BTU requirements by fuel usage

GroundUp · January 31

I think I already know how this is supposed to work, but I'm conflicted. I've got a building with an oversized boiler as it'll still cycle under full load, but I'm not sure how oversized it is. The walls are CMU with siding on the outside and finished drywall on the inside, but I have no idea what type of insulation (if any) was used and I can't tear the wall apart to find out. Running a heat loss calc based on a guess, the number is very similar to the existing boiler size (133k input). However, that number does not reflect the fuel usage. It's been averaging about 1300-1500 gallons of LP annually for the past 4 years. If I adjust the insulation values and air infiltration on the heat loss calc to match the fuel usage, the number is around 55k using 8800 HDD- so close to half. The building is approximately 3000SF of 60's construction in northern MN with 8ft CMU walls on a concrete slab with roughly R30 cellulose in the lid so 18 BTU/SF doesn't even seem realistic, if we're going by fuel usage. Am I calculating this wrong?

Yes I could measure the baseboard element and go from there, which I intend to do at some point, but this is bothering me now. I'm completely baffled at how the heat loss of a 60 year old block building could possibly be 18 BTU/SF when a lot of new construction I do is higher.

Jamie Hall · January 31

I have to admit that Im very wary of the fuel usage approach to calculating heat loss, although I do know that it has enthusiastic supporters. The problem, in my view -- and, as I say, others differ -- is that it has no way of adjusting for extreme events.

Now that said... you have three other approaches open to you for that building, and I would be worth the small amount of time to pursue at least two of them, if only as a giggle check. First you have the known cycling behaviour of the existing boiler and its net output. If you observe that cycling on one of the colder nights, you can get a very close approximation of the actual heat requirement of the building. The procedure is elementary. Take a time when the boiler is nicely up to temperature and cycling. Measure the total cycle length (boiler start to boiler start). Measure the on time within a cycle (boiler start to boiler stop). The actual heat output, to a very close approximation, is the boiler rated net output times the on time divided by the total cycle time. You don't need a boiler bigger than that -- but you do need that. Be sure to do this on a nice cold night! The second approach is to add up the radiation in the building and figure out how much power it can deliver. There is no point in a boiler which can push more power into the radiation than the radiation can deliver to the space. Note that this calculation, however, is affected by the water temperature for hydonics applications, so take that into account. The third, obviously, is to do as good a heat loss calculation as you can on the building -- and I think you have already done that.

GGross · January 31

When you are taking the fuel usage annually in the way you are doing you are getting an average heat loss instead of a design day heat loss. on those 18btu/sf new construction that is 18btu on design day, every other day it is less than that. If you could take just the fuel usage on the coldest day of the year you would get the design day heat loss, for this reason the fuel usage method is not always the best, but is a whole lot better than a guess. On the other hand you have multiple points of information now, you have the average heat loss, and you know what setting the aquastat is at, and also the length of baseboard. So once you calculate the total capacity of the baseboard at the aquastat temp, you can take both those numbers and come up with a correct size for the boiler

Hot_water_fan · January 31

You're right, 1500 gallons/year is not a high heat loss. Is the building kept at a constant indoor temp? What is that temp? It works best when the temp is constant and the temp you will use with the new boiler.

The fuel usage method is based on reality. So wind, cold, you name it, all of it is included in the fuel usage. If that makes you nervous, take that number and multiply it by 1.3 to pad it.

The ~~other methods~~ Manual J is not based on reality. You can do your best with it, but you will have values entered that are either 1. wrong or 2. guesses. Infiltration is the biggest guess but so is insulation!

GGross · January 31

Hot_water_fan said:
You're right, 1500 gallons/year is not a high heat loss. Is the building kept at a constant indoor temp? What is that temp? It works best when the temp is constant and the temp you will use with the new boiler.

The fuel usage method is based on reality. So wind, cold, you name it, all of it is included in the fuel usage. If that makes you nervous, take that number and multiply it by 1.3 to pad it.

The other methods are not based on reality. You can do your best with them, but you will have values entered that are either 1. wrong or 2. guesses. Infiltration is the biggest guess but so is insulation!

The heat loss number you calculate with a manual J is based on design day. ~~Annual fuel usage calculation will not tell you the heat loss on design day~~ which is critical when sizing equipment for other people. Both methods have a use, and both are based in reality

Jamie Hall · January 31

I know you love the fuel use method, @Hot_water_fan , but to state that measuring actual heat loss by measuring actual heat input under certain conditions (the cycle timing approach) is not based on reality is worse than misleading. That approach is, I will grant, based on a point in time measurement under steady or near steady state conditions, and has to be accompted as such, but is quite accurate.

Either that or the method we engineers have been using to determine power inputs, outputs, and efficiency for machines for the last couple of hundred years is wrong...

Hot_water_fan · January 31

The heat loss number you calculate with a manual J is based on design day. Annual fuel usage calculation will not tell you the heat loss on design day which is critical when sizing equipment for other people. Both methods have a use, and both are based in reality

No, that isn't true: you're taking an average Fuel/HDD then applying it to the max HDD you expect, aka design day. It is dependent on heat loss being pretty linear, but that's how the manufacturers set it up. A manual J would match the fuel usage method if the manual J's assumptions were right. The manual J is great for individual rooms and for new buildings.

Hot_water_fan · January 31

@Jamie Hall I consider the cycling method the same as the fuel usage method, so I agree with you (as I often do). I should have specified that I meant the Manual J is not reality based (often it occurs before the building is even built!!). I consider reading a meter the same thing as sitting in a basement manually timing a boiler. The meter is just automated, that's all.

GroundUp · January 31

This property is hours away and it's 80 degrees above design conditions right now, so sitting there to watch it is not an option. If I were there, I'd simply measure the BB element and be done with it. The question was regarding the fuel usage/heat loss/HDD equation, and it appears the jury is still out on this.

Zman · January 31

High mass construction like yours is challenging for heat loss calcs. If the house has always heated well using your baseboard lengths as a max will work. I have also taken to logging actual boiler run times by putting a state logger on a low-voltage CT switch right at the gas valve.

jesmed1 · January 31

GroundUp said:
This property is hours away and it's 80 degrees above design conditions right now, so sitting there to watch it is not an option. If I were there, I'd simply measure the BB element and be done with it. The question was regarding the fuel usage/heat loss/HDD equation, and it appears the jury is still out on this.

If I may add a data point, I'm a mechanical engineer who also does maintenance on our 4-unit condo building with 2 Weil-McLain oil fired boilers that burn about 1200 gallons of oil per year.

I have calculated our building's heat loss 4 different ways, including Manual J and based on fuel consumption, both for an entire season and for one month during the coldest part of the winter. All of those calculations have come in within 20% of each other, and I consider the calculation based on observed fuel consumption during one coldest month, along with the known HDD's for that month in my area, as the most accurate. The method I used for that calculation is here:

https://www.greenbuildingadvisor.com/article/replacing-a-furnace-or-boiler

Furthermore, I've verified those calculations by observing our boiler run times, with known oil input rates, on days when the outdoor temperature is at or below design temperature, and have concluded that, at least for this building, the above heat loss calculation based on annual fuel consumption has an error of around 10%, and the error for a heat loss calculation based on one month's fuel consumption during a cold month like January or February is close enough to zero to be negligible.

STEVEusaPA · February 1

jesmed1 said:
The method I used for that calculation is here:
.

I wish you would stop quoting that article as I, among others, feels like it was lifted from an article that came out a year earlier, by one of the best in the business @Robert O'Brien, mentioned here:
https://forum.heatinghelp.com/discussion/187373/heat-loss-calculation-accuracy

Original article:
https://hydronicshub.com/heat-loss-calculation-on-every-residential-boiler-replacement

I know of 2 installers (and me) that use @Robert O'Brien method exclusively since he published it and I made a spreadsheet of it, and according to them, used correctly, never let them down.
You do have to think about everything going on in the building to properly implement.

jesmed1 · February 1

STEVEusaPA said:

jesmed1 said:
The method I used for that calculation is here:
.
I wish you would stop quoting that article as I, among others, feels like it was lifted from an article that came out a year earlier, by one of the best in the business @Robert O'Brien, mentioned here:
https://forum.heatinghelp.com/discussion/187373/heat-loss-calculation-accuracy

I didn't realize there was a question of plagiarism with the article I posted, so I apologize if I inadvertently promoted such a thing. I would only observe that the Dana Dorsett article is quite a bit more detailed in the math, and considers things such as different balance points (60 F vs. 65 F) that were not discussed in the O'Brien article, which in my view makes the Dorsett article more informative and useful. YMMV.

GroundUp · February 1

@STEVEusaPA Using Robert's method, my numbers come back to 55k which is right where I was before I made this post. I did it a different way the first time, but came up with the same result. Are you actually using this in the field and finding it to be accurate for design days?

Hot_water_fan · February 1

The key is learning that heat loss is linear.

I'm fortunate to have hourly gas usage data. Here's how it shakes out:

1. Observe boiler method: if I timed the boiler on the 10 coldest days from 2018-2023, I'd get an average of 40,960 btus / HDD65. There's hot water and gas stove usage in here too, so it's less than that for those days for heating.

2. Bill method: 39,639 btus / HDD65

3. Regression Method with 1496 days: 38,158 btus / HDD65

As you can see, they're all extremely close.

WMno57 · February 1

Hot_water_fan said:
The key is learning that heat loss is linear.

