Newtonian Relaxation (Nudging)

lannymac · November 18, 2014, 22:30

I'm trying to implement nudging into OF. Nudging is described by: da/dt = G*(anudge - a) where a is the variable you want to nudge, anudge is what you are nudging it towards, and G is an inverse timescale.

Right now I'm using bouyantPimpleFoam and I want to nudge U, T, and P (or rho if need be). I'm having a lot of trouble implementing this into the code. Does anyone have any suggestions?

Cheers,
Landan

chriss85 · November 21, 2014, 05:32

I suppose this is similar to underrelaxation, but in this case you "relax" the field to some other value than the last calculation step. You'll have to add your modification of the field values somewhere, for example at the end of a time step or so.

Relaxation works like this:
U=f*U_n+(1-f)*U_(n-1)
with f = [0..1]

lannymac · November 28, 2014, 15:33

Thanks chriss85.

I was more so thinking of nudging the simulation to some measured meteorology. Do you think you're method should still work best?

Cheers

chriss85 · December 1, 2014, 01:52

No, underrelaxation is only used to increase numerical stability. You might be able to look at some relaxation code to figure out how to do your nudging though.

November 18, 2014, 22:30	Newtonian Relaxation (Nudging)	#1
lannymac New Member Join Date: Jun 2014 Posts: 2 Rep Power: 0	I'm trying to implement nudging into OF. Nudging is described by: da/dt = G*(anudge - a) where a is the variable you want to nudge, anudge is what you are nudging it towards, and G is an inverse timescale. Right now I'm using bouyantPimpleFoam and I want to nudge U, T, and P (or rho if need be). I'm having a lot of trouble implementing this into the code. Does anyone have any suggestions? Cheers, Landan

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
buoyantPimpleFoam - no relaxation	Tobi	OpenFOAM Bugs	1	January 14, 2014 17:14
Multiphase and Non Newtonian fluid	wanna88	FLUENT	0	May 27, 2013 00:09
Relaxation and convergence	sammi	Phoenics	0	March 20, 2008 03:32
Question on adjusting relaxation factor	CFD Rookie	Main CFD Forum	3	January 26, 2004 14:37
Relaxation factor	Moon	Siemens	1	June 13, 2003 11:13

November 21, 2014, 05:32		#2
chriss85 Senior Member Join Date: Oct 2013 Posts: 397 Rep Power: 18	I suppose this is similar to underrelaxation, but in this case you "relax" the field to some other value than the last calculation step. You'll have to add your modification of the field values somewhere, for example at the end of a time step or so. Relaxation works like this: U=fU_n+(1-f)U_(n-1) with f = [0..1]

November 28, 2014, 15:33		#3
lannymac New Member Join Date: Jun 2014 Posts: 2 Rep Power: 0	Thanks chriss85. I was more so thinking of nudging the simulation to some measured meteorology. Do you think you're method should still work best? Cheers

December 1, 2014, 01:52		#4
chriss85 Senior Member Join Date: Oct 2013 Posts: 397 Rep Power: 18	No, underrelaxation is only used to increase numerical stability. You might be able to look at some relaxation code to figure out how to do your nudging though.