Problems with y+ and residuals

KimmyMinny · June 9, 2022, 12:53

Hi!

I am a student and I am writing a graduate thesis on the topic: resistance of the KCS ship. I'm new to CFD and i don't have much experience.
When i run the simulation i notice huge y+ numbers on the upper part of the hull while at the lower part they seem good, except at the end of the rudder.
Is it the fault of the bad mesh or is it something else?

I'm posting a picture so you can see my problems.

Thank you very much!

koful · June 10, 2022, 10:28

Hi Kimmy,

Your are looking the y+ for all ship surface. You can scene it for under water part of the ship only. To do that you need to create a treshold from derived part, and you need to choose ship surface as a part, volume fraction of air as function and below min(0.5) as a treshold.

by the way, the residual image that you share is from very beginning of the simulation. It is possible to change at later stages of the simulations.

There is a tutorial for KCS geometry at the documentation of starccm. Did you checked that?

fluid23 · June 13, 2022, 08:29

It is impossible to determine if your issues are mesh-related without you sharing some details of your mesh. Maybe post a few images?

Generally speaking, y+ can be affected by many different things including mesh, but also density, viscosity and skin friction. You could see a steep gradient in y+ near the waterline where your fluid properties change abruptly, or a poorly constructed prism mesh might be shedding layers as it approaches the corner causing the wall distance to increase. The magnitudes of y+ you show are definitely suspect, but they also depend somewhat on your selection of turbulence and wall model.

As for your residuals, they don't look great, but koful is correct. You haven't run this solution nearly long enough to judge convergence. You will also likely find this to not be a steady-state problem, which can also play tricks with your residuals.

Long story short, if you can provide more information, you will likely get better answers.

I believe there are some excellent tutorial examples on ship hull resistance (at least one anyway). If you haven't worked through it, I highly suggest starting there before attempting your own model.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Cannot find functionObject file residuals	vava10	OpenFOAM Pre-Processing	2	November 14, 2020 12:21
Problems with residuals oscillations	spacecraft	STAR-CCM+	2	September 13, 2010 08:57
problems with residuals in fluent.	Tan	FLUENT	3	February 10, 2009 10:05
Problems with residuals	Martin	CFX	2	January 2, 2007 22:10
Problems with residuals??	Peter Kolb	FLUENT	0	November 29, 2000 03:48

June 10, 2022, 10:28		#2
koful Member Join Date: Jun 2014 Location: Turkey Posts: 43 Rep Power: 11	Hi Kimmy, Your are looking the y+ for all ship surface. You can scene it for under water part of the ship only. To do that you need to create a treshold from derived part, and you need to choose ship surface as a part, volume fraction of air as function and below min(0.5) as a treshold. by the way, the residual image that you share is from very beginning of the simulation. It is possible to change at later stages of the simulations. There is a tutorial for KCS geometry at the documentation of starccm. Did you checked that?

June 13, 2022, 08:29		#3
fluid23 Senior Member Matt Join Date: Aug 2014 Posts: 947 Rep Power: 17	It is impossible to determine if your issues are mesh-related without you sharing some details of your mesh. Maybe post a few images? Generally speaking, y+ can be affected by many different things including mesh, but also density, viscosity and skin friction. You could see a steep gradient in y+ near the waterline where your fluid properties change abruptly, or a poorly constructed prism mesh might be shedding layers as it approaches the corner causing the wall distance to increase. The magnitudes of y+ you show are definitely suspect, but they also depend somewhat on your selection of turbulence and wall model. As for your residuals, they don't look great, but koful is correct. You haven't run this solution nearly long enough to judge convergence. You will also likely find this to not be a steady-state problem, which can also play tricks with your residuals. Long story short, if you can provide more information, you will likely get better answers. I believe there are some excellent tutorial examples on ship hull resistance (at least one anyway). If you haven't worked through it, I highly suggest starting there before attempting your own model.