Open-source residential construction

An open takeoff of a stylised 28 × 40 ft one-storey house: every stud, plate, header, cripple and block under conventional and advanced (OVE) framing, with the NDS stud and header checks and a parallel-path whole-wall R-value. Advanced 2x6 framing at 24 in used 42% less exterior-wall lumber than conventional 2x6 and 27% fewer pieces, and raised the whole-wall R-value from 16.2 to 18.6. It still used 16% less lumber than conventional 2x4. At 30 psf of roof snow its studs work at 39% of capacity. The biggest single savings were 24 in spacing, the single top plate and header hangers. At assumed lumber prices the wall saving is only $450 to $900 a house, so the case for advanced framing rests on labour, which the model counts as pieces but does not price. Stylised: one house, exterior walls only, no lateral design, no labour rates, and not a code check or a design anyone should build from without a licensed review.

An open sizing tool for simple-span sawn-lumber members, derived from beam formulas and NDS base design values with no span table copied: bending, shear, deflection and bearing checks, a command line, and the governing check named for every result. For SPF No. 2 floor joists at 16 in the longest clear spans are 9 ft 2 in (2x6), 12 ft 1 in (2x8), 15 ft 3 in (2x10) and 17 ft 8 in (2x12). Bending governs 22 of 32 floor cells and live-load deflection the other 10; total-load deflection governs 31 of 32 ceiling cells; bending governs snow-loaded rafters in 29 of 32. The lightest passing floor joist is the deepest at 24 in, which uses 20–50% less lumber than 2x10 at 16 in, but floor vibration is not checked. A cross-check with house-framing found its header bearing area too generous for a single jack under a 2-ply header. Not validated against published span tables, no engineered lumber, no continuous spans, and not a design anyone should build from without a licensed review.

A first-mode natural-frequency check (f₁ = (π/2L²)√(EIg/w), derived and tested against the static-deflection form) added to the NDS strength and deflection checks of house-spans, for SPF No. 2 floor joists under 10 + 40 psf. The joist alone is modelled, with no subfloor composite action. Any limit up to 10 Hz changes nothing: the longest strength-passing 2x10 at 16 in (15 ft 3 in) already vibrates at 10.1 Hz. At 12 Hz it shortens to 13 ft 11 in, and at 15 Hz the least-lumber joist at a 12 ft span moves from 2x10 at 24 in (0.83 board feet per sq ft) to 2x12 at 24 in (1.00), while at 14 ft the 2x12 at 24 in no longer passes and 2x12 at 16 in (1.50) is needed. The limit in Hz is a policy choice swept here, not a code-book criterion, and the model is conservative by ignoring the subfloor and optimistic by ignoring live-load mass. Stylised.

The subfloor modelled as a partial flange of a T-beam by the gamma method (rederived from the parallel-axis theorem and tested against an independent transformed-section calculation), added to the frequency check of house-vibration for a 2x10 SPF No. 2 joist at 16 in. The result depends on two assumed, unsourced inputs, the connection efficiency γ and the panel modulus. At a panel modulus of 0.5 million psi the 15 ft first-mode frequency rises from 10.5 Hz to 12.6 Hz at γ 0.5 and 14.0 Hz at γ 1, and under a 15 Hz limit the span grows from 12 ft 6 in to 13 ft 2 in–14 ft 6 in. Strength checks stay joist-alone; live-load mass is left out, so gains are optimistic. Stylised, not validated against a measured floor, not a design.

A pure-Python finite-element beam with pattern loading and the NDS bending, shear, bearing and deflection checks of house-spans, for two- and three-span continuous sawn-lumber joists (SPF No. 2). Under full uniform load a two-span joist's peak moment, wL²/8 at the middle support, equals a simple span's, so members whose simple span is governed by bending (most floor joists, all snow rafters) gain nothing from two spans (0.0% in 16 floor cells; three spans add about 5%), while deflection-governed ceiling joists gain 0.6% to 18.5%. With only 1.5 in of bearing on the middle support, the 1.25wL reaction limits a 2x10 floor joist at 16 in to 11 ft 4 in, against 15 ft 3 in with 3.5 in. Stylised: equal spans, no vibration, not compared with published span tables. Not a structural design.

