Near compressible fluid - denstiy-based solver

Hooman · December 15, 2011, 12:22

Hi,
My problem involves the numerical analysis of a fluid (liquid) with low compressibility. I have varied the density through out my formulations. From how my equations were formed (with the variable density) I naturally used a denistiy-based solver. However I have just realized that for near compressible flows (low Mach number) a pressure-based solver is more efficient. I have found a few references as well. However my problem is also isothermal. I was wondering if anyone knew of any references where an isothermal near compressible flow is being solved using a density-based solver. Also if you knew any example of situations where this might occur, please let me know.
Regards,
Hooman

swetkyz · December 15, 2011, 12:56

Hooman,

If your fluid is only slightly compressible, a pressure based solver will still be applicable. You may also be able to solve the flow as incompressible since your fluid is a liquid and it is at a low speed.

Hooman · December 15, 2011, 16:08

yeah I know that. I was wondering if you knew of any isothermal examples?

Martin Hegedus · December 16, 2011, 23:37

Wouldn't the equation of state for the liquid state that if you have constant density and temperature, then the pressure would also be constant?

Hooman · December 17, 2011, 11:59

I'm using a density-based method, so the density varies.

Martin Hegedus · December 17, 2011, 14:48

Hi Hooman

Just to be clear, your flow is low speed isothermal. I'm not sure of the equation of state of your liquid, but I'm going to assume that it has the same form as the ideal gas law, i.e. p=k*rho*T. If this is the case, then p/p_inf=rho/rho_inf. As the Mach number decreases, and your pressure and density flucuations decrease, your system becomes stiff. This is a bigger issue with the right hand side than, I believe, how one solves it. Yes, solving a low Mach number problem may take more time. That's an issue with the solver. Unfortunately, the answer you converge on could be bad. That's an issue with the right hand side. Two things you can do. First, use Mach number preconditioning. Or, if your range of Mach numbers is small, increase your Mach number slightly. A Mach number of 0.001 and 0.01 or even 0.1 should produce similar answers, as long as the max Mach number in the field remains below 0.3 to 0.4.

I'm not aware of any isothermal examples. This is not to say they don't exist. It's just not my area of experience.

Hooman · December 17, 2011, 15:16

Thanks very much Martin.

Martin Hegedus · December 17, 2011, 21:04

Oh, when I said things would be similar, I meant cp and V/V_inf. Sorry. To get your true pressure, you'll need to take the cp value and multiply it by your true velocity, not the scaled up velocity resulting from the scaled up Mach number. To get your true density you'll need to use your equation of state. Of course I don't know your exact problem, so be careful with this idea of scaling up Mach number.

vicarious · December 19, 2011, 04:03

Dear Hooman,
if you are solving a low Mach number flow ( approximately lower than 0.3) you better consider incompressible and use pressure based solver, becuase the density based is not a high speed solver and one of the important issue caused by density based is the convergency, it means that your residuals may not drop and oscillate constantly.
I hope this would help. and by the way where are you from?
best regards.

December 15, 2011, 12:22	Near compressible fluid - denstiy-based solver	#1
Hooman Senior Member Join Date: Jun 2010 Posts: 111 Rep Power: 16	Hi, My problem involves the numerical analysis of a fluid (liquid) with low compressibility. I have varied the density through out my formulations. From how my equations were formed (with the variable density) I naturally used a denistiy-based solver. However I have just realized that for near compressible flows (low Mach number) a pressure-based solver is more efficient. I have found a few references as well. However my problem is also isothermal. I was wondering if anyone knew of any references where an isothermal near compressible flow is being solved using a density-based solver. Also if you knew any example of situations where this might occur, please let me know. Regards, Hooman

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Pressure based and Density based Solver	Xobile	Main CFD Forum	39	August 19, 2020 07:04
Using pressure based solver for compressible flows	aamer	Main CFD Forum	2	June 15, 2011 13:45
Problem Interface Solid Fluid with wall velocity Solver v12	hills1	CFX	2	October 12, 2009 06:36
pressure based solver	fadhil	FLUENT	2	May 28, 2008 07:29
interfacing a fluid solver with abaqus	Tuhin Rakshit	Main CFD Forum	0	June 16, 2005 11:03

December 15, 2011, 12:56		#2
swetkyz Member Zachariah Swetky Join Date: Aug 2011 Posts: 36 Rep Power: 15	Hooman, If your fluid is only slightly compressible, a pressure based solver will still be applicable. You may also be able to solve the flow as incompressible since your fluid is a liquid and it is at a low speed.

December 15, 2011, 16:08		#3
Hooman Senior Member Join Date: Jun 2010 Posts: 111 Rep Power: 16	yeah I know that. I was wondering if you knew of any isothermal examples?

December 16, 2011, 23:37		#4
Martin Hegedus Senior Member Martin Hegedus Join Date: Feb 2011 Posts: 500 Rep Power: 19	Wouldn't the equation of state for the liquid state that if you have constant density and temperature, then the pressure would also be constant?

December 17, 2011, 11:59		#5
Hooman Senior Member Join Date: Jun 2010 Posts: 111 Rep Power: 16	I'm using a density-based method, so the density varies.

December 17, 2011, 14:48		#6
Martin Hegedus Senior Member Martin Hegedus Join Date: Feb 2011 Posts: 500 Rep Power: 19	Hi Hooman Just to be clear, your flow is low speed isothermal. I'm not sure of the equation of state of your liquid, but I'm going to assume that it has the same form as the ideal gas law, i.e. p=krhoT. If this is the case, then p/p_inf=rho/rho_inf. As the Mach number decreases, and your pressure and density flucuations decrease, your system becomes stiff. This is a bigger issue with the right hand side than, I believe, how one solves it. Yes, solving a low Mach number problem may take more time. That's an issue with the solver. Unfortunately, the answer you converge on could be bad. That's an issue with the right hand side. Two things you can do. First, use Mach number preconditioning. Or, if your range of Mach numbers is small, increase your Mach number slightly. A Mach number of 0.001 and 0.01 or even 0.1 should produce similar answers, as long as the max Mach number in the field remains below 0.3 to 0.4. I'm not aware of any isothermal examples. This is not to say they don't exist. It's just not my area of experience.

December 17, 2011, 15:16		#7
Hooman Senior Member Join Date: Jun 2010 Posts: 111 Rep Power: 16	Thanks very much Martin.

December 17, 2011, 21:04		#8
Martin Hegedus Senior Member Martin Hegedus Join Date: Feb 2011 Posts: 500 Rep Power: 19	Oh, when I said things would be similar, I meant cp and V/V_inf. Sorry. To get your true pressure, you'll need to take the cp value and multiply it by your true velocity, not the scaled up velocity resulting from the scaled up Mach number. To get your true density you'll need to use your equation of state. Of course I don't know your exact problem, so be careful with this idea of scaling up Mach number.

December 19, 2011, 04:03		#9
vicarious Member Pedram Mojtabavi Join Date: Apr 2011 Location: Iran Posts: 66 Rep Power: 15	Dear Hooman, if you are solving a low Mach number flow ( approximately lower than 0.3) you better consider incompressible and use pressure based solver, becuase the density based is not a high speed solver and one of the important issue caused by density based is the convergency, it means that your residuals may not drop and oscillate constantly. I hope this would help. and by the way where are you from? best regards.