A digital signal processing-based electric energy metering data analysis method and system

Abstract

The present application relates to the technical field of data processing, and especially relates to a power metering data analysis method and system based on digital signal processing, which obtains power metering data signals in the operation process of a power system, performs frequency domain conversion on the power metering data signals to obtain corresponding spectrum images; determines the signal type of the power metering data signals according to the information entropy of the spectrum images, obtains the spectral transient characteristic coefficient and the spectral dynamic range factor of the power metering data signals if the signal type is a dense spectrum signal, so as to obtain the best window function for harmonic identification of the power metering data signals by using a window interpolation FFT algorithm, and identifies the harmonic components in the power metering data signals according to the best window function, which is used for power quality analysis of the power system; and the best window function of the window interpolation FFT algorithm is adaptively selected, which can eliminate the spectrum leakage and fence effect to the greatest extent and improve the identification and processing efficiency of the harmonics.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for analyzing electricity metering data based on digital signal processing. Background Technology

[0002] Electricity metering data refers to the data accurately measured and recorded using specific instruments and methods. It is the foundation for the operation, settlement, and management of the power system. Digital signal processing and analysis of electricity metering data is a complex, multi-layered process. Its core lies in extracting valuable electrical parameter information from raw voltage and current sampling data to achieve effective power grid analysis. Among these, power quality analysis is an important and critical area of ​​analysis, and harmonic processing is crucial in this area. Harmonics are sinusoidal wave components in the power system whose repetition frequency is an integer multiple of the fundamental frequency. The presence of harmonics significantly reduces power quality and causes various harms to the power grid, electrical equipment, and the overall operation of the power system.

[0003] While the traditional Fast Fourier Transform (FFT) can accurately analyze harmonics under ideal synchronous sampling, it is difficult to achieve strictly synchronous sampling in practice, leading to spectral leakage and the picket fence effect. Therefore, to solve this problem, existing technologies use windowed interpolation FFT algorithms. By weighting the sampled signal with a specific window function, the spectral leakage caused by truncation is reduced. At the same time, interpolation is performed on the windowed FFT results to correct the parameter estimation deviation caused by the picket fence effect. This effectively processes and eliminates harmonic interference in power quality analysis, enabling effective analysis of power quality direction based on power metering data.

[0004] However, due to the inherent limitations of window functions—any window function is a trade-off between the main lobe width and the side lobe suppression level—even with the use of windowed interpolation FFT algorithms, spectral leakage and picket fence effects cannot be completely eliminated. Therefore, it is necessary to weigh and select the appropriate window function based on the specific task and characteristics in order to balance the accuracy and efficiency of harmonic processing and provide the best reference and standard for the final power quality analysis.

[0005] Therefore, choosing the right window function to ensure the elimination of spectral leakage and picket fence effects when using the windowed interpolation FFT algorithm for power quality analysis has become an urgent problem to be solved. Summary of the Invention

[0006] In view of this, embodiments of the present invention provide a method and system for analyzing power metering data based on digital signal processing, in order to solve the problem of how to select a window function so that when using the windowed interpolation FFT algorithm for power quality analysis, the problems of eliminating spectral leakage and picket fence effect can be ensured.

[0007] In a first aspect, embodiments of the present invention provide a method for analyzing electricity metering data based on digital signal processing, the method comprising the following steps:

[0008] Acquire power metering data signals during the operation of the power system, and perform frequency domain conversion on the power metering data signals to obtain the corresponding spectrum image;

[0009] Based on the information entropy of the spectrum image, the spectral density characteristic coefficient of the energy metering data signal is obtained. The larger the information entropy, the larger the spectral density characteristic coefficient. Based on the spectral density characteristic coefficient, the signal type of the energy metering data signal is determined. If the signal type is a dense spectrum signal, the spectral transient characteristic coefficient of the energy metering data signal is obtained based on the amplitude distribution of each frequency in the spectrum image. Based on the frequency difference between the maximum and minimum amplitudes in the spectrum image, the spectral dynamic range factor of the energy metering data signal is obtained.

[0010] Based on the spectral transient characteristic coefficients and spectral dynamic range factor of the power metering data signal, the optimal window function is obtained when using the windowed interpolation FFT algorithm to identify harmonics in the power metering data signal. Based on the optimal window function, the harmonic components in the power metering data signal are identified for power quality analysis of the power system.

