Voltage sag detection method based on Kaiser window function optimization S transformation
By using the fast S transformation algorithm optimized by Kaiser window function in power drop detection, the problem that traditional methods cannot effectively detect voltage drop is solved, and higher detection accuracy and accuracy are achieved, suitable for complex power grid system environments.
Patent Information
- Application Number
- CN202410864469.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The traditional power drop detection method cannot effectively perform local detection and is sensitive to noise, making it difficult to accurately monitor voltage drop disturbances in complex power grid system environments.
Using the fast S transformation algorithm based on Kaiser window function, the Kaiser optimized window function is constructed and the S transformation algorithm is optimized to improve the energy aggregation performance of time-frequency integral transformation, and the precise detection of voltage drop signals is achieved.
It improves the accuracy and accuracy of voltage drop disturbance detection, reduces noise interference, and enhances the detection ability of the fundamental amplitude of the power grid signal.
Smart Images

Figure CN119003952A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of signal fault detection, and in particular relates to a voltage sag detection method based on Kaiser window function optimized S transformation. Background Art
[0002] The power system is one of the indispensable infrastructures in modern society, and power quality is one of the important factors affecting the normal operation and stable power supply of the power system. Power sag is a common fault in power quality problems and one of the main causes of power outages and equipment damage. Therefore, effective detection and analysis of power sag is of great significance to ensure the safe and stable operation of the power system.
[0003] Traditional power sag detection methods mainly rely on the instantaneous changes of voltage and current signals. However, the traditional Fourier transform method has some limitations, such as the need to transform the entire signal and the inability to perform local detection of the sag signal. In addition, the traditional Fourier transform method is sensitive to noise signals, and noise will interfere with the detection results of the sag signal. Therefore, in order to overcome the shortcomings of the traditional method, a fast KS transform algorithm S based on the Kaiser window is designed to detect the power sag signal in the power quality disturbance. Therefore, how to design a voltage sag detection method based on the Kaiser window function optimized S transform to address the difficulty of monitoring voltage sag disturbances in complex power grid systems is an urgent problem to be solved by those skilled in the art. Summary of the invention
[0004] In view of the defects in the prior art, the present invention provides a voltage sag detection method based on Kaiser window function optimized S transform.
[0005] The present invention comprises the following steps:
[0006] S1: By using the time- and frequency-dependent Kaiser window as the integral transform kernel function, an S-transform algorithm based on the Kaiser window function is constructed;
[0007] S2: Optimizing the Kaiser window function to obtain a Kaiser optimized window function suitable for power quality disturbance detection;
[0008] S3: constructing a fast KS transformation algorithm based on the Kaiser optimized window function using the Kaiser optimized window function;
[0009] S4: Detect voltage sag disturbance signal based on the fast KS transformation algorithm of the constructed Kaiser optimized window function;
[0010] Furthermore, based on the Kaiser window function S-transform algorithm, for a given signal to be measured x(t)∈L2 (R), the KS transformation KS(τ, f):R→C of the signal x(t) is defined as
[0011]
[0012] In the above formula, τ is the time shift factor that controls the position of the Kaiser window in the time domain; I 0 (·) is the first kind of zero-order modified Bessel function.
[0013] Further, in step S2, the Kaiser optimization window function expression is:
[0014]
[0015] In the above formula, η 1 and η 2 is the Kaiser window adjustment parameter; parameter η 1 ≥0 is used to adjust the rate of change of the window shape as the detection frequency changes; η 2 >0 can easily adjust the time-frequency resolution at a certain frequency; I 0 (·) is the first kind of zero-order modified Bessel function described by equation (7).
[0016] Further, in step S3, the Kaiser optimized window function is used to construct a fast KS transformation algorithm based on the Kaiser optimized window function, which specifically includes:
[0017] (1) Construct the KS transform based on Kaiser optimization window, expressed as
[0018]
[0019] (2) The continuous and discrete expressions of the KS transform of the Kaiser optimization window are calculated; the discrete expression implemented by the fast KS transform is:
[0020]
[0021] In the above formula, It is expressed as the discrete form of the Fourier transform of the signal frequency shift X(ξ+f); It is expressed as the window function w in formula (6) k (t, f) about the time τ to ξ spectrum W k The discrete form of (ξ, f).
