How the "Sensors" work.

When you create the Scene (makeScene line), bifacial_radiance defines the "sensors" or "linepoints" (linepts) that it will later query when you run the analysis.

The way the Radiance irradiance measurement works: the linepoints establish a vector facing a direction, and whichever surface they encounter first they will query for irradiance. So it doesn’t matter how far or close they are from the module, as long as they intercept the desired surface. By defualt, the bifacial irradiance sensors (linepoints) are created very close to the module surface's front and rear sides, to avoid impinging in the torque tube if the simulation has one. It is important to note that since this linepoints are just 'vectors', the separation to the module surface does not really affect the outcome of the rtrace command or the irradiance output as long as you are querying the right surface (at the right location), so you could manually place them further apart if needed for your own simulations.

You can select which module you will get data from with ModWanted and RowWanted. By Default, ModWanted and Rowwanted are located at the middle of the array. Looking at your irr_Whatevertitle.csv OUTPUT file on the results folder: if you have a 3 modules x 3 rows setup, it will query module 2 in row 2 (the mattype column on your CSV should say something like a1.1.... , for aMODWANTED.aROWWANTED but indexed starting on 0)

The sensor coordinates on your output csv files will follow the center line of the panel's slope, querying as many points as you requested in numsensors. If your module's azimuth is East, for this case you'll probably find that your Y values are all 0, since the center module will coincide with the 0,0 coordinate; and you will also notice that half your sensors in X will have positive values and the other half will have negative values, as it will sample from the bottom of the module up.