[0011] Preferably, determining the signal type of the energy metering data signal based on the spectral density characteristic coefficients includes:

[0012] A preset spectral density feature coefficient threshold is obtained. If the spectral density feature coefficient is greater than or equal to the spectral density feature coefficient threshold, the signal type of the power metering data signal is determined to be a dense spectral signal; if the spectral density feature coefficient is less than the spectral density feature coefficient threshold, the signal type of the power metering data signal is determined to be a sparse spectral signal.

[0013] Preferably, after determining the signal type of the energy metering data signal based on the spectral density characteristic coefficients, the process includes:

[0014] If the power metering data signal is a sparse spectrum signal, then the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to the Blackman window, and the harmonic components in the power metering data signal are identified using the windowed interpolation FFT algorithm for power quality analysis of the power system.

[0015] Preferably, obtaining the spectral transient characteristic coefficients of the power metering data signal based on the amplitude distribution of each frequency in the spectral image includes:

[0016] Based on the amplitude of each frequency in the spectrum image, the kurtosis of the spectrum image is calculated, and the difference between the kurtosis and a preset value is used as the spectral transient characteristic coefficient of the power metering data signal.

[0017] Preferably, obtaining the spectral dynamic range factor of the energy metering data signal based on the frequency difference between the maximum and minimum amplitude values ​​in the spectral image includes:

[0018] The maximum and minimum amplitude values ​​in the spectrum image are obtained, and the ratio between the absolute value of the maximum amplitude value and the absolute value of the minimum amplitude value is obtained. The ratio is converted into a decibel value using a preset logarithmic function. The opposite of the decibel value is used as the independent variable of an exponential function with the natural constant as the base, and the spectrum dynamic range factor of the power metering data signal is obtained.

[0019] Preferably, the step of obtaining the optimal window function for harmonic identification of the energy metering data signal using a windowed interpolation FFT algorithm based on the spectral transient characteristic coefficients and spectral dynamic range factor of the energy metering data signal includes:

[0020] If the transient characteristic coefficient of the spectrum is less than a constant 0 and the dynamic range factor of the spectrum is less than the preset dynamic range factor threshold, then the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to the Blackman window.

[0021] If the transient characteristic coefficient of the spectrum is greater than the constant 0, and the dynamic range factor of the spectrum is greater than the preset dynamic range factor threshold, then the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to a rectangular window.

[0022] If the transient characteristic coefficient of the spectrum is greater than or equal to a constant 0, and the dynamic range factor of the spectrum is less than or equal to a preset dynamic range factor threshold, then the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to the Hanning window or the Hamming window.

[0023] If the transient characteristic coefficient of the spectrum is less than or equal to a constant 0, and the dynamic range factor of the spectrum is greater than or equal to a preset dynamic range factor threshold, then the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to the Hanning window or the Hamming window.

[0024] In a second aspect, embodiments of the present invention provide an energy metering data analysis system based on digital signal processing, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements an energy metering data analysis method based on digital signal processing as described in the first aspect.

[0025] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0026] This invention acquires power metering data signals during the operation of a power system, performs frequency domain conversion on the power metering data signals to obtain a corresponding spectrum image; based on the information entropy of the spectrum image, it obtains the spectral density characteristic coefficient of the power metering data signal. The larger the information entropy, the larger the spectral density characteristic coefficient. Based on the spectral density characteristic coefficient, it determines the signal type of the power metering data signal. If the signal type is a dense spectrum signal, it obtains the spectral transient characteristic coefficient of the power metering data signal based on the amplitude distribution of each frequency in the spectrum image; based on the frequency difference between the maximum and minimum amplitude values ​​in the spectrum image, it obtains the spectral dynamic range factor of the power metering data signal; based on the spectral transient characteristic coefficient and the spectral dynamic range factor of the power metering data signal, it obtains the optimal window function for harmonic identification of the power metering data signal using a windowed interpolation FFT algorithm; based on the optimal window function, it identifies the harmonic components in the power metering data signal for power quality analysis of the power system. Among them, spectrum analysis is performed on the power metering data signal. Based on the spectrum characteristics (dense and sparse conditions, spectrum signal amplitude fluctuations, etc.), the window function in the existing windowed interpolation FFT algorithm is improved to adaptively select the optimal window function. This can maximize the algorithm performance, eliminate spectrum leakage and picket fence effect, and improve the efficiency of harmonic identification and processing. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart of a method for analyzing electricity metering data based on digital signal processing, provided in Embodiment 1 of the present invention. Detailed Implementation