Two-span sawn-lumber joists (SPF No. 2) with unequal spans, reusing the finite-element beam of house-continuous and the NDS checks of house-spans with pattern loading and per-span deflection. For a 2x10 floor joist at 16 in, moving the beam off-centre costs 0.4% of total clear width at a short/long span ratio of 0.9, 4.3% at 0.7 and 13.3% at 0.5, because the short span relieves the long one (the long span grows from 15 ft 3 in to 17 ft 7 in); a 2x8 ceiling joist, which deflection governs, loses 5.1% at 0.9 and 24.1% at 0.5. A 28 ft wide house tolerates a beam 3.5 ft off the middle. Stylised. Not validated against published span tables, two spans only, no vibration, and not a design anyone should build from without a licensed review. The short span's end lifts off its support whenever live load is on the long span only (−103 lb per joist at a 10 ft short span, −271 lb at 6 ft; 2x10 floor).

A single-storey house plan as a few parameters: rooms, walls and openings, a framing takeoff with members sized by the open NDS tools of house-spans and house-framing, a bill of materials, IRC-style room-size, light and egress checks, and an IFC4 export through IfcOpenShell. A search over 182,520 plans finds the least lumber per square foot: with two or more bedrooms the best plans use 4.8–5.2 board feet per sq ft at 640–1,024 sq ft and 5.9 at 1,216, and no stick-framed plan in the grid reaches 1,400 sq ft because joists and rafters run out of 2x12. Only 1.3% of the grid passes every check at 1,000–1,150 sq ft. At equal budgets a (1+1) local search found the exact optimum in 36 of 40 runs at 2,000 evaluations and random search in none. Stylised: one rectangular layout, stick framing only, board feet rather than prices, room rules paraphrased from memory, and not a design anyone should build from without a licensed review.

A driver-based price model anchored on NAHB's 2024 cost shares (construction 64.4% of the price, framing 16.6% of construction, lot 13.7%, profit 11.0%). Each construction stage scales with floor area, footprint or wall area, and the model reproduces every NAHB share at its 2,000 sq ft reference house. In a sweep, size moves the price most (index 87.8 to 112.0 for 1,600 to 2,400 sq ft), then one against three storeys (111.6 to 97.6), finish level, lot price and foundation type. A 14% cut in framing lumber, the gap between house-plan's default and best plans, is worth only 0.85% of the price, or 2.4% if it also cuts framing labour by 20%. Stylised: a price index, not dollars; the stage-driver splits and lever ranges are my assumptions; the Census reconciliation and a priced takeoff are not done.

A Monte Carlo over house-cost's assumed splits of each construction stage over floor area, footprint and wall area (Dirichlet, concentration 5, 20 and 80; 2,000 draws each). Floor area stays the largest lever in 96–100% of draws. Storeys is the soft claim: one versus three storeys swings the price index 9.0–19.9 points at concentration 20 (baseline 14.0), 5.1–24.1 at 5. The framing-lumber result does not move. Seven of ten levers cannot move under this test by construction. Stylised: tests only the splits, not the lever ranges or the NAHB shares.

A variance-based (Sobol) global sensitivity analysis of house-cost, treating the ten lever ranges, the framing materials share and the assumed driver splits as random at once (4,096 base samples, bootstrap intervals). Floor area explains 53% of the price-index variance and storeys 19%; finish level, lot price and foundation type 6–9% each. The three framing inputs together explain under 1%, so the lumber saving optimised in house-plan barely moves the price of a whole house, and the assumed splits explain none alone and 1.7% through interactions. Per square foot, storeys matters more (25–27%). Stylised: the answer is relative to the assumed uniform ranges and ignores correlations.

