A few weeks ago we announced the launch of Sense HomeCheck, a free beta service that uses Sense’s high resolution energy monitoring to diagnose when something may be wrong in your home. At the time of launch, we explained that investigating each incident would largely be a manual effort on the part of our data science team, but we also suggested that over time (and with lots of data) we would begin to automatically detect certain incidents.
Well, it didn’t take as long as we thought! We’re excited to announce the release of the first of many of these automated detectors. Sense can now automatically detect and alert homeowners when we identify problems with solar inverters. We’ve seen cases where these problems can result in losses of 50% or more of the potential solar production in a home. In fact, this new automated detection capability was inspired by a Sense user, who noticed that his home was not generating any solar power daily from 10am – 3:30pm during peak hours. With his permission, we took a closer look at the data, and found an identifiable pattern that has helped us craft new algorithms to automatically detect these same failures in the homes across our user base.
Diving into the data
The illustrations below show the two types of solar problems Sense detects: inverter cutoffs and undersized inverters. Of course, this is just the beginning. As more users submit HomeCheck incidents, we will continue to expand the set of problems automatically detected (not only solar, but also faults with house wiring, failing appliances, etc).
Figure 1 shows the two power legs of a 240V solar inverter over a four day span. This homeowner is losing more than 50% of their production because the inverter is turning off even before the day’s peak production. This can be caused by a variety of inverter failures and often goes unreported to homeowners.
Figure 1: four days of solar production showing inverter cutoff consistently happening early in the day.
While this is an extreme case, inverter cutoffs happen in many homes and in some cases result in significant loss of solar production.
In normal solar production, the total power produced follows the intensity of the sun which has a smooth curve up to a peak at the maximum production in the day (note the “mountain” in Figure 2). But, some homes have a well-defined maximum production which is seen as a flat top in the power production curve (contrast “the mountain” to the “plateau” in Figure 3. The dips in Figure 3 are due to passing clouds and do not indicate a problem). This maximum production is typically a limitation of the inverter or the breakers used for the feed-in power and means there is additional production available (the gap between “the mountain” and “the plateau”) which is being wasted by the undersized inverter.
Figure 2: normal solar production, note the curve as solar production reaches its max
Figure 3: a single day of solar production showing an inverter hitting maximum capacity, note the flat top
We hope that by automatically detecting solar inverter problems we can help you get the most out of your solar investment, improve your energy efficiency, and save on your electric bill. If you have other thoughts or suggestions of issues that would be useful for Sense to detect, please send us your feedback or submit an incident report so our data science team can investigate.