Downloaded from http://egsp.lyellcollection.

org/ by guest on February 11, 2019

Ground movements resulting from urban tunnelling: predictions and

effects

W. J. Rankin

ABSTRACT: The paper provides practical guidance for estimating the effects of 'soft ground' tunnelling in
urban areas upon existing structures and services. Various empirical approaches to the definition of the surface
settlement zone are summarized and the assessment of the magnitude and distribution of surface movements
is compared with case history data. A tentative risk classification related to settlement and maximum slope
criteria is proposed, which will allow rapid optimization of route adjustments and thereby identification of
those buildings particularly at risk and requiring a more detailed assessment.

Introduction all round ground movements which are manifest at the

surface in a trough form, to the sides and to the front
Tunnel construction in urban areas is likely to increase, of the advancing face, as shown in Fig. 1. This figure
in the long term, as the need for more direct and also indicates the notation adopted for both horizontal
environmentally acceptable transportation corridors and vertical movements. Two phases of movement are
expands. Such underground construction in soft ground, recognized, an immediate phase which accompanies
which in this context covers the spectrum from weak excavation and tunnel construction and a post construc-
rocks to soft clays, can cause modes and rates of defor- tion phase. The latter embraces time dependent move-
mation different from those normally experienced by ments arising from, for example, consolidation due to
structures responding to ground movements caused by pore-water pressure dissipation and creep, and a
the imposed structural loading. The strains induced b y number of other factors such as recompaction. Whereas
tunnelling are superimposed upon existing strains, and the immediate movements can be predicted with reason-
it is possible that where the existing strains in a structure able confidence, this is not the case for the subsequent
are at critical levels, which may not be apparent, the time dependent movements. The likely nature and
effect of additional strains could be out of all proportion causes of the two phases of movement are detailed in
to their magnitude. Table 1.
The form and magnitude of ground movements In the general design case it is necessary firstly, to
depends closely upon the geological setting and the forecast the extent of the zone of movements, and sec-
nature of the ground surrounding the tunnel. The spatial ondly, the magnitude of movement produced by the
extent and magnitude of ground surface movements works in that zone. Finally, it will be necessary to assess
caused by underground construction can be predicted any beneficial or detrimental modifying effects caused
with a reasonable degree of confidence for 'green field' by the presence of structures and services.
site conditions. These relatively simple predictions are In any underground construction there is also the
based upon empirical correlations derived from field possibility of catastrophic failure of the face, or gross
observations. However, the predicted near-surface losses of ground, which can occur for a variety of
movements may be modified by the presence of struc- reasons. These are characterized by singular, large, sud-
tures and services, and there is some evidence that the den and unrestrained ground movements. Case studies
pattern of sub-surface movements is dependent on the suggest that the highest potential for such losses occurs
method of tunnelling. either when unforeseen geological conditions arise such
This paper describes the available empirical methods as hnexpected interfaces between soil types, or when
for the prediction of the extent, form and magnitude contingency measures provided are inadequate: At the
of ground movements in differing soil conditions. As design stage it is not sufficient to predict routine ground
part of a wider study Lake et al. (1987) have compiled movements, it is also necessary to assess the implications
an extensive series of case history records and some and to minimize the risk of catastrophic loss of ground.
examples from this data bank are included to highlight
the variability of actual experience. A brief review of Surface trough
the damage criteria applicable to structures is included,
together with an example of how the predictions can Trough width. It has been observed that the shape of
provide a rational basis for alignment selection of tun- the surface trough, at right angles to the tunnel axis, in
nels in an urban area. the initial stages approximates to an inverted normal
Gaussian distribution curve as proposed by Peck (1969)
(Fig. 2) based on studies by Schmidt (1969). The
Ground movement prediction parameter iy is the distance of the point of inflexion
from tunnel centre line on this curve. The parameter iy
Tunnelling in soft ground is inevitably accompanied by has a mathematical relationship with the settlement at
From BELL, F. G., CULSHAW,M. G., CRIees, J. C. ~: LOVELL,M. A. (eds) 1988. Engineering Geology of Underground
Movements, Geological Society Engineering Geology Special Publication No. 5, pp. 79-92.
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

80 W.J. RANKIN

F1G. 1. Three dimensional shape of surface trough and tunnel F1G. 2. Idealized transverse surface settlement profile with nor-
co-ordinate system (after Yeates 1985). mal distibution (after O'Reilly & New 1982).

