SizingKitHeating · fabric and ventilation
Heat loss calculator in U-values, R-values and both unit systems
Free, no login: enter each element’s area and its U-value or R-value and this returns the building’s conductance and its design heat loss in BTU/h and kW, with the ventilation term derived from air changes rather than assumed. Any row can be stated four ways — U in BTU/(h·ft²·°F) or W/(m²·K), R in h·ft²·°F/BTU or m²·K/W — because the same calculation is written in two unit systems either side of the Atlantic and one metric R equals 5.678 US ones. It also names which element is carrying the loss, which is the part a single total hides.
- 100% free
- No signup
- U or R, US or metric
- BTU/h and kW together
- Names the dominant element
- Annual energy from degree days
Design conditions and the units you work in
21 °C and 70 °F are both in common use; they differ by 0.2 °F.
Negative figures are accepted; type the minus sign.
Internal volume, ft³ — floor area times the ceiling height, room by room where they differ.
Yours to supply: a blower-door result run through an infiltration model, or the figure your national method prescribes.
For the annual figure only. Leave it if you want the peak alone.
Default 0.65, and the published span is 0.6 to 0.8 — a third of a spread, on a figure that multiplies the whole annual answer.
Element schedule
Each row states its performance however you hold it — a transmittance or a resistance, in either system — and the schedule reports every row both ways.
| Element | Area ft² | Figure | Stated as | U, IP / SI | UA | Remove |
|---|---|---|---|---|---|---|
| 0.083 / 0.47 | 95.8 | |||||
| 0.026 / 0.15 | 26.3 | |||||
| 0.053 / 0.3 | 52.6 | |||||
| 0.32 / 1.82 | 57.6 | |||||
| 0.35 / 1.99 | 14 | |||||
| Ventilation and infiltration | 133 CFM | 0.5 ACH on the volume | 1.08 per CFM | 144 | ||
| Total UA | 390.4 | |||||
The heat loss
Design heat loss
24,594 BTU/h
7.21 kW
Building conductance
390.4
BTU/(h·°F) · 205.9 W/K
Per unit of floor
12.3
BTU/h per ft² · 38.8 W/m²
The term that decides this answer is ventilation and infiltration, at 37% of the total conductance. Halving it takes 4,536 BTU/h off the design load; halving everything else together takes 7,761 BTU/h. That comparison is the reason to compute a schedule rather than a single figure.
Annual heat delivered: 10,709 kWh, or 365 therm — heat into the house, not fuel into the appliance, which is larger by the seasonal efficiency. At the two ends of the published correction span, 0.6 and 0.8, the same building reads 9,885 kWh to 13,180 kWh.
What the answer rests on. Q = UA × ΔT at a design difference of 63.0 °F / 35.0 °C, which is a difference and is converted between the scales as one — a 63 °F difference is a 35 °C difference, not a 17 °C one. The ventilation row is 1.08 BTU/h per CFM per degree, which is 60 minutes an hour times the density of standard air times its specific heat, and it comes out identical whether it is reached through the airflow or straight from the air-change rate (144 and 144). Nothing here has a solar, occupancy or appliance term: a heating load deliberately takes no credit for gains that depend on somebody being home.
U and R, and the two R-values. They are reciprocals: U = 1/R, so an R-19 wall is U-0.053. What trips people is that “R-value” means two different quantities either side of the Atlantic. The US R is h·ft²·°F/BTU and the SI R is m²·K/W, and one SI unit is 5.678 US units — so a European “R-2.5” roof is a US R-14.2, and reading one as the other is a factor of nearly six. That conversion is generated on this page from the watt-per-BTU/h figure, the square feet in a square meter and the size of a Celsius degree, rather than typed in as 5.678. Building an assembly R from its layers is a separate operation, and it is where the framing takes back a fifth of what the insulation gives.
