Overheating assessments and CIBSE TM59 competence

CIBSE TM59 is a method, not a qualification. Nobody is licensed to run it, and nothing stops a graduate with a trial licence of a simulation package producing a compliant-looking report. So TM59 overheating assessment competence is judged on the inputs, the weather file and the room-by-room results, because those are the places a weak assessment shows itself.

If you are an architect commissioning or reviewing one, the useful skill is knowing which five questions to ask.

What TM59 requires of the modeller

TM59 sets out a design methodology for assessing overheating risk in homes using dynamic thermal modelling. For naturally ventilated dwellings it applies the adaptive criteria from TM52 to living areas, and a fixed criterion to bedrooms, where operative temperature should not exceed 26 degrees for more than 1 per cent of annual night hours between 10pm and 7am.

For dwellings that are mechanically ventilated or cooled, the method switches to a fixed temperature criterion applied across occupied rooms rather than the adaptive approach.

There is a third test that catches more schemes than either of those. TM59 includes a separate criterion for communal corridors in blocks of flats, which are frequently unventilated, surrounded by heated risers and ignored until the assessment is run. A scheme that passes in every flat and fails in the corridor is a common outcome, and the fix is usually ventilation design rather than glazing.

Dynamic thermal modelling software and operator skill

The model is only as good as the person building it. Two modellers with the same software, the same drawings and the same weather file routinely produce different answers, and the difference is almost always in the assumptions rather than the physics.

Ask which tool and which version, and ask whether it has been validated against recognised simulation test procedures. Then ask the harder question: how many residential schemes has this person modelled, and has any of their work been checked by a third party.

How the model is divided into thermal zones is where inexperience shows. A block modelled as one thermal zone per flat will behave differently from one modelled room by room, and TM59 results are reported per room. Single aspect flats, top floor flats and south or west facing rooms are the ones that fail, so the model has to include the worst cases rather than a representative sample chosen for convenience.

The simplified and dynamic routes compared

Approved Document O of the Building Regulations provides two routes to demonstrating that overheating risk has been limited in new residential buildings.

The routes are not equivalent in difficulty or in outcome. A scheme that fails the simplified method may pass the modelling route, which is why the modelling route exists. A scheme that fails the modelling route has a design problem.

Two points about Approved Document O that architects underestimate. It restricts reliance on openable windows where external noise would prevent occupants opening them, which pushes noisy urban sites towards mechanical ventilation and a different compliance strategy. And the modelling route under the Approved Document is not identical to a plain TM59 run, so read the Approved Document's own requirements rather than assuming a TM59 report satisfies it.

This is Building Regulations rather than planning. Planning overheating requirements are separate, and in London the London Plan policy on managing heat risk requires the cooling hierarchy to be addressed at application stage, which is why an Overheating Risk Assessment often sits alongside the Energy Statement in a submission.

Weather files and the assumptions that matter

The weather file is the single input with the largest effect on the result, and the one most often reported vaguely.

TM59 specifies the use of CIBSE Design Summer Year data for the location nearest the site, at a stated percentile and emissions scenario, with the hotter design summer years used for sensitivity rather than compliance. A report that says current weather data was used tells you nothing. It should name the location file, the design summer year, the percentile and the time period.

The other inputs that move the answer:

• Internal gains from occupancy, lighting, appliances and hot water distribution, which TM59 standardises so that results are comparable.
• Window opening logic, which TM59 defines by internal temperature and occupancy rather than leaving it to the modeller.
• Shading, including external obstructions from neighbouring buildings and any fixed shading devices modelled geometrically.
• Blinds and curtains, which should not be relied on unless there is a sound reason, because occupants do not operate them to a schedule.
• Ventilation path resistance through trickle vents, corridors and lobbies, rather than an assumed free area at the window.

Ask for a table of these assumptions in the report. A modeller who cannot produce one has not thought about them.

Where a planning condition requires verification

Conditions requiring an overheating assessment to be submitted and approved, and sometimes requiring post-construction verification or in-use monitoring, are becoming common on larger residential schemes.

If you are offered such a condition, read it as the person who will discharge it. Verification by monitoring means installing sensors, collecting a summer's data and reporting against criteria written for a model rather than for measurement, and the two do not map neatly onto each other. Agree what the comparison will be before accepting the condition.

Where the condition requires an updated assessment at a later design stage, make sure the obligation sits with somebody in the team at that stage. An assessment condition that falls due after the architect's appointment ends is a problem nobody owns.

Reviewing someone else's model

You cannot review a TM59 report from the summary table. Ask for five things.

1. The software and version, and confirmation of validation against recognised test procedures.
2. The weather file, named in full, with the percentile, scenario and time period.
3. The assumptions schedule covering gains, window opening, shading and blinds.
4. Results for every assessed room, not an average or a selection, with the criteria applied and the margin by which each passes.
5. The model file itself, so that a reviewer can rerun it.

Then check that the rooms that ought to be worst actually are. If a top floor, single aspect, west facing bedroom passes comfortably while a ground floor north facing room is marginal, something in the model is wrong.

Where the result matters commercially or the scheme is contentious, an independent second opinion on the technical evidence costs a fraction of redesigning a facade. Our professional directory is the route to a reviewer where a draft needs sign-off, and liability between the author and the submitting party should be settled in the appointment rather than afterwards.

What to ask for in the deliverables

Specify the report contents in the appointment, not at the end. A usable TM59 deliverable contains the modelled geometry with a thermal zone plan, the full assumptions schedule, the weather data reference, results per room against each criterion with the margin, a sensitivity run on the hotter design summer years, a schedule of the mitigation relied on with drawing references, and the model file.

Add one requirement that gets missed: a statement of what the design depends on. If compliance relies on a 1.2 m deep brise soleil, external blinds or a specific free ventilation area, that dependency has to be visible to the design team and to the contractor pricing the value engineering exercise. Mitigation that disappears in a cost saving round and reappears as a complaint in July is the predictable failure in this field.

For specialist technical reports, qualified professional input may be required. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.