CIBSE TM59 overheating: the test, criteria and weather files

CIBSE TM59 is the standard method for assessing overheating risk in homes, and it works by dynamic thermal modelling of every occupied room against a summer weather file. If a planning officer or a building control body asks for a TM59 assessment, they want hourly simulated temperatures for living rooms, kitchens and bedrooms across 1 May to 30 September, tested against two pass criteria. It is a modelling exercise with fixed rules on occupancy, internal gains and window opening, which is what makes results from different consultants comparable.

What TM59 tests and for which rooms

TM59 was published by the Chartered Institution of Building Services Engineers in 2017 as a design methodology for homes. It fixes the assumptions that used to vary between modellers: occupancy profiles for bedrooms and living areas, equipment and lighting gains, hot water pipe losses in flats with communal heating, and the rules on when a window can be assumed open.

The method asks for a representative sample of units rather than every flat in a block. A modeller normally picks the worst cases: top floor, single aspect, south or west facing, deep plan, small openable area, and units with a communal riser running through them. Any unit with an unusual glazing ratio should be added.

Every occupied room is assessed. Bedrooms get a separate night-time test because sleep quality is where overheating does real harm. Circulation space, plant rooms and stores are excluded, though communal corridors served by heating risers are covered by a separate part of TM59 that looks at corridor temperatures.

The criteria a residential scheme has to pass

For homes that rely on opening windows, TM59 borrows two criteria from CIBSE TM52 and applies them across the May to September period.

Both must pass. A scheme that clears the daytime test and fails the bedroom test has failed, and that is the more common outcome in flats with a single north or east elevation and limited cross ventilation.

Homes with mechanical ventilation and no reliance on open windows are tested differently: all occupied rooms should stay below 26 degrees for at least 97 per cent of occupied hours. That route looks easier on paper. But it is not, because you then have to demonstrate that the ventilation plant can shift the heat and that residents have no reason to open windows anyway.

Care homes, extra care housing and other buildings with occupants who cannot regulate their own environment are treated as vulnerable, and TM59 applies the tighter fixed 26 degree threshold rather than the adaptive one.

How Approved Document O interacts

Requirement O1 of the Building Regulations came into force in June 2022 and applies to new residential buildings in England. It gives two compliance routes. The simplified method caps glazing area by orientation and sets a minimum free area for purge and background ventilation, using a maximum and minimum table that depends on whether the location is high risk. The dynamic thermal modelling route uses TM59 as the analysis method, with amendments set out in the approved document itself.

That last point catches people out. Part O modelling is not identical to a planning TM59 run. The approved document changes some of the assumptions, particularly around which windows may be counted as openable at night, and it requires the assessment to record how excess heat is removed rather than only whether a temperature threshold is met.

Noise is the other difference. If external noise means windows cannot be opened at night, the modelling has to reflect that, and the acoustic evidence has to agree with the thermal evidence. Where the rating and background sound levels point to a facade that has to stay shut after 11pm, the overheating strategy has to deliver comfort with the windows closed. A Noise Impact Assessment and an overheating assessment that contradict each other is the single most common reason a technical consultee sends both back.

Adaptive comfort and weather files

The adaptive comfort model behind TM52 accepts that people in naturally ventilated homes tolerate warmer indoor conditions after a run of warm days. The comfort threshold moves with a running mean of outdoor temperature rather than sitting at a fixed number, so a July heatwave raises the acceptable indoor temperature slightly. That is why a room can sit at 27 degrees and still pass the daytime test, while the same room fails at 26.5 degrees in the bedroom test at night.

Weather files decide as much as the fabric does. CIBSE publishes Design Summer Year files for locations across the UK, in three variants representing a moderately warm summer, a short intense heatwave and a long period of sustained warmth. TM59 uses the first of these for the 2020s high emissions scenario as the baseline. Many London boroughs and some other authorities ask for a future weather file as a sensitivity test, typically the 2050s, and for the London Weather Centre file rather than a suburban one, because the urban heat island is real and measurable.

State the file you used, the year band, the emissions scenario and the percentile in the report. An assessment that does not name its weather file cannot be checked.

Design responses when the model fails

Failures cluster in predictable places, and the fixes have a rough order of cost.

1. Reduce solar gain first: shallower plan depth, smaller glazed area on west elevations, external shading such as brise soleil, balconies and reveals, and solar control glazing with a lower g-value.
2. Improve the ventilation path: dual aspect layouts, openable areas increased beyond the minimum, top-hung vents that can stay open securely at night, and cross ventilation through a hallway where the fire strategy allows it.
3. Reduce internal gains: insulate and locate communal heating risers away from bedrooms, use pipework with better insulation standards, and consider ambient loop or heat interface unit arrangements that run at lower temperatures.
4. Add thermal mass and night purge: exposed soffits and higher ceilings help, though only where night ventilation can actually happen.
5. Mechanical cooling last, and only where the London Plan cooling hierarchy or an equivalent local policy has been worked through in order.

Corridors deserve their own attention. A communal corridor with uninsulated heating and hot water pipes running along it can sit above 30 degrees for weeks, and it then acts as a heat source to every flat door on the landing.

What planners look for in an overheating section

Officers are not modellers. What they check is whether the assessment exists, whether it covers the worst units, whether it agrees with the acoustic and air quality evidence, and whether the mitigation is drawn rather than promised. Shading that appears in the text but not on the elevations will be picked up at condition stage, and by then it is a cost.

A useful overheating section in a planning submission does five things: names the method and the weather file, lists the sampled units and why they were chosen, gives the results per room against both criteria, sets out the mitigation with a reference to the drawing that shows it, and states the residual risk. Where the scheme relies on closed windows, say so plainly and cross refer to the acoustic report.

An Overheating Risk Assessment usually sits alongside the Energy Statement, and the two should share the same fabric assumptions. A scheme that claims a very low U-value and high airtightness for the energy calculation, then quietly assumes more infiltration in the thermal model, will be caught by a reviewer who reads both.

Overheating modelling has to be run by someone with the software and the judgement to set it up, so treat any AI-assisted draft as the structure and the policy framing around a modeller's outputs, never as a substitute for the simulation. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.

Enter your postcode on PlanScope to see whether overheating evidence is likely to be asked for on your site.