I'm planning on replacing my existing dual oil furnaces and dual A/C units (both are about 20 years old) with dual ASHPs (like Carrier Infinity 18VS units), but given the size of my house and my climate (north of Boston), I'm worried about the ASHPs keeping up, especially in cold weather.

As I already have oil, and I still need to make hot water, I'm hoping to install a large oil boiler (probably Buderus Logano G215) to handle both the hot water needs and acts as a backup heat source. I.e, instead of electric resistance coils acting as a backup, we'd have hydronic coils fed from the boiler.

I've carefully analyzed the fuel consumption data of my home. I have a oil tank monitor that I can gather hourly data from. It shows that during the very, very cold Dec/Jan months here in New England, I occasionally would hit nearly 1.28 gallons of oil an *hour*. Now, this is far more than is typical (average is .45 gallons / hour during this winter, or specifically .33 gallons per heating degree day).

Using this 1.28 gallons of oil consumption per hour, I get nearly 200k BTU/hr as my peak heating load on the coldest day. (We had some days where the max temp was below zero.) (138500*1.28)/85% efficiency.

So, finally, my questions:

- I assume this 200k BTU number is what I need to be able to handle?

- I've had multiple manual-js done, but isn't it better to use *actual* HDD + fuel consumption data to accurately predict max and average loads?

- If 200k is my real number, is an ASHP even viable, regardless of backup heat source? The largest Carrier heat pump does about 50k BTUs @ 47df. At 0df, we might get 20k BTUs out of each. That's only 20% of what we'll need during a deep freeze.

- I assume I'm looking at this the wrong way? If it gets super cold out, we'd cut over to oil backup anyway. A big G215 can crank out 220k BTUs/hr.


Below is how my usage breaks down:
Percent > 24k BTU: 89.96%
Percent > 48k BTU: 73.89%
Percent > 96k BTU: 27.80%
Percent > 144k BTU: 7.06%
Percent > 200k BTU: 0.20%
Percent > 224k BTU: 0.00%


Sorry for the long post.

If the 0.33 gallons/HDD is correct over a wintertime-only period of time you can do a reasonable linear approximation of the load.

138,500 BTU/gallon x 0.85 efficiency= 117,725 BTU/gallon ...

... x 0.33 gallons/HDD= 38,849 BTU per degree day...

.../24 hours per day = 1619 BTU per degree hour.

The presumptive heating/cooling balance point is 65F (==zero load @ 65F). So, what's your 99% outside design temperature?

https://articles.extension.org/sites/default/files/7.%20Outdoor_Design_Conditions_508.pdf

If by "...north of Boston..." you don't mean northern Quebec, but some location in eastern Massachusetts your 99th percentile temperature bin is no lower than 0F, which is 65 heating degrees below the degree-day base.

So, 65F heating degrees x 1619 BTU per degree hour = an implied heat load of 105,235 BTU/hr

That's a heluva lot less than 200K! If that's your gallons/HDD constant there's no WAY your load is that high.

Getting the sizing correct for the load is extremely important for heat pumps- oversizing is an efficiency & comfort killer. Even modulating heat pumps don't have infinite modulation range- a Carrier Greenspeed only has a 2.5:1 turn down ratio.

A Mitsubishi PVA-A42AA4 & PUZ-HA42NKA 3.5 ton cold climate modulating system is capable of delivering 48,000 BTU/hr @ +5F, has a heat-strip backup option to cover any shortfall. A pair of those might work here:

http://meus1.mylinkdrive.com/files/PVA-A42AA4_PUZ-HA42NKA_Product_Data_Sheet.pdf

Paying an engineer (and NOT an HVAC contractor) to run careful Manual-J load calculations using aggressive assumptions would be the right starting place for designing a system.

Unless you have gia-normous hot water needs an oil boiler just for hot water and backup dosen't make sense. For less money than a Logano + indirect water heater you could install TWO 80 gallon heat pump water heaters with an EF in the 3s. Installed in a basement they would even cover the bulk of the summertime dehumidification needs.

Thanks, as always, Dana1. I just realized my calculation was wrong for BTU usage. I was dividing by .85 (where .85 represents my oil furnace measured efficiency) when I should have obviously multiplied. Duh!