[0022] Furthermore, in step S4, the fast KS transform algorithm selection method based on the constructed Kaiser optimized window function adopts the fast Fourier transform and its inverse transform to replace the discrete Fourier transform and its inverse transform, and at the same time, non-characteristic frequency points can be eliminated, specifically including:
[0023] Step 1: uniformly sample the measured signal x(t) to obtain a discrete signal x[n] with a sampling length of N, where n=0, 1, 2, ..., N-1, and calculate the Fourier transform spectrum of x[n];
[0024] Step 2: For frequency point k, k=0,1,2,···,M-1, Select characteristic frequency points and eliminate non-characteristic frequency points to obtain a specific frequency shift vector X[l+k];
[0025] Step 3: Calculate the Kaiser window function spectrum W at the characteristic frequency point k through Fourier transform k [l, k];
[0026] Step 4: For frequency point k, calculate the frequency shift vector X[l+k] in step 2 and the window function spectrum vector W in step 3 k Hadamard product of [l, k];
[0027] Step 5: Perform an inverse Fourier transform from frequency l to time m on the product vector obtained in step 4 to obtain KS transform time domain information of the signal at frequency point k;
[0028] Step 6: Repeat steps 2 to 5 to obtain the KS transform time domain information corresponding to all detection frequency points of the signal, and construct the signal KS transform complex time-frequency matrix;
[0029] Step 7: Extract the amplitude and phase angle through KS transformation by using the formulas described in equations (19) and (20) to extract the time domain amplitude information and phase information of the signal at each characteristic frequency point;
[0030] The parameter selection method of the fast KS transform algorithm based on the constructed Kaiser optimization window function is analyzed. In the process of implementing the KS transform, the Kaiser window adjustment function is used as
[0031]
[0032] Among them, [NT s f]+1 represents the frequency point representing frequency f in the calculation. For example, when the detection frequency is f=0, the corresponding frequency point is [NT s f]+1=1, the detection frequencies reflected in the calculation results are all able to make NT sf is the frequency of an integer; the adjustment parameter η in the Kaiser window adjustment function 1 and η 2 Determine so that the Kaiser window adjustment function has α[f 0 ]∈[90,110] holds, where f 0 Indicates the fundamental frequency, ensuring that the KS transform has a better time resolution for the signal fundamental so as to accurately detect the fundamental amplitude.
[0033] The present invention provides a voltage sag detection method based on Kaiser window function optimized S transform, which has at least the following beneficial effects:
[0034] The present invention is based on the S transform (KS transform) algorithm of the Kaiser window function, which can improve the energy aggregation performance of the time-frequency integral transform and give full play to the performance of the window function.
[0035] The KS transform of the Kaiser optimization window constructed by the present invention has a Kaiser optimization window function with lower side lobes in the low frequency band and narrower main lobe width in the high frequency band, which is beneficial to improving the detection accuracy of the fundamental amplitude of the power grid signal, thereby improving the detection accuracy of the voltage sag disturbance. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 It is a flow chart of a voltage sag detection method based on Kaiser window function optimized S transform;
[0037] Figure 2 A schematic diagram of the KS transformation fast implementation process of a voltage sag detection method based on Kaiser window function optimized S transformation;
[0038] Figure 3 It is a comparison chart of the detection results of voltage sag disturbance by KS transform of a voltage sag detection method based on Kaiser window function optimized S transform and the control algorithm S transform;
[0039] in Figure 3 (a) Figure 3 (b) Figure 3 (c) and Figure 3 (d) are the detection results under different voltage sag disturbance parameters. DETAILED DESCRIPTION
[0040] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0041] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings.
[0042] like Figure 1 As shown, the present invention provides a voltage sag detection method based on Kaiser window function optimization S transform, which specifically includes the following steps:
[0043] Step S1: Using the time- and frequency-dependent Kaiser window as the integral transform kernel function, the S transform (KS transform) algorithm based on the Kaiser window function
[0044] Specifically, the voltage sag is manifested as a temporary drop in the amplitude of the voltage signal over a period of time, and the expression is:
[0045] u 2 (t) = A{1-α[ξ(tt 1 )-ξ(tt 2 )]}sin(ω 0 t)#(1)In the formula, A(1-α) is the amplitude of the voltage sag signal, 0.1≤α≤0.9; ξ(·) is the step function; t 1 is the start time of sag; t 2 is the end time of the temporary drop; T≤t 2 -t 1 ≤9T.
