Fitting Losses In Laminar Flow

[DKWINE]

There seems to be some confusion in the literature regarding fitting losses in laminar flow. Crane uses the laminar friction factor only with the actual pipe lengths, and uses the "turbulent flow" friction factor for any fittings. The Viking pump catalogue says to neglect fitting losses in laminar flow. Lots of other literature simply uses the laminar friction factor when converting fitting resistances into equivalent lengths of pipe.
My gut feel is that fittings probably offer little or no resistance to laminar flow and that Crane and Viking are probably correct. Thoughts??

 

[katmar]

I agree 100% that there is a lot of confusion in the literature on this subject, and unfortunately this is a particular weakness of Crane 410 (which is still a great resource).

There are two well known methods for predicting the K values for fittings under laminar flow conditions. These are Hooper's 2-K method (Chem Eng, Aug 24, 1981) and Darby's 3-K method (Chem Eng July 1999 and April 2001). Both of these show that the K values increase rapidly as the Reynolds number decreases below 2100. However, remember that the friction factor for straight pipe also increases rapidly as the Reynolds number decreases below 2100.

A few quick calculations show that the proportion of the pressure drop split between the pipe and the fittings remains much the same for turbulent or laminar flow. This means that if you cannot neglect the pressure drop due to the fittings in turbulent flow, you also cannot neglect them in laminar flow. So I would disagree with the Viking catalogue.

If you tried to do laminar flow calcs using turbulent flow K values for fittings then the K values would be much too small, and this would lead to the erroneous conclusion that the pressure drop in the fittings was negligible.

regards
katmar

 

[25362]

Some K values (as number of velocity heads) for pipe fittings in laminar flow, as functions of Re, appear on table 5-15 (Perry Ed VI).

 

[katmar]

Thanks 25362, I had forgotten about Perry. That is a much more accessible reference, and also shows how rapidly the K-values increase as the Reynolds Number decreases.