URDH: Using Sense to track oil usage in a boiler

I’m planning the eventual migration away from an oil boiler to a likely combination of solar + geothermal + heat pump + large thermal storage tank. Unfossilized Renewable Dream Home.

We have 20 (modestly sized) high temperature (180F) radiators in our circa 1940s converted 1870-ish barn that I see no reason to bypass or replace since the plumbing is all OK and done in the days when copper was likely better than today. The “in wall” radiators would also be expensive and disruptive to change. To me the heating system in this house is best left alone, maintained until failure or until something appropriate comes along. If Vattenfall’s heat pump were available in the US I would probably be making plans to change the boiler sooner rather than later.

Anyway, through a combination of having higher renovation priorities (removing wallpaper & plastering!) and laziness I haven’t installed an oil tank level gizmo (yet). The oil delivery company does a pretty good job of estimating consumption on HDD (Heating Degree Days) and prior usage so they get auto-deliveries into the 330 gallon tank at just about the right time.

In March 2021 I decided to put a Flex DCM on the whole boiler circuit, which includes the oil pump, burner and exhaust and 2 circulation pumps, one for the “single zone” loop through all the radiators and the other for a small domestic hot water tank.

The system is controlled by a Nest v3.6 thermostat with 2 remote (Nest) sensors.

OK, to the data.

Nest data is somewhat infuriating (yes, I should use ecobee) and the real meat is only available via Takeout. Even then, parsing it is pretty frustrating.

I’m using HDD data from NOAA but eventually will incorporate a Purple that needs to be installed on the North side of the house. Hyper local weather obviously ideal.

First things first I want to check if Sense and Nest correlate.

I would call that correlation!

But wait, before I go further, this leaves me somewhat confused due to the times when the boiler is on to heat domestic-use water (let’s call that hot water) vs for the radiators (I’ll call that hot glycol even though the percentage of glycol is very low). “NEST ON” is pretty much unrelated to hot water demand since it’s only separately activating for hot glycol demand. Hot water demand meanwhile triggers the boiler to run and so the Sense kWh are accumulating for that day, in heating season. The data here excludes NEST ON = 0. Can heating pros tell me what’s going on? Is hot water usage really that small relative to the hot glycol load? 1% perhaps? Can that data be extracted?

I’m assuming that by parsing the Summer data when all boiler activity is for hot water I can see a baseline of Sense kWhrs. My guess is that the cooincident glycol heating and hot water usage outweighs the infrequent separate heating for hot water only. The more inefficient the system the more linear the response??

I’ll post the HDD charts after I resolve the hot water vs hot glycol vs Sense usage.


Correlation for January 2022 even more impressive.

One can assume that’s because of the high usage.

848 minutes at 1 gallon/hr (the boiler’s burn rate) = 14 gallons, equivalent to 568 kWh.

add 4 kWh for the motors and controls if you want to be picky.

That’s ONE DAY.

1 Like

My natural gas bill in August is about 5-10% what it is during peak heating season. And that includes domestic hot water usage for a family of 4 and cooking appliances. 1% seems a bit on the low side, though. I suppose a smaller family (no kids = less water use) and a higher HVAC heating load (bigger and/or less efficient home = more fuel for heating) could shift the percentages enough to get below 5%.

For reference, my gas bill for the last 30 days listed an average temperature of 11 F., 1683 heating degree days, and 183.4 therms used. Gas company usually shows us as among the top 10% for efficiency and the house is about 3400 square feet with 1200 of that kept at “unoccupied” temperatures with the remainder kept at 65 F when occupied and 60 when away or asleep. Summer usage can be a low as 10 to 15 therms.

1 Like

Here my latest gas usage stats in case they help:

My lowest bill was in August of 2020 and was $56.78.

1 Like

Thanks for the reference @dylan.christopherson
Your house is similar in size in terms of conditioned space, and your thermostat is close enough to ours I suspect (yours is a tad higher overnight but for a few hours around meals we are set higher) that we can draw some interesting comparisons. I’ll ignore the likely differences in boiler efficiency.

You are using 183.4 therms for 1683 HDD.
That’s 183.4 x 29.3 kWh / 1683 HDD = 3.19 kWh/“degree”

I’m using (January 2022), according to NEST runtime on the boiler at 1 gallon/hr
293 gallons for 1290 HDD.
That’s 293 x 40.6 kWh / 1290 HDD = 9.22 kWh/“degree”


@brian5, not sure how big your place is but your house is at 9.0 kWh/“degree”

The chart below needs some outliers checked but roughly speaking it clearly shows the hot water heating baseline for 2021. For Nest = 0, Sense shows 0.416 kWh/day.

Based on Jan 2022 consumption and NEST stats above along with the total electrical consumption for that month (91.4 kWh) I get 293/91.4 = 3.2 gallons/kWh ==> 1.33 gallons(oil)/day for hot water in the warm seasons.

Of course I can also (roughly) check oil consumption based on delivery.

Now I can calculate the volume of hot water we are using if I make some guesses about the boiler efficiency and groundwater temperature.

1 Like

I think you can get a pretty close estimate on boiler efficiency just using stack temperature. For a standard non-condensing gas-fired unit, 85% is about as good as it gets. I expect fuel oil would be similar. Condensing units should be comparable to a modern forced air furnace at 92-93%.

Lowering your return glycol temp as low as possible is key to efficiency. For non condensing units this is usually 150-155 degrees F., but best to monitor for condensation if you’re under 160 on the return. That will help keep your stack temp low and your efficiency high.

FYI: this is all based on my experience with large industrial hot water fire tube boilers (natural gas), but should be applicable to smaller forced draft units.

Good points @dylan.christopherson, thanks.

From what I understand, quantifying stack temperature would only go so far though. Assessing that and O2 content of the combustion gas could give you the combustion efficiency but that wouldn’t really translate to heat exchange efficiency, especially for a small residential system. I will get out my Fluke though and poke around more on the stack.

That makes sense in terms of the goal with space heating: first transfer heat from oil combustion to glycol; then transfer heat from glycol (via radiation) to the air. So the lower the glycol return temperature, the better the heat exchange to the air has been. That doesn’t, though, quantify the efficiency without a lot of hard-to-get data.

What I’m thinking is: Looking at your house and doing some comparative math (above) I see you are perhaps 3X the raw efficiency. Obviously a lot of that has to be due to the better sealing and insulation of your house and I suspect (?) you know that due to your relative household efficiency given to you by the gas company. Maybe accounting for a more efficient gas boiler and your heating practices (not conditioning the whole space for example) we could reduce it to, say, my boiler being half the efficiency or within some believable realm. The trick is where to attribute the various inefficiencies when nothing is exactly quantified!

I’m going to start on something I can wrap my head around: the hot water supply. Standby.