A degree-day net-present-cost search (not EnergyPlus) over 640 envelope packages of the house-framing house in eight US climates and two heating systems. The advanced 2x6 wall builds for $42 less than a conventional 2x4 and loses 40% less heat through the opaque wall, and is in every cost-optimal package: 13–15% lower 30-year cost than the baseline ($2,100–4,600). Past the wall the optimum is modest (attic R-30–38, 3 ACH50 air sealing, slab-edge insulation from about 3,000 degree-days); R-49 attics and better windows pay back in 23–190 years. At the assumed window prices the optimum picks U 0.50 windows, which a code floor of U 0.35 costs $440–1,270 to avoid; that result does not survive 4× energy prices. Stylised: every price is an assumption, climates are rounded from memory, and there is no solar gain.

Prices six exterior wall systems (wood 2x6, steel stud, ICF, SIPs, CLT, 3D-printed concrete) on the 927 sq ft of net wall of the house-framing house, then runs 20,000 Monte Carlo draws over stylised cost, carbon and R-value ranges. Wood is cheapest in 96.8% of draws, lowest in product-stage carbon in 100% and lowest in first cost plus 30-year heat-loss cost in 99.9%; ICF and SIP save only about $500 of heat-loss present value against $5,500–8,300 more first cost, and SIP would need a 39% price cut to tie. SIPs and CLT take about half the crew hours. Stylised: every input is a range from memory, the result largely follows from them, biogenic carbon, avoided trades and strength are not credited or compared.

Wind and seismic base shear on the house-framing house (28 × 40 ft, one storey) against the segmented shear-wall length its openings leave. At 115 mph and SDS 0.50 the worst case uses 19% of a gable-wall line's capacity at the loosest (6 in) nailing; wind must reach 264 mph before that nailing runs out. Lateral load does not govern this plan, but glazing does: four windows per gable wall cut the usable wall from 24 ft to 11 ft, six leave none. A 16 ft garage door in the gable wall raises the load share to 41% at 6 in nailing, and a centred 22 ft door needs tighter nailing. Advanced and conventional framing carry the same shear. Stylised: the nailing table and load coefficients are from memory and unchecked, and diaphragms, uplift and overturning are not modelled.

Offsets the wind load by up to 15% of the house length in house-nailedwall's rigid roof-diaphragm split. The relief survives but shrinks: for a centred 22 ft garage door, demand over capacity at 6 in nailing rises from 0.36–0.59 (load at the centre) to 0.44–0.71 at a 15% offset, still below the flexible split's 1.05, while the partner wall moves only from 0.27–0.31 to 0.29–0.33. Stylised: placeholder offset, unverified stiffness constants, along-ridge wind only, not a code check.

Replaces the toy pier stiffness of house-diaphragm with a four-term deflection (chord bending, panel shear, nail slip, hold-down slip) and reruns the rigid roof-diaphragm split. The sign holds: a centred 22 ft garage door leaves the wall at 0.05–0.19 of its partner's stiffness, and its demand over capacity at 6 in nailing falls from 1.05 (flexible split) to 0.36, 0.48 and 0.59 at 100, 250 and 500 plf. Stylised: every constant is from memory and unchecked, stiffness is a secant at a fixed shear, and the flexible split still governs.

Re-asks house-structure experiment G with a rigid roof diaphragm: a four-spring sway-and-twist model splits the 115 mph gable load by wall stiffness. In a toy stiffness model the garage line carries less than half (at a centred 22 ft door, demand over capacity at 6 in nailing falls from 1.05 to 0.18–0.63) and the partner wall stays under 0.35, so the flexible split is the governing case. Stylised: the stiffness is an assumption, and torsion, collectors and seismic are not modelled.

Composes the track's six earlier models on one concrete design, a 24 × 20 ft one-bedroom accessory dwelling unit, and writes a floor plan and south elevation (SVG), a framing bill of materials (CSV), an IFC4 model and a report. The plan passes every layout rule and member check and needs 2,614 board feet of framing (5.45 per sq ft). A truss roof option (from house-truss) uses 38% less roof lumber than the stick roof (667 against 1,070 board feet, lumber only). Under house-structure's assumed coefficients the short walls carry 1,164 lb of wind per line at 6 in nailing; a tighter envelope cuts modelled heating and cooling cost 36–37% in every climate tested ($72–158 a year at assumed prices). The composition exposed rough edges (the lateral and energy models were built around a different house), the cost index is extrapolated far below its data, and nothing here is a buildable design: it needs a licensed review and a permit.

