steady solution, the lift fluctuate,but unsteady method,the lift is convergent

ti_hen · January 25, 2016, 10:48

case: flow over a naca0012 airfoil, re=20000, cfx, sst,AOA=10deg. In my simulation, when using steady solution, the monitor lift is fluctuating,I think it is right. But when I switch it in unsteady, and want to capture the wortex shedding, unfortunatly, the lift seems convergent. Y+ <1, time step is 0.0001, I can't find the reason. Can anyone help me? Thankyou.

ghorrocks · January 26, 2016, 01:52

FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

This sounds a case where you need to make sure your time discretisation schemes are accurate. Ensure your time step size is suitable (by a sensitivity check) and you are using second order time differencing.

ti_hen · January 26, 2016, 04:12

ghorrocks,Thankyou for your reply.
The transient scheme is second order backward euler. The time step 0.001,0.0001,0.00005,0.00001 have already tried,but there is still no vortex shedding.

ghorrocks · January 26, 2016, 05:44

A naca0012 airfoil at Re=20000 will have a large amount of laminar boundary layer on it. Are you using the turbulence transition model?

Also: at that Re the transient behaviour is probably a laminar separation bubble, not vortex shedding. This means it is not shedding vorticies into the flow but the separation bubble jiggling about.

ti_hen · January 26, 2016, 06:21

Yes, I use Gamma model transitional turbulence.It is really a laminar separation bubble.As you say, the bubble jigging, so the lift can also jiggling. but i can not observe that in my simulation.( why steady the lift is fluctuating? i still puzzled)

ghorrocks · January 26, 2016, 17:35

Have you checked the drag? It is probably moving more.

How much do you expect the lift to move around by?

ti_hen · January 29, 2016, 04:19

The drag is the same with lift. The flowfield is almost the same between 1.0s and 1.5s. Actually, the lift change very very small, for example, the lift is 0.6472 during more than 0.5s, only the fifth digit after the decimal point is changing. In steady solution ,the flowfield changing obviously.

ghorrocks · January 29, 2016, 04:51

So you are modelling the motion of the laminar separation bubble. That is what you wanted to do isn't it?

ti_hen · January 29, 2016, 04:55

Yes, but using transient solution, the laminar separation bubble is steady. I think there is must something wrong this my solution.

ghorrocks · January 29, 2016, 07:14

Yes, I agree. My suggestions are in my post #2.

January 25, 2016, 10:48	steady solution, the lift fluctuate,but unsteady method,the lift is convergent	#1
ti_hen New Member hen Join Date: Jan 2016 Posts: 5 Rep Power: 10	case: flow over a naca0012 airfoil, re=20000, cfx, sst,AOA=10deg. In my simulation, when using steady solution, the monitor lift is fluctuating,I think it is right. But when I switch it in unsteady, and want to capture the wortex shedding, unfortunatly, the lift seems convergent. Y+ <1, time step is 0.0001, I can't find the reason. Can anyone help me? Thankyou.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Lift is oscillating while using steady state solver.	jmmichie	Main CFD Forum	10	March 18, 2015 04:55
Unsteady Flow past circular cylinder-Re=100 (fluctuating drag and lift co-efficients)	Vino	Main CFD Forum	0	April 10, 2014 09:39
Unsteady vs Steady simulation - any difference?	Sally	Main CFD Forum	16	December 16, 2005 09:15
steady or unsteady? (in dpm)	winnie	FLUENT	1	April 28, 2003 11:30
About the difference between steady and unsteady problems	Lisa	Main CFD Forum	11	July 5, 2000 14:37

January 26, 2016, 01:52		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F This sounds a case where you need to make sure your time discretisation schemes are accurate. Ensure your time step size is suitable (by a sensitivity check) and you are using second order time differencing.

January 26, 2016, 04:12		#3
ti_hen New Member hen Join Date: Jan 2016 Posts: 5 Rep Power: 10	ghorrocks,Thankyou for your reply. The transient scheme is second order backward euler. The time step 0.001,0.0001,0.00005,0.00001 have already tried,but there is still no vortex shedding.

January 26, 2016, 05:44		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	A naca0012 airfoil at Re=20000 will have a large amount of laminar boundary layer on it. Are you using the turbulence transition model? Also: at that Re the transient behaviour is probably a laminar separation bubble, not vortex shedding. This means it is not shedding vorticies into the flow but the separation bubble jiggling about.

January 26, 2016, 06:21		#5
ti_hen New Member hen Join Date: Jan 2016 Posts: 5 Rep Power: 10	Yes, I use Gamma model transitional turbulence.It is really a laminar separation bubble.As you say, the bubble jigging, so the lift can also jiggling. but i can not observe that in my simulation.( why steady the lift is fluctuating? i still puzzled)

January 26, 2016, 17:35		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Have you checked the drag? It is probably moving more. How much do you expect the lift to move around by?

January 29, 2016, 04:19		#7
ti_hen New Member hen Join Date: Jan 2016 Posts: 5 Rep Power: 10	The drag is the same with lift. The flowfield is almost the same between 1.0s and 1.5s. Actually, the lift change very very small, for example, the lift is 0.6472 during more than 0.5s, only the fifth digit after the decimal point is changing. In steady solution ,the flowfield changing obviously.

January 29, 2016, 04:51		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	So you are modelling the motion of the laminar separation bubble. That is what you wanted to do isn't it?

January 29, 2016, 04:55		#9
ti_hen New Member hen Join Date: Jan 2016 Posts: 5 Rep Power: 10	Yes, but using transient solution, the laminar separation bubble is steady. I think there is must something wrong this my solution.

January 29, 2016, 07:14		#10
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Yes, I agree. My suggestions are in my post #2.