PDE solve

[marc400317]

Hi all, as part of the design of a catalytic combustor, I had completed a PDE system that governs the temperature distribution of air and ceramic as hot air flows through a honeycomb Ceramic cylinder. I am positive that the PDE system is right, however when plotting the result using MathCad, it doesn't look right, the plot is overshooting and not stable, the graph should be decaying exponentially. I found that by adjusting the xpts, and tpts ( the discretisation pts in PDEsolve) the graph tend to change but still haven't got the right results! my question would be, is there any relation between the Xpts and tpts to get the right answer?
I have attached summary of my work

Thanks !

 

[philipoakley]

There are edge limit effects on the number of points, including a boundary effect related to the combination of algorithms. the cubic spline interpolator can go very wrong some times.

I'm sure the guys on the Collaboratory

http://collab.mathsoft.com/~Mathcad2000

will be able to give help. Tom is very good, though JmG can be irracible.

Philip Oakley