Niskanen Matti, Duclos Aroune, Dazel Olivier, Groby Jean-Philippe, Kaipio Jari, Lähivaara Timo
Department of Applied Physics, University of Eastern Finland, Postal Box 1627, FIN-70211 Kuopio, Finland.
Laboratoire d'Acoustique de l'Université du Mans, LAUM - Unité mixte de recherche Centre national de la recherche scientifique, UMR CNRS 6613, Le Mans Université, Avenue Olivier Messiaen, F-72085 Le Mans Cedex 9, France.
J Acoust Soc Am. 2019 Oct;146(4):2596. doi: 10.1121/1.5129369.
The estimation of poroelastic material parameters based on ultrasound measurements is considered. The acoustical characterisation of poroelastic materials based on various measurements is typically carried out by minimising a cost functional of model residuals, such as the least squares functional. With a limited number of unknown parameters, least squares type approaches can provide both reliable parameter and error estimates. With an increasing number of parameters, both the least squares parameter estimates and, in particular, the error estimates often become unreliable. In this paper, the estimation of the material parameters of an inhomogeneous poroelastic (Biot) plate in the Bayesian framework for inverse problems is considered. Reflection and transmission measurements are performed and 11 poroelastic parameters, as well as 4 measurement setup-related nuisance parameters, are estimated. A Markov chain Monte Carlo algorithm is employed for the computational inference to assess the actual uncertainty of the estimated parameters. The results suggest that the proposed approach for poroelastic material characterisation can reveal the heterogeneities in the object, and yield reliable parameter and uncertainty estimates.
