Environmental Statement Appendix 14C Assessment of Ground ...... · Appendix 14.C Assessment of Ground-borne Noise and Vibration and Underwater Noise from the Tunnel Boring Machine

Environmental Statement Appendix 14C (6.3.14.3) Assessment of Ground-borne Noise and Vibration and Underwater Noise from the Tunnel Boring Machine

Silvertown Tunnel

Appendix 14.C Assessment of Ground-borne Noise and Vibration and Underwater Noise from the Tunnel Boring Machine

Document Reference: 6.3.14.3

THIS PAGE HAS INTENTIONALLY BEEN LEFT BLANK

Page 2 of 46

Silvertown Tunnel



Contents Contents .................................................................................................................... 3

List of Abbreviations ................................................................................................ 5

Glossary of Terms .................................................................................................... 7

1. INTRODUCTION ........................................................................................... 9

2. THE ALIGNMENT OF THE SILVERTOWN TUNNEL ................................. 11

3. BACKGROUND .......................................................................................... 13

4. PREDICTION OF VIBRATION AND NOISE FROM TUNNEL BORING MACHINES .............................................................................................................. 15

5. NUMERICAL MODELLING ......................................................................... 17

6. MODELLING ASSUMPTIONS USED ......................................................... 19

7. SIGNIFICANCE CRITERIA ......................................................................... 25

8. RESULTS .................................................................................................... 31

9. DISCUSSION .............................................................................................. 39

10. CONCLUSIONS .......................................................................................... 41

APPENDIX 1: THE FINDWAVE® MODEL ............................................................... 43

Page 3 of 46

Silvertown Tunnel



List of Tables

Table 6-1 Typical strata boundaries on the northern side of the Thames................. 19

Table 6-2 Typical strata boundaries on the southern side of the Thames ................ 20

Table 6-3 Soil properties assumed in the model ...................................................... 23

Table 7-1 Assessment thresholds for groundborne noise and vibration ................... 26

Table 7-2 Fish counts from the Environment Agency 2015 fisheries survey ............ 29

List of Figures

Figure 2-1 Overview of the Silvertown Tunnel Alignment ......................................... 11

Figure 4-1 Cutaway view of a large diameter tunnel boring machine (Herrenknecht) ................................................................................................................................. 15

Figure 6-1 Typical cross section through the model showing soil layers .................. 21

Figure 6-2 Typical cross section through the model under the Thames ................... 22

Figure 6-3 Material damping - loss factor assumed.................................................. 24

Figure 7-1 Examples of fish hearing thresholds ....................................................... 28

Figure 8-1 Direct computation of airborne sound level, dB(A) .................................. 32

Figure 8-2 Pseudo noise level derived from ground surface vibration, dB(A) ........... 32

Figure 8-3 Pseudo noise level to 80m distance – 14m ground cover, dB(A) ............ 33

Figure 8-4 Pseudo noise level to 80m distance – 24m ground cover, dB(A) ............ 34

Figure 8-5 Spectrum above Tunnel 14m Ground Cover........................................... 35

Figure 8-6 Spectrum above Tunnel 24m Ground Cover........................................... 35

Figure 8-7 Effect of Distance on Lasmax .................................................................... 36

Figure 8-8 Effect Of Distance on Vibration, mm/s .................................................... 36

Figure 8-9 Range Of Underwater Sound Levels According To Depth And Water Level. ........................................................................................................................ 37

Page 4 of 46

http://onelink.tfl.gov.uk/sites/dco/Shared%20Documents/008.%20DCO%20Submission%20-%20PRODUCTION/6.3%20Environmental%20Statement%20Appendices/App%2014/ES%20Appendix%2014C_clean.docx%23_Toc449715667















Silvertown Tunnel



List of Abbreviations

µPa micropascals

dB Decibel

DCO Development Consent Order

EPB Earth Pressure Balance

ES Environmental Statement

GNm-2 giga pascal

HS2 High Speed 2

Hz Hertz

kg/m3 Kilogram per metre squared

LAsmax A-weighted, Slow, Maximum, Sound Level.

LOAEL Lowest Observed Adverse Effect Level

m Metres

m/s Metre per second

mAOD Metres Above Ordnance Datum

mBGL Metres Below Ground Level

Page 5 of 46

http://www.acoustic-glossary.co.uk/frequency-weighting.htm

http://www.acoustic-glossary.co.uk/time-weighting.htm

http://www.acoustic-glossary.co.uk/definitions-l.htm%23lmax

http://www.acoustic-glossary.co.uk/sound-pressure.htm

http://www.acoustic-glossary.co.uk/sound-pressure.htm

Silvertown Tunnel



mm Millimetres

rms Root Mean Square

SOAEL Significant Observed Adverse Effect Level

TBM Tunnel Boring Machine

TfL Transport for London

VDV Vibration Dose Value

Page 6 of 46

Silvertown Tunnel



Glossary of Terms

Amplitude The objective measurement of the degree of change (positive or negative) in atmospheric pressure (the compression and rarefaction of air molecules) caused by sound waves

Calibrate Mark (a gauge or instrument) with a standard scale of readings.

