The International Light Rail Magazine
+44 1733 367610
Geoff@lrtap.com

Is LRT really the noisy neighbour?

Hong Kong’s tramway runs in as tight an environment as you can imagine, but in this instance – on Des Voeux Road West – the noise produced by the trams is largely drowned out by the environmental noise of busy diesel-engined delivery vehicles, buses and general traffic. (Image: Neil Pulling)

It has been my experience in 30 years of devising and promoting tram and light rail schemes that noise and vibration are often the ‘hot button’ issues that opponents will invoke in an attempt to activate popular opposition. This remains a fact even after the practical experience of new schemes worldwide shows that potential problems are often grossly exaggerated.

However, light rail and tramway noise is a complex issue and the technical aspects are often confusing; I hope that this article goes some small way towards helping campaigners and promoters counter the misguided, or mendacious, anti-tram propaganda.

What is noise?

It is helpful to start here with the commonly accepted dictionary definition: noise is a loud or unpleasant sound that is undesired and/or interferes with a person’s hearing of something. In everyday terms noise can disturb normal activities, such as holding a conversation, reading or sleeping, and distract concentration.

Sound is energy radiated by pressure waves, usually passing through air (at speeds of approximately 332m/s), but also passing through solids – rails, surrounding structures, the earth, etc. One of the problems in understanding noise is that it is measured on the Decibel (dB) scale which measures proportions rather than a linear measure (such as a foot rule) that measures absolute quantities. Thus the perceived effect of doubling the amount of energy in a sound – such as putting two identical sound sources close together – is quite small, it is much less than a doubling of the sound heard; such an increase would lead to a measured increase of around 3dB. Similarly, ten motorcycles close together will only be perceived as twice as loud as one motorcycle – they will certainly not be ten times as loud.

To add to the complexity of sound measured, the commonest measure used for overall assessment of noise impact is the ‘A-weighted decibel’ – dB(A). This weighting is a method of calibrating measurement so that it represents the experience of the human ear, which has varying responses to different pitches of sounds. The good news is that noise levels in dB(A) measure proportions so that a change of 10 dB(A) corresponds approximately to a doubling, or halving, of the noise; only a change of more than 3dB(A) can be detected by the human ear.

Judgement of volume varies from person to person and is dependent on the characteristics of the sound; for example, research conducted in the Netherlands has shown that for a given loudness road noise is generally more annoying than rail noise. Indeed some research has shown that light rail can be up to 4-5dB(A) louder than other highway noise before annoyance is experienced. The table (right) shows some examples of the dB(A) of various common experiences.

Yet further complication is added by the fact that general noise will consist of a background noise (e.g. the general rumble of traffic) and noisier peaks (the squeal of the hard application of a disc brake, the sound of a tram rounding a curve etc). An attempt is made to average these sounds out to produce an ‘equivalent continuous sound level’ or LAeq. It is this figure that will often be quoted in Environmental Impact Assessments.

Rail operating noise is different from that generated by other road transport in that it mainly comes from the interaction of steel wheels on steel track, whereas other vehicles produce a mix of various sounds. This includes internal combustion engine noise, brake squeal and the continuous hissing sound generated by rubber tyres flexing on a rough asphalt surface or the rumble associated with concrete paving. The requirements of the variety of worldwide accessibility regulations can also add to tramway/LRT operating noises by requiring vehicles to make ‘doors closing beeps’ and the escape of public address sounds from the car to the stop area. Although these are extremely helpful to many passengers, they can be a significant source of complaint from local residents.

The three main generators of noise on a light rail system are:

  • Noise arising from the wheel rolling along the rail-head; this can be made far worse by rough, worn or corrugated railheads and ‘flats’ on the wheels
  • Wheel ‘squeal’, partly generated when flanges engage with the gauge face of the rail or the check rail, but mostly arising from the ‘skip slip’ effect as heels that are locked together by a rigid axle pass through a tight curve
  • Impact noise, where wheels pass over discontinuities in the railhead – usually at turnouts but also at dropped or open joints.

The good news is that all of these issues can be easily managed on a well-built and operated system and will be considered below.

So how noisy is light rail?

In a typical urban environment, with both rail and highway traffic, the tramway can be up to 10dB less noisy than general traffic; however, as the overall level is a combination of all noise energy the tramway noise will probably not be noticeable. Adding a tram, which at low speeds may only generate around 60db(A) to a road which is already generating 60-70 dB(A)Leq will make little perceptible difference and no increase in annoyance.

LRVs will be noisier at higher speeds, yet the precise level of noise will depend on car type, alignment and type of track; grass, open ballasted or paved. This higher speed operation is more likely to happen out of central city areas where general traffic levels may be lower and the change may be perceptible to residents. Adding a tram with a peak sound of 70dB(A) measured at nearby buildings in a residential area where background noise levels may be in the 40-50dB(A)Leq level could increase noise beyond the 4-5dB(A) ‘bonus’ of rail over other road transport mentioned above.
Thus complaints are likely to arise in the areas with more affluent residents – who often have a greater propensity to complain.

