Density and Pressure based solver

Diger · May 1, 2018, 09:41

How is it possible, that with some flows the pressure based solver seems to struggle, but the density based solver converges and with other flows it's vice versa?

I'm running a very simple model at the moment in which I just have a 1m x 1m x 1m box and air is flowing in at the one side and exitting on the other with high velocity ~200m/s. In the middle I just have a pylon and want to see how the flow develops around this.

I also noticed that the energy-residuals are much larger in general using the density based solver compared to the energy-residuals using the pressure based solver.

What is the reason for this?

LuckyTran · May 2, 2018, 00:51

The density based solver tends to be unstable for low Mach number flows. In Fluent and many other software, there is preconditioning to stabilize it. On the other hand, the pressure-based solver struggles with compressible flows. Both are usable and both can give accurate results, but sometimes one works better than the other when you want to arrive at a result in a particular (e.g. fast) way.

In the density based solver, the equations are coupled and you solve them together. For example, the momentum and energy equation. For compressible flows there's fairly strong coupling between energy and momentum due to compressibility and density based solver shines here.

In the pressure based solver, two things should be noted. You segregate the equations and decouple the momentum and energy equations (which is why the pressure-based solver is not as good as solving compressible flows). But instead of iterating between the continuity and momentum equation to solve for the pressure & velocity, you solve for velocity with an assumed pressure and then you solve a pressure correction equation (this part makes it very cheap to solve for pressure & velocity).

As for why the residuals are so different. We say that we solve a set of equations but never really discuss what actual algorithm is used to solve it. That matters.

May 1, 2018, 09:41	Density and Pressure based solver	#1
Diger Senior Member Join Date: Nov 2015 Posts: 135 Rep Power: 11	How is it possible, that with some flows the pressure based solver seems to struggle, but the density based solver converges and with other flows it's vice versa? I'm running a very simple model at the moment in which I just have a 1m x 1m x 1m box and air is flowing in at the one side and exitting on the other with high velocity ~200m/s. In the middle I just have a pylon and want to see how the flow develops around this. I also noticed that the energy-residuals are much larger in general using the density based solver compared to the energy-residuals using the pressure based solver. What is the reason for this?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Density or Pressure based solver	Achu	FLUENT	1	April 27, 2018 05:34
External aerodynamics of an aircraft: pressure based solver or density based solver?	shereez234	FLUENT	2	April 30, 2015 16:30
Pressure based vs density based solver	newbie384	FLUENT	2	July 5, 2013 03:04
Pressure based and Density based Solver	Xobile	Siemens	1	November 30, 2004 22:13
Warning 097-	AB	Siemens	6	November 15, 2004 05:41

May 2, 2018, 00:51		#2
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,761 Rep Power: 66	The density based solver tends to be unstable for low Mach number flows. In Fluent and many other software, there is preconditioning to stabilize it. On the other hand, the pressure-based solver struggles with compressible flows. Both are usable and both can give accurate results, but sometimes one works better than the other when you want to arrive at a result in a particular (e.g. fast) way. In the density based solver, the equations are coupled and you solve them together. For example, the momentum and energy equation. For compressible flows there's fairly strong coupling between energy and momentum due to compressibility and density based solver shines here. In the pressure based solver, two things should be noted. You segregate the equations and decouple the momentum and energy equations (which is why the pressure-based solver is not as good as solving compressible flows). But instead of iterating between the continuity and momentum equation to solve for the pressure & velocity, you solve for velocity with an assumed pressure and then you solve a pressure correction equation (this part makes it very cheap to solve for pressure & velocity). As for why the residuals are so different. We say that we solve a set of equations but never really discuss what actual algorithm is used to solve it. That matters.