Part O overheating, explained: TM52, TM59, and the gap a model can't close

An overheating assessment is a promise made on a model. The building keeping that promise, through a real summer, with real people in it, is a separate thing entirely.

Since 2022, new homes in England have had to prove they will not overheat. Approved Document O — Part O — is the building regulation that sets the test. It applies to dwellings, residential care homes and other places people sleep, and it has to be satisfied before the building is signed off.

What Part O asks for

Part O gives two routes to compliance. The simplified method caps how much glazing a room can have and how much window can open, by orientation — design limits you either meet or you do not. The second route is dynamic thermal modelling: a simulation of the building run against a future summer and judged by the CIBSE methods — TM59 for homes, TM52 for non-domestic buildings.

TM59 — the test for homes

TM59 runs a year of simulated weather through the design and checks two things.

• Living rooms, kitchens and bedrooms must not run 1°C or more above the comfort threshold for more than 3 percent of occupied hours between May and September.
• Bedrooms carry a second, stricter rule: the temperature must not exceed 26°C for more than 1 percent of the night-time hours between 10pm and 7am — roughly 32 hours across the year.

TM52 — the test for non-domestic buildings

TM52 uses three criteria: the hours a room runs over the comfort temperature, a daily weighted measure of how far and how long it ran over, and an absolute ceiling of 4°C above the comfort temperature. A room has to fail no more than one of the three to be considered low risk.

All of this happens before the building exists

Every one of these numbers comes out of a simulation. The weather is assumed, the occupancy is assumed, the windows are assumed to be opened the way the model says. The assessment is a careful, regulated prediction of how the finished building should behave — but a prediction. It is signed off, filed, and rarely opened again.

Nobody returns the following August to check whether the bedrooms actually stayed under 26°C at night. The model said they would. The record stops there.

Phase 2 of Awaab's Law changes what that gap costs

From October 2026, excess heat joins damp, mould and excess cold as a hazard social landlords must act on to a legal timescale. That moves overheating from a design-stage calculation to an in-use duty. The question is no longer only whether a home overheats on paper, but whether it is too hot now, and whether you can show what you did about it. A modelling report written before handover does not answer that.

Where monitoring fits

A continuous temperature sensor measures the thing TM59 only predicted: the real temperature in the real room, through the real summer. It can count the hours a bedroom spends above 26°C at night and show whether the building is keeping the promise its assessment made. When a room drifts over, it is logged and flagged while there is still time to act — not discovered later in a complaint.

The point

Part O and the TM methods prove the design. They cannot prove the building. One is a promise made on a model; the other is the evidence of what the rooms did once people moved in. As in-use heat becomes a legal standard, it is the second that gets asked for.