CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Lax-Friedrichs flux (http://www.cfd-online.com/Forums/main/91088-lax-friedrichs-flux.html)

panou July 30, 2011 10:02

Lax-Friedrichs flux
 
Have anyone a paper which explain how lax-friedrich(approximate riemann solver) is derived?

Also if we suppose that (u+) is the values of a volume , (u-) the values of neighbour volumes and F the flux then the lax-friedrichs flux gives:

F=0.5*( (F+) + (F-) ) - a*( (u-) - (u+) )

This is applied by all faces of a initial volume regardless of its outward unit vector?

cfdnewbie July 30, 2011 12:45

There's nothing too fancy about the LF-flux....its the average, penalized by the jump to add a strong dissipative constant.... not sure about the original paper, anybody has a hint???

you compute your F by the formula given above, and then do int (F*n) dx to take the normal into account...

hope this helps a bit!

panou July 30, 2011 17:09

Yes , sure. The third term is the dissipative term. Simply , i did not know if it has a complex theory under of it (due to appearance of eigenvectors as dissipative multiplier). With other words, i want to learn what assumptions do for to solve approximate the Riemann problem and finally result at this simply relation.

For other cases as HLLC or HLL there is a assumption on a number of waves. Here?

cfdnewbie July 31, 2011 03:52

Quote:

Originally Posted by panou (Post 318168)

For other cases as HLLC or HLL there is a assumption on a number of waves. Here?

Hm, I'm only familiar with the LOCAL LF (also called Rusanov), where the a in your equation above is the local largest EV.... so that's your fastest wave...


All times are GMT -4. The time now is 06:50.