A Simulation Calculation Method for Partial Discharge Pulse Width Based on SG Smoothing Filtering

Abstract

This invention discloses a method for simulating and calculating the pulse width of partial discharge pulses based on S-G smoothing filtering. The method includes the following steps: generating a first signal, which is an analog signal comprising a base signal and a noise signal; preprocessing the first signal to obtain a second signal; initializing the S-G filter parameters and processing the second signal based on the S-G filter to obtain a third signal; selecting the peak pulse width interval of the third signal, calculating the pulse width ratio, and determining whether the constraints are met; repeating the above steps to verify the algorithm and iteratively optimize it. This invention can effectively suppress noise and improve the accuracy and reliability of pulse width calculation. Field tests and simulation experiments have verified the feasibility and effectiveness of the algorithm, providing strong support for partial discharge monitoring of power equipment. It can effectively distinguish the pulse widths of different types of partial discharge signals, which is helpful for further fault diagnosis and maintenance. It is applicable to partial discharge monitoring equipment from different types and manufacturers, and has universal applicability.

Description

Technical Field

[0001] This invention relates to the field of partial discharge detection technology, and in particular to a method for simulating and calculating the pulse width of partial discharge pulses based on SG smoothing filtering. Background Art

[0002] Partial discharge (PD) is a discharge phenomenon that occurs in localized areas of the insulation layer of high-voltage electrical equipment under the influence of a strong electric field. This phenomenon not only leads to the degradation of the insulation medium but may also cause more serious electrical faults. Pulse width is the duration of a high-level signal. Although the types and locations of partial discharges have been extensively studied, research on the pulse width characteristics of ultra-high frequency partial discharge pulses is relatively limited due to the sporadic and irregular generation of partial discharge pulse signals in power equipment and the complex on-site conditions. Furthermore, directly measuring these pulse widths presents numerous difficulties in practical applications.

[0003] Traditional partial discharge detection methods struggle to effectively distinguish between partial discharge signals and noise signals when processing signals in complex field environments, and their accuracy and reliability in pulse width calculation are limited. Therefore, an algorithm capable of effectively calculating the pulse width of ultra-high frequency partial discharge pulses is needed to improve the accuracy and reliability of partial discharge monitoring. Summary of the Invention

[0004] The purpose of this invention is to provide a simulation calculation method for partial discharge pulse width based on SG smoothing filtering, which can solve the problems of accuracy and reliability in calculating ultra-high frequency partial discharge pulse width.

[0005] The present invention solves the technical problem by adopting the following technical solution:

[0006] A method for simulating and calculating the pulse width of a partial discharge pulse based on SG smoothing filtering includes the following steps: generating a first signal, which is an analog signal including a base signal and a noise signal; preprocessing the first signal to obtain a second signal; initializing SG filter parameters and processing the second signal based on the SG filter to obtain a third signal; selecting the peak pulse width interval of the third signal, calculating the pulse width ratio, and determining whether the constraint conditions are met; repeating the above steps to verify the algorithm and iteratively optimize it.

[0007] Preferably, before generating the first signal, the feasibility of the partial discharge pulse width simulation calculation method is first calculated, the original pulse signal waveform is noise suppressed and denoised by using an SG filter, the data points of the original pulse signal are fitted with local polynomials, the smoothing value is estimated, and the noise suppression result is obtained.

[0008] Preferably, the signal preprocessing involves taking the absolute value of the first signal.

[0009] Preferably, the SG filter parameters include: window width and polynomial order, wherein the window width is selected as an odd number, and the window is selected from the original pulse signal waveform data points;

[0010] The SG filter performs smoothing and noise reduction on the second signal.

[0011] Preferably, the window width is in the range of 85 to 100, and the polynomial order is 3.

[0012] Preferably, the peak value range of the third signal pulse width is selected from 10% to 30%.

[0013] Preferably, the constraint condition is: the ratio of the obtained third signal pulse width to the first signal pulse width is less than or equal to 5%; if this condition is not met, the SG filter parameters are optimized and adjusted; if this condition is met, the process ends.

[0014] Preferably, the noise signal is Gaussian white noise.

