Traveling wave waveform detection method and device based on empirical mode decomposition and random Kaczmarz algorithm
Through the combination method of empirical modal decomposition and random Kaczmarz algorithm, the problem of modal aliasing effect and noise influence in traveling wave signal inversion is solved, and the precise detection of traveling wave waveform is realized, and the reliability of fault positioning and protection is improved. .
Patent Information
- Application Number
- CN202411209758.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-12-06
AI Technical Summary
The existing technology has modal aliasing effect, difficulty in selecting regularized parameters and noise influences in the process of inversion of traveling wave signal, resulting in inaccurate detection of traveling wave waveforms and ineffective isolation of faults, which may cause system oscillation and major power outages.
Using the method based on empirical modal decomposition (EMD) and random Kaczmarz algorithm, the transfer function model of the traveling wave sensor is constructed, and the quadratic traveling wave signals are decomposed collectively, and the inherent modes with similarity greater than the preset value are extracted. Components and use the random Kaczmarz algorithm to invert, reduce noise interference and calculation difficulty, and realize accurate detection of a traveling wave signal.
Effectively reduce the influence of modal aliasing effect on inversion accuracy and improve the accuracy of traveling wave waveform detection. The simulation and experimental results show that the primary traveling wave signal obtained by inversion is very similar to the simulation results, realizing the traveling waveform. Accurate detection improves the reliability of fault location and protection.
Smart Images

Figure CN119087124A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of traveling wave detection, and in particular relates to a traveling wave waveform detection method based on empirical mode decomposition and random Kaczmarz algorithm. Background Art
[0002] In recent years, traveling wave technology has become a research focus due to its advantages of fast response speed and not being affected by factors such as line distributed capacitance, system oscillation, and current sensor saturation. This technology analyzes and processes high-frequency transient traveling wave signals containing rich fault information, completes fault detection and positioning, and effectively improves the stability and safety of transmission network operation. The key to the practical application of traveling wave technology lies in the accurate detection of traveling wave signals. Traveling wave signals are wide-band transient signals that usually appear in the faulty power grid in the form of current and voltage. Conventional transformers are only suitable for measuring power frequency voltage and current, but traveling wave sensors can effectively transmit and detect high-frequency signals using Rogowski coils, and have the advantages of high measurement accuracy, fast response, and wide range. However, due to the physical structure of the Rogowski coil itself, the primary traveling wave signal is distorted after passing through the traveling wave sensor, and the traveling wave protection and positioning method may not be able to isolate the fault in time, which in turn causes system oscillation, decoupling, and even causes a major power outage. Therefore, how to accurately detect high-frequency transient traveling wave signals is an urgent problem to be solved in traveling wave positioning technology.
[0003] The existing method of solving the problem of secondary traveling wave waveform distortion by optimizing the sensor structure is not ideal. At present, the actual transient traveling wave signal in the faulty power grid is mainly obtained by traveling wave waveform inversion, that is, the real primary traveling wave signal is reversely derived from the known secondary traveling wave signal. Some existing technical literature uses Tikhonov and other regularization methods to solve the ill-conditioned problem in traveling wave inversion, but the selection of regularization parameters in the regularization method has a great influence on the inversion result.
[0004] The phase-frequency characteristics and amplitude-frequency characteristics of the traveling wave sensor based on the Rogowski coil are different in different frequency bands. The frequency domain information obtained by analyzing the non-stationary nonlinear traveling wave signal using Fourier transform can only reflect the average value of a certain frequency component in the entire signal, which is difficult to match the broadband transmission characteristics of the traveling wave sensor, so it is impossible to achieve local frequency domain analysis. Some existing technical documents use empirical mode decomposition (EMD), but the modal aliasing effect in signal decomposition will affect the inversion accuracy. Summary of the invention
[0005] The present invention provides a traveling wave waveform detection method based on empirical mode decomposition and random Kaczmarz algorithm, which solves the modal aliasing effect, regularization parameter selection and noise influence in the inversion process of the prior art, and realizes accurate detection of the traveling wave waveform of a primary traveling wave signal obtained by inverting a measured secondary traveling wave signal.