Coefficient A number or symbol multiplied with a variable or an unknown quantity in an algebraic term.

Eigenfrequencies One of the natural resonant frequencies of a system.

Fourier Transformation Encompasses a function of time (a signal) into the frequencies that make it up

Groundborne Vibrations that travel through the ground from a source and produce a noise as a result.

Homogeneous Term in physical chemistry and material science that refers to substances and mixtures which are in a single phase. I.e. Are of a similar kind or nature

Impedances Measure of the opposition that a circuit presents to a current when a voltage is applied.

Interpolated A method of constructing new data points within the range of a discrete set of known data points.

Isotropic Uniformity in all orientations

Page 7 of 46

Silvertown Tunnel



Lithology A description of its physical characteristics

Longitudinal Running lengthwise rather than across.

Mechanosensory Systems Pertaining to sensory activation in response to mechanical pressures or distortions

Oscillating Move or swing back and forth in a regular rhythm.

Vary in magnitude or position in a regular manner about a central point.

Pseudo Noise Level A signal similar tonoise which satisfies one or more of the standard tests for statistical randomness. Although it seems to lack any definite pattern, pseudorandom noise consists of a deterministic sequence of pulses that will repeat itself after its period.

Spurious Not being what it purports to be; false or fake.

Structureborne Noise is transmitted through solid structures

Transverse Situated or extending across something.

Tunnel Boring Machine A machine used to excavate tunnels with a circular cross section through a variety of soil and rock

Page 8 of 46

Silvertown Tunnel



1. INTRODUCTION 1.1.1 Rupert Taylor Ltd was instructed by Transport for London (TfL) to carry out a

study of the likely level of groundborne noise and vibration from the excavation of the proposed Silvertown Tunnels using a tunnel boring machine (TBM). This report gives the results of a study carried out to predict the likely levels of groundborne noise and vibration in properties above the alignment during the passage of the TBM. These are assessed against criteria proposed by the project. The derivation of criteria is outside the scope of this report.

1.1.2 The work originally reported in Issue 1 of this report has been extended to include predictions of underwater noise and its effects on fish.

1.1.3 Models were created to study the propagation of vibration from the tunnel face, with the TBM operating in the soils which are likely to occur along the alignment. The output of the modelling is an indication of likely ground vibration and associated groundborne noise at various depths and geological situations, as well as a prediction of the decay of vibration with distance. Where the TBM is operating underneath the Thames, models were created for the worst case with the shallowest ground cover, for both high and low water levels.

1.1.4 Vibration and groundborne noise are different aspects of the same phenomenon. Vibration is oscillating movement of the ground or other solid material. This may cause sound to be radiated from vibrating surfaces into either air or water. Vibration, if high enough in amplitude, may be perceived by the tactile sense. Re-radiated groundborne or structureborne noise is perceived by the sense of hearing at frequencies within the audible range. Vibration may be perceived at frequencies too low to be audible.

Page 9 of 46

Silvertown Tunnel




Page 10 of 46

Silvertown Tunnel



2. THE ALIGNMENT OF THE SILVERTOWN TUNNEL

2.1.1 An overview of the alignment is shown in Error! Reference source not found.. Potential vibration and noise from the TBM as perceived by human beings is limited to locations between each portal and the river foreshore. There are sensitive receptors near Silvertown Way in Newham, and in the vicinity of East Parkside and West Parkside in Greenwich.

Figure 2-1 Overview of the Silvertown Tunnel Alignment

Page 11 of 46

Silvertown Tunnel




Page 12 of 46

Silvertown Tunnel



3. BACKGROUND 3.1.1 TBMs have been used for tunnel excavation in London since the

construction of the Jubilee Line Extension to replace the former shield-and-backhoe method. Projects have included the Docklands Light Railway extension to Greenwich, Crossrail and the Channel Tunnel Rail Link (High Speed 1). In the Environmental Statements (ESs) for those projects it was considered that there would be no significant effect because although the passage of the machines would be audible the duration would be limited to a few days. Some complaints were received with respect to tunnelling for the Channel Tunnel Rail Link operating in harder soil types.