[0046] The Kaiser window is defined as a set of adjustable window functions composed of zero-order Bessel functions. By adjusting the control parameters, the proportion between the main lobe width and the side lobe attenuation can be freely selected. The discrete expression of the Kaiser window with a length of N is:
[0047]
[0048] Among them, α is the coefficient parameter for adjusting the ratio between the main lobe width and the side lobe attenuation; the spectrum of the Kaiser window function is expressed as
[0049]
[0050] The S transform (KS transform) algorithm based on the Kaiser window function, for a given signal to be measured x(t)∈L 2 (R), the KS transform KS(τ,f):R→C of the signal x(t) is defined as
[0051]
[0052] Step S2: Optimizing the Kaiser window function, which is suitable for the Kaiser optimized window function that can detect power quality disturbances;
[0053] Kaiser optimizes the window function. First, the adjustment function in the KS transformation Kaiser window is:
[0054]
[0055] Among them, η 1 and η 2 is the Kaiser window adjustment parameter; parameter η 1 ≥0 is used to adjust the rate of change of the window shape as the detection frequency changes; η 2 >0 can easily adjust the time-frequency resolution at a certain frequency. Therefore, the Kaiser optimization window function expression is:
[0056]
[0057] Among them, I 0 (·) is the first kind of zero-order modified Bessel function. Its power series expansion is
[0058]
[0059] Step S3: Analyze the Kaiser optimized window function and construct a fast KS transformation algorithm based on the Kaiser optimized window function;
[0060] First, the fast KS transformation algorithm based on Kaiser optimized window function uses the adjustment function described by formula (5) and the Kaiser optimized window function described by formula (6) as the transformation kernel function, and constructs the KS transformation based on Kaiser optimized window as follows:
[0061]
[0062]
[0063] Secondly, the continuous and discrete expressions of the KS transform of the Kaiser optimization window are calculated. In the discrete process of the KS transform, the sampling rate is recorded as f s , sampling time interval The sampling length is N. The KS transform is discretized, and accordingly, the time shift factor τ = mT s , m = 0, 1, 2, ..., N-1, time variable t = nT s , n = 0, 1, 2, ..., N-1, frequency variable k=0,1,2,···,N-1. The discrete form of 1-S transformation equation (8) is
[0064]
[0065] The discrete expression implemented by the fast KS transform is:
[0066]
[0067] in, It is expressed as the discrete form of the Fourier transform of the signal frequency shift X(ξ+f); It is expressed as the window function w in formula (9) k (t, f) about the time τ to ξ spectrum W k The discrete form of (ξ, f) is
[0068]
[0069] Then, in the implementation process of this method, the fast Fourier transform and its inverse transform are used to replace the discrete Fourier transform and its inverse transform. At the same time, the non-characteristic frequency points can be eliminated and only the characteristic frequency points such as fundamental and harmonic frequencies can be calculated to greatly improve the calculation efficiency and improve the algorithm application performance. The N-point discrete signal to be measured x[n] is expressed as
[0070] x[n]=[x 0 , x 1 , x 2 ,...x N-1 , ] T (12)
[0071] The Fourier transform result of x[n] is
[0072] X[1]=[X 0 , X 1 , X 2 , ....X N-1 , ] T (13)
[0073] Among them, for the frequency point k, k=0,1,2···,M-1, The frequency shift vector is expressed as
[0074] X[1+k]=[X k , X k+1 , X k+2 , ...X N-1 , X 0 ...X k-1 ] T (14)
[0075] At the same time, the Kaiser optimized window function spectrum W used at frequency point k k [l, k] can be expressed as
[0076]
[0077] According to the Fourier transform of KS transform, the form
[0078]
[0079] Among them, W k (ξ, f) is represented by the Kaiser window function w k The frequency spectrum of (τ, f) is shown in equation (17).
[0080]
[0081] Among them, I 0 (·) is the first kind of zero-order modified Bessel function. Its power series expansion is
[0082]
[0083] Then, for frequency point k, the frequency shift vector X[l+k] is compared with the window function spectrum W k [l, k] is Hadamard-multiplied, and then the inverse Fourier transform is performed on it from frequency l to time m, and the time domain information of the KS transform of the signal at the frequency represented by the frequency point k can be obtained, which is specifically:
[0084]
[0085] The implementation process of the fast KS transformation algorithm is as follows Figure 2 As shown, the specific steps can be summarized as follows:
[0086] Step 1: Uniformly sample the measured signal x(t) to obtain a discrete signal x[n] with a sampling length of N, n = 0, 1, 2, ..., N-1, and calculate the Fourier transform spectrum of x[n]
[0087] X[l],l=0,1,2,…,N-1;
[0088] Step 2: For frequency point k, k = 0, 1, 2, ···, M-1, Select characteristic frequency points and eliminate non-characteristic frequency points to obtain a specific frequency shift vector X[l+k];
[0089] Step 3: Calculate the Kaiser window function spectrum W at the characteristic frequency point k through Fourier transform k [l, k];
[0090] Step 4: For frequency point k, calculate the frequency shift vector X[1+k] in step 2 and the window function spectrum vector W in step 3 k Hadamard product of [l, k];
[0091] Step 5: Perform an inverse Fourier transform from frequency l to time m on the product vector obtained in step 4 to obtain KS transform time domain information of the signal at frequency point k;
[0092] Step 6: Repeat steps 2 to 5 to obtain the KS transform time domain information corresponding to all detection frequency points of the signal, and construct the signal KS transform complex time-frequency matrix.
