A natural electromagnetic pulse signal processing method based on K-line dynamic modeling

Abstract

This invention discloses a natural electromagnetic pulse signal processing method based on dynamic K-line modeling. First, the natural electromagnetic pulse signal is divided into time windows. Then, an integrated empirical mode decomposition algorithm is introduced to process each time window, extracting the intrinsic signal functions of each time window. Next, the amplitude values ​​of the intrinsic signal functions with the same number in each time window are set together, obtaining multiple sets which are then sequentially numbered. A K-line rectangle is constructed for each numbered set, and all numbered sets are arranged in ascending order of frequency to form a dynamic K-line chart. This enhances the interpretability of signal features, and the time-energy-frequency three-dimensional encoding mechanism of this dynamic K-line chart can intuitively reflect the spatiotemporal evolution of the pulse signal. Furthermore, the collected natural electromagnetic pulse signals are continuously subjected to high-precision analytical processing, and the signal change characteristics of the dynamic K-line chart are updated in real time, thus providing data support for urban geological disaster early warning and power grid safety protection scenarios.

Description

Technical Field

[0001] This invention relates to a natural electromagnetic pulse signal processing method based on dynamic K-line modeling, belonging to the field of signal processing technology. Background Technology

[0002] Urban natural electromagnetic pulse signals are an important component of the complex electromagnetic environment, and their sources include lightning activity (peak electric field strength reaching 10). 4 Electromagnetic radiation, including electromagnetic pulses (e.g., V / m), micro-vibrations in geological structures (such as the piezoelectric effect generated by fault friction), and electromagnetic radiation caused by underground fluid movement, is generated by various sources. These signals have wide frequency ranges (0.1Hz-100kHz), low signal-to-noise ratios (<10dB), and non-steady-state characteristics. They serve as both precursory information carriers for geological disaster early warning and key indicators for assessing urban power grid safety (electromagnetic pulses can affect the power grid). Traditional signal extraction and processing methods mainly rely on time-frequency analysis techniques (such as Fourier transform and wavelet analysis), but they have significant shortcomings in dynamic characterization, anti-interference, and real-time performance. On the one hand, conventional algorithms struggle to analyze the time-frequency coupling characteristics of transient pulses, resulting in insufficient modeling of the correlation between signal energy, phase, and duration. On the other hand, artificial noise from urban environments such as 5G base stations (3.5GHz band) and new energy charging piles overlaps with the spectrum of natural signals, causing filtering in existing methods to easily result in effective signal loss, thus failing to accurately obtain signal change characteristics.

[0003] In recent years, with the upgrading of sensor technology and the development of intelligent algorithms, signal processing has entered the dynamic analysis stage, but it still lacks the visual representation of multi-dimensional features. Furthermore, the need for minute-level real-time acquisition of natural electromagnetic pulse characteristic changes in scenarios such as urban geological disaster early warning and power grid security protection further highlights the necessity of constructing a dynamic processing framework.

[0004] Therefore, the research direction of this invention is to provide a new method for processing natural electromagnetic pulse signals, which can perform high-precision analysis and processing of collected urban natural electromagnetic pulse signals and dynamically update signal change characteristics in real time, so as to provide data support for urban geological disaster early warning and power grid safety protection. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention provides a natural electromagnetic pulse signal processing method based on K-line dynamic modeling. This method can perform high-precision analysis and processing on the collected urban natural electromagnetic pulse signals and dynamically update the signal change characteristics in real time, providing data support for urban geological disaster early warning and power grid safety protection scenarios.

[0006] To achieve the above objectives, the technical solution adopted by this invention is: a natural electromagnetic pulse signal processing method based on K-line dynamic modeling, comprising the following steps:

[0007] Step 1: Obtain a segment of natural electromagnetic pulse signal and divide it into multiple time windows.

[0008] Step 2: Perform modal decomposition on each time window in Step 1 to extract intrinsic signal functions, and number each intrinsic signal function sequentially from low frequency to high frequency, and record the dominant frequency corresponding to each number; at the same time, number the amplitude value corresponding to the dominant frequency of each intrinsic signal function.

[0009] Step 3: Collect the amplitude values ​​of the intrinsic signal functions with the same number in each time window after processing in Step 2, obtain multiple sets, and number them sequentially.