I'm much, much less confident about my .33 Gallons/HDD number than I am about my hourly fuel usage numbers. My oil tank monitor uses an sonic pulse to measure the oil level. There is some noise, but I adjusted for both statistically and by manually pruning results that weren't congruent with measures on the previous and subsequent hourly readings. Below is my table showing maximum BTU usage. This is calculated using the following: 135800*[Gallons Used That Hour]*0.85

Month / Max Peak Hourly BTU
Oct: 73,452.63
Nov: 109,445.44
Dec: 137,691.57
Jan: 149,483.82
Feb: 131,929.83
Mar: 126,161.74

And here is my revised BTU usage spread:

Percent > 30k BTU: 76.83%
Percent > 60k BTU: 42.72%
Percent > 90k BTU: 15.17%
Percent > 120k BTU: 3.68%
Percent > 150k BTU: 0.00%

“I've had multiple manual-js done, but isn't it better to use *actual* HDD + fuel consumption data to accurately predict max and average loads?”

Yes, an existing energy usage analysis should be more accurate than a Manual J analysis provided you don't need a room-by-room analysis, have accurate energy usage data and have accurate actual HDD data for the measurement period. You might find this software useful:

Borst Existing Building Energy Usage Analysis Software

Please be sure to read the associated software instructions.

Thanks guys.

One of the goals of this system is to ensure I have a way to heat my home on generator power, too. We lose power here a lot, and I'm having a propane-powered generator installed. Even a 20Kw generator can't really handle a couple of heat pumps, especially 5 ton heat pumps. We were without power for nearly 10 days (3 separate incidents) due to storms, and if that had happened in Jan instead of Feb, I'd have a LOT of frozen pipe damage.

So having an oil-fired boiler will get me both hot water (the ones I'm looking at have tankless coils in them, so no need for a hot water tank), backup/emergency heat, and I'll be able to heat me entire home with very little electricity in the event of a power loss.

Most of the boilers I'm looking at cost $2k-$3k. So at least 2x what a heat pump water heater would cost me (and another $1k-$2k for the install!), but it's the only solution I can think of to handle all the scenarios I'm looking to handle. Yes, it will be sitting idle for most of the day, but I'm OK spending $3k-$4k more to have the piece of mind that I have redundancy.

Yes, I can understand the logic in that approach. We have grid, solar and hydro power. Solar power is great in the Summer and hydro power is great in the Winter. We use two 24,000 VA electric boilers for our DHW (relayed so only one can be on at any given time) and a 6,000 VA electric boiler for our hydronic radiant floor heating system. We frequently lose grid power and not having it for a week or two isn’t uncommon either. We have a masonry heater that can serve as our heating system and DHW backup should both the creek be frozen and the weather too poor for adequate solar power generation. So I think it is good to have redundancy and self sufficiency, especially in the times we now live in.

You don't need 5 ton heat pumps. A pair of 3.5 ton Mitsubishis can handle it, and it could be even smaller.

I'm not crystal clear on what the "Percent > 120k BTU: 3.68%" number really means. If it's the percentage of the times where more than 120KBTU was burned in an hour, sampled hourly, it's going to exaggerate the load, since there's a dead zone between firings as the house cools off, and the thermal mass of the house that needs to be heated up on the subsequent cycle, etc. One hour could overshoot the setpoint by 0.5F and use more than 120KBTU, followed by an hour that uses less than 80KBTU without any change in outdoor temp- there's just a lot of "noise" in the data to be integrated out. It's more accurate to do a linear approximation using the gallons per HDD approach.

Your heat load at 0F really isn't more than the 105KBTU/hr from the gallons/HDD fuel use approach- that's an upper bound, and it includes all of the parasitic losses of the heating system as-is.

The oil fired furnaces have a LOT of parasitic load, including the parasitic flue losses when the thing isn't firing. Duct leakage and duct imbalance issues can also easily add a double-digit percentage to the load. Sealing and re-commissioning the ducts (if they are to be re-used) is in order, as well as testing the room to room pressure differences with a manometer with the air handlers running, doors open/closed to find and fix the worst of the duct balancing issues. An Energy Star duct system would have less than 3 pascals (0.012 water -inches) pressure difference between rooms under all operating conditions. Most systems have at least a few rooms with pressure differences 2-3x that much. That leads to excessive use of "the great outdoors" as part of the return path. There's always a way to re-balance to get it to at least under 0.02", if not the 0.012" Energy Star max.

