How well does it work?

The inverter displays instantaneous power (in Watts) and cumulative energy (in Kilowatt Hours) so its easy to check how well things are working.

The first day after my PV system was turned on happened to be the winter solstice (the shortest day when the sun is lowest in the sky).

Why is the peak not at noon? Kona is 6.1 degrees west of the center of our time zone so the sun’s zenith is at 12:24 [determined by adding (6.1 * 24 * 60 / 360) minutes to noon]. Also my roof is oriented 8 degrees west of south so the sun’s rays are normal to the panel at 12:56 [determined by adding (8 * 24 * 60 / 360) minutes to zenith]. So peak power is at about 1pm. The early afternoon reduction is due to clouds.

How good is the Sunlight Estimate?

The sizing of the PV system was based on the Sunlight Estimate, which in turn was based on energy estimates. Comparing estimated energy to actual 2007 energy readings from the inverter I get:

Comparing Estimated and Actual Energy
Calendar Month	Estimated AC Energy (kWh)	Actual AC Energy (kWh)	(Actual Energy / Estimated Energy) * 100
J	230	197	85.7%
F	231	244	105.6%
M	265	240	90.6%
A	252	220	87.3%
M	251	171	68.13%
J	246	160	65.04%
J	257	216	84.04%
A	273	209	76.56%
S	261	224	85.73%
O	252	247	98.21%
N	231	213	91.99%
D	229	196	85.70%
Total to date	2957	2537	85.80%

The original sunlight estimate based on PVWatts was 4.4h/panel/day. Based on actual generated power so far 3.8h/panel/day (4.4 * .85 ) would be a better estimate.

I notice by watching the display on the inverter that peak power generation is as expected, but the table above shows that total energy generated is less than expected. The difference is in the number of hours of sunlight, and by observation I would say this is due to clouds. As suspected the PVWatts solar energy simulation software does not seem to know about micro climates

A ‘How To’ for Grid-Tie PV on the Big Island of Hawaii