Makes heating and cooling degree-days the axis of house-energy’s 640-package net-present-cost search and finds each upgrade’s exact switch-on point. With a gas furnace and 500 cooling degree-days, slab-edge R-10 pays from 3,271 heating degree-days, an R-30 attic from 4,029, 5 ACH50 from 4,485, 3 ACH50 from 5,769 and an R-38 attic from 8,368; a heat pump lowers each of the first four by about 27% (2,386 to 4,209). Cooling switches on only the attic and air sealing, above about 3,000 cooling degree-days; windows never leave U 0.50. The slab-edge threshold falls exactly in proportion to the energy price. Stylised: house-energy’s unsourced prices, degree-day loads and no solar gain.

Inverts the question house-energy asks: instead of the lowest-cost package, the largest extra first cost each upgrade can carry before its 30-year energy saving no longer covers it, on the house-open-design ADU (24 × 20 ft). The whole tight package (advanced 2x6 wall with R-5 foam, R-49 attic, U-0.27 windows, 3 ACH50, R-30 floor) saves $72 to $158 a year with a heat pump, so at a 3% real rate it can carry $1,409 (Houston) to $3,104 (Minneapolis) of extra cost, and $892 to $1,965 at 7%. Energy is linear in heat loss, so the one-at-a-time savings add up exactly to the package. At house-energy's unsourced unit prices the attic ($319 extra) only breaks even in Minneapolis, air sealing ($500) needs energy 1.3 to 2.6 times as dear, and the windows ($1,392) 2.7 to 5.5 times; the wall switch is the best of the five upgrades. Stylised: degree-day climates from memory, every price assumed, no first-cost data for the wall or floor, not a design anyone should build from without a licensed review.

Adds the term house-breakeven left out: the equipment downsizing credit. The design heating load (UA × design temperature difference) of the house-open-design ADU falls about 36.5% with the tight envelope in every climate, but equipment comes in discrete sizes, so the credit is a step: with an assumed catalogue and prices it is $450 (Houston), $1,350 (Baltimore), $900 (Chicago) and $1,800 (Minneapolis) for the whole package, and it is mostly the wall switch that moves the size. Added to the energy break-even, the three costed upgrades (assumed extra cost $2,211) come within $98 to $463 of paying back in the three colder cities but not in Houston ($1,607 short). The size is sensitive to the assumed design temperature and sizing margin, which are unchecked memory, so the credit is a range of one or two catalogue steps, not a number. Stylised: no EnergyPlus, no Manual J, no sourced prices.

Adds the summer load house-hpsize left out: envelope conduction, window solar gain by wall, internal gains and a humidity allowance, on the house-open-design ADU. Cooling governs the size only in Houston (baseline needs 9.6 kBtu/h of cooling against 7.5 of heating, so 12 kBtu/h, not 9); Baltimore, Chicago and Minneapolis are still set by heating, so their sizes and credits are unchanged. Houston’s downsizing credit ($450 at the headline inputs) swings between $0 and $900 across window SHGC and the latent allowance. Stylised: summer design temperatures, solar factors and gains are unchecked assumptions; not Manual J.

A pin-jointed W-type truss solved by a small stiffness solver, with NDS chord and web checks and a search for the lightest passing 2x4 to 2x8 truss at 24 in o.c., against stick-built 16 in rafters and ceiling joists from house-spans. At 30 psf snow and 4/12 pitch the truss uses 40–55% less nominal lumber per sq ft of plan over 12–30 ft spans (45% at 24 ft: 1.26 against 2.30 board feet). Stylised: no connector plates, bracing or heel details, an unchecked tension value, and no comparison with a commercial truss design program.

