Cavity — series resistance

Total thermal resistance is the sum of the interior surface film, every material layer, and the exterior surface film. The U-value is the inverse of that total.

• Material conductivities (λ, W/m·K) come from a built-in library; pick Custom λ… to enter your own.
• Air gaps use fixed cavity resistances (not thickness ÷ λ) — horizontal 0.18, vertical 0.15 m²·K/W.
• Surface films are selectable for heat-flow direction and wind exposure; set to None for a surface-to-surface (R-value-only) result.
• Valid only for non-bridged, layered assemblies. For repeating thermal bridges use the Clear Field tab.

Assumes steady-state conditions and no moisture, convection, or 2-D/3-D bridging effects.

Clear Field — parallel-path correction

When an assembly contains repeating thermal bridges (studs, Z-girts, fasteners), each distinct cross-section is treated as an independent series path. Path U-values are area-weighted by their fractional areas, which must sum to 1.0.

• This is the upper limit of resistance — it assumes layers are isothermal and tends to over-predict U (conservative for bridged assemblies).
• The U-value increase reports how much the corrected U rises versus the cavity path (path 1): (U_corrected − U_cav) ÷ U_cav.
• Layer rows align across paths so the R-table can be read row-by-row.

The full method averages the upper and lower limits; this tool reports the upper limit only. For high-conductivity bridges (steel) consider a detailed 2-D/3-D calculation or χ/ψ point/linear transmittances.

Glazing — window assembly

Two interchangeable methods are offered via the toggle; both share the same geometry (glazing and frame areas derived from the diagram).

• Area-weighted EOG. The glazing is split into a centre-of-glass zone and an edge-of-glass band (a share of glazing area, default 10%). The edge U-value is the regression U_eog = a + b·U_cog with spacer coefficients — aluminium a=0.359/b=1.14, stainless a=0.182/b=1.04, warm-edge/TGI a=0.136/b=1.02. Each zone (COG, EOG, frame) is weighted by its fraction of the whole window. Pick Custom… to enter a measured U_eog directly.
• Geometry. A_pf = width × height (projected fenestration area). The glazing rectangle is the window inset by the frame border width on all four sides, so A_g = (width − 2·border)·(height − 2·border) and A_f = A_pf − A_g (frame fraction is derived and shown as a card). The glazing perimeter L_g = 2·(glazing width + glazing height) — so a tall, narrow window carries more edge length per unit area than a square one.
• Spacer ψ_g. Linear thermal transmittance in W/m·K. Typical values (non-metal frames): aluminium/metal ≈ 0.08, stainless ≈ 0.06, warm-edge/insulated ≈ 0.04; single glazing has no spacer (ψ = 0). Choose Custom ψ… (or drag the ψ slider) to enter a value from a spacer datasheet or a 2-D edge calculation.
• Window install ψ. An optional linear thermal bridge around the window-to-wall junction (W/m·K), graded very good → worse than typical, or Custom. Shown as a separate install ΔU = ψ·P ÷ A_pf over the outer window perimeter P — it is not folded into the headline U-window.
• Spacer edge ΔU reports ψ_g·L_g ÷ A_pf — the amount the spacer adds to the whole-window U-value.
• Selecting a frame preset fills the frame U-value; editing the U-value directly switches the type to Custom.

Fenestration benchmarks are whole-window U-factor targets. The climate-zone selector filters the zonal code-minimum target to the chosen zone; the non-zonal High-Performance Target is always shown. A simplified method — for ratings use certified whole-window values.

2D Detail — quasi-2D conduction

Solves the steady-state temperature field through a cross-section drawn as rectangles or polygons (orthogonal or angled). Shapes are rasterised onto a regular grid; each cell carries one conductivity, and a five-point control-volume balance (harmonic-mean face conductances) is assembled and solved by conjugate gradient. This computes the real 2-D heat flow that the Clear Field tab can only bound. A material background fill tiles the whole bounding box for simple layered sections; choosing Outside / void makes only the drawn shapes solid, so corners and angled exposed faces can be modelled.

