Partial discharge detection method and system based on synchronous compression wavelet-synchronous extrusion

Abstract

The invention relates to the technical field of power equipment state monitoring, and discloses a partial discharge detection method and system based on synchronous compression wavelet-synchronous extrusion. The method comprises the following steps: preprocessing a partial discharge signal; calculating a coefficient through continuous wavelet transform and estimating an instantaneous frequency and a chirp rate; synchronous compression is carried out to compensate the phase, and scale remarking is carried out; energy is reallocated to a frequency axis through synchronous extrusion, and a high-aggregation-degree time-frequency diagram is generated; extracting a peak domain based on the saliency spectrum, demodulating along a ridge line, and estimating event parameters; and finally, robust judgment and auditing output are completed by using the joint statistics. According to the method, the sensitivity and robustness of partial discharge signal detection under the conditions of low signal-to-noise ratio and strong interference are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of power equipment condition monitoring technology, specifically relating to a partial discharge detection method and system based on synchronous compression wavelet-synchronous squeezing. Background Technology

[0002] In existing technologies, non-invasive detection and feature extraction of partial discharge in power equipment mainly follow two technical routes: one focuses on time-domain / frequency-domain statistical feature methods, characterizing discharge activity through indicators such as envelope energy, pulse count, peak factor, and spectral kurtosis of high-frequency current (HFCT), ultra-high frequency (UHF), or ultrasonic channel signals; the other is based on time-frequency decomposition and time-frequency energy aggregation methods, using wavelet transform (WT), continuous wavelet transform (CWT), Hilbert-Huang transform (HHT), empirical mode decomposition (EMD), or synchronous compressed wavelet transform (SCWT) to reconstruct the time-spectrum and enhance the features of non-stationary partial discharge signals. The former can quickly detect discharge events when noise is low and pulses are sparse, but lacks the ability to distinguish multi-source interference and complex modulation components; the latter can more finely characterize the transient spectral evolution of partial discharge pulses, but suffers from problems such as limited resolution, mode aliasing, energy leakage, and high parameter sensitivity.

[0003] From the perspective of the resolution mechanism of time-frequency analysis methods, traditional wavelet transform is characterized by multi-scale and multi-band coupling. Its time-frequency energy concentration is limited by the selection of the mother wavelet and the scale factor, making it difficult to maintain high time and frequency resolution simultaneously under the condition of coexistence of high-frequency partial discharge components and broadband noise. Although synchronous compressed wavelet transform improves energy concentration through frequency redistribution, it is easily affected by instantaneous phase estimation errors during frequency compression, leading to spectral drift and frequency ambiguity in higher-order modulated signals. On the other hand, HHT and its variants rely on empirical decomposition processes, resulting in problems such as endpoint effects, mode aliasing, and non-reproducibility of results, making it difficult to achieve standardized quantitative comparisons.

[0004] In scenarios involving parallel acquisition of multi-channel signals (such as UHF, ultrasound, and high-frequency current), existing technologies typically employ independent analysis or simple energy fusion, lacking joint modeling of the time-frequency coherence and phase synchronization characteristics between channels. This makes it difficult to effectively identify the synchronization characteristics of cross-channel discharge events. Furthermore, existing modulation and demodulation methods often rely on fixed-bandwidth filtering or envelope demodulation, failing to dynamically adjust demodulation parameters based on the instantaneous frequency of the signal. This results in insufficient detection sensitivity for complex modulation or weak discharge signals.

[0005] From a system perspective, existing partial discharge detection processes lack a closed-loop feedback mechanism between noise suppression, feature extraction, and event recognition. Most algorithms only output the final detection result without providing verifiable intermediate physical quantities (such as instantaneous frequency, energy spectral density, and phase evolution curves), resulting in insufficient system interpretability and traceability. When signal characteristics, sensor placement, or operating environment change, model parameters need to be recalibrated or manually adjusted, leading to poor engineering adaptability.

