Complex broadband oscillation signal detection method based on multiple synchronous compression KST

Abstract

The invention discloses a complex broadband oscillation signal detection method based on multiple synchronous compression KST. A complex broadband oscillation signal containing an oscillation component, a harmonic component and a noise component is modeled; performing S transformation based on a Kaiser window on the obtained complex broadband oscillation signal model to obtain a time-frequency characteristic matrix, and performing multiple synchronous compression on the obtained time-frequency characteristic matrix; performing oscillation component extraction and reconstruction on the time-frequency characteristic matrix after multiple synchronous compression; and carrying out oscillation parameter detection on the reconstructed single oscillation component by using a Prony method. The method is suitable for accurate detection of broadband complex oscillation signals, and accurate detection can be carried out under the condition that oscillation, harmonic wave and noise interference are contained at the same time.

Description

Technical Field

[0001] This invention relates to the field of power system signal analysis and power quality detection technology, specifically to a method for detecting complex broadband oscillation signals based on multiple synchronous compression KST. Background Technology

[0002] With the dense integration of large-scale new energy generating units and power electronic equipment such as flexible DC transmission into the power grid, the dynamic behavior of the power system exhibits significant broadband, strong coupling, and high-dimensional nonlinear characteristics. Complex dynamic phenomena such as subsynchronous oscillations, supersynchronous oscillations, and even broadband coupled oscillations frequently occur in the system. These phenomena have a wide frequency range, strong time-varying characteristics, and rapid energy migration, posing a serious threat to the stable operation of the system and the safety of equipment. Traditional linear analysis methods based on steady-state models are insufficient to characterize the dynamic evolution mechanism of such broadband oscillations. Furthermore, during the detection of complex signals, mutual interference between the spectra of different components can occur, making it difficult to guarantee the reliability of the detection results.

[0003] In recent years, researchers have proposed improvement schemes to address the above problems. For example, synchronous compression transform is used to improve time-frequency convergence and simultaneously represent the time-frequency characteristics of the signal, enabling it to reflect the dynamic changes of the signal under test. However, it does not fundamentally solve the problems of spectral leakage and inter-spectral interference, and it still cannot accurately identify harmonic / interharmonic interference at frequencies close to the oscillation. Due to the lack of preprocessing of the sampled signal, existing methods cannot complete oscillation identification when noise interference and spectral interference coexist. The detection of complex broadband oscillation signals still requires further research. Therefore, there is an urgent need for a method for analyzing complex broadband oscillation signals that balances time-frequency resolution, anti-spectral aliasing, and noise immunity to meet the engineering requirements of new power systems that simultaneously contain oscillations, harmonics, and noise interference. Summary of the Invention

[0004] This invention provides a method for detecting complex broadband oscillation signals based on multiple synchronous compression KST, in order to solve the problem that existing technologies cannot achieve the analysis of complex broadband oscillation signals that simultaneously take into account time-frequency resolution, anti-spectral aliasing and anti-noise capabilities, and cannot meet the engineering requirements of new power systems that contain oscillations, harmonics and noise interference.

[0005] According to a first aspect, one embodiment provides a method for detecting complex broadband oscillation signals based on multiple synchronous compression KST, the method comprising: Modeling of complex broadband oscillating signals containing oscillatory components, harmonic components, and noise components; The obtained complex broadband oscillation signal model is subjected to Kaiser window-based S-transform to obtain the time-frequency characteristic matrix, and the obtained time-frequency characteristic matrix is ​​subjected to multiple synchronous compression. Oscillation components are extracted and reconstructed from the time-frequency characteristic matrix after multiple synchronization compression. The Prony method is used to detect the oscillation parameters of the reconstructed single oscillation component.

[0006] Furthermore, a model is developed for complex broadband oscillating signals containing oscillatory components, harmonic components, and noise components, specifically including: Complex wideband oscillation signal model A signal model that includes fundamental frequency, harmonic interference, noise interference, and high-frequency oscillations:

[0007] in, It is a wideband oscillation frequency. For harmonic frequencies, This represents noise interference, where t is a time parameter.

