62 results about "Spectral leakage" patented technology

The invention relates to the technical field of high-frequency signal modulation, and discloses a high-frequency carrier synchronization signal modulation system. A carrier distortion compensation module of the system constructs a model based on a historical transmission data set, captures a phase jitter parameter, a spectrum leakage component and a modulation pulse sequence of a transmitting end carrier in real time, and outputs a reconstructed baseband parameter; the multi-dimensional distortion analysis module compares the reconstructed baseband parameter with a receiving end demodulation baseband parameter through composite difference detection, and generates a channel-level distortion coefficient tensor; the topological positioning network module is used for positioning signal out-of-step physical nodes and generating a probability distribution thermodynamic diagram by combining impedance characteristics of transmission nodes and signal path delay information; and the adaptive modulation strategy module starts a multi-band carrier injection mode for the high-probability out-of-step nodes according to the abnormal probability gradient of the thermodynamic diagram, and applies a phase disturbance test to adjacent transmission links. The system can improve the adaptability and accuracy of high-frequency carrier signal synchronous modulation.

The invention discloses a new energy broadband test analysis management and control method and system, and relates to the technical field of new energy detection and intelligent management and control, and the method comprises the following steps: obtaining a multi-window splicing frequency spectrum of a load switching edge, superposing phase lock consistency verification in a splicing frequency spectrum result, generating a frequency spectrum leakage suspicion chart, and carrying out the boundary constraint for energy track extraction; under the constraint of a spectrum leakage suspicion chart, an energy center track, a group delay curve and a phase derivative sequence are extracted, and a cross-frequency migration fingerprint is constructed and is used as a basic evidence for misjudgment traceability. By means of spectrum suspicion identification, feature fusion modeling, misjudgment traceability, label correction, energy reverse intervention and the like, accurate identification and closed-loop regulation and control of false operation caused by spectrum leakage are achieved, high identification accuracy, low misjudgment rate, high anti-interference performance and quick response capacity are achieved, and the operation stability and safety of a new energy field station are remarkably improved.

The invention discloses a structural damping ratio identification method and device considering frequency spectrum leakage error compensation. The method comprises the following steps: firstly, obtaining a cut-off free attenuation signal under impact vibration, and carrying out Fourier transform on an autocorrelation function and a cross-correlation function of the signal to obtain a response power spectral density matrix influenced by a spectrum leakage error; singular value decomposition is carried out on the response power spectrum density function through a frequency domain decomposition method to obtain a damping ratio and an inherent frequency, a spectrum leakage error is calculated through the calculated inherent frequency and the damping ratio, and the error obtained through calculation is substituted into an output power spectrum. And the inherent frequency and the damping ratio are identified by using the power spectrum after the frequency spectrum leakage error is compensated, and the above process is iterated for multiple times, so that the frequency spectrum leakage error in damping ratio identification can be compensated. When the free attenuation data is used for identifying the damping ratio, spectrum leakage errors caused by data truncation can be made up, the signal-to-noise ratio of the data can be improved, the influence of noise can be reduced, and the identification precision of the damping ratio is improved.

The invention relates to a multi-parameter estimation algorithm combining a disturbance quantification model and a rooting method in interference measurement. In order to solve the problems that in the prior art, the disturbance degree of a sequence is difficult to quantify, and harmonic key parameters are influenced by spectrum leakage, the invention provides the following innovative scheme: firstly, a sequence multi-classification measurement entropy quantification model is constructed, and quantitative evaluation of the signal noise degree is realized; secondly, through a Hankel matrix construction and rooting parameter estimation method, realizing accurate estimation of the harmonic frequency of the real-value signal; meanwhile, an amplitude correction algorithm based on window function characteristics is designed, and the signal sampling condition is effectively judged; and finally, through matrix weighting operation of a phase demodulation function, efficient reconstruction of an optical surface shape is realized, the problems of spectrum leakage and disturbance quantification in a traditional method are solved, and the method has relatively high measurement precision.

A broadband signal spectrum analysis method based on spectrum leakage suppression and an improved FFT algorithm comprises the following steps: setting a broadband signal model comprising a fundamental wave frequency offset signal and a plurality of broadband oscillation signals according to broadband signal characteristics; based on a Fourier transform theory, analyzing a spectrum leakage generation mechanism of the broadband signal; a Nuttall-Kaiser combination window function is designed, and is used for suppressing the spectrum leakage of the broadband signal; the FFT algorithm is improved, and side lobe attenuation is suppressed by using the convolution window property so as to reduce the influence of spectrum leakage. According to the broadband signal spectrum analysis method based on spectrum leakage suppression and the improved FFT algorithm, spectrum leakage is effectively reduced while high frequency resolution of broadband signals is guaranteed, and meanwhile the method has the advantage of being high in measurement accuracy.

