Insolation is grossly underestimated


I have a split array of 21 panels with 16 at 232° and 5 at 42°. The Isolation calculated for mid Summer seems to be accurate but during the Winter the Isolation is underestimated by over 30%. I am trying to understand why.


Same here with 2 arrays, 18xnortheast and 18xwest:

And in october it’s a bit better, don’t have daily data from the summer yet:


The disparity seems to mostly affect systems with multiple orientations. In my case 16 SW and 5 NE and the 5 NE are on one string with 5 SW panels. The disparity is greatest midwinter (35%high) and least midsummer (0 - 6%). The low sun angles and panel efficiencies are most likely not well compensated in the calculation.


I find that suncalc derived insolation will return negative numbers for some orientations in winter. In my latest attempt, this is in node-red using the solarforcaster-plus node which in turn uses suncalc.

This is probably what’s happening under the hood here too.