What HDD assumption? Just use actual HDD data for the actual location and actual measurement time period.

http://www.weatherdatadepot.com/

The only assumption is what balance point to use and Dana already explained that very well.

Using real time indoor/outdoor temperatures and fuel monitoring with a numerical integration algorithm to calculate actual real time load and energy usage instead of using HDD data is an even more accurate method for accomplishing an existing building energy usage analysis. While some HVAC control systems have the capability of accomplishing this for you, this is otherwise typically beyond the average DIYer.

Some HDD data is also better than other HDD data too. Some HDD data is simply generated by using the 24-hour high minus the 24-hour low divided by two and the better HDD data is generated by using a more accurate temperature numerical integration approach. Here’s another place to get HDD data:

Degree Days

They also try to explain the differences in how HDD data is calculated:

Degree Days Calculation Methods

And if you want to really delve deep in how to use HDD data for energy consumption analysis, you may want study linear regression analysis too:

Linear Regression Analysis

My neighbor says that on the coldest days of last winter, his house couldn't get above 63F. I don't know if I'd consider that an unacceptable failure or just plug in a few electric heaters and not worry about it.

I think that accurate heat load analysis could be done with the right instrumentation/algorithms and several cold nights of data. But I have no proof of that. Worse case cooling load is harder, due to things like solar gain, cooking and "lots of people during a party".

If the temp on the coldest days of last winter was below the winter out door design temp this is just how it should be and you neighbor had a great success as the system was not grossly oversized as is more often the case. Since you only want the most accurate btu per deg number this energy usage analysis can be done during hot weather too since heat gain and loss is pretty linear with in and out door temp difference. If you have new construction or want a different kind of heating system then you do not have this energy usage analysis option and you have to do less accurate and more expensive manual J analysis.

While the heat gain/loss is proportional to delta T and the building envelop total R-value, please don’t forget that the R-value of some building materials can vary significantly with temperature. This should be properly accounted for in both Manual J and energy usage analyses.

@icfbound: Dana was wrong about the balance point not mattering. As I indicated (using degreedays.net as the basis for my numbers), the difference between base 65 and base 60, which is a very normal range for modern construction that is not at the ultra-high performance end of the spectrum, is over 20% in calculated HDDs all by itself here in central/coastal maine. Plug loads, building mass, local daily temperature swing ranges, and solar gain can all move the appropriate HDD base number (which you are calling a balance point) within that range for typical modern homes. That differential between the two base numbers will shrink in more mild climates, but still, an infiltration estimate has to be pretty far off to be that far off the mark too.. typically infiltration is around 20% of the total load, so to be off by enough to match the HDD potential inaccuracy you have to guess pretty wild. And it happens, absolutely! But not in the hands of a capable and experienced practitioner.

And that's just deciding what base to use. There are still all the other issues I already mentioned that can very significantly cloud the waters or even completely invalidate the case for fuel use analysis once you get past that one.

Fuel use analysis DOES work in a lot of homes. Block loads work better in a lot of homes, and not better in a lot of homes. Different tools for different available datasets, both getting you where you need to go.

If you need to analyze cooling you HAVE to use an MJ unless your client is submetering their HVAC electrical, and if you need to address system operation in any way for control or distribution improvements, room by room MJ is usually the correct tool.

Borst is right to note that many contractors do NOT use the right tool for the job. But the right tool for the job is not always FUA, which is not clear in the quote you posted, so I'm attempting to clarify that a bit further, that's all.

The importance of thermostat settings and heating degree-day base can be illustrated on a recent set of numbers I ran for house in Portland, ME.

The house a ~1400' 1940s Cape, heated with cast-iron propane and the owner was converting to natural gas. He wanted to spec a mod-con solution rather than converting his aging propane burner.

Before knowing that the house was kept in the low 60sF during the use period (and knowing nothing about plug load inputs, etc), the 65F fuel use calc implied a heat load of ~25,000 BTU/hr @ +2F (ACCA's 99% outside design temp for Portland ME.)

After learning the true low-60s indoor temp during the fuel-use period, I worst-cased it using a 55F balance point, then interpolated to a higher 70F interior temp, whe rose to ~32,250 BTU/hr.

That's a whopping 29% higher, but that's also most likely to be worse than reality, since with a ~62F interior temp and some occupants the balance point was likely to be higher than 55F, independent of any reasonable assumptions on plug loads and mammalian occupant output.

At the implied 32K number that's about 23 BTU/hr-ft^2 , which is pretty high for an insulated 2x4 house, even at the modest 1400' footprint, which suggested to me that the foundation may lack insulation &/or the house has a lot of parasitic air leakage. (Have yet to hear back on those details.) Reality might be closer to 30K, but it's probably still worth lowering the load for comfort reasons.

But even if we upsized 25% from the 32K number it would not change the boiler choices, but even if we sized it EXACTLY at 32KBTU/hr output the system would have margin at +2F outdoor temps, since the assumptions in the fuel-use calculation were all skewed to the high side, such as assuming 85% combustion efficiency on a 25 year old 140,000 BTU/hr input propane burner ~4x oversized for the load, and a balance point that was 7-8F under the interior temp in an unoccupied house with minimal plug loads, etc.

The TT Solo-110 he had been considering (as recommended by his heating contractor) has a min-fire output comparable to the 99% heat load, which seemed ridiculous on the face of it, but I ran the numbers for him (twice) anyway, and suggested he look for those heat leaks, since they're probably worth fixing just from a comfort point of view, and would likely be cost-effective on fuel use savings over the longer term. He has sufficient cast-iron radiation to run at the edge of condensing with the Solo-110, but it's ridiculous to have a modulating boiler if it literally never modulates. I recommended he insist on going with the smallest of the line Solo-60, but he might be better off finding a different contractor.

Bravo, but beneath your considerable talents. We do many small houses but the design software would be used to model room as well as radiant loads and the decision for me would be between a tank-type condensing water heater and a small condensing boiler. The maximum size was really a given.

And the maximum size is really a given for most "normal" sized homes built to IRC 2012 code minimums too, but in new construction you don't have a fuel use history from which to estimate the load.

To be sure it gets more complicated & error prone to use fuel-use derived number on large leaky Victorians in various states of retrofit air sealing & insulation, especially when the fuel-use derived numbers are close to the output limits of a potential heating source choice. But as the Portand ME case illustrates, upsizing by 25% from a fuel use derived number isn't likely to leave you in the cold very often even without adjusting for balance point & indoor temp, and upsizing by the ACCA recommended 1.4x would have put them at 35,000 BTU/hr which is above the interpolated worst-case calculation.

If the interpolated heat load had come in at say, 50K that's arguably too close to the ~52K non-condensing output of the Solo-60, but upsizing that by 25% to a 62.5K min-output would be well within the non-condensing output of an 80K boiler like a Burnham ALP-80 or Lochinvar KBN081, without nearly the modulating compromise he would have to take stepping up the Sol0-110. Those 0K boilers would be 1.5x oversized if he had radiation to support 95% combustion efficiency, but only 1.4x oversized at 87% combustion efficiency, neigher a comfort or efficiency disaster.

Bottom line, if you're close to a sizing boundary when using base 65F, finding out more information can fine-tune it. Barring a better informed and adjusted fuel use calc, upsizing by a modest 25% does not have much down-side risk. In the Portland case upsizing by 25% from the 25K number comes to 31.25K. If there were boiler that maxed out at 31.25K exactly, he would have been fully covered at the 97% temperature bin, if not the 99% bin, and may have had to potentially suffer letting it drop to 68F indoors instead of holding the line at 70F on design day. If he upsized by EXACTLY 25% from the 32K number, a 40K boiler would give him huge margin for the Polar Vortex events, and would not suffer from efficiency degradation.

It's unfortunate that within the Triangle Tube series they don't have something that splits the difference between the Solo-60 and the Solo-110, since the step-sizing on min-fire output between them is so huge. What I find completely pathetic is that an installer in Portland ME would be pushing hard to go with the Solo-110 in this house, since it's obvious that they didn't even bother to estimate radiation to min-fire output ratio, and even MORE obvious that they don't have a clue as to the magnitude of real heat loads. The non-condensing output of the Solo-60 would have left him at 37 BTU/hr-ft^2 ratio, and even the dumb " X feet of conditioned space times Y BTU/ft " rule of thumb estimators rarely use more than 35 BTU/hr-ft^2 for Y. Most of those guys I've encountered in New England use 25 BTU/hr-ft^2, and that's already a fairly gross oversizing factor in most cases. This particular contractor is a total hack on that front, which makes me wonder how well they make out on the rest of the system design. (If their plumbing skills looked OK I might still hire them, but I wouldn't let them do the whole installation without supervision or design oversight.)

On this, as with most things, we agree. It is pathetic and all too common for professionals to ignore available resources for proper sizing. Radiation and controls can overcome some of the down sides of over-sizing but right-sizing it the answer. However you come to that is of little concern to me. Have a recognized standard, like minimum codes, is useful.

As homes get tighter smaller boilers will prevail, HTP Contenders, Lochinvar Cadets and NTI TFT60s will rule.

In the future I think the trend to be higher modulation and a single boiler to cover 90% of the market such as the incomparable IBC VFC150 with one of the lowest outputs in the industry.

Until then, give me a Polaris with isolated sub-system on outdoor reset whenever the loads, micro-loads or lack of space demand it.

PS. Unlike Rob, I have designed several systems, with my clients full knowledge, that just met the design loads and fell short when design conditions were exceeded. I do not pad the Manual J8 nor find it exaggerates loads, cooling or heating, in any of the varied applications we use it for.

There is no doubt the minimum end of the boiler range there was obvious with or without any calculations. Even if the house was uninsulated entirely at that size footprint.

So lessons I'd point to here:

1- Fuel use and block loads are useless on anything under 1500 square feet for sure and for most homes, anything under 2000 sq ft, if you're sizing any boiler or furnace I'm aware of. Even uninsulated houses are pretty unlikely to break a 30 BTU/sqft average for the whole house, certainly not more than 35 in any but a tarpaper shack. They are also less sensitive to inaccuracies in any method because of the small real load you are starting with... even large percentages are small moves in absolute load. *Maybe* if you're trying to size an air source heat pump or something, such an analysis might have reason to be.

2- for larger homes where this kind of analysis is actually helpful, those percentages matter more. A 3000 sq ft home might be very comfortable with a solo 60. Or not.

Dana I'm not sure how you ended up only 29% off in this analysis either. Degreedays.net shows the last 12 month HDD for portland maine to be 7253 base65 vs 4659 base55, which means calculating at base 65 would give you about 50% more HDD for your calc than base 65 would. that's over 30% in the final load calc right there and then interpolating for a load that is 10% higher than they are actually using at this time, that differential should have been more like 40% off of your first calc, shouldn't it? Not that it matters here, I'm just curious how you ended up even as close as you did.

base temp isn't only about plugs and people too, of course, mass and solar gain is a big component we are trying to correct for. 2x4 without much solar gain wouldn't get much benefit there. ICF with solar design, different story. And the effects differ in different climates as well. A building being really good at a wide diurnal swings in milder weather can make the peak load appear lower when using FUA.. you're getting credit for situations that may not help you much in cold dark situations.

“ICF with solar design, different story. And the effects differ in different climates as well. A building being really good at a wide diurnal swings in milder weather can make the peak load appear lower when using FUA.. you're getting credit for situations that may not help you much in cold dark situations.”

Yes indeed, this was precisely the situation for the ICF passive solar home that we live in and have tested endlessly, LOL! The Manual J was close for the Winter/Spring loads, but the Manual J was way out of bed for the Summer/Fall loads. However, the energy usage analysis was nearly dead on for all seasonal loads and also agreed very well given what one would predict when properly accounting for the ICF seasonal thermal mass effective R-values.

Standard ICF construction conventional R-value is typically about R23.10 for ICF having 2.5 inch interior EPS insulation, 6 inch concrete core, and 2.5 inch exterior EPS insulation. For our standard ICF construction and using our diurnal Rogue River, Oregon outdoor temperature profiles, the effective R-value was R21.38 in Spring, R66.15 in Summer, R63.42 in Fall, and R21.07 in Winter. So this is an average annual ICF performance of R43.07 or 1.86 higher performance than the conventional R23.10 R-value. Here’s the software that I developed that was used to make these accurate ICF seasonal thermal mass effective R-value predictions:

Borst ICF Performance Software

This subject was also discussed at great length in this thread for those interested in better understanding ICF seasonal thermal mass effective R-value:

ICF Effective R-value Discussion

Posted By NRT.Rob on 13 Apr 2015 03:08 PM
There is no doubt the minimum end of the boiler range there was obvious with or without any calculations. Even if the house was uninsulated entirely at that size footprint.