The invention discloses an anisotropic noise generation method based on local tensor integral field feature modulation, and is applied to the technical field of computer texture generation. The method comprises the following steps: step 1, constructing a global tensor field of spatial anisotropy intensity and direction; step 2, calculating local tensor integral features in a support domain based on the support domain of a Gabor noise kernel function, the local tensor integral features including principal direction features and anisotropic ratio features; step 3, calculating a curvature penalty factor based on the deviation between the local tensor integral feature and the center tensor of the seed point; step 4, based on the local tensor integral features, performing weighted correction on the frequency domain center frequency and Gaussian envelope parameters of the Gabor kernel function; and step 5, generating a final noise texture based on the corrected kernel function. The Gabor kernel function parameter is subjected to weighted correction through the local tensor integral feature and the curvature penalty factor, modulation of the kernel function direction under the high-curvature tensor field is achieved, spectrum leakage, texture fracture and artifacts caused by inconsistent direction changes in the kernel function can be reduced, and the robustness of the kernel function is improved. Therefore, the topological continuity and the simulation fidelity of the anisotropic noise texture at the complex structure are improved.

The invention discloses a harmonic analysis method based on double buffer window function optimization, and relates to the technical field of harmonic analysis and signal processing of a power system, and the method comprises the steps of hardware initialization, double buffer initialization, FFT module parameter configuration, data acquisition, buffer switching, data extraction and windowing and FFT processing. Parallel data acquisition and processing are realized through a double-buffer ping-pong mechanism, spectrum leakage is suppressed by applying a Hanning window function, and FFT calculation is performed after signal amplitude is recovered through a compensation factor; according to the method, the data acquisition continuity and real-time performance are improved, the harmonic amplitude error is controlled within + / -0.5%, the hardware resource utilization rate is optimized, and the method is suitable for power grid electric energy quality monitoring.

The invention provides an anti-noise interference frequency parameter estimation method for high-precision optical coherence ranging, which comprises the following steps: acquiring a spectral interference signal of an optical low-coherence interference system and a reference light spectral signal output only by a reference arm, and converting the spectral interference signal into interference signals with wave number space arranged at equal intervals, carrying out spectral density amplitude normalization to obtain a standard interference signal so as to eliminate interference of light source spectrum fluctuation and non-uniform sampling on frequency estimation; hanning window modulation is applied to a standard interference signal, spectrum leakage is suppressed, spectrum analysis is performed on the modulated signal, a peak value and an adjacent secondary peak value of a spectrum main lobe are extracted, a frequency correction amount is calculated based on a peak value and secondary peak value amplitude ratio relation derived by a Hanning window spectrum mode function, and an accurate frequency of the interference signal is obtained; and calculating the displacement of the object to be measured according to the product of the accurate frequency and the spectral resolution of the optical low-coherence interference system in combination with the refractive index of the light in the measurement medium.

The invention discloses a generative adversarial network image enhancement method combined with frequency domain information, relates to the technical field of computer vision, and solves the problems of spectrum leakage, ringing artifacts, splicing strips and structural dislocation easily caused by the existing frequency domain image enhancement method. The specific scheme is as follows: receiving a degraded image and extracting multi-scale features and boundary section features; the method comprises the following steps of: performing overlapping partitioning and periodic continuation expansion on the basis of features, performing spectral analysis on an expansion block to generate a parameterized window function and performing windowing, then obtaining a preliminary reconstructed image through complex field correction and overlapping perception mixed inverse transformation, and finally performing optimization reconstruction under a generative adversarial framework through a multi-scale wavelet decomposition and sub-band fusion decoder to obtain a reconstructed image. According to the method, the consistency of space geometric precision and texture details can be kept while artifacts and seams are inhibited, and the applicability and stability of image enhancement in a complex scene are improved.

