Timo Lähivaara

Learn More
We study the feasibility of data based machine learning applied to ultrasound tomography to estimate watersaturated porous material parameters. In this work, the data to train the neural networks is simulated by solving wave propagation in coupled poroviscoelastic-viscoelastic-acoustic media. As the forward model, we consider a high-order discontinuous(More)
Groundwater is a hugely important resource in both developed and developing countries and makes up over 90% of the world’s freshwater. The idea behind our work is to use wave scattering in saturated porous media to quantify aquifer state and hydraulic properties. We initially focused on solving related inverse problem using the Specfem software to model(More)
Time-dependent wave fields are often approximated using wave equations. Due to the oscillatory nature of wave phenomena, numerical approximations of wave equations are challenging. Approximating wave problems with tolerable accuracy requires the use of a relatively dense spatial discretization. For standard numerical techniques, such as the low-order finite(More)
A method to characterize macroscopically homogeneous rigid frame porous media from impedance tube measurements by deterministic and statistical inversion is presented. Equivalent density and bulk modulus of the samples are reconstructed with the scattering matrix formalism, and are then linked to its physical parameters via the(More)
  • 1