PVWatts is a website run by the National Renewable Energy Lab (NREL) which can be used to estimate how much energy a given solar array might produce, based on historical mesurements of insolation. If you ever have a site analysis done, the installer will probably use it to estimate your potential for annual production, and you might wonder how close the the mark it’ll be. Now that I’ve had solar PV for almost 2 years, it’s interesting to look at how close the estimates were for my situation.
Once you choose your location, you input a few other bits of information such as array size (in kW), orientation (tilt & azimuth), and a derate factor (more on that later). If it’ll be a tracking array rather than a fixed array, you give it that information as well. I have 11 230W panels, due south, flush mounted to a roof pitched at about 33 degrees. So, I entered:
- DC Rating (kW): 2.53
- DC to AC Derate Factor: 0.817
- Array Type: Fixed Tilt
- Array Tilt (degrees): 33
- Array Azimuth (degrees): 180
There are a couple of items to note. First, the DC rating. Although the nameplate wattage of all my panels adds up to 2530W, the Enphase M190 microinverter under each one has a max output of 190W (in practice, it gets up to 199W). So even on a perfect day, the array maxes out at about 2190W:
However, this doesn’t happen very often, and when not “clipping” like this, the modules are performing as 230w panels would with any other inverter, so I left it at the nameplate value.
Second, the Derate Factor. This is essentially a measure of the aggregate efficiency of DC to AC conversion. The default in PVWatts is 0.77, but Enphase recommends a slightly higher value of 0.817 due to stated efficiencies of their system.
If you look at the graph above, some months outperform, others underperform. In the first full year, I produced 3209 kWh, and PVWatts estimated 3448 – so I produced about 93% of the estimate.
There are a couple of reasons for this, I think, which are related. PVWatts assumes no shading (although it can be incorporated into the Derate). In reality, I do have some trees and some snow. If you look at the winter months, I fairly drastically underperform the estimates, and I think snow is the likely culprit here, despite my best snow-clearing efforts.
And year round, I do have just a little shade from my neighbor’s trees – this was known at the time of install, because my installer used a Solar Pathfinder [amzn] (see here for how it works, a very cool tool) for the site survey, and it put efficacy right at about 93% after accounting for shading.
So overall, I’d say that if a site assessment is done properly, you should absolutely be able to trust the estimates, based on my experience.