[0010] Step 4: Select the first numbered set and calculate the maximum, minimum, mean, and median amplitudes within that set.

[0011] Step 5: Use the amplitude data obtained in Step 4 to create the initial electromagnetic signal K-line rectangle corresponding to the first numbered set.

[0012] Step 6: Calculate the amplitude distribution density inside the initial electromagnetic signal K-line rectangle to obtain the corresponding distribution probability; and fill the K-line rectangle with different distribution probabilities using different grayscale values ​​to finally form the K-line rectangle of the first numbered set.

[0013] Step 7: Repeat steps 4 to 6 to form a rectangular frame for all numbered sets of candlesticks.

[0014] Step 8: Arrange all the numbered sets of K-line rectangles according to frequency from smallest to largest to form a dynamic K-line chart of this segment of natural electromagnetic pulse signal.

[0015] Step 9: Subsequently, acquire natural electromagnetic pulse signals at different time periods, and repeat steps 1 to 8 in sequence to continuously obtain dynamic K-line charts of natural electromagnetic pulse signals at different time periods.

[0016] Furthermore, in step two, the intrinsic signal functions are numbered IMF1 to IMF1 in order from low frequency to high frequency. n The corresponding main frequency is denoted as f. im1 ~f imn Each major frequency contains M amplitude values, denoted as A1 to A2. m .

[0017] Furthermore, in step two, mode decomposition is performed for each time window to extract intrinsic signal functions, specifically as follows:

[0018] ① Detect extreme points: For the electromagnetic signal E(t) in the current time window, detect all local maxima and minima to form an extreme value sequence {p}. max (t)} and {p min (t)}.

[0019] ② Generate the envelope by connecting the maxima and minima using cubic spline interpolation to generate the upper and lower envelopes p. max (t) and p min (t).

[0020] ③ Calculate the average envelope

[0021] ④ Subtract the average envelope from the original electromagnetic signal of the current time window to obtain the intermediate signal E. k (t)=E k-1 Let IMF be the intrinsic signal function, and check E(t)-m(t). k (t) Does it satisfy the IMF condition: that is, the difference between the number of extreme points and the number of zero-crossing points is ≤1, and the mean of the local envelope is zero; if not, repeat steps ① to ④ until E k (t) satisfies the IMF conditions.

[0022] ⑤ Separate the current IMF into the k-th IMF, denoted as IMF. k =E k (t), at which point the residual term

[0023] ⑥ Repeat steps ① to ⑤ above to process the residual term r(t) until it becomes monotonic or constant, thereby obtaining the remaining IMF and completing the extraction of the intrinsic signal function of the current time window.

[0024] Furthermore, step five specifically involves: using the width of the window defined in step one as the width of the rectangular frame, and taking the 25% median value A. 1med The lower limit and the 175% median A 1med The upper limit is the value, and the difference between the two is the length of the rectangle. Draw a perpendicular line from the midpoint of the bottom width to the minimum value A. 1min Draw a perpendicular line from the midpoint of the width at the top to the maximum value A. 1max , take the mean A 1avg and median value A 1med The rectangle markings indicate that the initial electromagnetic signal K-line rectangle has been created.

[0025] Furthermore, step six specifically involves: dividing the interior of the initial electromagnetic signal K-line rectangle into n consecutive ΔA intervals based on the amplitude from the lower limit to the upper limit, denoted as A... Δ1 <A Δ2 <...<A Δn Let L be the total number of amplitude values ​​in the first set IMF1 that satisfy the upper and lower limits of the initial electromagnetic signal K-line rectangle; then, count the data frequencies L1, L2...Ln in each ΔA interval, and then calculate the distribution density of amplitude values ​​in each ΔA interval, ρ. i =Li / L, i = 1, 2…n; thus obtaining ρ1, ρ2…ρ_n for multiple consecutive intervals of ΔA. n After normalization, the probability distribution ρ of each ΔA interval is obtained. 1Nor ρ 2Nor …ρ nNor The probability distribution of each ΔA interval is displayed using different grayscale values ​​and filled into the corresponding interval positions of the K-line rectangle, thus finally forming the K-line rectangle of the first numbered set.