3.1.2 During the construction of the Dublin Port Tunnel, a road tunnel in which a much larger TBM was used to excavate limestone and glacial till, there were strong complaints about noise and vibration which occurred at much higher levels than was the case for the rail tunnels in London. This led to greater attention being paid to noise and vibration from TBM working in subsequent projects, such as Dublin Metro North. Monitoring of groundborne noise and vibration was carried out during TBM working on the Crossrail Project in London. Detailed predictions of vibration and groundborne noise were not carried out prior to the Dublin Port Tunnel, and since then they have been done for Metro North, Corrib onshore pipeline, Thames Tideway Tunnel and Silvertown Tunnel. Groundborne noise and vibration was specifically considered in the High Speed 2 (HS2) ES with the conclusions that there would be no significant adverse effects.

3.1.3 The factors which influence the generation and propagation of vibration and groundborne noise from TBMs are primarily the amount of energy required to cut the soil (or rock in the case of Dublin) and the propagation characteristics of the soil. Rotational speed, cutter head type and face pressure have a much smaller effect. The energy requirement is a function of the tunnel diameter and the operating characteristics of the machine. The proposed Silvertown tunnel is likely to be approximately 12.5m diameter. The diameter of the Jubilee Line Extension TBM face was 4.9m, Dublin Port Tunnel was 11.8m, Crossrail 7.1m and High Speed 1 London tunnels were 8.11m diameter.

3.1.4 Soil type is a major influence, with London Clay being soft enough for the main noise from the TBM to be machinery on the TBM. At the other extreme, excavating through rock generates a large amount of noise and vibration due to the cutting effect itself. The previous tunnels mentioned were in a variety of lithologies. Dublin is carboniferous limestone below glacial till. The tunnels

Page 13 of 46

Silvertown Tunnel



in London are in London Clay, Gravel, Lambeth Beds, Chalk and Thanet sands. The Silvertown tunnel will be bored through clay, gravel and sand.

3.1.5 The tunnels referred to above were bored using earth pressure balance (EPB) machines. In some cases slurry machines are used, but current information is that an EPB machine will be likely to used for the Silvertown Tunnel construction. This is a worst case assumption as slurry machines are mainly used in highly permeable unstable terrain where vibration generated by the cutting action at the face would be very low.

3.1.6 The tunnels referred to above were bored using earth pressure balance machines. In some cases slurry machines are used, but current information is that an EPB machine will be likely to used for the Silvertown Tunnel construction. This is a worst case assumption as slurry machines are mainly used in highly permeable unstable terrain where vibration generated by the cutting action at the face would be very low.

Page 14 of 46

Silvertown Tunnel



Figure 4-1 Cutaway view of a large diameter tunnel boring machine (Herrenknecht)

4. PREDICTION OF VIBRATION AND NOISE FROM TUNNEL BORING MACHINES

4.1.1 The prediction of vibration and groundborne noise from tunnel boring machines has to begin with measured field data which is used to calibrate the output of a model for predicting the spatial spread of the vibration (which in turn may also cause groundborne noise). In stiff or hard soils the source is concentrated at the cutter face. In soft soil, groundborne noise may be radiated from the entire length of the TBM, which can reach lengths of 100m or so (see cutaway illustration in Figure 4-1).

Page 15 of 46

Silvertown Tunnel




Page 16 of 46

Silvertown Tunnel



5. NUMERICAL MODELLING 5.1.1 The predictions of vibration and groundborne noise were carried out using

the Rupert Taylor Finite Difference Time Domain model FINDWAVE.

5.1.2 The model used for this study predicts, in the time domain, the three-dimensional vibration velocity of the tunnel face and surrounding lithology. The time-domain results are transformed into the frequency domain to give 1/3 octave frequency spectra, and overall sound levels in dB(A) and vibration units.

5.1.3 The model has been calibrated by using the model to predict vibration from the Crossrail TBM, and back-fitting the results from field measurements obtained during the tunnel drive.

5.1.4 The approach has been to set up a group of generic models in a selection of soil conditions and produce cross-sectional plots of vectored soil velocity from which, subject to the application of transfer functions to buildings, ground surface predictions can been made.

5.1.5 FINDWAVE is a finite difference time-domain numerical model for computing the propagation of waves in elastic media. Full details of the model are given in Appendix I. The excitation is provided from a random array of impulses applied to the tunnel face. The model predicts, in the time domain, the dynamic behaviour the medium surrounding the tunnel face.

5.1.6 The model has a cell size of 100mm in the lateral and vertical directions, and 100mm in the longitudinal direction (along the tunnel). A time step of 1/131072 seconds was used. The model was run for a time period of 1 second. Output from the model consists of time series of the velocity of transverse and longitudinal sections through the model, which are subjected to frequency transformation and expressed as 1/3 octave band spectra. A model with a cell size of 250mm x 250mm was also used to study the effect of distance out to 80m.