[0006] The root causes of these shortcomings are twofold: firstly, traditional wavelet and HHT-type methods fail to fully utilize the local structural features of the signal in the time-frequency-phase space and lack a coherent enhancement mechanism under multi-resolution conditions; secondly, while existing synchronous compression methods can improve frequency concentration, they are not linked with the modulation and demodulation mechanism, and still suffer from high-frequency detail leakage and energy mismatch. Consequently, under conditions of strong noise or multi-source interference, it is difficult to simultaneously achieve both sensitivity and robustness in partial discharge detection. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of existing partial discharge detection methods, such as low sensitivity and poor robustness under strong noise, narrowband interference, and complex modulation environments, and to provide a partial discharge detection method and system based on synchronous compressed wavelet-synchronous squeezing. This method aims to achieve the following objectives: (1) Significantly improves the concentration of time-frequency energy and enhances the visibility of weak transient signals; (2) Reduce dependence on fixed bandpass filtering and detection threshold, and improve adaptability; (3) Achieve effective compensation for chirp modulation and group delay effects; (4) Supports synchronous analysis and fusion of multi-channel signals, and outputs complete and auditable parameters and evidence chains.

[0008] To solve the above-mentioned technical problems, the present invention is implemented as follows: This invention provides a partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing for detecting partial discharge signals in power equipment. The method includes the following steps: Signal preprocessing step: Pre-filtering and detrending the raw partial discharge signal collected from the power equipment to obtain a standardized signal. The pre-filtering includes bandpass filtering and adaptive notch filtering, wherein the passband of the bandpass filter is set to the service bandwidth. And satisfies the Nyquist constraint. The adaptive notch filter is designed for known narrowband interference fundamental frequencies. Suppressing its harmonics, among which, This is the lower limit frequency of the selected service bandwidth for the partial discharge signal in the current detection channel; This is the upper limit frequency of the selected service bandwidth for the partial discharge signal in the current detection channel; The sampling rate is used; the detrending process employs median filtering or low-order polynomial fitting to remove low-frequency drift components from the signal; the time-frequency analysis step involves performing continuous wavelet transform on the standardized signal to obtain wavelet coefficients. and in amplitude mask Internal estimation of instantaneous angular frequency and chirping rate ,in, The scaling factor (scaling parameter) in continuous wavelet transform is used to control the scaling of wavelet basis functions in the frequency domain, thus corresponding to different analysis frequency bands; The shift factor (time shift parameter) in continuous wavelet transform is used to control the position of the wavelet basis function in the time domain, corresponding to the time position of the signal being analyzed. The amplitude threshold is set based on signal-noise statistics; the synchronous compression step is based on the chirp rate. Phase compensation is performed on the wavelet coefficients, and scale recalculation is applied to achieve pre-shrinking of the scale domain, resulting in corrected wavelet coefficients. Synchronous extrusion step: based on the instantaneous angular frequency The energy carried by the modified wavelet coefficients is reassigned from the scale domain to the frequency axis to generate a synchronous squeeze time-frequency map with high energy concentration. ,in For synchronous extrusion time-frequency diagram The target frequency variable in the data is used to carry the instantaneous angular frequency components reconfigured from the scale domain. Event detection steps: Construct a saliency spectrum based on the time-frequency graph, extract peak regions through morphological processing and connectivity markers, and perform adaptive demodulation and envelope extraction along the ridge line in each peak region to estimate the arrival time, center frequency, bandwidth, and energy of the event; Joint decision steps: Form a joint decision statistic based on the super-background energy statistics of the peak region and the multi-resolution consistency score, and compare it with a threshold to complete the robust identification and auditable output of real partial discharge events.

[0009] Furthermore, in the synchronous extrusion step, the modified wavelet coefficients are adjusted before energy reconfiguration. The joint robust weights are applied and calculated as follows: ; in, The amplitude threshold is set based on the noise statistics of the partial discharge signal; The chirp suppression parameter is set based on the maximum acceptable chirp rate of the partial discharge signal of the power equipment; It is a monotonically increasing threshold function.

