A power system frequency measurement data preprocessing method and system

Abstract

The present application relates to the technical field of data preprocessing, and provides a power system frequency measurement data preprocessing method and system, which comprises: obtaining measurement data and dividing into several intervals; for each interval, detecting pulse noise points, determining the size of a first window according to the ratio of the number of pulse noise points to the total number of all data points in the interval, and performing median filtering on the detected pulse noise points in the interval by using the first window according to the size of the first window; selecting a second window, and suppressing Gaussian noise by weighted mean filtering according to the similarity between any data point in the second window and the center point of the window. The three factors of noise reduction effect, feature protection and calculation speed are considered.

Description

Technical Field

[0001] This invention belongs to the field of data preprocessing technology, and in particular relates to a method and system for preprocessing frequency measurement data in power systems. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] The integration of high-proportion renewable energy sources and high-proportion power electronic equipment has significantly changed the structure and operation of the power system, with one of the most prominent features being a reduction in system inertia. Due to the inherent uncertainties and randomness of renewable energy generation, and the increasing application of various virtual inertia control strategies to provide active power support to the grid under frequency deviations, the time-varying characteristics of the system's equivalent inertia have become increasingly complex. Therefore, it is necessary to quantitatively assess the power system inertia based on frequency measurement data, monitor the system inertia level in real time, and provide a basis for the safe and stable operation of the power system.

[0004] With the large-scale integration of new energy generating units and power electronic equipment into the power grid, a significant amount of polymorphic noise is introduced into the power system phasor measurement system. This noise is then incorporated into the frequency measurement results through processes such as frequency domain convolution or least squares estimation in frequency calculation. Furthermore, during communication transmission in wide area measurement systems (WAMS), new impulse interference is introduced, such as when the communication carrier is transmitted via power lines. When a large amount of frequency measurement data from different sources is collected in the inertia assessment system, it contains both small, random fluctuations and larger deviations or outliers. The former can be considered background noise (described by Gaussian noise), while the latter is typically considered random impulse noise. Before applying frequency measurement data, it is necessary to first consider how to suppress the mixed noise in the original data. Especially for online monitoring of equivalent inertia, quasi-steady-state data under normal power system operating conditions is required. The frequency / active power data exhibits small fluctuations and weak fluctuation characteristics similar to Gaussian noise. This makes the inertia estimation results more susceptible to mixed noise interference, leading to inertia information distortion in the absence of suitable data denoising algorithms.

[0005] Therefore, considering the accuracy requirements of online inertia estimation in power systems, the requirements for frequency data preprocessing methods can be summarized as follows: 1) The online inertia estimation system needs to meet real-time requirements, which places high demands on the computational speed of the algorithm. Many traditional filters, such as Kalman filtering, Wiener filtering, and wavelet filtering, have a large computational load and cannot meet the requirements of online real-time calculation; 2) The raw data may contain mixed noise composed of impulse noise and Gaussian noise, and the data preprocessing algorithm used should have the ability to suppress both types of noise; 3) While filtering out noise, the detailed features of the data should be fully preserved. If the frequency fluctuation features are severely lost, the effectiveness of the subsequent inertia estimation algorithm will be greatly reduced; 4) Frequency is an inertial quantity and generally does not undergo abrupt changes. Even in system transient scenarios, the frequency curve is relatively smooth. Low-pass linear filtering methods such as mean filtering and local regression smoothing are suitable for filtering Gaussian noise, but lack robustness to random impulse noise. Time-domain geometric filters, represented by median filtering, are highly sensitive to outlier data and have a strong suppression effect on impulse noise that deviates from the normal data sequence, but they are prone to causing loss of data waveform details. When impulse noise and Gaussian noise coexist in the measurement data, it is difficult to achieve satisfactory noise reduction results by using median or mean filtering alone. Summary of the Invention