5.1.7 At this stage, not enough detail is known about the structure and foundations of potentially affected sensitive buildings to enable detailed predictions to be made for specific receptors. A high resolution model was set up to look at the effect of the presence of a building (a brick-built house on a slab foundation) on the levels of airborne sound likely to occur within rooms due to ground-transmitted vibration at acoustic frequencies.

Page 17 of 46

Silvertown Tunnel




Page 18 of 46

Silvertown Tunnel



6. MODELLING ASSUMPTIONS USED 6.1.1 This geotechnical data assumed in the modelling are shown in Table 6-1 and

Table 6-2.

Table 6-1 Typical strata boundaries on the northern side of the Thames

Formation Typical Soil Description

Top (mAOD)

Bottom (mAOD)

Top (mBGL)

Bottom (mBGL)

Thickness (m)

Min Max Min Max Min Max Min Max Min Max

Made Ground Brick rubble, ash, sand

2.13 5.72 -0.91 2.57 0.0 0.0 0.91 6.2 0.91 6.2

Alluvium Silty CLAY with pockets of peat

-0.91 2.57 -3.95 -0.48 0.91 6.2 3.66 9.45 1.22 4.5

River Terrace Deposits

Silty Sandy GRAVEL

-3.95 -0.48 -10.96 -6.88 3.66 9.45 10.36

16.0 5.95 8.38

London Clay Formation

Stiff silty CLAY

-10.96 -6.88 -16.93 -11.86

11.58 16.0 14.02 22.65 0.9 6.8

Harwich Dense black -16.93 -14.48 -22.76 -15.39

-17.53 22.65 18.44

Formation Pebbles 28.48 1.02 5.83

(GRAVEL) Lambeth Group Very dense

pale green / blue SAND

-22.76 -6.88 -35.26 -18.8 10.36 28.48

24.3 40.6 8.9 14.8

Upnor Formation

Silty fine to medium SAND

-35.26 - -37.41 - 40.6 - 42.75 - 2.15 -

Thanet Formation

Very dense silty fine SAND

-37.41 -18.8 -45.9 -29.5 24.3 42.75

35 49.38 10.7 12.5

Chalk N/A -45.9 - N/A N/A 49.38 - N/A N/A N/A N/A

Page 19 of 46

Silvertown Tunnel



Table 6-2 Typical strata boundaries on the southern side of the Thames

6.1.2 The tunnel behind the TBM was assumed to be lined with concrete with properties as given below, along with the properties assigned to the lithology.

6.1.3 Figure 6-1 and Figure 6-2 are a typical cross section through the model showing example soil layers.

Formation Typical Soil Description

Top (mAOD)

Bottom (mAOD)

Top (mBGL)

Bottom (mBGL)

Thickness (m)

Min Max Min Max Min Max Min Max Min Max

Made Ground Brick rubble, ash, sand

1.35 5.28 -9.22 1.76 0.0 0.0 1.0 14.50

1.0 14.50

Alluvium Silty CLAY -3.23 1.76 -5.95 -1.1 1.0 8.1 3.2 10.3 1.45 7.7

River Terrace Deposits

Silty Sandy GRAVEL

-5.84 -1.1 -8.74 -4.43 3.2 10.3 6.6 13.9 1.6 4.4

London Clay Formation

Stiff CLAY -9.22 -4.43 -22.3 -16.54 6.6 14.5 18 26.04 9 17.9 (P)

Harwich Formation

Very dense GRAVEL

-20.76 -19.48 -25.48 -20.54 14.5 26.04 15.02 30.64

0.52 5.17

Lambeth Group

Very dense pale green / blue SAND

-25.48 -20.15 -40.08 -27.83 15.02

30.64 30.85

45.24 5 15.83

Upnor Silty fine to -38.97 -36.37 -40.47 -39.33 30.85 44.25 33.81 45.25

1.5 2.96

Formation medium

SAND

Thanet Very -40.47 -39.33 -52.52 -50.44 33.81 45.75 47.0 56.88

10.02 13.19

Formation siltySAND

Chalk N/A -52.52 -50.44 N/A N/A 47.0 56.88 N/A N/A N/A N/A

Page 20 of 46

Silvertown Tunnel



Figure 6-1 Typical cross section through the model showing soil layers

Page 21 of 46

Silvertown Tunnel



Figure 6-2 Typical cross section through the model under the

6.1.5 The modulus assumptions used are relevant to the extremely small strains involved in groundborne noise and vibration, and are not necessarily the same as those used for civil engineering purposes. The property D is the compressive modulus, given by:

D=2G(1−σ)/(1-2σ)

Page 22 of 46

Silvertown Tunnel



where σ is Poisson’s ratio and G is shear Modulus.