The invention discloses a modal modeling method for a tendon-driven continuum robot. The method comprises the following steps: carrying out differential modeling on a configuration under a SE (3) Lie group-se (3) Lie algebra framework; the strain field is subjected to spectrum parameterization and is decomposed into active / passive orthogonal subspaces, and an inner product adopts energy weighted Hilbert measurement containing material and section parameters. The active mode is constructed by a linear continuous Jacobian of a tendon path; legendre polynomial expansion is adopted under the collinear layout, and complex Fourier harmonic expansion is adopted under the non-collinear layout; the passive mode is selected according to a spectrum slot position mutual exclusion principle so as to ensure that the passive mode is orthogonal to the active mode in energy and suppress spectrum leakage. Lie group integration and spectral integration are adopted for numerical value realization; newton-Rafson iteration is used in the quasi-static state, and generalized-alpha time integration or Newmark-beta integration is used in the dynamics. The method has high precision and robustness under nonlinear large deformation and multi-tendon coupling conditions, and is suitable for scenes such as software operation, minimally invasive intervention and precise detection.

The application provides a baseband radio frequency signal data feedback processing method based on vector modulation, which first receives a baseband signal data set transmitted by a target radio frequency link in a preset frequency band, which contains multiple signal frame original modulation parameters and corresponding radio frequency feedback signal waveforms, then calculates feedback error indicators composed of phase offset, amplitude distortion and spectral leakage factor according to the radio frequency feedback signal waveforms, quantizes the modulation error, analyzes the feedback signal waveforms in the time-frequency domain, extracts error feature vectors, inputs the pre-trained model for cross-frame correlation learning, generates a dynamic adjustment parameter sequence containing time sequence dependence, and finally constructs a closed-loop feedback control link based on the dynamic adjustment parameter sequence and the original modulation parameters, so as to realize accurate processing and optimized control of the baseband radio frequency signal and effectively improve the signal transmission performance.

The invention discloses a self-adaptive overlapped framing and variable window phase measurement method, device and system based on signal state perception, and belongs to the technical field of digital signal processing and phase measurement. Aiming at the technical problems that a fixed overlapping ratio and a single window function are adopted in an existing phase measurement method, resolution ratios of a time domain and a frequency domain cannot be considered under a time-varying signal, and frequency spectrum leakage and phase jump are easily introduced, the invention provides a joint self-adaptive preprocessing method. The method comprises the following steps: firstly, framing collected signals, and judging whether the signal state is a steady state or an unsteady state according to smoothness detection of absolute value sums of adjacent sampling points; according to the determined signal state, adaptively configuring an overlapping rate between adjacent data frames and window function type parameters, and then, before fast Fourier transform is carried out, introducing a window function coherent gain compensation mechanism to eliminate an amplitude reference error caused by window type switching; and finally calculating the phase difference based on the corrected spectrum result. According to the method, the overlapping framing and window function parameters are adaptively adjusted according to the signal state, the phase measurement continuity is ensured, the phase measurement precision and stability under the multi-frequency signal are improved, and the method has good real-time performance and engineering application value.

The embodiment of the present application discloses a spectrum analysis method, device and signal processing device, and the embodiment of the present application introduces a convolutional neural network to process discrete real part signals and discrete imaginary part signals in a frequency domain on the basis of an existing cycle graph spectrum analysis method, in the convolutional neural network, a convolution operation is performed by using frequency domain correlation characteristics of a window function used in a time domain, so that the problem of frequency domain distortion such as spectrum leakage, fence loss and the like caused by time domain processing of the window function in the frequency domain can be eliminated to a certain extent, and meanwhile, the accuracy of spectrum analysis can be effectively improved by combining the trained convolutional neural network.

The invention discloses a method for quickly estimating the reflection intensity of a defect of a buffer layer of a long-distance high-voltage cable. The method is mainly suitable for quickly and accurately estimating the reflection intensity of an ablation defect of the buffer layer of the long-distance high-voltage cable. The method comprises the following specific steps: (1) respectively testing the head end and the tail end of a defective cable to obtain head end reflection coefficients of the two ends; (2) obtaining frequency spectrum amplitude and pseudo frequency information at the corrected defect; (3) solving to obtain a defect position and a signal propagation attenuation constant; and (4) solving to obtain the reflection intensity of the defect of the buffer layer of the high-voltage cable at any position. The method is simple and rapid to operate and high in calculation accuracy, eliminates the influence of spectrum leakage, the fringe effect and the defect position on the calculation of the reflection intensity to a certain extent, can rapidly and accurately estimate the reflection intensity of the ablation defect of the buffer layer of the high-voltage cable at any position, and improves the calculation accuracy. The method has important reference value for timely mastering the severity of the defects of the buffer layer of the high-voltage cable and determining the priority level of defect processing.