[0006] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:
[0007] A traveling wave waveform detection method based on empirical mode decomposition and random Kaczmarz algorithm, comprising:
[0008] S1, construct the transfer function model of the traveling wave sensor;
[0009] S2, performing collective empirical mode decomposition on the secondary traveling wave signal actually detected and output by the traveling wave sensor to obtain N intrinsic mode components; then using normalized energy similarity, extracting the intrinsic mode components with similarity greater than a preset value, and recording them as valid intrinsic mode components;
[0010] S3, for each valid intrinsic mode component, the random Kaczmarz algorithm is used to perform inversion to obtain the corresponding intrinsic mode component of the primary traveling wave signal;
[0011] S4, superimposing all the natural mode components obtained in step S3 to obtain an inverted primary traveling wave signal.
[0012] Furthermore, step S1 constructs a transfer function model of the traveling wave sensor, and the method is: simulating a power grid fault through a power grid fault traveling wave generator; connecting a non-inductive resistor and a capacitive voltage transformer in series on the output line of the power grid fault traveling wave generator, and sleeved a traveling wave sensor on the equivalent capacitor of the capacitive voltage transformer; measuring the terminal voltage of the non-inductive resistor through a digital oscilloscope, which is a primary traveling wave signal; measuring the terminal voltage of the wave sensor output through a digital oscilloscope, which is a secondary traveling wave signal; using the measured primary traveling wave signal and secondary traveling wave signal as the input and output of the transfer function model, respectively, to solve and obtain the transfer function model of the traveling wave sensor.
[0013] Furthermore, step S1 constructs the transfer function model of the traveling wave sensor as follows:
[0014]
[0015] Where H(s) represents the transfer function model, s = jω, C 1 , C 2 , C 3 , C 4 are the coefficients of the transfer function model.
[0016] Furthermore, step S2 performs collective empirical mode decomposition on the secondary traveling wave signal, including:
[0017] S211: Add the Gaussian white noise n for the mth time to the secondary traveling wave signal y(t) detected by the traveling wave sensor m (t), forming a new time series signal y after superimposing noisem (t);
[0018] S212: For the new time series signal y m (t) Perform empirical mode decomposition to obtain the corresponding N intrinsic mode components j=1,2,…,N and a residual component
[0019] S213: Repeat steps S211 and S212 M times, sum and average the j-th natural mode component obtained M times, and use it as the j-th natural mode component of the secondary traveling wave signal y(t)
[0020]
[0021] S214: Performing ensemble empirical mode decomposition on the secondary traveling wave signal y(t) to express it as:
[0022]
[0023] In the formula, each natural mode component and the remaining component R es The components of different frequency bands from high to low constitute the secondary traveling wave signal y(t) in sequence.
[0024] Furthermore, the normalized energy similarity calculation method is specifically as follows:
[0025] S221: Calculate each natural mode component And the energy of the secondary traveling wave signal y(t):
[0026]
[0027] In the formula, E j represents the jth intrinsic mode component The energy, E y represents the energy of the secondary traveling wave signal y(t), represents the jth intrinsic mode component The i-th sample data, represents the i-th sampling data of the secondary traveling wave signal y(t), and T is the total data length of the secondary traveling wave signal;
[0028] S222: Normalize the energies calculated in S221:
[0029]
[0030] In the formula, E j,norm Energy E j The normalized value of E y,norm Energy Ey The normalized value of The energy E corresponding to all natural mode components j and the energy E corresponding to the secondary traveling wave signal y(t) y The average value, std() represents the standard deviation calculation;
[0031] S223: Calculate the normalized energy E of each natural modal component j,norm The normalized energy E between the energy of the secondary traveling wave signal y,norm Similarity:
[0032]
[0033] In the formula, r j represents the normalized energy similarity between the jth intrinsic mode component and the secondary traveling wave signal.
[0034] Furthermore, the preset value for comparing similarity in step S2 is calculated using the following formula:
[0035]
[0036] In the formula, is the mean of the normalized energy similarity between the secondary traveling wave signal and all IMF components, r j represents the normalized energy similarity between the jth intrinsic mode component and the secondary traveling wave signal.
