Vibration from railways and the assessment thresholds

A railway vibration assessment has to answer two separate questions from one source. Will occupants feel the vibration, and will they hear it. Feeling it is assessed against human response criteria expressed as a vibration dose value. Hearing it is a groundborne noise question, where energy transmitted through the ground makes floors and walls radiate a low rumble inside a room that is otherwise quiet. A scheme can pass one and fail the other, and acoustic glazing does nothing for either.

Architects usually meet this after the plan is drawn, which is late.

Two effects from one source

Passing trains put energy into the ground. It travels as surface and body waves, attenuating with distance and changing character with the soil, then couples into a building's foundations and up through its structure.

Perceptible vibration is the response of people to movement they can feel, typically through the floor, and it is dominated by frequencies below about 80 Hz. Groundborne noise is the sound radiated by the building's own surfaces as they vibrate, usually audible as a low rumble, and it is what occupants of flats near tunnels complain about most.

Airborne noise from the railway is a third thing entirely, controlled by the building envelope. It can be solved with glazing and ventilation. The other two cannot.

Distinguish them explicitly in the assessment, because a report that reports only a vibration dose value has answered half the question, and one that reports only an internal noise level has answered the other half.

How vibration dose values are used

The standard method for human response uses the vibration dose value, which accumulates the fourth power of the weighted acceleration over a period, so it is sensitive to occasional high events as well as to continuous exposure. Values are calculated separately for a 16 hour day and an 8 hour night.

The relevant British Standard sets out ranges of vibration dose value against the probability of adverse comment in residential buildings, with lower values giving a low probability of adverse comment, intermediate values making adverse comment possible, and higher values making it probable. Night-time ranges are roughly half the day-time ones. Take the figures from the current edition of the standard rather than from an older report, and state the edition you used.

Two practical points about this. The dose value depends on the assessment period and the number of train movements within it, so a night-time value on a freight route can exceed the daytime value on the same site. And the criteria are about adverse comment, not about damage. Cosmetic damage thresholds are a different question and are assessed against different, much higher, levels.

Measuring and predicting at an undeveloped site

On a site with no building on it, the assessment is a prediction based on measurements at the ground surface.

Measure with triaxial accelerometers at the proposed building footprint, at more than one distance from the track where the site is deep, over a period covering the full range of services including the heaviest and the slowest. A three hour survey that catches only passenger units on a route that carries night freight will understate the result badly.

Ground conditions govern propagation, so the geotechnical picture is part of the acoustic work rather than parallel to it. Soil stiffness, layering, water table and rockhead depth all change the transfer from track to building, which is one reason a Geotechnical Assessment should be commissioned early enough to inform it.

The prediction then applies a transfer function from ground surface to building foundation and up through the structure. That step carries the largest uncertainty in the whole exercise, and an honest report says so and applies a margin rather than quoting a single number to two decimal places.

Isolation at building level

Where predicted levels exceed the criteria, the response is isolation, and the options run from low cost and limited to expensive and effective.

Foundation type matters. Piles can couple a building more strongly to deeper, stiffer strata than a raft would, so the foundation decision has an acoustic consequence that the structural engineer will not raise unless asked.

Note what does not work. A barrier or bund reduces airborne noise and has no effect on groundborne transmission, which is a distinction worth keeping clear alongside what an acoustic barrier actually achieves.

Layout does more than any of it. Putting circulation, storage, plant and commercial uses on the elevation nearest the track, and bedrooms furthest from it, changes the problem from an engineering one into a planning one.

Talking to the infrastructure operator

Development near an operational railway brings the infrastructure owner into the process as a consultee, and its interests are not the same as yours.

It will want to know about crane oversail, excavation close to the boundary, surface water discharge towards the track, fencing, and anything that could affect the stability of the railway or its assets. Those are safety matters and they proceed through the operator's own asset protection process alongside the planning application.

Two things are worth asking for early: information on current and anticipated future services on the route, since a freight path added later changes the source, and any existing vibration data the operator holds for that section. Also ask whether track renewal is planned, because renewed track with modern resilient fastenings can reduce vibration at source, and a scheme designed against a worn track may be conservative.

Reporting and condition wording

State the criteria, the edition of the standard, the measurement and prediction method, the uncertainty, and the mitigation required to meet the criteria, with the mitigation described well enough to be priced.

Ask for condition wording that names an internal performance standard to be achieved and requires verification testing before occupation, rather than one that requires a scheme of works to be approved. A performance condition survives design development. A condition tied to a specific isolation product does not, and it generates a variation application when the supplier changes.

Verification testing after completion should be carried out in the completed rooms, with the building finished and the trains running normally, at points agreed in advance.

A Noise Impact Assessment covering airborne sources should sit alongside this rather than absorb it, because the criteria, the methods and the mitigation are different.

When the site is simply unsuitable

Sometimes the honest conclusion is that residential use on part of the site does not work.

Full base isolation on a small scheme can cost more than the uplift from the units it protects, and a development that proceeds without it puts the problem onto future occupiers who have no remedy. The railway was there first, and an occupier who complains about it will be told so.

Where that is the answer, say it in the report and change the use on the affected part of the site. Commercial, storage, parking or landscape on the track frontage is not a failure of the scheme. It is the scheme responding to the constraint.

This is general guidance, not legal advice; check the position with your local planning authority, the infrastructure operator and a qualified acoustician before relying on it. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.