How can we accurately compare air flow data in an open system?

[xXDeltaXx]

Hi, I have a question relating to normalizing air flow data in an open system…

The system comprises of a regulator feeding 2barg line pressure to a valve (controlling the downstream flow), a variable restriction, a second valve (normally open) and a flow meter. Pressure readings are taken upstream and downstream of the restriction – these generate the line pressure and differential pressure readouts.

Because of the layout, the restriction in the pipe work generates backpressure as the flow increases – as such, a ‘0 barg’ test can end up with 1 barg of line pressure…

We apply a calculation to the data to ‘normalize’ it at the generated temperatures and pressures.
This is the equation to normalize the data used: QN = Q × (Pa / Pref) × (Tref / T) - sourced from
The problem is that we wish to compare the data to other data sets on a uniform scale (set temperature and set pressure)…
As the generated data is produced with a variable line pressure, I presume that another calculation is needed to remove this variability before entering the uniform values into the normalizing calculation…??

For example:
I wish to compare the following data at 0.5barg line pressure and 15°C to create a flow curve
0 l/min, 0mbar dP at 0barg line pressure, 10°C
20 l/min, 226mbar dP at 0.2bar line pressure, 9°C
42 l/min, 502mbar dP at 0.38bar line pressure, 9°C
The equation can be applied to each line – generating a ‘normalized’ value.
The desired conditions can be applied to each line – generating ‘normalized’ values… but would these really be accurate for comparison as the dP was generated with different line pressures?

(I'm finding this a little difficult to explain exactly what I mean...)

Thanks for your time

 

[xXDeltaXx]

Hopefully this will explain a little better...

http://img268.imageshack.us/img268/5466/query1.jpg
(Illustrative results)
The actual data shows an increase in line pressure and a decrease in temperature during the test where the flow rate and differential pressure are recorded.
The ideal conditions are set at a static temperature and line pressure.
The Nm³/hr in each data table shows the difference changing these values makes.

My query is: because the temperature and line pressure recorded are not static, are the generated results accurate? Or does the calculation work regardless as each point is individually converted? - the fact that the data is part of a data set and across different temperatures and pressures does not matter...