Interference identification and suppression method for partial discharge test of high-altitude extra-high voltage transformer

Abstract

The invention relates to a high-altitude extra-high voltage transformer partial discharge test interference identification and suppression method, and belongs to the technical field of high-altitude region electrical equipment insulation state detection. In order to solve the technical problems that interference signals and real discharge signals are difficult to distinguish and the signal-to-noise ratio is low in the partial discharge test of the extra-high voltage transformer in the high-altitude environment, the method combines instantaneous zero-crossing density analysis and wavelet packet multi-scale feature extraction; according to the method, a multi-dimensional criterion fusing energy features, zero-crossing features, time-domain shape features and altitude correction factors is constructed, an adaptive interference discrimination rule is established, a frequency domain and time domain combined suppression strategy is adopted, various interference signals are effectively identified and suppressed, meanwhile, waveform and amplitude information of real partial discharge pulses is reserved to the maximum extent, and the suppression precision is improved. Therefore, the signal-to-noise ratio and the reliability of a partial discharge measurement result are remarkably improved, and key technical support is provided for accurate evaluation of the insulation state of the high-altitude extra-high voltage transformer.

Description

Technical Field

[0001] This invention belongs to the field of insulation condition detection technology for electrical equipment in high-altitude areas, and relates to a method for identifying and suppressing interference during partial discharge tests of ultra-high voltage transformers in high-altitude areas. Background Technology

[0002] As a key component of long-distance, high-capacity power transmission systems, the reliability of the insulation condition of ultra-high voltage (UHV) transformers directly affects the safe and stable operation of the entire power grid. Partial discharge detection is an important method for assessing the insulation condition of transformers and has been widely adopted in equipment operation status evaluation. However, conducting partial discharge tests on UHV transformers at high altitudes presents unique technical challenges. Due to the significant reduction in ambient air pressure, air density, and air ionization characteristics, corona discharge is more likely to occur on the external insulation of the equipment, leading to a sharp deterioration of the electromagnetic environment at the test site.

[0003] When ultra-high voltage transformer withstand voltage tests and partial discharge (PD) tests are conducted simultaneously, the test circuit is often mixed with various types of interference signals, mainly including corona discharge interference, interference caused by poor grounding, spatial electromagnetic radiation interference, transient interference from switch operation, and various random noises. Especially at high-altitude substations, the amplitude and frequency of background interference are often comparable to or even greater than the actual partial discharge pulse signal, severely obscuring the effective partial discharge signal and posing great difficulties for the accurate interpretation of test results.

[0004] Current anti-interference methods mainly follow two technical routes. The first approach focuses on engineering measures, aiming to reduce the coupling degree of external interference signals by improving measurement wiring methods, adding voltage equalization shielding devices, and optimizing grounding system design. The second approach focuses on signal processing algorithms, employing multi-bandpass or band-stop filtering techniques, wavelet or wavelet packet denoising methods, and empirical mode decomposition (EMD) and other time-frequency analysis tools to suppress interference components in the frequency domain or time-frequency domain. These traditional methods have shown some effectiveness in suppressing continuous narrowband interference such as carrier communication and high-frequency protection, as well as some white noise. However, they are difficult to effectively distinguish and accurately suppress pulse-type interference signals, especially corona interference and loop self-discharge interference, which are highly similar to the spectral characteristics of partial discharge pulses. Existing methods easily lead to amplitude attenuation and waveform distortion of the real partial discharge signal, seriously affecting the accurate measurement of partial discharge values ​​and the accurate judgment of insulation defects.

[0005] Therefore, for partial discharge tests of UHV transformers in special high-altitude environments, there is an urgent need for an innovative method that can incorporate high-altitude environmental parameters and comprehensively identify and suppress various types of interference, so as to significantly improve the signal-to-noise ratio of partial discharge test signals and the reliability of measurement results, and provide key technical support for the accurate assessment of the insulation status of UHV transformers in high-altitude areas. Summary of the Invention

[0006] In view of this, the purpose of the present invention is to provide a method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A method for identifying and suppressing interference during partial discharge tests of ultra-high-voltage transformers at high altitudes includes the following steps: S1. Acquisition of high-altitude environmental parameters and raw signal acquisition: Acquire the altitude h and air density ρ(h) of the test site, and acquire the discrete time series x(i) of the partial discharge test signal; S2. Coarse segmentation and non-pulse noise removal based on instantaneous zero-crossing density: The discrete time series x(i) is subjected to coarse segmentation based on instantaneous zero-crossing density to remove non-pulse noise intervals and obtain suspected pulse intervals. S3. Single pulse precise localization based on wavelet packet: In the suspected pulse interval, single pulse precise localization is performed based on wavelet packet decomposition to obtain a single pulse; S4. Construction of multi-scale feature vectors: Construct a multi-dimensional feature vector for each single pulse. The multi-dimensional feature vector includes energy features, zero-crossing features, temporal shape features, and altitude-based correction features. S5. Discrimination between interference pulses and partial discharge pulses: Discrimination between interference pulses and partial discharge pulses is based on the multidimensional feature vectors. S6. Interference Suppression and Signal Reconstruction: Suppress pulses identified as interference, retain partial discharge pulses, and reconstruct the signal.