These guys say its not linear.
https://journals.ametsoc.org/view/journals/apme/40/8/1520-0450_2001_040_1413_datael_2.0.co_2.xml
I found that article in the references for this:
https://en.wikipedia.org/wiki/Heating_degree_day#Problems

Calculations using HDD have several problems. Heat requirements are not linear with temperature,[7] and heavily insulated buildings have a lower "balance point". The amount of heating and cooling required depends on several factors besides outdoor temperature: How well insulated a particular building is, the amount of solar radiation reaching the interior of a house, the number of electrical appliances running (e.g. computers raise their surrounding temperature) the amount of wind outside, and what temperature the occupants find comfortable. Another important factor is the amount of relative humidity indoors; this is important in determining how comfortable an individual will be. Other variables such as precipitation, cloud cover, heat index, building albedo, and snow cover can also alter a building's thermal response.

The Spanish study was about modeling predicted electrical demand. We do a lot of that in Texas now. We are cutting it too close there.
All of this is way over my head, which is good because it makes me think. I'll have to read more about this. I think i'ts good to have different models though. The hard part is to know which model to use when they disagree. Kind of like a weather forecast.

GGross · February 1

WMno57 said:

Hot_water_fan said:
The key is learning that heat loss is linear.
These guys say its not linear.
https://journals.ametsoc.org/view/journals/apme/40/8/1520-0450_2001_040_1413_datael_2.0.co_2.xml
I found that article in the references for this:
https://en.wikipedia.org/wiki/Heating_degree_day#Problems
Calculations using HDD have several problems. Heat requirements are not linear with temperature,[7] and heavily insulated buildings have a lower "balance point". The amount of heating and cooling required depends on several factors besides outdoor temperature: How well insulated a particular building is, the amount of solar radiation reaching the interior of a house, the number of electrical appliances running (e.g. computers raise their surrounding temperature) the amount of wind outside, and what temperature the occupants find comfortable. Another important factor is the amount of relative humidity indoors; this is important in determining how comfortable an individual will be. Other variables such as precipitation, cloud cover, heat index, building albedo, and snow cover can also alter a building's thermal response.
The Spanish study was about modeling predicted electrical demand. We do a lot of that in Texas now. We are cutting it too close there.
All of this is way over my head, which is good because it makes me think. I'll have to read more about this. I think i'ts good to have different models though. The hard part is to know which model to use when they disagree. Kind of like a weather forecast.

A few things here, the way that @Hot_water_fan has described the calculation does take into account the external factors. The other thing here is this is a bad reference that wikipedia has used as the quoted source is talking about electricity usage overall being non-linear with temperature outside, which makes sense as electricity does more than heating. I may also be misunderstanding this as I didn't really understand how they were calculating based on usage in the first place lol and I couldn't say that heat loss is 100% linear or anything, but looking at my load calculations and changing the outdoor temp its close enough to get an accurate size for equipment, and still miles better than guessing

Jamie Hall · February 1

Oh dear. Now we really are getting off into the weeds. Just don't. Conductive heat loss through a medium such as a wall is, in fact, linear. Fine. However, the skin temperature of the wall on both the source and sink sides is what determines that -- and neither of those skin temperatures can be reliably predicted, except under very controlled conditions. This is particularly true of the exterior surface of the wall. The variation of that skin temperature is influenced by a number of factors -- which are not linear, not even close. Then there is infiltration, which is also not even remotely linear, being related to wind speeds and orientation of the structure and any external influences.

And that's just for starters.

Don't cut things too close...

Hot_water_fan · February 1

@WMno57 I like that source.

and heavily insulated buildings have a lower "balance point". The amount of heating and cooling required depends on several factors besides outdoor temperature: How well insulated a particular building is, the amount of solar radiation reaching the interior of a house, the number of electrical appliances running (e.g. computers raise their surrounding temperature) the amount of wind outside, and what temperature the occupants find comfortable. Another important factor is the amount of relative humidity indoors; this is important in determining how comfortable an individual will be. Other variables such as precipitation, cloud cover, heat index, building albedo, and snow cover can also alter a building's thermal response.

I think they're missing some points here:
1. heavily insulated buildings have a lower "balance point" Duh. This is adjusted by using a lower base temp. That's a free change that takes 1 second to make. That's easy for an individual building to adjust for, but the paper is discussing a larger analysis, so would be harder to incorporate.
2. How well insulated a particular building is This is clearly already included in the usage (ie a well insulated house will have a lower BTU/HDD compared to the same sized uninsulated house). Not sure what point they're making here.
3. the amount of solar radiation reaching the interior of a house True, if you're about to reduce solar radiation reaching the interior. I think for heating, it matters very little: the days are already shorter, the sun is lower, etc.
4. the number of electrical appliances running (e.g. computers raise their surrounding temperature) Again, true but how practical is this? Does a manual J account for the computer usage changing? No.
5. the amount of wind outside This is assumed to be "average" with fuel usage. Not 0.
6. what temperature the occupants find comfortable Another good point. You'll have to establish a typical indoor temp for the historical data and assume it stays the same in the future, or pad the number.