[0010] Furthermore, the adaptive demodulation and envelope extraction in the event detection step include: in a single peak region Inner, ridge line Depend on Confirm; demodulated signal ,in Ridge The corresponding instantaneous angular frequency, The imaginary unit; envelope ,in A low-pass filter set according to the partial discharge envelope bandwidth; the arrival time of the event is determined by... Sure.

[0011] Furthermore, the parameter estimation in the event detection step includes: center frequency. By Peak Domain The frequency values ​​of all points within the time-frequency plot are used to determine their relative positions. The bandwidth is obtained by weighted averaging of the amplitude values. By Peak Domain The second moment of the internal frequency relative to the center frequency is calculated.

[0012] Furthermore, the joint decision statistics in the joint decision step The calculation is as follows: ; in, To characterize the approximate generalized likelihood ratio statistic of the peak region superbackground energy, The multi-resolution consistency score is obtained by changing the quality factor of the wavelet transform and calculating the overlap of the peak domain projection under different settings.

[0013] This invention also provides a partial discharge detection system based on synchronous compressed wavelet-synchronous squeezing for executing the aforementioned partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing, used for partial discharge detection of power equipment, comprising: a signal acquisition interface for receiving partial discharge signals from at least one sensor of the power equipment; a processor coupled to the signal acquisition interface; and a memory storing a computer program, which, when executed by the processor, is configured to perform the functions of the following program units: a preprocessing unit for pre-filtering and detrending the partial discharge signal to obtain a standardized signal; a wavelet transform and parameter estimation unit for calculating the continuous wavelet transform coefficients of the standardized signal and estimating the instantaneous angular frequency and chirp rate; a synchronous compression unit for performing phase compensation and scale recalculation on the wavelet coefficients; a synchronous squeezing unit for reconfiguring energy from the scale domain to the frequency axis to generate a time-frequency diagram; a peak detection and demodulation unit for constructing a saliency spectrum, extracting peaks and ridges, and performing adaptive demodulation to obtain event parameters; and a decision and output unit for completing event identification based on joint statistics and outputting auditable results.

[0014] Furthermore, the signal acquisition interface is compatible with at least one of ultra-high frequency (UHF), ultrasonic, or high-frequency current (HFCT) sensors.

[0015] Furthermore, when performing energy reconfiguration, the synchronous compression unit is configured to apply the joint robust weights to the wavelet coefficients after synchronous compression.

[0016] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described thereon.

[0017] The present invention also provides an online partial discharge monitoring system, including the aforementioned partial discharge detection system, and a back-end server for time alignment and coherent fusion of the output results of multiple detection systems.

[0018] Compared with the prior art, the advantages of this invention are as follows: (1) By using SCWT and SST in a cascaded manner and combined with local chirp compensation, the high concentration of time-frequency energy is achieved, which significantly improves the detectability of weak partial discharge signals under strong interference.

[0019] (2) By adopting a multi-resolution modulation and demodulation and peak domain connectivity mechanism, adaptive analysis of complex modulation signals is realized, reducing the dependence on preset bandwidth and threshold, and improving the robustness and applicability of the system.

[0020] (3) By combining statistical decision and multi-quality factor consistency verification, the false alarm rate was effectively controlled, and the stability of the test results under different working conditions was guaranteed.

[0021] (4) The output parameters are complete and the intermediate evidence is traceable, supporting result verification and system self-verification, which greatly improves the credibility and maintainability in engineering applications.