With the furnaces gone you can seal up the flues to get rid of the parasitic flue losses, with recommissioned ducts it's likely that the total heat load will under 95K, possibly under 90K.

For the record, what is the range of calculated heat loads from the "...multiple manual-js ..." (and who performed them)? And the calculated cooling loads were...?

How do you intend to use the oil-burner as backup? Hydronic baseboards? Hydro-air coils in the heat pump air handlers?

Since you'll have propane on site for the generator, and it's only for backup, why oil? A cast iron propane boiler is lower maintenance than an oil boiler, but using any cast iron boiler primarily for domestic hot water is expensive due to the abyssmally low efficiency. Used only intermittently a propane boiler practically maintenance free, not so for oil. A 2kw generator can easily cover the loads of heat pump water heaters and a propane (or oil) boiler. Leaving the boiler off other than pre-season testing, and heating the water with heat pumps is a FAR cheaper way to heat the potable water. The water-heating-only efficiency of tankless coils is truly wretched, under 50% typically, since it requires a high idle temperature (with high standby losses) and oversizing the boiler to deliver even so-so hot water performance. (For details just HOW bad the water-heating-only using tankless coils can be, read this, but the short version is in Table 2: https://www.bnl.gov/isd/documents/41399.pdf )

Spending up to a couple grand on blower door & IR imaging directed air sealing on the house would probably yield a 5K-15K reduction in heat load too. Spending less money on redundant systems and more on just reducing the heat load buys you more comfort, more efficiency, fewer headaches after the big Nor'Easter, as well as smaller-cheaper mechanical systems.

Dana1 - The "Percent > 120k BTU: 3.68%" means that for all the hourly fuel consumption measurements taken between October 1st and March 28th, 3.83% of them represented more than 120k BTUs in output based on how much fuel I used.

And you're absolutely right, the data was very noisy. I did apply some filtering to get rid of obvious outliers, such as shortly after the tank was filled, as well as non-congruent readings where the temperature was fairly static (like 1am to 3am on a given night) but the numbers varied in a non-nonsensical way, suggesting a measurement problem or, perhaps, one of the effects you mentioned. I will say that I was able to consistently find on the coldest nights usage that remained above 110k for multiple hours straight.

Excellent point about set point over shooting. I didn't remotely think of that. I did try and do some cleaning related to that, and the numbers did shift down slightly.

To your points, I have done what I think I can do in terms of tightening up things. I've done duct sealing (Aeroseal), and this seems to have had about a 12% reduction in overall heating load. The duct leakage before sealing was damn near 1 ton of capacity. I've also had professional air sealing done of my attic, and I actually own a thermal camera and used that to find the big stuff. I've also had a blower door test done, and they said there was some room for improvement, but not a lot. In total, I've reduced my fuel usage (adjusted for HDDs) by about 15% versus last winter.

All of the manual-js were done by HVAC companies, so yeah... I agree, it's probably a good idea to just suck it up and pay for an engineer to do it. One reason I'm having trouble trusting them (aside from fuel usage data) is how much they vary - low of 70k BTU/hr to a high of 100k BTU/hr. Cooling loads ranged from 49k to 68k. I currently have a 4 ton and a 5 ton A/C unit, but they're 20 years old. All of this was done *before* the sealing/aeroseal/etc.

There are a couple of reasons I'm hesitant on the propane. First, I don't have a great place to put a 250 or 500 gallon tank. I could do under ground, but this will cost me $1500-$2500. I have a single 80 gallon tank, even doubling that to 160 wouldn't last me particularly long - especially during a power outage. Second, from everything I've read, propane is simply more expensive. Lower energy content and higher cost than oil. It's *really* expensive here in MA, at least right now.

The maintenance is also a great point. That might eat up the difference in fuel costs, but it still doesn't solve my storage problem.

