scholarly journals An ADMM-Newton-CNN numerical approach to a TV model for identifying discontinuous diffusion coefficients in elliptic equations: Convex case with gradient observations

2021 ◽  
Author(s):  
Wenyi Tian ◽  
Xiaoming Yuan ◽  
Hangrui Yue
2006 ◽  
Vol 258-260 ◽  
pp. 360-365
Author(s):  
M. Stasiek ◽  
Andreas Öchsner

A numerical approach for the segregation of atomic oxygen at Ag/MgO interfaces is presented. A general segregation kinetics is considered and the coupled system of differ- ential equations is solved due to a one-dimensional finite difference scheme which accounts for concentration-dependent diffusion coefficients. Based on a model oxide distribution, the influence of the concentration-dependency is numerically investigated and compared with the solution for constant coefficients. In addition, the numerical approach allows for the consider- ation of general boundary conditions, specimen sizes and time-dependent material and process parameters.


2017 ◽  
Vol 3 (2) ◽  
pp. 199-217
Author(s):  
Fayssal Benkhaldoun ◽  
Mohammed Seaid ◽  
Amadou Mahamane

AbstractThe major emphasis of this work is the derivation of a posteriori error estimates for the mixed finite volume discretization of second-order elliptic equations. The estimates are established for meshes consisting of simplices on unstructured grids. We consider diffusion problems with nonhomogeneous diffusion coefficients. The error estimates are of residual types and are formulated in the energy semi-norm for a locally postprocessed approximate solutions. The estimates are fully computable and locally efficient that they can serve as indicators for adaptive refinement and for the actual control of the error. Numerical results are shown for two test examples in two space dimensions. It is found that the proposed adaptive mixed finite volume method offers a robust and accurate approach for solving second-order elliptic equations, even when highly nonhomogeneous diffusion coefficients are used in the simulations.


Sign in / Sign up

Export Citation Format

Share Document