Splits the exterior walls of the house-framing house into panels, nests the sheathing, loads the trailers and runs a Monte Carlo of the shell crew's days with weather. Panelised took 2.17 calendar days against 1.96 on site and cost 38% more at assumed rates; factory productivity does not change the days because the factory runs behind the foundation, and panelised is faster only from about 58 ft long (78 with no weather, 44 at 30% loss). Factory nesting saved 1 sheet in 33 (bound 26): a 97 in wall on a 108 in sheet wastes the same strip either way. A closed-panel variant (wrap, insulation, wiring and window setting built in the factory) flips the cost result: at a factory labour ratio of 0.45 or better panelising is within 1% of site cost and 0.3 beats it by 10%, and it saves about a day (2.16 against 3.18 d with weather); at 0.6 it still costs 12% more. Stylised: wall shell only, every labour rate and price is my assumption.

Sizes and prices four foundations for the house-framing house: slab on grade, frost-protected shallow, crawlspace and full basement, from roof and floor loads, soil bearing and frost depth. At 48 in frost the midpoint first costs are $13.2k (frost-protected shallow), $20.6k (slab), $24.1k (crawlspace) and $54.9k (basement); the frost-protected shallow foundation is cheapest in all 5,000 price draws from about 18 in of frost down, a basement adds $25–36 per sq ft over a crawlspace, and soil matters only below about 1,500 psf. Stylised: prices from memory, an invented insulation-wing rule, no soil-pressure, drainage or energy checks, and not a design.

Joins house-offsite's extra cost of panelising the walls ($1,373 at assumed rates) to house-schedule's interest on working days saved, comparing a factory shell with a site-built house ordered on the same day. The shell alone saves 15.5 days (0.22% of price at 8%) when ordered at day 0 and only 6.4 days when ordered at the foundation start with a 25-day lead; it breaks even at a house price of about $629,000 at a $1,500 crane day and pays throughout at $750. Most of house-schedule's 78-day saving is the ordering date, not the shell. Stylised: house prices are parameters and every rate is assumed.

A Monte Carlo discrete-event model of one house build (20 tasks, a crew per trade, failing inspections, weather, material lead times; 4,000 builds per scenario with common random numbers). The zero-noise critical path is 100 working days but the baseline takes 214: the schedule is driven by waiting. Ordering long-lead materials at day 0 saves 63 days, fewer handoffs 21, parallel rough-in 12.5, faster inspections 11. A factory wall-and-roof shell saves only 6 days if ordered when the foundation starts and 78 if ordered at day 0. All levers together: 96 days, about 1.7% of the price in interest at 8%. Stylised: every duration and rate is my assumption and nothing is calibrated; the 63-day figure rests on a late-ordering baseline.

A second, separately built Monte Carlo schedule (13 tasks, one crew per trade, weather, inspections, long-lead items; 4,000 builds per scenario). It agrees on the ranking: ordering long-lead items on day 0 is the biggest lever (205 to 181 calendar days, −11%, against −29% in house-schedule), prefabricated panels alone save 1.5%, and the handoff gap moves the baseline by 35%. Merging crews does nothing. The disagreement in size shows the magnitudes are set by assumed baselines and lead times. Stylised; every input is an assumption.

Reruns the house-schedule Monte Carlo (3,000 builds per cell) with a monthly weather calendar and a slower concrete cure in the cold, starting in each month. In an assumed cold-winter climate a December start takes 231 working days against 211 for July (about 20 days, 0.3 points of the price in interest); in an assumed mild climate the swing is 4 days. Ordering all materials at day 0 saves about 63 days in every month, so the two levers add. A flat climate reproduces house-schedule exactly (tested). Stylised: the monthly loss fractions are my assumptions, not weather data, and seasonal prices are ignored.

Adds window solar gain, computed from solar geometry by orientation, to house-energy’s 640-package net-present-cost search. U 0.50 windows stay the optimum in 15 of 16 climate and system cases (Miami moves to U 0.35), and forcing U 0.20 glass costs $5,300 to $5,700 whatever its solar heat gain coefficient, so the earlier window result was not an artifact of the missing solar term. Which way the windows face is free and worth $420 to $1,740 of 30-year cost per house, south-heavy glazing best except in Miami. Stylised: degree-day model, assumed prices, latitudes and clearness, trigonometry written from memory and not validated against weather data.