The present invention relates to a method for recovering signal spectral peak information in the case of spectral leakage. Existing methods have defects such as low accuracy in recovering spectral peak information or requiring a long sampling time to improve the recovery accuracy when dealing with spectral leakage problems. The present invention first performs downmixing and then filtering and decimation on the sampled intermediate-frequency sequence; then adjusts the decimated signal to obtain two sequences, and records the spectral line numbers corresponding to the extreme points of the spectral modulus of one of the sequences; then starting from the first spectral line number, calculates the relevant phase difference and performs iterative processing; when the processing result meets the threshold, jumps out of the iteration, completes the recovery of the frequency and amplitude information of the spectral peak corresponding to this number, and repeats the above process for the next spectral line number until all spectral line numbers are processed. The present invention can achieve the purpose of accurately positioning the true frequency of the signal and improving the accuracy of the spectral peak amplitude, and is more suitable for short-time signals.

The invention belongs to a signal direction finding method, and provides a signal direction finding method based on a space-time modulation metasurface and a related device in order to solve the technical problems that in the prior art, a traditional direction finding system has frequency spectrum leakage and poor direction finding precision, and related research for achieving direction-of-arrival estimation by modulating transmission amplitude lacks. When the space-time modulation metasurface receives a signal transmitted by a signal source, square waves of modulation array elements in the sub-arrays are estimated through a space-time modulation model, the space-time modulation model represents each sub-array signal in combination with a modulation frequency and a channel amplitude phase of each sub-array, and each sub-array signal is obtained; and estimating the square wave of the modulation array element in the sub-array through the relative phase difference of the + 1 harmonic and the + 2 harmonic of the represented sub-array signal. Then, according to square waves of modulation array elements in the sub-arrays, sub-array signals are recovered, and spatial frequency spectrum estimation is carried out on the sub-array signals to obtain the direction of arrival. And a non-uniform period modulation scheme is adopted to improve the estimation precision of the direction of arrival.

The present application relates to the technical field of signal processing, in particular to a kind of ECG signal noise suppression method and system based on singular spectrum analysis mean strategy, method includes: obtaining ECG noise signal sequence, ECG noise signal sequence is converted to frequency domain and obtains ECG signal spectrum;Frequency response matrix is constructed based on ECG noise signal spectrum, target function is established based on frequency response matrix with the orientation of inhibiting spectral leakage, the target vector is obtained by solving target function, the optimal window is determined according to target vector;Multiple noise suppression signals are generated based on ECG noise signal sequence and optimal window, noise suppression ECG signal is generated based on multiple noise suppression signals;The present application can suppress noise low-frequency component, to improve noise suppression effect.

The invention discloses a subspace joint parameter estimation algorithm for constructing an analytic signal based on Hilbert transform in high-precision multi-surface phase-shift interference, belongs to the technical field of high-precision optical interference measurement, and aims to solve the problems of inaccurate frequency estimation caused by spectrum leakage and a fence effect in a traditional algorithm. And the problem that a modern spectrum estimation algorithm cannot directly process real-value light intensity signals is solved. The method comprises the following steps: firstly, performing Hilbert transform on a real-value light intensity sequence to construct an analysis signal; then, a Hankel matrix is constructed based on the analysis signal, and subspace decomposition is carried out; further, taking the obtained high-precision frequency as prior information, and performing joint correction on the amplitude and the phase by using a double-section phase difference correction method; and finally, reconstructing the surface topography according to the corrected phase. According to the invention, the limitation of spectrum leakage and the like is overcome, the joint fine positioning of frequency, amplitude and phase is realized, and the precision, robustness and measurement repeatability of interference parameter demodulation are obviously improved.

The present application discloses a fault recording triggering method, device, electronic device and storage medium, which belong to the field of fault recording control technology. The method collects and analyzes cached recording signals with a data window, constructs a standard signal and compares it with the original waveform signal, quantifies the waveform energy distortion characteristics through the waveform amplitude deviation area, effectively avoids defects such as spectrum leakage and parameter sensitivity, and realizes efficient capture of short-term transient faults; uses cached historical recording data to establish recording trigger conditions, and can further dynamically adjust the trigger conditions as signal acquisition proceeds, significantly reducing the risk of false triggering caused by environmental noise or instantaneous disturbances.

