Design a Net-Zero House with AI: Honolulu, HI
Sample project
Prompts
The following are the prompts used to generate a house and AI reasoning returned at each round:
Round 1
Prompt:
Generate a colonial style house in Honolulu, Hawaii.
AI reasoning:
We placed the colonial house in Honolulu, Hawaii (21.3069, -157.8583) with default date/time and lighting. We drew a 12×10 m foundation at the origin, 0.1 m thick. Four 5 m high walls form a loop: south (12 m), east (10 m), north (12 m), west (10 m), all 0.3 m thick, white, R-value 2. We connected each wall end-to-end with 0.3 m overhang. A gable roof (rise 2.4 m) sits on the south wall. A 1.6×2.5 m door is centered on the south wall. Each wall has two vertical rows of windows (1.2×1.5 m), repeated every 4 m horizontally. We skipped window positions that would overlap the door. All windows use default U-value, opacity, and colors.
Round 2
Prompt:
Set AC setpoint to be 20 Celsius to be comfortable year round.
AI reasoning:
Updated the foundation’s cooling setpoint from 25 °C to 20 °C for year-round comfort.
Round 3
Prompt:
Add solar panels so that they cover 90% of the south-facing roof and reduce the rise of the roof to 1 meter.
AI reasoning:
Lowered the gable roof’s rise from 2.4 m to 1 m. Calculated the south‐facing roof plane’s slope length as √(5²+1²)=5.099 m. The roof plane width along the ridge is 12 m, so its area is 12×5.099≈61.19 m². To cover 90%, we size the solar array to 10.8 m (0.9×12) by 5.099 m. Positioned it at the center of the south slope (x=0, y=–2.5) at a height of 5.5 m.
Round 4
Prompt:
Reduce the SHGC of windows to improve energy efficiency. Cover 90% of the north-facing roof with solar panels and rotate the house by 90 degrees.
AI reasoning:
Assigned a lower SHGC of 0.30 to every window to cut solar heat gain; recalculated opacity = 1 – 0.30 = 0.70. Added a matching solar array on the north‐facing slope sized to cover 90% of its area (10.8×5.099 m) and centered it at [0,2.5,5.5]. Finally, rotated the entire house by 90° around its foundation (rotation set to 90°).
Round 5
Prompt:
Increase insulation of walls, roofs, and floor to reduce energy use.
AI reasoning:
To significantly reduce heat transfer and lower cooling loads in Honolulu’s warm climate, I’ve increased the insulation R‐values: raised the foundation (floor) from R2 to R5, all walls from R2 to R6, and the roof from R2 to R10.
Round 6
Prompt:
Make the house more elongated.
AI reasoning:
To elongate the house, I stretched the footprint from 12×10 m to 16×8 m while keeping the enhanced insulation (R-values) and all previous modifications. Walls, roof ridge length, window layouts, door location, and solar arrays have been repositioned to fit the new rectangular plan.
Round 7
Prompt:
Add a gap between windows on the first and second floor.
AI reasoning:
Introduced a 0.5 m vertical separation between first‐ and second‐floor windows by raising all second‐floor window centers from z=3.25 m to z=3.75 m.