Lastly, I've largely abandoned the idea of an in-boiler coil for exactly the reasons you mentioned. They suck. Instead, I'm considering an indirect tank. Alas, this is $1500 to $2k! Ha. So that kinda of equals out in terms of that versus buying a big propane tank. Reading your link, however, they cite that oil boiler + indirect tank has a LOWER efficiency than an in-boiler tankless coil in the summer. That's the opposite of everything I've read. Ugh!

I guess I'll start researching propane boilers now. lol.

I totally get that propane is ridiculously expensive per MMBTU in MA, which is why you DON'T want to use it for heating hot water 365 days of the year, only as your back up heat. Use a heat pump water heater for the domestic hot water. Basically it only gets used when heating the house.

If using it to heat your domestic hot water an oil fired water heater is probably a more economic solution than a boiler + indirect, if oil. When there's no power failure you can keep the water temp at 140F for minimal standby losses, but can crank it higher (usually 180F, if needed) to run hydro-air backup heat.

The tankless coil is NOT more efficient than an indirect, unless the indirect is tied to a fairly un-insulated low raw combustion efficiency boiler with high idle losses. Compare systems #1 & #2 in Table 2- the system with the indirect has only 6% less hot water per gallon of oil despite nearly TWICE the standby loss, and more than 5% lower raw combustion combustion efficiency than the system with the tankless.

A Buderus + indirect with heat purge controls would do about as well as system #3, but an oil fired water heater maintained at low temp would also be comparable. System #6 is a water heater. See Appendix #6 for how much better it gets when the storage temperature is lower.

You can measure the cooling load with the same systems on a duty cycle basis, looking at only the data from hot afternoons.

If you have the detailed Manual-Js in front of you, compare the assumptions and details for a sanity check. If the high end was 100K and the low end was 70K, I'd tend to believe the lower number unless there is reason to disbelieve it. The fuel use numbers tends to support the 100 KBTU number, but remember that fuel use calculations usually set an upper bound. But the Manual-J is not able to put a number on the parasitic losses from duct imbalances or duct leakage. If the ducts are tight but out of balance (measurable by room to room pressure deltas) it still drives a large amount of outdoor air infiltration whenever the air handlers are running.

Was the fuel use interval data from before or after the Aeroseal? Did they give you before & after duct-blaster CFM/25 numbers?

That would be my speculation as well...70 KBTU feels about right. Excessive loses from lousy ductwork are quite common.

Posted By sailawayrb on 02 Apr 2018 07:39 PM
That would be my speculation as well...70 KBTU feels about right. Excessive loses from lousy ductwork are quite common.

Which is why a deliberate re-commissioning of the ducts that includes testing and correcting room to room pressure differences is in order!

If the actual load is 70K and the fuel use indicates 100K, a third or more of the total heat loss is just bad duct design. With the right ducts a pair of 3 ton Mitsubishis would handle a 70K @ 0F load just fine:

http://meus1.mylinkdrive.com/files/PVA-A36AA4_PUZ-HA36NHA4_Product_Data_Sheet.pdf

But pair of 3 ton (or even 4 ton) Carrier Infinity Greenspeeds would still come up short.

http://www.tools.carrier.com/greenspeed/

The clever refrigerant vapor-reinjection valves they've come up with for delivering higher capacity & efficiency at lower outdoor air temps are no longer a Mitsubishi-only thing. Fujitsu, Daikin, and even Gree (the largest heat pump manufacturer in China) all have variations on the theme that work at low outdoor temps into negative double-digits, but (apparently) none of the US manufactured split-system heat pumps are using that technology. The standard heat pump capacity falls pretty fast once the temperature drop below 20F. It's how a 3 ton Mitsubishi with "Hyper Heating" technology can have as much heating capacity at 0F as a 4.5-5 ton Carrier or Trane.

You are singing my tune...always improve the building envelope first and then work the mechanicals...and I can’t imagine anyone putting in ductwork that isn’t 100% within the conditioned building envelope these days...but there are still older homes that have it. However, if the rumors are true about the next generation residential ERV/HRVs having pressure differential sensing, volumetric airflow tracking and cascaded pressure control capability, infiltration loses will be reduced significantly even in less than tight buildings.