Table 6-3 Soil properties assumed in the model

Material Shear Modulus, G, GNm-2

Compression Modulus, D, GNm-2

Density, ρ kg/m3

Air 0 0.00014 1.18

Concrete 11.6 31.33 2400

Made Ground 0.068 0.266 1500

Alluvium 0.068 0.266 1500

Gravel 0.027 1.035 2000

London Clay 0.0735 4.41 1700

Lambeth Group 0.58 5.9 2100

Thanet Sand 0.452 5.79 1930

Chalk 0.3 0.845 1950

6.1.6 The loss factor assumed in each case was frequency dependent as shown in Figure 6-3. It should be note that this represents solely material damping within the soil. Additional losses occur in propagation through layered ground, and these are separately computed within the model. Published literature sometimes includes all loss effects within the loss factor, leading to a figure higher than that given in Figure 6-3.

Page 23 of 46

Silvertown Tunnel



Figure 6-3 Material damping - loss factor assumed

Page 24 of 46

Silvertown Tunnel



7. SIGNIFICANCE CRITERIA Significance Criteria – Human Beings

7.1.1 Vibration is assessed in the UK by means of the Vibration Dose Value (VDV) which is defined in to BS 6472-1:2008 “Guide to evaluation of human exposure to vibration in buildings”. There is no equivalent standard for groundborne noise, but the convention has been established that groundborne noise is assessed in terms of the maximum A-weighted sound level using the index LpASmax. VDV is based on weighted acceleration and is dependent on the number and duration of events. However, if vibration velocity does not exceed a continuous rms level of 0.5mm/s throughout an 8-hour night it will not be likely to exceed 0.1ms-1.75 VDV.

7.1.2 It is anticipated that the tunnel boring machine would travel at approximately 10 metres per day, which needs to be taken into account in determining significance criteria. However, it is recognised that there is no mitigation available for groundborne noise, other than temporary re-housing.

7.1.3 Crossrail did not have a significance threshold for TBM groundborne noise, it was explicitly excluded from the sources covered by the 40 dB LAmax figure. The passage of the TBM on Crossrail went largely without incident, and was not much above 40 dBA at 14m depth (based on monitoring at Sussex Gardens) and less at greater depths.

7.1.4 The Silvertown TBM will be much larger, with a face area some three times that of Crossrail. As far as noise policy thresholds are concerned, it is considered that Significant Observed Adverse Effect Level (SOAEL) is 45 dB LAsmax and the Lowest Observed Adverse Effect Level (LOAEL) at 35 LAsmax. Table 7-1 sets out the project’s criteria for vibration and groundborne noise.

Page 25 of 46

Silvertown Tunnel



Table 7-1 Assessment thresholds for groundborne noise and vibration

Ground-borne noise

Lowest Observed Adverse Effect Level LpASmax dB 35

Significant Observed Adverse Effect Level LpASmax dB 45

Vibration

Lowest Observed Adverse Effect Level

VDVday m/s-1.75 0.2 VDVnight m/s-1.75 0.1

Significant Observed Adverse Effect Level

VDVday m/s-1.75 0.8 VDVnight m/s-1.75 0.4

Significance Criteria – fish

7.1.5 Underwater sound is expressed using a different decibel scale to that used for airborne sound, with a reference level of 1µPa rather the 20µPa in the case of airborne sound.

7.1.6 Fish can sense waterborne noise through both auditory and mechanosensory systems1. Hearing threshold curves for various species of fish have been obtained and over the mid frequency range are typically about 75 dB re 1µPa or higher.

7.1.7 Most research has centred on the relationship between sound pressure levels and hearing damage in fish. Sounds that were lower than 180 dB (re 1 µPa) and sounds that were not on continuously had no apparent impact on the sensory cells of the ear2.

7.1.8 Most research has centred on the relationship between sound pressure levels and hearing damage in fish. Sounds that were lower than 180 dB (re 1

1 Clark, Joseph A., Young, Jane A. Bart, Amrit and Zohar, Yonathan, Physiological effects of infrasonic noise on captive fish, Journal of the Acoustical Society of America, October 1996 – Vol 100, Issue 4, p 2709. 2 Hastings MC, Popper AN, Finneran JJ, Lanford PJ, Effects of low-frequency underwater sound on hair cells of the inner ear and lateral line of the teleost fish Astronotus ocellatus. Journal of the Acoustical Society of America – March 1996, Vol 99, Issue 3. pp1759-1766

Page 26 of 46

Silvertown Tunnel



µPa) and sounds that were not on continuously had no apparent impact on the sensory cells of the ear.