[0029] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0030] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0031] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0032] See Figure 1 This is a flowchart of a method for analyzing electricity metering data based on digital signal processing, provided in Embodiment 1 of the present invention. Figure 1 As shown, the method may include:

[0033] Step S101: Obtain the power metering data signal during the operation of the power system, perform frequency domain conversion on the power metering data signal, and obtain the corresponding spectrum image.

[0034] According to the Nyquist sampling theorem, the sampling frequency is set... At least higher than the highest frequency of the signal being analyzed To prevent aliasing distortion, it is often twice the value of [the standard value], but in practical applications, it is usually [the standard value]. To allow for a safety margin, the sampling frequency can be set according to the required accuracy for the implementation scenario; there are no restrictions here. The signal to be analyzed refers to the electricity metering data signal during power operation, including but not limited to indicators such as voltage and current.

[0035] Therefore, in this embodiment of the invention, based on a set sampling frequency, wired (e.g., RS-485) or wireless (e.g., public wireless network, low-power wireless) methods are used to collect power metering data during the operation of the power system. This allows for the acquisition of power metering data signals over a period of time (e.g., one hour, one day), ensuring the stability and real-time performance of data transmission. It should be noted that before the power metering data signal is digitized (ADC), a hardware low-pass filter is used to filter out signals higher than [the specified value]. Irrelevant high-frequency components are filtered to prevent aliasing.

[0036] Raw electricity metering data signals often contain noise, anomalies, or missing values. Data cleaning is crucial for improving the quality of these signals. Data cleaning includes: handling missing values ​​(deletion or imputation using methods such as mean, median, top-to-bottom values, or interpolation); handling outliers (identifying and processing outliers caused by equipment malfunction or operational errors), commonly using statistical methods (e.g., standard deviation, quartiles) and machine learning methods (isolation forest, local outlier); and handling duplicate data (checking for, deleting, or merging duplicate records). It's worth noting that data cleaning is an existing technology and will not be discussed in detail here.

[0037] After obtaining the cleaned electricity metering data signal, the electricity metering data signal is frequency domain converted to obtain the corresponding spectrum image. The horizontal axis of the spectrum image represents frequency, and the vertical axis represents amplitude. Frequency domain conversion is an existing technology and will not be described in detail here.

[0038] Thus, the frequency domain information of the electricity metering data signal is obtained, which is the spectrum image.

[0039] Step S102: Based on the information entropy of the spectrum image, obtain the spectral density characteristic coefficient of the power metering data signal. The larger the information entropy, the larger the spectral density characteristic coefficient. Determine the signal type of the power metering data signal based on the spectral density characteristic coefficient. If the signal type is a dense spectrum signal, obtain the spectral transient characteristic coefficient of the power metering data signal based on the amplitude distribution of each frequency in the spectrum image. Obtain the spectral dynamic range factor of the power metering data signal based on the frequency difference between the maximum and minimum amplitudes in the spectrum image.

[0040] The main mathematical tool in digital signal processing is the Fourier transform. The Fourier transform studies the relationship between the entire time domain and frequency domain. However, when using computers to implement engineering test signal processing, it's impossible to measure and process infinitely long signals; instead, a finite time segment is used for analysis. The approach is to extract a time segment from the signal, then perform periodic extension processing on the extracted signal time segment to obtain a virtual infinitely long signal. Then, mathematical processing such as Fourier transform and correlation analysis can be performed on the signal. After an infinitely long signal is truncated, its spectrum is distorted; the energy originally concentrated at f(0) is dispersed into two wider frequency bands (this phenomenon is called spectral energy leakage). To reduce spectral energy leakage, different truncation functions can be used to truncate the signal. The truncation function is called a window function, or simply a window. A window function is a signal with a finite width in the time domain. Leakage is related to the side lobes of the window function's spectrum. If the height of the side lobes approaches zero, and the energy is relatively concentrated in the main lobe, it can more closely approximate the true spectrum. Therefore, different window functions can be used to truncate the signal in the time domain.