[0093] Step 7: Extract the amplitude and phase angle through KS transformation. The time domain amplitude information and phase information of the signal at each characteristic frequency point are extracted by the formulas described by equations (19) and (20) respectively.
[0094]
[0095] and
[0096]
[0097] Finally, the parameter selection method of the fast KS transform algorithm based on the constructed Kaiser optimization window function is analyzed. In the process of implementing the KS transform, the Kaiser window adjustment function is used as
[0098]
[0099] Among them, [NT s f]+1=1 indicates the frequency point representing the frequency f in the calculation. For example, when the detection frequency is f=0, the corresponding frequency point is [NT s f]+1=1, in particular, the detection frequencies reflected in the calculation results are all able to make NT s f is the frequency of an integer. For the time-frequency analysis of power quality disturbance signals, the adjustment function parameter determination of the Kaiser optimization window in the KS transformation is analyzed. The adjustment parameter η in formula (21) 1 and η 2 Determine so that equation (21) has
[0100] α[f 0 ]∈[90,110] (22)
[0101] Established, where f 0 In the specific implementation process, formula (22) can ensure that the KS transform has a better time resolution for the signal fundamental wave so as to accurately detect the fundamental wave amplitude. In addition, by adjusting the parameter η 1 and η 2 Adjustment can further adjust the rate at which the KS transform time-frequency resolution changes adaptively as the detection frequency changes.
[0102] Step S4: Detecting the voltage sag disturbance signal based on the fast KS transformation algorithm of the constructed Kaiser optimization window function;
[0103] The KS transformation algorithm based on the optimized Kaiser window is tested on the voltage sag disturbance signal, and the sampling rate is set to f s =10240Hz, the sampling length is N = 1024. The control parameter η of KS transformation 1 and η 2 They are set to 11000 and 100 respectively. The start and end time of the voltage sag disturbance and the signal amplitude after the sag are shown in Table 1.
[0104] Table 1 Voltage sag disturbance parameters
[0105]
[0106]
[0107] The signal expressions are shown in equations (23) to (26).
[0108] u 1 (t)={1-0.8[ε(t-0.01)-ε(t-0.05)]}sin(2π×50t), t∈[0,0.1]#(23)
[0109] u 2 (t)={1-0.5[ε(t-0.03)-ε(t-0.08)]}sin(2π×50t), t∈[0,0.1]#(24)
[0110] u 3 (t)={1-0.3[ε(t-0.02)-ε(t-0.07)}sin(2π×50t), t∈[0,0.1]#(25)
[0111] u 4 (t)={1-0.4[ε(t-0.04)-ε(t-0.09)]}sin(2π×50t), t∈[0,0.1]#(26)
[0112] The S transform is developed based on the short-time Fourier transform.
[0113] Based on the time-correlated window function introduced by the transformation, the time-frequency related Gaussian
[0114] The window function makes the transform have local time-frequency characteristics and variable resolution characteristics. It is a typical linear time-frequency analysis method that can process non-stationary signals. Therefore, S transform is used.
[0115] as a control method for this method.
[0116] Figure 3The results of the voltage sag signal amplitude detection using the KS transform algorithm and the control method S transform algorithm are shown. The detection results of the two algorithms are represented by solid lines and dotted lines, respectively.
[0117] The voltage sag start and end time and sag amplitude information in Table 1, combined with Figure 3 The detection results show that the amplitude detection results of the voltage sag disturbance signal by this method are consistent with the actual signal amplitude results.
[0118] The degree of agreement is higher, while the detection results of the traditional S-transform algorithm have a large deviation from the actual amplitude.
[0119] In order to characterize the degree of deviation between the detection result and the actual signal amplitude, three error metrics, namely mean square error (MSE), mean absolute error (MAE) and root mean square error (RMSE), are used to describe the detection error. Table 2 shows the different error metrics of KS transform and the control method S transform algorithm for the four examples of sag signals in equations (23) to (26). It can be seen that the detection error of the proposed KS transform algorithm is better, indicating that the amplitude detection result of the voltage sag test signal by this method is more consistent with the actual amplitude result.
