How Accurate is Sense? Summary of Sense Accuracy Measurements vs Utilities

Unfortunately for myself (I’ve had it for a few years now) the “depends” answer is what Sense technical support hides behind on a regular basis. For all of 2023 81.3% of our energy use representing 22,974.2 kWh is tagged under “Other”. ie Sense does not recognize any of this energy use as a device. To claim that Sense is accurate compared to the utility is futile compared to such a large portion of use being unaccounted for. What the energy is being used for is the whole purpose of purchasing a Sense. If your bill is slightly off of what the utility says it is, without full detection, that inaccuracy is moot.

Welcome to the discussion @rycamtl.
Accuracy vs. utility is important for 2 reasons:

  1. It’s the basis of all future measurements, so double checking is helpful. Some home monitoring products do less well, especially when it comes to merging net usage and solar.
  2. As the original post relates, many people discover energy issues with this simple comparison. One user discovered a whole second main panel they weren’t aware of. Another discovered they were getting billed against the meter of their next-door neighbor. Several discovered how frequently their smart meter was doing poorly predicted estimated reads or how irregularly the utility was reading their meter.

Detection accuracy is entirely different metric and varies considerably based on the devices in your house, plus the amount of noise in your house, plus a few other factors. This blog post gives potential users and existing users a better idea of what to expect, plus tools and knowledge to understand what’s happening.

My home and devices come in at about 20% Always On and 20% Other, with 60% Detected (though some detections are conflated between a couple of devices). I was able to reduce Other. The methodology is explained here.

One of the big challenges is that many of the new “energy-efficient” devices don’t have clear on and off cycles, that most of the current detection relies on. But I would imagine that Sense is evolving new detection techniques to “see” and “detect” that kind of usage.

My last month electric bill compared to Sense:

My utility company:

Screenshot from 2024-03-08 08-22-36

Screenshot from 2024-03-08 08-24-01

According to Sense:

To grid: 1132.8 kWh
From grid: 387.3 kWh
That should mean I would have had a surplus into the grid of 1132.8-387.3 = 745.5 kWh
But my utility charged me for 238 kWh more from the grid.

- 745.5 kWh Sense
+387.3 kWh Utility

That is a big difference
Any one who is thinking about getting a Sense with a so called AoI all-in-one aka hybrid inverter:
DON"T DO IT , it doesn’t work

Even though they have 1 CT directly over the mains all the time, the way they calculate the solar usage completely destroys the real usage to/from the utility company.

When you look at the products specs:


It nowhere says that it doesn’t work with certain solar setups.

1 Like

There should definitely be a warning about using Sense on any system with 3 energy sources (grid, solar, and battery), since Sense only has 2 pairs of CTs. Sense is only configurable for a couple of multiple-source configurations.

1 Like

I don’t agree with your wording.
The sense unit ALWAYS has 2 CT’s on the mains.
That is the same spot where the (smart) meter of your utility measures.
No matter what setup you have: without solar, with supply side solar, with load side solar!

Those readings should 100% match with what your utility says
And IRL it doesn’t with certain setups.
Not because they didn’t have the info, but the way Sense calculates their things with solar only !

The fact that Sense is not willing to change the way they calculate stuff is plain utterly wrong and utterly frustrating (to me at least)

ADDED LATER
If they than add a warning that with certain solar setups, they are not able to calculate the SOLAR PART correctly, that would be acceptable in my book.
And for clear technical reasons, every load behind a hybrid inverter can not be detected by the Sense unit. If you can’t detect it, you can not measure it.

since Sense only has 2 pairs of CTs. Sense is only configurable for a couple of multiple-source configurations.

I get the idea that you think that is limited because the sense doesn’t have a 3rd set of CT’s but that is irrelevant. The whole thing is DC coupled, nothing an AC CT could measure there.