[0041] Based on the underlying operations and logic of window functions, it is known that the narrower the main lobe width, the higher the frequency resolution, which allows for the differentiation of closely spaced frequency components. However, a narrower main lobe width also sacrifices the performance of side lobes, meaning that high side lobe width leads to severe spectral leakage. Therefore, the selection of a window function essentially involves a trade-off between frequency resolution (main lobe width) and spectral leakage (side lobe suppression). Different signal characteristics require different trade-off decisions. That is, if extremely high frequency resolution is desired, leakage will inevitably be severe; similarly, if minimal or no leakage is desired, the frequency resolution will be lower. Therefore, in this embodiment of the invention, the frequency domain information of the electricity metering data signal is analyzed to select the optimal window function for harmonic identification.

[0042] Firstly, the spectral density feature quantization is used. The initial selection of the window function is based on the density or sparseness of frequency components in the spectral image of the energy metering data signal. Specifically, for sparse signals, some main lobe width can be sacrificed for better sidelobe suppression; for dense signals, high frequency resolution (narrow main lobe) is needed to distinguish adjacent frequencies. Specifically, the spectral density feature coefficient of the energy metering data signal is obtained based on the information entropy of the spectral image. Lower information entropy indicates that the energy of the energy metering data signal is more concentrated on a few frequencies, corresponding to a sparser spectrum; conversely, higher information entropy indicates that the frequency components in the spectral image are more complex and evenly distributed, corresponding to a denser spectrum. Therefore, in this embodiment of the invention, the information entropy of the spectral image is obtained, and this information entropy is used as the spectral density feature coefficient of the energy metering data signal. The larger the information entropy, the larger the spectral density feature coefficient.

[0043] The formula for calculating the spectral density characteristic coefficients of the electricity metering data signal is as follows:

[0044]

[0045] in, These are the spectral density characteristic coefficients. This indicates the number of amplitude values ​​in the spectrum image. This represents the k-th amplitude. This represents the probability of the k-th amplitude occurring.

[0046] Then, the signal type of the power metering data signal is determined based on the spectral density characteristic coefficient, thereby determining the spectral density of the power metering data signal. In existing technologies, under general signal analysis scenarios, an information entropy range of 3-4 is generally considered to be low entropy. Therefore, in this embodiment of the invention, the spectral density characteristic coefficient threshold is preferentially set to 4. If the spectral density characteristic coefficient is less than the threshold, the signal type of the power metering data signal is determined to be a sparse spectral signal. Furthermore, the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm can be set as the Blackman window, and the harmonic components in the power metering data signal can be identified using the windowed interpolation FFT algorithm for power quality analysis of the power system.

[0047] If the spectral density characteristic coefficient is greater than or equal to the spectral density characteristic coefficient threshold, the signal type of the electricity metering data signal is determined to be a dense spectrum signal, requiring high frequency resolution (narrow main lobe) to distinguish adjacent frequencies. However, if the electricity metering data signal is transient or contains short-term impulses, the window function is required to better capture the start and end of the signal in the time domain, typically requiring a window with a narrow main lobe. Conversely, if the electricity metering data signal contains frequency components with significantly different amplitudes, the window function is required to have low side lobes in the frequency domain to prevent strong signals from drowning out weak signals, requiring a window with a wide main lobe. Therefore, when determining that the electricity metering data signal is a dense spectrum signal, further refinement is needed to quantify the signal characteristics under different conditions, and then optimize the final selection of the optimal window function.

[0048] The specific detailed processing methods are as follows:

[0049] (1) Based on the amplitude distribution of each frequency in the spectrum image, obtain the transient characteristic coefficient of the power metering data signal, which is used to quantify the “sharp” or “flat” distribution shape of the power metering data signal, and reflect whether it contains short-term impacts or sudden changes (such as fault impacts or pulses).

[0050] Specifically, based on the amplitude of each frequency in the spectrum image, the kurtosis of the spectrum image is calculated, and the difference between the kurtosis and a preset value is used as the spectral transient characteristic coefficient of the power metering data signal.

[0051] The formula for calculating the instantaneous characteristic coefficients of the spectrum is:

[0052]

[0053] in, Represents the transient characteristic coefficients of the spectrum. This indicates the number of frequencies in the spectrum image. This represents the amplitude of the j-th frequency in the spectrum image. This represents the average amplitude of all frequencies in the spectrum image.

[0054] It should be noted that, Kurtosis, used to characterize a spectral image, indicates that the signal amplitude distribution is close to a normal (Gaussian) distribution when the kurtosis approaches 3. When the kurtosis is greater than 3, it is theoretically a peaked distribution, with a distribution curve steeper than a normal distribution, indicating the presence of transient impulses or abnormal peaks in the signal. Conversely, when the kurtosis is less than 3, the distribution curve is flatter than a normal distribution, indicating a more uniform signal lacking prominent peaks. Therefore, if... A value close to 0 indicates that the spectrum image is close to a normal distribution. A value less than 0 indicates that the spectrum image does not have a peak state. A value greater than 0 indicates that the spectral image has a peak state. At this time, by using a window function with a narrow main lobe, the instantaneous moment and shape can be accurately captured, avoiding blurring of time-domain details.

[0055] (2) Based on the frequency difference between the maximum and minimum amplitude values ​​in the spectrum image, obtain the dynamic range factor of the power metering data signal, which is used to quantify the amplitude difference between the strongest and weakest frequency components in the power metering data signal in order to assess the potential risk of frequency leakage.

[0056] Specifically, the maximum and minimum amplitude values ​​in the spectrum image are obtained, and the ratio between the absolute value of the maximum amplitude value and the absolute value of the minimum amplitude value is obtained. The ratio is converted into a decibel value using a preset logarithmic function, and the opposite of the decibel value is used as the independent variable of an exponential function with the natural constant as the base to obtain the spectral dynamic range factor of the power metering data signal.

[0057] The formula for calculating the spectral dynamic range factor is:

[0058]

[0059] in, Represents the dynamic range factor of the spectrum, in decibels. This represents an exponential function with the natural constant as its base. This indicates the linear amplitude ratio. Convert to decibels. This represents the maximum amplitude value in the spectrum image. The value represents the minimum amplitude in the spectrum image, and | represents the absolute value sign.

[0060] It should be noted that since the amplitude of voltage or current signals is usually directly measured in spectrum analysis, a factor of 20 is used in the formula. The decibel value is used instead of a linear amplitude ratio primarily because decibels offer advantages: compressing the ultra-large dynamic range, simplifying calculations, and facilitating graphical display and analysis. A larger decibel value indicates a more significant difference in amplitude between strong and weak components. This requires the window function to have low sidelobes in the frequency domain, i.e., a wider main lobe. Therefore, using... Perform inverse proportional normalization, so that The value range is between 0 and 1, and the closer it is to 0, the wider the main lobe needs to be; conversely, the closer it is to 1, the narrower the main lobe needs to be.

[0061] This completes the further refinement of the spectrally dense power metering data signal, yielding the spectral transient characteristic coefficients and spectral dynamic range factor of the power metering data signal.

[0062] Step S103: Based on the spectral transient characteristic coefficients and spectral dynamic range factor of the power metering data signal, obtain the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm. Based on the optimal window function, identify the harmonic components in the power metering data signal for power quality analysis of the power system.

[0063] In signal processing, windowing is used to reduce spectral leakage and the picket fence effect. When a signal is truncated, if the truncated signal is not periodic, spectral leakage occurs, meaning the signal energy is dispersed across multiple frequency points. Windowing reduces spectral leakage by multiplying the truncated signal by a smooth window function at both ends, causing the signal to gradually decrease to zero at the truncation point. Commonly used window functions include the rectangular window, the simplest window function, which directly truncates the signal but has a relatively large spectral leakage.

[0064] The Hanning Window has lower side lobes and is suitable for most cases; the Hamming Window is similar to the Hanning Window but has better side suppression; the Blackman Window has lower side lobes but increased main lobe width; the Kaiser Window can achieve a balance between main width and side suppression by adjusting parameters.

[0065] Therefore, after obtaining the spectral transient characteristic coefficients and spectral dynamic range factor of the power metering data signal, the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm can be obtained based on these coefficients. The specific method is as follows:

[0066] In the first case, if the transient characteristic coefficient of the spectrum is less than the constant 0 and the dynamic range factor of the spectrum is less than the preset dynamic range factor threshold, a wider window function is required. Therefore, the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to the Blackman window.