7.1.9 It is not known what effect noise has in the range between 75 and 180 dB, on such matters as the behaviour of fish in, for example, swimming to their spawning grounds. Rivers and oceans are, however, naturally noisy, with heavy rain producing sound pressure levels up to 110 dB. Shipping can produce noise levels over 105 dB.

7.1.10 Fish species also have pressure-sensitive cells that are sensitive to low frequencies (typically 10 Hz to 30 Hz) and near-field pressure changes, limited to an area immediately surrounding the fish. This allows the fish to detect the presence of other fish in close proximity (such as in schooling behaviour) or assists in predatory avoidance. The sensitivity of the pressure-sensitive cells, however, is at an insufficient level to detect the TBM operation.

7.1.11 Fish are known to have complex and diverse inner ear structures with a broad range of hearing sensitivities, generally weighted toward lower frequencies. Some species, such as some members of the herring family can hear across a wider range of frequencies and quieter sounds, these species are sometimes referred to as hearing specialists. Others, particularly flatfish, e.g. dab and plaice, hear over a narrower range of frequencies and have much lower sensitivities to sound (they have no swim bladders); salmon and trout fall into this latter group also, which are sometimes referred to as hearing generalists. Some examples of fish hearing thresholds3 are given in Figure 7-1 Examples of fish hearing thresholds.

3 Nedwell, J.R., B. Edwards, A.W.H. Turnpenny, and J. Gordon. 2004. Fish and marine mammal audiograms: A summary of available information. Prepared by Subacoustech Ltd., Hamphire, UK.Report 534 R 0214.

Page 27 of 46

Silvertown Tunnel



Figure 7-1 Examples of fish hearing thresholds

7.1.12 Data obtained from the Environment Agency regarding the results of a fisheries survey along the river Thames have recorded the species shown in Table 7-2.

Page 28 of 46

Silvertown Tunnel



Table 7-2 Fish counts from the Environment Agency 2015 fisheries survey

7.1.13 The results reported below show that underwater sound levels caused by the TBM are likely to be well above the threshold of hearing of fish, and also likely to be well above ambient sound levels although significantly below hearing damage thresholds. The effect this will have on the behaviour of fish can be assumed to be that they will tend to swim away from the source of the sound.

Page 29 of 46

Silvertown Tunnel




Page 30 of 46

Silvertown Tunnel



8. RESULTS

Groundborne Vibration and Noise

8.1.1 Groundborne noise is commonly estimated by prediction of the vibration of a surface such as a floor and converting that to what has become known as “pseudo noise level” using a formula4 originally proposed by Kurzweil:

Lp=Lv-27 dB

where Lp is the airborne sound pressure level and Lv is the velocity level in dB re 1 nanometre per second. There are some indications5 that this may be an overestimate in some cases, by an amount of the order of 5 dB.

8.1.2 The first model run was carried out to consider the relationship between ground vibration and sound pressure level inside an example residential building – a brick structure on a concrete raft foundation. Figure 8-1 shows the actual A-weighted sound pressure level computed for a height of 1.25m above ground. Figure 8-2 shows pseudo noise level computed by the subtraction of 27 dB from the ground surface vertical velocity level (the highest zone plotted in these figures is 40-41 dBA). It shows that if the pseudo noise level approach is used based on free vibration of the ground surface, the prediction would be a maximum of 41 dB LASmax although the level is highly dependent on room dimensions and position within the room, with highest levels near the walls and in the corners of the rooms. Bearing in mind the fact that bedheads are usually close to walls, these areas are highly relevant. Figure 8-2 shows pseudo noise 44-45 dB, suggesting that the Kurzweil formula may well be a worst case.

8.1.3 Given the sensitivity of the airborne noise predictions to room dimensions and location, it is considered appropriate to use pseudo noise level in this study generally.

4 Kurzweil, L.G. Ground-borne noise and vibration from underground rail systems, Journal of Sound and Vibration (1979) 66(3), 363-370 5 Association of Noise Consultants, Measurement and Assess of Groundborne Noise and Vibration, 2012

Page 31 of 46

Silvertown Tunnel



Figure 8-1 Direct computation of airborne sound level, dB(A)

Figure 8-2 Pseudo noise level derived from ground surface

8.1.4 In order to study the effect of distance out to the order of 80m to the side of the tunnel a coarser grained model was used with a cell size of 250mm

Page 32 of 46

Silvertown Tunnel



Figure 8-3 Pseudo noise level to 80m distance – 14m ground cover, dB(A)

presented within Figure 8-1 and Figure 8-2. The corresponding spectra are shown in Figure 8-3 and Figure 8-4

8.1.5 The effect of distance on both the spectrum the overall level of groundborne noise and the rms vibration velocity is shown in Figure 8-7 and Figure 8-8. The vibration velocity was predicted to be very low – a maximum of 0.016 mm/s which for continuous operation would not exceed 0.03 m/s-1.75 VDVday or 0.025 m/s-1.75 VDVnight.

