Question to Residuals

Dexonin · June 9, 2019, 07:01

Hey Guys,
I just finished my first real simulation with an amount of 6541099 cells and 500 iterations. To judge how good my simulation is, I wanted to have a look at the residual plot. And after seeing this my question is, why is there so much ups and downs. My lowest residual was 0.006, which is quiet high i belive.

https://gyazo.com/79ea9332f4ae41a43f3278413a003512

So i switched the normalization option from "Auto" to "Off", as told in another thread. Now I have much better residuals, as shown here:

https://gyazo.com/944f64e450d9efb5ba88713a187dd4fe

So my question is, what exactly does this normalization option do? And is my simulation "good"?

Thanks in advance

flotus1 · June 9, 2019, 13:37

Look at the graphs again and squint a bit

They are exactly the same, just with an offset to each line. So none of them is better than the other. The CCM+ manual explains quite well how residuals are normalized. IIRC they use the maximum value over the first 5 iterations for normalization.

If you want to judge how good your simulation result is in terms of iteration error, at the very least you should be monitoring the quantities you are interested in. Looking exclusively at residuals will mislead you.
There should be a lengthy thread about this topic buried somewhere in the forum: Convergence

June 9, 2019, 07:01	Question to Residuals	#1
Dexonin New Member Daniel Join Date: May 2019 Posts: 7 Rep Power: 7	Hey Guys, I just finished my first real simulation with an amount of 6541099 cells and 500 iterations. To judge how good my simulation is, I wanted to have a look at the residual plot. And after seeing this my question is, why is there so much ups and downs. My lowest residual was 0.006, which is quiet high i belive. https://gyazo.com/79ea9332f4ae41a43f3278413a003512 So i switched the normalization option from "Auto" to "Off", as told in another thread. Now I have much better residuals, as shown here: https://gyazo.com/944f64e450d9efb5ba88713a187dd4fe So my question is, what exactly does this normalization option do? And is my simulation "good"? Thanks in advance

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
How to stop iterating when residuals fall down enough	CRI_CFD	OpenFOAM Running, Solving & CFD	4	December 15, 2014 07:11
Unanswered question	niklas	OpenFOAM	2	July 31, 2013 17:03
Strange residuals of the Density Based Solver	Pat84	FLUENT	0	October 22, 2012 16:59
A question about the residuals	matteo_gautero	OpenFOAM Running, Solving & CFD	3	March 14, 2008 10:41
question	K.L.Huang	Siemens	1	March 29, 2000 05:57

June 9, 2019, 13:37		#2
flotus1 Super Moderator Alex Join Date: Jun 2012 Location: Germany Posts: 3,427 Rep Power: 49	Look at the graphs again and squint a bit They are exactly the same, just with an offset to each line. So none of them is better than the other. The CCM+ manual explains quite well how residuals are normalized. IIRC they use the maximum value over the first 5 iterations for normalization. If you want to judge how good your simulation result is in terms of iteration error, at the very least you should be monitoring the quantities you are interested in. Looking exclusively at residuals will mislead you. There should be a lengthy thread about this topic buried somewhere in the forum: Convergence