TRANSPORT AND ROAD RESEARCH LABORATORY Department of Transport

RESEARCH REPORT 53

GROUND VIBRATION CAUSED BY CIVIL ENGINEERING

WORKS

by B M NEW MSc PhD

The views expressed in this Report are not necessarily those of the Depa~ment of Transport

Ground Engineering Division Highways and Structures Department Transport and Road Research Laboratory Crowthorne, Berkshire, RG11 6AU 1986

ISSN 0266-5247

Ownership of the Transport Research Laboratory was transferred from the Department of Transport to a subsidiary of the Transport Research Foundation on 1 st April 1996.

This report has been reproduced by permission of the Controller of HMSO. Extracts from the text may be reproduced, except for commercial purposes, provided the source is acknowledged.

CONTENTS

Page

Abstract 1

1. Introduction 1

2. The effects of vibration 1

2.1 Peak Particle Velocity (PPV) as a measure of damage potential 1

2.2 Current damage criteria 3

2.3 Human perception and intrusion criteria 7

3. The prediction of vibration 8

3.1 Non-explosive sources 8

3.2 Explosive sources 10

4. Blasting trials 11

4.1 Measurement procedures and equipment 11

4.2 Data processing and presentation 13

4.2.1 Square and cube root scaling methods 13

4.2.2 Site specific scaling using multiple regression analysis 14

5. Contractual specification and vibration control 15

6. Acknowledgements 17

7. References 17

8. Appendix 19

© Crown Copyright 1986 Extracts of the text may be reproduced,

except for commercial purposes, provided the source is acknowledged.

GROUND VIBRATION CAUSED BY CIVIL ENGINEERING WORKS

A B S T R A C T

Investigations have been undertaken to improve techniques for the prediction and control of ground vibration caused by civil engineering construction works. Contemporary damage and nuisance criteria are reviewed and factors affecting the input and propagation of ground vibration are discussed. Field data from numerous sites in the UK are summarised to provide some guidance as to the relative importance of various sources (eg, traffic, piling, tunnelling, blasting).

A methodology for trial blasting is given with recommendations for the deployment and specification of the equipment. Data processing and presentation format is described and a 'site specific' scaling method given which provides improved correlation between peak particle velocity and scaled distance. The use of peak particle velocity as a damage-hazard specifier is discussed in relation to other dynamic parameters which also effect damage to structures.

Options regarding the distribution of vibration associated risks between Employer and Contractor are considered in the context of legal and contractual obligations.

1 I N T R O D U C T I O N

This paper addresses the problem of the specification, measurement and control of ground vibration which is caused by civil engineering works. Vibration from construction sources will generally be of a temporary nature, but the disturbance caused may result in permanent damage to property and substantial nuisance to the local population. Either factor may lead to restraints on the working method that result in additional costs or even, in extreme circumstances, curtailment of activity.

Blasting and piling operations have in the past been the cause of greatest concern, but in recent years construction works have utilised larger plant as economic pressures have forced greater emphasis on mechanised rather than labour intensive techniques. These developments have resulted in the use of machines that dissipate large amounts of energy, in the form of ground vibrations and noise, into the environment.

At the present time pre-construction vibration prediction analyses are often rather hit and miss affairs and in many instances they are not carried out

at all. This lack of prior knowledge can lead to severe restrictions being placed on the Contractor during the actual works.

Most vibration associated problems may be expressed as two separate questions:

(i) what level of vibration will be produced by the proposed construction works? This will depend on the construction method and the seismic propagation characteristics of the site.

and (ii) what is the acceptable level of vibration? This will depend on the type of structures at risk and the sensitivity of the local population to nuisance.

The answers to both these questions are always site specific, although an important initial appraisal of hazard may be made which is based on experience from other sites. Reference to case history information will often greatly assist planning operations although where excavation by blasting is required in an urban environment it may often be prudent to carry out trial blasts as part of the site investigation programme. Procedures are given here which enable trial blasting works to be rationally planned and for the data gained to be presented and used in the most beneficial manner.

The determination of 'acceptable' vibration levels is in many ways more difficult to deal with owing to its subjective nature, particularly with regard to nuisance. A wide variety of contradictory advice on 'safe' vibration levels is available. The increasing influence of the environmental lobby has resulted in some standardising authorities insisting on unreasonably low levels of vibration from blasting. On the other hand explosive manufacturing companies and blasting contractors do occasionally seem rather optimistic in their assessment of the durability of structures (and people) subject to dynamic loading. The review and discussion given below seeks to evaluate and reconcile these viewpoints and provide guidance for working damage and nuisance criteria.

2 THE EFFECTS OF V I B R A T I O N

2.1 PEAK PARTICLE VELOCITY (PPV) AS A M E A S U R E OF D A M A G E POTENTIAL

Vibration induced damage thresholds are usually expressed in terms of peak particle displacement, velocity or acceleration and sometimes include a frequency dependent factor. Harmonic vibrations may

be described by any two of the fol lowing parameters; f requency, peak particle displacement, peak particle velocity or peak particle acceleration. The 'peak particle' preface indicates that it is the maximum value associated with the motion of a particle at a point in the ground (or on a structure) that is considered.

It is most usual for peak particle velocity (PPV) to be used as it has been found to be the best correlated wi th case history data of damage occurrence and has a theoretical underpinning inasmuch as the strain induced in the ground is proportional to the particle velocity. A l though PPV is widely used to quantify the damaging potential of a vibration it must be recognised that 'velocity', of itself, cannot induce damaging forces. Such forces are generated in structures by both : - -

(a) differential displacements which give rise to distortion as the structure fol lows movement of the ground upon which it is founded

or(b) change in the ground particle velocity vector (magnitude or direction) which produces inertial forces upon the structure.

In practice the structure will be subjected to both 'distort ion' and 'inertial' mechanisms at the same time and these will be superimposed upon pre existing stresses and strains from other causes; damage will occur when the combined effects exceed the tolerance of the structure. Vibration can also give rise to longer term ground movements (eg by compact ion) which may also contr ibute materially to structural distress; this mechanism is discussed elsewhere (New, 1978) and is not considered in this paper.

For convenience Figure 1 separates the distortional and inertial factors and considers particle motions normal and parallel to the surface. The ground distortion (Figures la and b) could be attributed to vertically polarised shear waves or the vertical component of Rayleigh surface waves whilst the dilatation in this case (Figures l c and d) results from a compressional wave type. It can be shown (see Appendix) that the shear strain )/ imposed by the ground distortion at foundat ion level is a function of the particle velocity (V) and the velocity of (shear) wave propagation cs. That is ), = V/cs. Similarly the strain (~) caused by dilatations depends on the particle velocity and the compressional wave velocity Cp. That is ~ = V/cp. This form of calculation will be appropriate for any structure which closely fol lows the movement of the ground eg (pipes, tunnels, foundations).

For most buildings however the actual distortions will be heavily dependent on the dynamic response of the structure. The natural frequencies and damping characteristics of a building will determine the strains imposed by the ground motions. To simplify, the case

is considered where the ground motion wavelength is long compared to the length of the structure.

A force is required at the interface of structure and ground to move the building either up and down or side to side, Figures lb and d. These forces will depend on the effective mass of the structure and the acceleration imposed by the ground wave motions. Now the acceleration (a) is a function of particle velocity (V) and the frequency (f) of the motion (a = 2 ~ V for harmonic motion). The actual forces transmitted within the structure are dependant on its particular response characteristics but for a given structure the ground motion parameters concerning damage hazard are particle velocity and frequency.

In practice structural damage may result from the complex interaction of the mechanisms shown and it is important to emphasise that whilst PPV may be the single most valuable parameter to observe, the frequency and propagation velocity of the ground motions must also be considered. A more detailed discussion of structural response using single degree of freedom models is given by Dowding (1985).

The dependence of distortion induced in structures on wave propagation velocity has been recognised by Langefors and Kihlstrom (1978), who give risk of damage thresholds which are closely proportional to the wave propagation velocity of the ground upon which the structures are founded (see Table 1). Other velocity criteria given below are varied with respect to the frequency of the vibrations which will assist the prediction by making empirical allowance for the response characteristics of common urban structures.

When Vibration damage is discussed little if any allowance is generally made for the characteristic propagation velocity of the wave motionS. For instance, consider a group of vibrations comprising compressional waves followed by shear waves of slightly lower peak particle velocity and similar frequency. It will be the PPV of the compressional arrivals that will be used to assess the risk of damage despite the fact that the dynamic distortions from the later shear wave arrivals (of slightly lower PPV) may be significantly larger (owing to their lower propagation velocity). Further, the wave propagation velocity in rock may easily be ten times that associated with a soil; for a given PPV and frequency this means that the ground distortions will be ten times greater for a structure founded on soil than that for one on rock.

Some instruments used to monitor blast vibrations only show PPV levels and not the full time history of the vibration wave packet. Where instruments show the wave packet in detail it should usually be possible to estimate the frequency of the motions and to differentiate between compressional and shear/surface wave arrivals by allowing for their relative times of arrival and discontinuities in the waveforms.

2

(a) Short wavelength distortion

Structure distortion ~ " 7

Ground profile L 1 / "~ I without vibration [ / with vibration ~ | |

Ground particle +Wave velocity C s-~

T t T t ra jectory--~ :

(b) Long wavelength distort ion (inertial effect) Structure ~ Inertial displacement ~ ~ - - ~ ~ l de fo rma t i on

. . . . . . . . - ]~-- - - __ - i x ~ - ~ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \

Verticle force on X -- X 1 = mea = me21"rfV

T T T

(c) Short wavelength dilatation

Structure elongation ~-~

[ F[ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ -

(d) Long wavelength di latat ion (inertial effect)

Structure displacement 4-~ Inertial n I deformation

~ \ \ \ \ \ \ \ \ \ \ \ N N N \ \ \ \ N \ \ \ \ \ \ \ X \ \ \

Horizontal force on X -- X 1 = mea = m e 2TrfV

Wave velocity Cp i _ _ = 1 N -o H~ ~ Particle r - j -,

trajectory

1 ~ Ground particle displacement

Fig. 1 D e f o r m a t i o n and inert ia l fo rces d u e t o g r o u n d w a v e m o t i o n

T A B L E 1

Risk of damage in ordinary dwelling houses with varying ground conditions (Langefors and Kihlstrom, 1978)

Sand, shingle, clay under ground water

level Moraine, slate, soft

limestone

Hard limestone, quartzy sandstone

gneiss, granite, diabase Type of damage

Wave velocity m/s 300-1500 2000-3000 4500-6000

4-18

6-30

Vibration velocity mm/s

35

55

80

115

8-40

12-60

70

110

160

230

No noticeable cracks

Insignificant cracking (threshold value)

Cracking

Major cracks

Although it is essential to recognise that the risk of structural damage is also dependant on frequency and propagation velocity it seems likely that vibration damage criteria will continue to be related empirically with peak particle velocities. An important pragmatic influence on this is that there is little reliable information available on the damage induced in structures related to measured dynamic stresses and strains. The data that are available on distortions of a quasi-static nature are not readily interpreted in terms of dynamic movements as the effective strength of a material is often critically dependent on the rate at

which the load is applied. (For instance, the dynamic tensile strength of rock may easily exceed its static value by an order of magnitude).

2.2 CURRENT D A M A G E CRITERIA

The establishment of precise or universal criteria that define vibration damage thresholds is not possible and expert judgment based on specific site knowledge and previous case history data will often

3

be necessary. The possibility or degree of damage that results from vibration will depend on the nature of the source, the transmission characteristics of the intervening geological strata and the inherent strength and response characteristics of the subject structure. The dynamics involved will usually be of a complex nature and many of the variables that control the resulting structural motions are likely to be unknown. Any suggested damage criteria are, of necessity, a compromise on which engineering judgments may be based; they must not be regarded as hard and fast rules.

Standardizing authorities throughout the world have experienced difficulty in defining acceptable standards to be supported by legislative powers. No British or International Standard defines vibration thresholds for damage to structures, although such a document is in the course of preparation.

Attempts to establish safe levels of vibration have, not unnaturally, tended to be of a conservative nature and the original German Standard DIN4150 (1938, revised 1984) is an example that is generally held to be overcautious and unworkable. Recently revisions have been proposed to this document and some revised guide values are given in Table 2.

The Swiss Association Standards (1978) again seem rather cautious (see Table 3). They provide a guide to acceptable levels from blasting or traffic/ machinery forms of excitation for various types of structure. Persson et al (1980) reported vibration limits enforced in Swedish cities (Table 4). Langefors and Kihlstrom (1978) provided a risk assessment for ordinary dwelling houses with varying ground conditions (Table 1). Westine et al (1978) and Attewell and Fry (1983) have investigated the effects of explosive detonations on buried pipelines and vibration effects on subsurface structures have also been considered (Dowding, 1985; New, 1984).

Siskind et al (1980) have produced a comprehensive account of structure response and damage produced by ground vibration from surface mine blasting. This work provides damage probability analyses for various conditions and Figure 2 is, in part, based on their 'alternative safe blasting level criteria' for residential structures. The 'safe' levels indicated fit British case history information well and it is suggested that these levels provide a useful basis for risk assessment. It must be emphasised that there are many recorded

A

E E

O

n

100 80 60

40

20

10 8 6

4

2

1 0.8 0.6

0.4

0.2

0.1

Safe levels of blasting vibration for houses

'Drywall' ,,- fro? / / " "

y 'Plaster' finish

• Disturbing or annoying

-t Perceptible

I

Imperceptible

I I I I I

2 4

Barely perceptible (transient)

6 810 20

Frequency (Hz)

I I I

40 60 80

Fig.2 Safe blasting and human perception vibration thresholds

00

T A B L E 2

Guide values for peak particle velocity during transient shaking (DIN 4150)

Structure Type

Offices and industrial premises

Domestic houses and similar constructions

Other building sensitive to vibrations

Peak particle velocity guide values (mm/s)

Foundations

<10 Hz

20

5

3

10-50 Hz

20-40

5-15

3-8

50-100 Hz*

40-50

15-20

8-10

Top storey on wall at floor

level (all frequencies)

40

15

* At frequencies higher than 100 Hz a higher guide value is allowable.