Page 33 of 46

Silvertown Tunnel



Figure 8-4 Pseudo noise level to 80m distance – 24m ground

cover, dB(A)

Page 34 of 46

Silvertown Tunnel



Figure 8-5 Spectrum above Tunnel 14m Ground Cover

Figure 8-6 Spectrum above Tunnel 24m Ground Cover

Page 35 of 46

Silvertown Tunnel



Figure 8-7 Effect of Distance on Lasmax

Figure 8-8 Effect Of Distance on Vibration, mm/s

Page 36 of 46

Silvertown Tunnel



Figure 8-9 Range Of Underwater Sound Levels According To Depth And Water Level.

Underwater sound

8.1.6 The results of the underwater model are shown in Figure 8-9 for the worst case location in the vicinity of the TBM, close to the position of the face.

Page 37 of 46

Silvertown Tunnel




Page 38 of 46

Silvertown Tunnel



9. DISCUSSION 9.1.1 The results of the predictions of groundborne from the Silvertown Tunnel

TBM drive noise are approximately 5 dB greater in level than was the case for the Crossrail tunnel drive. This is partly due to the increased face diameter, which for all other things unchanged would be expected to cause an increase of 5 dB. Boring in the Lambeth Group beds is likely to have a greater effect than boring in the London Clay but this is offset by the effect of the layered ground.

Page 39 of 46

Silvertown Tunnel




Page 40 of 46

Silvertown Tunnel



10. CONCLUSIONS 10.1.1 In summary the results were as follows, together with the relevant

significance criteria:

Groundborne noise: 45 dB LpASmax, equal to SOAEL

Vibration: 0.03 ms/-1.75 VDVday, less than LOAEL (0.2 ms/-1.75 VDVday)

0.025 ms/-1.75 VDVnight, less than LOAEL (0.1 ms/-1.75 VDVday)

Underwater sound: 140 dB re 1 µPa, well above fish hearing thresholds but below damage risk threshold of 180 dB.

10.1.2 The study has shown that the level of groundborne noise is unlikely to exceed the SOAEL even if there were sensitive receptors directly above the tunnel. Actual sensitive receptors are located to the side of the tunnel where the levels will be lower and the LOAEL will be reached at about 75m to the side of the tunnel regardless of tunnel depth.

10.1.3 The LOAEL for vibration is unlikely to be exceeded at any sensitive receptor.

10.1.4 With regard to underwater sound, the predicted levels are well above the hearing thresholds of at least some fish species, but are well below hearing damage thresholds. The effect of the noise is likely to be temporary and to cause fish to move away from the vicinity of the TBM.

10.1.5 There is uncertainty associated with the predictions in that the soil properties have been derived from library data for the characteristics of the strata found in the ground investigation carried out for this project, and not from field measurements as the relevant measurements were not included in the investigation. The method used for deriving groundborne noise levels from predicted ground vibration is, however, conservative.

Page 41 of 46

Silvertown Tunnel




Page 42 of 46

Silvertown Tunnel



APPENDIX 1: THE FINDWAVE® MODEL

A.1 INTRODUCTION

The wave equation in differential form is as follows A.1.1

(1)

for the x axis, with corresponding equations for the y and z axes, where x, y, z and x, η, ζ are displacements in three orthogonal axes; ρ λ and µ are Lamé constants and ρ is the density. The Lamé constant µ is also known as the shear modulus, G. The Lamé constant λ is also known as the coefficient of dilatation and is given by

where σ is Poisson’s ratio.

Equation (1) can be stated in finite difference form by replacing the A.1.2differential operator with the approximation

x∂∂x ≈ (x[i][j][k] - x[i-1] [j][k])/ ∆x (2)

For ∆x →0 these two forms are identical.

For a homogeneous, isotropic medium with a finite value for ∆x, ∆y and ∆z, A.1.3elastic wave propagation can be computed using the finite difference substitution of equation (2)

Effectively, the process is as follows, for each axis, i, j and k. The example A.1.4given is for axis i. Each point p(i,j,k) lies at the corner of a rectangular cell and is assigned a mass equal to one eighth of the sum of the eight contiguous cells as well as a displacement and velocity. The displacement and velocity is interpolated for each intermediate “virtual” point p(i+d,i+d,k+d) where d=0 or 0.5.

