Use neural network in photoacoustic measurement of thermoelastic properties of aluminum foil

•Neural networks (NNs) are developed for treatment of photoacoustic (PA) experiments.•High precision of thermo-elastic PA characterization of aluminum has been achieved.•The extended measurement range and NNs enable increased capacity of PA measurement. The objective of this paper is to present meth...

Full description

Saved in:

Bibliographic Details
Published in:	Measurement : journal of the International Measurement Confederation Vol. 199; p. 111537
Main Authors:	Djordjević, К.Lj, Galović, S.P., Popović, M.N., Nešić, M.V., Stanimirović, I.P., Stanimirović, Z.I., Markushev, D.D.
Format:	Journal Article
Language:	English
Published:	Elsevier Ltd 01-08-2022
Subjects:	Aluminum Inverse problem Linear coefficient of thermal expansion Minimum volume cell Neural networks Photoacoustic Thermal diffusivity Photoacoustic Aluminum Neural networks Thermal diffusivity Linear coefficient of thermal expansion Minimum volume cell Inverse problem
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	•Neural networks (NNs) are developed for treatment of photoacoustic (PA) experiments.•High precision of thermo-elastic PA characterization of aluminum has been achieved.•The extended measurement range and NNs enable increased capacity of PA measurement. The objective of this paper is to present methodology of the precise and reliable determination of thermal diffusivity and linear coefficient of thermal expansion of the photoacoustic signal recorded (obtained) using open photoacoustic cell where thickness of the sample served as a control parameter. The methodology was based on the application of neural networks that were trained on numerical experiments and optimized by adding Gaussian noise to the training base that corresponded in percentage to maximum measurement uncertainty. By comparing the predictions of the neural network with theoretical fitting curve for experimental results for the aluminum sample that was 197 μm thick, it was shown that the proposed methodology achieves high precision in the determination of thermoelastic and geometrical properties of the sample.
ISSN:	0263-2241 1873-412X
DOI:	10.1016/j.measurement.2022.111537