[0006] To address the technical problems mentioned above, this invention provides a power system frequency measurement data preprocessing method and system. By detecting noise, the method separates impulse noise from Gaussian noise in the original data. The separated impulse noise and Gaussian noise are suppressed by improved adaptive median filtering and adaptive weighted mean filtering, respectively, to achieve a balance between noise reduction effect, feature protection, and calculation speed.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] The first aspect of the present invention provides a method for preprocessing power system frequency measurement data, comprising:

[0009] Acquire measurement data and divide it into several intervals;

[0010] For each interval, impulse noise points are detected. The size of the first window is determined based on the ratio of the number of impulse noise points to the total number of data points in the interval. Based on the size of the first window, the median filtering of the detected impulse noise points in the interval is performed using the first window.

[0011] Select the second window, and suppress Gaussian noise by using weighted mean filtering based on the similarity between any data point in the second window and the center point of the window.

[0012] Furthermore, for each interval, the median absolute deviation algorithm is used to detect impulse noise points.

[0013] Furthermore, if the center data point of the first window is an impulse noise point, then the median of all impulse noise points within the first window is used to replace the value of the center data point.

[0014] Furthermore, if the data point at the center of the first window is not an impulse noise point, then the value of the data point at the center of the window remains unchanged.

[0015] Furthermore, the similarity is represented by a weighted coefficient, which is determined by the difference between any data point within the second window and the center point of the window.

[0016] Furthermore, the method for suppressing Gaussian noise is as follows:

[0017] Calculate the average of the weighting coefficients for all data points within the second window;

[0018] Collect the data points in the second window whose weighting coefficients are greater than the average value, and calculate the weighted average of the data points in the set.

[0019] The value of the center data point in the second filter window is replaced by a weighted average.

[0020] Furthermore, the weighting coefficients are:

[0021]

[0022] Where i represents the center data point in the second filtering window, s represents the s-th data point in the second filtering window, h(i) represents the value of the i-th data point, and Z is the scale parameter.

[0023] A second aspect of the present invention provides a power system frequency measurement data preprocessing system, comprising:

[0024] The data acquisition module is configured to acquire measurement data and divide it into several intervals;

[0025] The median filtering module is configured to: detect impulse noise points in each interval, determine the size of the first window based on the ratio of the number of impulse noise points to the total number of data points in the interval, and perform median filtering on the detected impulse noise points in the interval based on the size of the first window.

[0026] The Gaussian filtering module is configured to: select a second window, and suppress Gaussian noise by weighted mean filtering based on the similarity between any data point in the second window and the center point of the window.

[0027] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the power system frequency measurement data preprocessing method described above.

[0028] A fourth aspect of the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the power system frequency measurement data preprocessing method described above.

[0029] Compared with the prior art, the beneficial effects of the present invention are:

[0030] This invention provides a method for preprocessing frequency measurement data in power systems. It separates impulse noise and Gaussian noise in the original data through noise detection. The separated impulse noise and Gaussian noise are suppressed by improved adaptive median filtering and adaptive weighted mean filtering, respectively, so as to achieve a balance between noise reduction effect, feature protection and calculation speed.

[0031] This invention provides a method for preprocessing power system frequency measurement data, which focuses on sub-intervals in the data. It only needs to search the MAD of all data values ​​in a certain interval to classify all data points in that interval, thus greatly reducing the detection time. At the same time, this method does not require a pre-set threshold value and has stronger adaptability.

[0032] This invention provides a preprocessing method for frequency measurement data in power systems. The improved adaptive median filtering method only filters the detected noise points, which can effectively protect uncontaminated signal points. The filtering window length can be adaptively selected according to the degree of noise contamination in a certain interval of the data, which can balance noise suppression and protection of data details.

