Non-laminar, non-fully developed flow in a duct

[Tatonka]

I've recently been given the task of determining the percent increase in pressure loss for a system that will have several elbows added. The flow is very high (it's the supply for a gas-turbine engine). I hypothesized that by taking some peak measurements, I could approximate the current loss and analytically determine the pressure loss of the new system using the computed flow rate. After taking some measurements with an anemometer, it became painfully evident that the peak velocity did not lie in the center of the round duct, but rather right on the edge. I assume that this means that the flow is not fully developed, and therefor non-laminar. This basically renders all of my assumptions invalid and doesn't allow me to use the tabulated local loss coefficients to calculate the pressure loss due to the added elbows. Is there any way to model or approximate the pressure loss of these elbows without having to hire a contractor to do a CFD study? Ballpark would be nice...

 

[quark]

Tatonka,
I think practically you are on right track and theoretically on wrong one. Flow in duct or pipelines in general is turbulent. Thatswhy you are getting max. velocity at the boundary.(but it should be almost same through out entire cross section except at the very boundary). All the pressure loss calculations will be based on turbulent flow only if not specified.

 

[235zzzo]

Just enforce Quark's reply, which is obvious rigth. In a gas-turbine or any turbo-machine system, think always hard and fully developed turbulent regimes!

The inflexions in the main flow current lines impose by those elbows, just introduce different velocity distributions, in the mean motion or the main flow, but don't change the flow nature, which means the fluctuations components,(the turbulent stresses/Reynolds stresses) are kept present, that means always a flow of a turbulent nature!
zzzo