High-voltage switch cabinet partial discharge ultrasonic detection denoising method

Abstract

The invention relates to a high-voltage switch cabinet partial discharge ultrasonic detection denoising method, and belongs to the field of power system and electrical equipment detection. The method comprises the following steps: converting a collected one-dimensional signal into a Hankel matrix, decomposing the signal by using an improved singular value decomposition method, performing Fourier transform on a noisy signal to obtain a frequency power spectrum of the noisy signal, then performing difference transform, separating narrow-band interference in the signal by using a threshold method and setting the narrow-band interference to zero, and finally reconstructing the matrix. The signal after the narrowband interference is removed can be obtained; dividing a noise interference mode and an effective signal mode by calculating an autocorrelation energy coefficient of each IMF and setting a critical coefficient; and performing threshold de-noising on the decomposed noise interference mode by using an improved threshold method, and superposing the processed noise interference mode and the effective signal mode to obtain a de-noised partial discharge signal. According to the invention, the signal-to-noise ratio of the partial discharge signal is improved, and the stability and readability of the signal are enhanced.

Description

Technical Field

[0001] The present invention belongs to the field of power system and electrical equipment detection, and relates to a method for denoising partial discharge ultrasonic detection of high-voltage switchgear. Background Technique

[0002] In the power system, as an important part of control and protection equipment, the health status of high-voltage switchgear is directly related to the safe and stable operation of the power system. The partial discharge phenomenon inside the high-voltage switchgear is one of the key factors leading to the decline of insulation performance, equipment aging and even failure. The ultrasonic signal generated by the partial discharge phenomenon can be used as an important basis for monitoring and diagnosing the state of high-voltage switchgear. However, due to the complex operating environment of high-voltage switchgear, the collected ultrasonic signals are often interfered by various noises, which are mainly composed of narrowband interference and white noise, seriously affecting the signal analysis and processing, reducing the signal-to-noise ratio (SNR) of the partial discharge signal, and making it difficult to accurately identify the fault characteristics.

[0003] Existing denoising techniques, such as wavelet transform denoising, filter denoising, etc., although have achieved certain effects in some applications, still have some limitations. For example, although wavelet transform denoising can provide multi-scale signal analysis, it often needs to rely on experience when selecting thresholds and processing strategies, and the processing effect on non-stationary signals is limited. Filter denoising is simple and easy to implement, but it is difficult to adapt to the time-varying characteristics of signals and noises, especially when the noise characteristics are unknown or changing. In addition, these methods may introduce signal distortion during denoising, affecting subsequent signal analysis and feature extraction. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a method for denoising partial discharge ultrasonic detection of high-voltage switchgear, which can effectively remove narrowband noise and white noise, improve the signal-to-noise ratio of the partial discharge signal, and enhance the stability and readability of the signal.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] A method for denoising partial discharge ultrasonic detection of high-voltage switchgear, comprising the following steps:

[0007] Using the improved singular value decomposition method to remove the narrowband interference of the collected partial discharge signal;

[0008] Using the complete ensemble empirical mode decomposition method CEEMD to decompose the signal into a series of intrinsic mode functions IMF;

[0009] Calculating the autocorrelation energy coefficient of each IMF and setting the critical coefficient to divide the noise interference mode and the effective signal mode;

[0010] Use an improved threshold method to perform threshold denoising on the noise interference mode;

[0011] Superimpose the processed noise interference mode and the effective signal mode to obtain the denoised partial discharge signal.

[0012] Furthermore, the improved singular value decomposition method includes the following steps:

[0013] Convert the collected one-dimensional signal F(t) = {f(1), f(2), f(3), ……, f(N)} into a Hankel matrix;

[0014] Perform singular value decomposition on the Hankel matrix;

[0015] Perform Fourier transform on the noisy signal to obtain its frequency power spectrum;

[0016] Take the absolute value of the difference between the amplitudes of two adjacent spectral points to form a new spectrum denoted as ΔP(n), that is, the difference transformation;

[0017] Introduce the classical threshold formula Determine the number of narrowband interference noises

[0018] Set the first 2η singular values decomposed to zero;

[0019] Reconstruct the Hankel matrix.

[0020] Furthermore, the complete ensemble empirical mode decomposition method CEEMD includes the following steps:

[0021] Preprocess the signal by adding positive and negative paired white noises;

[0022] Use the empirical mode decomposition EMD to decompose the preprocessed signal.

[0023] Furthermore, the calculation method of the autocorrelation energy coefficient is as follows:

[0024] Calculate the autocorrelation function R m (τ) = E[m(t)m(t + τ)];

[0025] Normalize the autocorrelation function

[0026] Calculate the second norm of the normalized autocorrelation function ρ m That is, the autocorrelation energy coefficient.

