Getting the measure of things

Measuring noise, part 4

In this blog I will look at the dBA unit in use in real situations. Let’s start with a graph of the noise that results from a pass by a TGV train; this graph comes from the US Department of Transport Federal Railroad Administration (US FRA) manual High-Speed Ground Transportation Noise and Vibration Impact Assessment which I first introduced in my blog of 17 May. In that document the graph reproduced below is figure 2-3 on page 2-5.

TGV pass at 180mph (source: USFRA)

The first thing to notice is that the graph specifies the distance from the track at which the noise is being measured; this distance is 25 metres although it is expressed as 82 feet on the graph to conform to normal practice in the United States. Obviously this distance has to be stated, since our day to day experience tells us that the measured noise power will decrease with distance from the track; 25 metres is the standard distance for making such measurements. It is also assumed that there is level ground between the measuring point and the track and that no noise mitigation is being used.

The second point to make is that the speed of the train is 180 mph, which is about 290 kph. This is much lower than the initial maximum speed of HS2 at 360 kph, but I have used the TGV graph because there appears to be very little data available at ultra high speeds. However what is particularly significant is that the TGV is travelling below the 300 kph threshold at which aerodynamic noise starts to dominate and the increase of noise with speed is steeper.

The horizontal axis of the graph is time and the vertical axis is dBA, so the vertical axis is logarithmic not linear. The vertical axis also starts at 65 dBA, not zero. This is because the noise event is considered as starting when the sound level first crosses a specified noise threshold (in this case 65 dBA) and finishing when the level returns below this threshold again. One way of setting this threshold is to consider it as when the noise first becomes noticeable above the background noise, but other approaches may also be used.

The trace shows how the noise measured varies instantaneously with time as the train is passing. Acoustic engineers usually denote the level of noise on the dBA scale by the symbol “L”. The peak level of L, Lmax, is 93 dBA in this example. So how loud is that?

One way of looking at it is that a level of 93 dBA is two billion (2,000,000,000) times the reference level power of 0 dBA, which is nominally the threshold of hearing. That is probably not very helpful, other than illustrating how the logarithmic scale compresses the huge variations that are encountered in the sound world. Another comparison is that this level is in the same power region as a petrol lawnmower from about a metre away.

I used the term “in the same power region” since the human ear is not a precise measuring tool and one should not be too fussy about the odd dB. For example, an increase of 3 dB is barely perceptible to the human ear, 5 dB is clearly noticeable, 10 dB is about twice as loud and 20 dB is about four times as loud.

So will HS2 at 360 kph or 400 kph be perceptibly louder than the TGV train in our example graph? Since I have not been able to find any data of peak noise power at ultra high speeds, I have resorted to a graph published by HS2 Ltd as figure 4 on page 45 of Appendix 5.4 to the Appraisal of Sustainability Main Report (which can be found here); this graph is reproduced below.

How noise varies with speed (source: HS2 Ltd)

There is an important caveat that I must make in using the HS2 Ltd graph. The vertical scale is not an instantaneous measurement, it is noise averaged over the day and this accounts for the lower dBA values compared with the TGV graph (I will cover this way of measuring noise in a future blog in this series). This also means that the two vertical scales in the two graphs for TGV and HS2 that I have included in this blog, although both marked “dBA”, are not directly comparable. However I believe that it is valid to use the HS2 Ltd graph to look at the relative increase in noise that results from increasing the speed.

One other thing that I wish to point out on the HS2 Ltd graph is that it illustrates very nicely the 300 kph threshold that I have referred to above. You can see that the curve kinks at this speed and the section to the right of the kink is steeper as the aerodynamic noise dominates.

The HS2 Ltd graph shows that there is about a 5 dB increase in noise power when the speed increases from 290 kph to 360 kph. The curve doesn’t continue past 360 kph, so it is necessary to extrapolate it. Doing this indicates that the further increase in noise from 360 kph to 400 kph is possibly around 4 dB. So the answer to the question that I posed above, is that a train pass at 360 kph will be louder than one at 290 kph and also that the difference, at around 5 dB, will be “clearly noticeable”. At 400 kph the difference from 290 kph will, at 9 dB or so, appear “about twice as loud”.

At these higher speeds the noise from a high speed train at 25 metres is heading into similar territory as a jet aeroplane flying over at a height of about 300 metres (1,000 feet).

In my next blog I will look at a way of expressing the noise nuisance of a train pass, or other noise event, as a whole, rather than just considering the peak power.

Note: In this blog I have used “dBA” where an absolute sound power level is being referred to and “dB” for the difference between (ratio of) two sound power levels on the dBA scale. I have done this not to confuse, but because I have always employed this usage and believe it to be correct.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: