Their answer might be "because that's what Manual J came up with". Might be worth looking into why it is so far off.

Jon, I referenced in an earlier post that I have entered my info into an online free manual J calculator and came up with the same as they are coming up with: about 75k.

Manual J's can be far off, actual usage data is not.

Yes, the email you had drafted is right on target.

Sorry Dana.....I guess we both have names that can go either way........:)

Posted By Tyler23 on 25 May 2011 01:58 PM
Dana & Doc, thanks.

Here is an email I would like to send to the first company (who bid both air source and geo). Please let me know if this looks correct:

Based on info I found here (link to degreedays.net), there were 3,506 heating degree days (HDD) from 12/1/10 – 4/30/11 for Ridgefield. During that time, we used 16,920 KWH. That equates to 4.83 KWH/HDD, or 16,468 BTU/HDD (4.83 x 3412.3) or 686.16 BTU per heating degree hour (16,468/24). At a base temperature of 65 degrees and an outside temp of 15 deg, we would need to 34,308 BTU/hr to cover that. So that’s 34,308/12,000 = 2.859 tons of capacity.

The electricity we used during that time also included all lighting, appliances and hot water, so not everything was going to heat. But to be on the safe side, if we assume a 25% increase in heating load and all electric used was for heat, we could add 8,577 BTU/hr of capacity to 42,885/12,000 = 3.57 tons.

So I guess my question is: if we needed only 35k BTUs to heat our home this past winter, why would we need 72k (6 tons) to do so moving forward?

Going up 25% from the measured number and assuming all power went to heat would be a mistake.  Surely you bathed and kept the refrigerator on during that period!?! 

Average single family non-heating power use in the PNW is on the order of 800-1000kwh/month. Those who heat water with electricity run to the high side of that average.  It's safe to assume that yours is at least HALF that, or 500kwh/ month unless you live in the dark and never bathe. Most of the heat from hot water went down the drain, but the much of the rest of the background load went into heating the house, and a good fraction of the background is distributed over time of day. If we discount the average for all uneveness factors by half, you're still pretty conservative on how much of the peak design load is being supported by the background use.

Take at least 2500kwh (5 months @ 500kwh/month ) off the 16,920kwh number before running the calc and you end up with 14420/3506= 4.11kwh/HDD, or 14,023 BTU/HDD or 584BTU/degree-hour.  Then you're looking at truly 29,200BTU/hr @ +15F, or 26,280 BTU/hr @ 20 F.  Bumping the 15F number by 25% yields 36,500 BTU/hr, and 3-tons is still guaranteed to be slightly larger than optimal, even for the "before" picture on envelope improvements.  That should be a hard upper limit for the compressor, or you could end up close to 2x oversized after taking all of the lower-hanging improvements.  Since even 20F is colder than the 97.5th percentile heating design temp, adding 10% to the 26,280 BTU/hr number, 28,908 BTU/hr would be an even better hard upper bound.

But whether 2.5 tons or 3.0 tons, either would:

A: keep you warm enough, with modest use of aux heat,

...and...

B: run fewer & longer cycles than a bloated 4-5-6 ton system, yielding  higher average efficiency (and lower maintenance.)

BTW: The 99th percentile design temp for Portland is +17F, the 97.5th is +23F, so designing to 20F is still quite conservative for compressor sizing.  Less than 2% of heating season hours would require aux heat. And even if you disabled the aux heat you wouldn't be cold, since most of those hours occur before dawn when most people are abed.  It would only fall truly short during 25 year extreme cold snap events, but that's why you design in some auxiliary heat. 

Dana, yes, I understand all that. But I had to make a point.....and thanks to you, it is a pretty strong one. I have floated this out to one of the companies I've already received a bid from; I have not heard back (not sure if I will). I also pitched this over to the estimator who came out last Friday. He was surprised by the numbers and receptive to using that data to size. I have not received bid from them yet; still waiting on those.

Good luck on finding the right contractor and the right system, and I'm sure they are out there. There is also a lot of old-school thinking, and sloppy estimating out there that drive system sizing out of proportion, but fortunately you've been able to head this one off at the pass. Most residential heating systems in the US are more than 50% oversized for the load, probably half are more than 2x oversized. (Mine was about 4x oversized before I redesigned the system and scrapped the fully functional but piggy high mass boiler. The central AC in this place is something like 8-9x oversized- I rarely run it- a truly ridiculous system, if Ive ever seen one.)

If yours were a 2- stage condensing gas furnace the effect of 2x oversizing on efficiency would be slight, and the up-front cost relatively small. But on a high ticket system like geo it pays to right-size it.

This is also true for modulating-condensing hydronic boilers: If the min-mod is more than half the 97.5th percentile design temp load, efficiency suffers, and maintenance goes up, but the impact on up front system cost is still far less than oversized geo. An extra 20-30KBTU of burner is still usually less than a grand.

So the conclusion is that a properly done Manual J can be off by 100%?

