|
[Sponsors] |
November 7, 2008, 12:12 |
About the jacobian
|
#1 |
Guest
Posts: n/a
|
Hi all,
I have a weird, maybe stupid, question about the jacobian. if g = u*dudx + v*dudy (convective termes in NS equations) i.e. dgdu = d(u*dudx + v*dudy)du so how am I supposed to derive for example dudx with respect to u? I know how to do it in Finite Elements but right now I'm applying it in Finite Differences. Thanks in advance |
|
November 7, 2008, 12:39 |
Re: About the jacobian
|
#2 |
Guest
Posts: n/a
|
I'm not sure your math is correct. The typical approach to the convective terms is to write them in a conservative form, so that
g = 0.5*d(u**2)/dx + 0.5*d(v**2)/dy to use your example. Then the flux terms are defined as g1 = 0.5*u**2 , g2 = 0.5*v**2 and the Jacobian is defined as the derivatives of the flux terms with respect to the variables (u,v). Not sure if this helps. |
|
November 7, 2008, 13:25 |
Re: About the jacobian
|
#3 |
Guest
Posts: n/a
|
Do you mean
g = 0.5*d(u**2)/dx + 0.5*d(u*v)/dy (change is second term)? |
|
November 7, 2008, 15:29 |
Re: About the jacobian
|
#4 |
Guest
Posts: n/a
|
My bad - that's what I get for not getting my glasses. Thanks.
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
What's the use of Jacobian in Grid Generation? | Pete | Main CFD Forum | 5 | April 6, 2014 21:23 |
Jacobian of 1D Euler Equation of Gas Dynamics | Simplicity. | Main CFD Forum | 3 | September 6, 2011 14:12 |
Viscous Flux Jacobian | bearcat | Main CFD Forum | 10 | March 11, 2010 17:14 |
how to compute the jacobian of a line | ztdep | Main CFD Forum | 0 | June 1, 2007 07:04 |
Multigrid and Jacobian | biga | Main CFD Forum | 2 | July 25, 2001 23:08 |