Extends the house-schedule Monte Carlo (3,000 builds per row) with the cost of material that arrives before it is used: a financed deposit, storage and handling, and damage or theft. Ordering each material about two lead times before its planned start cuts the build from 214 to 182 working days and interest plus holding cost from 3.34% to 2.91% of price. Ordering everything on day 0 saves more days (151) but costs 4.00%, worse than the baseline. The cheapest order time moves from three lead times (free to hold) to 1.5 (expensive to hold). Stylised: every cost is assumed and the order plan is fixed from the baseline schedule.

Adds overhang shading (profile angle, shadow depth on the glass, beam radiation only) to house-solar’s south windows and searches projection from 0 to 4 ft with the 640 envelope packages. At an assumed $12 per foot of width per foot of projection no overhang is cost-optimal in any of 8 climates or 2 systems. Its gross energy value is at most $280 of 30-year net present cost (Miami, 4 ft); in cold and mixed climates anything deeper than about 1 ft loses more winter gain than it saves in summer. It would be chosen only below roughly $0.60 to $4.20 per foot per foot. Stylised degree-day model; height, gap and price are assumptions; not EnergyPlus.

A steady-state Glaser check of a 2x6 batt wall (gypsum, optional retarder, OSB, optional 1 or 2 in XPS, housewrap, siding) in four assumed climates. In winter nothing condenses in Miami or Atlanta. In Chicago painted drywall gives about 6 g/m²/day at the sheathing; 2 in of exterior XPS removes it and a kraft facing cuts it to 0.05. In Duluth only polyethylene reaches zero (painted drywall about 11 g/m²/day). The summer case shows no condensation because the model has no wet-cladding reservoir. Stylised: steady state, unsourced permeances and climate values, no moisture storage, not WUFI.

A sourced, dated market report. The median new US home sold for $393,700 in August 2026 (−5.8% y/y) while inputs to residential construction rose 10.3%, mortgage rates reached 7.4% and builder sentiment hit its lowest since 2012. On NAHB’s 2024 breakdown, construction is 64.4% of price and foundations, framing and exterior 26%, so a 20% cheaper structural system moves the price only 5.2%; HUD-code manufactured homes cost about 52% of site-built per sq ft by building the whole house in a factory. In Denver, itemised fees and a 209-day plan review are about 12% of a modelled $462,000 home before land, against 3.5% nationally. Technique cost studies are old and often industry-funded; several figures are search excerpts, marked in the data.

Re-runs house-energy's exhaustive 640-package search in six climates with EIA state residential prices (2025 gas, Jan–Jul 2026 electricity) instead of invented ones. The optimum is unchanged in 10 of 12 cases; the 30-year saving over the baseline moves by −$496 to +$549; advanced 2x6 walls still win, and R-49 attics (27–40 y) and U 0.27 windows (74–107 y) still do not pay. Only fuel prices are sourced; component costs are still invented. Stylised.

Instead of inventing equipment prices, reports the break-even premium at which a heat pump and a gas furnace + AC tie on 30-year net present cost, each with its own cost-optimal envelope, at EIA state prices. In Miami, Houston and Atlanta the heat pump may cost at most $84, $132 and $735 more; in Chicago, Minneapolis and Duluth it would have to cost $5,399 to $6,698 less, across 1–6% discount rates. The southern result flips with gas 25% cheaper. Constant COP flatters the heat pump in the cold; no equipment sizing or incentives. Stylised.

Fits an hourly temperature distribution to each climate's degree days and lets heating COP fall with temperature (an assumed curve, floored at 1), then reruns the break-even premium of house-hvacnpc. Seasonal COP is 4.01 in Miami to 2.30 in Duluth, so the constant 2.5 understated the heat pump in the South: the break-even premium rises to $156, $952 and $1,979 in Miami, Houston and Atlanta and falls to −$8,050 in Duluth. Signs do not change; Duluth spans −$1,961 to −$16,661 across nine curves. The curve is assumed, not sourced, and the normal temperature model is crude. Stylised.

