Guide
How are construction noise levels predicted under BS 5228-1?
Section 61 · Updated
Short answer
A BS 5228-1 prediction starts with how much noise each machine gives out, then allows for the distance to the property and how much of the working period the machine runs. The levels from all the machines are added together and compared with a threshold that depends on how noisy the area already is, as in the ABC method, one of the standard's example methods. The platform follows the standard's sound power method for the predictions in an application, and says what it leaves out.
The steps in words
The BS 5228-1 method predicts the level at a property from each machine on the site. Annex F of the standard gives the methods. One starts from levels measured 10 metres from similar work, one starts from each machine's sound power level, and two more cover plant that moves about the site or along a haul road. This guide explains the sound power method in words, and does not copy the standard's text or tables.
The platform uses the sound power method for the predictions in an application.
- Sound power level. Each machine has a sound power level, in decibels. It is the noise the machine gives out, whatever the distance. Annex C of BS 5228-1 gives typical levels for many types of plant as the level 10 metres away, and says that adding 28 decibels to one of those gives the sound power level. The levels were measured on particular machines, and the standard says a real machine can be quieter or louder, depending on its make, its maintenance and how it is used.
- Platform guidanceDistance. Sound spreads out as it travels. For hard ground, the sound power method treats the sound as spreading over half a sphere, and the platform uses that. The formula is below. Each time the distance doubles, the level falls by about 6 decibels. The standard has a second formula for soft ground, such as grass, which gives lower levels from 25 metres out. The platform does not use it.
- On-time. A machine that runs for part of the working period gives a lower average level than one that runs for all of it. The standard reads the adjustment from a chart, and it works out at 10 times the base 10 logarithm of the share of the period the machine runs. A breaker used for 3 hours of a 10-hour day has an on-time of 30%, which lowers its level by about 5 decibels.
- Adding machines. Identical machines at one point add 10 times the base 10 logarithm of their number. Levels from different machines are combined the same way. Levels from separate machines are not added like ordinary numbers: two equal machines come to about 3 decibels more than one (70 and 70 make about 73, not 140).
Level at the property (LAeq) = sound power level (LwA) - 20 x log10(distance in metres) - 8
If the works have separate activities that run one at a time, the platform gives each property the level of its loudest activity. If some activities run together, it adds up all the plant.
What the prediction leaves out
The platform's predictions in an application make no allowance for soft ground, screening by barriers or buildings, reflection from the front of a building, absorption by the air, or any mitigation you plan to use. Leaving out soft ground and screening gives a higher prediction. Leaving out reflection gives a lower one: where the point of interest is 1 metre from the front of a building, BS 5228-1 adds 3 decibels to the free-field level, and the platform does not, so its level there is about 3 decibels lower than the standard's method gives.
Table E.1 does not say whether its thresholds are free-field levels or levels at the front of a building. The standard's worked examples predict the level at the front of the nearest building, with the 3 decibels added.
Screening is anything between the machine and the property that blocks the direct path of the sound, such as a barrier, a hoarding or a building. It lowers the level at the property. Screening you plan to use does not change the prediction in the application. The platform's separate noise model does count screening by buildings and by the mitigation you place.
The predictions are estimates. Before you save an application, the platform asks you to confirm that you know the predicted levels are estimates for a competent person to check before the application goes to the council.
The ABC method in words
Annex E of BS 5228-1 gives examples of ways to judge whether construction noise could have a significant effect. It is an informative annex, and it says it gives examples only. One of its examples, the ABC method, sets a threshold for the noise from the site, using Table E.1. The table gives example thresholds at dwellings and says it applies to residential receptors only. The threshold depends on how noisy the area already is. In words:
- Take the ambient noise, which is the noise already there before the works start, as an LAeq for the period you will work in. Round it to the nearest 5 decibels.
- Compare the rounded level with the Category A value for that period. If it is below that value, the area is Category A. If it is equal, the area is Category B. If it is above, the area is Category C.
- Each category has its own threshold for each period, and quieter areas get lower thresholds. If the noise from the site, rounded to the nearest decibel, is above the threshold, a potential significant effect is indicated.
- If the ambient level is above the Category C value for the period, the test is different. A potential significant effect is indicated when the site noise raises the total level for the period by more than 3 decibels.
- A potential significant effect is not the last word. The standard says the assessor then weighs other things about the project, such as how many people are affected and how long the noise lasts and what it is like, to decide whether there is a significant effect.
The platform compares every property it lists with these thresholds, schools and offices included. That goes beyond the table, which is for dwellings, and is the platform's choice. The platform also compares levels to a tenth of a decibel, not the nearest decibel, so a level a fraction above a threshold counts as over. That is the platform's choice too, and it counts more properties as over, not fewer.
The values for each category and period are in the standard, and this guide does not copy them. Check them in your copy of BS 5228-1, or with your council.
On the platform, a level you measured yourself is used before the noise map's estimate of the ambient noise.
Periods
BS 5228-1 splits the week into periods, each with its own threshold: daytime, evenings and weekends, and night. The clock times are in the standard (Annex E, Table E.1).
The platform's impact bands
The platform shows each property as high, medium or low. High means the predicted level is over the threshold. Medium means it is not over the threshold but is within 5 decibels below it. Low means it is more than 5 decibels below. These bands are the platform's guidance. They are not terms from the standard.
Where the standard sits in the law
In England, the Control of Noise (Code of Practice for Construction and Open Sites) (England) Order 2015 approves BS 5228-1:2009, as amended by Amendment No. 1, for giving guidance on appropriate methods for minimising noise from construction and open sites (the Order, article 2). The Order applies in England only. In Wales, the Welsh Ministers' order of 2017 approves the same edition (S.I. 2017/81, article 2). In Scotland, the Scottish Ministers' order of 2002 approves Part 1 of BS 5228 in its 1997 edition, with Amendment No. 1, and Parts 3, 4 and 5 (S.S.I. 2002/104, article 2), so ask your council which edition it expects.
This guide follows the 2009 edition, as amended in 2014 (BS 5228-1:2009+A1:2014), and not the 1997 edition.
Subject to the rest of section 72, regard is to be had to any relevant provision of a code of practice approved under section 71 when reading the words “best practicable means” (section 72(6)).
Related guides
Start an application
The Section 61 platform prepares an application and the evidence behind it. You check it and send it to your council yourself. Start an application.
Sources
- BS 5228-1:2009+A1:2014, Code of practice for noise and vibration control on construction and open sites, Part 1: Noise (British Standards Institution)A paid standard. These guides describe its method in words and copy none of its tables or limit values.
- The Control of Noise (Code of Practice for Construction and Open Sites) (England) Order 2015, S.I. 2015/227Article 2 approves BS 5228. Article 1(2) says the Order applies in England only.
- The Noise from Audible Intruder Alarms (Wales) (Revocation) and Control of Noise (Codes of Practice for Construction and Open Sites) (Wales) Order 2017, S.I. 2017/81 (W. 30)Article 2 approves BS 5228 Parts 1 and 2. Article 1(3) says the Order applies in relation to Wales.
- The Control of Noise (Codes of Practice for Construction and Open Sites) (Scotland) Order 2002, S.S.I. 2002/104Article 2 approves BS 5228 Part 1 (1997) and Parts 3, 4 and 5. Article 1(2) says the Order extends to Scotland only.
- Control of Pollution Act 1974, section 72: “Best practicable means”
- The Section 61 platformWhere a guide describes what the platform does or suggests, it says so. Those are the platform's own choices, not law.
The legislation cited was read on legislation.gov.uk on 26 October 2026.
This is guidance, not legal advice. Check with the council.