[0008] Furthermore, the obtained complex broadband oscillation signal model is subjected to an S-transform based on the Kaiser window to obtain the time-frequency characteristic matrix, specifically including: For complex wideband oscillation signals The Kaiser window-based S-transform is performed as follows:

[0009] Where t is the time parameter and f is the frequency parameter. Let R be the time-shift parameter, and R be the obtained time-frequency characteristic matrix; The Kaiser window function has the following specific expression:

[0010] Where w is the Kaiser window function and T is the data period. For the expansion of the zeroth-order Bessel function of the first kind, As a regulatory factor ,in , These are control parameters.

[0011] Furthermore, the obtained time-frequency characteristic matrix is ​​subjected to multiple synchronous compression, specifically including: First, the obtained time-frequency feature matrix is ​​subjected to a synchronous compression process:

[0012] in, It is the compressed frequency factor; Represents the Dirichlet function; Indicates the time-frequency matrix in The instantaneous frequency estimation at point A is expressed as follows:

[0013] in j Represents the imaginary unit. The obtained time-frequency characteristic matrix, For time shift parameters, This indicates taking the partial derivative; The formula for repeated synchronous compression m times, derived through calculation, is as follows:

[0014] in, This represents the time-frequency characteristic matrix after m synchronous compressions; The m-fold frequency estimate is expressed as follows:

[0015] in, Indicates the actual phase of the signal. and Let represent its first derivative and second derivative, respectively; t For time parameters, f This is a frequency parameter.

[0016] Furthermore, the oscillation components of the time-frequency characteristic matrix after multi-synchronization compression are extracted and reconstructed, specifically including: The oscillatory components of the time-frequency characteristic matrix after multiple synchronization compression are extracted using ridge features, as shown in the following formula:

[0017] in, It retrieves the coordinates of the maximum value; This represents a single frequency trajectory extracted from the time-frequency feature matrix; This is the trajectory with the kth oscillation maximum frequency, i.e., the ridge line; This refers to all frequency trajectories, where n takes values ​​from 0, 1, ..., N-1. This represents the time variable corresponding to the frequency ridge.

[0018] Furthermore, the oscillation components of the time-frequency characteristic matrix after multi-synchronization compression are extracted and reconstructed, specifically including: The time-domain signal is obtained by reconstructing the k-th oscillation separately:

[0019] in, The time-domain signal is obtained by reconstructing the k-th oscillation separately. ; This represents the key information of the time-frequency characteristic matrix obtained after m-time synchronous compression through ridge feature extraction; w(0) The value of the window function at 0.

[0020] Furthermore, the Prony method is used to detect the oscillation parameters of the reconstructed single oscillation component, specifically including: The Prony method is used to accurately detect the amplitude A, frequency f, and damping factor of the reconstructed single oscillation component. .

[0021] According to a second aspect, one embodiment provides a complex broadband oscillation signal detection system based on multiple synchronous compression KST, the system comprising: The signal modeling module is used to model complex broadband oscillating signals that contain oscillating components, harmonic components, and noise components. The KST module with multiple synchronous compression is used to perform Kaiser window-based S-transform on the obtained complex wideband oscillation signal model to obtain the time-frequency characteristic matrix, and then perform multiple synchronous compression on the obtained time-frequency characteristic matrix. The oscillation component extraction and reconstruction module is used to extract and reconstruct the oscillation components of the time-frequency characteristic matrix after multiple synchronization compression. The oscillation parameter detection module is used to detect the oscillation parameters of the reconstructed single oscillation component using the Prony method.

[0022] According to three aspects, one embodiment provides an electronic device, the device comprising: a processor and a memory; The memory is used to store one or more program instructions; The processor is configured to run one or more program instructions to perform the steps of a complex wideband oscillation signal detection method based on multiple synchronous compression KST as described in any of the preceding claims.

[0023] According to a fourth aspect, one embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a complex broadband oscillation signal detection method based on multiple synchronous compression KST as described in any of the preceding claims.

[0024] This invention provides a method for detecting complex broadband oscillation signals based on multiple synchronous compression KST, which integrates preprocessing techniques based on Kaiser window S-transform and synchronous compression transform, combined with ridge extraction, time-domain reconstruction and Prony parameter identification. The specific benefits are as follows: 1) This method first utilizes the excellent sidelobe suppression capability and flexible main lobe width adjustment characteristics of the Kaiser window to obtain the initial time-frequency representation under the S-transform framework, effectively suppressing spectral leakage and cross-interference caused by strong interference components such as harmonics; 2) Subsequently, synchronous compression transform is applied to redistribute energy in the KST results, significantly sharpening the time-frequency ridges and overcoming the shortcomings of energy diffusion in traditional time-frequency analysis. Thus, the time-frequency ridges of the target oscillation mode are clearly separated and accurately extracted in the background of high noise and harmonics. 3) Based on the extracted pure ridge line, the high-quality time-domain oscillation signal is reconstructed, which not only significantly reduces noise and harmonic interference, but also the reconstructed signal contains only the essential features of the target oscillation mode; 4) Finally, the Prony method is used to identify the parameters of the reconstructed pure oscillation signal. The inherent noise reduction effect of this method and the characteristics of the reconstructed signal together ensure that the order of the Prony model can be stably determined to be 2, thereby realizing the accurate detection of complex wideband oscillation signals. Attached Figure Description

[0025] Figure 1 A flowchart illustrating a method for detecting complex broadband oscillation signals based on multiple synchronous compression KST, as provided in one embodiment of the present invention; Figure 2 The high-frequency oscillation test results are provided in a method for detecting complex broadband oscillation signals based on multiple synchronous compression KST, according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the logic structure of a complex wideband oscillation signal detection system based on multiple synchronous compression KST, provided as an embodiment of the present invention. Detailed Implementation

[0026] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0027] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0028] The first embodiment of this invention provides a method for detecting complex broadband oscillation signals based on multi-synchronous compression KST. Addressing the limitation that existing multi-synchronous squeezing transform (MSST) has a fixed time-frequency resolution and cannot be flexibly adjusted according to the frequency band under test, this embodiment improves MSST by utilizing Kaiser window S-transform (KST), namely, multi-synchronous compression KST (MSSKST). The time-frequency matrix of the signal under test is obtained through MSSKST, and then ridge extraction is used to group and reconstruct each oscillation component. Finally, the Prony method is used to achieve high-precision detection of complex broadband oscillations containing multiple interferences. The following section combines... Figure 1 Please provide a detailed explanation.

[0029] like Figure 1 As shown, in step S100, a complex broadband oscillation signal containing oscillation components, harmonic components, and noise components is modeled.

[0030] The above steps specifically include: modeling a broadband noisy signal x(t) that contains oscillation components, harmonic components, and noise components.

[0031] Let the signal model be... For a signal model that includes fundamental frequency, harmonic interference / noise interference, and high-frequency oscillations:

[0032] in, It is a wideband oscillation frequency. For harmonic frequencies, This represents noise interference. The time parameter is discretized during testing. , where k is the number of samples and T is the sampling period.

[0033] In this embodiment, the signal model above is discretized at the sampling frequency. The discretization process preserves key frequency domain information, laying the foundation for subsequent high-precision time-frequency analysis.

[0034] like Figure 1 As shown, in step S200, the obtained complex wideband oscillation signal model is subjected to an S-transform based on the Kaiser window to obtain a time-frequency characteristic matrix, and the obtained time-frequency characteristic matrix is ​​subjected to multiple synchronous compression.

[0035] The above steps specifically include: S210, perform KST on the signal to be measured, using the following formula:

[0036] Where t is the time parameter and f is the frequency parameter. Let R be the time-shift parameter, and R be the time-frequency characteristic matrix obtained by KST. The specific expression for the Kaiser window function is as follows:

[0037] Where w is the Kaiser window function and T is the data period. For the expansion of the zeroth-order Bessel function of the first kind, As a regulatory factor ,in , The control parameters are set to 1600 and 10 respectively.

[0038] The window function length in the Kaiser Transform (KST) can automatically adjust according to the frequency band, making it more suitable for detecting wide-band oscillation signals. Through its frequency-adaptive window function, it achieves high frequency resolution at high frequencies and high time resolution at low frequencies, enabling it to reveal the local time-frequency characteristics of non-stationary signals containing a wide range of frequency components more clearly than the fixed-resolution Short-Time Fourier Transform. The Kaiser window function, due to its lower main lobe near the fundamental frequency, can mitigate the spectral leakage at that frequency; its better spectral discrimination at high frequencies prevents aliasing of oscillation and harmonic spectra, resulting in superior performance in wide-band oscillation detection.

[0039] This embodiment applies an S-transform based on a Kaiser window to the discrete sampled signal to generate a time-frequency matrix and a time-frequency characteristic matrix. The Kaiser window automatically adjusts its length according to the frequency band, and the high stopband near the fundamental frequency attenuates to suppress harmonic spectral leakage. It overcomes the limitations of the fixed window function of the traditional STFT, significantly reduces cross-term interference in a strong harmonic background, and simultaneously maintains transient oscillation capture capability and interference suppression performance.

[0040] S220 performs multiple synchronous compressions on the time-frequency characteristic matrix to improve energy concentration and frequency resolution.

[0041] First, a synchronous compression process is performed to obtain the time-frequency feature matrix from KST:

[0042] in, It is the compressed frequency factor; Represents the Dirichlet function, Representing the time-frequency matrix The instantaneous frequency estimation at point A is expressed as follows:

[0043] in j Represents the imaginary unit. The obtained time-frequency characteristic matrix, For time shift parameters, This indicates taking the partial derivative; The formula for repeated synchronous extrusion m times, derived through calculation, can be written as:

[0044] in, This represents the time-frequency characteristic matrix after m synchronous compressions. The m-fold frequency estimate is expressed as follows:

[0045] in, Indicates the actual phase of the signal. and Let represent its first derivative and second derivative, respectively; t For time parameters, f This is a frequency parameter.

[0046] The multiple synchronous compression transform, based on the synchronous squeezing transform, gradually condenses the fuzzy time-frequency energy through an iterative redistribution strategy, effectively overcoming the limitations of the Heisenberg uncertainty principle. Combined with the advantages of KST, it jointly solves the spectral leakage problem. It generates high-resolution time-frequency representations in strongly time-varying signals and noisy environments; simultaneously, it strictly preserves signal reconstruction capabilities, supports single-component mode decomposition, and requires no prior knowledge of the signal's frequency modulation rules or additional parameters, making it suitable for real-time processing in complex scenarios. The high-resolution time-frequency characteristics ensure subsequent reconstruction of individual oscillation components, thus determining the order of the Prony method, reducing the computational burden, and effectively addressing the shortcomings of the Prony method.

[0047] This embodiment performs a multiple synchronous squeezing operation on the obtained time-frequency feature matrix, iteratively redistributing energy along the frequency axis to the instantaneous frequency trajectory. This process compresses the energy distribution diffused on the time-frequency plane, sharpening the blurred time-frequency ridges into high-energy-density curves. MSST increases the energy concentration several times, eliminates the residual energy diffusion of KST, provides a super-resolution time-frequency characterization for the separation of weak oscillatory components, and has a significant effect on preventing spectral interference.

[0048] like Figure 1 As shown, in step S300, the oscillation components of the time-frequency characteristic matrix after multiple synchronization compression are extracted and reconstructed.

[0049] The above steps specifically include: The oscillation information of the time-frequency matrix after multiple synchronous compression is extracted using ridge features, as shown in the following formula:

[0050] in, It retrieves the coordinates of the maximum value; This represents a single frequency trajectory extracted from the time-frequency feature matrix; This is the trajectory with the kth oscillation maximum frequency, i.e., the ridge line; This refers to all frequency trajectories, where n takes values ​​from 0, 1, ..., N-1. This represents the time variable corresponding to the frequency ridge.

[0051] After obtaining the time-frequency characteristic matrix coordinates of a single oscillation, it is reconstructed to obtain the single oscillation signal in the time domain:

[0052] in, The time-domain signal is obtained by reconstructing the k-th oscillation separately. ; This represents the key information of the time-frequency characteristic matrix obtained after m-time synchronous compression through ridge feature extraction; w(0) The value of the window function at 0.

[0053] Amplitude thresholding is applied to the MSSKST time-frequency matrix of the signal for filtering to suppress weak components. By synchronously extracting the time-frequency features of the oscillation components, key information is focused and noise is suppressed. The effective components of the signal appear as continuous ridges with concentrated energy in the time-frequency domain, while noise energy is dispersed. By extracting and reconstructing these high-energy ridge regions, the principal components of the signal can be effectively separated from background noise, accurately obtaining the individual oscillation component under test. This method does not require a pre-defined noise model and is suitable for unknown and complex noise environments. By capturing the physical essence of the signal, it achieves strong noise-resistant signal identification while maintaining interpretability.

[0054] This embodiment extracts the time-frequency ridge of the target oscillation mode from the MSSKST time-frequency matrix and reconstructs a pure, single-oscillation time-domain signal through inverse transformation. The reconstructed signal completely removes harmonics and noise, retaining the essential characteristics of a single oscillation mode, thus simplifying the subsequent Prony analysis object into an ideal oscillation model.

[0055] like Figure 1 As shown, in step S400, the oscillation parameters of the reconstructed single oscillation component are detected using the Prony method.

[0056] The above steps specifically include: This embodiment uses the Prony method to identify the reconstructed individual oscillating time-domain signal. The specific principle of the Prony method is as follows: The reconstructed single oscillatory component is fitted using an exponential function:

[0057] in, For unknown coefficients, Let p be an unknown complex number, and let p be the order of the fitted model. This represents the fitted signal. Since an oscillation component was reconstructed separately in the previous processing, the order is fixed at 2, thus solving the problem of determining the order.

[0058] This exponential fit can be viewed as a solution to the difference equation, leading to the Prony method equation:

[0059] in The left-hand matrix r is an intermediate parameter, and the relational expression generated by the construction method is defined as follows, where x represents the actual sampled value. Indicate its conjugate value:

[0060] in To minimize the error between the fitted result and the true value, the following definition is made:

[0061] Solve this equation to obtain the intermediate parameters. and minimum error .

[0062] Then use polynomials Solving for Prony extremes .

[0063] Transform the exponential model into unknown parameters The system of linear equations:

[0064] in, This is the signal estimate. It is obtained using least squares. . Let be the Prony pole, and let the matrix on the left side of the equation be the Vandermonde matrix formed by .

[0065] Finally, based on the obtained parameters, the amplitude of the oscillation signal is obtained. ,frequency and attenuation factor :

[0066] in, The sampling time interval, This indicates how to obtain the imaginary part of a complex number. This indicates obtaining the real part of a complex number.

[0067] This embodiment locks the Prony order to 2 based on the single-mode characteristics of the reconstructed signal, completely avoiding the traditional Prony order estimation problem. Furthermore, the reconstructed signal has a low noise level, which is beneficial for the Prony method to accurately identify the various parameters of the oscillation.

[0068] This invention combines KST and MSST to further enhance the energy concentration of the time-frequency matrix, suppress spectral leakage, and improve frequency resolution. This allows signals containing multiple oscillation information to be reconstructed in separate groups, which reduces noise and interference from other components and helps the Prony method to accurately determine the model order, reducing the error risk and computational burden caused by order misjudgment.