4

T A B L E 3

Swiss standard for vibration in buildings

Frequency Blasting induced Traffic or machine Type of structure bandwidth Hz PPV mm/s induced PPV mm/s

Steel or reinforced concrete structures such as factories, retaining walls, bridges, steel towers, open channels, underground tunnels and chambers

Buildings with foundation walls and floors in concrete, walls in concrete or masonry, underground chambers and tunnels with masonry linings

Buildings with masonry walls and wooden ceilings

Objects of historic interest or other sensitive structures

10-60 60-90 10-30 30-60

10-60 60-90 10-30 30-60

10-60 60-90 10-30 30-60

10-60 60-90 10-30 30-60

30 30-40

18 18-25

12 12-18

8 8-12

12 12-18

8 8-12

5 5 -8

3 3 -5

T A B L E 4

Some typical vibration limits enforced in Sweden when the foundation is on hard rock. Valid for short duration construction blasting (Persson et al, 1980)

Object

Concrete bunker Steel reinforced

High rise apartment block Modern concrete or steel frame design

Underground rock cavern roof Hard rock, span 15-18 m

Normal block of flats Brick or equivalent walls

Light concrete building

Swedish National Museum Building structure Sensitive exhibits

Computer center Computer supports

Circuit breaker control room

Limiting vibration parameter (peak value)

Amplitude Velocity Acceleration mm mm/s mm/s 2

0.4

0.1

20O

100

70-100

70

35

25 5

2.5

0 .5-2

instances of particle velocities well in excess of those indicated not having damaged structures. Conversely, if a structure is in very poor condition, even the slightest vibration can cause inherent weaknesses to become apparent.

Research is being carried out on response spectra techniques intended to improve prediction of vibration damage to structures (Siskind et al, 1980; Walker et al, 1982). These techniques use measurements of the mass, stiffness (or natural frequency) and

damping characteristics of the subject structures to assess their likely response to vibration. This approach is very effect ive and should be employed where specific structures are at risk and the additional investigations are financially acceptable.

In some respects the conventional form of damage criteria already incorporate an important element of response spectra techniques. For instance, the safe level of PPV (Figure 2) reduces considerably at frequencies less than 40 Hz. This coincides with the predominant frequencies (5 -30 Hz) associated with the response of residential structures.

Besides damage to man-made structures, consideration must sometimes be given to potentially unstable soil or rock condit ions in the vicinity of construct ion works. Cohesive soils are unlikely to be adversely affected by vibration, whereas loose sands may be caused to settle. In certain circumstances l iquefaction may take place and result in large ground sett lements. A 'state of the art' report on soil dynamics (and its application to foundat ion engineering) has been published by Yoshimi et al (1977) and it provides a useful background to the subject wi th an extensive bibliography. Seed and Goodman (1964) and Sarma (1975) have provided analyses concerned wi th the earthquake stability of soil slopes.

Blasting vibrations may also impose substantial dynamic loading on nearby rock slopes, which may result in rock fall or total slope failure. The analysis of rock slope stability has received considerable attention (Hoek and Bray, 1977), but most texts tend to consider only the 'static' (gravitational, hydrostatic, etc) Ioadings that affect stability. Al though it is reasonably straightforward to calculate the imposed dynamic loads due to a given vibration, it is in practice diff icult to assess their effects on the overall stability of the slope. A l lowance for even quite moderate values of PPV will predict major destabilizing dynamic stresses in the stability equations. Calculations of this kind often give very low values for the slope's ' factor of safety" that are not evidenced in the field. This may be explained, at least in part, by considering the oscillating character of the imposed vibrations in relation to the frictional properties of the discont inuity plane taken as the failure surface. The displacements that result from blast vibrations will almost invariably be very much smaller than the persistence (or 'wavelength') of the major interlocking asperities that contr ibute to the effect ive friction along the assumed failure surface. Thus, in practice, the vibrations may not produce suff icient relative movement, on either side of the discontinuity, for the asperities to 'ride over' each other and result in slope failure. Further research into this problem would seem to be desirable.

Reports of damage to structures have sometimes been associated wi th unexpectedly low vibration levels, but close reading of ten reveals that the

vibrations had been monitored in only one or two directions and almost certainly the maximum particle velocity that occurred had not been detected. To determine the maximum particle velocity it is essential to measure the vibrations in three mutually perpendicular directions and to establish the resultant value by vector summation. In the past it has been common practice to monitor construction blast vibrations in one direction only. This practice seems to fol low that for measuring the effects of quarry blasting, where the direction of the predominant vector component may be predicted with some confidence. There is, however, often a fundamental difference in the type of wave motion of concern during nearby construction in comparison with that normally associated with quarry workings. Quarrying will usually involve relatively large rounds of explosives at substantial distances from residential structures. It is therefore usual to measure these vibrations at many tens or hundreds of metres from the source. At these distances and with a near- surface source the predominant ground vibrations will be due to surface wave motions with some refracted body wave arrivals. The direction of these motions has become well understood and in practice it has been found satisfactory (and expedient) to measure vertical or vertical and radial motions only. This practice is not acceptable for construction blasting close to structures. In this situation it will often be body wave motions that will predominate. It is rarely possible to predict with any confidence the direction of maximum particle velocity in these circumstances and it is therefore essential to use triaxial transducer arrays.

Where ground vibration is of a continuous nature potential damage thresholds should be set at rather lower levels than those discussed above for transient vibrations. How much lower is a matter for some speculation, as little research has been carried out to determine the effects of continuous vibration on urban structures. Some guidance may be provided, however, by analogy with damage criteria associated with road traffic induced vibration. Whiffin and Leonard (1971) reviewed traffic induced vibration and concluded that 'architectural' damage may occur at PPV in excess of 5 mm/sec and that structural damage may take place at PPV in excess of 10 mm/sec. It should also be noted that much of the vibration induced in structures from road traffic is transmitted as sound through the air rather than vibration through the ground.

Criteria recommended by the Swiss Association of Standardisation (1978) for sources of a continuous nature are given in Table 3. Here too 'traffic' and 'machine' sources are grouped together. The Department of Transport specify site procedures for the use of explosives and blasting in the Specifications for Road and Bridge Works (Department of Transport, 1986).

TABLE 5

Example vibrations (PPV) in building during normal use

Resultant PPV (ram/s)

Modern masonry dwelling Old dwell ing house (Thick, Vibration source Modern steel framed office house l ime-mortar masonry)

Normal footfalls Foot stamping Door slams Percussive drilling

0.02-0.2 0.2-0.5 10-15

5-25

0.05-0.5 0 .3-3 .0 11-17 10-20

0.02-0.3 0.15-0.7

3 - 9 10-15

Table 5 is included to put construction induced vibrations into an 'ambient' environmental perspective. The table lists a range of resultant particle velocities measured in three examples of common types of structure. The transducers were located at various positions on walls usually within 1-4 m from the source. Clearly it would be difficult to justify a limiting level for construction induced vibrations which is lower than that to which the structure is subjected in normal use. Temperature changes and other environmental variations will also impose potentially damaging strains but these are difficult to compare with vibrations because their timescale is usually much longer and failure mechanisms are likely to be significantly affected.

In reviewing the 'safe vibration levels' given over the past few decades it is generally apparent that the more recent the publication the lower the 'acceptable' value has become. This trend does not appear to be supported by much new field evidence of damage at lower particle velocities and indicates a shifting social climate rather than a change in engineering values.

2.3 H U M A N PERCEPTION AND INTRUSION CRITERIA

The fears expressed concerning vibration damage are often a result of the extreme sensitivity of the human body to vibration, especially in the low-frequency range (1-100 hz). Human reaction is more likely to be influenced by previous experience and understanding than by the actual level of vibration itself; a person's state of health, temperament and age will all contribute to this reaction. Jackson (1967), Soliman (1968) and Guignard and Guignard (1970) have published useful works on human response to vibration.

The work of Reiher and Meister (1931) has stood the test of time very well and is useful in defining in quantitative terms subjective descriptions of human perception of vibration (see Figure 2). ISO Standard 2631 (1978) and BSl BS 6472 (1964) both provide valuable guidance on acceptable levels of human exposure to vibration.