The application discloses a kind of long-distance high-voltage cable buffer layer defect reflection intensity fast estimation method, this method is mainly suitable for the quick and accurate estimation of long-distance high-voltage cable buffer layer ablation defect reflection intensity.The specific steps of the method are as follows: (1) test at the defect cable first end and the end, obtain the first end reflection coefficient of two ends;(2) obtain the frequency spectrum amplitude and pseudo frequency information after correction at the defect;(3) solve to obtain the defect position and signal propagation attenuation constant;(4) solve to obtain the reflection intensity of high-voltage cable buffer layer defect at any position.The method is simple, fast, and the calculation accuracy is high, to a certain extent, eliminates the influence of spectrum leakage, aperture effect and defect position on the calculation of reflection intensity, can quickly and accurately estimate the reflection intensity of high-voltage cable buffer layer ablation defect at any position, has important reference value for timely grasping the severity of high-voltage cable buffer layer defect and determining the priority of defect treatment.

The present invention discloses a method for measuring power frequency phasors in urban substations, comprising the following steps: S1, establishing a mathematical model of components to be analyzed in the power supply system of the urban substation, and inputting the signal components to be measured into the mathematical model; S2, fitting the power frequency phasor parameters using linear Bessel interpolation to obtain the fundamental frequency, amplitude, and phase; S3, using the fitting result in S2, compensating for a negative spectral image through an iterative process to reduce spectral leakage of the negative image; S4, constructing an energy threshold inequality, and judging whether the ratio of the energy of the interharmonic component to the signal component to be measured meets the threshold based on the result of S3, and directly outputting the result if it meets the threshold; if not, repeating S2-S4 until the threshold is met.

This invention relates to the field of exosome quantitative analysis technology, and discloses a multi-parameter joint quantitative analysis method and apparatus for exosomes based on the fluorescence intensity of encoded microspheres. The method includes: performing three-dimensional spatial attitude inversion to obtain a three-dimensional attitude characterization vector and constructing a feature point diffusion function; performing spatial deconvolution processing using the feature point diffusion function, and combining it with a visibility weight calculated from the longitudinal depth deviation for brightness gain compensation to obtain intrinsic feature intensity; constructing a channel intermodulation matrix to perform signal decoupling operations, obtaining decoupling intensity and mapping it to quantified abundance. This invention solves the problems of projection distortion, defocus blurring, and spectral leakage caused by random attitude, eliminates steric hindrance effects and system nonlinear interference, significantly improves the accuracy and repeatability of exosome detection, and achieves accurate analysis of high-fidelity biological quantitative information.

The application discloses a Ku-band satellite chain downlink signal interference suppression method, system, device and medium, and relates to the technical field of wireless communication signal processing. First, an OFDM signal from a satellite chain system is received, the OFDM signal is segmented, and fast Fourier transform is performed on the segmented signal to convert the segmented signal into a frequency domain to obtain a frequency domain signal. Then, an FCME algorithm is used to preliminarily locate a narrow-band interference frequency of the frequency domain signal to obtain an interference range. Further, a frequency offset value is calculated by using an interference frequency determined by DTFT, and then, spectrum shifting is performed on the frequency domain signal according to the frequency offset value, and interference alignment is completed through the spectrum shifting, so that interference suppression of the Ku-band satellite chain downlink signal is realized while spectrum leakage is avoided.

The present application discloses a method, device and medium for optimizing laser vibration measurement signals, belonging to the technical field of laser vibration measurement signal optimization. The method includes: collecting vibration signals on the tunnel lining surface; performing spectrum leakage suppression processing on the vibration signals based on the third-order five-term maximum sidelobe attenuation self-convolution window to obtain suppressed signals; performing variational mode decomposition and noise reduction on the vibration signals through an improved chicken swarm optimization algorithm to obtain noise-reduced signals; inputting the suppressed signals and the noise-reduced signals into a Transformer model to obtain optimized vibration signals; extracting time-frequency domain feature parameters of the optimized vibration signals based on the short-time Fourier transform and wavelet packet decomposition algorithms to obtain time-frequency feature parameters; processing the time-frequency feature parameters based on an adaptive filtering algorithm to obtain optimized data; and performing real-time correction on the optimized parameters through a compensation calibration algorithm to obtain calibrated vibration signals. The present application achieves the technical effect of optimizing laser vibration measurement signals through the above method.

