Deep Neural Network Application in the Phase-Match Calibration of Gas–Microphone Photoacoustics

In this paper, a methodology for the application of neural networks in phase-match calibration of gas–microphone photoacoustics in frequency domain is developed. A two-layer deep neural network is used to determine, in real-time, reliably and accurately, the phase transfer function of the used microphone, applying the photoacoustic response of aluminum as reference material. This transfer function was used to correct the photoacoustic response of laser-sintered polyamide and to compare it with theoretical predictions. The obtained degree of correlation of the corrected and theoretical signal tells us that our method of phase-match calibration in photoacoustics can be generalized to a photoacoustic response coming from a solid sample made of different materials.