• U-value (repeating section). The heat per square metre of wall face per degree (W/m²·K). Use it when the slice is a repeating wall section — it is the proper replacement for a Clear Field estimate, and it is what flows into the Library/Envelope. For a one-off corner or junction (which doesn't tile), the heat-flow figure is the more honest output.
• Heat flow. The actual heat (in watts) passing through a one-metre length of the drawn section at the indoor/outdoor temperatures you set — a tangible heat-loss number for the junction. Unlike the U-value it scales with the temperature difference, so re-solve after changing the temperatures. (This is the quantity the solver integrates directly; the U-value is derived from it by dividing through the section's face area.)
• What f_Rsi means. The temperature factor rates how warm the coldest interior surface stays, on a 0–1 scale: (coldest surface − outdoor) ÷ (indoor − outdoor). 1 means the surface is as warm as the indoor air (no bridge); 0 means it is as cold as outdoors. Higher is safer — a low value marks a cold spot where condensation or mould is most likely. As a rule of thumb, values below about 0.7 are a concern in cold climates.
• Draw, then paint. Shapes start material-less (hatched). Pick a material in the top bar, click the 🖌 Paint brush, then click each shape to fill it. The background fill (in Boundary conditions) defaults to an air gap; choose Outside / void to model only the drawn shapes (corners, angled faces).
• Boundary conditions. Surfaces exposed to air are convective (air temperature applied through R_si/R_se); everything else is an adiabatic symmetry cut — place those where no heat flows sideways (mid-stud, mid-cavity). Draw the interior and exterior boundaries as snapping lines (click vertices along the faces, double-click or Enter to finish), so a warm surface can wrap a corner or follow an angled face; the Eraser removes a line. The heat-flow direction (used for the area 2-D U, the air-cavity thickness, and the default interior film R_si) is read from where you placed interior vs. exterior — there is no separate axis setting.
• Air cavities are modelled as an equivalent conductivity derived from the cavity's drawn thickness and a fixed cavity resistance — internal convection and radiation are not modelled. Vapour-control / barrier layers are treated as thermally transparent.
• Condensation — climate-aware. The dew point (and an 80%-RH mould line) is found from the interior temperature and RH by the Magnus approximation, then placed in the field. The vapour-drive direction is read from the temperatures: interior warmer (T_in > T_out) drives vapour outward — risk toward the exterior; interior cooler drives it inward — risk toward the interior. The climate zone sets the severity baseline (Zone 4 and above → Condensation risk, below 4 → Low condensation risk). Draw a vapour-control line on the warm side and the verdict becomes Risk mitigated as long as the dew-point / mould isotherm never reaches the warm side of it; if it does cross, the baseline risk stands. With no climate zone selected the tool falls back to the direct coldest-surface-vs-dew-point check.

A simplified educational 2-D method (orthogonal geometry, temperature-independent conductivity, no radiation/convection in cavities). The condensation logic is a steady-state screening aid following PNNL BASC cold-weather guidance — not a transient hygrothermal (WUFI-style) analysis. For certified thermal-bridge values use a validated 2-D/3-D thermal-bridge solver.

Envelope

Combines individual assemblies into a single envelope figure. Each component contributes U×A to the total heat-loss coefficient; the area-weighted envelope U is that total divided by the total area.

• Add components by loading a saved construction (series / parallel / window) or pulling the assembly currently live in a tab. The U-value is taken from that construction; enter the area it covers.
• Thermal bridges are optional. Linear bridges use a linear transmittance ψ (W/m·K) over a length; point bridges use a point transmittance χ (W/K) over a count. Both add to UA but not to ΣA, so they raise the effective envelope U.
• U-values are snapshots taken when the component is added — re-add or edit a row if the source assembly changes.

A first-order fabric heat-loss estimate. It excludes ground-coupling adjustments, air infiltration/ventilation, and dynamic effects — for compliance use a full envelope/energy model.

Library

The library stores any construction you save from the Series, Parallel-path or Window tabs. It persists in this browser only (nothing is uploaded), so clearing site data or switching machines starts it fresh — use the per-tab JSON Save/Load to move a construction between browsers.

• Each entry shows its live-recomputed U-value and can be renamed inline.
• Projects are simple labels for grouping. Create one, then assign constructions to it with the per-row dropdown; deleting a project returns its constructions to Unsorted.
• + Env adds a construction to the Envelope tab as a component (a U snapshot — editing the library entry later won't change components already added).