The application discloses an FID signal frequency estimation method and system based on improved interpolation DFT, and relates to the technical field of signal processing.The application adds a maximum sidelobe attenuation window to an input signal, derives an initial frequency estimation value under the condition of considering negative spectrum components, and significantly reduces the interference of negative spectrum leakage on frequency estimation on the basis of the initial frequency estimation value; on the basis, the application is no longer limited to discrete integer frequency points, but uses a discrete time Fourier transform (DTFT) to perform spectrum interpolation, thereby effectively alleviating the inherent fence effect of a fast Fourier transform (FFT), and making the frequency estimation no longer limited to the discrete interval of spectrum lines. A high-precision frequency estimation model further constructed can maintain consistent estimation performance in a full frequency band, and overcomes the defects of the existing IpDFT method that the precision fluctuates due to frequency changes. Therefore, higher estimation precision, smaller negative spectrum residual influence, and stronger robustness in a full frequency band can be realized in a complex electromagnetic environment.

The invention relates to a method, a medium and a device for calculating an anisotropic seabed interface two-dimensional roughness spectrum, and belongs to the technical field of underwater sound. The method comprises the following steps: firstly, carrying out interpolation and gridding on a measured seabed micro-landform three-dimensional point cloud map to obtain a seabed micro-landform three-dimensional digital elevation map; then translation and whitening processing are carried out, and vertical large-scale offset caused by submarine topography and laying of a measurement system is eliminated; a two-dimensional window function is added to the whitened seabed interface roughness, spectrum leakage is reduced, and energy correction of the window function is carried out; estimating the power spectrum density by adopting a sliding window periodogram method to obtain a seabed interface two-dimensional roughness spectrum; and finally, performing linear fitting on the two-dimensional roughness spectrum scatter diagram under a logarithmic coordinate system, and calculating the spectrum index and the spectrum intensity of the two-dimensional roughness spectrum. The invention also provides a medium and a device for operating the method. According to the method, the three-dimensional statistical characteristics of the anisotropic seabed interface fluctuation can be truly reflected, and accurate model input parameters are provided for seabed scattering intensity forecast.

The invention discloses a broadband impedance spectroscopy cable defect positioning method and system based on improved Z transformation, and the method comprises the steps: enabling the input impedance of each section of a cable to serve as the load impedance of a previous section according to an equivalent distribution parameter circuit of a defective cable, and carrying out the forward deduction from the tail end of the cable section by section, and finally obtaining the head end input broadband impedance spectrum of the defective cable. And carrying out improved Z transformation considering spectrum leakage suppression on the head-end input broadband impedance spectrum of the defective cable, converting the head-end input broadband impedance spectrum of the defective cable from a frequency domain to a time domain, and positioning the cable defect through a wave crest in the time domain. According to the method, spectrum leakage suppression processing is carried out on a traditional Z transformation method in broadband impedance spectrum analysis, and the accuracy of cable defect positioning can be remarkably improved.

The invention discloses a resistive leakage current detection protection method of an intelligent circuit breaker, and belongs to the technical field of electric leakage detection. The method comprises the following steps: synchronous sampling; shaping a window function; performing high-resolution spectrum analysis; estimating a Kalman state; generating a fault evolution index; inverse time delay decision making; and tripping execution. According to the method, a dynamic fault evolution index is taken as a core, online joint estimation of the resistive leakage current amplitude and the change rate of the resistive leakage current amplitude is realized in a circuit breaker for the first time through high-resolution spectral analysis, Kalman state estimation and a multi-dimensional nonlinear inverse time limit curved surface, and protection delay is generated in a self-adaptive manner according to the online joint estimation. Meanwhile, milliampere-level sensitivity and working condition self-adaption reliability are met, and the problems of FFT spectrum leakage, fixed threshold misjudgment and non-self-adaption in time delay in the prior art are solved.

This invention discloses a fault identification method and system for wheel-rail conveyors based on deep learning. By introducing a dynamic harmonic sensing sinc convolutional layer based on physical priors, this invention can actively extract discrete harmonic components that match the characteristics of fault harmonics, effectively suppressing strong background noise. Secondly, by utilizing an offset field predictor and deformable convolution, it accurately models and corrects Doppler nonlinear frequency shift distortion caused by sensor movement. Furthermore, it employs dynamic pooling sampling to generate Doppler invariant feature maps and splits time and frequency token sequences for cross-attention weighting, enabling it to focus on transient impact signals as short as tens of milliseconds, avoiding spectral leakage and global average dilution effects caused by long-window STFTs. Finally, this invention achieves high-precision identification of three core faults: wheel flange wear, bearing failure, and fastener loosening, thus exhibiting strong engineering adaptability and generalization ability.