Apart from helping to define thresholds of perception and annoyance, tolerance scales alone do not provide sufficient information for defining limits for construction generated vibrations as they are generally applicable to situations in which vibration is an accepted part of the environment. A dif ferent type of criterion has to be considered in areas where vibration does not normally occur or is at a very low level. Vibration may then be considered as intrusive. It is the unpredictabil i ty and unusual nature of a source rather than the level itself that is likely to result in complaints. The effect of intrusion tends to be psychological rather than physiological and is more of a problem at night when occupants of buildings expect no unusual disturbance from external sources.

A second type of involvement of people wi th vibrations is in interpreting the effect on buildings or their contents. Not surprisingly, this is particularly true where the person concerned is the owner. Even the slightest disturbances from an unusual source may excite anxiety and draw attention to minor cracking of plaster or similar effects that were pre- existing or may otherwise have remained unnoticed.

It is the author's experience that when vibration levels from an unusual source exceed the human threshold of perception (PPV, 0 .2 -0 .3 mm/sec) complaints may occur. In an urban situation serious complaints are probable when PPV exceed 3 mm/sec. Complaints will increase with the duration of the vibration and will be most severe when the ' intrusion' occurs outside normal work ing hours, particularly at night.

These values apply to vibration of a cont inuous or semi-continuous nature. Swedish experience (Persson et al, 1980) indicates that tolerance levels may be considerably higher during blasting when the vibrations occur as infrequent shocks of brief duration. Peoples tolerance will also be improved provided that the origin of the vibration is known in advance and no damage is done. Wiss and Parmelee (1974) have shown that levels of perception are considerably higher when the vibration is of a transient rather than cont inuous nature. They indicated that vibrations at 1.5 mm/sec are 'barely perceptible' and it is not until levels exceed

22 mm/sec that they become 'strongly perceptible'. These values relate to their 'mean subject response', their threshold levels being rather lower.

In assessing human response to construction induced vibrations it should not be assumed that reason will always prevail. For instance, although a person may be shaken vigorously for several hours during a coach journey wi thout considering complaint they may not tolerate less than one tenth this vibration level for just a few seconds a day during blasting works. The wri ter has found that a useful method of allaying the fears of home owners is to show them the worrying vibrations (on a f loor or wall of their house) being recorded on a chart recorder and ask them toc lose the door during the recording. In many cases the vibrations from the door closing will be seen to be considerably greater than those from the construction source and the owner's worries will be abated.

3 T H E P R E D I C T I O N OF V I B R A T I O N

3.1 NON-EXPLOSIVE SOURCES

Construction induced vibration may be predicted either by analogy with information from other sites or by carrying out site trials prior to the start of works. Figure 3 provides a summary of measurements taken by the TRRL which indicates the relative effects of various construction related sources. Fuller details of vibrations caused during tunnell ing works in urban areas are given elsewhere (New, 1982).

The Figure provides general guidance only as condit ions that affect the input and transmission of vibrat ion will vary considerably from site to site. This form of presentation shows the likely relative intensities from various sources and highlights the fact that even a 1 kg charge of dynamite wil l produce larger vibrations than those from most 'continuous' types of source. Fatigue effects that result from repeatedly taking a structure through a stress cycle wil l , however, be more likely to occur under condit ions of 'continuous' vibration. Acceptable levels wil l therefore be related to the nature and duration as well as the maximum level of the vibration.

Spectral analysis of the particle veloci ty data indicates that most of the energy from these sources lies in the f requency band 10-100 Hz. However it is not unusual to encounter circumstances where considerably higher frequencies predominate particularly close to the source. For instance the tunnell ing machines caused vibration peaking at frequencies from 200-400 Hz a few metres from the face and railway trains in the rock tunnel yielded signif icant energies at frequencies over the bandwidth of 200-800 Hz.

The vibration caused by the collapse of large structures is often the cause of concern particularly where other structures or buried services are nearby. For instance the demolition of a tower block (Plate 1) above the London Underground railway has been recently reported by New (1983) and Skipp (1984) gives information regarding vibrations caused by the collapse of cooling towers. A recently introduced method of demolition utilises the expanding power of a slurry caused by chemical reaction. This material takes the form of a powder which when mixed with water and poured into suitably drilled holes expands with considerable force, over a period of several hours, breaking out the concrete or other material. Recently it was necessary to remove a large steel/concrete machine foundation in a laboratory at TRRL. Environmental considerations precluded the use of explosives and excessive periods of noise and vibration from pneumatic breakers were highly undesirable. The process caused very little noise or vibration (less than 1 mm/s was observed) which was attributed to the cracking of the concrete and opening of the fissure. The product is described as a demolition agent but may be considered for rock excavation in special circumstances (eg pre splitting) although this may not be economically viable at present.

Another non-explosive method of excavation and demolition uses pressurised carbon dioxide for dislodging coal and other materials. This technique is claimed to be particularly suited for use in the presence of inflammable vapours and gases.

A

E E

o

O .

100 80 60

40

20

10 8 6

4

1 0.8 0.6

0.4

0.2

0.1 1 2 4 6 810 20 40 6080100

Distance (m)

Fig.3 Relative PPV from various sources

8

ms i f

m m ~ m

D

~ t ~ ~ £

|

Plate 1 Demolit ion of a tower block above the London Underground

Neg. no. E 2 2 / 8 2

9

3.2 EXPLOSIVE SOURCES The magnitude and spatial decay characteristics of blast-induced vibration have been extensively explored by numerous authors. For instance, the US Bureau of Mines publication by Siskind et al (1980) gave a full discussion of ground vibration and its effects on some fully instrumented test structures. Their report is generally concerned with quarry blasting, but it provides a broad background and bibliography for the whole subject area. Oriard (1979) gave a very useful and reasoned account of urban blasting practice and philosophy based on many years of experience in the USA. Ambraseys and Hendron (1968) provided a description of the dynamic behaviour of rock masses with special reference to blasting vibrations whilst Skipp (1984) gives a wide range of information on explosive and other man made vibrations.

Case history records of blasting beneath British conurbations have been given by Pakes (1976) and Ashley and Parkes (1976). Both papers described difficulties associated with vibration damage and intrusion in residential areas and discussed ways of minimising disturbance to the local population. More recently, Persson et al (1980) have described blasting techniques used to minimise such problems in Sweden.

300

-~ lOO E E > 3O

10

r r 3

1

0.3

0.1

Heavily confined charges (eg early delays of a burn-cut tunnel round, some pre-splitting)

'Upper b

i . experience IPPV tends~

"to decrease with rock \ strength and charge \ confinement)

1 10 100 ½ ½

riM (m/kgf)

Fig.4 Peak particle velocity as a function of scaled distance - range of observed field values

Much of the data on blast vibration has come from the mining and quarrying industries, and great care must be exercised if this information is to be used in a civil engineering context. Mineral extraction blasts generally comprise large rounds which are detonated at sites fairly remote from habitation. Construction blasts often involve much smaller rounds but may be very close indeed to vulnerable environments. Because of important differences in the character of the signif icant wave motions, measuring systems wi th a more demanding specification may be required for civil engineering applications. For instance, triaxial transducer arrays should be used and the frequency response of the system should often be able to cope wi th at least 500 Hz.