The annual figure is an estimate, and it says so. The empirical correction factor C_D of the ASHRAE degree-day method, which accounts for internal gains, solar gain and the fact that equipment does not run at its rated efficiency at part load. Published guidance for it runs from about 0.6 to 0.8. The weakest number in this file, and the reason the degree-day method is a fuel-bill estimate rather than an engineering result. It is sensitive to how tight and how well-insulated the building is and to how the thermostat is operated. A modern house with large internal gains sits at the bottom of the range or below it; the modern replacement for the whole method is a variable-base degree-day or an hourly simulation. Degree days themselves are the mean-temperature method at base 65 °F. The assumption behind 65 °F dates from a much leakier housing stock. A modern well-insulated house has a balance point nearer 55–60 °F, and comparing its fuel use against HDD 65 overstates the weather's effect. Variable-base degree days, computed at the building's own balance point, are the honest version. The peak figure above is an engineering result; this one is a fuel-bill guess with a named uncertainty, and turning it into gallons of a delivered fuel multiplies that uncertainty rather than removing it.
How to work out a building's heat loss from its fabric
Areas and U-values, then the air, then the temperature difference.
List every surface between heated space and unheated space
Walls, roof or top ceiling, ground floor, glazing and doors, each with its area measured on the warm side. A surface between two heated rooms is not on the list — nothing crosses it — and neither is a party wall to a heated neighbor, which is why a mid-terrace house has a far smaller schedule than a detached one of the same floor area.
State each element's performance in whatever form you hold it
A window sticker gives you a U-value, an insulation label gives you an R-value, and a European specification gives you either one in metric. Set the basis on the row and the schedule does the conversion; it prints every element's U in both systems so you can check it against whichever document you have open.
Add the air, then read the dominant term rather than the total
Air changes times volume gives the airflow, and each cubic foot per minute costs 1.08 BTU/h for every degree of difference. When the answer appears, the sentence under it names the element carrying the largest share of the conductance and compares halving that against halving everything else — which is the difference between a retrofit that works and one that does not.
Technical specifications
| Ways to state an element | Four per row: U in BTU/(h·ft²·°F), U in W/(m²·K), R in h·ft²·°F/BTU, R in m²·K/W. Rows in one schedule can mix all four. |
|---|---|
| Unit conversion | 1 BTU/(h·ft²·°F) = 5.678 W/(m²·K), generated on the page from 0.29307107 W per BTU/h, 10.7639 ft² per m² and the 1.8 ratio between a Fahrenheit and a Celsius degree — not stored as a rounded decimal. |
| Ventilation term | 1.08 BTU/h per CFM per °F, itself 60 min/h × 0.075 lb/ft³ × 0.24 BTU/(lb·°F). Airflow comes from air changes: CFM = volume × ACH ÷ 60, so 16,000 ft³ at 0.5 ACH is 133 CFM and 144 BTU/(h·°F) of conductance. |
| Outputs | Design heat loss in BTU/h and kW, building conductance in BTU/(h·°F) and W/K, and intensity in BTU/h per ft² and W/m² — every figure in both systems, side by side, with no toggle. |
| Annual estimate | 24 × UA × heating degree days × correction, at a base of 65 °F. The correction defaults to 0.65 and the page prints the same building at 0.6 and at 0.8, because published guidance spans that range and the spread is a third of the answer. |
| Temperature handling | Negative outdoor design temperatures are accepted and the difference is converted between scales by the degree ratio alone: a 63 °F difference is a 35 °C difference. Running a difference through the absolute conversion instead would give 17 °C, and it is the single most common bug in this calculation. |
| Ground floors | Treated as a plain U × A against the outdoor design temperature, which overstates a slab. Real ground contact loses heat around the perimeter into soil that sits well above outdoor air; if your method gives a perimeter figure or an effective U for the floor, enter that instead. |
| Terms deliberately absent | Solar gain, occupants, appliances and thermal mass. A design heating load takes no credit for any of them, because the hour it is sized for is a still, dark one with nobody at home. |
Frequently asked questions
My insulation says R-2.7 and yours says R-15 for the same thing. Which is it?
Both, in different units — and the ratio between them is 5.678. The US R-value is measured in h·ft²·°F/BTU and the metric one in m²·K/W, and every one of the three underlying units differs, so the numbers cannot be compared without converting. This catches people out constantly, because unlike most unit clashes there is no symbol to warn you: both are written as a bare number after the letter R.
Is heat loss the same thing as a heating load?
It is the largest part of one, and a load adds two things to it. The first is distribution loss — ducts or pipework running outside the heated envelope give up heat before it arrives, and in a vented attic that is a substantial fraction. The second is pickup: a system that is allowed to set back overnight has to make up the shortfall as well as hold the setpoint, and some methods add an explicit allowance for it. Steady-state fabric and ventilation loss is what this page computes, and it is the floor.