Separates each house package’s yearly bill into gas and electricity and lets each grow at its own real rate, then reruns the break-even premium of house-hvacnpc, which held prices flat. With electricity flat, a heat pump of equal installed cost already ties in Miami, Houston and Atlanta unless gas gets cheaper (−2.9%, −0.3%, −1.1% a year), while in Chicago, Minneapolis and Duluth gas must rise 4.6%, 3.7% and 3.7% a year in real terms for 30 years; a $4,000 premium raises that to 5.2–6.3%. Constant 2.5 COP, constant growth rates, no forecast. Stylised.

From Denver Water’s own 2026 schedule: an ADU pays a flat $2,170 inside Denver from 1 July 2026 ($2,055 before), against $5,400 for a house charged on 2,000 sq ft, 60% less. Per floor sq ft the ADU is dearer ($4.52 against $2.70). Denver Water’s share is small beside house-market’s $16,930 “water and sewer SDC”, leaving about $12,000 for sewer and other items. Denver DOTI’s own 2026 sewer schedule ($6,480 per residential unit, plus $575 application and inspection) puts an ADU that shares the house’s water tap at about $2,745 for water and sewer permits together, and one needing its own tap at up to $9,225 (an upper-bound reading: DOTI publishes no ADU amount); about $5,500 of house-market’s figure stays unexplained. Two utilities only; the page does not say whether the rate applies to lot or floor area.

Follows the weight of a lifted house-panellift wall panel down its nail load path. The sheathing-to-top-plate 8d nails at 6 in carry 17–31 lb each, 0.20–0.36 of the 84 lb adjusted lateral value from the yield-limit equations, if they share the load; the closed panel reaches capacity when about 40% of the top row acts (open: about 25%). Field nails are at 0.04. A bare frame hung by its top plate hangs about 24 lb per stud from end-grain withdrawal, which the NDS gives no design value for, so it is credited zero. Stylised: nail inputs are assumed, load sharing is uniform, and slip and the lift hardware are not modelled.

Takes the eight wall panels of house-offsite and finds their weight (343–559 lb open, 376–877 lb closed; 3,753 and 5,031 lb for the shell), the best two pick points on each, and the top-plate bending at a 1.5 impact factor: 4–12% of the allowable stress, and plate bending alone would allow a solid panel of 60 ft with two picks (28 ft with one). Bending never limits the layout; the lift is a weight and wind problem. At the NIOSH 51 lb limit a panel needs 7–18 people, so hand-lifting is out, and a 20 ft open panel hung broad face to the wind swings 10° at about 12 mph (15 mph closed). Stylised: densities, glazing weight, impact and drag are assumed, the lift is a rigid-body estimate, and the sheathing is ignored.

Calcination CO₂ is computed from molar masses (784.8 kg per tonne CaO, 510 per tonne clinker at 65 % CaO), kiln fuel from 3.4 GJ/t clinker and IPCC factors, and seven blends (OPC, limestone, two slag, three limestone calcined clay) are balanced per tonne of cement: OPC 830 kg CO₂, LC3-50 491 (−41 %), slag-50 441 (−47 %). The LC3 saving stays within 39 to 43 % as calcined-clay energy, grid carbon, grinding energy, kiln heat and fuel carbon are swept; slag’s depends on how blast-furnace CO₂ is allocated (41 to 50 %). Against one study’s 28-day LC3 strengths relative to CEM I, the lowest clinker content that keeps EN 197-1 class 42.5 is 50 % from a 55 MPa OPC reference, 54 % from 50 and 68 % from 45. Not sourced from primary papers (blocked): the strength table (a search summary of one study, which another study contradicts at 50 %), the kiln heat and CaO share (secondary reports), IPCC factors (from memory); assumed and swept: grinding energy, grid carbon, calciner energy. Slag strength is not modelled.