Pease Studio

Beta

Heat Balance Studio

Envelope heat balance for early-stage design

Open Heat Balance Studio
SourceMIT
  • Eighteen building types from the PNNL prototype models
  • Five envelope surfaces, by U-value or R-value
  • Gains and 24-hour schedules per type, all of them editable
  • A design day derived from ten years of hourly weather
  • Balance-point temperature and wall-to-floor ratio
  • Runs in the browser — no account, no upload

Can this building be designed to need no heating at all?

That question is usually answered far too late, once there is a model and a mechanical system and a great deal of committed geometry. But the things that decide it — how much wall there is per square foot of floor, how much of that wall is glass, how much heat the occupants and their equipment already make — are all fixed in the first sketch.

Heat Balance Studio puts the two sides against each other hour by hour on a cold design day. Pick a building type, name a US city, enter five envelope surfaces, and see where the gap is — or what you have in hand if there is not one.

Heat Balance Studio — envelope heat loss vs internal heat gain A building section in the style of a hand sketch. Five terracotta arrows point outward from the envelope, labelled walls, windows, roof, ground floor and exposed floor, each with a heat loss in watts. Three green arrows point inward from within the building, labelled people, lighting and equipment, each with a heat gain in watts. Hatched ground runs beneath and beside the building, with grade drawn as a single line. WALLS 1200 W WINDOWS 900 W ROOF 700 W GROUND FLOOR 400 W EXPOSED FLOOR 300 W PEOPLE 900 W LIGHTING 500 W EQUIPMENT 700 W HEAT LOSS — ENVELOPE HEAT GAIN — INTERNAL

The two sides of the balance. Five envelope surfaces losing heat outward, the occupants and their equipment pushing back — and the arrows are sized by the numbers, so the shape of the problem is visible before any of it is read.

Eighteen buildings, and the schedules are the half that matters

Pick Medium Office or Warehouse or Full Service Restaurant and the occupancy density, the equipment load, the lighting and all four 24-hour schedules arrive together, from the PNNL prototype building models. The section drawing changes with the choice; six massings cover the eighteen types.

The schedules are the half people underestimate. An apartment sits near full occupancy at five in the morning where an office sits at zero — and since the verdict is decided between 04:00 and 07:00, that difference moves the answer further than any of the densities do. A picker that changed only the densities would be the wrong half of the problem.

Lighting is the one number that does not come from those models. They are the 90.1-2004 vintage and their lighting runs well above current code — office 1.02 against 0.64 W/ft² — and in a tool asking whether a building can heat itself, overstated lighting flatters every answer. So lighting comes from the 90.1 Building Area Method instead, and the citation on each field says which is which.

Everything is editable. Change a number and the badge stops claiming the published value and reads edited — but the building stays the building, and so does its drawing.

The design day is derived, not looked up

Name a city and it pulls ten years of hourly weather, takes the coldest 0.4% of hours for the design minimum, and builds the day’s shape from the average of the cold days in the record. You can drop in your own .epw instead, and a .ddy beside it supplies the published ASHRAE minimum — though the shape still comes from the file’s own cold days, because the ASHRAE heating design day is flat and a flat day hides the hour that matters.

That hour is almost always before dawn. Weather data comes from ERA5 via Open-Meteo, and the file you open is read in the browser rather than uploaded.

The finding is usually the spread

The headline is a balance-point temperature: the outdoor temperature below which the building needs heat, where lower is better. A building with a schedule does not have one balance point, though — it has a band, and the band is the interesting part. The same office can sit at 10 °F at peak gain and 61 °F overnight: comfortably self-heating at lunchtime, nowhere near it before dawn.

Which is why the verdict lands between 04:00 and 07:00, and why an equipment row that drops to zero at night flatters every building that has one. The twenty-four hour chart shows loss and gain against each other for every hour, marks the worst one, and draws the Passive House 3.2 Btu/h·ft² line behind it for scale. Work in IP or SI, one at a time.

What it leaves out, and which way that cuts

This is a sensible-heat, envelope-only screen. There is no latent load, no thermal mass, and no ventilation or infiltration.

That last one is the largest single omission, and it means a passing result is optimistic. Solar gain is left out too, which cuts the other way — but the two do not cancel, and the optimistic one is larger in almost every case. The tool states all of this on the page rather than in a dismissible modal, which is the right way round: the assumption you cannot see is the one that misleads you.

Kitchens, laundries and machine rooms are left out of the equipment density. Their load is cooking and washing, most of which leaves through an exhaust hood, and a tool with no exhaust would count all of it as heat in the room. Left in, a primary school’s kitchen — 1,808 ft² of a 74,000 ft² building — carries 77% of its entire equipment load and the school stops needing heat at all.

What it is for

Everything runs in the browser. No account, nothing uploaded, nothing kept.

This is beta, and it is a theoretical screen built on historic weather data — it does not replace a formal heat loss calculation performed to ASHRAE standards. What it will do is tell you, in the week when the massing is still an argument rather than a drawing, whether the building you are describing is anywhere near heating itself, and which surface is costing you the most.