This is a read only copy of the old FEniCS QA forum. Please visit the new QA forum to ask questions

System on non-linear PDEs

0 votes

Hi, I am new to fenics and I need to solve a challenging non-linear system of PDEs. The system has functions A,B and u which are all functions of x and t. It looks like:

\begin{equation}
-(A^{-1} \sqrt{AB}x^2 u_t)_t + (B^{-1}\sqrt{AB}x^2 u_x)_x + \frac{dV(u)}{du} \sqrt{AB}x^2 = 0
\end{equation}
\begin{equation}
A_x = A\frac{B - 1}{x} + x(A u_x^2 - A B u^2 + B u_t^2)
\end{equation}
\begin{equation}
B_x = -B\frac{B - 1}{x} + x B (\frac{B}{A} u_t^2 + u_x^2 + B V(u))
\end{equation}
where $V(u) = (1-exp(-u))^2$.
In the end I would like to solve the system on a square mesh and have the following boundary conditions: $u(t,x=0) = 0.1$, $u(t,x=xmax) = 0$, $u(t=0,x) = u(t=tmax)$, $B(t,x=0) = 1$, $A(t,x=0) = 0$.

I would greatly appreciate if someone would help me getting started with this problem.

asked Aug 4, 2015 by niklas FEniCS Novice (120 points)
edited Aug 5, 2015 by niklas

1 Answer

0 votes

Have you looked at the demos on the FEniCS website?

http://fenicsproject.org/documentation/dolfin/1.6.0/python/demo/index.html

answered Aug 4, 2015 by aldenpack FEniCS User (1,450 points)

I have looked at some demos, but I have not managed to figure out how to formulate my system of non-linear PDEs.
The Cahn-Hillard demo is nonlinear and in the mixed formulation, perhaps I can generalize this, eventhough it is first order in time.

...