Heat Loss Calculator — BTU/h

A house loses heat surface by surface, and every surface obeys the same small law: area times temperature difference, divided by R-value. Enter your walls, windows and ceiling with their areas and R-values, set the temperature difference you are designing for, and this calculator totals the loss in BTU per hour — the number furnace sizing starts from.

One Law, Three Surfaces

Conduction obeys Fourier's law wearing building clothes: Q = A × ΔT ÷ R. Area and temperature difference push heat out; the assembly's R-value — the summed thermal resistance of its layers — holds it back. The three surfaces this calculator prices are the classic audit because their R-values live in different worlds: a framed wall at R-13, a decent attic at R-30 and upward, and glazing stranded at R-1 to R-3 however modern the frame, because glass is a thermal hole with a view. That spread is the whole diagnostic game: the per-surface breakdown the calculator prints almost always indicts the same suspect, and retrofit money follows the indictment. Floors over unheated space, and the air-leakage term the note prices qualitatively, complete a real load calculation — Manual J in the trade — of which this conduction core is the largest and most honest part.

Worked Example: The Indictment of the Windows

The defaults describe a modest, ordinarily insulated house on a 40-degree design difference (70 inside, 30 outside). Walls: 1,000 sq ft at R-13 lose 1,000 × 40 ÷ 13 ≈ 3,077 BTU/h. Ceiling: 1,200 sq ft at R-30 loses 1,600. Windows and doors: a mere 150 sq ft at R-2 lose 150 × 40 ÷ 2 = 3,000 BTU/h — fifteen percent of the audited area producing thirty-nine percent of the loss, and effectively tying the entire wall system. Total: 7,677 BTU/h of fabric loss. The retrofit moral writes itself and survives every fuel price era: replacing those windows with R-3 units cuts their line to 2,000 and the whole-house figure by an eighth, while doubling the wall insulation — the far larger project — saves less than a fifth of the wall line. Chase the biggest quotient, not the biggest area.

Design Temperature: Choosing the ΔT Honestly

The temperature difference is a climate decision disguised as arithmetic. Design practice uses the local 99% winter design temperature — the cold all but a few hours of the year stay above — against a 68–70°F interior, so Minneapolis designs near ΔT 75 and Atlanta near ΔT 45. Overshoot it and the furnace oversizes (short-cycling, poor humidity control, wasted capital); undershoot and the house cannot hold temperature on the design night. The calculator takes ΔT as an input because it is a lookup against your climate's tables, not a universal constant; the worked example's 40 degrees is a mild-climate placeholder wearing no authority. Weather files and Manual J tables hold the real figure for your county.

Internal Gains: The House Heats Itself a Little

One credit stands against the loss, and honest load work acknowledges it: the occupants and their appliances are heaters. People shed roughly 400 BTU/h each at rest-to-light activity; cooking, lighting, electronics and hot water pour most of their energy into the rooms as heat; even the furnace's own distribution adds its share. Design calculations traditionally ignore internal gains for sizing (the design night must be survivable with the house empty and cold-soaked), but energy modelling counts them, which is why a tight modern house can hold comfort on a fraction of its design load through most of the season. The calculator follows the design convention and prints the gross fabric loss; the reader modelling annual bills rather than sizing equipment should know the credit exists and is theirs to estimate.

The Slab Confesses Differently: Perimeter Loss

One surface breaks the area law, and an honest audit names it: a slab on grade loses heat primarily at its exposed edge, so its loss is priced per linear foot of perimeter (an F-factor times perimeter times ΔT) rather than per square foot of floor. Typical F-factors run from roughly 0.7 for an uninsulated edge down toward 0.35–0.4 with vertical edge insulation — which is why a modest strip of foam at the slab's rim outperforms heroic insulation under its middle, the heat's escape route being the rim's short path to the outdoors. This calculator's three surfaces exclude the slab deliberately; a house on a slab should add the perimeter line mentally (a 120-foot perimeter at F 0.55 and ΔT 40 loses about 2,640 BTU/h — not a rounding error), and a basement's walls price by their own below-grade rules, where the ground itself is the insulation's ally.