[0022] (5) The system architecture is flexible and applicable to various sensing channels such as UHF, ultrasound, and HFCT, and has good cross-platform migration and edge deployment capabilities. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein: Figure 1 A flowchart of the partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing provided by the present invention; Figure 2The structural block diagram of the partial discharge detection system based on synchronous compressed wavelet-synchronous squeezing provided by the present invention is shown. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] The terms "first," "second," etc., used in this specification are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class, without limiting the number of objects; for example, the first object can be one or more. Furthermore, in the specification, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0026] Please see Figure 1 As shown, this embodiment of the invention provides a partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing for detecting partial discharge signals in power equipment, including the following steps: Step S1, signal preprocessing: pre-filtering and detrending processing are performed on the original partial discharge signal collected from the power equipment to obtain a standardized signal. The pre-filtering includes bandpass filtering and adaptive notch filtering, wherein the passband of the bandpass filter is set to the service bandwidth. And satisfies the Nyquist constraint. The adaptive notch filter is designed for known narrowband interference fundamental frequencies. Suppressing its harmonics, among which, This is the lower limit frequency of the selected service bandwidth for the partial discharge signal in the current detection channel; This is the upper limit frequency of the selected service bandwidth for the partial discharge signal in the current detection channel; The sampling rate is used; the detrending process employs median filtering or low-order polynomial fitting to remove low-frequency drift components from the signal; step S2, time-frequency analysis: continuous wavelet transform is performed on the standardized signal to obtain wavelet coefficients. and in amplitude mask Internal estimation of instantaneous angular frequency and chirping rate ,in, The scaling factor (scaling parameter) in continuous wavelet transform is used to control the scaling of wavelet basis functions in the frequency domain, thus corresponding to different analysis frequency bands; The shift factor (time shift parameter) in continuous wavelet transform is used to control the position of the wavelet basis function in the time domain, corresponding to the time position of the signal being analyzed. The amplitude threshold is set based on signal-noise statistics; Step S3, synchronous compression: based on the chirp rate Phase compensation is performed on the wavelet coefficients, and scale recalculation is applied to achieve pre-shrinking of the scale domain, resulting in corrected wavelet coefficients. Step S4, Synchronous Extrusion: Based on the instantaneous angular frequency The energy carried by the modified wavelet coefficients is reassigned from the scale domain to the frequency axis to generate a synchronous squeeze time-frequency map with high energy concentration. ,in For synchronous extrusion time-frequency diagram The target frequency variable in the data is used to carry the instantaneous angular frequency components reconfigured from the scale domain. Step S5, Event Detection: Construct a saliency spectrum based on the time-frequency graph, extract peak regions through morphological processing and connectivity markers, and perform adaptive demodulation and envelope extraction along the ridge line in each peak region to estimate the arrival time, center frequency, bandwidth, and energy of the event; Step S6, Joint Decision: Form a joint decision statistic based on the super-background energy statistics of the peak region and the multi-resolution consistency score, and compare it with a threshold to complete the robust identification and auditable output of real partial discharge events.

[0027] In step S4, the modified wavelet coefficients are adjusted before energy reconfiguration. The joint robust weights are applied and calculated as follows: ; in, The amplitude threshold is set based on the noise statistics of the partial discharge signal; The chirp suppression parameter is set based on the maximum acceptable chirp rate of the partial discharge signal of the power equipment; It is a monotonically increasing threshold function; For the natural constant Exponential function operators with base 0.

[0028] In step S5, adaptive demodulation and envelope extraction include: in a single peak region Inner, ridge line Depend on Confirm; demodulated signal ,in Ridge The corresponding instantaneous angular frequency, The imaginary unit, The independent variable for demodulating the time integral in the exponential term. Envelope. ,in A low-pass filter set according to the partial discharge envelope bandwidth; the arrival time of the event is determined by... Sure.

[0029] Parameter estimation includes: center frequency By Peak Domain The frequency values ​​of all points within the time-frequency plot are used to determine their relative positions. The bandwidth is obtained by weighted averaging of the amplitude values. By Peak Domain The second moment of the internal frequency relative to the center frequency is calculated.

[0030] In step S6, the joint judgment statistic... The calculation is as follows: ; in, To characterize the approximate generalized likelihood ratio statistic of the peak region superbackground energy, The multi-resolution consistency score is obtained by changing the quality factor of the wavelet transform and calculating the overlap of the peak domain projection under different settings.