[0037] Furthermore, step S3 uses the random Kaczmarz algorithm to invert each intrinsic modal component, specifically:
[0038] S301: Based on the transfer function model of the traveling wave sensor, obtain the Toeplitz matrix A generated by the unit impulse response of the traveling wave sensor, and convert the lth effective natural mode component Let it be matrix y l , create matrices A and y l The relationship is:
[0039] y l =Ax l
[0040] In the formula, x l for y l The intrinsic mode components obtained by inversion;
[0041] S302: Data initialization, specifying the maximum number of iterations K, assigning an initial value of the number of iterations k = 0, and assigning an initial value of the iterations x l (0) =0;
[0042] S303: Update k=k+1; when the number of iterations k<K, in the kth iteration, according to the probability P i (k) Select the i-th row in matrix A:
[0043]
[0044] In the formula, represents the square of the Frobenius norm of matrix A; is the i-th row matrix of matrix A in the k-th iteration, For the matrix The transpose of
[0045] S304: The intrinsic modal component x of the previous iteration l (k-1) Projection onto the hyperplane On the update solution vector, we get vector x l (k) :
[0046]
[0047] In the formula, Representative Matrix The square of the 2-norm of ; y l,i for y l The i-th row of
[0048] S305: When the maximum number of iterations K is reached or the iteration convergence condition is met, the natural modal component x at the current number of iterations is output. l (k) As y l The corresponding inverted intrinsic modal component x l ;
[0049] The iterative convergence condition is:
[0050]
[0051] In the formula, x l (k) for y l Invert the intrinsic modal components of the k-th iteration output; x l (k-1) for y l Invert the natural mode components of the k-1th iteration output; p is a given positive integer.
[0052] A traveling wave waveform detection device based on empirical mode decomposition and random Kaczmarz algorithm, comprising:
[0053] The transfer function model building module is used to: build a transfer function model of the traveling wave sensor;
[0054] The modal decomposition and extraction module is used to: perform collective empirical mode decomposition on the secondary traveling wave signal actually detected and output by the traveling wave sensor to obtain N intrinsic modal components; then use normalized energy similarity to extract the intrinsic modal components with similarity greater than a preset value and record them as valid intrinsic modal components;
[0055] The component inversion module is used to: perform inversion on each valid intrinsic mode component using the random Kaczmarz algorithm to obtain the corresponding intrinsic mode component of the primary traveling wave signal;
[0056] The component synthesis module is used to superimpose all the inherent mode components obtained by the component inversion module to obtain an inverted primary traveling wave signal.
[0057] Compared with the prior art, the advantages of the present invention are:
[0058] The present invention uses EEMD decomposition to decompose the secondary traveling wave signal actually detected by the traveling wave sensor into components of different frequency bands, and then extracts the effective components through normalized energy similarity. The signals of each frequency band are all stable signals and have local characteristics, which effectively reduces the influence of modal aliasing effect on inversion accuracy; then the signal components of different frequency bands are inverted using the random Kaczmarz algorithm, which only needs to use the inversion algorithm, and no longer uses the forward transfer function model of the traveling wave sensor to derive its inverse function, which effectively reduces the calculation difficulty and noise interference. Simulation and experimental results show that the waveform similarity between the primary traveling wave signal obtained by the frequency division inversion of the present invention and the primary traveling wave signal obtained by simulation is extremely high, and accurate detection of the traveling wave waveform can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] Figure 1 It is a schematic flow diagram of the method of the present invention;
[0060] Figure 2 is a circuit model diagram in an embodiment of the present invention;
[0061] Figure 3 It is the simulated primary and secondary traveling wave waveform in the embodiment of the present invention;
[0062] Figure 4 is a schematic diagram of the result of EEMD decomposition of a secondary traveling wave in an embodiment of the present invention;
[0063] Figure 5 is a schematic diagram of the IMF' component obtained by inversion in an embodiment of the present invention;
[0064] Figure 6 It is a schematic diagram of the final inversion waveform obtained by synthesizing the IMF' components in the embodiment of the present invention. DETAILED DESCRIPTION
[0065] The following is a detailed description of an embodiment of the present invention. This embodiment is based on the technical solution of the present invention, and provides a detailed implementation method and a specific operation process to further explain the technical solution of the present invention.
[0066] like Figure 1 As shown, the traveling wave waveform detection method based on EEMD decomposition and random Kaczmarz algorithm of the present invention comprises the following steps:
[0067] S1, obtain the primary traveling wave signal generated by the power grid fault and the secondary traveling wave signal output by the traveling wave sensor, and build the transfer function model of the traveling wave sensor by measuring the primary traveling wave signal and the secondary traveling wave signal.
