Noise reduction of continuous recorded laser AE waveform by time-frequency analysis

A laser AE method is non-contact measurement of AE by using laser interferometer sensor. However, the sensitivity of laser interferometer is lower than conventional piezoelectric sensor. In this study, a method to improve signal-to-noise ratio of the output waveform of interferometer is investigated. A new system was developed to record and analyze the waveform continuously. Continuous waveform was transformed into time-frequency-magnitude data by short time Fourier transform (STFT) or wavelet transform. After noise component in this time-frequency magnitude data is cut, inverse STFT or WT is applied to obtain clear AE signals in time domain. Simulated AE signal generated by thermal stress of YAG laser pulse was analyzed and detection capability of AE event was improved compared to the conventional signal processing method.