Pressure drop through Tee/ branch fitting

Regarding pressure drop for flow through a tee the pressure drop factors given in Crane, Flow of Fluids is K=20fT for flow through run and K=60fT for branch.

I require clarifications on the following:

1)Is there any limitation on tee size for using the above. Say if I have a 10 inch. pipe with a branch of 1" dia. would I be right in using the K factors as above.

2) If there is a flow division between run and branch, how do I compute the pressure drop? What are the velocities to be used for run and branch?

Thanks in advance for your feedback.

The Crane L/D figures are intended to be used for line sized tees, not reducing tees.

If you had a 1" tee coming off a 10" line, I'd treat it as an entrance loss rather than as a tee.  When the change in size isn't so extreme, say a 6" by 6" by 4" tee, I've used the 60 L/D figure for a 4" tee as a conservative estimate of the pressure losses.  Unless this is a significant loss in your system resistance, the effect will be minor.

For estimating losses when you have splits in direction, my usual reference is "Mechanical Design of Process Systems", 2nd edition, Volume 1 "Piping and Pressure Vessels" by A. Keith Roscoe.

TD2K:

Thanks for your reply.For the first part I have been following the approach you have indicated.

You mention the change in size is not so extreme. Can you quantify this?
In the case of flow through run of 10" line (which has 1" branch), I presume the pressure drop is neglected and not calculated using 20fT.

With best regards

I can't quantify it easily as it depends on the area ratio AND the split of the flow between the branch and the run so it's a bit of an interative solution.  

You would be best off trying to get access to a copy of this reference and running the numbers.  The example you are looking for is Figure 1-10K, post your email and I can scan and send you a copy of this page.

TDK:

Thanks and would appreciate your scanning and sending me a copy of the page referred in your reply. My email address is batraarun@hotmail.com

I have a similar query - I have flow through a 1.5" line and branch of 1" but to make matters more complicated, there is a 0.8" orifice at the beginning of the branch.

     |                | < pipe
     |                |
     |                |___________
     |                |                  < branch
     |                  < orifice
     |                |___________
     |                |
     |                |
     |                |

(hope this makes it clearer!)

Can I just 'ignore' the orifice and if so, should I treat it as an entrance loss or a branch loss?

Regards,

James

Hmm, interesting question.  Haven't actually had to rationalize what I'd do for this case.  Unfortunately, I don't think I have any references that deal with 'fitting to fitting' losses other than the discussion of losses through a 180 return bend versus 2, 90 degree elbows that Crane goes through.

I definitely would NOT ignore the orifice, it's just the combination of the orifice AND the tee that I'm not sure how to handle (rather than ignore the orifice, I'd likely ignore the 'flow through branch' tee first).

While I'm thinking, let me lob the ball back into your court James?  Is the purpose to try and estimate the maximum flow through the system (eg, you want a conservatively 'low' figure for the resistance) or the minimum flow you can expect to get (in which case, you want a conservatively high flow)?

Thanks for the reply.  The junction in question is part of an existing design for a ventilation system for animal cages.  I have been asked to look at ways of improving the system in general so that it is more efficient.  I hope that by finding the head loss in the existing system I can show how much more efficient my new design would be.  So, in answer to your question, I am looking for a conservatively 'high' figure for the resistance.

I have been thinking about the streamlines through the juntion and how I could break the junction down into seperate resistances.  What do you think of the idea of treating it as flow through a 1.5" line (as before) but a 0.8" branch, followed by an sudden expansion to 1"?

EAYVJER:

I would approach the problem as addition of the folowing losses:

1) Loss through branch fitting
2) Sudden contraction from 1.5 to 1 inch. diameter
3) Orifice loss

This would of course mean considering the full flow through the branch.This would give the max. loss for the branch flow.In case of partial flows it may not be that easy to determine the losses.

The problem is certainly intersting and if you find some way of estimating the loss, I would like to know.

eayvjer,

Can you rearrange the configuration to put the orifice well downstream from the branch connection?  As shown, the flow in the main line will produce erratic and hard-to-evaluate behaviors.  (The orifice at the branch connection cannot be ignored.)  All normal orifice loss computations and factors presume a well established upstream flow.  With the orifice located well downstream, its loss characteristics can be predicted with significant accuracy.  Operational stability and predictability alone are quite sufficient reasons to implement this change in configuration.  
  
By moving the orifice downstream, the losses at the branch can then be treated as either an entrance loss or as a branch flow loss.  (Both should be considered.)  Actual performance tests can provide a basis for determining which method is most suitable under the important operating conditions.  The velocity in the main line and the velocity in the branch are both important.  (You may be able to devise a proportioning scheme to rationally apply both methods in a mathematical model of the system.)

Thank you all for your thoughts and suggestions!

Arunb; I don't think adding all the losses is the correct method.  True it would give a high value for the loss, but surely too high?  I have tried drawing streamline diagrams to help picture in my mind what the flow is doing and it seems that most of the losses would occur after the orifice, where flow 'expands' after the orifice.  In some ways this would seem to 'replace' the losses due to the contraction.  As far as I can see, the only way to do it is to make an 'guess'timate based on the individual losses.  If I hear of any clever method - I'll let you know!

Ccfowler;  I am trying to estimate the losses in a current system so I cannot move the orifice.  I have been asked to improve on the current system design, and one option I have been looking at is to move the orifice (or even remove it).   Regarding treating the tee as a branch or an entrance loss - I am using a low flow rate and the flow is laminar.  I assume it would be better to use entrance losses in this case?

Many thanks again for all your help!

Regards,

James.

Arunb, sorry for the delay on getting that page to you.  I should be able to send it tomorrow night.