[0031] Combined Figure 2 As shown, the present invention also provides a partial discharge detection system based on synchronous compressed wavelet-synchronous squeezing for executing the aforementioned partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing, used for partial discharge detection of power equipment. The system includes a signal acquisition interface 1, a processor 2, a memory 3, a preprocessing unit 31, a wavelet transform and parameter estimation unit 32, a synchronous compression unit 33, a synchronous squeezing unit 34, a peak detection and demodulation unit 35, and a decision and output unit 36. The signal acquisition interface 1 is used to receive partial discharge signals from power equipment from at least one sensor.

[0032] Specifically, the signal acquisition interface 1 is adapted to at least one of ultra-high frequency (UHF), ultrasonic, or high-frequency current (HFCT) sensors. The processor 2 is coupled to the signal acquisition interface 1. The memory 3 stores a computer program, which, when executed by the processor 2, is configured to perform the following functions: a preprocessing unit 31, used to pre-filter and de-trend the partial discharge signal to obtain a standardized signal; a wavelet transform and parameter estimation unit 32, used to calculate the continuous wavelet transform coefficients of the standardized signal and estimate the instantaneous angular frequency and chirp rate; a synchronous compression unit 33, used to perform phase compensation and scale recalculation on the wavelet coefficients; a synchronous squeezing unit 34, used to reconfigure the energy from the scale domain to the frequency axis to generate a time-frequency graph; a peak detection and demodulation unit 35, used to construct a saliency spectrum, extract peaks and ridges, and perform adaptive demodulation to obtain event parameters; and a decision and output unit 36, used to complete event identification based on joint statistics and output auditable results.

[0033] It should be further explained that, when performing energy reconfiguration, the synchronous compression unit 34 is configured to apply the joint robust weight to the wavelet coefficients after synchronous compression.

[0034] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described thereon.

[0035] The present invention also provides an online partial discharge monitoring system, including the aforementioned partial discharge detection system, and a back-end server for time alignment and coherent fusion of the output results of multiple detection systems.

[0036] The preferred embodiments of the present invention will now be described in detail. The specific details described herein are for ease of understanding and not intended to limit the scope of the invention.

[0037] Example This embodiment provides a partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing, the specific steps of which are as follows: Step S1, Signal Preprocessing Acquire partial discharge signals from power equipment; these signals can originate from sensor channels such as UHF, ultrasonic, or HFCT. For the input discrete signals... Preprocessing includes: Bandpass filtering: The passband is set to the service bandwidth. Satisfying the Nyquist constraint ; Adaptive notch filtering: targeting known narrowband interference fundamental frequencies And its harmonics, using the transfer function as The notch filter is used for suppression, where the transfer function It can be expressed as follows: ; in, For digital filters in Complex variables in a domain The number of harmonic orders. The harmonic order is... Sampling rate, For the notch pole mode length and .

[0038] Detrending processing: Median filtering or low-order polynomial fitting is used to remove low-frequency drift components from the signal to obtain a detrended signal. ; Robust standardization: based on the median and median absolute deviation The signal is standardized to obtain a standardized signal: ; in It is a numerically stable term and .

[0039] It should be noted that the median and median absolute deviation They are respectively expressed by the following formulas: ; .

[0040] The above processing conditions include: the bandpass filter satisfies the Nyquist constraint and the passband ripple limit; the notch filter's zeros and poles form conjugate pairs to ensure stability and amplitude-frequency monotonicity; and the detrending window and polynomial order satisfy the residual whitening criterion. The final result... As input for subsequent steps.

[0041] Step S2, Time-Frequency Analysis For standardized signals Perform synchronous compressed wavelet transform: 1. Calculate the continuous wavelet transform coefficients: ; in For the selected mother wavelet, For scale parameters, These are the translation parameters.

[0042] 2. In amplitude mask Internal estimation of local instantaneous angular frequency: ; in This indicates taking the imaginary part. The amplitude threshold is set based on noise statistics. For translation variables The first-order partial derivative operator is used to calculate the local rate of change of wavelet coefficients with time position.

[0043] 3. Estimate local chirp rate: .