So lessons I'd point to here:

1- Fuel use and block loads are useless on anything under 1500 square feet for sure and for most homes, anything under 2000 sq ft, if you're sizing any boiler or furnace I'm aware of. Even uninsulated houses are pretty unlikely to break a 30 BTU/sqft average for the whole house, certainly not more than 35 in any but a tarpaper shack. They are also less sensitive to inaccuracies in any method because of the small real load you are starting with... even large percentages are small moves in absolute load. *Maybe* if you're trying to size an air source heat pump or something, such an analysis might have reason to be.

2- for larger homes where this kind of analysis is actually helpful, those percentages matter more. A 3000 sq ft home might be very comfortable with a solo 60. Or not.

Dana I'm not sure how you ended up only 29% off in this analysis either. Degreedays.net shows the last 12 month HDD for portland maine to be 7253 base65 vs 4659 base55, which means calculating at base 65 would give you about 50% more HDD for your calc than base 65 would. that's over 30% in the final load calc right there and then interpolating for a load that is 10% higher than they are actually using at this time, that differential should have been more like 40% off of your first calc, shouldn't it? Not that it matters here, I'm just curious how you ended up even as close as you did.

base temp isn't only about plugs and people too, of course, mass and solar gain is a big component we are trying to correct for. 2x4 without much solar gain wouldn't get much benefit there. ICF with solar design, different story. And the effects differ in different climates as well. A building being really good at a wide diurnal swings in milder weather can make the peak load appear lower when using FUA.. you're getting credit for situations that may not help you much in cold dark situations.

The only fuel use numbers I used of those provided by the owner was the period from 6 December through 23 March, based on the fill-up dates & quantities, during a period of sustained heating load, and at best so-so solar gain. I didn't bother using other periods since these were the fill-ups with the best sustained heat load, which would be impacted the least by other factors. Annual fuel-use/HDD is a TERRIBLE place to start, since hot water heating during the shoulder seasons & summer and the solar heating during the shoulder seasons skew the results too much.  The best accuracy comes from winter-only fuel use, since the solar gain and hot water use factors as a percentage of the energy use is much much smaller.  (Without a closer address I used the weather station at the Portland airport, daily HDD downloaded from degreedays.net.)

During December there was part-time occupancy, including some water heating use, but that's pretty much "in the noise"- no way did any of it drive the average balance point as low as 55F,  or exaggerate the heat load with excessive water heating.  Using 55F as the balance point for a 62F interior temp and 85% efficiency on the propane boiler it came up with about 512 BTU per degree hour.  Projecting the same BTU/degree-hour forward to estimate the load at 70F does not introduce a large error, despite a steeper than reality slope.  If anything these methods inflated the real heat load somewhat, but not more than 10%. It's a sufficiently solid stake in the ground to lean against.  The guy later acknowledged that the foundation is not yet insulated, which is probably a primary factor driving the higher than usual BTU/ft ratio numbers. 

High mass houses and solar-optimized houses are special cases too, eh?  I wouldn't use a fuel use calc on a PassiveHouse or a solar tempered mass house. But Manual-J doesn't do them much justice either.  Site & orientation & shading-factor specific DOE-2/BeOpt simulation (at a minimum) is called for in that type of house.  But those aren't typical house either.

The typical US house built before 2000 (the youngest house you'd be looking for replacement HVAC equipment for)  is under 2400', almost (but not quite) all of which would have a heat load well within the capacity of a Solo-60, but verifiable via (wintertime) fuel-use/HDD data.  But my sense is they sell far more Solo-110s into the replacement boiler market than Solo-60s.    But fuel use methods works fine on conventional 3000'+  houses too.  BTU/hr-ft^2 ratios decline somewhat in bigger houses (on average), but that's only the vaguest sanity-check to look at.

