[hmasenger]

Hi all
I am modeling flow in a butterfly valve which is installed in a pipe .I need to know the pressure drop accruing in the valve by variation in flow velocity.
I don’t know what value I should set for the pressure outlet BC that makes it possible to measure the pressure drop.
Is it all right if I just set the domain's reference pressure to[ 1 atm ] and set pressure for outlet BC to [ 0 atm ] which is about 20 times of pipe diameter downstream of valve and then measure the average pressure difference between locations in 5times of pipe diameter upstream and 10 times of diameter downstream of valve?


Best regards

[ghorrocks]

There is no "correct" answer for this, just many different approaches. What you have done makes sense and is easily comparable to published results. But you have also included a fair amount of pipe before and after the vlave, and this may introduce its own pressure loss. So you measured pressure loss is a combination of the valve and pipe pressure loss. And then there is the entrance condition - if you have assumed plug flow then you will have high pressure losses in the first bit as the boundary layers develop.

But these are all details and most engineers do nt worry about them or have simple ways of correcting for them. It is up to you as to whether this is important or not.

Is the simulation incompressible?

[hmasenger]

your answer was very helpful .thanks for that.
yes the flow is incompressible(water@20) and fully developed so I am considering 10% for turbulent intensity(i am not sure that's enough)
Actually the valve is a 1200mm diameter butterfly valve so I thing pipe friction head loos in that distance is negligible and the reason for taking 10D and 5D distances for pressure drop in literature is for including all eddy dissipation in measurements . I am studding cavitation and pressure drops in various water speeds.although effect of valve opening in cavitation and maybe a new disk shape for decreasing the cavitation for my msc degree.
for inlet BC i am going to use developed velocity profile published in fluid mechanic references

[ghorrocks]

That sounds good, your approach sounds sensible then. Make sure you do sensitivity analysis on all tunable variables to make sure things are under control.

[oj.bulmer]

One of the major reasons of measuring the pressure drop at those distances (5D, 10D) is to have a measure of "lost energy" in overcoming hydraulic resistance of the valve. And energy is indicative of total of static pressure (p) and dynamic pressure (1/2* density * v^2).

If you plot a graph of pressure in streamwise direction, you'll see that static pressure starts falling before the valve, it goes to minimum in wall region and then recovers. At the same time, dynamic pressure changes as well. If we take only (static) pressure drops close to the valve, these distances chosen may involve the drop and recovery of the pressure, where some pressure is still being converted into velocity or vice versa. Hence it won't be a complete representation but only a half one.

To mitigate this, you can take total energy (static or dynamic pressure) and thus have an adequate indication of energy loss across the valve. But this is a bit tedious. Or simply, let the static pressure recover, which it should typically do over 10D distance downstream. While before the valve, the pressure starts falling before a shorter length prior to valve, hence 5D is sufficient.

OJ

[hmasenger]

tanks OJ .It was really helpfull .