Convergence criteria (II)

PrandtlGlawert · August 21, 2009, 19:55

Hi everybody:

Shall i consider a fluctuation between two constant (approximately) values as a convergence (in the semi-value of the fluctuation between peak and valley) without assuming a great mistake?

Thx in advance!!

Giuki · August 22, 2009, 05:09

That depends on the convergence history and the amplitude of the convergence I would say. It depends which residual you are considering, maybe you could check the contours and see if somewhere an irregularity occurs, trying to adapt the mesh at those spots and see whether you still have fluctuations. Also you might check the skewness of your mesh...

PrandtlGlawert · August 22, 2009, 06:27

It´s fluctuating around its third-fourth significative number, whose order is 10x^-3. Enough for my Cl, Cd and Cm (i think...)

Ty m8!!

Giuki · August 22, 2009, 06:32

Is it the residual which is fluctuating or your Cl, Cd you are monitoring?

PrandtlGlawert · August 22, 2009, 19:40

Here is an example of the residuals. Almost constant, as you can see... Fluctuation of Cd, Cl and Cm is what expected (anything but weird

), peak-valley-peak-valley-peak-valley and so on...

Thank you m8!!

tstorm · August 22, 2009, 21:41

In my experience residuals on the order of 1e-5 are small enough to consider it converged. I would consider the case converged in your case, assuming the amplitude of the oscillations aren't too large.

That said, I've seen residuals less than 1e-5 for Cl for a circulation-control airfoil and yet the magnitude of oscillations for Cl were on the order of 1 (for an airfoil with an expected Cl of around 6.) In that case, the problem turned out to be three-fold: highly skewed cells in areas of high gradients, too large of cells in these areas, and large area changes. Have you tried grid adaption based on gradients?

Another thing to check would be the under-relaxation factors and courant number. Try increasing all of those to see if it either causes divergence or convergence.

PrandtlGlawert · August 23, 2009, 06:13

Ty m8s, this has been (is being) useful for me

Cheers!!

csri07 · August 23, 2009, 20:28

Hello Everyone

Could any one please help me with vibrational Analysis.

I am carrying Vibrational Analysis on a Fuel Rail, if i want to add mass of the fuel during Modal analysis, how can i do ??

Please help me out.

August 21, 2009, 19:55	Convergence criteria (II)	#1
PrandtlGlawert New Member Daniel LLanes Join Date: Aug 2009 Posts: 11 Rep Power: 16	Hi everybody: Shall i consider a fluctuation between two constant (approximately) values as a convergence (in the semi-value of the fluctuation between peak and valley) without assuming a great mistake? Thx in advance!!

August 23, 2009, 20:28	Vibrational Analysis	#8
csri07 New Member Sri Join Date: Aug 2009 Posts: 1 Rep Power: 0	Hello Everyone Could any one please help me with vibrational Analysis. I am carrying Vibrational Analysis on a Fuel Rail, if i want to add mass of the fuel during Modal analysis, how can i do ?? Please help me out. Last edited by csri07; August 23, 2009 at 20:59.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Convergence criteria	PrandtlGlawert	FLUENT	1	August 13, 2009 18:06
Convergence criteria	PrandtlGlawert	Main CFD Forum	1	August 13, 2009 17:13
star-ccm+ convergence criteria	star	Siemens	2	January 14, 2009 04:57
Convergence Criteria	edwin	FLUENT	1	February 14, 2008 19:24
Convergence criteria	Gaurav	Siemens	0	February 3, 2006 04:53

August 22, 2009, 05:09		#2
Giuki New Member Giuki Cael Join Date: May 2009 Location: Belgium/The Netherlands Posts: 13 Rep Power: 16	That depends on the convergence history and the amplitude of the convergence I would say. It depends which residual you are considering, maybe you could check the contours and see if somewhere an irregularity occurs, trying to adapt the mesh at those spots and see whether you still have fluctuations. Also you might check the skewness of your mesh...

August 22, 2009, 06:27		#3
PrandtlGlawert New Member Daniel LLanes Join Date: Aug 2009 Posts: 11 Rep Power: 16	It´s fluctuating around its third-fourth significative number, whose order is 10x^-3. Enough for my Cl, Cd and Cm (i think...) Ty m8!!

August 22, 2009, 06:32		#4
Giuki New Member Giuki Cael Join Date: May 2009 Location: Belgium/The Netherlands Posts: 13 Rep Power: 16	Is it the residual which is fluctuating or your Cl, Cd you are monitoring?

August 22, 2009, 19:40		#5
PrandtlGlawert New Member Daniel LLanes Join Date: Aug 2009 Posts: 11 Rep Power: 16	Here is an example of the residuals. Almost constant, as you can see... Fluctuation of Cd, Cl and Cm is what expected (anything but weird ), peak-valley-peak-valley-peak-valley and so on... Thank you m8!!

August 22, 2009, 21:41		#6
tstorm Member Join Date: Apr 2009 Posts: 78 Rep Power: 17	In my experience residuals on the order of 1e-5 are small enough to consider it converged. I would consider the case converged in your case, assuming the amplitude of the oscillations aren't too large. That said, I've seen residuals less than 1e-5 for Cl for a circulation-control airfoil and yet the magnitude of oscillations for Cl were on the order of 1 (for an airfoil with an expected Cl of around 6.) In that case, the problem turned out to be three-fold: highly skewed cells in areas of high gradients, too large of cells in these areas, and large area changes. Have you tried grid adaption based on gradients? Another thing to check would be the under-relaxation factors and courant number. Try increasing all of those to see if it either causes divergence or convergence.

August 23, 2009, 06:13		#7
PrandtlGlawert New Member Daniel LLanes Join Date: Aug 2009 Posts: 11 Rep Power: 16	Ty m8s, this has been (is being) useful for me Cheers!!