1) Compute pressure gradient

( ) 2

222

2

2

2

2

2

2

2

2

tzxyxxzyx ∂∂

=

∂∂

∂+

∂∂∂

+∂∂

++

∂∂

+∂∂

+∂∂ xρζηxµλxxxµ

( )σσλ21

2−

=G

Page 43 of 46

Silvertown Tunnel



2) Compute shear force gradient 3) Accelerate p(i,j,k) by ∆v=F/ρ ∆t where F is the sum of the

force 1 & 2 and ρ is the density assigned to the point and v is the point velocity.

4) Displace p(i,j,k) by ∆x=∆v*∆t where x is the point displacement and t is one time step.

5) repeat from step 1

The geometric part of wave propagation is completely represented by this A.1.5process. Further terms are required to represent damping. Of several possible terms, the inclusion of a coefficient by which the velocity is multiplied produces a loss factor which decreases within increasing frequency (and gives rise to an excess attenuation per unit distance which is independent of frequency). A viscous damping term can be used, by including a force proportional to acceleration multiplied by a coefficient. However, many materials exhibit hysteretic damping, or damping with other types of frequency dependence. To model these effects it is necessary to include an algorithm which implements Boltzmann’s strain history method where

s(t) = D1ε(t) - ∫∞

∆∆∆−0

)()()( tdttt ϕε

where ϕ(∆t) = τ

τ/2 teD ∆−

is an after-effect function, D2 is a constant and τ is a relaxation time. D1 is a modulus, s(t) is stress and ε(t) is strain. By combining several after-effect functions with different values of D2 and τ any relationship between loss factor and frequency may be represented. Note that in the frequency domain the integral has a real and imaginary part, with the result that the value of the modulus is reduced by the inclusion of the relaxation terms. Depending on the choice of the constants and relaxation times, the stiffness of a resilient element will be frequency-dependent, and the value of D1 must be adjusted at the same time that D2 and τ are selected to give the required dynamic stiffness. This method has been implemented in the version of FINDWAVE® used for this study.

A.2 BOUNDARIES

For modelling finite objects fully surrounded by space, the boundaries can be A.2.1represented by assigning zero-valued elastic moduli to the space provided that the acoustic load of the air in an airspace can be neglected. If radiation

Page 44 of 46

Silvertown Tunnel



into air is to be modelled, or if an infinite or semi-infinite medium such as the ground is required, it is necessary to minimise the effect of reflections from the boundaries. For a train tunnel, where distances to be modelled are small compared with the length of the train, the z-axis boundaries are dealt with by creating a model exactly one rail vehicle (or unit of several coupled rail vehicles) in length, and then connecting the ends of the model together to create an infinitely long train. This is done by copying the cell displacements and velocities from one end of the model to the other end at the end of each time-step.

For the other boundaries in the x- and y-axes, the potential problem of A.2.2spurious reflections from model boundaries is overcome by the use of an impedance matching technique. This effectively assigns to the cells which are required to be non-reflective on the boundaries of the model the properties of a massless viscous damper such that

∆

∆−

+−= − txv

DcK o

0

1 )(''ρ

xxρ

ωη

where η is the loss factor (dimensionless), K’’ is the imaginary part of a complex spring stiffness in which the real part is zero, ω the angular frequency, ρc the characteristic impedance of the medium, x0 and x-1 are the displacements of cell points 0 and –1 where the boundary is at cell 0, ρ is the density of the cell contents and v0 is the velocity of cell 0. Over 95% absorption is achieved across the spectrum.

A.3 INPUT DATA

The only input data required for the model are the masses of each cell, plus A.3.1the shear modulus and the compression modulus, and the loss factor. Otherwise, all secondary parameters such as wave speeds, impedances etc. are automatically generated by the finite difference algorithm. The only other input relates to methods of approximating actual structure shapes using the orthogonal grid.

The output of the model consists of a file containing the displacement and/or A.3.2velocity of one or more selected cells.

The time steps used are of the order of 30 to 60 microseconds, and the A.3.3model is run for either 16384 or 32768 steps to give a signal length of just under 1 second.

Page 45 of 46

Silvertown Tunnel



The resulting discrete time series can then be subjected to discrete fourier A.3.4

transformation to yield frequency spectra.

Note that, whereas in the acoustical analogy, the impedance of air varies A.3.5little (except close to sources such as points), so that in most cases power is proportional to velocity squared, in elastic media, velocity transfer functions do not directly convey information about power transmission, and velocity at the receiver, in a low impedance medium, can be higher than velocity near the source, in a high impedance medium, even when there are power losses between the source and the receiver.

VALIDATION

The finite difference algorithm is validated by creating models of structures A.3.6for which algebraic solutions are available and comparing the eigenfrequencies and decay rates. For Timoshenko beams, plates, thin and thick cylinders the eigenfrequencies are correctly predicted.

Page 46 of 46