[0067] In the second case, if the transient characteristic coefficient of the spectrum is greater than the constant 0 and the dynamic range factor of the spectrum is greater than the preset dynamic range factor threshold, a narrower window function is used. Thus, the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to a rectangular window, which has the core feature of high frequency resolution but relatively serious sidelobe leakage.

[0068] In the third case, if the transient characteristic coefficient of the spectrum is greater than or equal to a constant 0, and the dynamic range factor of the spectrum is less than or equal to the preset dynamic range factor threshold, then it is considered that the frequency transient change and amplitude change characteristics of the power metering data signal are not particularly obvious. In this case, the best window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to the Hanning window or the Hamming window. Its core feature is that it is relatively more comprehensive and balanced, and can achieve a better balance between frequency resolution and sidelobe suppression.

[0069] In the fourth case, if the transient characteristic coefficient of the spectrum is less than or equal to a constant 0, and the dynamic range factor of the spectrum is greater than or equal to the preset dynamic range factor threshold, then the frequency transient change and amplitude change characteristics of the power metering data signal are not particularly obvious. In this case, the optimal window function for harmonic identification of the power metering data signal using the windowed interpolation FFT algorithm is set to the Hanning window or the Hamming window. Its core feature is that it is relatively more comprehensive and balanced, and can achieve a better balance between frequency resolution and sidelobe suppression.

[0070] It should be noted that the value range of the spectral dynamic range factor threshold is 0-1. In this embodiment of the invention, the spectral dynamic range factor threshold is set to 0.5. The reason for this setting is to balance accuracy and computing power. An excessively large or small spectral dynamic range factor threshold imposes stricter or more lenient requirements on the power metering data signal, which may lead to more power metering data signals being classified into the first or second case. The window functions used in the first and second cases are relatively complex and have poor overall balance. A spectral dynamic range factor threshold of 0.5 can make the power metering data signal be classified into the third or fourth case as much as possible. In this scenario, the window function (Hamming window or Hanning window) is commonly used and is the default window function in the windowed interpolation FFT algorithm. This window function has a stronger balance for signals with relatively common features, that is, it can achieve a good balance between frequency resolution and sidelobe suppression, while having strong computing power and low complexity.

[0071] Thus, the optimal window function was obtained when the power metering data signal had a dense spectrum. The window function in the windowed interpolation FFT algorithm was then set as the optimal window function, and the windowed interpolation FFT algorithm was used to analyze the harmonic components (the content of each harmonic and the total harmonic distortion (THD)) in the power metering data signal. This assessed the degree of harmonic pollution and its impact on the power grid and equipment, enabling effective elimination of harmonic interference during power quality analysis. This allows for effective analysis of power metering data in the direction of power quality, such as power system fault detection.

[0072] It is worth noting that the focus of this invention is on how to optimally select the window function in the windowed interpolation FFT algorithm to improve the elimination effect of harmonic components in the signal. However, the analysis of power quality direction based on the harmonic-eliminated signal is an existing technical method and will not be described in detail here.

[0073] Based on the same inventive concept as the above method, this embodiment of the invention also provides an energy metering data analysis system based on digital signal processing, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described energy metering data analysis methods based on digital signal processing.

[0074] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. 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. Such 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, and should all be included within the protection scope of the present invention.

Claims

1. A digital signal processing based electric energy metering data analysis method, characterized by, The method comprises: acquiring an electric energy metering data signal in the operation process of a power system, performing frequency domain conversion on the electric energy metering data signal to obtain a corresponding frequency spectrum image; obtaining a frequency spectrum density characteristic coefficient of the electric energy metering data signal according to the information entropy of the frequency spectrum image, the greater the information entropy, the greater the frequency spectrum density characteristic coefficient, determining the signal type of the electric energy metering data signal according to the frequency spectrum density characteristic coefficient, if the signal type is a dense frequency spectrum signal, obtaining a frequency spectrum transient characteristic coefficient of the electric energy metering data signal according to the amplitude distribution of each frequency in the frequency spectrum image, obtaining a frequency spectrum dynamic range factor of the electric energy metering data signal according to the frequency difference between the maximum amplitude and the minimum amplitude in the frequency spectrum image; obtaining the best window function for harmonic identification of the electric energy metering data signal by using a windowed interpolation FFT algorithm according to the frequency spectrum transient characteristic coefficient and the frequency spectrum dynamic range factor of the electric energy metering data signal, identifying the harmonic component in the electric energy metering data signal according to the best window function, and using the same for power quality analysis of the power system.