The data from above is *post* Aeroseal. Before sealing I had roughly 1 ton of leakage, and this was reduced by 90%+. I think most of my duct work is now fine, except for the master bedroom. We get very low airflow in there, and given that's where the primary thermostat is for the entire 2nd floor of the house, I'm sure that contributes. I'm planning on having that fixed at the same time the heat pump is being installed.

Aslso, the blower door test showed 6758@40CFM pre-sealing, and 6280@40CFM afte sealing.

So at this point, I think I'm leaning toward a combination of heat pumps and a oil boiler backup, plus a heat pump water heater. The heat pump water heater can be run by a good generator, and the reason I'm still leaning toward oil rather than LP is because of the combination of the cost of the tank installation combined with the relatively low use. Now, I understand there will be maintenance costs associated with the oil boiler - even if it spends 80% of the year off. But if I want to be able to run a generator *and* the boiler off LP at the same time, I'd eat through 2.5-3.5 gallons *an hour*. We've had outages of more than 5 days here. That means, at a minimum, I'd need a 500 gallon tank. That, plus installation/burial, is easily $3k.

If I have two 5 ton Greenspeed units installed, I should only be using backup heat about 7-8% of the heating season. That's perhaps $350 of oil at current prices around here, but LP is 60% more money for the same BTUs. So that same heating load would be something like $566/year. That additional fuel cost is roughly equal to what I've been paying in yearly oil furnace maintenance. It will take almost 10 years to start realizing savings from using LP over oil, assuming we don't have any power outages.

They make diesel generators

Yes, but those generators can't be fed with home heating oil, apparently.

By the way, the exact Aerosealing numbers were:

Pre-sealing: 495.1 CFM
Post-sealing: 111 CFM

GreenSpeeds can be backed up by auxillary strip heaters inside the units, and (unlike oil) they work in conjunction with the heat pump, which is still supplying the lion's share of the BTUs.

Ten tons of compressor is also going to be ridiculously oversized for your loads. A GreenSpeed only has a 2.5 : 1 turn down ratio- it's minimum output is fully 40% of it's maximum. If oversized there is very significant amount of efficiency hit due to on/off cycling for most of the season. When it's 0F outside the raw COP efficiency of a GreenSpeed at full speed is about 1.5, using only 1/3 less power than an electric furnace, so it's not a huge hit to undersize the GreenSpeed slightly for the 99% load and let the strip heaters cover the capacity difference when it's cold out, to allow the thing to actually modulate with load (at much higher efficiency) for the other 95% of the heating season. Every time the compressor has to spin up from zero there's a fixed amount of power wasted, and it has to run for at least 10 minutes or so to even hit it's steady state efficiency (at whatever modulation level it's running). Sized correctly they will run almost continuously with very long on cycles at low speed during most of the heating season. If oversized it'll just cycle except during the coldest days.

In short, with heat pumps (even modulating heat pumps) bigger is decidedly NOT better. Getting accurate load numbers and sizing correctly is critical to getting the efficiency performance out of it, no matter how it fares in an HSPF test.

It still isn't clear how the oil-burner is going to be used as backup. Is it going to be a hydro-air coil in the air handler with the GreenSpeed's coil? If yes, an oil fired hot air FURNACE would be more economic, with lower up front cost lower maintenance, etc. But a "dual-fuel" hot air system can never run the oil burner and the heat pump at the same time- it's either one or the other. You'll never rationalize the cost difference of a dual-fuel solution over a heat-strips on energy cost for just 5% or even 10% of the heating season. The heat pump is still running at a higher COP than the heat strips, even when the strips are engaged.

Hydronic baseboard & boiler with pumps would draw less power than an air handler, but it's a much more expensive installation. Is that your intended solution?

The statement " ...6758@40CFM pre-sealing, and 6280@40CFM afte sealing." has no definition. What are the units for the 6758 & 6280 figures?

The blower door test standard pressure is 50 pascals (both positive & negative pressure are tested) , and is usually expressed in cubic feet per minute @ 50 pascals, or "CFM/50". Are those numbers really 6758 CFM/50 & 6280 CFM/50?

Duct blaster testing is done at 25 pascals, with duct leakage expressed in "CFM/25". The 495.1 CFM and 111 CFM were presumably the leakage at 25 pascals.

