My Sense Success Story - Payback in 4 months

#1

So, I moved in to a new house in November ’17. All electric but with 2-stage Geothermal heat pump (GTHP), so while I expected higher total utility bills than my previous gas/electric house, I assumed the GTHP would keep total usage down.

December / January hit hard here (Ohio) with multiple stretches with many days averaging single digits. So, with the continuously running GTHP (~24/7), and a few nights where the supplemental electric coils (stage 3) heat kicked on electric usage was high ($$$$). The first >$400 electric bill quickly justified the cost of the Sense meter.

I installed Sense in about 20 minutes over Christmas weekend. It detected 7 devices within 1 week. Since then, it has detected 20 devices total. It is still missing a few big items (ex. not consistently finding GTHP, differentiating the Stage 1, 2 on GTHP, and associated blowers, etc.) but most can be identified through elimination (ex. My ‘Other’ usage is 90% GTHP).

I quickly started chasing the energy hogs that I could limit.

Even before the water heater (WH) was detected, I found there was a significant difference in what my GTHP was using, and my total usage. Using Sense data, I tracked this down to my electric WH, but didn’t know why.

Digging in to the plumbing, I found a recirculating pump (I never had one before) set to solve the first world problem of waiting 30 seconds for hot water at the taps. The problem is, the installers decided it was a good idea to leave this operating 24/7/365. So, aside from the oversized pump (~140W) running full time, this was continuously circulating hot water through my pipes, and causing WH to cycle around the clock trying to heat my crawlspace.

I turned off the recirc pump, and immediately saw a drop. Oddly enough though, we noticed some odd mixing of hot and warm water, and still excessive cycling of WH (ex. Cycling on every 1-2 hours, even overnight). With that data in hand, I dug in to the plumbing further, and found there was no check valve on the recirc loop, therefore even with the pump off the water was thermocycling through the return line, and dumping cooled water back to my WH 24/7. Two isolation valves later, and WH runs on average twice per day (Now 2-3 kWh vs. 17-20 kWh per day). Immediate cost savings of ~$60/mo. and ~$720/year.

Even if I was able to find this without Sense, it would have taken me weeks to know there was an issue, let alone identify and troubleshoot it.

Identification / Change Savings

  • Improving GTHP controls / settings $200 / year
  • Always on loads – 600 W down to 280 W $330 / year
  • HW recirculating pump (electric for pump) $150 / year
  • WH recirc line thermosyphoning – Isolate. $60 / month $720 / yr

I had previously been tracking individual circuits in my panel with an ammeter, which helped identify a few drains but was painfully slow, and the single point in time data completely missed any trends. I’ve only had solid data reported for 3 months now, and have already more than justified the purchase price. I estimate the payback period is <4 months for me based on the savings above, even after troubleshooting and investigating to find the major energy hogs.

7 Likes
#2

On a related note, I have also been able to trend my daily electric usage with total degree days, with surprising accuracy (R^2=0.89)

This allows me to forecast monthly usage based on weather forecast, and track efficiency gains from any changes. In the example above, Correcting the WH and recirc line usage, shows a notable shift in the trendlines that I can then use to calculate specifically how much a change has saved in kW and $ over a month, year, etc.

I would second all previous posts recommending integration with weather data, and allowing user access to Sense data through API, XML, or other format so we can work with it outside Sense app.

4 Likes
#3

Nice work ! FYI - You can pull data from Sense web app today if you use Selenium, or any of its derivatives, like RSelenium, to harvest the usage pages. I posted a sample script in this thread.