when to stop my airfoil Transient Simulation ?!

losiola · June 10, 2019, 06:56

Hello Foamers,
i am making 2D airfoil simulation (transient simualtion using pimpleFoam ) and i want to know when to stop the simualation ?

Someone told me an answer that i have to calculate L = U*n*dt and compare it to my domain but i didnt understand why? or from where did he get that formula!

can you please help me with this

Many Thanks

hxaxtma · June 11, 2019, 03:20

Hi Iosiola,

since you did not mention your setup in detail I can only give you some general remarks:

1) First watch your residuals, a simuation with residuals lower 1e-04 for pressure and momentum is in the industrial sense well converged

2) More important: Watch your physics, which means monitor the forces of your blade or some points in the wake of your blade. If the integral forces reach saturation on a steady flow regime, they get constant, if the regime is per se unsteady, the reassamble kind of a periodic behaviour. At this stage, you can only observer proper statistics by sampling over 2 periods.

3) L = U*n*dt. I guess your colleague meant by this formula in words: how many times the flow passes through the domain. From enginnering point of view there are some "rule of thumbs" about that, depending on your setup. For channel flow, people say you need 5-10 L or fan simulation 5. As I mentioned, these are only rule of thumbs. So WATCH YOUR PHYSICS -> see point 2

4) Another integral value I do often choose for observing the point of saturation is the integral of the turbulent kinetic energy (https://en.wikipedia.org/wiki/Turbulence_kinetic_energy) over the whole domain. After certain time steps it shows off as kind of constant or perdiodic bevahiour. This monitoring parameter is more phyiscal to me, since all the velolicty fluctuations are used.

So, I hope this helps you out,
best

June 10, 2019, 06:56	when to stop my airfoil Transient Simulation ?!	#1
losiola Member Os Join Date: Jun 2017 Posts: 80 Rep Power: 8	Hello Foamers, i am making 2D airfoil simulation (transient simualtion using pimpleFoam ) and i want to know when to stop the simualation ? Someone told me an answer that i have to calculate L = Undt and compare it to my domain but i didnt understand why? or from where did he get that formula! can you please help me with this Many Thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
CFX/Fluent- Underprediction of Lift Coeffcient in 2D Airfoil Simulation	DanielBarreiro	Main CFD Forum	4	September 2, 2021 02:16
Need help for the 2D airfoil simulation using OpenFoam	losiola	OpenFOAM Running, Solving & CFD	0	October 10, 2018 10:22
Transient Simulation in CFX	ashtonJ	CFX	1	September 28, 2018 01:18
restarting paused transient simulation using reactingFoam	JMDag2004	OpenFOAM Running, Solving & CFD	1	August 10, 2015 10:15
Boundary Conditions - Transient Simulation	miki256	CFX	2	May 18, 2012 01:22

June 11, 2019, 03:20		#2
hxaxtma Senior Member Join Date: Jan 2014 Posts: 179 Rep Power: 12	Hi Iosiola, since you did not mention your setup in detail I can only give you some general remarks: 1) First watch your residuals, a simuation with residuals lower 1e-04 for pressure and momentum is in the industrial sense well converged 2) More important: Watch your physics, which means monitor the forces of your blade or some points in the wake of your blade. If the integral forces reach saturation on a steady flow regime, they get constant, if the regime is per se unsteady, the reassamble kind of a periodic behaviour. At this stage, you can only observer proper statistics by sampling over 2 periods. 3) L = Undt. I guess your colleague meant by this formula in words: how many times the flow passes through the domain. From enginnering point of view there are some "rule of thumbs" about that, depending on your setup. For channel flow, people say you need 5-10 L or fan simulation 5. As I mentioned, these are only rule of thumbs. So WATCH YOUR PHYSICS -> see point 2 4) Another integral value I do often choose for observing the point of saturation is the integral of the turbulent kinetic energy (https://en.wikipedia.org/wiki/Turbulence_kinetic_energy) over the whole domain. After certain time steps it shows off as kind of constant or perdiodic bevahiour. This monitoring parameter is more phyiscal to me, since all the velolicty fluctuations are used. So, I hope this helps you out, best