[0026] Furthermore, step eight specifically involves: arranging all numbered sets of IMFs into K-line rectangles according to their frequencies from smallest to largest, where the horizontal axis represents the dominant frequency value of the intrinsic signal function and the vertical axis represents the amplitude value; continuously connecting the median and mean values ​​of adjacent intrinsic signal functions (IMFs) to form a median curve and a mean curve, ultimately forming a dynamic K-line chart of the natural electromagnetic pulse signal segment.

[0027] Compared with the prior art, the present invention has the following advantages:

[0028] (1) Intrinsic signal extraction: This invention first divides a natural electromagnetic pulse signal into time windows, and then introduces an integrated empirical mode decomposition algorithm to process each time window and extract the intrinsic signal function of each time window, thereby effectively separating interference noise and reducing the influence of noise on the pulse signal.

[0029] (2) Multidimensional signal fusion: This invention utilizes the K-line chart technology in the financial field to provide a new approach for dynamic modeling. Combining the characteristics of time-domain pulse waveforms and frequency-domain energy distribution, the amplitude values ​​of intrinsic signal functions with the same number in each time window are first set together to obtain multiple sets and number them sequentially. K-line rectangles are constructed for each set of numbers. All K-line rectangles of numbered sets are arranged in ascending order of frequency to form a dynamic K-line chart of natural electromagnetic pulse signals, which enhances the interpretability of signal characteristics. Furthermore, the time-energy-frequency three-dimensional encoding mechanism of this dynamic K-line chart can intuitively reflect the spatiotemporal evolution of pulse signals.

[0030] (3) Real-time dynamic update: This invention performs high-precision analysis processing by dynamically encoding the period, amplitude and spectrum parameters of natural electromagnetic pulse signals, and updates the signal change characteristics of dynamic K-line charts in real time, thereby providing data support for urban geological disaster early warning and power grid safety protection scenarios. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the intrinsic signal function obtained after processing a segment of natural electromagnetic pulse signal in this invention;

[0032] Figure 2 This is a schematic diagram of the initial electromagnetic signal K-line rectangle for establishing a single number set in this invention;

[0033] Figure 3 This is a dynamic candlestick chart of the natural electromagnetic pulse signal generated after processing by this invention. Detailed Implementation

[0034] The present invention will be further described below.

[0035] like Figure 1 As shown, the present invention includes the following steps:

[0036] Step 1: Obtain a sampling rate of F s The natural electromagnetic pulse signal is obtained and divided into multiple time windows, each with a length of M points, thereby obtaining T1~T n There are several time windows, each with a duration of T = M / F. s .

[0037] Step 2: Perform mode decomposition on each time window in Step 1 to extract the intrinsic signal function, specifically as follows:

[0038] ① Detect extreme points: For the electromagnetic signal E(t) in the current time window, detect all local maxima and minima to form an extreme value sequence {p}. max (t)} and {p min (t)};

[0039] ② Generate the envelope by connecting the maxima and minima using cubic spline interpolation to generate the upper and lower envelopes p. max (t) and p min (t);

[0040] ③ Calculate the average envelope

[0041] ④ Subtract the average envelope from the original electromagnetic signal of the current time window to obtain the intermediate signal E. k (t)=E k-1 Let IMF be the intrinsic signal function, and check E(t)-m(t). k (t) Does it satisfy the IMF condition: that is, the difference between the number of extreme points and the number of zero-crossing points is ≤1, and the mean of the local envelope is zero; if not, repeat steps ① to ④ until E k (t) satisfies the IMF condition;

[0042] ⑤ Separate the current IMF into the k-th IMF, denoted as IMF. k =E k (t), at which point the residual term

[0043] ⑥ Repeat steps ① to ⑤ above to process the residual term r(t) until it becomes monotonic or constant, thereby obtaining the remaining IMF and completing the extraction of the intrinsic signal function of the current time window.

[0044] Then, from low frequency to high frequency, each intrinsic signal function is sequentially numbered as IMF1 to IMF2. n And record the main frequency f corresponding to each number. im1 ~f imn Each major frequency contains M amplitude values, denoted as A1 to A2. m .

