Method for detecting fatigue crack in structure using long short-term memory network-based spectral noise reduction and nonlinear ultrasonic modulation, and system therefor

Abstract

A method for detecting fatigue cracks in a structure and a system therefore are disclosed. A vibration detection element detects vibration of the structure generated when the structure is excited at low and high frequencies and generates a measurement signal in the time domain. Based on a trained long short-term memory (LSTM) network, a signal of the next time step is predicted from the measurement signal of the current time step measured by the vibration detection element. The measurement signal and the prediction signal are combined to create a new combined signal with an increased length in the time domain. The combined signal is Fourier transformed to reduce noise and enhance the information of the measured signal in the frequency domain, and the Fourier transformed signal is used to obtain a spectral density function and to detect cracks in the structure.