Numerical exploration of a system of reaction-diffusion equations with internal and transient layers (Q2572110)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Numerical exploration of a system of reaction-diffusion equations with internal and transient layers |
scientific article |
Statements
Numerical exploration of a system of reaction-diffusion equations with internal and transient layers (English)
0 references
14 November 2005
0 references
A model for chemical reaction between two species \(A\), \(B\) is considered especially from computational point of view. This chemical reaction is of the type that it has two phases and the first one with by product \(C\) is several orders of magnitudes faster than the other. From the mathematical point of view we can model it by a system of three nonlinear reaction-diffusion equations of the type \[ \begin{aligned} u_t & = u_{xx} -\lambda uv -uw,\\ v_t& =v_{xx}-\lambda uv,\\ w_t & =w_{xx} +\lambda uv -uw,\end{aligned} \] in \( \Omega=(0,1)\), where \(u,v,w\) represent the unknown concentration of species A, B and the byproduct of the first reaction \(C\). Dirichlet and Neumann boundary conditions are prescribed: \[ u=\alpha, \;v_x =0, \;w_x =0 \text{ at } x=0, \] \[ u_x=0, \;v=\beta, \;w_x =0, \text{ at } x=1, \] and the initial condition \[ u(x,0) = u_{\text{ini}}(x), \;v(x,0) = v_{\text{ini}}(x), w(x,0)= w_{ini}(x). \] Due to the appearance of the large factor \( \lambda>>1\) in reaction-diffusion equations, this model has features of singularly perturbed problems. Convergence of a finite difference discretization for the stationary problem is well known as well as the existence of a steady state solution and the convergence as \( \lambda \rightarrow \infty\). Using the knowledge about the behavior of the stationary problem, the authors want to investigate the behavior of the time-dependent problem in this article. The authors use a finite element method with linear basis functions for space discretization. For simplicity they use uniform mesh. Thus using the method of lines, for solving the ODE system in time they use ode15s solver in MATLAB with automatic step size and order control. In simulation examples, they are interested mainly on internal layers of a solution which appear also in the steady state problem. The first group of simulated examples are of the type when \( \lambda = 10^6\) and different initial conditions are chosen. Figures that show concentrations \(u,v,w\) in several time steps and then the reaction rate \( q=\lambda uv\) are included. The motion of the interface between regions dominated by species \(A\), or \(B\) is also presented. The results obtained confirm the conjecture about the smooth behavior of the interfaces and their monotone coalescence. The width of the internal layers are of the same order of magnitude as at the steady state problem. The results validate that the transient solution tends to the steady state.
0 references
reaction-diffusion system of equations
0 references
internal layer
0 references
transient layer
0 references
method of lines
0 references
finite element method
0 references
simulation
0 references
smooth behavior of the interfaces
0 references
monotone coalescence
0 references
0 references
0 references
0 references
0 references