Low volumetric flow rate measurement

[wrath90210]

Hi everyone, I'm currently a mechanical engineer interning at an HVAC company. I need to measure the resistance of a fluid through ductwork. The CFD shows a linear curve through a plenum when plotting R vs volumetric flow rate. For example, a straight rectangular box 1 1/16” by 2.5” and 3ft long has air passes through it. The machine I use can measure pressure drops at different spots by knowing the input pressure first. Then measuring the second pressure it computes the flow rate. It also takes density and air temperature into consideration.
When running tests of 10 cubic feet per minute or lower the resistance is not constant. The CFD analysis we run shows a constant resistance value that converges. The problem is measuring this in the real world and matching it up to the CFD. Of course the CFD could be incorrect, but the problem has been known for some time.
The resistance is given by R=(DeltaP)/(volumetric flowrate^2)
When plotting 10CFM to 1CFM and calculating each resistance the curve is not a constant value as it should be. The error is the measuring equipment. At HIGH volume flowrates there is no problem, but still does occur when measuring many twists and turns through a pipe.

CMM is cubic m^3/min
as CFM is ft^3/min

Curves and data

P3 and P4 are two pressure taps. They are each a delta in pressure. Just two different locations on the plenum.

I am also aware of minor losses and major losses and the math I have done is accurate to what the machine finds.



Any ideas how to help the R vs CMM curve match the CFD of my system? I just cannot get accurate measurements at low flow rates.
Thank you all!

 

[wrath90210]

Bump...

I have made a large box (plenum) that I place the duct inside. The outside of the duct is exposed to atm, while the other side is inside the box. The box is airtight. I then know the flowrate into the box and know the P inside the box. So the R value can be obtained. But I still have the same result. Evern trying to remove entrance effects seems hopeless. Does anyone know how CFD programs find R values in flows?

 

[jlefevre76]

Without understanding the setup better, I can at least tell you that perhaps, at lower flow rates, your delta P approaches the resolution of your pressure taps, making it difficult to distinguish the actual measurement from the instrument error. You can average the delta P over time, but if the delta P is on the order of the instrument error, you are unlikely to get a good measurement regardless.