[0008] Furthermore, in S1, the air density ρ(h) is obtained through a standard atmospheric model or by looking up a table.

[0009] Furthermore, in S2, the coarse segmentation based on instantaneous zero-crossing density includes: dividing the sequence x(i) into sliding windows of length N, where N is a positive integer, and calculating the instantaneous zero-crossing density of each window. Its definition is:

[0010] Where sgn is the sign function.

[0011] Furthermore, S2 also includes: setting an adaptive zero-crossing threshold. It is adjusted based on altitude h. ,in As a preset baseline threshold, This is a preset altitude correction factor. The standard air density at sea level; when When, the corresponding window is set to zero or marked as an invalid range; when At that time, it is retained as a suspected pulse interval.

[0012] Furthermore, in S3, the single-pulse precise localization based on wavelet packet decomposition includes: performing wavelet packet decomposition on the suspected pulse interval to obtain the sub-signal coefficients of each frequency band, and determining the pulse's time support interval based on the joint criterion of local energy and local zero-crossing density.

[0013] Furthermore, in step S4, the multidimensional feature vector includes the high-frequency energy ratio. Band entropy H, high-frequency zero-crossing density vector, time-domain zero-crossing density Ascent time Half-peak width Peak Repetition rate and phase aggregation Multiple features in.

[0014] Furthermore, the high-frequency energy ratio The result calculated using wavelet packet decomposition is:

[0015] in The normalized energy of the subband is HF, which is a preset set of high-frequency subbands; the frequency band entropy H is calculated as follows:

[0016] in It brings normalized energy to the child.

[0017] Furthermore, S4 also includes correcting the features based on altitude h: the high-frequency energy ratio is corrected to... The high-frequency zero-crossing density vector is corrected to ,in and This is the preset altitude correction function.

[0018] Furthermore, in step S5, the discrimination includes calculating the interference index. and partial discharge index ; Among them, the interference index The calculation is as follows:

[0019] Partial discharge index The calculation is as follows:

[0020] in ~ and ~ For preset weighting coefficients, The steepness of the rising edge is the indicator; when If the pulse is detected in time, it is considered an interference pulse; otherwise, it is considered a partial discharge pulse. This is a threshold function based on altitude.

[0021] Furthermore, in S6, the interference suppression includes: for narrowband interference, attenuating the main interference frequency band coefficient in the wavelet packet domain; for wideband interference, replacing it with adjacent interference-free signal segments in the time domain; and maintaining the original waveform of the partial discharge pulse.

[0022] The beneficial effects of this invention are as follows: (1) This invention creatively combines instantaneous zero-crossing density analysis with wavelet packet multi-scale feature extraction, breaking through the limitations of traditional methods that rely solely on single-domain analysis. By constructing a multi-dimensional criterion that integrates energy features, zero-crossing features, temporal shape features, and altitude correction factors, a systematic interference identification and suppression algorithm framework is formed. This multi-feature fusion analysis method can more comprehensively characterize signal characteristics and significantly improve the identification accuracy of various types of interference, especially impulse interference.

[0023] (2) This invention introduces an altitude correction function, enabling the algorithm to automatically adjust its sensitivity to signal features based on different altitudes and air density conditions. This design makes the method particularly suitable for plateau regions with frequent corona interference and complex electromagnetic environments, demonstrating good environmental adaptability.

[0024] (3) The present invention employs a joint frequency domain and time domain suppression strategy, which can effectively suppress interference while maximally protecting the amplitude information and waveform structure of the real partial discharge signal. This method avoids the signal distortion problem that is easily caused by traditional filtering techniques, and ensures the accuracy of partial discharge measurement and the integrity of waveform characteristics.

[0025] (4) This invention provides reliable technical support for the insulation condition assessment of UHV transformers in high-altitude areas. By improving the signal-to-noise ratio of partial discharge test signals and the reliability of measurement results, this method can provide more accurate basis for key diagnostic steps such as insulation defect judgment and initiation discharge voltage determination, thereby ensuring the safe and stable operation of the UHV power grid.