[0068] This embodiment is used to construct a primary traveling wave signal and a secondary traveling wave signal of a transfer function model, which are simulated and generated by a power grid fault traveling wave generator: a power grid fault is simulated by a power grid fault traveling wave generator; a non-inductive resistor (this small resistor is not only used for traveling wave signal measurement, but also has the function of current limiting, and the general resistance value is 10Ω~100Ω) and a capacitive voltage transformer are connected in series on the output line of the power grid fault traveling wave generator, and a traveling wave sensor is mounted on the equivalent capacitor of the capacitive voltage transformer; the terminal voltage of the non-inductive resistor is measured by a digital oscilloscope, which is the primary traveling wave signal; the terminal voltage of the wave sensor output is measured by a digital oscilloscope, which is the secondary traveling wave signal.
[0069] The measured primary traveling wave signal and secondary traveling wave signal are used as the input and output of the transfer function model respectively, and the transfer function model of the traveling wave sensor is solved.
[0070] S2, perform collective empirical mode decomposition on the secondary traveling wave signal actually detected and output by the traveling wave sensor to obtain N inherent modal components (IMF components) in different frequency bands; then use normalized energy similarity calculation to extract the inherent modal components with similarity greater than a preset value and record them as valid inherent modal components.
[0071] The ensemble empirical mode decomposition of the secondary traveling wave signal includes:
[0072] S211: Add the Gaussian white noise n for the mth time to the secondary traveling wave signal y(t) detected by the traveling wave sensor m (t), forming a new time series signal y after superimposing noise m (t);
[0073] S212: For the new time series signal y m (t) Perform empirical mode decomposition to obtain the corresponding N intrinsic mode components j=1,2,…,N and a residual component
[0074] S213: Repeat steps S211 and S212 M times, sum and average the j-th natural mode component obtained M times, and use it as the j-th natural mode component of the secondary traveling wave signal y(t)
[0075]
[0076] S214: Performing ensemble empirical mode decomposition on the secondary traveling wave signal y(t) to express it as:
[0077]
[0078] In the formula, each natural mode component and the remaining component R es The components of different frequency bands from high to low constitute the secondary traveling wave signal y(t) in sequence.
[0079] The specific calculation method of normalized energy similarity includes:
[0080] S221: Calculate each natural mode component And the energy of the secondary traveling wave signal y(t):
[0081]
[0082] In the formula, E j represents the jth intrinsic mode component The energy, E y represents the energy of the secondary traveling wave signal y(t), represents the jth intrinsic mode component The i-th sample data, represents the i-th sampling data of the secondary traveling wave signal y(t), and T is the total data length of the secondary traveling wave signal;
[0083] S222: Normalize the energies calculated in S221:
[0084]
[0085] In the formula, E j,norm Energy E j The normalized value of E y,norm Energy E y The normalized value of The energy E corresponding to all natural mode components j and the energy E corresponding to the secondary traveling wave signal y(t) y The average value, std() represents the standard deviation calculation;
[0086] S223: Calculate the normalized energy E of each natural modal component j,norm The normalized energy E between the energy of the secondary traveling wave signal y,norm Similarity:
[0087]
[0088] In the formula, r j represents the normalized energy similarity between the jth intrinsic mode component and the secondary traveling wave signal.
[0089] In addition, the preset value used to compare the similarity is calculated using the following formula:
[0090]
[0091] In the formula, is the mean of the normalized energy similarity between the secondary traveling wave signal and all IMF components, r j represents the normalized energy similarity between the jth intrinsic mode component and the secondary traveling wave signal.
[0092] If the normalized energy similarity between a certain intrinsic mode component and the secondary traveling wave signal is greater than the preset value, that is, r j >Th, and record the eigenmode component as the effective eigenmode component.
[0093] S3, for each valid intrinsic mode component, the random Kaczmarz algorithm is used to perform inversion to obtain the corresponding intrinsic mode component of the primary traveling wave signal.
[0094] Among them, the random Kaczmarz algorithm is used to invert each intrinsic mode component, specifically including:
[0095] S301: Based on the transfer function model of the traveling wave sensor, obtain the Toeplitz matrix A generated by the unit impulse response of the traveling wave sensor, and convert the lth effective natural mode component Let it be matrix y l , create matrices A and y l The relationship is:
[0096] y l =Ax l
[0097] In the formula, x l for y l The intrinsic mode components obtained by inversion;
[0098] S302: Data initialization, specifying the maximum number of iterations K, assigning an initial value of the number of iterations k = 0, and assigning an initial value of the iterations x l (0) =0;
[0099] S303: Update k=k+1; when the number of iterations k<K, in the kth iteration, according to the probability P i (k) Select the i-th row in matrix A:
[0100]
[0101] In the formula, represents the square of the Frobenius norm of matrix A; is the i-th row matrix of matrix A in the k-th iteration, For the matrix The transpose of
[0102] The Kaczmarz algorithm selects each row of the matrix for calculation in turn, which will affect the iteration speed. The random Kaczmarz algorithm selects a row according to probability. Randomly selecting rows can break the regularity of matrix A, prevent the algorithm from falling into a local optimal solution, and help improve the convergence speed of the algorithm.
[0103] S304: The intrinsic modal component x of the previous iteration l (k-1) Projection onto the hyperplane Update the solution vector to get vector x l (k) :
[0104]
[0105] In the formula, Representative Matrix The square of the 2-norm of ; y l,i for y l The i-th row of
[0106] S305: When the maximum number of iterations K is reached or the iteration convergence condition is met, the natural modal component x at the current number of iterations is output. l (k) As y l The corresponding inverted intrinsic modal component x l ;
[0107] The iterative convergence condition is:
[0108]
[0109] In the formula, x l (k) for y l Invert the intrinsic modal components of the k-th iteration output; x l (k-1) for y lInvert the natural mode components of the k-1th iteration output; p is a given positive integer.
[0110] S4, superimposing all the natural mode components obtained in step S3 to obtain an inverted primary traveling wave signal.
[0111] A specific application example of the traveling wave waveform inversion method proposed in this embodiment is as follows:
[0112] (1) The transfer function model of the traveling wave sensor is constructed as follows:
[0113]
[0114] (2) Figure 2 As shown in the figure, a simulation model of a 500kV transmission line is established by using the simulation software Simulink. In this model, the length of the transmission line MN is 300km, and the sampling frequency is set to 1MHz. First, a single-phase grounding fault is set at 100km of the 500kV transmission line MN, and a primary traveling wave signal is obtained by setting different transition resistances Rg and appropriate noise interference. Then, the primary traveling wave signal is passed through the voltage traveling wave sensor simulation model to obtain a simulated secondary traveling wave signal. The waveforms of the simulated primary and secondary traveling wave signals are shown in Figure 3 As shown, the waveforms of the primary and secondary traveling wave signals obtained by simulation are obviously different.
[0115] (3) Perform EEMD decomposition on the secondary traveling wave signal obtained by the transformation to obtain 8 IMF components arranged from high to low frequency, and then extract the waveform of each IMF component, such as Figure 4 shown.
[0116] (4) The effective IMF components are inverted in turn using the random Kaczmarz algorithm to obtain the IMF' components of the primary traveling wave signal, such as Figure 5 shown.
[0117] (5) Finally, all the inverted IMF' components are superimposed to obtain the inverted primary traveling wave signal. The inversion result is as follows: Figure 6 shown.
[0118] The simulation results show that the present invention effectively solves the problem of secondary traveling wave distortion of the dedicated traveling wave sensor. The obtained inverted primary traveling wave has a high similarity with the measured primary traveling wave, and the inversion result is close to the actual waveform. The method is not affected by the modal aliasing effect. The error between the fault information of the primary traveling wave obtained by inverting the present invention and the fault information of the measured primary traveling wave is small. Using the obtained inverted primary traveling wave for positioning and protection analysis can effectively improve the positioning accuracy and protection reliability.
[0119] The above embodiments are preferred embodiments of the present application. Ordinary technicians in this field can also make various changes or improvements on this basis. Without departing from the overall concept of the present application, these changes or improvements should fall within the scope of protection required by the present application.
Claims
1. A traveling wave waveform detection method based on empirical mode decomposition and random Kaczmarz algorithm, characterized in that: include: S1, construct the transfer function model of the traveling wave sensor; S2, performing collective empirical mode decomposition on the secondary traveling wave signal actually detected and output by the traveling wave sensor to obtain N inherent mode components; Then, the normalized energy similarity is used to extract the intrinsic modal components whose similarity is greater than a preset value and recorded as valid intrinsic modal components; S3, for each valid intrinsic mode component, the random Kaczmarz algorithm is used to perform inversion to obtain the corresponding intrinsic mode component of the primary traveling wave signal; S4, superimposing all the natural mode components obtained in step S3 to obtain an inverted primary traveling wave signal.
2. The traveling wave waveform detection method according to claim 1, characterized in that: Step S1 constructs a transfer function model of the traveling wave sensor, and the method is: simulating a power grid fault through a power grid fault traveling wave generator; connecting a non-inductive resistor and a capacitive voltage transformer in series on the output line of the power grid fault traveling wave generator, and sleeves a traveling wave sensor on the equivalent capacitor of the capacitive voltage transformer; measuring the terminal voltage of the non-inductive resistor through a digital oscilloscope, which is a primary traveling wave signal; measuring the terminal voltage output by the wave sensor through a digital oscilloscope, which is a secondary traveling wave signal; using the measured primary traveling wave signal and secondary traveling wave signal as the input and output of the transfer function model, respectively, to solve and obtain the transfer function model of the traveling wave sensor.
3. The traveling wave waveform detection method according to claim 1, characterized in that: Step S1 constructs the transfer function model of the traveling wave sensor as: Wherein, H(s) represents the transfer function model, s=jω, and C1, C2, C3, and C4 are coefficients of the transfer function model, respectively.
4. The traveling wave waveform detection method according to claim 1, characterized in that: Step S2 performs collective empirical mode decomposition on the secondary traveling wave signal, including: S211: Add the Gaussian white noise n for the mth time to the secondary traveling wave signal y(t) detected by the traveling wave sensor m (t), forming a new time series signal y after superimposing noise m (t); S212: For the new time series signal y m (t) Perform empirical mode decomposition to obtain the corresponding N intrinsic mode components and a remaining component S213: Repeat steps S211 and S212 M times, sum and average the j-th natural mode component obtained M times, and use it as the j-th natural mode component of the secondary traveling wave signal y(t) S214: Performing ensemble empirical mode decomposition on the secondary traveling wave signal y(t) to express it as: In the formula, each natural mode component and the remaining component R es The components of different frequency bands from high to low constitute the secondary traveling wave signal y(t) in sequence.
5. The traveling wave waveform detection method according to claim 1, characterized in that: The normalized energy similarity calculation method is as follows: S221: Calculate each natural mode component And the energy of the secondary traveling wave signal y(t): In the formula, E j represents the jth intrinsic mode component The energy, E y represents the energy of the secondary traveling wave signal y(t), represents the jth intrinsic mode component The i-th sample data, represents the i-th sampling data of the secondary traveling wave signal y(t), and T is the total data length of the secondary traveling wave signal; S222: Normalize the energies calculated in S221: In the formula, E j,norm Energy E j The normalized value of E y,norm Energy E y The normalized value of The energy E corresponding to all natural mode components j and the energy E corresponding to the secondary traveling wave signal y(t) y The average value, std() represents the standard deviation calculation; S223: Calculate the normalized energy E of each natural modal component j,norm The normalized energy E between the energy of the secondary traveling wave signal y,norm Similarity: In the formula, r j represents the normalized energy similarity between the jth intrinsic mode component and the secondary traveling wave signal.
6. The traveling wave waveform detection method according to claim 1, characterized in that: The preset value used for comparing similarity in step S2 is calculated using the following formula: In the formula, is the mean of the normalized energy similarity between the secondary traveling wave signal and all IMF components, r j represents the normalized energy similarity between the jth intrinsic mode component and the secondary traveling wave signal.
