2 Replies Latest reply on Jul 20, 2011 12:42 PM by Eamon Ronan

    Flow coefficient of a gas valve - hoping Joe Galliera or similar can help

    Eamon Ronan

      Hi folks,

      This is pretty basic as as I've just started with Flow but am struggling with some fundamentals of what I need it to do.

      Basically, I'm trying to set up Flow to replicate flow coefficient tests in semiconductor gas valves, as opposed to water, although granted they're similar enough.  My baselines are experimental test results - although some may be questionable.

      I'm trying to get the CFD to match these - the CV result for say the demo ball valve will be different is done with water as opposed to a gas.

       

      So the rig, or model now that I've SW, is simply set up to create a 1 psig pressure drop across a valve by 'choking' the downstream (e.g.  60 psig inlet and 59 psig outlet) and measuring the volume flow rate at the outlet.

       

      What I'm really struggling with is firstly whether I need to 'do something' to the units/results to standardise them or change them becuase of a feature of SW Flow?

      Secondly, the issue of whether to measure the volume flow rate at the inlet or the outlet.  If I've looked at threads and examples correctly, I see some done differently - maybe because water is incompressible.

      And indeed thirdly, I just can't seem to find real explanations of some of the parameters, conditions, etc.  e.g.  I have been using Static Pressure at inlet and outlet, and just don't know if that's right because I don't really understand the differences between environmental, static and total on the set-up.  The same really for mass flow versus volume flow (that's where some issues on standardised measures really started to throw me).

       

      I running away with changing some of the above and now back-tracking through the ANSI and SEMI specs ......but any advice even on where to get the best explanations of total/static/etc and whether there are nuances in the CFD that need units to be changed - it would be a great help.

        • Re: Flow coefficient of a gas valve - hoping Joe Galliera or similar can help
          Joe Galliera

          Hi Eamon,

           

          First, I had just found your post/question while doing a search online and saw my name.  I'm floored by having me called out to help, now let's see if I can deliver.  Hopefully being two months late is not an issue (btw my email is in my user profile).

           

          To get Cv for valves using gases, the calculation is different than for liquids as you pointed out.  Primarily has to do with the relative compressibility of gases as compared to liquids.

           

          I'm going to offer two links to help you out:

          1) Cv equation for gas at both low and high pressure drops: http://www.steam-toolbox.com/flow-coefficients-d_277.html

          2) Swagelok's site has a nice PDF with the same equations and with more detail on how this is done.

           

          -Joe

            • Re: Flow coefficient of a gas valve - hoping Joe Galliera or similar can help
              Eamon Ronan

              Hi Joe,

              Hey - thanks for responding - very good of you.  I had looked at some of your blogs, including I think those 2 attachments that you've referenced before? 

              We've only recently gotten SW and it's all based on the FLOW so it's been hugely fustrating over the past 2-3 months. 

              Trying to work through the issues, we've had good local SW support and that's starting to necessarily extend out to SW-CFD experts in the US and France. 

              Basically, we were trying to set up simulations as a replication of a physical flow co-efficient valve test, and were getting the answers close to expectations on some products (old test results).  But the results of other valves were miles off and that opened a whole can of worms.  At the start, we hadn't even realised that SW was measuring volume flow rate rather than standardised conditions.  So we were really not setting up correctly but we've certainly (since my post) made some progress regarding parameters, criterion setting, manual refinement, setting up 'dummy lids' as measurement sufaces and especially adaptive meshing. 

              I thought it would be simple .....  Set an inlet of 60 psig, Outlet of 59psig, Measure a pressure drop and Measure a flow rate and then equation out to SCFM and CV.   But it didn't work out that way because of incorrect CFD set up, parameters, etc. 

              Still not there really -- for example I don't really understand some differences between static, total, dynamic pressure and why static is meant to be the appropriate one. 

              But trying some cross-checks now all we've learned against the other method of setting an inlet volume flow rate for air, and an outlet static pressure, taking measurements at surfaces 'set distances' from the test specimen (and a load of info goals just to see what's happening).  It should work out at same CV as the set inlet and outlet method.  If so, I just hope that SW support will 'rubber-stamp' the CFD set-up process in the context of the objective  -  CV test replication.  Or, maybe there will be nuances that come out for very large flow, or convoluted flows, or low pressure drops, etc - that's the concern. 

              But at this stage, there can't be much else. 

              I might post up where this ends?   - which hopefully will be in 2-3 weeks as I run some more simulations. 

              If there's anything esle specific to this gas-flow through valve type scenario that you think I should be paying particular attention to, a reply post would be great. 

              Thanks again for responding and the posts.

              Kind regards

              Eamon

              Picture1.png