The form of the PPV prediction equations is given in the next section and in Figure 4 the PPV v 'scaled d istance' (square root format) is used to show the wide variation of PPV which may be expected from site to site under various blasting condit ions. This graph is based on experience from various sites in the United Kingdom and is similar to that observed in America by Oriard (1979). The amount of energy from the explosive which is transmitted as vibration is af fected by numerous factors. It will depend on the type of explosive charge and condit ions imposed by the rock mass. The degree of conf inement (constrict ion) imposed will be a major factor but other effects such as explosive/rock impedance match and s temming and initiation procedures will also be of importance (Dowding, 1985; New, 1984).

Generally the upper bound line in Figure 4 will apply when an explosive is heavily confined by a strong rock mass and fragmentation is poor. This will often occur during the initial delays of a tunnel round or where the burden is excessive. In general the more energy used in fragmenting and throwing the rock the less will be transmitted as ground vibration. Where blasting procedures allow considerable energy to be vented to atmosphere (eg plaster shooting, poor stemming) this will also reduce ground vibration levels although air overpressures may then begin to dominate damage assessments.

It has been considered that pre-splitting techniques also give rise to high levels of ground vibration (Oriard, 1979) and this is likely to be true where the pre-split is created by a row of holes tightly packed with explosive. In this case there is minimal fragmentation until the subsequent bulk blast removes the burden beyond the pre-split plane. However a method of pre-split blasting has been developed (Matheson, 1983) which uses air decoupled charges in the plane of the finished rock cut. This method can provide a relatively undamaged face which requires a minimum of maintenance, a factor of considerable importance in the design of major road schemes. A further advantage is that the decoupled charges cause considerably less vibration than similar charges tightly packed. Figure 5 summarises the results of a series of trials carried out by TRRL (New, 1984) to determine the effect of various degrees of decoupling. (The decoupling ratio is expressed as the

10

A

E E v >

0_

rr

100

10

0.1

0.01

Coupled explosives

(Trimobel and SG80)

Tr imobel decoupled

; 75ram hole

Tr imobel decoupled

; 100mm hole

. D e c o u p l i n g rat io

I I

10 100

r /M ½ ½ ( m / k g f )

Fig.5 Peak particle velocity as a function of scaled distance with varying decoupling ratio

ratio of the drillhole to explosive diameter.) These trials were carried out in strong schists at the site of a new road scheme which required extensive rock excavation. Blasting gelatine in plastic sleeves was centred in the drillholes using cruciform spacers with the holes tightly stemmed in the top one metre. The annular space around the explosive substantially reduced the pressure applied to the drillhole and created a mismatch in the effective rock/explosive impedences. These factors significantly reduced the vibration caused by a given charge weight without materially changing the spectral distribution (hence the propagation characteristics) of the energy transmitted by the rock mass.

4 BLASTING TRIALS

4.1 MEASUREMENT PROCEDURES AND EQUIPMENT

Trial blasts should be carried out where initial desk studies show that nearby structures could be at risk. Costs and inconvenience will generally be minimised if the trials are carried out as part of the normal site investigation programme for the works. This will also ensure that the information gained will be of assistance in the preparation of tender documents and bids for the contract.

If trial blast data are available to contractors at the tender stage, it wil l provide valuable guidance as to the maximum charge weight per delay that may be used during the works. Where no such information is provided the usual course of action is for the contractor to use the initial blasts to determine his own 'safe' charge weights with sometimes unexpected consequences. Clearly, the pre-knowledge provided by a trial blast is invaluable to all parties concerned.

The trials should be designed with a clear concern for the factors which wil l influence the induced PPV during the excavation works, and wherever possible the materials and techniques used during the trials should simulate those expected for the full-scale excavations. This wil l include:

(i) the type of explosive,

(ii) the drillhole diameter and depth,

(iii) the coupling of the explosive to the rock (tamping, stemming, water in-hole, hole/ explosive diameter ratio),

(iv) the confinement imposed by the rock mass on the explosion

(v) multiple hole shots with appropriate hole spacing and detonating accessories,

(vi) similar shot point/structure ranges and paths,

(vii) wherever possible full-scale charge weights (although the trial blast sequence should commence with small charges),

(viii) the ranges at which PPV is measured should be as broad as possible, not only centred on those of specific interest,

(ix) where millisecond delay detonators are used the minimum delay period should be sufficient to al low the vibrations from a given delay to die away before the arrivals from the next,

(x) changes in site conditions between the time of the trials and the works must also be considered. For instance, are there any new sensitive structures to be introduced or will seasonal changes in the water table have any effects?

The vibrations from the trial charges should be measured at several distances in line or lines away from the source, the transducer locations being chosen in the light of preliminary site investigation data to allow for possible directional variations in the propagating characteristics of the rock mass.

It is recommended that plenty of data points be obtained. The cost of firing addit ional charges, or obtaining measurements at additional ranges, will be small compared to the mobilisation costs involved and will probably provide valuable information, or as a minimum provide greater confidence in the derived predictive equations.

11

Neg. no. CR604/80/4a

Plate 2 Site of trial blasting for a new road tunnel

Plate 2 shows a typical blasting trial site. A major road tunnel is to be blasted through the hillside to the right of the plate and careful consideration had to be given with regard to the safety of the adjacent main railway line tunnel. Trial charges were fired in the vicinity of the new tunnel portal and vibration measurements taken at various locations on the hillside and in the rail tunnel. The blasting was directed from the portable laboratory shown in the left foreground. Further seismic measurements at this site provided detailed data on ambient vibration levels and subsequent wavepacket analysis provided information on the dynamic properties of the rock mass.

Transducers are readily available with output voltages proportional to particle velocity and geophone types are self-generating (need no power supply) and are ruggedly designed for field use. Measurement of particle velocity allows single-process derivation of acceleration, by differentiation, and of displacement by integration if required.

Measurement of particle velocity should be made in three mutually perpendicular directions. This allows the calculation of the resultant particle velocity, VR,

by the vector summation of the component velocities (V~, V2 and V3):

VR = (V~ 2 + V22 + V32) 1/2 (1)

Vibration measuring instruments that display the resultant peak particle velocity directly on a chart record are available. If such an instrument is not used, the three vectors must be summed by 'hand' or computer processing. The true resultant is obtained by summing the three component values at simultaneous times. The 'pseudo resultant', sometimes referred to, is obtained by summing the maximum value obtained for each component during the period of the vibrations.

It is vital for the specification of the instrument chosen to be appropriate to the vibrations that it is to record, particularly in terms of frequency response and sensitivity. For instance, some instruments are limited to an upper frequency bound of 200 Hz. This type of equipment may be satisfactory but will, of course, be insensitive to vibrations above 200 Hz that could be present. High frequencies (200-500 Hz) wil l often be encountered in the region close to construction works.

12

Range = 10.5m

,°°It 80

I I I

L '~ 40

2O

0 i

0 400 8oo Frequency (Hz)

Range = 23.3m

0 400 800 Frequency (Hz)

v

LU

rr

Range 45.0m

100 L

8 0 -

6 0 -

4 0 -

2 0 -

0 0

Range 140.7m

L

1 O0 200 0 100 200 Frequency IHz) Frequency (Hz)

Fig. 6 Effect of source distance on energy spectral density ( lkgf charge)

Rocks and soils tend to act as low pass filters, that is, low frequency vibrations are subject to less attenuation during propagation than higher frequencies. This effect is particularly noticeable in the weathered geologic materials close to the surface rather than at depth where weathering effects are not present and the in-situ stress field is higher. In general these frequency dependent losses cannot be satisfactorily explained in terms of a linear frictional mechanism and scattering effects are likely to be significant. Figure 6 shows how the energy spectral density from a 1 kg dynamite charge (fired in a strong psammitic rock) varied when measured at various ranges. Note how, close to the charge, the maximum energy is in the 200-300 Hz band and significant energy is present at up to 1 kHz. This changes rapidly away from the near field and at 23 m the energy peak occurs at about 100 Hz. At 140 m the energy is contained in a relatively narrow band between 20 and 60 Hz. Thus equipment suitable for measurements at ranges of 25 m and beyond (in this particular case) may not be suitable for use close to the source. Where doubt exists with regard to the frequencies present expert opinion should be sought and trial measurements made with systems sensitive to high frequencies.