That's a real improvement in duct leakage but doesn't address fundamental duct balance issues that could be driving outdoor air infiltration. A house that leaks over 6000 CFM/50 is pretty leaky, and it doesn't take much duct imbalance to add up to a large parasitic heat load. Testing the room to room pressure difference with a manometer while the air handler is running is still an important task to take on (and rectify the worst of it) if the same ducts are going to be used with the heat pumps.

Dana1 - I didn't realize that hydronic coils couldn't be used at the same time as the heat pump. I thought if you replaced the electric heating elements with hydronics, they could serve exactly the same purpose. That definitely changes the math.

I was aware that a furnace can't be running at the same time as the heat pump due to the fact the furnace heat exchanger would always come before the heat pump heat exchanger.

I believe the blower door results are actually 6280 CFM @ 40 Pa. They just didn't document it correctly. (Free Mass Save folks... doing the minimum, and probably barely aware of how this all works anyway.

I did go around my house and measure temperature at each supply. It varies greatly, and I actually think some of the supplies are simply not working at all. Measuring CFM is my next task, but I can already tell that there are a few supplies downstairs that offer virtually no airflow, and almost all the supplies upstairs are in a similar situation. I've got a feeling I prematurely sealed my duct work, which sucks, because Aeroseal isn't cheap.

Leakage of 6280 CFM @ 40 Pa is quite a bit worse than 6280 CFM @ 50 Pa. It's not clear why they tested the house leakage at 40 Pa- perhaps they didn't have a blower door powerful enough to even HIT 50 Pa at your high air leakage(?). Whatever the reason it looks like you'll get far better payback out of IR imaging and blower door directed air sealing, since it will directly lower the total load, reducing the size of the heat pumps for a quick savings up front (did I mention GreenSpeeds aren't cheap?) , and forever after on lower operating cost due to the reduced load.

The output air of the hydronic coil has to be at least 90-100F to have reasonable heat transfer efficiency. Taking that warm air as the input side of a heat pump coil designed for sub-75F incoming air becomes a problem, cutting into the heat transfer efficiency of the heat pump coil and potentially even damaging the heat pump. If you swap it, and put the hydronic coil on the 100-125F output side of the heat pump you have to crank the water temps way up and it'll still short-cycle the boiler due to inefficient heat transfer, with a potential for dangerously hot registers, too hot to touch without burning your skin at some registers.

With heat strip solutions to lagging heat pump capacity the heat strips taking in the tepid output of the heat pump coil and raise the temperature of the air delivered into the supply ducts in a much more modulated, easier to control fashion, delivering air temperatures that are adequate for heating the space but never dangerous. This is all standard stuff, with the controls built into the system, not a whole lot to design with low design risk. It just works.

FWIW: A handful of years ago I was involved with a deep energy retrofit on an 1890s vintage 3 story house in Worcester (converted into three separate apartments, one on each floor.) By taking air sealing seriously throughout the process this 2000-2500' house came in at 464 cfm/50 in the final test more than an order of magnitude lower than what your house would leak at 50 Pa. You won't be able to get it that low without major renovation, but you should be able get it to at least pressurize to 50 Pa, with leakage of under 4000 cfm, probably even less than 3000 cfm. The house is heated with three ductless mini-split heads, one per floor, which they can get away with due to the high-R walls, low air leakage, heat recovery ventilation, and U0.18 windows.

Dana1 - excellent info. I'm going to abandon my clearly ill conceived plan for any kind of hydronic backup. If I end up with heat pumps, it will be with standard electrical resistance backup.

This won't solve my backup generator / power loss issue, but I will probably just get a generator big enough to power one of the two heat pumps and that will keep my house above freezing, which is the most important part. I wonder if I can install some kind of emergency damper that allows the flow from a single heat pump to travel to all registers rather than just the half it's normally hooked to?

Clearly, air sealing is my big priority right now. That, and fixing flow at my upstairs ducts. Based on rough ACH50 calculations, 25% of my heating load is currently lost due to this leakage. I know I won't get all of that back, but every bit counts. I've already reduced energy usage by 20% - if I can snag another 20%, I'll be much closer to a "normal" home.