I think the biggest missing piece is that thermostats are centrally located. Maybe a house has a few zones, but not infinite zones. Balancing will not be perfect. If the thermostat is satisfied prior to an appliance delivering the full heat load into the house, a bigger appliance doesn't solve that.

Jamie Hall · February 2

Actually, solar radiation reaching the wall does make a difference -- which can be substantial. What it does is alter the effective temperature of the outer surface of the wall, which in turn affects the heat loss. The worst part of it is that the wall material and finish has a major influence...

PRR · February 8

The issue with fuel-logging is that it averages over a longer time, and in domestic heating we NEED to get through that one COLD night.

I have connected a plain old 60cps(Hz) clock to the blower wires to get the duty-cycle of an oversize furnace. Check once a day. If it runs 18 hours (6 hours but you know it was 12 too) in 24 hours then a 75% duty cycle. Or 9 hours in 12 hour night. Actually mine never reached 50% duty so the furnace was twice as big as needed. I didn't wait for the very coldest ("design") day but could estimate the difference. The new downsize furnace can run 22 hours straight if we let the house cool and try to recover on a design day.

My new trick: you can get 24V AC hour meters. They are not common. I hunted much of a year to find one. Today I see this one.
https://www.ebay.com/itm/393612700564
Wire across the gas valve (the load is insignificant) or relay.

You can wire a 120V AC hour-meter across the motor (flame or house, makes little difference in simple systems) but after the one experiment with the clock I decided line-power was not the best idea (and too much work to do right). You can run a 24V DC hour-counter on 24V AC control power with a rectifier, but if you have to ask you probably shouldn't.

Sylvain · February 8

The degree days method is a good one.
It is extensively explained here: https://www.degreedays.net/
with the limitations of the method and how to do it correctly.

As it is statistical, it takes into account mean wind and solar effects on that period.
The points below and above the line show this.

In the equation found: y= 0.3816 x + 0.6688 , the number 0.6688 shows hot water and cooking gas usage which remains more or less stable on the whole year.

The only problem I see, is, for the heat losses calculation, one has to make an hypothesis about the efficiency of the present boiler.
The coefficient 0.3816 must be corrected for this efficiency to get heat losses.

But if the new boiler has more or less the same efficiency as the old one, it is not necessary to calculate the heat losses but only the fuel usage for the degree day of the design day: 0.3916 * HDD(design day).

WMno57 · February 8

PRR said:
The issue with fuel-logging is that it averages over a longer time, and in domestic heating we NEED to get through that one COLD night.

Right. As of late, we seem to be having milder winters. But we get a "Polar Vortex" event every couple of years.
I really liked all of your post. It deserved both a smiley face and light bulb. The clock idea is cool, so I went with light bulb.

Hot_water_fan · February 8

@PRR sounds like you reinvented the meter!

Above, you’ll see that the “one cold night” heat loss is actually not any different than the average heat loss on a Btu/HDD basis.

Jamie Hall · February 8

I'm not going to debate the various methods of determining heat loss. They all have advantages -- and disadvantages. I will add one caution: if you are looking at a steam system installed in a house which has had significant envelope upgrades, you can arrive at a structure heat loss which is considerably smaller than what was there when the radiators were installed. If you now size the boiler to the structure heat loss, rather than the installed radiation, you can end up with a grossly undersized boiler. Somewhat undersized may not be a problem. Grossly undersized, however, and you are guaranteed to have balancing problems, if not flat out no heat problems in parts of the structure -- so be a little cautious.

Sylvain · February 8

For the full benefit of the degree days method, one has to collect the data long before the necessity to change the boiler.
Although, the example of the annual bill hereabove shows that the result is not very different.

The boiler has to be tuned as best as possible. The evaluation of heat losses with a 60% boiler efficiency while assuming 80% efficiency will give a substantial error.

Another benefit of regular use of degree days is the possibility to detect that something has gone astray with the efficiency.

About steam heating, knowing the heat losses after house improvement might help one to evaluate the possible reduction of installed EDR in conjunction with boiler size reduction.

Hot_water_fan · February 8

The boiler has to be tuned as best as possible. The evaluation of heat losses with a 60% boiler efficiency while assuming 80% efficiency will give a substantial error.

True, but it's in the safe direction. If you're assuming 80% efficiency, and it turns out to be 70% efficient, you'll be oversized, not undersized.

BTU requirements by fuel usage

Comments

Categories