[0026] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0027] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein: Figure 1 This is a technical roadmap for the present invention; Figure 2 To obtain the PD signal of the substation with noise; Figure 3 This is the denoised PD signal of the substation. Detailed Implementation

[0028] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0029] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0030] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0031] The purpose of this invention is to provide an interference identification and suppression method for partial discharge tests of ultra-high voltage transformers at high altitudes. By combining instantaneous zero-crossing density with wavelet packet multi-scale analysis, a multi-dimensional feature vector is constructed that integrates energy characteristics, zero-crossing characteristics, temporal shape characteristics, and altitude correction parameters. An adaptive interference discrimination rule is established, and a joint frequency-domain and time-domain suppression strategy is adopted. This effectively suppresses corona and other interference signals in high-altitude environments while preserving the true partial discharge pulse waveform as much as possible. The technical roadmap is as follows: Figure 1 As shown, it mainly consists of seven steps.

[0032] Step 1: Acquisition of high-altitude environmental parameters and raw signal collection Before conducting the withstand voltage superimposed partial discharge test on the UHV transformer, the altitude h of the test site is obtained based on the geographical information of the test site or on-site meteorological data. The corresponding air density ρ(h) is obtained according to the standard atmospheric model or by looking up a table, and it is used as the environmental input parameter for the subsequent algorithm.

[0033] The measurement circuit is arranged according to relevant partial discharge test standards. High-frequency current sensors or coupling capacitors are installed at locations such as the transformer neutral point grounding lead, bushing end screen, or detection impedance to collect partial discharge measurement signals of the transformer during the test. The data acquisition device performs analog-to-digital conversion to obtain a discrete time sequence x(i), which serves as the raw signal for subsequent processing in this invention.

[0034] Step 2: Coarse segmentation and non-impulse noise removal based on instantaneous zero-crossing density To reduce the computational cost of subsequent wavelet packet analysis and initially eliminate pure noise segments, this invention first performs coarse segmentation processing on the original signal based on zero-crossing density. The sequence x(i) is divided into sliding windows of length N, and the instantaneous zero-crossing density is calculated for the signal within the k-th window. Its definition is .

[0035]

[0036] in, The sign function is used to characterize the frequency of sign changes of the signal within the window, thereby reflecting the degree of local oscillation.

[0037] Based on this, a baseline threshold is set. Based on the altitude correction factor γ, an adaptive zero-crossing threshold is constructed that varies with altitude according to the air density parameter ρ(h). .

[0038]

[0039] in, The threshold value represents the standard air density at sea level. When the test environment is at a higher altitude and air density decreases, the threshold value... Adjustments were made accordingly to adapt to changes in noise structure at high altitudes.

[0040] when When the signal within the window is determined to be mainly high-frequency non-pulse noise, it can be considered as having no effective partial discharge information, and the window is set to zero or marked as an invalid region; when If a pulse component is detected within a window, that window is classified as a suspected pulse cluster. By merging adjacent suspected pulse cluster windows, several suspected pulse intervals are obtained, providing a time range for the next step of fine wavelet packet analysis.

[0041] Step 3: Single-pulse precise localization based on wavelet packets Within the suspected pulse interval, this invention employs wavelet packet multi-scale decomposition to achieve precise localization of individual pulses. For each suspected interval... By selecting a preset mother wavelet function (such as the Daubechies wavelet db4) and a decomposition level L, wavelet packet decomposition is performed to obtain the coefficient sequence of each frequency band sub-signal. , where j is the sub-band number.

[0042] Local energy and local zero-crossing density are calculated across several high-frequency subbands. A combined criterion of "relatively high local energy and moderate zero-crossing density" is used to collaboratively determine the start and end sample positions of pulses across multiple subbands. By sliding detection of local energy peak values ​​and zero-crossing density changes across each subband, the time support interval of a single pulse can be determined, thereby further subdividing the original suspected pulse cluster into a series of single pulses. It also records the phase information of each pulse relative to the power frequency voltage.

[0043] Step 4: Construction of Multi-Scale Feature Vectors After completing the single-pulse precise positioning, the present invention performs each pulse... Multi-scale feature extraction is performed to construct a multi-dimensional feature vector for interference detection. The pulse is decomposed into L levels in the wavelet packet domain to obtain the coefficients of each sub-band. The energy of each subband is calculated as shown in Equation 3.

[0044]

[0045] The summation range covers all sample points within the sub-band. Then, the normalized energy is calculated.

[0046]

[0047] Based on this, the high-frequency energy ratio is defined.

