Numerical Towing Tank Velocity Ramp

lava12005 · July 30, 2012, 23:52

Hi all,

I am simulating a towing tank and would like to do a simulation where the model start from rest (0m/s) to a test speed, say 1m/s.
I would like to know on how to ramp up the velocity in using the VOF model?

I tried to do a simple simulation where there is no model at all, and using the boundary condition as those specified in the DFBI Boat tutorial. Except that I change the Velocity Magnitude boundary condition using a field function of $Time

I am using dt of 0.01s, so after one time step, I expect the speed in the whole domain should be 0.01 m/s. But it is not and even after 100 iteration it is not converging to 0.01m/s

Does anyone can help me in this?

sanjay · July 31, 2012, 03:17

Try to run for more iterations, 100 is too less.
I had carried out similar towing tank experiments for flow visualization, but didn't try numerical simulation of it. I am looking forward do the same

regards
www.aerosapien.blogspot.com

lava12005 · July 31, 2012, 03:29

Hi Sanjay, that 100 iterations is still within the 1 time step. Isn't is considered to be excessive already for just 1 time step?

Does it has something to do with the boundary condition? Maybe the hydrostatic pressure outlet doesn't really allow this? Any suggestion?

Sideshore · July 31, 2012, 04:12

I've done this in the past.

I've specified a velocity ramp at inlet boundary. And you need to add momentum to the fluid in the domain while accelerating the fluid.

Good luck!

ryancoe · July 31, 2012, 09:27

Sideshore is correct. You need to prescribe a motion to your body (and the domain surrounding it) not the flow at the inlet. Your setup is more akin to water tunnel increasing speed, than a carriage doing so.

You should be able to use a field function to set the time varying velocity of the body.

lava12005 · August 17, 2012, 00:27

Hi all,

Yes, thanks to the suggestion I manage to ramp the velocity on the domain (just a water channel without any object) by using the User Defined Vertex, grid velocity method and using the user field function to setup the time varying motion.

Now If I want to have an object being ramp by the carriage (free to heave and trim) is this still possible?
Normally for fixed speed, I create an Overset mesh and prescribed a DFBI rotation and translation to it, but I don't know how to ramp the velocity in this region (since we have prescribed the motion as DFBI).

Henry Arrigo · September 5, 2012, 06:13

You can start with a velocity namely 0.1 m/s and run the case until it converges. then stop the simulation and change the velocity to namely 0.2 m/s and then run the case again. try to repeat these sequences until you reach the 1 m/s.
or if you have any idea about the needed power at 1 m/s you can use the thrust power instead of velocity.

lava12005 · September 14, 2012, 11:15

I finally able to do it via adding gravity in the flow direction (according to the ramping acc). Anyway just asking, if say we are using momentum source option (constant) to accelerate the fluid, but since there is 2 diff fluid here (air and water) doesn't it makes the air is accelerated more than the water?

Sideshore · September 17, 2012, 03:03

You have to make the force you apply dependent on the mass you want to accelerate.

I think in Star-ccm+ you define a volumetric momentum force. So the force applied to each cell does not automatically depend on the mass in the cell. You can use the volume fraction to calculate the mass in the cell.

July 30, 2012, 23:52	Numerical Towing Tank Velocity Ramp	#1
lava12005 Senior Member KHB Join Date: Aug 2010 Location: Singapore Posts: 118 Rep Power: 15	Hi all, I am simulating a towing tank and would like to do a simulation where the model start from rest (0m/s) to a test speed, say 1m/s. I would like to know on how to ramp up the velocity in using the VOF model? I tried to do a simple simulation where there is no model at all, and using the boundary condition as those specified in the DFBI Boat tutorial. Except that I change the Velocity Magnitude boundary condition using a field function of $Time I am using dt of 0.01s, so after one time step, I expect the speed in the whole domain should be 0.01 m/s. But it is not and even after 100 iteration it is not converging to 0.01m/s Does anyone can help me in this?

