Water heater, not detecting auto-heat cycles?

My water heater was detected within 2 weeks of installing SENSE in November 2019 and appeared to be a full and accurate detection until recently.

What I discovered this weekend through a process of deduction is that my Water Heater turns on for 10-20 minutes approximately every 12 hours. I turned my hot water tank off at the breaker for a day to be sure it was in fact the tank.

I can only assume this is normal behaviour from a hot water tank given its regularity. It sort of makes sense that it would need to do this to maintain the set temperature.

This automated usage has never been detected as a device.

Screenshots attached to illustrate:

The Water Heater device screenshot shows a 2 hour period this afternoon when I flipped the breaker back on for my hot water tank. It took this long to heat up after being off for a day.

The other screenshot shows the main graph with the 2 hour long spike on the left and the 13 minute long spike from 7:20pm to 7:33pm on the right.

This spike on the right wasn’t detected.

Can anyone confirm my assumption about this being a normal auto-heating of the tank?

Would it be worth opening a ticket with support to have this looked at?

A water heater’s thermostatic control doesn’t distinguish between wether the temperature of the water drops due to hot water being used, and so cold water replenishing the tank needs to be heated, or when the temperature drops due to heat loss through the imperfect tank insulation. Meaning: all tank heating is automatic, triggered by a temperature drop.

What’s different between the longer duration heat cycles and the shorter ones is simply that water is being used in the former. Again, your tank doesn’t really “know” that water is being used.

So, why is Sense missing those short cycles that keep the tank at the required temperature? Short answer is that the detection algorithm is not perfect. That said, I have seen better than 95% tracking on my tank when I compare Sense-native detection to Sense ground-truth … by dedicating CTs to the water tank.

It’s likely that certain tank cycles could be missed due to some concurrent usage of other devices that are creating some “noise” for Sense … specifically when the tank cycle goes on. You might try and correlate the times with potential noise sources. The period you posted: dinner time? Cooktops. Ovens. Fridge cycles.

i’m not an expert on hot water heaters, but is it possible its the difference of upper vs lower element?

My understanding of electric hot water is that the upper element usually fires first to heat the water closest to the outlet. Then when that hits temp, the lower element turns on to heat the bottom of the tank.

So in a hot water heater that is being actively used, the top element does the heavy lifting until the water demand stops and it can catch up the upper part of the tank, and eventually the lower element turns on to finish off the heating cycle.

For a tank that isn’t used, presumably the bottom of the tank is going to cool before the upper portion of the tank does, so in that scenario, the lower element may turn on out of sequence to what Sense has id’d as “Hot Water Heater”, so the Sense doesn’t ID it.

This could be a terrible suggestion, but just a thought.

1 Like

Not a terrible suggestion at all, I think it’s very likely you are right about that.

Even though the elements are usually the same it’s my understanding that they will present different loads because:

  • Resistive elements are never identical.
  • The different water strata they are immersed in are at different temperatures. The lower element is cooled more by the water intake so will present a (slightly) different load.
  • Elements often accumulate lime scale and other deposits that modify their resistance.
  • The lower element will normally run longer during water demand and so present Sense with a different signature (at least from a temporal aspect).