[0048]

[0049] in This represents a pre-selected set of high-frequency sub-bands. To characterize the degree of energy dispersion across the frequency bands, this invention further calculates the band entropy to describe the degree of energy dispersion across each frequency band.

[0050]

[0051] Regarding the zero-crossing feature, this invention calculates the zero-crossing density in the overall time domain of the pulse. This is used to reflect the oscillation characteristics of the pulse waveform; at the same time, the zero-crossing density is calculated on several high-frequency sub-bands to form a high-frequency zero-crossing density vector.

[0052] in Indicates the first Zero crossover density of each high-frequency subband.

[0053] Regarding temporal shape features, this invention extracts the peak value for each pulse. Ascent time Half-peak width Parameters such as polarity and waveform symmetry are used to describe the shape characteristics of a pulse in the time domain. Among them, The maximum absolute amplitude of the pulse. This is the time required for the pulse to rise from the initial threshold to the peak value. This represents the time width of the pulse at half its peak value.

[0054] To reflect the statistical behavior of pulses within the power frequency cycle, this invention statistically analyzes the occurrence frequency of similar pulses or the interval between adjacent pulses within a multi-cycle range to obtain a repetition rate index. Simultaneously, the phase distribution of each pulse relative to the power frequency voltage is statistically analyzed, and the phase concentration index is calculated. It is used to characterize whether pulses are concentrated in a specific phase interval. Considering the significant impact of high-altitude environments on the corona and partial discharge spectral structure and triggering behavior, this invention adjusts the high-frequency energy ratio based on the altitude h of the test site. and high-frequency zero-crossing vector Introducing correction function Perform weighted processing.

[0055]

[0056] Thus, the high-frequency energy characteristics after altitude correction were obtained. and high-frequency zero-crossing features This allows the characteristic quantities to reflect the differences between interference signals and partial discharge signals under different altitude conditions. Finally, a multidimensional feature vector is constructed for each single pulse.

[0057] This feature vector comprehensively reflects various attributes of the pulse in the frequency domain, time domain, and statistical distribution, providing a basis for the distinction between interference and partial discharge.

[0058] Step 5: Distinguishing between interference pulses and partial discharge pulses Based on the obtained multidimensional feature vectors, this invention constructs an interference index. and partial discharge index In a typical implementation, the interference index and the partial discharge index can be expressed as:

[0059] in, It is a weighted average of high-frequency zero-crossing densities. The rise edge steepness index is composed of the pulse peak value and rise time. These are the weight coefficients obtained through training with simulated or labeled sample data.

[0060] Based on simulation and measured data from high-altitude tests, this invention sets a decision threshold function for operating conditions at different altitudes. When a pulse satisfies When, the pulse is identified as an interference pulse; when When this occurs, the pulse is identified as a partial discharge pulse. For pulses identified as interference, characteristics such as high-frequency energy ratio and bandwidth entropy can be combined to further classify them into narrowband interference and broadband interference, so as to select different suppression strategies.

[0061] Step Six: Interference Suppression and Signal Reconstruction After determining the type of interference, this invention employs frequency domain and time domain processing methods to suppress interference. For pulses identified as narrowband interference, the main interference frequency band is identified in the wavelet packet domain, and a coefficient attenuation factor is applied to the corresponding sub-band to reduce the coefficient of that sub-band by a predetermined ratio, while the coefficients of the remaining sub-bands remain unchanged. This selective frequency band attenuation suppresses narrowband interference while preserving the effective spectral components of partial discharge to the maximum extent.

[0062] For pulses identified as broadband interference or loop self-discharge, the time interval containing the pulse is replaced by an adjacent interference-free signal segment in the time domain. The interpolation method can be linear interpolation, piecewise cubic spline interpolation, etc., to ensure that the signal has good continuity and smoothness at the replacement point and avoid artificially introducing new artifacts.

[0063] For signals identified as partial discharge pulses, this invention does not modify their wavelet packet coefficients and time-domain waveforms, maintaining their original structure and amplitude for subsequent partial discharge quantity calculation and phase distribution analysis.

[0064] Finally, all processed coefficients or time-domain segments were reassembled in chronological order and inverse wavelet packet transform was performed to obtain the interference-suppressed partial discharge test signal. This signal exhibits a clear pulse morphology and significantly reduced background noise in the time domain. The partial discharge spectrum in the phase distribution diagram shows a more concentrated discharge band, providing a reliable basis for subsequent diagnostic analysis. To more comprehensively evaluate the denoising performance of the algorithm of this invention, an ultra-high voltage transformer was selected as the test object at a high-altitude substation. During its withstand voltage superimposed partial discharge test, a high-frequency current transformer was used to collect the partial discharge signal from the core grounding point. The measured results are as follows: Figure 2 As shown, the original signal contains a large number of mixed broadband and narrowband interference components, the amplitude of which is comparable to that of a partial discharge pulse, and the spectrum overlaps significantly. If the partial discharge quantity is directly calculated and the phase is analyzed, it will be difficult to obtain effective conclusions.