Instruments that record peak particle velocity only may be left for many days without attention and are particularly useful for routine monitoring during construction. Systems that provide high-speed recordings that show the full vibration waveform will, however, often prove useful for diagnostic purposes during initial works. Periodic checking and recalibration of site measuring equipment is essential for reference purposes. Bollinger (1971) and Stagg and Engler (1980) have dealt with the theory and choice of vibration measuring equipment and Jaeger and Cook (1976) and Kolsky (1963) have provided an appropriate background on strain wave propagation in rock and other solids.

4.2 D A T A PROCESSING A N D P R E S E N T A T I O N

4.2.1 Square and cube root scaling methods

Broadly, the ground motions resulting from a blast will depend upon the weight of explosive fired, the distance between the explosion and the observation point and the rock mass characteristics. The effect of each of these factors is complex and at present a satisfactory theoretical approach for calculating the form of these motions has not been developed. Therefore, scaling of field measurements is used almost exclusively to predict the magnitude and character of the vibrations from explosions.

A wide range of field data is available and several similar empirical approaches are in common use. The principal variables are usually related by an equation of the form:

PPV = KM"r -~ (2)

where PPV is the peak particle velocity, M is the charge weight and r is the distance from the explosion. The constants K, e and /~, are dependent on condit ions imposed by the site and the type of explosion. Two special cases of this formulation are most commonly used:

PPV = K( r /~ -M)-° (square root scaling) (3)

and PPV= K(r /~-M)-n (cube root scaling) (4)

where again n is an empirical constant. Values of K and n typically range between 700-2000 and 1 .5 -2 respectively (for M in kg and r in metres).

The 'square root ' and 'cube root" scaling methods both al low simple graphical presentation of the derived site laws. These laws are shown graphically

13

o

O.

K J ~ p p v = K(r/M1/2 or 1/3)

S,o e - -

- - n

Scaled distance r/M1/2 or 1/3"

*note for site specific scaling the exponent 1/2 or 1/3 will become (Z/~

Fig.7 Scaled distance site law format

(Figure 7) as a straight line (on a log-log plot) with a slope of - n and an intercept of K at unit scaled distance.

Because of the difference in the forced scaling of r to M these two methods lead to different predictions of peak particle velocity based on the same field measurements. Where extrapolation beyond the bounds of the field data is not required the predictions will be similar whichever method is used. However, where prediction is required beyond the range of charge weights or distances covered by the trial blasts, significant differences may result between the two methods. This situation is clearly unsatisfactory, as is the somewhat arbitary scaling of r to M inherent in both methods.

It has been argued that these empirical laws should be 'shaped' by dimensional analysis (Ambraseys and Hendron, 1968; Newmark, 1968). The dimensional analysis results in 'cube root' scaling laws for explosions of different sizes in the same medium. This approach has led to equation 4 being used extensively (Sauer et al, 1964; Newmark and Haltiwanger, 1962). Note that the size of the charge, M, is given in terms of explosive weight rather than energy, this being valid as energy is directly proportional to charge weight. If, however, data are scaled from various types of explosive the relative strengths of the explosives should then be taken into account.

It has been suggested (Ambraseys and Hendron, 1968) that square root scaling has 'no basis in dimensional analysis' and that cube root scaling should be used if estimates of motions are required which necessitate extrapolations beyond the limits of existing data. However, experimental evidence (Devine, 1966; Snodgrass and Siskind, 1974; Devine and Duvall, 1963) indicates that in certain situations "square root" scaling does normalise the data very well. That is, the correlation coefficients were higher using r / V ~ scaling than for r/~/-M scaling.

In fact because site conditions rarely comply with the assumptions made for the dimensional analysis

neither scaling method is strictly appropriate and the best estimate of relative scaling between r and M is site-specific.

It is usual to present blast vibration data in scaled distance graphical format with a 'best fit' straight line obtained by linear regression analysis. A great volume of data in this form is available in the literature and almost without exception the peak particle velocity is well represented by a power law decay with scaled travel distance. That is, the measured PPV decay can be represented by a straight line with negative slope on a log-log plot, although the actual slope and intercept values may vary considerably from site to site and with different blasting conditions (see Figure 4 for range of observed field values).

More complex forms of equation have been proposed in an attempt to include frictional dissipation effects. Such equations take the form

PPV = K(r/k/--M)- nexp( - ar)

where a is the spatial attenuation coefficient.

This type of equation allows a non-linear regression line to be fitted to the data thereby improving its correlation coefficient. However as the majority of site data are well fitted by linear equations (on log-log plots) the additional complication is unlikely to find wide application. Moreover this formulation is only appropriate at a single harmonic frequency which is an unacceptable assumption for construction sites particularly close to the source (see Figure 6). Also calculations based on the known frictional properties of rocks indicate that losses from this particular mechanism are unlikely to be significant at ranges of interest from most construction sources (New, 1984).

4.2.2 Site specific scaling using multiple regression analysis

Having obtained field data relating peak particle velocity to explosive charge weight and range, it is clearly important to process and present the information in the best and most useful manner. The arbitary forced scaling of r to M implicit in equations 3 and 4 may be unsatisfactory to some extent, but does allow simple graphical presentation of the data in an easily usable format. However, the exploitation of equation 2 allows site specific scaling of r to M which will enable predictions to be made using equations better correlated with field data. The improved correlation is the direct consequence of the more versatile equation with three variables (PPV, M and r) rather than two (PPV) and 'scaled distance', rive).

The constants K, ~ and/3 in equation 2 may be determined by transforming the equation and applying

14

a three-variable multiple linear regression analysis to the data as follows:

from eq 2 (PPV = KMar -~)

log PPV=Iog K + , log M- /3 log r (5)

In this form 'best' values for K, a and/3 may be calculated based on the usual regression criteria that the sums of the squares of the deviations shall be minimised. For simple regression (two variables) these deviations are taken as deviations from a straight line whereas for this three variable analysis they are represented by deviations from the plane, KABC, shown in Figure 8.

Although the projection shown in Figure 8 is useful in visuaiising the equations 5 or 6, it is not appropriate for routine graphical presentation of data. It is possible, however, to plot the data, using the constants and exponents calculated by the multiple regression, in the convenient PPV v 'scaled distance' format. By transforming equation 2 to the form

PPV= K(r/M ~/~) -~ (6)

(where the term r /M "/~ is the 'scaled distance') the data may be presented in the same convenient format as for square or cube root scaling (see Figure 7) but without the forced scaling relationship. General comparisons of varying site law regressions may require reduction of the data to a unified scaled distance.

The coefficient of determination for the three variable regression will always be better than that obtained

> Q. D.

