I’ve come up on the year anniversary of having the photovoltaic system installed, so I downloaded my PG&E usage information (which is available on an hourly time-step) for the last year, and did some quick Python/Pandas analysis.
Here’s a chart showing our average hourly use over that time period. It shows both the effect of the solar panels (the negative numbers during the middle of the day, when the PV system was putting power onto the grid) as well as the effect of having to plug in the EV (which we usually do after the TOU rate declines at 9 pm, which accounts for the usage peak between 9 pm and 10 pm. (Click here to go to an interactive version if the chart doesn’t work on your browser.)
https://plot.ly/~SSShupe/122
Interestingly, although the system was supposedly sized to produce the estimated amount of power I’d use in a year, summing up the usage data shows that I had a net “take” from the grid of about 1220 kilowatts. That’s about 2 month’s worth of usage for the average household, probably about a month or so’s worth for someone with an EV. Whether this result was due to a mis-estimation of usage or was weather-related is uncertain (although it didn’t seem to me to be a particularly cloudy last year).
I’m planning to do a few other analyses just for fun (usage by day-of-the-week, and by month), and it will be interesting to see if the system’s performance differs next year (although you can’t directly access the system’s output from the PG&E data, since it only gives a “net” PV-output-minus-usage figure).
Categories: Python, technology