window are smaller than a threshold Th when the point B is reached, namely taking the point B as an end point of the signal waveform and recording the point B.
4. The method for monitoring the partial discharge of the high-voltage switch cabinet according to claim 1, wherein the step 7 specifically comprises the following steps:
step 7-1: selecting and signal x from overcomplete atomic pool 2 (t) the most residual related atom, the signal can be decomposed into two parts, the component on the best matching atom and the residual component:
x 2 (t)＝<x 2 (t)，g γ0 >g γ0 +R 1 x 2 (t)
step 7-2: and repeatedly decomposing the residual signal after the optimal matching for many times:
R n x 2 (t)＝<R n x 2 (t)，g γn >g γn +R n+1 x 2 (t)
and 7-3: after k times of decomposition, obtaining a sparse decomposition formula of the signal after the residual meets the requirement:
and 7-4: and processing partial discharge standard maps of different power generation tests by using a sparse decomposition formula, establishing a sparse overcomplete dictionary base, and inputting the sparse overcomplete dictionary base to a BP neural network data processing terminal.
5. The method for monitoring the partial discharge of the high-voltage switch cabinet according to claim 1, wherein the step 8 specifically comprises the following steps:
step 8-1: the quality factor Q is adjusted so that,
alpha and beta are low-pass and high-pass scale factors, and r is redundancy;
step 8-2: when Q is 1, low resonance decomposition; when Q is 3, high resonance decomposition is carried out;
step 8-3: in the resonance sparse decomposition process, different decomposition layer numbers L correspond to different central frequencies f c The TEV sensor detects electromagnetic wave signals of 1 MHz-100 MHz, and the non-contact ultrasonic sensor detects ultrasonic wave signals with the center frequency of 20 kHz-200 kHz;
step 8-4:
f s the frequency of the original sampled signal x (t);
and 8-5: obtaining the de-noised oscillation and periodic signals x decomposed by high resonance G (t), and a low resonance resolved characteristic information-rich impact signal x D (t)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210646834.4A CN115061018A (en) | 2022-06-09 | 2022-06-09 | High-voltage switch cabinet partial discharge monitoring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210646834.4A CN115061018A (en) | 2022-06-09 | 2022-06-09 | High-voltage switch cabinet partial discharge monitoring method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115061018A true CN115061018A (en) | 2022-09-16 |
Family
ID=83199428
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210646834.4A Pending CN115061018A (en) | 2022-06-09 | 2022-06-09 | High-voltage switch cabinet partial discharge monitoring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115061018A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117110818A (en) * | 2023-10-25 | 2023-11-24 | 江苏沙洲电气有限公司 | Novel high-voltage switch cabinet partial discharge detection method and system |
| CN118444113A (en) * | 2024-07-08 | 2024-08-06 | 江西华莱电技术有限公司 | Partial discharge detection method of switch cabinet and high-voltage switch cabinet |
-
2022
- 2022-06-09 CN CN202210646834.4A patent/CN115061018A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117110818A (en) * | 2023-10-25 | 2023-11-24 | 江苏沙洲电气有限公司 | Novel high-voltage switch cabinet partial discharge detection method and system |
| CN118444113A (en) * | 2024-07-08 | 2024-08-06 | 江西华莱电技术有限公司 | Partial discharge detection method of switch cabinet and high-voltage switch cabinet |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chan et al. | Self-adaptive partial discharge signal de-noising based on ensemble empirical mode decomposition and automatic morphological thresholding | |
| CN115061018A (en) | High-voltage switch cabinet partial discharge monitoring method | |
| Zhang et al. | A novel wavelet transform technique for on-line partial discharge measurements. 1. WT de-noising algorithm | |
| Cunha et al. | A new wavelet selection method for partial discharge denoising | |
| Satish | Short-time Fourier and wavelet transforms for fault detection in power transformers during impulse tests | |
| Chang et al. | Denoising of partial discharge signals in wavelet packets domain | |
| Ashtiani et al. | Feature-oriented de-noising of partial discharge signals employing mathematical morphology filters | |
| Wang et al. | Optimization of wavelet and thresholding for partial discharge detection under HVDC | |
| CN102624349B (en) | Harmonic noise and white-noise interference eliminating method with low distortion to initial data | |
| CN114492538A (en) | Local discharge signal denoising method for urban medium-voltage distribution cable | |
| Wang et al. | Analysis of partial discharge signal using the Hilbert-Huang transform | |
| CN108089100A (en) | The detection method of small current neutral grounding system arc light resistance earth fault | |
| Zhang et al. | Suppression of UHF partial discharge signals buried in white-noise interference based on block thresholding spatial correlation combinative de-noising method | |
| CN114609515A (en) | GIS ultrahigh frequency partial discharge detection interference suppression method based on sequential hierarchical signal processing | |
| CN117520916A (en) | Cable partial discharge signal denoising method | |
| CN110287853B (en) | Transient signal denoising method based on wavelet decomposition | |
| CN108020761B (en) | A kind of Denoising of Partial Discharge | |
| Long et al. | A novel automatic pulse segmentation approach and its application in PD-induced electromagnetic wave detection | |
| CN112149626B (en) | Engine knock characteristic frequency extraction method | |
| Xin et al. | De-noising of GIS UHF Partial Discharge monitoring based on wavelet method | |
| Naderi et al. | Application of wavelet analysis to the determination of partial discharge location in multiple-α transformer windings | |
| Yang et al. | An Improved EMD Algorithm based on Complete Complementary Adaptive Noise | |
| Li | Research on extraction of partial discharge signals based on wavelet analysis | |
| Zhuo et al. | Denoising and features extraction of distorted waveform based on mean-dilation operator and mean-erosion operator | |
| CN105974283B (en) | Cable partial discharge feature extraction method based on wavelet packet survival index singular entropy |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |