Author |
Message |
Paolo Di Blasi (unknown089)
New member Username: unknown089
Post Number: 1 Registered: 05-2008
| Posted on Friday, May 30, 2008 - 06:54 am: | |
I've a problem to solve this kind of differential's equazion: [url]http://img140.imageshack.us/my.php?image=eqdiffea8.jpg[/url] where U, H, V=\0 are knowndata , phi=0, and g it's a theta1 function my problem is that i haven't the condition on the first derivate (cauchy condition) but i've two condition on the position (dirichelet conditions), this problem should be solved by shooting technique,i.e. try with an hypothetic cauchy condition and one of the two dirichelet condition and check the value of the function where we have the other dirichelet condition . thanks so much |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 1124 Registered: 06-2003
| Posted on Friday, May 30, 2008 - 01:03 pm: | |
If you have a value condition at each end of the interval, simply pose the problem as a 1D boundary value problem with a VALUE(g) condition at each end. No shooting is required.
|
Paolo Di Blasi (unknown089)
New member Username: unknown089
Post Number: 2 Registered: 05-2008
| Posted on Monday, June 02, 2008 - 05:21 pm: | |
sorry but i try to solve this differential equation but it doesn't work, i check the brackets and they're correct... what kind of error is this... |
Paolo Di Blasi (unknown089)
Junior Member Username: unknown089
Post Number: 3 Registered: 05-2008
| Posted on Monday, June 02, 2008 - 05:28 pm: | |
the value of the data v,u,h ,are arbitrary, |
Marek Nelson (mgnelson)
Moderator Username: mgnelson
Post Number: 40 Registered: 07-2007
| Posted on Monday, June 02, 2008 - 06:24 pm: | |
1) In a 1D problem, you need to use POINT VALUE not VALUE for your boundary conditions. If you do this, you get a slightly more descriptive error message : "Log of a negative number (-17382.8)". 2) In your equation, gamma1 = ln[(1+g')/(1-g')]. This will be taking the logarithm of a negative number anytime g'>1 or g'<-1. This is causing your error. With your boundary values of g and the domain size you have given, g' must always be greater than 1 somewhere. So your system has no solution as posed. 3) Try defining gamma and gamma1 thru gamma3 as named parameters. This will make the equation cleaner and will allow you to use MIN and MAX to put a clamp on (1+g') and (1-g') before taking the natural logarithm.
|
|