This summer I finally bowed to my wife’s wishes and had solar panels installed on our garage.
They make no sense economically—our electricity bill is only $420/year and the panels cost us about $20k—for a payback period of about 45 years (it will be a little less after tax rebates). Of course, that ignores the maintenance costs—the inverters aren’t designed to last 45 years and will need replacing well before then, and the panels need to be cleaned once or twice a year.
We don’t use a lot of electricity (usage before solar installed):
Our solar installation has 8 Sunviva Optimus 270-60-4-1B0 60-cell panels, each nominally 270 W, but the highest output I’ve seen from one is 237W. The observed peak is less than the rating, because the panels are not getting 1000 W/m2 at 25°C, which is what the panels are tested at. The panels are not optimally oriented to the sun, so they get somewhat less light per square meter (multiply by the cosine of the angle from the optimum). Even if they were getting 1000 W/m2, with that much sun they’d get pretty warm, and the maximum power drops by about 1/8th to below 240W at 50°C.
One interesting observation is that peak power occurs right after dips due to cloud cover—I think that the panels warm up quickly in full sun, and the solar cells are less efficient when they are warm.
The Enphase M250 microinverters are rated for 240W continuous (250W peak), so are a good match for the panels.
The panels are not oriented optimally for solar power, because it is cheaper for installation and maintenance to mount them flat on the garage roof and buy more panels than to make complicated roof racks to get them to the optimal orientation. The garage is oriented so that one half of the roof faces southeast and the other half northwest. The highest peak power is on the southeast half, but the maximum energy generation is on the northwest half, because morning fog limits the sunlight from the east, but afternoons are usually clear, so the northwest panels generate power for longer. Around 1:20pm, both sets of panels are illuminated, and I get a peak power of about 1.57kW (~210W from each southeast panel and ~185W from each northwest panel).
The west-facing panels generate more power and more valuable power, because I have time-of-use service, and peak power costs 30¢–32¢ a kWh (90¢–92¢ on SmartDays™), while morning part-peak costs only 18¢–21¢.
The panels are generating about 10kWh per day (at the beginning of August—I expect about 30% of that in winter on good days, and far less on rainy days). Note: total insolation for Santa Cruz is about 7.41 kWh/m2/day in June and only 1.99 kWh/m2/day in December [http://www.gaisma.com/en/location/santa-cruz-california.html], so minimum winter production at 27% of peak summer production seems reasonable. The average insolation over the year is 4.764 kWh/m2/day, so I expect to average 7 kWh/day over the year (maybe a bit less for bad weather),
Based on these estimates, I expect to generate about 2.5MWh a year, and I use about 2.6MWh a year, so there will be only small net usage or surplus. PG&E will buy any surplus electricity for about 4¢ a kWh. An astute follower of the solar market may wonder why our system cost so much. Well—it’s complicated. We were putting the solar panels on our garage roof, but the shingles there were within 5 years of their expected lifespan, so we ended up reroofing the garage. And the city wouldn’t give us a permit for the solar installation unless we replaced the 60-year-old fuse box, added bigger grounding rods, put the electric feed to the house above the roof line, … . So we ended up with a much bigger project than just adding solar, coming to about $10/W, rather than the more expected $5/W.
I’m not certain about the details of PG&E’s net metering scheme, but I think that if we use electricity at peak times on one day, we get charged full price for it, even if we generate excess electricity at peak times on other days. It would be fairer to have net computations take time-of-use and time-of-generation into account, so that net peak use is computed, with any surplus credited to part-peak, then net part-peak computed, with any surplus to off-peak, then final net computed. But I don’t think that there is any day-to-day carryover of peak surplus power. I’ll have to check the net energy metering contract very carefully to figure out the exact rules.
Of course, total energy use for the house is much larger, since we heat the house, heat water, cook, and dry clothes with natural gas. My natural gas use is about 433 therms/year, which translates to 12.7MWh, about 4.9× my electricity use. So the solar panels don’t make the house energy neutral. Even if we do generate some surplus electricity, switching some usage from natural gas to electricity doesn’t make economic sense, as natural gas costs about $1.28/therm or 4.4¢/kWh, only slightly more than PG&E would pay for the surplus electricity.