Does a temperature-dependent heat-pump COP change how much more a heat pump can cost than gas before gas wins?
research/house-coldcop
Does a temperature-dependent heat-pump COP change how much more a heat pump can cost than gas before gas wins? Seasonal COP 4.01 (Miami) to 2.30 (Duluth) under an assumed curve, so the constant 2.5 understated the South: break-even premium $156, $952, $1,979 in Miami, Houston, Atlanta and −$8,050 in Duluth; signs unchanged, size curve-dependent. Stylised.
What it found
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.
A figure from its own output
Paper, code and results
More in Open-source residential construction
- How fast must gas prices rise against electricity before a heat pump ties a gas furnace on 30-year cost?How fast must gas prices rise against electricity before a heat pump ties a gas furnace on 30-year cost? At equal installed cost the heat pump already wins in Miami, Houston and Atlanta unless gas gets cheaper; in Chicago, Minneapolis and Duluth gas must rise 4.6%, 3.7% and 3.7% a year in real terms for 30 years. Stylised.
- What does Denver Water's 2026 tap fee cost a 480 sq ft backyard cottage (ADU) against a house?What does Denver Water's 2026 tap fee cost a 480 sq ft ADU against a house? A flat $2,170 for the ADU against $5,400 for a house charged on 2,000 sq ft, from the utility's own schedule; Denver's own sewer schedule adds $575 (shared water tap) or up to $7,055 (own tap), so about $2,745 to $9,225 in all.
- Do the nails carry a hung wall panel, and how few can act before they fail?Do the nails carry a hung wall panel? Sheathing-to-top-plate nails take 17–31 lb each (0.20–0.36 of capacity) if they share the load; the closed panel fails when under about 40% of the row acts, and a bare frame's stud end nails get no credit.
- Can a crane or telehandler lift the house's wall panels, and does anything but weight and wind limit it?Can a crane or telehandler lift the house's wall panels? Weights of the eight
house-offsitepanels (343–559 lb open, 376–877 lb closed), best pick points and top-plate bending (4–12% of allowable with two picks): bending never limits the layout, weight and wind do. Hand-lifting needs 7–18 people; a 20 ft open panel swings 10° at about 12 mph. Stylised, every density and coefficient assumed. - How much CO2 does clinker substitution save, and how far can it go before a cement loses its strength class?Mass, energy and CO2 balance of a tonne of cement: OPC 830 kg CO2, LC3-50 491 (41% lower, 39-43% over every swept input), 50% slag 441; the clinker can fall to 50% and keep class 42.5 only if the OPC reference has 55 MPa of strength, to 54% at 50 MPa.
- How much lumber, labour and heat does advanced framing save in a house's exterior walls, and does it still carry the roof?How much does advanced (OVE) framing save in a house's exterior walls? Stylised 28 × 40 ft house: 42% less wall lumber and 27% fewer pieces than conventional 2x6, whole-wall R 16.2 to 18.6, studs at 39% of capacity under 30 psf snow; the lumber is worth only $450–900 a house, so labour carries the case. Not a code check. First project of track 12.
All 41 Open-source residential construction projects →
Other research topics
Proper scoring & elicitation28Effort & peer prediction9Verification games & slashing19Detection & tolerances14Markets as routers13Prediction-market design9Paying contributors8Decentralised training systems23Digital twins & sim-to-real164LLM chip design loops40Robot manufacturing loops53Ambient AI assistant46Self-replicating manufacturing34Simulation-first design64Engineered clothing45Task-specific AI accelerators40Algorithmic chip-design loops42Open-source purpose-built robots44The software-engineering loop43Adversarial human imitation41Open-source residential construction41Autonomous multi-angle capture34Sensing, imaging and 3D31Energy supply and dense energy storage39Robot societies among people31Agentic co-design of low-cost robots27Laundry-hanging robots25Machine learning on qualitative data23Robots on the building site27Civilizations & group psychology1History & human sciences1
Produced with AI assistance (Claude). Results are from the code in this repository and are simulated or stylised unless the abstract says otherwise; nothing here implies affiliation with or endorsement by any lab, researcher or company cited. MIT licensed.