Assessing noise from a takeaway extract fan

The extract fan, not the cooking, is what gets a takeaway refused on noise grounds. A kitchen canopy pulling several thousand cubic metres an hour needs a fan that runs every hour the premises is open, including late at night when the background sound level in a residential street has dropped to almost nothing, and it sits on a rear elevation a few metres from someone's bedroom window. That combination is the whole problem.

The assessment that answers it is not complicated, but it has to use the right numbers. Most weak submissions fail because the supplier's figure was copied into a report without anyone checking whether it was a sound power level or a sound pressure level, or at what distance, or whether it included the duct.

Why extract fans dominate objections

A commercial kitchen extract system runs at a constant level for long periods, often past 23:00, and steady mechanical sound at night is the single most reliable source of complaint in mixed-use streets. Cooking odour attracts objections too, and odour is dealt with in its own assessment, but noise is what environmental health measures and what conditions control.

Three characteristics make fan noise worse than its decibel figure suggests. It is continuous, so there is no respite. It often carries tonality from blade passing frequency, which the assessment method penalises. And it is frequently mounted high on a rear wall with a clear line of sight to upper-floor windows, so there is no screening.

Sound power against sound pressure in supplier data

This is the distinction that decides whether the assessment is worth anything.

Sound power level, written as Lw, describes the source itself and does not vary with distance. Sound pressure level, written as Lp, is what a receiver experiences and falls with distance. Manufacturers publish both, sometimes on the same page, and a figure of 78 dB(A) means nothing until you know which it is.

For a point source radiating over a reflecting ground plane, the conversion follows a standard relationship: the sound pressure level equals the sound power level minus 20 times the logarithm of the distance in metres, minus about 8 decibels. An illustrative worked case, using a fan with a sound power level of 85 dB(A). At 10 metres the free-field sound pressure level is roughly 85 minus 20 minus 8, so about 57 dB(A). Add a correction for a facade reflection and for any second source, then apply the acoustic feature corrections in the assessment method. The arithmetic is illustrative only; the real calculation uses octave band data.

What to ask the supplier for, in writing:

• octave band sound power levels for the fan at the duty point, not at free delivery
• separate figures for inlet, outlet, and casing radiated sound
• the duty point actually selected, with static pressure
• attenuator insertion loss by octave band, for the specific unit proposed

A fan selected at the top of its curve is both inefficient and loud, and a system whose ductwork resistance was underestimated will end up there.

Duct breakout and the parts people forget

Noise escapes a kitchen extract system at four places, and submissions routinely address one.

There is noise from the discharge terminal, which is the obvious one. There is noise radiated through the casing of the fan itself. There is breakout through the walls of the ductwork, which matters enormously where a lightweight spiral duct runs up an external wall past bedroom windows, because the duct becomes a long line source. And there is structure-borne transmission through the fan mountings into the building fabric, which is the source of the complaints about a hum that cannot be heard outside.

The fixes are all straightforward if designed in:

1. Attenuators on both sides of the fan, sized by octave band rather than by overall dB(A).
2. Anti-vibration mounts and flexible connections at every duct-to-fan junction.
3. Acoustic lagging to any duct run passing a sensitive receptor, with the specification and the mass per unit area stated.
4. A discharge terminal chosen for low regenerated noise, pointing away from windows, at high level.
5. Rigid duct rather than lightweight spiral where the run passes close to habitable rooms.

Retrofitting any of these on an occupied building costs several times what it costs at design stage, and the lagging is usually the item that turns out to be impossible because there is no room for it.

Attenuators, plenums and fan speed

Fan speed is the underused lever. Sound power from a fan rises steeply with rotational speed, so a modestly oversized fan running slower on an inverter is much quieter than a smaller fan working hard, at the same airflow. Where an operator can accept a slightly larger plant footprint, this is the lowest-cost acoustic measure available.

Attenuator selection has to match the spectrum. A unit that performs well at 500 Hz and above does little for a low-frequency rumble at 63 and 125 Hz, which is exactly the content that passes through a party wall into a bedroom. Ask for insertion loss by band and check it against the fan's own spectrum.

A plenum box at the discharge, sized to slow the air, reduces regenerated noise and helps with odour dispersion at the same time. The duct velocity target matters for both: high velocities generate noise at every bend and damper.

Assessment against the background at the receptor

The recognised method for this kind of plant is BS 4142:2014+A1:2019. It compares a rating level, which is the specific sound level of the plant plus corrections for acoustic features such as tonality, impulsivity and intermittency, against the background sound level measured as the LA90 at the receptor over the relevant reference time interval.

Two points decide most disputes. The reference time interval is one hour during the day and 15 minutes at night, and the night-time background in a quiet residential street can be very low, so a fan that is inaudible at lunchtime can produce a large excess at 23:30. And the measurement position has to be at the receptor, meaning the nearest bedroom window, not the site boundary or the middle of the yard.

In broad terms, the standard treats a rating level around 5 decibels above background as an indication of an adverse impact and a difference of around 10 decibels or more as an indication of a significant adverse impact, with the context of the area taken into account. Our note on unattended noise monitoring duration, position and validity covers how to get a defensible background figure in the first place.

Where the premises is in a busy high street the background may be high enough that compliance is straightforward. Where it backs onto gardens, expect to design to a rating level at or below background at night, and expect the environmental health officer to say so.

Coordinating noise with the odour case

One system, two assessments, and they pull in different directions. Odour control wants high discharge velocity and a tall termination above the roofline. Noise control wants low velocity, attenuation and a terminal that does not fire at windows.

Resolve the conflict on paper before submission rather than in a consultation response. A short design note showing the extract duty, duct sizes and velocities, the attenuator schedule, the filtration train and the termination point serves both cases. The Odour Assessment and the Ventilation and Extraction Statement then describe the same system rather than two different ones, which is the mistake that generates a request for further information.

What to include in the submission

For a hot food premises in or near residential use, a submission that gets through validation and consultation contains:

• a noise assessment to BS 4142:2014+A1:2019 with the measured background at the nearest sensitive receptor, the time periods covered, and the rating level with corrections shown
• the plant schedule with octave band sound power data at the duty point, from the manufacturer
• the attenuator and lagging specification, with insertion loss by band
• a drawing showing duct route, termination height and direction, on elevation and plan
• the proposed operating hours for the extract system, which may need to be shorter than the opening hours
• a commissioning commitment: a post-installation verification measurement submitted to the council

That last item is worth volunteering. Officers are wary of predictions because they have seen installations that did not match the report, and offering a verification condition turns a leap of faith into a checkable obligation.

Expect conditions restricting extract operating hours, requiring the system to be installed and maintained as approved, requiring verification measurements, and sometimes setting a numerical limit at a stated position. Read a numerical limit carefully before agreeing to it, because a limit expressed at a facade includes every other sound source at that point. Our piece on where planning control ends and licensing control begins explains why a premises licence does not settle the planning question.

This is general guidance, not legal advice; check the position with your local planning authority, your environmental health team or a qualified acoustician before relying on it. Noise assessments of this kind are normally expected from a suitably qualified author, and professional review is strongly advised.

Draft your ventilation and noise documents on PlanScope, then send them for professional review. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.