Heat Pumps Read the Same Ledger

The total has acquired a new audience in the heat-pump era: a heat pump is sized to the same design load (its capacity falls as the outdoor temperature drops, so the load at the design point is exactly the figure its selection turns on), and its seasonal efficiency is an argument conducted across the whole load curve this page samples at one point. The retrofit corollary matters more: every BTU/h removed from the design load by insulation is a BTU/h of heat-pump capacity not purchased, at the coldest, least efficient moment of its year. The window line that indicted itself in the worked example is thus twice worth fixing for the electrifying household — once for the fuel it was wasting, once for the compressor it was silently upsizing.

Related Tools

More from the Construction Calculators hub: the Floor Area Ratio Calculator for the size of house being heated, and the Snow Load Calculator for the same roof's other winter burden.

Frequently Asked Questions

How do I size a furnace from this number?

As a first approximation only: the design load is this conduction total plus infiltration and any floor and duct terms — commonly 1.25 to 1.5 times the fabric figure in ordinary existing housing — and the furnace's output rating (input × efficiency) should meet that load without gross oversizing. A 7,677 BTU/h fabric loss might land a real design load near 10,000–11,500, served by a nominal 40,000-input furnace at high efficiency with room to spare — the gap between load and available unit sizes being why modulating furnaces exist. A contractor's Manual J is the deliverable; this page is the sanity check that catches its absence.

What R-value should my walls and ceiling be?

Climate-zone codes answer officially (ceilings R-30 to R-60 ascending with cold, walls R-13 to R-21 plus continuous layers in the north), but the calculator's own lesson is directional: returns diminish as R grows — going R-13 to R-19 saves a third of a wall's loss; R-30 to R-60 in a ceiling halves an already small line — while the window line dominates at almost any code-compliant specification. Model your candidate upgrades surface by surface here before pricing them; the quotient ranking is the retrofit plan.

Why isn't infiltration included in the total?

Because it is computed differently (air changes per hour × volume × air's heat capacity factor) and varies more violently between houses than any fabric term — a blower-door-tested tight house and a leaky farmhouse of identical fabric can differ twofold in total load on this term alone. Hiding it inside a fixed multiplier would print false precision. The note prices its typical range (20–40% on top in leaky construction, much less in tight) so the conduction total can be scaled by the reader who knows their house's character — or measured, via a blower door, by the reader who wants the real figure.

Does this work for cooling loads too?

The conduction law is symmetric — summer ΔT drives heat inward by the same formula — but cooling design is dominated by what this page omits: solar gain through glass (often the largest single term), internal gains and latent (humidity) load. A summer run of this calculator with the ΔT reversed gives the conduction component only, useful for envelope comparisons and dangerously incomplete as an air-conditioner sizing figure. Manual J's cooling chapters exist for the rest; the page's silence on them is deliberate.

What is U-factor, and how does it relate?

U is R's reciprocal — U = 1/R — the conductance the window industry quotes (a U-0.30 window is about R-3.3). Where a product gives U, invert it for this calculator's R input. Assemblies complicate slightly: a wall's rated R is its cavity value, while its whole-wall effective R (through studs and plates) runs lower, which is why serious audits derate framed walls by 15–25% — another honesty adjustment the surface inputs let you make yourself.

How accurate is this against a full Manual J?

The conduction core should land close — the law is the law — but a Manual J adds room-by-room granularity, orientation and shading, infiltration measurement, duct location losses and equipment derating, and typically finishes within its own few percent of the truth. This page's role in the ecosystem is the feasibility figure and the retrofit ranking: where the heat goes, in what order to fix it, and whether a contractor's proposed unit is in a sane county. Precision beyond that is purchased with the full calculation, and for a furnace that will run twenty winters, it should be.