valve flow simulation

quiqui · June 29, 2014, 23:51

Hi,
I am simulating a valve moving, watching the delta_p between inlet and outlet.
Inlet is constant mass flow rate (water), outlet is constant static pressure.
inlet and outlet are extended to have fully developed flow.

As the valve moves, the inlet pressure is changing.
At close position, delta_p from simulation is 450 psi.

But the experiment only gave 250 psi.

we don't know where the problem is. Any body have any idea?

thanks

Boris_M · June 30, 2014, 08:26

Can you explain what the closed position is?
Is it fully closed valve without any flow going through?
Then the error would be in the boundary condition. In your measurements you will not be able to fix a mass flow rate as nothing is flowing. If you check your pump curve you will see that at very high pressure the volume or mass flow rate is 0. to the difference you would probably measure is the value at that curve in your pump curve. So the maximum pressure difference I wuld say.
In general the cosed case is no flow at all so you don't really nead CFD is there is not flow there is nothing to calculate.

Does this help?
Boris

quiqui · June 30, 2014, 15:43

sorry I didnot clarify the close position, the "close" position means partially closed, there is still flow going through.

In the experiment, as the valve is in the closing process, pump cannot maitain the constant mass flow rate, but it should not cause so big difference for the pressure.

Boris_M · July 1, 2014, 02:41

Are you sure that it shouldn't make a big difference?
How much is the mass flow rate the pump is delivering compared to what you set?
What is your pressure reading, not the delta but the actual pressure for inlet and outlet?
Did you get any warning in the solver monitor about negative pressure?

Reason I'm asking is if you use a liquid you might get very low pressure (in fact negative pressure) at a high mass flow rate and a narrow opening which would indicate cavitation is appearing.

Can you show the mesh in the gap of your model?
At such small gaps the mesh is very important to resolve the flow and get the pressure loss right.

Boris

quiqui · July 1, 2014, 12:46

for simulation, at inlet, I put 500 gpm water, outlet pressure is 1000 psi.
got inlet pressure 1420psi at partially close position.

experiments, inlet 500 gpm ~485 gpm, outlet pressure is 1000 psi. the inlet pressure is below 250 psi.

we checked the geometry, viscosity, kept them the same.

I reran the simulation using 485 gpm, but got inlet 396 psi, which is still much higher than the test data.
In the simulation, all the pressure is above 1000 psi. no caivation appears.

the partially close position is less than half of the fully closed. so there is enough gap for the water to go though.
thanks

Boris_M · July 2, 2014, 03:21

In order to see if your simulation results resolve the flow filed correctly you can try to use a cut plot and deactivate the interpolation and activate the mesh with the pressure and velocity as parameters. This will show how big the jumps are between the values in the cells.
For example if you have a finde mesh in the narrow gap of the valve opening that might be good resolved then it still doesn't mean that the flow behind is good resolved which can propagate and cause large deviations. So if you have large jumps between to neighboring cells such as from red or orange to blue on your color scale, then you should refine the mesh. For anything that is to coarse in the flow field but you don't know where the high gradients are you should use the adaptive mesh revinement during the solver is running.

But I cannot help you more without any images of the mesh and the non-interpolated plots to tell if that might really the case. I understand that you might not be able to share it here but then you will have to contact your support.

Boris

will321321 · July 7, 2014, 22:05

Hi,

I am not an expert on CFD but I would try to help, Correct me if im wrong somewhere, so on the experiment the pump is on constant pressure at constant speed so it reads 250 always on the inlet it doesn't change the variable is the pressure at the Outlet so you get a specific pressure drop at that mass flow which when constricted mass flow at the outlet becomes less.

Now for the simulation you are putting 500gpm in the inlet which, in my understanding, the software try to deliver at the same mass flow to the outlet thus the simulation compensates pressure in the inlet in order to deliver mass flow at a certain resistance.

I suggest using the measured mass flow at the Outlet and measured static pressure at inlet.

Hope I helped Thx.

