|Paola Gervasio - DICATAM - University of Brescia - paola.gervasio_at_unibs.it|
Scientific Computing course (italian)
PhD programme in Civil and Environmental Engineering, International cooperation and Mathematics (DICACIM)
NUMERICAL METHODS FOR PARTIAL DIFFERENTIAL EQUATIONS
Ph.D. Programme in
Civil and Environmental Engineering, International
cooperation and Mathematics (DICACIM)
|Lecture (1h)||Introduction to PDEs.||
|Lecture (3h)||Elliptic problems and Elements of Functional Analysis. Strong form of 2nd order elliptic problems for d=1. The space L^2. Functionals and forms on normed spaces. Distributions and differentiation of distributions. The Sobolev space H^1. Weak form of the Poisson problem with homogeneous Dirichlet boundary conditions for d=1. Non-homogeneous Dirichlet conditions for d=1. 2nd-order elliptic problems with Neumann boundary conditions for d=1. Lax-Milgram lemma.||
Finite Elements approximation of 1D elliptic problems.
3a. Galerkin formulation of the discrete problem. The algebraic formulation of the discrete problem. The Cea Lemma. Estimate of the continuity and of the coercivity constants for the bilinear form a(u,v)=(u',v')+sigma(u,v).
3b. Linear finite elements: the Lagrangian basis, construction of the mass and stiffness matrices.
3c. Construction of the right hand side. Gauss-Legendre quadrature formulas. Quadratic finite elements. The Lagrange composite interpolation (of degree 1 and of degree r gt 1). Estimate of the error between the continous solution of the pde and the discrete solution.
3d. The 2nd order elliptic problem with non-homogeneous Dirichlet conditions and with Neumann conditions. Connectivity matrix.
|29/04/2020 h. 9.00- 12.00 Laboratory (3h)||MATLAB - FEM 1d for elliptic problems.||
Elliptic problems for d ≥ 2. Finite Elements discretization.
Weak formulation of 2nd order elliptic problem. Triangulations. P1 and Q1 FEM. Matrix assembling. Quadrature formulas. Cholesky factorization to solve the linear system.
|Lecture (1h)||A short review on Spectral Element Methods.||
Approximation of Parabolic problems.
Strong form of the parabolic equations (heat equation). Semidiscrete weak form and Galerkin approximation. Semidiscrete numerical solution.
Approximation of the first order Cauchy problem by Euler and Crank-Nicolson methods.
Theta-method for the discretization of the heat equation.
Absolute stability and convergence of the theta-method applied to the heat equation.
|08/05/2020 h. 15.00 - 17.00 Laboratory (2h)||MATLAB PDEtoolbox||
|15/05/2020 h. 9.00 - 11.00 Laboratory (2h)||MATLAB discretization of the heat equation.||
Advection Diffusion Reaction problems.
Strong and weak formulation. Galerkin approximation.
1D case with linear FEM: the Peclet number and the bound on h to avoid numerical oscillations.
Upwind scheme and centered scheme with artificial diffusion.
2D case: an overview on artificial diffusion, streamline diffusion and Galerkin Least Squares methods.
|15/05/2020 h. 11.15-12.30 Laboratory (1h+ 15')||MATLAB advection diffusion problems.||
|Paola Gervasio - May 2020|