This post is a bit overdue, but I'll make up for it with a slew of new charts. I've started collecting solar radiation data from a nearby weather station and am using that to estimate the efficiency of my panels on a daily basis. I've also incorporated some additional comparisons between past data (yearly and monthly) to see how my system performance may be changing over time.
First up, an overview of performance for the month (click any of the charts below to see a larger version). Compared to last month my electricity use is down, while solar production is up significantly. I only needed to buy a net of 22 kWh from Seattle City Light this month. Compared to last year my used is significantly lower (likely due to my lack of a roommate this year) while production is down slightly. The system still produced more energy than estimated for the month.
Next up are a couple of new charts comparing production on a daily basis between this year, last year, and the estimate. The month got off to a slow start, lagging slightly behind the estimate and significantly behind last year, but improved noticeably in the latter half of the month thanks to a few particularly sunny days.
System efficiency averaged 13.5%, which just goes to show how much energy could really be provided by the sun with more efficient solar cells. My panels are rated for a peak efficiency of 15.5% while the inverter should peak around 97%, so on a good day I'd hope to see a system efficiency around 15%. For comparison, last year my median system efficiency was 12.7%.
Zooming in a bit, the panels seem to be generating roughly the same amount of power every day, though the lower numbered panels are lagging a bit. Most (but not all) of these panels face West. Given the low total amounts this time of year I'm not too worried about a 10% difference in production.
Here's a better look at production and efficiency as they relate to panel direction. It appears that on high production days the West-facing panels noticeably lagged their East-facing counterparts. I'd have to go back and check the daily weather numbers (perhaps in a future post?) but I'd speculate that this is due to sunny mornings and cloudy afternoons.
Drilling all the way down to the individual panels there's an interestingly large range of performance. In the charts below a long vertical smear of color corresponds to a large difference in performance, while a tight grouping represents more consistency. On days with low production the panels all have very similar performance, but on higher production days there's a large range. The plot on the right is a scatter plot of the efficiency of each panel versus the amount of energy that it was exposed to, measured daily. It's interesting that there seems to be a downward trend, with lower efficiency on days with more available energy. I'm not sure what to make of that.