METHOD FOR FREQUENCY-TIME CORRELATION ANALYSIS OF DIGITAL SIGNALS Russian patent published in 2020 - IPC G01R23/16 

FIELD: physics.

SUBSTANCE: invention relates to digital signal processing and can be used for analysis of signals of various origin when solving tasks of non-destructive testing and diagnostics of equipment based on correlation analysis. Method of frequency-time correlation analysis of noisy signals includes direct Fourier transform in form of fast Fourier transform of two input signals, determining a complex-conjugate value of direct conversion results of one of signals, pairwise multiplication of obtained complex signals of direct Fourier transform with complex-conjugated values of the Fourier direct Fourier transform of the second signal, generation of m signals M^k from the obtained product P_j according to the expression. Obtained signals M^k are subjected to inverse Fourier transform. Fourier-time correlation function is determined from results of inverse Fourier transformation, according to the graph of which correlation of harmonic components of signals at different frequencies is determined.

EFFECT: reduced noise effect on resultant values of frequency-time correlation function.

1 cl, 3 dwg