POA Orientation

[BHCS]

I'm working on a utility scale solar field with 3 central inverters each covering a third of the site.  From the PVSyst each sub array shows varying azimuth's.  Took the average azimuth for each sub array (ranges from 7 degrees east to 10 degrees west).  Three POA sensors in the field - one for each sub array (one per inverter)

When installing POA sensors in the field do you set the POA so it matches the azimuth of the sub array it is representing?  Since not every module will be at that average tilt for its subarray i would assume you need to figure a way to set the azimuth in the field and align your POA sensor/support pipe in that same orientation.  Also when setting the tilt (20 degrees) would that tilt be measured in the true north-south direction or is the tilt really measures in the direction of the azimuth.

Is there something I'm missing?

[André Mermoud]

If you want to measure the irradiance in the POA, you should choose an orientation for the solarimeter, corresponding to the average orientation of the concerned set of tables. 

However in your case it would be much better to measure the irradiance in the horizontal plane. In this way you only need one only solarimeter, and you can put more attention to the maintenance of this device.  You could also consider measuring the diffuse component (DHI) as well, for improving the transposition. 

The accuracy of the transposition model in different orientation is sufficienly accurate.