Implementing Dual-Fuel house heating (Heat pump)

Looking to upgrade my old Air Conditioner, I decided to go for a dual-fuel heating solution. Heat-pump for cooling and moderate heating with my existing natural gas furnace for heating at low temperatures.

Update 14 Apr 2023 – some clarifications, updated running data & running costs. (70% heating on HP & ROI < 10 years)

The setup:

I have an existing forced-air heating system with a semi-recent natural gas furnace and an old air conditioner. While the air conditioner worked decently, it was quite inefficient (SEER 10) and nearing its end of life. I live in Ottawa, Canada which has temperatures ranging anywhere between -30C (-22F) and +30C (86F) so both heating and cooling requirements are fairly high.

Equipment Options:

I have a few friends that recently replaced their older furnaces with cold climate heat pumps. Although they would be the most effective for reducing carbon footprint, they are quite expensive and still need a backup solution for low temperatures. Besides, I have a working furnace that has a lot of life left in it.

My real focus was to replace the air conditioner and see how much more a heat-pump would be. A heat-pump is essentially an air conditioner that is run in reverse, cooling the outside and heating the inside. In this setup, a heat exchanger needs to be installed in the duct work above the furnace (same thing for an Air Conditioner) and a unit installed outside. The heat-pump cannot provide all of the required heating in very low temperatures unless it is very large so the thermostat has to be configured so it works until a certain setpoint, after which the natural gas furnace provides the heating.

Finally I could have just opted to go with an air conditioner only.

Heat Pump Sizing

With the approximate price of the Heat Pump being about 35% more than an Air Conditioner, I figured it would likely at least break-even in terms of cost and reduce my carbon footprint quite a bit.

I had a home energy audit conducted as part of an eco-home renovation program so I had some estimates with regards to how much heating and cooling my house would require. I also had some data from tracking my Nest data via the API during the heating season so I could estimate what my actual requirements were.

With a system that is too small you run the risk of not being able to keep up with heating/cooling demand and having the system run continuously without being able to maintain the desired temperature.

With a system that is too big, you run the risk of short-cycling which puts additional wear on the system and can create some temperature imbalances or reduced humidity management.

As I was going for a cheaper single stage heat-pump, I had to balance my heating and cooling requirements. With a dual-stage or continual stage, the system can ramp up or down so if you have too much cooling in the summer, it can just run at a lower stage and avoid short cycling. In my climate, heating loads are 2-3x the cooling load, so there is no risk of sizing too small for heating.

As heat-pumps essentially transfer heat from one location to another, the more temperature difference, the less effective they are. In other words, as the temperature outside gets colder, the heat pump can’t pump as much heat inside. This is why there is a cross-over temperature in a dual-fuel system where the heat pump just can’t keep up anymore (output-based cross-over temperature).

It also means that heating gets less efficient as the outside temperature drops. So it takes more electricity (and costs more) as the temperature drops. So at a certain point, it becomes less expensive to heat using the regular furnace than the heat pump (economic-based cross-over temperature).

The lower the cross-over temperature is, the more cost savings and more carbon reduction you can get. So sizing correctly is important to make sure it is both economical and eco-friendly.

Essentially I had 3 scenarios to consider, 1 estimated loads pre-eco retrofit, 2 estimated loads post-eco retrofit and 3 existing sizing/nest data. On top of that, do I use the economic cross-over to just run when it is cheaper or, do I try and use the output cross-over to save more CO2 emissions?

Sizing procedure and calculations

I calculated what each scenario would give and essentially figured slightly oversizing for cooling loads would give the best cross-over temperatures. I created an excel sheet with the various options and my calculated loads to see where my cross-over points would be and corresponding cooling requirements/output. (Horizontal axis is in F, vertical is in BTU/h)

To do this, you basically have to plot a line between where you assume there will be no heating required up to where you need the most heating (design temperature – 99th percentile of how cold it can get). In this graph you can see my worst case, where the heating load is 80kBTU, the optimal post-retrofit load at 40kBTU and in the middle, what my estimates from existing equipment/data gives around 50kBTU. This line basically maps the increase in heating you need assuming it is linear. For my region we assume no heating at 60F (15.6C) and the 99th design temp at -7.24F (-21.8C). I looked up the design temperature based on historical data for my location.

The same kind of procedure for the cooling where I used a range between 80% and 125% of the estimated cooling load. The cooling data typically only has one data point so the target is to size to fall in between that range. On this side however, I had a bit of an issue with the estimated loads. My existing AC was already beyond these ranges and on hot days can run 12hrs in a day or more. So while I wouldn’t say it is undersized, it certainly is not *that* oversized to fall out of the range, at least in my mind.

The next step is to plot the curves of the actual heat pump choices. In this case, I used what data was available on the spec sheets. There is some better data available from the manufacturer but they don’t provide it unless you create an account as a dealer/supplier etc. so this is the best I could do. Specs usually have a cooling capacity, and two heating capacities one at 47F (8.3C) and the other at 17F (-8.3C). As noted before, typical heat pumps have less capacity at lower temperatures and also are less efficient. (However some cold climate heat pumps have more output at lower temperatures so using only these 2 data points may not be a good estimate).

Output Cross-over temperature

Now that we plotted the heating load and the outputs of the various sizes, we can calculate the output-based cross-over temperatures. Essentially that is the temperature where the heating load line crosses the output line. At this point (if the temperature was stable for hours), your heat pump would run 100% of the time and not be able to provide enough heat so the house would slowly cool down.