[0065] like Figure 3 As shown, after the measured signal is processed by the method of this invention, it can be observed that: after coarse segmentation with instantaneous zero crossover density, a large number of pure noise intervals are effectively eliminated, and the main pulse clusters are retained; through fine wavelet packet analysis and multidimensional feature discrimination, interference pulses such as corona are accurately identified and suppressed, and the partial discharge pulses are more prominent in the denoised time-domain waveform, with a complete waveform shape.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for identifying and suppressing interference during partial discharge tests of ultra-high-voltage transformers at high altitudes, characterized in that: Includes the following steps: S1. Acquisition of high-altitude environmental parameters and raw signal acquisition: Acquire the altitude h and air density ρ(h) of the test site, and acquire the discrete time series x(i) of the partial discharge test signal; S2. Coarse segmentation and non-pulse noise removal based on instantaneous zero-crossing density: The discrete time series x(i) is subjected to coarse segmentation based on instantaneous zero-crossing density to remove non-pulse noise intervals and obtain suspected pulse intervals. S3. Single pulse precise localization based on wavelet packet: In the suspected pulse interval, single pulse precise localization is performed based on wavelet packet decomposition to obtain a single pulse; S4. Construction of multi-scale feature vectors: Construct a multi-dimensional feature vector for each single pulse. The multi-dimensional feature vector includes energy features, zero-crossing features, temporal shape features, and altitude-based correction features. S5. Discrimination between interference pulses and partial discharge pulses: Discrimination between interference pulses and partial discharge pulses is based on the multidimensional feature vectors. S6. Interference Suppression and Signal Reconstruction: Suppress pulses identified as interference, retain partial discharge pulses, and reconstruct the signal.

2. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 1, characterized in that: In S1, the air density ρ(h) is obtained through a standard atmospheric model or by looking up a table.

3. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 1, characterized in that: In step S2, the coarse segmentation based on instantaneous zero-crossing density includes: dividing the sequence x(i) into sliding windows of length N, where N is a positive integer, and calculating the instantaneous zero-crossing density of each window. Its definition is: Where sgn is the sign function.

4. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 3, characterized in that: S2 further includes: setting an adaptive zero-crossing threshold. It is adjusted based on altitude h. ,in As a preset baseline threshold, This is a preset altitude correction factor. The standard air density at sea level; when When, the corresponding window is set to zero or marked as an invalid range; when At that time, it is retained as a suspected pulse interval.

5. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 1, characterized in that: In step S3, the single-pulse precise localization based on wavelet packet decomposition includes: performing wavelet packet decomposition on the suspected pulse interval to obtain the sub-signal coefficients of each frequency band, and determining the pulse's time support interval based on the joint criterion of local energy and local zero-crossing density.

6. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 1, characterized in that: In step S4, the multidimensional feature vector includes the high-frequency energy ratio. Band entropy H, high-frequency zero-crossing density vector, time-domain zero-crossing density Ascent time Half-peak width Peak Repetition rate and phase aggregation Multiple features in.

7. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 6, characterized in that: The high-frequency energy ratio The result calculated using wavelet packet decomposition is: in The normalized energy of the subband is HF, which is a preset set of high-frequency subbands; the frequency band entropy H is calculated as follows: in It brings normalized energy to the child.

8. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 6, characterized in that: The S4 also includes feature correction based on altitude h: the high-frequency energy ratio is corrected to... The high-frequency zero-crossing density vector is corrected to ,in and This is the preset altitude correction function.

9. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 1, characterized in that: In step S5, the discrimination includes calculating the interference index. and partial discharge index ; Among them, the interference index The calculation is as follows: Partial discharge index The calculation is as follows: in ~ and ~ For preset weighting coefficients, The steepness of the rising edge is the indicator; when If the pulse is detected in time, it is considered an interference pulse; otherwise, it is considered a partial discharge pulse. This is a threshold function based on altitude.

10. The method for identifying and suppressing interference during partial discharge tests of high-altitude ultra-high voltage transformers according to claim 1, characterized in that: In step S6, the interference suppression includes: for narrowband interference, attenuating the main interference frequency band coefficient in the wavelet packet domain; for wideband interference, replacing it with an adjacent interference-free signal segment in the time domain; and maintaining the original waveform of the partial discharge pulse.

Images