[0033] This invention provides a preprocessing method for power system frequency measurement data. The adaptive weighted mean filtering method employed achieves good denoising effects for both smooth and volatile sections of the curve. Because the numerical differences between points in smooth sections are small and their similarity is relatively close, a good smoothing effect can be obtained. For volatile sections, the numerical values ​​near the edge of the section have higher similarity and participate in the weighted mean filtering, while data points far from the edge are relatively excluded due to lower similarity. Therefore, this method effectively reduces the feature loss phenomenon that often occurs in standard mean filtering. Attached Figure Description

[0034] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0035] Figure 1 This is a schematic diagram of the overall scheme of a power system frequency measurement data preprocessing method according to Embodiment 1 of the present invention;

[0036] Figure 2 This is a flowchart of a power system frequency measurement data preprocessing method according to Embodiment 1 of the present invention;

[0037] Figure 3 This is a schematic diagram of the improved WSCC 9-node system structure according to Embodiment 1 of the present invention;

[0038] Figure 4 This is a frequency offset diagram of Embodiment 1 of the present invention;

[0039] Figure 5 This is a diagram showing the equivalent inertia estimation results of Embodiment 1 of the present invention. Detailed Implementation

[0040] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0041] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0042] Example 1

[0043] This embodiment provides a method for preprocessing frequency measurement data in a power system.

[0044] This embodiment provides a power system frequency measurement data preprocessing method. Considering the accuracy requirements of power system inertia assessment, noise detection is first performed to separate impulse noise and Gaussian noise in the original data. The separated impulse noise and Gaussian noise are suppressed by improved adaptive median filtering and adaptive weighted mean filtering, respectively, so as to achieve a balance between noise reduction effect, feature protection and calculation speed.

[0045] This embodiment provides a method for preprocessing power system frequency measurement data, such as... Figure 1 and Figure 2 As shown, it includes the following steps:

[0046] (1) Read in the raw data (i.e., frequency measurement data) online. Using the idea of ​​local detection, divide the data of length M into S intervals. The k-th (k = 0, 1, ..., S-1) interval is denoted as U. k ;

[0047] (2) The median absolute deviation (MAD) algorithm is used to segment and detect the data sequence, and impulse noise is marked. Let g i =1 to mark impulse noise points, otherwise set g i =0;

[0048] (3) The size of the first filtering window is adaptively determined based on the degree of contamination by impulse noise; specifically, the number of noise points and the interval U are used. k The ratio r of the total number of all data points k To characterize the degree of noise pollution in this interval, and based on r k The size of the median filter window is adaptively determined.

[0049] (4) Determine the first window length l k Then, this first window is used for interval U. k The detected impulse noise points are processed, g i Points with a value of 0 retain their original values; when processing impulse noise points, the value within the first window (g) is considered. i All data points with a value of 1 are grouped into a set B. i , use B i The median value of the elements replaces the value of the center data point of the window;

[0050] (5) Next, Gaussian noise is processed by selecting a second window of size d (where d is an odd number) for filtering. The set of data within the window is C. k ;

[0051] (6) Find C k The weighting coefficient c corresponding to any data point within the range i,s Then, the average value T of the weighted coefficients corresponding to all data points within the window is calculated. i ;

[0052] (7) The weighting coefficient ratio T in the second window i Large datasets constitute a new set P i Find P i The weighted average value m(i) of each data point in the second filtering window is used to replace the value h(i) of the center data point in the second filtering window, thereby achieving the suppression of Gaussian noise in the data.

[0053] In step (1), in interval U k Let f(i) be the value of any data point i within the interval. The set consisting of all data values ​​in this interval is:

[0054] D k ={f(i)|i∈U k ,k=0,1,,S-1} (1)

[0055] In step (2), MAD is defined as set D. k The median of the absolute deviations of the data points from the median:

[0056] MAD(D k )=amed(|f(i)-med(D k )|),f(i)∈D k (2)

[0057] In the formula, med(D) k ) represents D k The median is the median value, where 'a' is the proportionality coefficient, which depends on the sample distribution. If the values ​​of f(i) follow a normal distribution, then a = 1.4826. MAD represents the median of the differences between the data and the median.