June 29, 2014, 23:51	valve flow simulation	#1
quiqui New Member Join Date: May 2013 Posts: 20 Rep Power: 12	Hi, I am simulating a valve moving, watching the delta_p between inlet and outlet. Inlet is constant mass flow rate (water), outlet is constant static pressure. inlet and outlet are extended to have fully developed flow. As the valve moves, the inlet pressure is changing. At close position, delta_p from simulation is 450 psi. But the experiment only gave 250 psi. we don't know where the problem is. Any body have any idea? thanks

July 1, 2014, 12:46		#5
quiqui New Member Join Date: May 2013 Posts: 20 Rep Power: 12	for simulation, at inlet, I put 500 gpm water, outlet pressure is 1000 psi. got inlet pressure 1420psi at partially close position. experiments, inlet 500 gpm ~485 gpm, outlet pressure is 1000 psi. the inlet pressure is below 250 psi. we checked the geometry, viscosity, kept them the same. I reran the simulation using 485 gpm, but got inlet 396 psi, which is still much higher than the test data. In the simulation, all the pressure is above 1000 psi. no caivation appears. the partially close position is less than half of the fully closed. so there is enough gap for the water to go though. thanks Last edited by quiqui; July 1, 2014 at 13:52.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Control Valve Flow Simulation	Ashraf Sharara	FLUENT	1	September 3, 2017 09:58
Ball valve flow in solidworks	nikesh	FloEFD, FloWorks & FloTHERM	10	December 10, 2013 08:40
Valve simulation with spring - FSI? Help!	farianka	CFX	1	April 17, 2011 18:04
Divergence detected in Valve simulation	mehn	STAR-CCM+	5	July 23, 2009 08:38
oscillating valve flow	Matthias Finke	Main CFD Forum	2	September 10, 2003 13:59

June 30, 2014, 08:26		#2
Boris_M Disabled Join Date: Jul 2009 Posts: 616 Rep Power: 23	Can you explain what the closed position is? Is it fully closed valve without any flow going through? Then the error would be in the boundary condition. In your measurements you will not be able to fix a mass flow rate as nothing is flowing. If you check your pump curve you will see that at very high pressure the volume or mass flow rate is 0. to the difference you would probably measure is the value at that curve in your pump curve. So the maximum pressure difference I wuld say. In general the cosed case is no flow at all so you don't really nead CFD is there is not flow there is nothing to calculate. Does this help? Boris

June 30, 2014, 15:43		#3
quiqui New Member Join Date: May 2013 Posts: 20 Rep Power: 12	sorry I didnot clarify the close position, the "close" position means partially closed, there is still flow going through. In the experiment, as the valve is in the closing process, pump cannot maitain the constant mass flow rate, but it should not cause so big difference for the pressure.

July 1, 2014, 02:41		#4
Boris_M Disabled Join Date: Jul 2009 Posts: 616 Rep Power: 23	Are you sure that it shouldn't make a big difference? How much is the mass flow rate the pump is delivering compared to what you set? What is your pressure reading, not the delta but the actual pressure for inlet and outlet? Did you get any warning in the solver monitor about negative pressure? Reason I'm asking is if you use a liquid you might get very low pressure (in fact negative pressure) at a high mass flow rate and a narrow opening which would indicate cavitation is appearing. Can you show the mesh in the gap of your model? At such small gaps the mesh is very important to resolve the flow and get the pressure loss right. Boris

July 2, 2014, 03:21		#6
Boris_M Disabled Join Date: Jul 2009 Posts: 616 Rep Power: 23	In order to see if your simulation results resolve the flow filed correctly you can try to use a cut plot and deactivate the interpolation and activate the mesh with the pressure and velocity as parameters. This will show how big the jumps are between the values in the cells. For example if you have a finde mesh in the narrow gap of the valve opening that might be good resolved then it still doesn't mean that the flow behind is good resolved which can propagate and cause large deviations. So if you have large jumps between to neighboring cells such as from red or orange to blue on your color scale, then you should refine the mesh. For anything that is to coarse in the flow field but you don't know where the high gradients are you should use the adaptive mesh revinement during the solver is running. But I cannot help you more without any images of the mesh and the non-interpolated plots to tell if that might really the case. I understand that you might not be able to share it here but then you will have to contact your support. Boris

July 7, 2014, 22:05		#7
will321321 New Member Join Date: Apr 2014 Posts: 29 Rep Power: 12	Hi, I am not an expert on CFD but I would try to help, Correct me if im wrong somewhere, so on the experiment the pump is on constant pressure at constant speed so it reads 250 always on the inlet it doesn't change the variable is the pressure at the Outlet so you get a specific pressure drop at that mass flow which when constricted mass flow at the outlet becomes less. Now for the simulation you are putting 500gpm in the inlet which, in my understanding, the software try to deliver at the same mass flow to the outlet thus the simulation compensates pressure in the inlet in order to deliver mass flow at a certain resistance. I suggest using the measured mass flow at the Outlet and measured static pressure at inlet. Hope I helped Thx.