Envelope UA Compliance: Method

UA Trade-Off, Layer U-Factors, Degree Days and the Glaser Method

Energy codes give two ways to show that a building's envelope is good enough: the prescriptive path, where every assembly meets the R-value or U-factor of a table, and the total-UA path, where the heat-loss coefficients of the whole envelope are added up and the sum must not exceed what the same building would have at the table values. The second is what the DOE's REScheck and COMcheck programs compute, and it is the method of this tool: an envelope with better walls and a worse floor can comply by trading one against the other. This page sets out the four calculations the tool makes, the sources of its constants, and what it does not do.

The total-UA comparison

For each assembly i with area A_i and overall U-factor U_i (inside and outside air films included, the whole-window value for fenestration), UA_i = U_i A_i is the steady heat flow per degree of indoor-to-outdoor temperature difference. The proposed building complies by total UA when

Σ U_p,i A_i ≤ Σ U_r,i A_i,

the right-hand side being the "code building": the same areas at the required U-factors. The REScheck user's guide describes exactly this: "a simple U-factor × Area (UA) calculation for each building assembly … if the total UA of your building is less than or equal to the total UA of the code building, your building complies". A slab-on-grade floor enters both sums as its F-factor (heat loss per unit length of exposed perimeter per degree) times the perimeter. The trade-off margin printed is (UA_r − UA_p)/UA_r: positive when the envelope beats the code building. The tool also flags every row that exceeds its own requirement, because the prescriptive path would reject it and because a code may cap how far one assembly can be traded (the IECC, for instance, keeps mandatory fenestration U and SHGC backstops).

Three descriptive quantities come with the comparison. The area-weighted U, Σ U A / Σ A, over the whole envelope, the opaque part and the fenestration. The window-to-wall ratio, the area of windows and glazed or opaque doors over the gross wall area (net wall plus those openings); ASHRAE 90.1's prescriptive path and the commercial IECC limit it (40 % and 30 % in the editions the DOE COMcheck slides describe, with the limit tradable in some editions and hard in others), the residential IECC does not. And the area-weighted solar heat gain coefficient of the glazing against the area-weighted requirement, which is how the codes state their SHGC rule.

Required U-factors. The IECC and ASHRAE 90.1 tables are copyrighted and the tool does not ship them. It carries one table that a government publication reprints: slide 15 of the DOE Building Energy Codes Program deck Energy Code Compliance Paths (2019) shows IECC Table R402.1.4, the residential U-factor alternative, for the eight assembly columns and seven climate-zone rows. A blank required cell takes that value for the selected zone and type; a value typed in the cell always wins, so the analyst can enter the adopted edition's numbers (the 2021 IECC changed several fenestration and wall values, as the DOE's 2021 webinar slides list). The same slide prints the small worked example the tool reproduces: in zone 5 a framed wall at U 0.060 is R-16.67 overall; R-20 of cavity insulation alone would be 1/20 = 0.05 before sheathing and films are added.

An assembly U-factor from its layers

Through a layered assembly the resistances add: R = R_{in film} + Σ R_layer + R_{out film}, U = 1/R. The ORNL Insulation Fact Sheet says so in one sentence, and warns in the next that framing bridges the insulation so the assembly's U is worse than the insulation's. The tool handles the framing by the parallel-path method of the REScheck technical support document (PNNL-14864-1-2, eq. A.1): the opaque area is split into a fraction f at the framing and 1 − f at the cavity, each path a series of resistances, and the paths add in parallel,

U = f / (R_common + R_framing) + (1 − f) / (R_common + R_cavity),

where R_common is every layer that both paths share (films, finishes, sheathing, continuous insulation), R_cavity the insulation or air in the stud space, and R_framing the wood across the same depth (the TSD uses 1.25 per inch: 4.38 for a 2×4, 6.88 for a 2×6). The TSD's framing fractions are 25 % of the opaque wall at 16 in on centre and 22 % at 24 in (from the 1993 ASHRAE Handbook), 7 % for ceiling joists at 24 in. The method is accurate unless a layer is highly conductive; steel studs short-circuit the insulation far more than their area suggests, and the TSD applies correction factors to the cavity R instead. For a metal-framed assembly use an entered U from a tested or modelled source, or the ORNL factor below as a rough figure.

Beside the parallel-path value the tool prints the ORNL fact-sheet shortcut (ORNL/TM-2008/017, Table 1): the "reduced clear wall" sums the layers from outside film to inside film, multiplies by 0.75 for wood framing or 0.40 for metal framing, and then adds any foam sheathing. It reproduces the fact sheet's own Table 9 composite U-values (a R-13 2×4 wall 0.0812, R-21 plus R-5 foam 0.0429) and is a check on the parallel path, not a replacement for it.

The layer library. Every value in the material list is printed in a public-domain document and is quoted in the units of its source (h·ft²·°F/Btu): the air films, gypsum board, plywood, siding and empty-cavity values of ORNL/TM-2008/017 Table 1 and of PNNL-14864-1-2 Tables A.1 and A.9, blown fiberglass at R-2.5 per inch (TSD, A.1.2), and wood at 12 % moisture content from the Wood Handbook's Table 4-7, whose values follow its eq. 4-15, k = G (B + C x) + A. Insulation is entered at its rated R (the fact sheet notes the FTC label requirement), continuous insulation likewise, and anything else (masonry, concrete, glass, metal, air spaces) is an entered layer: ASHRAE Fundamentals chapter 26 is where those tested values are read, and the tool does not reproduce its table. In SI the entered R-values are in m²·K/W and the library values are converted on the page.