[0015] A method for calculating the pulse width of partial discharge involves acquiring the partial discharge signal from the field, preprocessing the partial discharge signal, then performing smoothing and denoising processing based on an SG filter, setting the peak range of the partial discharge signal, and calculating the pulse width.

[0016] The partial discharge pulse width simulation calculation method based on SG smoothing filtering proposed in this invention can effectively suppress noise and improve the accuracy and reliability of pulse width calculation. The feasibility and effectiveness of the algorithm have been verified through field tests and simulation experiments, providing strong support for partial discharge monitoring of power equipment. It can effectively distinguish the pulse width of different types of partial discharge signals, which is helpful for further fault diagnosis and maintenance. It is applicable to partial discharge monitoring equipment of different types and manufacturers and has universal applicability. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the simulation algorithm flow of the present invention;

[0018] Figure 2 This is a schematic diagram of the original pulse signal waveform of the present invention;

[0019] Figure 3 The waveform after smoothing filtering by the SG filter of this invention;

[0020] Figure 4 This is a schematic diagram of the simulation analysis and calculation of the pulse signal with a pulse width of 61ns in this invention;

[0021] Figure 5 This is a schematic diagram of the simulation analysis and calculation of the pulse signal with a pulse width of 90ns in this invention;

[0022] Figure 6 This is a schematic diagram of the simulation analysis and calculation of the pulse signal pulse width of 89.6ns in this invention;

[0023] Figure 7 This is a schematic diagram of the simulation analysis and calculation of the pulse signal pulse width of 91.6ns in this invention;

[0024] Figure 8 This is a schematic diagram of the simulation analysis and calculation of the pulse signal pulse width of 89.8ns in this invention;

[0025] Figure 9 This is a schematic diagram of the simulation analysis and calculation of the pulse signal with a pulse width of 66ns in this invention;

[0026] Figure 10 This is a schematic diagram of the insulation discharge pulse width analysis and calculation of the present invention;

[0027] Figure 11 This is a schematic diagram of the tip discharge pulse width analysis and calculation of the present invention;

[0028] Figure 12 This is a schematic diagram illustrating the surface discharge pulse width analysis and calculation of the present invention;

[0029] Figure 13 This is a schematic diagram illustrating the calculation of the suspended discharge pulse width in this invention. Detailed Implementation

[0030] The technical solution of the present invention will be further described below with reference to the embodiments and accompanying drawings.

[0031] Example 1

[0032] Due to the influence of complex working conditions on site, the partial discharge pulse signal received by the UHF sensor is often mixed with noise signals. In order to reduce the impact of noise on the partial discharge pulse signal, it is necessary to suppress and denoise this noise. This invention selects Savitzky-Golay (SG) filter for denoising and smoothing. The SG filter estimates the smoothing value by performing local polynomial fitting on the partial discharge pulse waveform data points. The feasibility of the partial discharge pulse width calculation method is evaluated based on the denoising results.

[0033] In a further implementation of this embodiment, the SG filter parameters include: window width and polynomial order, the window width is selected as an odd number, and the window is selected from the original pulse signal waveform data;

[0034] A further implementation of this embodiment, specifically the feasibility calculation, involves fitting the original partial discharge waveform data points within the window using a k-1 degree polynomial, as shown in the formula:

[0035] y i =α0+a1i+αz i 2 +…+α k-1 i k-1 (1)

[0036] In the formula, i = (-m, -m+1, ..., 0, 1, ..., m-1, m), a j (j = 0, 1, 2, ..., k-1) are the fitting parameters;

[0037] In a further implementation of this embodiment, the k-variable linear equation system includes: n k-1 degree polynomials used to solve for k fitting parameters, with a window width range of n≥k;

[0038] In a further implementation of this embodiment, if n = k, the fitting parameters are solved using linear algebra; if n > k, the least squares method is used.