[0044] Step S3, synchronous compression Based on chirp rate Perform phase compensation and construct a local kernel function: ; Calculate the wavelet coefficients after phase correction : ; Implementation standards redefined: ; in For the scale rescaling coefficient, when Boundary clipping is applied when the scale exceeds the effective scale range; Output corrected wavelet coefficients .

[0045] Step S4, synchronous extrusion 1. Set the frequency grid Coverage of service bandwidth, of which, For synchronous extrusion, the frequency axis on the frequency graph is the first... Each frequency grid center (frequency sampling point) is used to discretely divide the frequency range within the service bandwidth; The parameter is the total number of frequency grids set, that is, the number of discrete grids (sampling points) on the frequency axis.

[0046] 2. Apply robust weights to the coefficients: ; in It is a monotonically increasing threshold function. For robust weight parameters, This is the chirp suppression parameter.

[0047] 3. A synchronous extrusion time-frequency diagram was obtained by implementing energy reallocation: Continuous form: ; in, For Dirac function, The normalization constant is The factor is used to compensate for scale measures and ensure energy conservation.

[0048] Discrete implementation: Quantization to the nearest frequency grid point And accumulate: .

[0049] Step S5, Event Detection 1. Background estimation and significance spectrum construction: ; ; in quantile and , It is a numerically stable term and .

[0050] 2. Peak region extraction: using Define peak region mask The candidate peak region set was obtained through morphological operations and connectivity labeling. .

[0051] 3. Ridge extraction and adaptive demodulation: In each peak region Internal ridge line determination: ; For standardized signals Demodulate: ; Extracting the envelope: ; in This is a low-pass filter operator.

[0052] 4. Parameter estimation: Arrival time: ; Center frequency: ; bandwidth: ; Event energy: ; Morphological indicators: Ridge curvature and duration . Step S6, Joint Judgment Calculate the approximate generalized likelihood ratio statistic: ; Calculate the multi-resolution consistency score , which is the overlap of the peak projection under different wavelet quality factor settings; Joint judgment statistics are formed: ; when When a valid partial discharge event is identified, all parameters and intermediate evidence are output, including... The decision threshold is determined by statistical analysis of historical data showing no partial discharge events. The distribution is determined based on the preset false alarm rate.

[0053] Implementation effect verification To verify the effectiveness of this invention, a dataset containing 3000 verified partial discharge events was collected at a 550kV substation for testing. Under a signal-to-noise ratio of -12dB, the detection rate of this invention reached 96.3%, with a false alarm rate of 0.7%. Compared with the traditional CWT method, the time-frequency resolution is improved by approximately 2.8 times, and the average single detection time is 48ms, meeting the requirements for near real-time monitoring.

[0054] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0055] Furthermore, it should be noted that the scope of the methods and systems in the embodiments of the present invention is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. In addition, features described with reference to certain examples may be combined in other examples.

[0056] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the scope of protection of the present invention, and all of these forms are within the protection scope of the present invention.

Claims

1. A partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing, used for partial discharge signal detection in power equipment, characterized in that, Includes the following steps: Signal preprocessing steps: The raw partial discharge signal collected from the power equipment is pre-filtered and de-trended to obtain a standardized signal. The pre-filtering includes bandpass filtering and adaptive notch filtering, wherein the passband of the bandpass filter is set to the service bandwidth. And satisfies the Nyquist constraint. The adaptive notch filter is designed for known narrowband interference fundamental frequencies. Suppressing its harmonics, among which, This is the lower limit frequency of the selected service bandwidth for the partial discharge signal in the current detection channel; This is the upper limit frequency of the selected service bandwidth for the partial discharge signal in the current detection channel; The sampling rate is used; the detrending process employs median filtering or low-order polynomial fitting to remove low-frequency drift components from the signal; the time-frequency analysis step involves performing continuous wavelet transform on the standardized signal to obtain wavelet coefficients. and in amplitude mask Internal estimation of instantaneous angular frequency and chirping rate ,in, The scaling factor in continuous wavelet transform; The translation factor in continuous wavelet transform; The amplitude threshold is set based on signal-noise statistics; the synchronous compression step is based on the chirp rate. Phase compensation is performed on the wavelet coefficients, and scale recalculation is applied to achieve pre-shrinking of the scale domain, resulting in corrected wavelet coefficients. Synchronous extrusion step: based on the instantaneous angular frequency The energy carried by the modified wavelet coefficients is reassigned from the scale domain to the frequency axis to generate a synchronous squeeze time-frequency map with high energy concentration. ,in For synchronous extrusion time-frequency diagram The target frequency variable in the event detection process is as follows: A saliency spectrum is constructed based on the time-frequency graph. Peak regions are extracted through morphological processing and connectivity markers. Adaptive demodulation and envelope extraction are performed along the ridges within each peak region to estimate the arrival time, center frequency, bandwidth, and energy of the event. A joint decision step is also performed: A joint decision statistic is formed based on the super-background energy statistics of the peak regions and the multi-resolution consistency score. This statistic is then compared with a threshold to complete robust identification and auditable output of real partial discharge events.