If a wintertime fuel use calculation of heat load is marginal for the max output it's useful to find out more rather than blindly going with that derived number, or blindly upsizing it.  But for giving the guy something to push back on his contractor with it's much quicker than a Manual-J or I=B=R calc, and it's far more clear: The house simply can't need a boiler with 3-4x the output of the fuel-use derived number, and the Solo-60 has about 2x the output at max-fire of the fuel use derived number(s).

"High mass houses and solar-optimized houses are special cases too, eh? I wouldn't use a fuel use calc on a PassiveHouse or a solar tempered mass house. But Manual-J doesn't do them much justice either. Site & orientation & shading-factor specific DOE-2/BeOpt simulation (at a minimum) is called for in that type of house. But those aren't typical house either."

Wait a minute. My Manual 'J' program does site orientation, shading-factor and more. Passive house, NetZero you name it. And I would want a Bosch 57 with a minimum input of 12.9 mbtuh.

thanks for the explanation Dana. I should have realized you were looking at a more targeted date range. That does help a lot to nix the shoulder season varability noise...

My point with big houses was simply that 20% of a big number is a bigger number than 20% of a small number, so the inaccuries matter more the larger the home gets. I definitely agree that solo 60s/EFT50s and other 45-60k boilers are not seen "in the wild" NEARLY enough.

Taking a look at the two load numbers calculated from fuel use, the 25K was obviously too low due to the fact that the HDD base was higher than the average thermostat setpoint. The 32.3K number was a intentional overestimate using an unrealistically low base temperature and an optimistic boiler efficiency.

That's a ~7000 BTU/hr bracket that the true heat load MUST live within. If you simply split the difference and called the heat load 29,000 BTU/hr you'd be inside of 10% of the real number. But it didn't make a whit of difference in this case on the boiler choice, so there was point to fine tuning it any further. If it happened that the numbers straddled the non-condensing output of a given boiler it would be worth taking a closer look, or specifying an 80K boiler from some other vendor. This would work for pretty much any sized house- the more information you have, the closer you can squeeze the brackets.

Morgan- do you actually go out an measure all of the site factors with sufficient precision to get that out of the Man-J software? The DOE2/BeOpt stuff does an OK job of estimating the true insolation based on the very local weather data when calculating annualized fuel use. It's not too much more sophisticated than Manual-J, but may be necessary to reconcile a measured fuel use number for a high-mass solar tempered house.

The Bosch 57 is a better fit for that guy's house, but he was already hooked in with a Triangle Tube installer. I'm not sure how much local support there is for the Bosch 57 in down-east Maine, but if he has any problems getting the Solo 60 dialed in he can just fire the idiot that was pushing the Solo-110 and call Rob Brown, eh? :-) He (thankfully) wasn't micro-zoning the place, and has more than adequate column radiator type heat emitters, except for the 15' stick of fin-tube on a separate zone in the basement, a zone that he never actually uses. Hopefully his hack installer can give him good enough service to make it work with what is pretty much a 1-zone high mass system- it doesn't get much more straightforward than that. But I did suggest he find another contractor- whether it was ignorance or apathy that had him recommending the -110, it's not a good sign.

Lucky to have you too. Well done.

HEY! Portland is NOT "Down East" maine! SLANDER. LIBEL!!!!!

Posted By NRT.Rob on 15 Apr 2015 08:49 AM
HEY! Portland is NOT "Down East" maine! SLANDER. LIBEL!!!!!

Yeah, Portland is practically part of the Peoples Republic of Massachusetts!   (or it was, at one time...)

Fuel use calcs are getting tougher for me as people are less in tune to how many cords of what kind of wood they went through in order to avoid burning propane. But I can tell you manual J has certainly let me down on more than one occasion when the house didn't perform at code minimum.
Recently noticing a clients auxiliary heat was over active, I suggested they get with an energy auditor to see if there was low hanging fruit they could tighten up. They hadn't noticed there was a huge problem as their bills were still much less than propane.
During the blower door test, the auditor found 20ACHs/hour meaning their 1980 home was as tight as an 1880 home!
Had I actual usage numbers to begin with I would have already known the place was leaky, but again they used a wood burner for the past few years and did not keep track of consumption.