Why is there no term for sunshine or for the people in the building?
Because heating equipment has to work on the night that has neither. Design conditions for heating are chosen at a percentile of the coldest hours, which fall before dawn, and a calculation that leaned on solar gain would size a system that fails at exactly the hour it exists for. Internal gains are real and do reduce annual fuel use, which is why the degree-day estimate lower down carries a correction factor that partly represents them — but they are excluded from the peak on purpose.
How do I enter a basement, or a floor on the ground?
Not as a wall, because the temperature on the other side is not outdoor air. Soil a few feet down sits far above the winter design temperature and the heat path runs sideways to the perimeter rather than straight down, so a slab entered at its construction U-value against a 60-degree difference will read much too high. If your national method has a perimeter loss coefficient in BTU/h per linear foot per degree, or an effective floor U, use that figure in the row and label it accordingly.
What air change rate should I put in?
A measured one if you can get it. A blower-door test gives air changes at 50 pascals, which is not the natural rate — dividing by a climate- and height-dependent factor of roughly 15 to 25 is the usual correlation, and that factor is not on this site. Failing a test, national heating-design methods publish their own infiltration figures by construction type and exposure, and those are what the field expects. Anything you pick out of the air is the largest uncertainty in the whole schedule.
The annual figure does not match my gas bill. Which is wrong?
Probably neither, and the correction factor is where the difference lives. It absorbs internal gains, solar gain, part-load efficiency and how the thermostat is actually operated, and published guidance for it runs from about 0.6 to 0.8 — so the method's own honest precision is around plus or minus 15% before your building is considered. Then the bill covers water heating and cooking as well, and the meter is billed in therms of fuel while this figure is heat delivered into the rooms.
Do I buy a boiler or a furnace at exactly this number?
You buy one that makes at least this much output at your design condition, which is a different and larger input rating. Efficiency separates the two — an appliance is sold on the fuel it burns and delivers less than that into the house — and the sizing method then allows a margin above the load rather than an exact match. Both of those steps are equipment questions rather than building ones.
U, R, and why the whole calculation is one line
Steady-state building heat loss is Q = UA × ΔT and nothing else. Everything a heat-loss calculation does is work out what UA is: every surface’s area multiplied by its transmittance, plus the air moving through the building, plus whatever the distribution gives up on the way. That single line is why the schedule above is worth building element by element rather than reaching for a per-square-foot figure — UA is additive, so each row is independently checkable, and the moment you have it you also have the answer at every other outdoor temperature, the annual energy through degree days, and the balance point at which a heat pump stops keeping up. One conductance, several questions.
The two things calculators get wrong here are both arithmetic rather than physics. The first is averaging R-values across a wall: resistances in parallel do not average, and a framed wall has two paths through it — one through the insulation and one through the studs. Converting each path to a conductance, averaging those by area, and inverting back is the only form that is right, and it is why a 2×4 wall with an R-13 batt performs near R-11.8 rather than R-15.5. The second is converting the temperature difference as though it were a temperature. A ΔT carries no offset between the scales, because both of its ends carry the same one and it cancels; run a 63-degree difference through the reading conversion and you get 17 instead of 35, and the resulting load is out by more than the entire ventilation term. That is the whole reason this site keeps two separate temperature conversions and never lets a page reach for the wrong one.
The output is deliberately a conductance as well as a load, because the conductance is what a retrofit changes and the load is only its consequence. A schedule that shows ventilation at 40% of UA is telling you that no amount of loft insulation will move the answer much, and one where a single-glazed elevation carries a third of the fabric loss is telling you where the money goes. What happens to the number next is a separate question: the furnace page turns this output figure into an input rating and an equipment size, the room-by-room schedule does the same physics with a cooling half attached, and the square-foot rule is what this page exists to replace.
Where the element schedule is held
Every number on this page is worked out by JavaScript running in the tab you are reading it in. Nothing you type — loads, lengths, nameplate ratings, the rates your utility charges you — is uploaded, logged or kept, which is also why the calculators carry on working in a mechanical room with no signal.
Areas, U-values and design temperatures describe a specific building, and none of them is transmitted or retained — the schedule is state in an open tab and nothing more, so use the copy button if you want to keep it.