Transient Simulation in ANSYS CFX

ylm054 · December 2, 2009, 21:10

I am running a transient simulation where the boundary condition is a pressure that oscillates with a cosine function (Womersley flow).

At the moment, I enter boundary conditions as columns (1st column time, 2nd column static pressure) in a data file and import it into a user function.

I then call the user function from the prompt box under B.C>Static value etc.

Is there a more elegant method of doing this? In the current method, will ANSYS interpolate between BC pressure values at each time step, or not?

Thank you

ghorrocks · December 4, 2009, 05:32

You can use a CEl expression. Something like:

Pressure = 100 [Pa] * cos(t/1[s])

hesamking · June 4, 2014, 21:32

Hi,
I am trying to model the flow around a helmeted head (around and inside the head-helmet gap). for that and to simulate a blast-like flow I have used two domains, a small one at an initial high pressure of 860 kPa with an initial velocity of 500 m/s and a very large domain at ambient setting. I have defined an interface between the two domains and can get an overpressure with a similar pattern as that happens in blast.
When I plot the velocity vectors, at each time steps I can see that half of the vectors move backwards and it doesn't seem to be showing the correct behaviors.
I am really confused and I don't know if I am wrong anywhere or it is as it should be. I am doing 3D. I have attached a picture of the velocity vectors at one of the time steps plotted on the symmetry plane.
I have also attached the pressure contours at the same time step which also shows my domain.v2.jpg

m7.jpg
Could you please take a look and help me.
Thanks

ghorrocks · June 5, 2014, 21:02

This is because your left hand boundary condition is not correct.

Also - you should not use multiple domains for this. Use a single domain with an initial condition.

hesamking · June 6, 2014, 14:55

Thanks Horrocks,
Actually the reason I used two domains was that the pressure time history plot obtained with that had a pattern very similar to that of the real blast modeling done in FE software such as Ls-Dyna. I have attached the plot.
As you mentioned I already reduced my domains to one, put the domain at ambient setting (p = 1 atm and T= 25 C and initial velocity of 0 m/s) and then introduced an inlet at the left boundary and defined that as supersonic with the Total Energy option. I chose the "Cartesian Velocity and Pressure" option and defined inlet flow with V= 500 m/s and P= 860 kPa.
The velocity vectors and streamlines looks fine as expected. However, I can't get the pattern I want from the pressure history. I have attached this one too.
Could you please also take a look at the velocity vectors and tell me the reason for the small blue spots (vectors) at the back of the helmets and why we are seeing teh velocity vectors such that in there? The vectors are plotted on a plane very close to the symmetric plane on the inside.
So I appreciate your help.
Thanks
domain2.jpg

two-domain pressure history plot.jpg

single domain pressure history plot.jpg

single domain velocity vectors.jpg

ghorrocks · June 9, 2014, 20:07

I think you need to do a little work on basic compressible flow modelling. If you spend a little time doing something like shock wave flow modelling you will find it quite valuable. It has analytical answers so you have exact solutions to see how accurate you are. You will also start understanding issues of numerical accuracy for compressible flows.

Everything you are seeing looks like normal compressible flow artefacts to me. So model a basic situation with known answers to develop a simulation technique which is accurate.

Bhavadip · April 24, 2016, 10:19

i am working on shell an tube heat exchanger for that how many domain i have to draw
1. shell fluid
2. copper fluid
3. tube side fluid

is this correct??

please guide me for the transient analysis of this problem

ghorrocks · April 25, 2016, 21:05

The CFX and workbench tutorials cover the basics for a simulation like this.

December 2, 2009, 21:10	Transient Simulation in ANSYS CFX	#1
ylm054 New Member Join Date: Dec 2009 Posts: 10 Rep Power: 16	I am running a transient simulation where the boundary condition is a pressure that oscillates with a cosine function (Womersley flow). At the moment, I enter boundary conditions as columns (1st column time, 2nd column static pressure) in a data file and import it into a user function. I then call the user function from the prompt box under B.C>Static value etc. Is there a more elegant method of doing this? In the current method, will ANSYS interpolate between BC pressure values at each time step, or not? Thank you Last edited by ylm054; December 2, 2009 at 21:11. Reason: no tags

June 4, 2014, 21:32	Velocity vectors in high pressure flow simulation around head-helmet assembly	#3
hesamking Member Hesam Moghaddam Join Date: Mar 2012 Posts: 49 Rep Power: 14	Hi, I am trying to model the flow around a helmeted head (around and inside the head-helmet gap). for that and to simulate a blast-like flow I have used two domains, a small one at an initial high pressure of 860 kPa with an initial velocity of 500 m/s and a very large domain at ambient setting. I have defined an interface between the two domains and can get an overpressure with a similar pattern as that happens in blast. When I plot the velocity vectors, at each time steps I can see that half of the vectors move backwards and it doesn't seem to be showing the correct behaviors. I am really confused and I don't know if I am wrong anywhere or it is as it should be. I am doing 3D. I have attached a picture of the velocity vectors at one of the time steps plotted on the symmetry plane. I have also attached the pressure contours at the same time step which also shows my domain.v2.jpg m7.jpg Could you please take a look and help me. Thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
2D Simulations in ANSYS CFX	CFDLife	CFX	10	August 1, 2008 15:04
transient simulation	Tad	CFX	1	November 30, 2003 17:25
Transient Rotor/Stator Simulation in TASCFLOW	Deepak Ganga	CFX	3	June 27, 2003 03:47
ANSYS to acquire CFX	Fred	FLUENT	0	February 18, 2003 22:04
ANSYS to acquire CFX	Fred	Siemens	0	February 18, 2003 22:03

December 4, 2009, 05:32		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	You can use a CEl expression. Something like: Pressure = 100 [Pa] * cos(t/1[s])

June 5, 2014, 21:02		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	This is because your left hand boundary condition is not correct. Also - you should not use multiple domains for this. Use a single domain with an initial condition.

June 6, 2014, 14:55		#5
hesamking Member Hesam Moghaddam Join Date: Mar 2012 Posts: 49 Rep Power: 14	Thanks Horrocks, Actually the reason I used two domains was that the pressure time history plot obtained with that had a pattern very similar to that of the real blast modeling done in FE software such as Ls-Dyna. I have attached the plot. As you mentioned I already reduced my domains to one, put the domain at ambient setting (p = 1 atm and T= 25 C and initial velocity of 0 m/s) and then introduced an inlet at the left boundary and defined that as supersonic with the Total Energy option. I chose the "Cartesian Velocity and Pressure" option and defined inlet flow with V= 500 m/s and P= 860 kPa. The velocity vectors and streamlines looks fine as expected. However, I can't get the pattern I want from the pressure history. I have attached this one too. Could you please also take a look at the velocity vectors and tell me the reason for the small blue spots (vectors) at the back of the helmets and why we are seeing teh velocity vectors such that in there? The vectors are plotted on a plane very close to the symmetric plane on the inside. So I appreciate your help. Thanks domain2.jpg two-domain pressure history plot.jpg single domain pressure history plot.jpg single domain velocity vectors.jpg

June 9, 2014, 20:07		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	I think you need to do a little work on basic compressible flow modelling. If you spend a little time doing something like shock wave flow modelling you will find it quite valuable. It has analytical answers so you have exact solutions to see how accurate you are. You will also start understanding issues of numerical accuracy for compressible flows. Everything you are seeing looks like normal compressible flow artefacts to me. So model a basic situation with known answers to develop a simulation technique which is accurate.

April 25, 2016, 21:05		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	The CFX and workbench tutorials cover the basics for a simulation like this.