[0058] From the following formula for D k The data points are used for discrimination, let g i =1 to mark impulse noise points, otherwise set g i =0:

[0059]

[0060] When D k If the value of a data point satisfies the greater than inequality in equation (3), it is determined to be an impulse noise point. b is the cutoff factor, which is selected according to the value of a and the probability distribution type of the data set. It is 2 within the 95% confidence interval of the normal distribution and 3 within the 99% confidence interval of the normal distribution.

[0061] Step (3) for interval U k In the middle, satisfy g i The number of data points with a value of 1 is counted, and the number of noise points is compared with the interval U. k The ratio r of the total number of all data points k To characterize the degree of noise pollution in this interval, and based on r k The size of the first filtering window is adaptively determined. Window length l k for:

[0062]

[0063] In the formula, p1 and p2 are constants greater than 0 and less than 1. Their values ​​have a significant impact on the filtering results, and existing research often sets them arbitrarily. This invention determines the empirical values ​​of p1 and p2 through experiments: Gaussian noise with a fixed standard deviation and impulse noise with different noise densities are added to the simulation data. p1 and p2 are set to values ​​between 0 and 1 (granularity of 0.01). Under each value, the normalized mean squared error (NMSE) and peak signal-to-noise ratio (PSNR) are compared with the ideal data. The values ​​of p1 and p2 corresponding to the minimum NMSE and maximum PSNR are taken as the optimal values. Experiments were conducted on different curves using this method, and it was found that when p1 and p2 are set to 0.1 and 0.3 respectively, good noise reduction effects can be obtained for different curves.

[0064] In step (4), the window length l is determined. k Then, this first window is used for interval U. k The detected impulse noise points are processed, g i Points with a value of 0 retain their original values. When processing impulse noise points, the value within the window (g) needs to be considered. i All data points with a value of 1 are grouped into a set B. i , use B i The median value of the elements replaces the value of the center data point of the window. In the interval U... k After adaptive median filtering, any data point with median value f(i) has the following value:

[0065]

[0066] In step (5), a second window of size d (where d is an odd number) is selected for filtering. If the filtering window is centered at i, the set of data within the window is C. k :

[0067]

[0068] In C k The weighting coefficient c corresponding to any data point in the dataset i,s In other words, it is determined by the difference between its value h(i+s) and the value h(i) of the center point:

[0069]

[0070] In the formula, Z is the scaling parameter, and its value has a certain impact on the suppression effect of Gaussian noise. Z is used to control the smoothness of the data curve. When Z is large enough, the weighting coefficient c i,sWhen the value of Z is approximately 1, the algorithm uses standard mean filtering, and some points on the curve may lose their original features due to over-smoothing. When the Z value is too small, the smoothing effect is too weak, and a good denoising effect cannot be achieved. To select a suitable Z value, this invention conducted similar experiments on different simulation data. The tests showed that when Z is between 120 and 150, a good smoothing effect can be obtained for different curves.

[0071] As can be seen from equation (7), the larger the weighting coefficient corresponding to a data point within the second filter window, the greater the similarity between the value of that data point and the value of the center point within the window. Therefore, the reconstructed curve after weighting and averaging these data values ​​retains better features. The average of the weighting coefficients corresponding to all data points within the second filter window is calculated as follows:

[0072]

[0073] In step (6), the weighting coefficients are compared with T. i Large datasets form new sets:

[0074]

[0075] In step (7), find P. i The weighted average of the data points:

[0076]

[0077] By replacing the value h(i) of the center data point in the second filtering window with m(i), Gaussian noise in the data can be suppressed.

[0078] This embodiment uses Figure 3 The improved WSCC 9-node system shown is used as an example to verify the effectiveness of the method proposed in this invention. The key parameter settings for the wind farm connected to the grid at node 5 are shown in Table 1.