[0045] Step 3: Collect the amplitude values ​​of the intrinsic signal function with the same number in each time window after processing in Step 2, obtain multiple sets, and number them sequentially; taking the first numbered set as an example, it consists of T1 to T... n The amplitude A corresponding to the same IMF1 number in each time window 11 A 12 …A 1n composition.

[0046] Step 4: Select the first numbered set and calculate the maximum amplitude A within that set. 1max Minimum value A 1min Mean A 1avg Median A 1med The mean is The median value is calculated by arranging the amplitude values ​​in ascending order. If the number of amplitudes is odd, the middle amplitude value is taken as the median value. If the number is even, the average of the two middle amplitude values ​​is taken as the median value.

[0047] Step 5: Using the amplitude data obtained in Step 4, construct the initial electromagnetic signal K-line rectangle corresponding to the first IMF1 set, as shown below. Figure 2 As shown, specifically: the width of the window is used as the width of the rectangle based on the width defined in step one, and the median value A is taken as 25%. 1med The lower limit and the 175% median A 1med The upper limit is the value, and the difference between the two is the length of the rectangle. Draw a perpendicular line from the midpoint of the bottom edge of the rectangle to the minimum value A. 1min Draw a perpendicular line from the midpoint of the width of the rectangle to the maximum value A. 1max , take the mean A 1avg and median value A 1med The rectangle markings indicate that the initial electromagnetic signal K-line rectangle has been created.

[0048] Step Six: Divide the initial electromagnetic signal K-line rectangle into n consecutive ΔA intervals according to the amplitude from the lower limit to the upper limit, denoted as A. Δ1 <A Δ2 <...<A ΔnLet L be the total number of amplitude values ​​in the first set IMF1 that satisfy the upper and lower limits of the initial electromagnetic signal K-line rectangle; then, count the data frequencies L1, L2...Ln in each ΔA interval, and then calculate the distribution density of amplitude values ​​in each ΔA interval, ρ. i =L i / L, i = 1, 2…n; thus obtaining ρ1, ρ2…ρ_n for multiple consecutive intervals of ΔA. n After normalization, the probability distribution ρ of each ΔA interval is obtained. 1Nor ρ 2Nor …ρ nNor The probability distribution of each ΔA interval is displayed using different grayscale values ​​and filled into the corresponding interval positions of the K-line rectangle, thus finally forming the K-line rectangle of the first IMF1 set.

[0049] Step 7: Repeat steps 4 to 6 to form the K-line rectangles for all numbered IMFi sets.

[0050] Step 8: Arrange all numbered sets of IMFs (IMFs) into K-line rectangles according to frequency from smallest to largest. The horizontal axis represents the dominant frequency of the IMF, and the vertical axis represents the amplitude. Connect the median and mean values ​​of adjacent IMFs to form median and mean curves, as shown below. Figure 3 As shown, the final result is a dynamic candlestick chart of this segment of natural electromagnetic pulse signal.

[0051] Step 9: Subsequently, acquire natural electromagnetic pulse signals at different time periods, and repeat steps 1 to 8 in sequence to continuously obtain dynamic K-line charts of natural electromagnetic pulse signals at different time periods.

[0052] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for processing natural electromagnetic pulse signals based on dynamic K-line modeling, characterized in that, Includes the following steps: Step 1: Acquire a segment of natural electromagnetic pulse signal and divide it into multiple time windows; Step 2: Perform modal decomposition on each time window in Step 1 to extract intrinsic signal functions, and number each intrinsic signal function sequentially from low frequency to high frequency, and record the dominant frequency corresponding to each number; at the same time, number the amplitude value corresponding to the dominant frequency of each intrinsic signal function. Step 3: Collect the amplitude values ​​of the intrinsic signal function with the same number in each time window after processing in Step 2, obtain multiple sets, and number them sequentially. Step 4: Select the first numbered set and calculate the maximum, minimum, mean, and median amplitudes within that set; Step 5: Use the amplitude data obtained in Step 4 to create the initial electromagnetic signal K-line rectangle corresponding to the first numbered set; Step 6: Calculate the amplitude distribution density inside the initial electromagnetic signal K-line rectangle to obtain the corresponding distribution probability; and fill the K-line rectangle with different distribution probabilities using different grayscale values ​​to ultimately form the K-line rectangle of the first numbered set. Step 7: Repeat steps 4 to 6 to form a rectangular frame for all numbered sets of candlesticks; Step 8: Arrange all the numbered sets of K-line rectangles according to frequency from smallest to largest to form a dynamic K-line chart of this segment of natural electromagnetic pulse signal; Step 9: Subsequently, acquire natural electromagnetic pulse signals at different time periods, and repeat steps 1 to 8 in sequence to continuously obtain dynamic K-line charts of natural electromagnetic pulse signals at different time periods.

