Undergraduate Thesis of Sergio Carrasco
|Career||Mathematical Civil Engineering, Universidad de Concepción|
|Thesis Title||New Mixed Finite Element Methods for the Coupled Convective Brinkman-Forchheimer and Double-Diffusion Equations|
In this thesis we introduce and analyze new Banach spaces-based mixed finite element methods for the stationary nonlinear problem arising from the coupling of the convective Brinkman- Forchheimer equations with a double diffusion phenomenon. Besides the velocity and pressure variables, the symmetric stress and the skew-symmetric vorticity tensors are introduced as auxiliary unknowns of the fluid. Thus, the incompressibility condition allows to eliminate the pressure, which, along with the velocity gradient and the shear stress, can be computed afterwards via postprocessing formulae depending on the velocity and the aforementioned new tensors. Regarding the diffusive part of the coupled model, and additionally to the temperature and concentration of the solute, their gradients and pseudoheat/pseudodiffusion vectors are incorporated as further unknowns as well. The resulting mixed variational formulation, settled within a Banach spaces framework, consists of a nonlinear perturbation of, in turn, a nonlinearly perturbed saddle-point scheme, coupled with a usual saddle-point system. A fixedpoint strategy, combined with classical and recent solvability results for suitable linearizations of the decoupled problems, including in particular, the Banach-Neˇcas-Babuˇska theorem and the Babuˇska-Brezzi theory, are employed to prove, jointly with the Banach fixed-point theorem, the well-posedness of the continuous and discrete formulations. Both PEERS and AFW elements of order ℓ ě 0 for the fluid variables, and piecewise polynomials of degree ď ℓ together with Raviart-Thomas elements of order ℓ for the unknowns of the diffusion equations, constitute feasible choices for the Galerkin scheme. In turn, optimal a priori error estimates, including those for the postprocessed unknowns, are derived, and corresponding rates of convergence are established. Finally, several numerical experiments confirming the latter and illustrating the good performance of the proposed methods, are reported.
|Thesis Director(s)||Sergio Caucao, Gabriel N. Gatica|
|Thesis Project Approval Date||2023, March 01|
|Thesis Defense Date||2023, August 03|
|PDF Tesis||Download Thesis PDF|