"Bumping the 15F number by 25% yields 36,500 BTU/hr, and 3-tons is still guaranteed to be slightly larger than optimal, even for the "before" picture on envelope improvements. That should be a hard upper limit for the compressor, or you could end up close to 2x oversized after taking all of the lower-hanging improvements."

Do not forget that at the peak of heating season a 3 ton geo will yield something like 2 tons of btus due to lowest seasonal ewts.
Also: No contractor in his right mind would design a system around a 65F winter set point. Occupants change or same occupants age and get on blood thinners....

All that said we've spent 3 months commenting that six tons is too much. I would want to reconcile load vs use descrepencies but without other data I'd lean towards 4 tons and again air source on the light zone. Geo will likely never pay. Plan B might be a small wtw with hydronic air handlers.

Electric rates and local conditions due impact sizing decisions, so you do want to arrive at a compromise with installing contractor. You want to try to avoiding "owning" the design and performance.
j

The seasonal temp differences in western WA aren't nearly as big as in the upper midwest or northeast. Discounting it fully 33% for seasonal ewt is probably too conservative ,but I'd defer to the local pro/designer on just what that discount number should be. (Mini-split air source output is simpler to determine.)

jonr- If done purely objectively with good data on actual interior temp, R, U, and infiltration rate values Manual-J won't overshoot by anything like 100%. It's a crude model, and when every assumption & value is padded a bit "to be conservative" it can easily be that far off. Using an inappropriate 99.5 percintile outside design temp rather than 97.5% or 99% can add quite a bit too. Favoring any Manual-J calc over actual energy use (when it can be determined) is always the wrong approach, at least for the whole-house sizing. Zone & room balance can still be pretty good with a Manual-J, even with the assumptions skewed toward oversizing.

Dana,
I added to that post while you were typing, with the second observation that a local contractor would be crazy to design at 65F winter set point.
I would add that EWT is a product of loop design not "seasonal temp differences". So if minimum 30F EWT is the design then the 33% discount, as you call it, would be accurate.
If not driven by cooling load, loops that are not approaching low 30's significantly oversized, which means consumer could have paid less for the installation.
j

With these discussions on sizing...
what are the options for calculating
heating load from usage?

I have 2 approaches....
1. I have a 20-year old boiler, 187,500 BTU/hr.
On some of the 24-hour periods that I sampled,
my boiler was on just 6 minutes/hour.

Does that mean my heating load was just 18,750 BTU/hr?

2. I do have the gas bills.   Can I calculate it from Therms?



On the flip side, I have two 12K A/Cs in the first floor,
that each is way oversized.
During construction, I am having a 2T MultiAqua installed
in the Kitchen addition, and similarly upstairs later.

(The multi-aqua is the only thing that must be done
before sheetrock :-)

Thanks
Seth

Tyler.....any news on this yet?

Thanks for the clarification,vis a vis EWT & loop design joe- I'd misconstrued the definitions completely (mea culpla!)

Base 65 is fine in a residential app on a not-huge house, since the internal heat sources from bodies & electronics easily fills the 3-5F difference. A heating system only needs to support the part of the load not maintained by the other heat sources in the building. A sleeping adult human is good for something like 400-500 BTU/hr, active they're worth more. The refrigerator and the Tivo add up to another 1000 BTU/hr, it all adds up, and it's real. But even using base 70F rather than 65 is not going to make a huge difference with a 25F design temp. If making the presumption that all winter use is heat, that also pads it considerably, since a bunch of it went down the drain as tepid water, and some of it was outdoor lighting, etc.. If they're planning to upgrade the building envelope, even that would result in oversizing, so base-65 is totally NOT insane. The 10% that was hot water flushed away would more than make up the difference.

Pacha (Seth)i: If the 187.5K is the input number, you're not discounting for the burner's steady-state combustion efficiency. If you take the DOE output and divide it by the input BTUs, that gives you the amount of heat delivered to the system for each unit of source-fuel energy. You can definitely work backward from there with therms of gas use correlated to heating degree data for the exact dates between meter-readings.

eg: Let's say your input # is 187.5K, with a DOE output of 155.5K, that's 155.5/187.5= 0.83 or 83% steady state thermal efficency. Then you look at the billing for 2 consecutive winter months, say between Decebemer 27th and February 25th, and it came out to be 628 therms. At 83% efficiency that means ~521 therms ended up in the house, the rest went up the flue.

Then you look up the base-65 HDD on degreedays.net for those months and add up the daily degree days between those two dates and it came out to be 1393 degree-days.

So you used 521 therms of heat over 1393 HDD which is 0.374 therms per degree day...

... or 37400 BTU per HDD...

... which is 1558 BTU per heating degree-hour.

Then you look up the design heating temperatures for the local area (google the exact phrase "outside heating design temperature" along with a city near you. If you find a couple of different numbers separated by 5-10 degrees the lower one is usually the 99% number, the higher the 97.5% number, based on 25 or 50 year weather history. Let's say the 97.5% number is -17F (no you're not in Kansas anymore Dorothy, you've landed in Fergus Falls, MN). Subtract -17F from your HDD base:

65F- (-17F) = 82 degrees

and multiply by yourBTU per degree hour number:

82 x 1558= 127,750 BTU/hr

Great news- you're only 21% oversized!

If you're like most homes, that 155K-out burner is actually only supporting a derived design load of 57K, and you're actually 270% oversized (great news- you're good down to - 150F!) which means even the 57K number is overstated due to the lower as-operated efficiency of the boiler.

But it's a "do not exceed" stake in the ground, a measurement, and usually much closer to reality than any Manual-J or IBR type heat loss estimator.

Idle losses on oversized boilers are huge, and the cycling losses also not insignificant. That 83% boiler at 2.7x oversizing is probably doing no better than 75% as-operated.

There's a handy boiler modeling tool available developed by the folks at Brookhaven Nat'l Labs that could tell you fairly accurately what your as-used efficiency and design condition heat load is, if you're interested (there's a learning curve to using it correctly, and it has bugs...) It's based on the studies outline in this study:

http://www.nora-oilheat.org/site20/uploads/FullReportBrookhavenEfficiencyTest.pdf

But do the BTU/HDD and design-temp calc first see how bad the oversizing really is.

I'm interested in how accurate this would be. For example, a windy or sunny month would shift values - but how much?

Solar gain can skew the result a bit if there's a lot of glazing and it's a highly sunny place (not much of a concern in most of the US though.) Unless it's a tight, well insulated place designed for heating season passive gain the error would be in the low single-digit percentages. Clearly a single month has a higher error from both weather, occupancy rate, and even the time of day variance of when the meter was read, but over 2-3 months those aspects all shrink.

Monitoring my own place over the past half-decade or so there can be a ~2% variance month-to month mid-winter, maybe 5% during the shoulder seasons, but using even April/May or October/November data rather than mid-winter to do such a calc yields less than a 5% difference. Most of that difference is ascribable to the larger fraction that water heating & cooking play relative to the space-heating loads. Manual-J methods tend to shoot 20%+ higher than the fuel/HDD calc even when I've done them extremely care. But at the temperatures I'm running the heating system and the radiation I have, were the Manual-J correct it would have fallen short this past season when it hit -8F, (my 97.5% design temp is +4F, the 99% condition is 0F, and going by Manual-J estimates the radiation would barely be breaking-even at the 97.5% condition) yet it clearly still had margin at -8F (and plenty of it) which was consistent with the fuel/HDD data calculation.

One still has to be cognizant of the other heating source issues though- use of wood-burners or other supplemental heat would skew the result badly.

"Base 65 is fine in a residential app on a not-huge house, since the internal heat sources from bodies & electronics easily fills the 3-5F difference."

Dana,
any load software worth its salt asks for # of occupants and misc. load (mostly kitchen appliances).

so again, i suggest a contractor would be ill advised to design at 65*f.

joe

And again using an energy-use/HDD measurement methodology rather than a heat loss estimation tool, base 65 is not crazy,especially if any fraction of that energy use went into heating hot water. The air-infiltration error/padding alone makes a bigger difference with heat loss software than using base 65 vs. 70, and I've yet to see a heating contractor who uses a blower door test to hit that factor on the money. (Although Morgan Audetat aka "BadgerMountainMN" might, in some cases.) Heat loss software isn't exactly a building simulator. The simple-arithmetic kwh/HDD is it's not a guess, model, simulation, or estimation, it's a measurement, and puts an upper bound on the heat load unless one derates from that for the "other uses" factors (which can still be done, if you're careful about it.)

But I guess we'll have to agree to disagree on this one, eh? ;-)

FWIW: There are a number of ongoing retrofit studies and field performance analysis of the heating-mode performance of ductless R410A heat pumps in the Pacific Northwest climate(s), funded by the regional utilities & governments. The results are generally favorable for mini-splits, eg:

http://www.bpa.gov/energy/n/emerging_technology/pdf/BPA-Report_Ductless-Heat-Pump-June2009_FINAL.pdf (Goldendale is a up river and a bit cooler in winter than the Portland/Vancouver region.) Not bad performance for a cheap little box, if you can stand the look of the interior units.

While 100% off is outrageous, 20% high might account for worse case short term, non-HDD conditions - very windy, reasonable recovery from setback, etc that don't show up in normal utility bills. Or simply uncertainty in the model. But with geothermal, the typical case is what's important and the unusual case is inexpensively addressed with backup heat.

Dana,
there are many ways to skin a cat, and many scenarios where some things won't work (i.e. calc based on usage history when I look at a plan of a home yet to be built).
My biggest point is most homeowners like it higher than 65* in a home so a contractor darn well better plan for more than that.
Further the contractor that fails to "over deliver" can get in trouble as soon as a clients behavior changes (i.e. I like it 72F now since I started on blood thinners and my heat bill doubled). Remember when dealing with electric auxiliary, 10% more load can easily double a heating bill.

I sure agree to disagree on this thread, but the picture is bigger. I have concerns other that require me not to design too tightly.
j