[0120] Table 2 K-S transformation and comparison algorithm S transformation voltage sag signal amplitude detection error (%)
[0121]
[0122] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0123] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A voltage sag detection method based on Kaiser window function optimized S transform, characterized in that: The following steps are involved: S1: By using the time- and frequency-dependent Kaiser window as the integral transform kernel function, an S-transform algorithm based on the Kaiser window function is constructed; S2: Optimizing the Kaiser window function to obtain a Kaiser optimized window function suitable for power quality disturbance detection; S3: constructing a fast KS transformation algorithm based on the Kaiser optimized window function using the Kaiser optimized window function; S4: The voltage sag disturbance signal is detected based on the fast KS transformation algorithm of the constructed Kaiser optimized window function.
2. The voltage sag detection method based on Kaiser window function optimized S transform according to claim 1, characterized in that: The S transform algorithm based on Kaiser window function, for a given signal to be measured x(t)∈L 2 (R), the KS transform KS(τ,f):R→C of the signal x(t) is defined as In the above formula, τ is the time shift factor that controls the position of the Kaiser window in the time domain; I0(·) is the first kind of zero-order modified Bessel function.
3. The voltage sag detection method based on Kaiser window function optimized S transform according to claim 1, characterized in that: In step S2, the Kaiser optimization window function expression is: In the above formula, η1 and η2 are Kaiser window adjustment parameters; parameter η1 ≥ 0 is used to adjust the rate of change of the window shape with the change of detection frequency; η2>0 can conveniently adjust the time-frequency resolution at a certain frequency; I0(·) is the first-kind zero-order modified Bessel function.
4. The voltage sag detection method based on Kaiser window function optimized S transform according to claim 1, characterized in that: In step S3, the Kaiser optimized window function is used to construct a fast KS transformation algorithm based on the Kaiser optimized window function, which specifically includes: (1) Construct the KS transform based on Kaiser optimization window, expressed as (2) The continuous and discrete expressions of the KS transform of the Kaiser optimization window are calculated; the discrete expression implemented by the fast KS transform is: In the above formula, It is expressed as the discrete form of the Fourier transform of the signal frequency shift X(ξ+f); It is expressed as the window function w in formula (6) k (t,f) about the spectrum W from time τ to ξ k The discrete form of (ξ,f).
5. The voltage sag detection method based on Kaiser window function optimized S transform according to claim 1, characterized in that: In step S4, the fast KS transform algorithm selection method based on the constructed Kaiser optimized window function uses fast Fourier transform and its inverse transform to replace discrete Fourier transform and its inverse transform. At the same time, non-characteristic frequency points can be eliminated, specifically including: Step 1: uniformly sample the measured signal x(t) to obtain a discrete signal x[n] with a sampling length of N, n=0, 1, 2, ···, N-1, and calculate the Fourier transform spectrum of x[n]; Step 2: For frequency point k, k=0,1,2,···,M-1, Select characteristic frequency points and eliminate non-characteristic frequency points to obtain a specific frequency shift vector X[l+k]; Step 3: Calculate the Kaiser window function spectrum W at the characteristic frequency point k through Fourier transform k [l,k]; Step 4: For frequency point k, calculate the frequency shift vector X[l+k] in step 2 and the window function spectrum vector W in step 3 k Hadamard product of [l,k]; Step 5: Perform an inverse Fourier transform from frequency l to time m on the product vector obtained in step 4 to obtain KS transform time domain information of the signal at frequency point k; Step 6: Repeat steps 2 to 5 to obtain the KS transform time domain information corresponding to all detection frequency points of the signal, and construct the signal KS transform complex time-frequency matrix; Step 7: Extract the amplitude and phase angle through KS transformation by using the formulas described in equations (19) and (20) to extract the time domain amplitude information and phase information of the signal at each characteristic frequency point; The parameter selection method of the fast KS transform algorithm based on the constructed Kaiser optimization window function is analyzed. In the process of implementing the KS transform, the Kaiser window adjustment function is used as Among them, [NT s f]+1 represents the frequency point representing frequency f in the calculation. For example, when the detection frequency is f=0, the corresponding frequency point is [NT s f]+1=1, the detection frequencies reflected in the calculation results are all able to make NT s f is the frequency of an integer; the adjustment parameters η1 and η2 in the Kaiser window adjustment function are determined so that the Kaiser window adjustment function has α[f0]∈[90,110] at the fundamental frequency, where f0 represents the fundamental frequency, ensuring that the KS transform has a better time resolution for the signal fundamental to accurately detect the fundamental amplitude.