7. The traveling wave waveform detection method according to claim 1, characterized in that: Step S3 uses the random Kaczmarz algorithm to invert each intrinsic modal component, specifically: S301: Based on the transfer function model of the traveling wave sensor, obtain the Toeplitz matrix A generated by the unit pulse response of the traveling wave sensor, and convert the lth effective natural mode component Let it be matrix y l , create matrices A and y l The relationship is: y l =Ax l In the formula, x l for y l The intrinsic mode components obtained by inversion; S302: Data initialization, specifying the maximum number of iterations K, assigning an initial value of the number of iterations k = 0, and assigning an initial value of the iterations x l (0) =0; S303: Update k=k+1; when the number of iterations k<K, in the kth iteration, according to the probability P i (k) Select the i-th row in matrix A: In the formula, represents the square of the Frobenius norm of matrix A; is the i-th row matrix of matrix A in the k-th iteration, For the matrix The transpose of S304: The intrinsic modal component x of the previous iteration l (k-1) Projection onto the hyperplane On the update solution vector, we get vector x l (k) : In the formula, Representative Matrix The square of the 2-norm of ; y l,i for y l The i-th row of S305: When the maximum number of iterations K is reached or the iteration convergence condition is met, the natural modal component x at the current number of iterations is output. l (k) As y l The corresponding inverted intrinsic modal component x l ; The iterative convergence condition is: In the formula, x l (k) for y l Invert the intrinsic modal components of the k-th iteration output; x l (k-1) for y l Invert the natural mode components of the k-1th iteration output; p is a given positive integer.
8. A traveling wave waveform detection device based on empirical mode decomposition and random Kaczmarz algorithm, characterized in that: include: The transfer function model building module is used to: build a transfer function model of the traveling wave sensor; The modal decomposition and extraction module is used to: perform collective empirical modal decomposition on the secondary traveling wave signal actually detected and output by the traveling wave sensor to obtain N inherent modal components; Then, the normalized energy similarity is used to extract the intrinsic modal components whose similarity is greater than a preset value and recorded as valid intrinsic modal components; The component inversion module is used to: perform inversion on each valid intrinsic mode component using the random Kaczmarz algorithm to obtain the corresponding intrinsic mode component of the primary traveling wave signal; The component synthesis module is used to superimpose all the intrinsic mode components obtained by the component inversion module to obtain an inverted primary traveling wave signal.
9. The traveling wave waveform detection device according to claim 8, characterized in that: The normalized energy similarity calculation method is as follows: S221: Calculate each natural mode component And the energy of the secondary traveling wave signal y(t): In the formula, E j represents the jth intrinsic mode component The energy, E y represents the energy of the secondary traveling wave signal y(t), represents the jth intrinsic mode component The i-th sample data, represents the i-th sampling data of the secondary traveling wave signal y(t), and T is the total data length of the secondary traveling wave signal; S222: Normalize the energies calculated in S221: In the formula, E j,norm Energy E j The normalized value of E y,norm Energy E y The normalized value of The energy E corresponding to all natural mode components j and the energy E corresponding to the secondary traveling wave signal y(t) y The average value, std() represents the standard deviation calculation; S223: Calculate the normalized energy E of each natural modal component j,norm The normalized energy E between the energy of the secondary traveling wave signal y,norm Similarity: In the formula, r j represents the normalized energy similarity between the jth intrinsic mode component and the secondary traveling wave signal.
10. The traveling wave waveform detection device according to claim 8, characterized in that: The component inversion module uses the random Kaczmarz algorithm to invert each intrinsic mode component, specifically: S301: Based on the transfer function model of the traveling wave sensor, obtain the Toeplitz matrix A generated by the unit pulse response of the traveling wave sensor, and convert the lth effective natural mode component Let it be matrix y l , create matrices A and y l The relationship is: y l =Ax l In the formula, x l for y l The intrinsic mode components obtained by inversion; S302: Data initialization, specifying the maximum number of iterations K, assigning an initial value of the number of iterations k = 0, and assigning an initial value of the iterations x l (0) =0; S303: Update k=k+1; when the number of iterations k<K, in the kth iteration, according to the probability P i (k) Select the i-th row in matrix A: In the formula, represents the square of the Frobenius norm of matrix A; is the i-th row matrix of matrix A in the k-th iteration, For the matrix The transpose of S304: The intrinsic modal component x of the previous iteration l (k-1) Projection onto the hyperplane Update the solution vector to get vector x l (k) : In the formula, Representative Matrix The square of the 2-norm of ; y l,i for y l The i-th row of S305: When the maximum number of iterations K is reached or the iteration convergence condition is met, the natural modal component x at the current number of iterations is output. l (k) As y l The corresponding inverted intrinsic modal component x l ; The iterative convergence condition is: In the formula, x l (k) for y l Invert the intrinsic modal components of the k-th iteration output; x r is the real inversion signal; p is a given positive integer.
