| Author |
Message |
    
sue
Member
Username: sue
Post Number: 7
Registered: 06-2010
| | Posted on Thursday, September 09, 2010 - 04:02 pm: | |
Can you advise me on how to compute a function, F, which is a line integral across coordinate x from x=0 to x=L at each value of z (in a 2D cartesian x,z coordinate system)?
In other words:
F = LINE_INTEGRAL(G,'Feature1')
where:
Feature Start "Feature1" (0,z) line to (L,z)
in which variables are apparently not permitted.
Many thanks! |
rgnelson
Moderator
Username: rgnelson
Post Number: 1404
Registered: 06-2003
| | Posted on Thursday, September 09, 2010 - 08:38 pm: | |
"FEATURE" generates a gridding line in the mesh, so it is impossible to place one at "each value of z", since z is a continuum. You could select a few representative z values and run a FEATURE at each one, with a corresponding integral definition.
Or, since "F = Integral wrt x of G" implies dx(F)=G, you can declare F to be another variable and give it the equation "F: dx(F)=g" with VALUE(F)=0 along the boundary x=0. An ELEVATION of F along x=L will show you the distribution in z.
If you are using FlexPDE version 6, you can present this equation after a THEN qualifier to decouple it from the main solution (See "Equations, sequencing" in the Help Index).
|
sue
Member
Username: sue
Post Number: 8
Registered: 06-2010
| | Posted on Friday, September 10, 2010 - 08:13 am: | |
Thank you. I chose to declare another variable as you suggested. The problem now is that the new equation is 0=0 at time=0. This is a true statement for the given problem. Any suggestions for how to address this? Can I somehow wait until t=to+t_increment to solve that equation? |
rgnelson
Moderator
Username: rgnelson
Post Number: 1406
Registered: 06-2003
| | Posted on Saturday, September 11, 2010 - 12:42 pm: | |
It's impossible to guess without seeing the script. Can you post it?
|
|