[0039] In a further implementation of this embodiment, the above n polynomials are combined to obtain the operation matrix as follows:

[0040]

[0041] Simplified to an overdetermined system of equations:

[0042] Y (2m+1)×1 =X (2m+1)×k -A kx1 +E (2m+1)×1 (3)

[0043] The least squares solution to A is:

[0044]

[0045] The filter value for Y is:

[0046]

[0047] In the formula, B = X·(X T ·X) -1 ·X T B is the filter coefficient matrix, which is determined by the X matrix alone. The B matrix is ​​a (2m+1)×(2m+1) order matrix. The SG smoothing fitting equation is obtained based on the coefficient matrix.

[0048] In a further implementation of this embodiment, the window width of the SG smoothing filter is set to 2m+1, and the number of original waveform pulse data points n is also 2m+1.

[0049] Further implementations of this embodiment, such as... Figure 2 The image shows the original pulse signal waveform, which is then processed by SG smoothing filter as follows: Figure 3As shown, noise in the partial discharge pulse signal is effectively suppressed and the waveform trend is enhanced, improving the readability of the data. Therefore, the partial discharge pulse width calculation algorithm is feasible.

[0050] Example 2

[0051] like Figure 1 As shown, a method for simulating and calculating the pulse width of partial discharge pulses based on SG smoothing filtering includes the following steps:

[0052] S1, select the base signal;

[0053] S2, add Gaussian white noise;

[0054] S3, Signal preprocessing: All data are taken as positive values;

[0055] S4, SG filter: Initializes SG filter parameters and performs smoothing and noise reduction processing;

[0056] S5, Set signal start and end positions: Select S% of the signal peak value as the start and end positions;

[0057] S6, Calculate the pulse width ratio: Calculate the ratio of the obtained pulse width to the base signal pulse width, and determine whether the constraint condition is met. Constraint condition: The ratio of the obtained pulse width to the base signal pulse width is less than or equal to 5%.

[0058] S7, if not satisfied, optimize and adjust the SG filter parameters;

[0059] S8, if satisfied, end;

[0060] S9, Iterative optimization: Select base signals with different pulse widths, repeat the above steps, and verify the algorithm.

[0061] In a further embodiment of this example, the first signal is an analog signal, including a base signal and a noise signal. The base signal includes surface discharge, floating discharge, insulating discharge, tip discharge, and 15 combinations thereof, as shown in Table 1. The noise signal is Gaussian white noise.

[0062] Table 1 lists 15 types of base signals.

[0063]

[0064] Further implementations of this embodiment, such as... Figure 4 As shown, the base signal pulse width is set to 61ns, and Gaussian white noise is added to generate the first signal to simulate the partial discharge signal in the field. The partial discharge monitoring device is used for sampling.

[0065] In a further implementation of this embodiment, the signal needs to be preprocessed before SG filtering. All data are taken as positive values ​​to obtain the second signal, which preserves the integrity of the first signal data and facilitates subsequent processing.

[0066] In a further implementation of this embodiment, the SG filter parameters include: window width and polynomial order. The window width is used to adjust the number of data points during local fitting. It is chosen to be an odd number so that there are an equal number of data points near the center point. The larger the value, the more obvious the smoothing effect. The polynomial order determines the complexity of the polynomial used to fit the data. Lower orders of 1 or 2 are suitable for simple data structures, while higher orders of 3 or 4 are suitable for complex data structures.

[0067] In a further implementation of this embodiment, through simulation experiments, it was found that when the window width is between 85 and 100, the polynomial order is selected as order 3, the mode is selected as "nearest" or "interp", and the signal peak value S = 20% is selected as the start and end position, the ratio of the calculated pulse width to the base signal pulse width is less than or equal to 5%, and the signal start and end position is the pulse width peak range.

[0068] In a further implementation of this embodiment, the signal start and end position S is set to 20% in the simulation experiment. Based on on-site conditions and experience, the range of S can be adjusted between 10% and 30%, specifically 10%, 15%, and 30%. A window width of 99 yields the best results, and its value is optimized and adjusted within the range of 85 to 100 based on the partial discharge signal collected on-site. The polynomial order is 3.

[0069] In a further implementation of this embodiment, the pulse width analysis results of the third signal in the simulation experiment are shown in the figure. The blue box represents the calculated pulse range, and the calculated pulse width is 61.8 ns. (61.8-61)÷61×100%=1.31%<5%, indicating that the simulation calculation method proposed in this embodiment is feasible and effective.