.J

Plane KABC is defined by: Log PPV = Log K +E'Log M --/~Log r

I I

I I I - I

S,ope-- / /

I I

/

Log r

Fig. 8 3-D site law format

A o~ E E > o_ 0. c

rr

100

10

0.1

0.01

• Upper bound PPV

" ~ i / MO'65)-- 1"81

PPV = 4487(r/MO'65) -1'81

I I

10 100 r/M O'6s(m/kgfO.65)

Fig.9 Example of particle velocity data with site specific scaling

using the two-variable analysis for square or cube root scaling. (Except where, by chance, ~//3 is equal to 1/2 or 1/3 when the correlation will, of course, be numerically the same.) Routine statistical tests of the significance of the correlations should be carried out and confidence limits calculated for the predictive equation if required.

Figure 9 shows a typical data set from a recent series of trial blasts at the site of a major rock excavation for a new road. Note that the 'upper bound' line is drawn parallel to the regression line just above the data points. The calculation of safe charge weight based on a maximum observed PPV may be made on a safety factor approach rather than by reference to a statistical 'confidence limits' description. For instance, the occupier of a structure might be happier with a statement that charge weights are restricted to less than, say half that associated with possible damage (based on the trial blasts), rather than being informed that there was a 2% probability that the exceptable vibration level would be exceeded.

5 CONTRACTUAL SPECIFICATION AND VIBRATION CONTROL

Figure 10 provides a f low diagram of questions which should be asked at the planning stage of major construction projects. These procedures have been

15

fol lowed for road construction projects and found to provide a useful basis for the investigation of vibration associated diff icult ies.

A t present there are three basic approaches with regard to the contractual specification of vibration control. They are as fol lows:

(a) No mention of vibration in the contract,

(b) The inclusion of a 'no damage, no intrusion' clause intended to absolve the Client of any responsibility, and

(c) The imposition of a limiting value for vibration usually defined in terms of resultant PPV at some distance or structure.

Method (a) may be adopted when it is the considered opinion of the Client and his Engineer that there are no possible vibration problems. It is also used widely by default and can lead to serious problems. Notably these problems wil l result in claims made under clause 12 of the Institution of Civil Engineers Conditions of Contract (5th Edition), and will be based on additional costs necessarily incurred by the Contractor due to circumstances unforeseen at the t ime of tender.

Method (b) is often used where the Client is not inclined to incur the additional expense of extending his site investigation to cover vibration. The inclusion of the 'no damage/no intrusion' clause may seem to be of assistance to the Client by passing vibration linked risks to the Contractor. This approach has two main drawbacks. Firstly, the Contractor may still feel that he can recover his additional costs through the Clause 12 claims procedure. Secondly, bidders for the Contract may feel obliged to increase their prices to al low for the vibration associated losses for which they feel liable. Neither circumstance is conducive to obtaining the best job at the lowest final cost.

Method (c) should involve an initial desk study perhaps fol lowed by blasting trials as found necessary. Where trial blasting is carried out as suggested in Section 4 and an appraisal of environmental hazard is made, a considerable element of uncertainty may be removed from many blasting contracts. Where full disclosure of trial blasting results is made to bidders an otherwise unquantified element of contractual risk is removed and bids may therefore be made wi th less allowance for vibration damage contingencies.

It must be recognised that the specification of vibration criteria by the Client does result in some extent of risk-sharing wi th his Contractor. This circumstance fol lows recent trends within the construction industry which suggest that such risk sharing may be mutual ly beneficial to Employer and Employee (CIRIA, 1978).

I Will vibration ~---~- No be caused ?

Yes Sensitive people ~ or structures No in vicinity ?

Yes I Desk stu'dy. Is /

damage/intrusion ~ No possible ? J

Yes i

I Site specific trials / Will damage/ ~ - N o ~ intrusion occur ? /

Yes

I odi,v I v0cuate I1 design and/or Irestore II construction I pr°perty afterl I method I cOnstructiOn I |

I Proceed with I construction

Cancel I project

Fig.lO Vibration hazard - planning f low diagram

The Engineer should also be aware of the legal aspects relating to vibration from construction sites. In statutory law vibration is included in the term 'noise' and dealt with by Part III of the Control of Pollution Act, 1974. These powers are usually exercised by Environmental Health Officers within Local Authorities. In particular Section 60 of the Act deals with the control of noise and vibration from construction sites and makes the provision that a local Authority may serve a notice imposing restrictive requirements prior to, or during construction operations. Section 61 allows 'prior consent' to be obtained based on the provision of information regarding the works, the method and proposed steps to be taken to minimise noise and vibration. Such a 'prior consent' may be of assistance if problems occur during the course of the works but will not necessarily constitute any ground of defence against proceedings related to nuisance suffered by the occupier of premises. Offences under the Act will be dealt with in the Magistrates Court which may impose an initial fine plus a subsequent daily fine. Noise and vibration may often be dealt with more quickly through Common Law procedures, 'Private' and 'Public' nuisance actions may be taken to the High Court where damages may be awarded, an injunction served, or in certain circumstances the offender committed to prison.

As part of his contractual duties to the Client it will be the Engineer's role to act as the supervising authority in relation to vibration control and

16

measurement. If suitable expertise and equipment are not available in 'in house' he should employ a specialist sub-consultant with the appropriate experience. The trial blast data will, of course, only describe that part of the site in the immediate vicinity of the trials. Many factors will influence blast wave propagation and careful measurements should be made during all construction blasts. The handling of damage claims will be eased where such records are available. Structural surveys with photographic records, made prior to the commencement of the works, may also prove valuable in the settlement of claims which may arise.

Some Clients and Engineers are attracted by the idea of imposing unjustifiably low limits of vibration on their Contractors. These limits are a form of regressive conservatism which is to be avoided: A common example is the use of the 'method of halves' whereby a specifying engineer halves the limit set by his predecessor on a similar job. These methods inevitably impose unreasonable restraints on the Contractor, thereby increasing costs to the Client. A large proportion of major civil engineering works are sponsored by the public sector and, although safe limits must be imposed to protect the local people, overconservative limits and unnecessary restrictions will be a charge on the community as a whole. It is therefore vital that authorities responsible for setting vibration limits do so on an informed rather than on an arbitrary basis.

6 ACKNOWLEDGEMENT

The work described in this Report was carried out in the Ground Engineering Division (Division Head: Dr M P O'Reilly) of the Highways and Structures Department (Department Head: Mr N W Lister) of TRRL. The author gratefully acknowledges the valuable contributions made by the employers, consulting engineers and contractors without whose permission and cooperation the site measurements presented in this paper could not have been obtained.

7 REFERENCES

AMBRASEYS N N and HENDRON A J Jr. (1968). Dynamic behaviour of rock masses. In Rock mechanics in engineering practice Stagg K G and Zienkiewicz 0 C eds (London, etc: Wiley), 203-36.

ASHLEY C and PARKES D B. (1976). Blasting in Urban Areas. Tunnels Tunnell, 8, No. 6, Sept., 60-7.

ATTEWELL P B and FRY R H. (1983). The effects of explosive detonations and mechanical impacts of adjacent buried pipelines. In 'Europipe 83' Conference, Basle, paper 16, 123-128.

BOLLINGER G A. (1971). Blast vibration analysis (Carbondale, II1: Southern Illinois University Press), 147p.

BRITISH STANDARDS INSTITUTION. (1984). Guide to evaluation of human exposure to vibration in buildings (1 Hz-80 Hz) BS 6472.

CIRIA. (1978). Tunnell ing--improved contract practices Report 79. Construction Industry Research and Information Association, London.

DEVINE J F. (1966). Avoiding damage to residence from blasting vibrations. Highway Res. Record, No. 135, Highway Res. Board, Natl. Res. Council-- Natl. Acad. Sci.

