On Mon, Feb 9, 2009 at 2:37 PM, Ionut Vancea <[email protected]> wrote:
>
> Hello, Daniel,
>
> On Mon, Feb 9, 2009 at 6:29 PM, Daniel Wheeler
> <[email protected]> wrote:
>>
>> If I replace the var**3 with h**3 the system seems stable, but uninteresting.
>>
>> ##eq = TransientTerm() == -DiffusionTerm((var**3, 1)) +
>> DiffusionTerm(-exp(-var) * var**3 + 3 * b / var)
>> eq = TransientTerm() == -DiffusionTerm((h**3, 1)) +
>> DiffusionTerm(-exp(-var) * h**3 + 3 * b * h**3 / var**4)
>>
>> In the email you wrote above you have:
>>
>> \partial_t \phi=-\nabla\{\phi^3\nabla[\triangle \phi
>> -\partial_\phi f(\phi)]\}
>>
>> where f(\phi) = \frac{b}{2\phi^2}-e^{-\phi}
>>
>> Why did you replace the phi^3 with a constant?
>
> Yes, the above equations are correct. I just changed the name of
> "\phi" variable with "h"
> in the code, I dont't hink that I replaced phi^3 with h.
In the code you sent h**3 is being used as one of the fourth order
diffusion coefficients. In the code the value of h is fixed to be 5.
When you wrote the equations with tex, in one of the earlier emails,
the diffusion coefficient is phi**3, which is the solution variable.
Have I misunderstood something?
The tex equation that you wrote is:
\partial_t \phi=-\nabla\{\phi^3\nabla[\triangle \phi
-\partial_\phi f(\phi)]\}
The coded equation that you wrote is:
eq = TransientTerm() == -DiffusionTerm((h**3, 1)) +
(fDerivative.getFaceGrad() * h**3).getDivergence()
and h is set to 5.0. These are different, hence my confusion.
--
Daniel Wheeler