July 31, 2012, 09:27		#5
ryancoe Member Ryan Coe Join Date: Jun 2010 Location: Albuquerque, NM Posts: 98 Rep Power: 15	Sideshore is correct. You need to prescribe a motion to your body (and the domain surrounding it) not the flow at the inlet. Your setup is more akin to water tunnel increasing speed, than a carriage doing so. You should be able to use a field function to set the time varying velocity of the body. __________________ Ryan

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Difficulty in calculating angular velocity of Savonius turbine simulation	alfaruk	CFX	14	March 17, 2017 06:08
Velocity profile disturbance due to loss coefficient	rks171	Main CFD Forum	3	May 25, 2012 17:30
Numerical wave tank	Bridget	FLUENT	0	March 27, 2006 16:09
Numerical Method	Andrew Hayes	Main CFD Forum	5	June 17, 2005 09:17
Convection velocity of Coherent structures	Jongdae Kim	Main CFD Forum	3	February 5, 2002 04:04

July 31, 2012, 03:17		#2
sanjay Member aerosapien Join Date: Sep 2010 Posts: 59 Rep Power: 15	Try to run for more iterations, 100 is too less. I had carried out similar towing tank experiments for flow visualization, but didn't try numerical simulation of it. I am looking forward do the same regards www.aerosapien.blogspot.com

July 31, 2012, 03:29		#3
lava12005 Senior Member KHB Join Date: Aug 2010 Location: Singapore Posts: 118 Rep Power: 15	Hi Sanjay, that 100 iterations is still within the 1 time step. Isn't is considered to be excessive already for just 1 time step? Does it has something to do with the boundary condition? Maybe the hydrostatic pressure outlet doesn't really allow this? Any suggestion?

July 31, 2012, 04:12		#4
Sideshore New Member Jan Willem Krijger Join Date: May 2010 Posts: 8 Rep Power: 15	I've done this in the past. I've specified a velocity ramp at inlet boundary. And you need to add momentum to the fluid in the domain while accelerating the fluid. Good luck!

August 17, 2012, 00:27		#6
lava12005 Senior Member KHB Join Date: Aug 2010 Location: Singapore Posts: 118 Rep Power: 15	Hi all, Yes, thanks to the suggestion I manage to ramp the velocity on the domain (just a water channel without any object) by using the User Defined Vertex, grid velocity method and using the user field function to setup the time varying motion. Now If I want to have an object being ramp by the carriage (free to heave and trim) is this still possible? Normally for fixed speed, I create an Overset mesh and prescribed a DFBI rotation and translation to it, but I don't know how to ramp the velocity in this region (since we have prescribed the motion as DFBI).

September 5, 2012, 06:13		#7
Henry Arrigo Senior Member Henry Arrigo Join Date: Jun 2010 Location: Italy Posts: 100 Rep Power: 15	You can start with a velocity namely 0.1 m/s and run the case until it converges. then stop the simulation and change the velocity to namely 0.2 m/s and then run the case again. try to repeat these sequences until you reach the 1 m/s. or if you have any idea about the needed power at 1 m/s you can use the thrust power instead of velocity.

September 14, 2012, 11:15		#8
lava12005 Senior Member KHB Join Date: Aug 2010 Location: Singapore Posts: 118 Rep Power: 15	I finally able to do it via adding gravity in the flow direction (according to the ramping acc). Anyway just asking, if say we are using momentum source option (constant) to accelerate the fluid, but since there is 2 diff fluid here (air and water) doesn't it makes the air is accelerated more than the water?

September 17, 2012, 03:03		#9
Sideshore New Member Jan Willem Krijger Join Date: May 2010 Posts: 8 Rep Power: 15	You have to make the force you apply dependent on the mass you want to accelerate. I think in Star-ccm+ you define a volumetric momentum force. So the force applied to each cell does not automatically depend on the mass in the cell. You can use the volume fraction to calculate the mass in the cell.