[0079] Table 1. Wind Farm System Parameters

[0080] Parameter name Parameter value Parameter name Parameter value <![CDATA[Rated voltage V n > 575V <![CDATA[Excitation inductance L m > 2.9 PU <![CDATA[Rated power P n > 1.5MW <![CDATA[Inherent inertia H DFIG > 5.04s <![CDATA[Stator resistance R s > 0.023 pu <![CDATA[Initial blade pitch angle β0]]> 0 <![CDATA[Stator inductance L s > 0.18pu Number of wind turbine units (m) 6 <![CDATA[Rotor resistance R r > 0.016pu <![CDATA[Wind farm capacity S m > 9MW

[0081] To verify the effectiveness of this invention in improving the accuracy of inertia estimation results, random load fluctuations and large disturbances of different amplitudes were combined, and Gaussian noise and impulse noise were added to the simulation data to simulate data collected during actual operation. The power grid COI frequency fluctuation curve and the data curve after adding the mixed noise set by the section are shown below. Figure 4 As shown.

[0082] To demonstrate the effectiveness and superiority of the data preprocessing method of this invention, both the invented method and the control method were used to preprocess simulation data with added noise before inertia estimation. The control method used median filtering preprocessing. The simulation length was 80 seconds, and the inertia estimation window length was set to 5 seconds. The inertia estimation results after data preprocessing using the invented method and the control method are as follows: Figure 5 As shown in Table 2. Table 2 shows the model identification accuracy during the estimation process. (Summary) Figure 5 As can be seen from Table 2, the accuracy of inertia estimation after preprocessing using the method of the present invention is significantly improved.

[0083] Table 2. Model identification accuracy during the estimation process

[0084]

[0085] Most existing methods focus on individual data points in a data sequence. Classifying each data point requires taking the median or extreme value of all values ​​within its neighborhood, which is computationally time-consuming. This embodiment, however, focuses on sub-intervals within the data. It only needs to search the MAD (mean squared deviation) of all data values ​​within a given interval to classify all data points within that interval. This significantly reduces detection time. Furthermore, this method does not require a pre-defined threshold value, making it more adaptable.

[0086] The improved adaptive median filtering method of this invention only filters the detected noise points, which can effectively protect the uncontaminated signal points; the length of the filtering window can be adaptively selected according to the degree of noise contamination in a certain interval of the data, which can balance noise suppression and protection of data details.

[0087] The adaptive weighted mean filtering method used in this invention achieves good denoising results in both smooth and volatile sections of the curve. This is because the numerical differences between points in the smooth section are small, c... i,s A closer match yields better smoothing results; for segments with drastic fluctuations, the c value corresponding to the edge point of that segment should be closer. i,s Larger data points participate in weighted mean filtering, while data points far from the edge values ​​are affected by their corresponding c values. i,s The smaller the feature size, the more it is excluded, thus effectively reducing the feature loss phenomenon that often occurs in standard mean filtering.

[0088] Example 2

[0089] This embodiment provides a power system frequency measurement data preprocessing system, which specifically includes:

[0090] The data acquisition module is configured to acquire measurement data and divide it into several intervals;

[0091] The median filtering module is configured to: detect impulse noise points in each interval, determine the size of the first window based on the ratio of the number of impulse noise points to the total number of data points in the interval, and perform median filtering on the detected impulse noise points in the interval based on the size of the first window.

[0092] The Gaussian filtering module is configured to: select a second window, and suppress Gaussian noise by weighted mean filtering based on the similarity between any data point in the second window and the center point of the window.

[0093] It should be noted that each module in this embodiment corresponds one-to-one with each step in Embodiment 1, and their specific implementation processes are the same, so they will not be repeated here.

[0094] Example 3

[0095] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the power system frequency measurement data preprocessing method described in Embodiment 1 above.

[0096] Example 4

[0097] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the power system frequency measurement data preprocessing method described in Embodiment 1 above.

