Authors: C.P. Coelho, S. Desai, D. Freeman and J.K. White
Affilation: Massachusetts Institute of Technology, United States
Pages: 549 - 552
Keywords: diffusion constant, parameter extraction, image data, convection diffusion
In this paper we present a robust and geometry independent method for the estimation of diffusion coeffients of solutes from image data. Our diffusion coefficient estimation algorithm works in two steps. First, given the device geometry and the pressure boundary conditions, we calculate the fluid velocity in the device by solving a Stokes Flow problem. Then, using the velocity and the concentration at the device inlet and outlet ports we calculate the solute concentration in the device by solving the convection diffusion problem with a three dimensional finite differences solver. The diffusion coefficient is estimated by using Newton's method to minimize the square error between the measured concentration data and the concentration estimates produced with our simulation tool.