[0070] In a further implementation of this embodiment, a fault partial pulse signal collected from the field by a monitoring device is used for verification. The parameters of the SG filter are set as follows: window width: 99, polynomial order: 3, mode: "nearest". 20% of the signal peak value S is selected as the start and end positions of the partial discharge pulse waveform. The analysis results are as follows: Figure 5-9 As shown in Table 2, the calculated pulse width values ​​are 90ns, 89.6ns, 91.6ns, 89.8ns, and 88ns. The pulse width values ​​fluctuate around the average value of 89.8ns, and the maximum deviation is less than 5% (2.0%), which further illustrates the feasibility and effectiveness of the calculation method in this embodiment.

[0071] Table 2 Calculation result of partial discharge pulse width value

[0072]

[0073] Example 3

[0074] This embodiment provides a method for calculating the pulse width of partial discharge. The method involves acquiring the partial discharge signal in the field, preprocessing the partial discharge signal, smoothing and denoising it based on an SG filter, setting the peak range of the partial discharge signal, and calculating the pulse width.

[0075] In a further implementation of this embodiment, different types of on-site partial discharge pulse signals are analyzed, such as... Figure 10-13 As shown, four typical partial discharge fault pulse signals of high-voltage electrical equipment, namely insulation discharge, tip discharge, surface discharge, and floating discharge, were selected. Using the calculation method of Example 3, their pulse widths were obtained as 189.2, 267.8, 142.9, and 389.8, respectively. These values ​​conform to the waveform characteristics of their corresponding partial discharge pulses, and the values ​​have a high degree of differentiation, indicating that different types of partial discharge pulse widths can be well identified, further demonstrating the accuracy of this embodiment.

[0076] The order of the above embodiments is for ease of description only and does not represent the superiority or inferiority of the embodiments.

[0077] 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 foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for simulating and calculating the pulse width of partial discharge pulses based on SG smoothing filtering, characterized in that, The algorithm includes the following steps: generating a first signal, which is an analog signal comprising a base signal and a noise signal; preprocessing the first signal to obtain a second signal; initializing SG filter parameters and processing the second signal based on the SG filter to obtain a third signal; selecting the peak pulse width interval of the third signal, calculating the pulse width ratio, wherein the pulse width ratio is the ratio of the difference between the pulse width of the third signal and the pulse width of the first signal to the pulse width of the first signal; determining whether a constraint condition is met, wherein the constraint condition is: the ratio of the difference between the pulse width of the third signal and the pulse width of the first signal to the pulse width of the first signal is less than or equal to 5%; if not met, optimizing and adjusting the SG filter parameters; if met, ending the algorithm; repeating the above steps to verify the algorithm and iteratively optimize it. Before generating the first signal, the feasibility of the partial discharge pulse width simulation calculation method is calculated. The original pulse signal waveform is noise suppressed and denoised by using an SG filter. The data points of the original pulse signal are fitted with a local polynomial to estimate the smoothing value and obtain the noise suppression result. The preprocessing of the first signal involves taking the absolute value of the first signal. The peak value range of the third signal pulse width is selected from 10% to 30%.

2. The method for simulating and calculating the pulse width of partial discharge pulses based on SG smoothing filtering according to claim 1, characterized in that, The SG filter parameters include: window width and polynomial order, wherein the window width is selected as an odd number, and the window is selected from the original pulse signal waveform data points; The SG filter performs smoothing and noise reduction on the second signal.

3. The method for simulating and calculating the pulse width of partial discharge pulses based on SG smoothing filtering according to claim 2, characterized in that, The window width ranges from 85 to 100, and the polynomial order is 3.

4. The method for simulating and calculating the pulse width of partial discharge pulses based on SG smoothing filtering according to claim 1, characterized in that, The noise signal is Gaussian white noise.

5. A method for calculating the pulse width of a partial discharge pulse, characterized in that, The partial discharge signal is acquired, preprocessed, and then smoothed and denoised using an SG filter. The peak range of the partial discharge signal is set, and the pulse width is calculated.

Images