[0098] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0099] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0100] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0101] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0102] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0103] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for power system frequency measurement data preprocessing, characterized in that, The method comprises the following steps: acquiring measurement data and dividing the data into several intervals; detecting impulse noise points in each interval, determining the size of a first window according to the ratio of the number of impulse noise points to the total number of data points in the interval, and performing median filtering on the detected impulse noise points in the interval using the first window according to the size of the first window; selecting a second window, and suppressing Gaussian noise by weighted mean filtering according to the similarity between any data point in the second window and the center point of the window; processing Gaussian noise, selecting a second window with a size of d for filtering, and taking the set of data in the window as Ck; calculating the weighted coefficient ci,s corresponding to any data point in Ck, and then calculating the average value Ti of the weighted coefficients corresponding to all data points in the window; constructing a new set Pi by taking data points in the second window with a weighted coefficient greater than Ti, calculating the weighted average value m(i) of the data points in Pi, and replacing the value h(i) of the center data point in the second filtering window with m(i) to achieve the suppression of Gaussian noise in the data; for the weighted coefficient ci,s corresponding to any data point in Ck, the difference between the value h(i+s) and the value h(i) of the center point is determined: where i denotes the center data point within the second filter window, s denotes the s-th data point within the second filter window, h(i) denotes the numerical value of the i-th data point, and Z is a scale parameter.

2. The power system frequency measurement data preprocessing method of claim 1, wherein, for each interval, the absolute deviation median algorithm is used to detect impulse noise points.

3. The power system frequency measurement data preprocessing method of claim 1, wherein If the center data point of the first window is an impulse noise point, the median value of all impulse noise points in the first window is used to replace the value of the center data point.

4. The power system frequency measurement data preprocessing method of claim 1, wherein If the center data point of the first window is not an impulse noise point, the value of the center data point remains unchanged.

5. The power system frequency measurement data preprocessing method of claim 1, wherein, The similarity is represented by a weighted coefficient, which is determined by the difference between any data point in the second window and the center point of the window.

6. A method for power system frequency measurement data preprocessing as recited in claim 5, characterized by, The method for suppressing Gaussian noise comprises the following steps: calculating the average value of the weighted coefficients of all data points in the second window; constructing a set by taking data points in the second window with a weighted coefficient greater than the average value, calculating the weighted average value of the data points in the set, and replacing the value of the center data point in the second filtering window with the weighted average value. The method comprises the following steps:

7. A power system frequency measurement data preprocessing system characterized by, a data acquisition module configured to acquire measurement data and divide the data into several intervals; a median filtering module configured to detect impulse noise points in each interval, determine the size of a first window according to the ratio of the number of impulse noise points to the total number of data points in the interval, and perform median filtering on the detected impulse noise points in the interval using the first window according to the size of the first window; a Gaussian filtering module configured to select a second window and suppress Gaussian noise by weighted mean filtering according to the similarity between any data point in the second window and the center point of the window; processing Gaussian noise, selecting a second window with a size of d for filtering, and taking the set of data in the window as Ck; calculating the weighted coefficient ci,s corresponding to any data point in Ck, and then calculating the average value Ti of the weighted coefficients corresponding to all data points in the window; constructing a new set Pi by taking data points in the second window with a weighted coefficient greater than Ti, calculating the weighted average value m(i) of the data points in Pi, and replacing the value h(i) of the center data point in the second filtering window with m(i) to achieve the suppression of Gaussian noise in the data; ​ In the Ck, any data point corresponding to the weighting coefficient ci,s, by its value h (i + s) and the center point value h (i) difference to determine: where i denotes the center data point within the second filter window, s denotes the s-th data point within the second filter window, h(i) denotes the numerical value of the i-th data point, and Z is a scale parameter.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the steps of the power system frequency measurement data preprocessing method in any one of claims 1-6.

9. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the steps of the power system frequency measurement data preprocessing method in any one of claims 1-6. The processor executes the program to implement the steps of the power system frequency measurement data preprocessing method in any one of claims 1-6.

Images