Here the difference between the more efficient 3 Ton heat pump was a bit more efficient at higher temperatures when heating and more efficient at cooling but a bit less output.

Nevertheless, there is a huge range of cross-over temperatures and heating load diversion between the scenarios. With a variety of house sealing and window replacement I had done and with the data I had, I was fairly certain it would be closer to the higher % in these ranges (lower cross-over temperature).

Economic Cross-over temperature calculation

I also calculated my economic-cross over temperature based on the tiered electricity pricing (and lower cost on weekends), as well as the cost of natural gas. Essentially I blended the rate for electricity as a rough estimate and looked at the coefficient of performance. The temperature when it would start being more expensive on average ie, the economic cross-over temperature was something like 3F, much below the output cross-over temperature. (for 6 hours a day on weekdays its actually more expensive to run the heat pump than gas, but the average over a week is quite a bit lower – you could fine-tune it for cost savings, but I rather save the planet than a few cents and the trouble).

I also looked at the average cost by integrated the coefficient of performance and temperature (basically how much efficiently the system would have from when I need heating down to the cross-over point and weigh the number based on how much annual heating there would be at each step). My averaged economic cross-over was something much below whatever the output-cross over would be. In fact, since most of the heating would happen at efficient ranges, it would never be more expensive to run the heat pump on average.

Sizing choice

With the economic and output based cross-overs in mind. It was fairly clear to me that the 3 Ton size was correct, although it was a bit oversized for cooling loads, the huge difference between the 2.5 and 3 in the heating season seemed worth it (and the cost was only around 5% more for upsizing). Going to 3.5 or 4 ton was not really an option as it would risk being way too oversized for cooling. With a 2-stage or variable stage, this would have been something to consider, since the heat-pump can run in a lower output stage during cooling, you can oversize without having to worry about it too much. However, my original justification was getting a new Air Conditioning + adding heating as bonus, so I’m going the cheap single-stage route.

Cost savings in using dual-fuel?

I calculated a range of savings (running costs) I could expect based on the estimated heating load and costs. Depending on the load scenario and using a blended electricity rate, comparing that to the cost of the furnace running on gas etc. I calculated running cost savings in the 10-20% range.

That’s great for running costs but in terms of capital return? Adding in heating functionality to the AC would pay itself off in 15-30 years. Well that’s not great seeing as that could be more than the life of the system, but the price of gas is forecast to keep climbing with increases in carbon taxation and extraction costs. On top of that, this unit also reduces my cooling running costs by 40% (my current AC is super inefficient) so it will pay itself off faster than that grim forecast. I decided to go for it anyways, divert some CO2 and save the planet!

Dual-fuel heat-pump system in use

I bought the system and got it installed, no issues getting it up, they just did a drop-in where my old AC was, piped in the lines, installed a new coil on top of the furnace, wired it all up with my Nest and good to go!

Defrost mode

The system was installed and configured so the gas furnace would turn on when the heat pump needed to defrost. This happens when some frost is detected on the outside unit (since it gets colder to “pump” heat inside, the temperature can go below the dew point and lead to frost which reduces the efficiency/effectiveness of the unit.). To get rid of the frost, the unit runs as an Air conditioner to heat the outside coils and remove the frost, this does cool the coil in the air ducts so heating them up with the gas furnace reduces the defrost time, and avoid “cooling” the house during this temporary defrost. Yes this does create some extra CO2, but it also extends the range/time where the unit can operate so overall likely a net benefit.

Cross-over temperature & findings

With the nest thermostat dual-fuel setup, I set the cross-over temperature below my calculations as an initial trial to see how things would go.

The heating is certainly much slower than when the natural gas furnace runs. It’s a slow and steady kind of heating vs a fast rise. It’s something to note and something to keep in mind (especially if you have a large temperature rise like if the house was in eco-mode while on a trip). But I would say its different rather than bad, the temperatures in certain areas of my house are more consistent with this type of heating.

My cross-over temperature is also quite a few degrees lower than my calculations (which is a good thing!). I was in the range of -6.5C and at -9C, it seems to at least maintain temperature (perhaps slow rise). I haven’t had temperatures much below that for a long time since it was installed so difficult to say if I could push it lower.

Essentially with this lower temperature, more of my heating is diverted to the heat pump, less carbon emissions and more cost savings. This might not always be the case but I did forecast my cross-over temperature to be on the better side of my estimates so it’s fairly aligned (but better) than what I was expecting.

My calculated range of 24%-60% of heating being taken over by the heat pump is actually closer to 70%! That’s quite a bit better than I was expecting & should translate to better running costs too!

I also took a look at my estimates for electricity costs. I used a blended rate as a first pass. We have tiered pricing; every weekday the cost varies depending on the time of day: low 12 hours, medium 6 hours, high 6 hours. On weekends and holidays the cost stays low. My blended rate just assumed heating would happen equally at all cost tiers when in actual fact a bigger proportion happens in low, off-hours. It ends up being 8% less than I had originally calculated this along with the higher % of heating diverted translates to running cost savings being 40% better and the return dropping to about 10 years. With the forecast increase in costs of gas & this not being the limit of the system, the heat-pump will pay itself off even sooner than that!

I haven’t had a cooling season yet with this higher sized unit. My previous one was a bit smaller but did run quite a lot during high demand days. Although due to its age, it may not have been running as effectively as it should have, time will tell.