[0027] The beneficial effects of the present invention are as follows:

[0028] ​(1) By improving the singular value decomposition method and the CEEMD autocorrelation judgment threshold denoising method, narrowband interference and white noise in the signal can be effectively removed, and the signal-to-noise ratio of the partial discharge signal can be improved.

[0029] (2) After removing the noise interference, the stability and readability of the partial discharge signal are significantly enhanced, which is beneficial to subsequent signal analysis and feature extraction, thereby improving the accuracy of fault diagnosis.

[0030] (3) By denoising the partial discharge signal, the fault characteristics can be more clearly identified, thereby improving the efficiency and accuracy of fault diagnosis.

[0031] (4) By promptly detecting and handling the partial discharge faults inside the high-voltage switchgear, equipment aging, insulation performance degradation and even faults can be effectively prevented, thereby ensuring the safe and stable operation of the power system.

[0032] Other advantages, objectives and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be described in detail preferably with reference to the accompanying drawings, where:

[0034] Figure 1 is the flowchart of partial discharge denoising;

[0035] Figure 2 is the flowchart of CEEMD decomposition. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] The following uses specific specific examples to illustrate the embodiments 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 also 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 drawings provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0037] Among them, the attached drawings are only for illustrative purposes, showing only schematic diagrams rather than actual physical diagrams, and should not be construed as limiting the present invention; for better illustration of the embodiments of the present invention, some components in the attached drawings will be omitted, enlarged or reduced, which do not represent the dimensions of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the attached drawings may be omitted.

[0038] In the attached 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 there are terms such as "upper", "lower", "left", "right", "front", "rear", etc. indicating the orientation or positional relationship, they are based on the orientation or positional relationship shown in the attached drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the positional relationship in the attached drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to specific circumstances.

[0039] Please refer to Figure 1 and Figure 2 , the present invention includes the following steps:

[0040] 1. Use the improved singular value decomposition method to remove the narrowband interference of the collected partial discharge signals

[0041] 1.1 Construct the Hankel matrix

[0042] Convert the collected one-dimensional signal F(t) = {f(1), f(2), f(3), ……, f(N)} into a Hankel matrix, that is

[0043]

[0044] where: f(n) represents the nth data point collected, N represents the number of collected data, m + n - 1 = N, and take n = N / 2.

[0045] Perform singular value decomposition on the matrix H, that is:

[0046] H = USV T (2)

[0047] u and v are orthogonal matrices of m×m and n×n respectively, and S is a non-negative diagonal matrix, that is:

[0048] S = diag(a 1 , a 2 , …, a i , …, a r )(3)

[0049] where: ai denotes the singular value and satisfies a 1 > a 2 > … a r , r = min(m, n).

[0050] 1.2 Singular value extraction of the noise signal and the effective signal

[0051] Perform Fourier transform on the noisy signal to obtain its frequency power spectrum, and take the absolute value of the difference between the amplitudes of two adjacent spectrum points to form a new spectrum, that is, the difference transform, denoted as:

[0052]

[0053] Introduce the classical threshold formula as follows:

[0054]

[0055] Determine the number of narrowband interference noises η:

[0056]

[0057] 1.3 Reconstruct the Hankel matrix

[0058] Set the first 2η singular values decomposed to zero to obtain a new non - negative diagonal matrix S', and reconstruct the Hankel matrix, that is, H = US'V T , and a new one - dimensional partial discharge signal sequence x(t) without narrowband interference can be obtained.

[0059] 2. Use the CEEMD autocorrelation - based threshold denoising method to remove white noise from the signal

[0060] 2.1 Use the EMD decomposition method to decompose the signal into a series of IMFs

[0061] Perform I - times of EMD decomposition on the noisy signal x(t), and define the sum of averages as IMF 1 (t):

[0062]

[0063] Among them, x(t) is the partial discharge signal to be decomposed, ω i is the white noise added at each stage of the decomposition process, ε k is the addition ratio of white noise at each stage, and E j (·) is the j - th mode obtained after the original partial discharge signal is decomposed by EMD.

[0064] For k = 1, 2…K, calculate the k - th order residual r k (t):

[0065] rk r(t) = r k-1 r(t) - IMF k r(t)(9)

[0066] Extract r using EMD k r(t) + ε k E k r(ω i )'s IMF 1 component of r(t), and calculate the overall average value to obtain the IMF of the target signal k+1 r(t):

[0067]

[0068] Repeat the screening until the residual can no longer be decomposed to obtain the final residual R(t):

[0069]

[0070] So far, the decomposition is completed, and the target signal x(t) can be expressed as:

[0071]

[0072] 2.2 Calculate the autocorrelation energy coefficient to distinguish the effective signal mode and the noise interference signal mode

[0073] Calculate the autocorrelation function of each IMF, that is

[0074] R m R(τ) = E[m(t)m(t + τ)](13)

[0075] Normalize the autocorrelation function, that is

[0076]

[0077] Calculate the second norm of the normalized autocorrelation function ρ m (τ), called the autocorrelation energy coefficient, that is

[0078]

[0079] 2.3 Use the improved threshold method to perform threshold denoising on the noise interference mode

[0080]

[0081] Among them, T j is the threshold of the component IMF j , and the calculation method is:

[0082]

[0083] where median(·) is the function to calculate the absolute median of each modal component, and N is the number of sampling points.