It must also be remembered that creating space for rail-borne vehicles may lead to the relocation of general road traffic closer to buildings. Experience in a number of cities has been that this was the main cause of predicted increases in noise. The French example of reducing road space for general traffic can thus have the multiple benefits of finding space for the tramway, further improving modal split by reducing the attractiveness of private car use, improving air quality as well as reducing noise. The UK approach of minimising reductions in highway space is attempting to force a ‘quart into a pint pot’ and will often yield disappointing results.

Mitigation of noise

It is depressing to see that many Environmental Impact Statements advocate the use of large stretches of noise barriers to mitigate LRT operating noise. Barriers can be visually intrusive, they take up open space, are costly to install and maintain, are a target for vandals, and often fail to achieve the hoped-for results. It is always better to reduce or eliminate the problem at source and much can be done by careful planning and construction.

A well thought-out alignment that allows efficient use of cant on curves, allows good transition and minimises reverse curves can help. It is often claimed that easier curve radii can help, but it is this author’s experience that curves of 30m-plus radius can be more likely to generate squeal than tighter radii. The tighter curves will be negotiated at slower speeds and can achieve very satisfactory results; I have noticed that two of the tightest curves on second-generation tramways – Wellesley Road/George Street, Croydon and St Andrew Street/Queen Street, Edinburgh (both UK) – are remarkably quiet in operation if the driver takes them at the correct speed.

Rail-heads can be noticeably rough when first laid; a combination of surface irregularities created during the mill rolling of the rails and corrosion during storage can create a poor surface. Systems that grind their entire track to a uniform shape and smoothness after commissioning and trial running but before commercial service often achieve a far better result. Similarly, an exercise to ensure that all car wheels are true and blemish-free before commercial operations can be beneficial.

Maintaining rail-head condition and wheel circularity can be an expensive business and the task can cut away the work-hardened surfaces – far better to start with good rails and wheels and ensure that drivers operate vehicles in a way that will minimise subsequent damage. Corrugation of the rail-head can be a very difficult problem; over the years many explanations for – and cures of – the problem have been propounded, with varying degrees of success. Minimising skip/slip conditions may help, as can different degrees of wheel and rail damping.
One unfortunate effect on new systems is that all cars will come as one batch, with virtually identical ride qualities and harmonic responses; which means that once one car has started a rail corrugation all other vehicles will respond to the corrugation frequency in the same manner and spread it further down the track, leading to a ‘corrugation epidemic’.

The dreaded ‘wheel squeal’

Wheel squeal is a particular problem, both for the noise-sensitive resident and for the accountant who has to find the money to repair track wear. Flange engagement with the gauge face of the rail (and with the check on some curves) is a source of some squeal and may be mitigated by careful matching of the wheel and rail profiles. This may seem to be an obvious point, but it is surprising how often it is overlooked.

Wear on the rail gauge face can be reduced by some use of friction modifiers; oil, grease sticks or track water sprays; it can be repaired by weld deposits, often with the beneficial result that the weld build-up can work-harden to a smoother surface than the initial rail. However not all rails are suitable for weld repairs and some forms of rail encapsulation can fail quite dramatically if subject to local high temperatures during weld repairs, again it should be the responsibility of the design engineers and contractors to specify and install the appropriate materials.

The wheel layout of the vehicle can also have a significant effect. It is well-known that long wheelbase trucks can cause binding on curves, but trucks with only limited horizontal rotation capacity – particularly when combined with long car overhang – can cause significant nosing and tracking problems. More conventional bogies with greater degrees of rotational freedom can help, but they are difficult to combine with the fashionable cult of the 100% low-floor car.

Fixing two wheels onto a rigid axle locks them together so that they can only rotate at the same speed. This causes problems on curves, where the length of the railhead to be traversed by each wheel is significantly different. On main line railways with large radius curves the conicity of the wheels can mitigate the problem, but this option is of little help on tight rail curves in urban environments; on a circular curve with a radius of 20m to the inner rail the difference in rail-head length can be nearly 4.5m, which for a wheel of 600mm diameter can mean a difference of well over two revolutions! This inevitably sets up a skip and slip episode which can lead to noise, wear and corrugation. This can only be effectively mitigated either by use of a differential (as in the Bremen and Berlin AEG type low-floor cars) or by the use of separately controlled independent wheels; SET in Derby (UK) has a major programme to develop such a traction package and it is to be hoped that this will be a major step forward in eliminating the problem.

As mentioned earlier, discontinuities of the rail-head are a major source of noise, their elimination can result in major benefit. Where track is laid in grooved rail, the use of flange-running crossings will support the wheel through the discontinuity, avoiding wheel drop and pounding of the nose of the common crossing. Where flange-running crossings cannot be used, particularly on conventional open track, a swing-nose crossing can be used to eliminate the gap. Swing-nose crossings are not cheap to buy or maintain but they can be effective; for example, they are used for the turnouts at Canary Wharf DLR station, UK – a particularly noise sensitive area.

Even the best point and crossing work will generate some noise and vibration, so the location of this special work needs to be carefully considered; does the crossover need to be at this precise spot, or can it be moved a few metres away to a less sensitive location? Would a pre-selector point (as at Church Street, Croydon) mitigate problems? It must be emphasised that good tramway design is a holistic and iterative process and all participants must be brought into the task if the best results are to be achieved and maintained.

This feature originally appeared in Tramways & Urban Transit – March 2015 issue (927).