Degree-day energy, design heat loss and the balance point

A heating degree day is the amount by which a day's mean temperature falls below the base, 65 °F (18.3 °C) in the NOAA normals; summed over a year it is a measure of how much heating the climate demands. If the building's loss were proportional to the indoor-outdoor difference and heating were needed whenever the outdoors fell below the base, the annual heating energy would be

Q_h = 24 × UA × HDD,

with UA the sum of the envelope UA and the infiltration conductance. Infiltration carries ρ c_p per unit volume of air exchanged: 0.018 Btu per ft³ per °F (0.075 lb/ft³ × 0.24 Btu/lb·°F, the familiar 1.08 Btu/h per cfm per °F) or 1.2 kg/m³ × 1,006 J/kg·K in SI, times the air changes per hour and the conditioned volume. The fuel input is Q_h/η at the seasonal efficiency, converted to sold units with the EIA heat contents (a therm is 100,000 Btu by definition, electricity 3,412 Btu/kWh, heating oil 138,500 and propane 91,452 Btu per gallon) and priced. The cooling figure 24 × UA × CDD is the conduction and infiltration part only; the solar and internal gains that dominate a cooling load are outside this estimate, so it is a lower bound to be used for comparing envelopes, not for sizing.

The design heat loss is UA × (T_set − T_design) at the winter design temperature, the number a heating plant is sized to. The balance-point temperature T_bal = T_set − q_int/UA is the outdoor temperature at which the internal gains alone hold the set point; above it no heating is needed. A well-insulated house with normal gains balances several degrees below 65 °F, and its true heating degree days are those at base T_bal, fewer than the published base-65 figure, which is why the degree-day estimate tends to overstate the heating of tight, well-insulated buildings. The station list fills the 1991–2020 NCEI normals (ANN-HTDD-NORMAL and ANN-CLDD-NORMAL of the by-station files) for twelve airports; any other site's values are entered from the same product.

The Glaser (dew-point) condensation check

Vapour diffuses through an assembly from the high-vapour-pressure side to the low, and condenses wherever the vapour pressure it would reach exceeds the saturation pressure at the local temperature. The Glaser construction makes the check in steady state: the temperature falls through the layers in proportion to their thermal resistance, the vapour pressure in proportion to their vapour resistance (the reciprocal of the permeance; a perm is one grain of water per hour per square foot per inch of mercury, 57.2 ng/(s·m²·Pa)), and the two profiles are compared interface by interface. Saturation pressure comes from the Hyland–Wexler equations of the site's psychrometric library, over ice below 0 °C. If the linear vapour-pressure line crosses the saturation curve, condensation forms at the interface where the excess is largest; the line is cut back to saturation there, and the rate of condensation is the vapour flux arriving from inside minus the flux leaving to the outside, each flux being a pressure difference over the vapour resistance of that side. The tool reports the plane, the rate and the amount accumulated over 30 days of design weather, the last as a comparison between designs: design conditions persist for days, and a seasonal moisture balance needs an hourly model with storage (NIST's MOIST program, SP 917, is the public-domain example).

Two rules from USDA Agriculture Information Bulletin 373 frame the result: a vapour retarder on the warm side should have 0.25 perm or less, and the cold-side layers should be at least five times more permeable than the warm side, so that what gets in can get out; a layer under 1 perm on the cold side traps the vapour. NIST SP 917 gives the permeances of finishes (latex paint 5 to 15 perm, oil paint 1 to 3, vapour-retarder paint under 1, vinyl wallpaper 0.5) that the tooltip repeats; the default plywood, foam and polyethylene entries are representative values and the adopted material data or ASHRAE Fundamentals chapter 26 should replace them in a real check. The dew point of the indoor air is printed as well: any surface colder than it, a window frame or a thermal bridge, condenses regardless of the wall's layers.

Reading the results and the limits

Table 1 is the compliance table: each row's UA, its own pass or fail, and the totals with the verdict. Figure 1 draws the two UA sets side by side so the trade is visible. Table 2 and the equations under it show how the layer U-factor was built; the value is written into the chosen assembly row when the box is ticked. Table 3 and Figures 2 and 3 are the condensation check: the dashed dew-point line of Figure 2 touching the temperature line, or a ratio above 1 in Table 3, means condensation at that interface. Table 4 is the annual estimate.

What the tool does not do: it is not REScheck or COMcheck and prints no certificate; the required values of an adopted edition, its mandatory backstops, its rules for crawl spaces, basements, raised trusses and steel framing, and its performance (simulation) path are the analyst's to apply. The parallel-path method is for wood framing; the ORNL metal factor is a rough figure. Thermal bridges at corners, band joists and window frames are not modelled. The degree-day method ignores solar gain, thermal mass and part-load equipment behaviour, and the Glaser check ignores moisture storage, air leakage (which moves far more moisture than diffusion), solar drying and the day-night cycle. Each is the standard first estimate, and each is reproduced in the test script from the public documents named above.