[0084] Superimpose the processed noise interference mode and the effective signal mode to obtain the denoised partial discharge signal:

[0085]

[0086] Embodiment

[0087] 1. Data acquisition

[0088] Use an ultrasonic sensor to collect the ultrasonic signals generated by partial discharge in the high-voltage switchgear. The sampling frequency is 50 kHz, and the sampling time is 1 second, obtaining a signal sequence containing 50,000 data points.

[0089] 2. Remove narrowband interference using the improved singular value decomposition method

[0090] Convert the collected signal sequence into a Hankel matrix and perform singular value decomposition. Calculate the threshold T = 0.6745 and determine the number of singular values h containing noise signals as 25. Set the first 50 singular values to zero and reconstruct the Hankel matrix to obtain the signal sequence after removing narrowband interference.

[0091] 3. Remove white noise using the CEEMD autocorrelation judgment threshold denoising method

[0092] Perform CEEMD decomposition on the denoised signal sequence to obtain 10 IMFs and a residual term. Calculate the autocorrelation energy coefficient of each IMF and set the critical coefficient α = 0.1. Divide the IMFs with autocorrelation energy coefficients less than 0.1 into noise interference modes, and the rest into effective signal modes.

[0093] Perform threshold processing on each noise interference mode, and superimpose the effective signal mode and the denoised noise interference mode to obtain the final denoised signal sequence.

[0094] 4. Result analysis

[0095] Perform spectral analysis on the signal before and after denoising. It can be seen that the denoised signal effectively removes narrowband interference and white noise, and the signal-to-noise ratio is significantly improved.

[0096] Perform time-domain analysis on the denoised signal. It can be seen that the stability and readability of the signal are enhanced, which is beneficial for subsequent signal analysis and fault diagnosis.

[0097] This embodiment verifies the effectiveness of the ultrasonic detection denoising method for partial discharge in high-voltage switchgear proposed in the patent. This method can effectively remove narrowband interference and white noise in the signal, improve the signal-to-noise ratio of the signal, enhance the stability and readability of the signal, and provide a reliable signal basis for the diagnosis of partial discharge faults in high-voltage switchgear.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the present technical solution, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for denoising partial discharge ultrasonic detection of high-voltage switch cabinet, characterized in that: The following steps are involved: The improved singular value decomposition method is used to remove the narrowband interference of the collected partial discharge signal; The signal is decomposed into a series of intrinsic mode functions (IMFs) using the complete set empirical decomposition method (CEEMD). Calculate the autocorrelation energy coefficient of each IMF and set the critical coefficient to divide the noise interference mode and the effective signal mode; The improved threshold method is used to perform threshold denoising on the noise interference mode; The processed noise interference mode is superimposed on the effective signal mode to obtain the denoised partial discharge signal.

2. A method for denoising partial discharge ultrasonic detection of high-voltage switch cabinet according to claim 1, characterized in that: The improved singular value decomposition method comprises the following steps: Convert the collected one-dimensional signal F(t) = {f(1), f(2), f(3), ..., f(N)} into a Hankel matrix; Perform singular value decomposition on the Hankel matrix; Perform Fourier transform on the noise-stained signal to obtain its frequency power spectrum; The absolute value of the difference between the amplitudes of two adjacent spectrum points is taken to form a new spectrum, which is recorded as ΔP(n), i.e., difference transformation; Introducing the classic threshold formula Determining the amount of narrowband interference noise Set the first 2η decomposed singular values ​​to zero; Reconstruct the Hankel matrix.

3. A method for denoising partial discharge ultrasonic detection of high-voltage switch cabinet according to claim 1, characterized in that: The complete set empirical decomposition method CEEMD comprises the following steps: The signal is preprocessed by adding positive and negative paired white noise; The preprocessed signal is decomposed using empirical mode decomposition (EMD).

4. A method for denoising partial discharge ultrasonic detection of high-voltage switch cabinet according to claim 1, characterized in that: The autocorrelation energy coefficient is calculated as follows: Calculate the autocorrelation function R of each IMF m (τ) = E[m(t)m(t+τ)]; Normalize the autocorrelation function Calculate the normalized autocorrelation function ρ m The second norm of (τ) That is, the autocorrelation energy coefficient.

Images