[0069] To verify the detection performance of this invention, the sampling frequency was set to 10.2 kHz, the sampling window length was set to 3 power frequency cycles (corresponding to a 50 Hz fundamental frequency, with a window length of 60 ms), and the total signal duration was 1 s. The test signal was designed as follows:

[0070] Harmonics are introduced into the fundamental frequency signal and oscillation signal to increase the density of instances and introduce noise interference. To compare the accuracy of different methods under different noise levels, the added noise is used... The signal-to-noise ratio (SNR) was varied from 30 dB to 80 dB in 5 dB steps. The comparison methods were the Multitone Filter (MTFIR) algorithm and the basic Prony method. MSSKSTP is used to abbreviate the method proposed in this embodiment. Mag.error, Freq.error, and Damp.error represent the amplitude error, frequency error, and damping factor error in the simulation test, respectively. To prevent errors caused by fluctuations, fifteen points were tested at each SNR, and their box plots were plotted. The final test results are as follows: Figure 2 As shown, the method proposed in this embodiment can accurately detect the oscillation component in the range of 30~80dB, and its detection accuracy is higher than that of the two comparative algorithms.

[0071] Corresponding to the aforementioned method for detecting complex broadband oscillation signals based on multiple synchronous compressed KST, this invention also discloses a system for detecting complex broadband oscillation signals based on multiple synchronous compressed KST, such as... Figure 3 As shown, it specifically includes: The signal modeling module is used to model complex broadband oscillating signals that contain oscillating components, harmonic components, and noise components. The KST module with multiple synchronous compression is used to perform Kaiser window-based S-transform on the obtained complex wideband oscillation signal model to obtain the time-frequency characteristic matrix, and then perform multiple synchronous compression on the obtained time-frequency characteristic matrix. The oscillation component extraction and reconstruction module is used to extract and reconstruct the oscillation components of the time-frequency characteristic matrix after multiple synchronization compression. The oscillation parameter detection module is used to detect the oscillation parameters of the reconstructed single oscillation component using the Prony method.

[0072] It should be noted that for a detailed description of the complex broadband oscillation signal detection system based on multiple synchronous compression KST provided in the embodiments of the present invention, please refer to the relevant description of the complex broadband oscillation signal detection method based on multiple synchronous compression KST provided in the embodiments of the present invention, which will not be repeated here.

[0073] In addition, embodiments of the present invention also provide an electronic device, the device comprising: a processor and a memory; the memory being used to store one or more program instructions; the processor being used to execute one or more program instructions to perform the steps of a complex wideband oscillation signal detection method based on multiple synchronous compression KST as described in any of the preceding embodiments.

[0074] It should be noted that for a detailed description of an electronic device provided in the embodiments of the present invention, please refer to the relevant description of a complex wideband oscillation signal detection method based on multiple synchronous compression KST provided in the embodiments of this application, which will not be repeated here.

[0075] In addition, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a complex broadband oscillation signal detection method based on multiple synchronous compression KST as described in any of the preceding claims.

[0076] It should be noted that for a detailed description of the computer-readable storage medium provided in the embodiments of the present invention, please refer to the relevant description of the complex wideband oscillation signal detection method based on multiple synchronous compression KST provided in the embodiments of this application, which will not be repeated here.

[0077] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. A method for detecting complex broadband oscillation signals based on multiple synchronous compression KST, characterized in that, The method includes: Modeling of complex broadband oscillating signals containing oscillatory components, harmonic components, and noise components; The obtained complex broadband oscillation signal model is subjected to Kaiser window-based S-transform to obtain the time-frequency characteristic matrix, and the obtained time-frequency characteristic matrix is ​​subjected to multiple synchronous compression. Oscillation components are extracted and reconstructed from the time-frequency characteristic matrix after multiple synchronization compression. The Prony method is used to detect the oscillation parameters of the reconstructed single oscillation component.

2. The method for detecting complex broadband oscillation signals based on multiple synchronous compression KST as described in claim 1, characterized in that, Modeling complex broadband oscillating signals containing oscillatory components, harmonic components, and noise components specifically includes: Complex wideband oscillation signal model A signal model that includes fundamental frequency, harmonic interference, noise interference, and high-frequency oscillations: ； in, It is a wideband oscillation frequency. For harmonic frequencies, This represents noise interference, where t is a time parameter.