DEVINE J F and DUVALL W I. (1963). Effect of charge weight on vibration levels for millisecond delayed quarry blasts. Earthquake Notes, Eastern Section, Butt. Seismol. Soc. Am. 34, No. 2, 17.

DEPARTMENT OF TRANSPORT. (1986). Specification for Road and Bridge Works. 6th Ed. HMSO.

DIN 4150. (1938). Vibrations in buildings--effects on structures. Deutsch Norm. April 1984.

DOWDING C H. (1985). Blast vibration monitoring and control. Prentice-Hall, N J, USA. 297 pp.

GUIGNARD J C and GUIGNARD E. (1970). Human response to vibration--a critical survey of published work. University of Southampton, Institute of Sound and Vibration Research Memo 373.

HOEK E and BRAY J W. (1977). Rock slope engineering, revised 2nd edn (London: IMM, 1977), 402 p.

INTERNATIONAL STANDARDS ORGANISATION. (1978). Guide for the evaluation of human exposure to whole body vibration and shock. ISO 2631. 2nd Edition.

JACKSON M W. (1967). Thresholds of damage due to ground motion. Proceedings international symposium on wave propagation and dynamic properties of earth materials, University of New Mexico, 961-9.

JAEGER J C and COOK N G W. (1976). Fundamentals of rock mechanics, 2nd edn (London: Chapman and Hall; New York: Wiley, a Halsted Press Book), 585p.

KOLSKY H. (1963). Stress waves in solids (New York: Dover Publications) 223p.

LANGEFORS U and KIHLSTROM B. (1978) The modern technique of rock blasting, 3rd edn (New York: Wiley, Halsted Press), 438 p.

17

MATHESON, G D. (1983). Pre-split blasting for highway rock excavation. Transport Road Res. Lab. Rep. 1094. (TRRL, Crowthorne)

NEW, B M. (1978). The effects of ground vibration during bentonite shield tunnelling at Warrington. Transport Road Res. Lab. Rep. 860. (TRRL, Crowthorne)

NEW, B M. (1982). Vibrations caused by underground construction. Proc. Tunnelling 82, (London: IMM), 217-229.

NEW B M. (1983). Explosive demolition works above a railway tunnel. Tunnels and Tunnelling, July, 15, 7, 18-20.

NEW B M. (1984). Explosively induced ground vibrations in civil engineering construction. PhD Thesis, University of Durham.

NEWMARK N M. (1968). Problems in wave propagation in soil and rock. Proc. Int. Soc. Wave Propagation and Dynamic Properties of Earth Materials, Univ. of New Mexico Press, Albuquerque, pp 7-26.

NEWMARK N M and HALTIWANGER J D. (1962). Principles and practices for design of hardened structures. Tech. Rep. No. AFSWC-TDR-62-138. Air Force Special Weapons Center, Kirtland Air Force Base, New Mexico.

ORIARD L J. (1979). Modern blasting in an urban setting. Atlanta Research Chamber Applied Research Monographs, interim report. Rep. no. UMTA- GA-06-0007-79-1, June.

PAKES G. (1976). Edinburgh sewage disposal scheme; tunnelling work. In Tunnelling '76 Jones M J ed. (London: IMM), 3-15.

PERSSON P A et al. (1980). Underground blasting in a city. In Subsurface space: proceedings of the international symposium (Rockstore '80), Stockholm Bergman M.ed. (Oxford, Pergamon, 1980), vol. 1, 199-206.

REIHER H and MEISTER F J. (1931). Human sensitivity to vibrations. Forschung auf dem Gebiete des Ingenieurwesen, 2, no. 11, 381-6. (German text)

SARMA S K. (1975). Seismic stability of earth dams and embankments. Geotechnique, 25, no. 4, 743-61.

SAUER F M, CLARKE G B and ANDERSON D C. (1964). Nuclear geoplosics-- Part IV-- Empirical analysis of ground motion and cratering. Stanford Res. Inst. for the Defence Atomic Support Agency, Report DASA-1285 (IV).

SEED H B and GOODMAN R E. (1964). Earthquake stability of slopes of cohesionless soils. J. Soil Mech. Found Div. ASCE, 90, Nov. 43-73.

SISKIND D et al. (1980). Structure response and damage produced by ground vibration from surface mine blasting. Rep. Invest. US, Bur. Mines 8507, 74 p.

SKIPP B O. (1984). Dynamic ground movements-- man-made vibrations. In Ground movements and their effect on structures. Ed Attewell, P B and Taylor, R K. Surrey University Press.

SNODGRASS J J and SlSKIND D E. (1974). Vibrations from underground blasting. RI 7937, US Bureau of Mines, Washington, USA.

SOLIMAN J T. (1968). A scale for the degrees of vibration perceptibility and annoyance. Ergonomics, 2, 101-22.

STAGG M S and ENGLER A J. (1980). Measurement of blast-induced ground vibrations and seismograph calibration. Rep. Invest. US, Bur. Mines 8506, 62 p.

SWISS ASSOCIATION OF STANDARDISATION. (1978). Effects of vibrations on constructions. Seefeldstrasse 9, CH8008, Zurich, Switzerland.

WALKER S, YOUNG P A, and DAVEY P M. (1982). Development of response spectra techniques for prediction of structural damage from open-pit blasting vibrations. Trans. Instn Min. Metall. (Sect. A: Min. industry), 91, A55-62.

WESTINE P S, ESPARZA E D and WENZEL A B. (1978). Analysis and testing of pipe response to buried explosive detonations. Southwest Research Institute, San Antonio, Texas. American Gas Association No. L51378.

WHIFFIN A C and LEONARD D R. (1971). A survey of traffic induced vibrations. Transport Road Res. Lab. Rep. 418. (TRRL, Crowthorne)

WlSS J F and PARMELEE R A. (1974). Human perception of transient vibrations. J. struct. Div. ASCE. Div ASCE, 100, ST4, April, 773-87.

YOSHIMI Y et al. (1977). Soil dynamics and its application to foundation engineering: state of the art report. In Proceedings 9th conference on soil mechanics and foundation engineering, Tok.yo, vol. 2, 605-50.

18

APPENDIX

STRESS A N D STRAIN IN AN ELASTIC SOLID Consider two particles, within an elastic solid, separated by a distance dx. Let an incident shear wave, of velocity cs (propagating in the x direction parallel to a line joining the particles) cause a distortion of amplitude dA normal to its direction of propagation.

Then the shear strain, y = d A / d x (A.1)

Now the shear stress, T= G7 (A.2)

and the shear wave velocity Cs= (G/~o) ½ (A.3)

where G is the shear modulus and ~o the rock density. The wave propagation velocity may also be expressed as the rate of change of distance (x) with respect to time (t)

That is, cs= dx/dt (A.4)

Substituting A.4 in A.1

dA y = -~-/Cs (A.5)

dA As -~- is the particle velocity, V,

y = V/cs (A.6)

Also, combining A.2, A.3 and A.6

T = Vcs~o (A.7)

Similarly, consider the effects of a compressional wave of velocity Cp which causes a dilatation between the particles of amplitude dA

Then the compressional strain, ~ = dA /dx (A.8)

Now the stress o = (~ +2G)~ (A.9)

and the compres~i~)nal wave velocity, Cp = [(X + 2G/Lo] (A. 10)

where Z and G are Lam~s parameters.

Now Cp = dx /d t (A.11 )

.'.~ = V/cp (A.12)

and o = VCpO (A.13)

Printed in the UK for HMSO by Hobbs the Printers of Southampton (2736) Dd8222661 3/86 G426

19