2. The natural electromagnetic pulse signal processing method based on K-line dynamic modeling according to claim 1, characterized in that, In step two, the intrinsic signal functions are numbered sequentially from low frequency to high frequency as IMF1 to IMF2. n The corresponding main frequency is denoted as f. im1 ~f imn Each major frequency contains M amplitude values, denoted as A1 to A2. m .

3. The natural electromagnetic pulse signal processing method based on K-line dynamic modeling according to claim 1, characterized in that, In step two, modal decomposition is performed for each time window to extract intrinsic signal functions, specifically as follows: ① Detect extreme points: For the electromagnetic signal E(t) in the current time window, detect all local maxima and minima to form an extreme value sequence {p}. max (t)} and {p min (t)}; ② Generate the envelope by connecting the maxima and minima using cubic spline interpolation to generate the upper and lower envelopes p. max (t) and p min (t); ③ Calculate the average envelope ④ Subtract the average envelope from the original electromagnetic signal of the current time window to obtain the intermediate signal E. k (t)=E k-1 Let IMF be the intrinsic signal function, and check E(t)-m(t). k (t) Does it satisfy the IMF condition: that is, the difference between the number of extreme points and the number of zero-crossing points is ≤1, and the mean of the local envelope is zero; if not, repeat steps ① to ④ until E k (t) satisfies the IMF condition; ⑤ Separate the current IMF into the k-th IMF, denoted as IMF. k =E k (t), at which point the residual term ⑥ Repeat steps ① to ⑤ above to process the residual term r(t) until it becomes monotonic or constant, thereby obtaining the remaining IMF and completing the extraction of the intrinsic signal function of the current time window.

4. The natural electromagnetic pulse signal processing method based on K-line dynamic modeling according to claim 1, characterized in that, Step five specifically involves: using the width of the window divided in step one as the width of the rectangle, taking the 25% median as the lower limit and the 175% median as the upper limit, and the difference between the two as the length of the rectangle. Then, draw a perpendicular line from the midpoint of the lower edge width of the rectangle to the minimum value, and draw a perpendicular line from the midpoint of the upper edge width to the maximum value. Mark and display the mean and median values ​​in the rectangle to complete the initial electromagnetic signal K-line rectangle.

5. The natural electromagnetic pulse signal processing method based on K-line dynamic modeling according to claim 1, characterized in that, Step six specifically involves dividing the interior of the initial electromagnetic signal K-line rectangle into n consecutive ΔA intervals based on the amplitude from the lower limit to the upper limit, denoted as A... Δ1 <A Δ2 <...<A Δn Let L be the total number of amplitude values ​​in the first set that satisfy the upper and lower limits of the initial electromagnetic signal K-line rectangle; then, count the data frequencies L1, L2...Ln in each ΔA interval, and then calculate the amplitude distribution density ρ in each ΔA interval. i =L i / L, i = 1, 2…n; thus obtaining ρ1, ρ2…ρ_n for multiple consecutive intervals of ΔA. n After normalization, the probability distribution ρ of each ΔA interval is obtained. 1Nor ρ 2Nor …ρ nNor The probability distribution of each ΔA interval is displayed using different grayscale values ​​and filled into the corresponding interval positions of the K-line rectangle, thus finally forming the K-line rectangle of the first numbered set.

6. The natural electromagnetic pulse signal processing method based on K-line dynamic modeling according to claim 1, characterized in that, Step eight specifically involves arranging all numbered sets of K-line rectangles according to frequency from smallest to largest, where the horizontal axis represents the dominant frequency value of the intrinsic signal function and the vertical axis represents the amplitude value. The median and mean values ​​of adjacent intrinsic signal functions are continuously connected to form the median curve and the mean curve, ultimately forming a dynamic K-line chart of the natural electromagnetic pulse signal.

Images