3. The method for detecting complex broadband oscillation signals based on multiple synchronous compression KST as described in claim 1, characterized in that, The obtained complex broadband oscillation signal model is subjected to an S-transform based on the Kaiser window to obtain the time-frequency characteristic matrix, which specifically includes: For complex wideband oscillation signals The Kaiser window-based S-transform is performed as follows: ； Where t is the time parameter and f is the frequency parameter. Let R be the time-shift parameter, and R be the obtained time-frequency characteristic matrix; The Kaiser window function has the following specific expression: ； Where w is the Kaiser window function and T is the data period. For the expansion of the zeroth-order Bessel function of the first kind, For the regulating factor: ,in , These are control parameters.

4. The method for detecting complex broadband oscillation signals based on multiple synchronous compression KST as described in claim 3, characterized in that, The obtained time-frequency characteristic matrix is ​​subjected to multiple synchronous compression, specifically including: First, the obtained time-frequency feature matrix is ​​subjected to a synchronous compression process: ； in, It is the compressed frequency factor; Represents the Dirichlet function; Indicates the time-frequency matrix in The instantaneous frequency estimation at point A is expressed as follows: ； in j Represents the imaginary unit. The obtained time-frequency characteristic matrix, For time shift parameters, This indicates taking the partial derivative; The formula for repeated synchronous compression m times, derived through calculation, is as follows: ； in, This represents the time-frequency characteristic matrix after m synchronous compressions; The m-fold frequency estimate is expressed as follows: ； in, Indicates the actual phase of the signal. and Let represent its first derivative and second derivative, respectively; t For time parameters, f This is a frequency parameter.

5. The method for detecting complex broadband oscillation signals based on multiple synchronous compression KST as described in claim 4, characterized in that, The oscillatory components of the time-frequency characteristic matrix after multi-synchronization compression are extracted and reconstructed, specifically including: The oscillatory components of the time-frequency characteristic matrix after multiple synchronization compression are extracted using ridge features, as shown in the following formula: ； in, It retrieves the coordinates of the maximum value; This represents a single frequency trajectory extracted from the time-frequency feature matrix; This is the trajectory with the kth oscillation maximum frequency, i.e., the ridge line; This refers to all frequency trajectories, where n takes values ​​from 0, 1, ..., N-1. This represents the time variable corresponding to the frequency ridge.

6. The method for detecting complex broadband oscillation signals based on multiple synchronous compression KST as described in claim 5, characterized in that, The oscillatory components of the time-frequency characteristic matrix after multi-synchronization compression are extracted and reconstructed, specifically including: The time-domain signal is obtained by reconstructing the k-th oscillation separately: ； in, The time-domain signal is obtained by reconstructing the k-th oscillation separately. ; This represents the key information of the time-frequency characteristic matrix obtained after m-time synchronous compression through ridge feature extraction; w(0) The value of the window function at 0.

7. The method for detecting complex broadband oscillation signals based on multiple synchronous compression KST as described in claim 1, characterized in that, The Prony method is used to detect the oscillation parameters of the reconstructed single oscillation component, specifically including: The Prony method is used to accurately detect the amplitude A, frequency f, and damping factor of the reconstructed single oscillation component. .

8. A complex broadband oscillation signal detection system based on multiple synchronous compression KST, characterized in that, The system includes: The signal modeling module is used to model complex broadband oscillating signals that contain oscillating components, harmonic components, and noise components. The KST module with multiple synchronous compression is used to perform Kaiser window-based S-transform on the obtained complex wideband oscillation signal model to obtain the time-frequency characteristic matrix, and then perform multiple synchronous compression on the obtained time-frequency characteristic matrix. The oscillation component extraction and reconstruction module is used to extract and reconstruct the oscillation components of the time-frequency characteristic matrix after multiple synchronization compression. The oscillation parameter detection module is used to detect the oscillation parameters of the reconstructed single oscillation component using the Prony method.

9. An electronic device, characterized in that, The device includes: a processor and a memory; The memory is used to store one or more program instructions; The processor is configured to run one or more program instructions to perform the steps of a complex wideband oscillation signal detection method based on multiple synchronous compression KST as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of a complex broadband oscillation signal detection method based on multiple synchronous compression KST as described in any one of claims 1 to 7.

Images