CV by FEA simulation

[Albator]

Hello all,

I'm new on this forum, and quite new also in the valve world.

Since last july, I (try to) handle a Fluid Finite Elements Analysis software to do flow calculations through the valves I design. In particular I'd like to use it to determine CV.

But the more I try to understand, the less I do...

I found a lot of tables with CV Max values for different sizes and classes, but I'm sooooooooooooooooo far to obtain them by simulation.

To give an example, I get approximately a CV of 200 for a 2 inches ball valve, but I should have about 500.
And I have numerous cases like this.

The models I use for simulations are based on the standards ISA S75-01 and -02, so I add pipes at the inlet and outlet of my CAD model.

Then I tried with a variety of limit conditions (pressure at inlet and/or outlet, flow at inlet and/or outlet,...) but nothing seems coherent.

Maybe someone tried something similar and should help me?

I can't ask anything in particular, because I ask myself so much questions about this.

Thanks a lot in advance for the good knight(s) who maybe will help me.

I begin to lose hope...

 

[Albator]

Maybe I should add some details, in order to have some directions where to go:



The general test is the following: I add pipes at the inlet and outlet of my CAD model (2 times the ND at the inlet and 6 times the ND at the outlet, following ISA S75-02), then I apply some pressure at the inlet of the inlet pipe, and less pressure at the outlet of the outlet pipe.

Then I do the simulation with water, and I measure the volume flow rate in US gpm that I obtain at the outlet of the outlet pipe.

I also add cut planes to get the pressure at the inlet of the valve (thus the outlet of the inlet pipe if you understand me), and at the outlet of the valve (i.e. inlet of outlet pipe). So I have the pressure drop of the valve only, not of the length of pipes.

First question: do you think it is wrong?
In ISA S75-02, the pressure taps are placed at the inlet and outlet of valve+pipes. But if I do this, the CV I obtain is again diminished, because the pressure drop is greater.


Then, I divide the flow in gpm by the square root of the pressure drop in psi. The value is the simulated CV (right isn't it?).
I also measure the outlet flow in m³/h and the pressure drop in bar, then apply the same type of formula with a conversion factor, and obtain the same value.

But this value of simulated CV is far less than the values I found everywhere.


Does anyone see a big mistake in my work method?

I work on this with a collegue who as some years of experience in valve engineering, and he absolutely don't understand more than me...

 

[bcd]

For starters, you need at least 10 pipe diameters downstream of the valve in order for the flow to fully develop in the solution. You might as well put the same upstream as well. By having the "ends" of you model so close to the valve, the boundary conditions you are setting at the inlet and outlet are influencing the flow calculations inside the valve. When you run the solution with much longer pipes, you then take the 4 pressure measurements at the wall of the pipe 2 diameters upstream and 6 downstream just where the pressure tap would be. Average the numbers at each location and calculate the Cv. You should be much closer. Keep in mind that the test results are allowed to vary by as much as +/- 5%. Realistically, unless you are using a very high-end CFD package, you're doing great if you get within 10%.

bcd

 

[Albator]

Thanks a lot!

I'll try this right now, and be back to let you know how it works!