2. The partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing according to claim 1, characterized in that, In the synchronous extrusion step, the modified wavelet coefficients are adjusted before energy reallocation. The joint robust weights are applied and calculated as follows: ； in, The amplitude threshold is set based on the noise statistics of the partial discharge signal; The chirp suppression parameter is set based on the maximum acceptable chirp rate of the partial discharge signal of the power equipment; It is a monotonically increasing threshold function.

3. The partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing according to claim 1, characterized in that, The adaptive demodulation and envelope extraction in the event detection step include: in a single peak region Inner, ridge line Depend on Confirm; demodulated signal ,in Ridge The corresponding instantaneous angular frequency, The imaginary unit; envelope ,in A low-pass filter set according to the partial discharge envelope bandwidth; the arrival time of the event is determined by... Sure.

4. The partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing according to claim 3, characterized in that, The parameter estimation in the event detection step includes: center frequency. By Peak Domain The frequency values ​​of all points within the time-frequency plot are used to determine their positions on the time-frequency plot. The bandwidth is obtained by weighted averaging of the amplitude values. By Peak Domain The second moment of the internal frequency relative to the center frequency is calculated.

5. The partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing according to claim 1, characterized in that, Joint judgment statistics in the joint judgment step The calculation is as follows: ； in, To characterize the approximate generalized likelihood ratio statistic of the peak region superbackground energy, The multi-resolution consistency score is obtained by changing the quality factor of the wavelet transform and calculating the overlap of the peak domain projection under different settings.

6. A partial discharge detection system based on synchronous compressed wavelet-synchronous squeezing for performing the partial discharge detection method based on synchronous compressed wavelet-synchronous squeezing as described in any one of claims 1-5, for partial discharge detection in power equipment, characterized in that, include: A signal acquisition interface for receiving partial discharge signals from power equipment from at least one sensor; The processor is coupled to the signal acquisition interface; The memory stores a computer program, which, when executed by the processor, is configured to perform the functions of the following program units: a preprocessing unit for pre-filtering and detrending the partial discharge signal to obtain a standardized signal; a wavelet transform and parameter estimation unit for calculating the continuous wavelet transform coefficients of the standardized signal and estimating the instantaneous angular frequency and chirp rate; and a synchronization compression unit for performing phase compensation and scale recalculation on the wavelet coefficients. Synchronous squeezing unit is used to reallocate energy from the scale domain to the frequency axis to generate a time-frequency map; The peak detection and demodulation unit is used to construct the saliency spectrum, extract peak regions and ridges, and perform adaptive demodulation to obtain event parameters; The decision and output unit is used to identify events based on joint statistics and output auditable results.

7. The system according to claim 6, characterized in that, The signal acquisition interface is compatible with at least one of ultra-high frequency (UHF), ultrasonic, or high-frequency current (HFCT) sensors.

8. The system according to claim 6, characterized in that, When performing energy reconfiguration, the synchronous compression unit is configured to apply joint robust weights as defined in claim 2 to the synchronously compressed wavelet coefficients.

Images