2. The method of claim 1, wherein, The determination of the signal type of the electric energy metering data signal according to the frequency spectrum density characteristic coefficient comprises: obtaining a preset frequency spectrum density characteristic coefficient threshold value, if the frequency spectrum density characteristic coefficient is greater than or equal to the frequency spectrum density characteristic coefficient threshold value, determining that the signal type of the electric energy metering data signal is a dense frequency spectrum signal, and if the frequency spectrum density characteristic coefficient is less than the frequency spectrum density characteristic coefficient threshold value, determining that the signal type of the electric energy metering data signal is a sparse frequency spectrum signal.

3. The digital signal processing based electric energy metering data analysis method as claimed in claim 2, wherein, After the determination of the signal type of the electric energy metering data signal according to the frequency spectrum density characteristic coefficient, the method comprises: if the signal type of the electric energy metering data signal is a sparse frequency spectrum signal, setting the best window function for harmonic identification of the electric energy metering data signal by using a windowed interpolation FFT algorithm as a Blackman window, and identifying the harmonic component in the electric energy metering data signal by using a windowed interpolation FFT algorithm, and using the same for power quality analysis of the power system.

4. The digital signal processing based electric energy metering data analysis method as claimed in claim 1, wherein, The obtaining of the frequency spectrum transient characteristic coefficient of the electric energy metering data signal according to the amplitude distribution of each frequency in the frequency spectrum image comprises: calculating the kurtosis of the frequency spectrum image according to the amplitude of each frequency in the frequency spectrum image, and taking the difference between the kurtosis and a preset value as the frequency spectrum transient characteristic coefficient of the electric energy metering data signal.

5. The digital signal processing based electric energy metering data analysis method as claimed in claim 1, wherein, The obtaining of the frequency spectrum dynamic range factor of the electric energy metering data signal according to the frequency difference between the maximum amplitude and the minimum amplitude in the frequency spectrum image comprises: obtaining the maximum amplitude and the minimum amplitude in the frequency spectrum image, obtaining the ratio between the absolute value of the maximum amplitude and the absolute value of the minimum amplitude, converting the ratio into a decibel value by using a preset logarithmic function, taking the reciprocal of the decibel value as the argument of an exponential function with a natural constant as the base, and obtaining the frequency spectrum dynamic range factor of the electric energy metering data signal.

6. The digital signal processing based electric energy metering data analysis method as claimed in claim 1, wherein, The best window function for harmonic identification of the electric energy metering data signal by using the windowed interpolation FFT algorithm is obtained according to the spectral transient characteristic coefficient and the spectral dynamic range factor of the electric energy metering data signal, and the best window function comprises: if the spectral transient characteristic coefficient is less than the constant 0 and the spectral dynamic range factor is less than the preset spectral dynamic range factor threshold, the best window function for harmonic identification of the electric energy metering data signal by using the windowed interpolation FFT algorithm is set as the Blackman window; if the spectral transient characteristic coefficient is greater than the constant 0 and the spectral dynamic range factor is greater than the preset spectral dynamic range factor threshold, the best window function for harmonic identification of the electric energy metering data signal by using the windowed interpolation FFT algorithm is set as the rectangular window; if the spectral transient characteristic coefficient is greater than or equal to the constant 0 and the spectral dynamic range factor is less than or equal to the preset spectral dynamic range factor threshold, the best window function for harmonic identification of the electric energy metering data signal by using the windowed interpolation FFT algorithm is set as the Hanning window or the Hamming window; if the spectral transient characteristic coefficient is less than or equal to the constant 0 and the spectral dynamic range factor is greater than or equal to the preset spectral dynamic range factor threshold, the best window function for harmonic identification of the electric energy metering data signal by using the windowed interpolation FFT algorithm is set as the Hanning window or the Hamming window.

7. A digital signal processing based electric energy metering data analysis system comprising a memory, a processor and a computer program stored in the memory and running on the processor, characterized in that, The processor executes the computer program to realize the steps of the electric energy metering data analysis method based on digital signal processing in any one of claims 1-6.

Images