TABLE 1. Causes of Ground Displacement

Displacement Nature of Loss Cause of Loss

Phase

Initial loss of material into the face; elastic and/or plastic

deformations and runs or
flows of soil;
loss over the shield; police plates, over-cutters,
or bead;
over-excavation,
ploughing, yawing or
negotiating curves;
pushing aside boulders;
build up of grout on tailskin;
loss on, or after, erection of when soil void not
lining (at tail); completely filled;
delays in erection of
lining or in grouting;

Time loss with time as heading void collapse;

Dependent advances; lining deflection;
additional loss with time; volume decrease of soil;
consolidation of soil;
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

GROUND MOVEMENTSRESULTINGFROM URBAN TUNNELLING 81

points along the settlement profile, and can be used

conveniently as an indicator of trough width. From an
analysis of selected case history data, Glossop (1978)
and O'Reilly & New (1982, 1983) suggest that for most /
purposes an empirical equation relating the parameter /
ir to the depth to tunnel axis z o may be adopted, namely: KEY / n
9 Gronulor soils / / x
x Cohesive soils / /
iy = (1) $. Residuol s o i l s / s i l t s ond fills x/ /" J
o Mixed soils d; /~
the value of K being dependent upon the soil type and ~o-
varying from 0.4 for stiff clays to 0.7 for soft silty clays.
In granular soils, a surface settlement trough of a similar g x /o A
form is obseryed, but often with a more irregular and X . ,.0 e~'//: ho{f trough width

deeper central zone. O'Reilly & New (1982) have indi-

cated that for tunnel depths of less than 10 m in granular ~ 5,
soils, K varies from 0.2 to 0.3.
The relationship between iy and z o for an enlarged
data base for various soil conditions is shown in Fig. 3. ..//x /.~../// -

0
tronsverse distonce to point of inflexion, iy(m) o ~b io ~b
0 K) 20 half trough widthlm)

KEY Fro. 4. Case history data for trough shape.

9 Gronulor soils
io x Cohesive soils
LX Residual soils/silts ond fills
r~ n Mixed soils The centrifuge test results reported by Kimura & Mair
(1981) for clay soils, indicate that the settlement trough
width corresponds to K = 0.5 in Equation 1, which
confirms the field evidence. The results of their work
suggest that the width of the trough is generally indepen-
v dent of the degree of support within the tunnel and
3
._=
\\ x~ \\\ therefore independent of the tunnel construction
zo- \\ ~ \\ technique. A typical pattern of equivalent surface and
e\ x \\ sub-surface displacement from their model tests in clay
\ ~ rl\\ can be seen in Fig. 5.
o
\\ \'O n
9 \\ ~ ;Y89176 9 9 , ~ ",~ "s ' '~" ~ ~" ~- ~ ~ j t, l, ....
5(3- x iy=O.Szo
iy =0.4 zo
I d r o s e " 9

F~. 3. Case history for trough shape related to tunnel depth.

..... "" i._)

From inspection, it would appear that for initial practical 9 . 9 , 9 9 9 , 9 9 9 . 9 9 , 9 . . 9 9 . 9 . 9 .

9 9 9 . 9 9 9 9 9 , 9 9 . 9 . . 9 . 9 9 9 . . ,
estimation purposes the distance to the point of inflexion
of the settlement trough from the tunnel centre line can 9 9 9 . 9 , 9 , 9 . 9 r , 9 . 9 . 9 . . , 9 . 9 ,

be assumed to be approximately equal to half the depth

to the tunnel. The case history data (Fig. 4), also indicate
that the overall initial trough width to the detectable
limits of surface settlement is approximately equal to
six times the trough width parameter iy, or about three FIG. 5. Soil displacements around model tunnel in clay (after
times the depth of the tunnel. Kimura & Mair 1981).
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

82 w . J . RANKIN

For a complex of tunnels in stiff clays, a reasonable,

initial prediction of the settlement profile can be
achieved by adding the individual settlement troughs,
t~ 2 T , 2 (r sec~+(C+r) ton ~ ~,
except in the case of close parallel tunnels, station
enlargements or complex interlinking tunnels. How-
ever, this method is not valid for parallel tunnels in soft
to firm clays or granular soil, where the ground can be
weakened significantly by the passage of the first tunnel.
-~wZ;---"[C§ (,+sin+ , ] t o n 'li/#
Trough length, longitudinal profile. The typical shape
COS 1//
of the settlement profile, parallel to the tunnel axis, is
shown in Fig. 8. The first signs of movement at the
~'= 4~* "2-~ surface occur at a distance of about one or two times
T the depth of the tunnel ahead of the face; a small heave
has sometimes been noted at this extremity. The initial
displacement is normally 80% to 90% complete at a
similar distance behind the face. The equations for the
solution of the shape of the longitudinal profile assume
FIG. 6. Generation of transverse settlement profile (after Cord- a cumulative normal distribution function. Attewell &
ing & Hansmire 1975). Hurrell (1985) indicate that for most practical design
cases the assumption that the longitudinal trough length
parameter, ix, equals the transverse trough width
parameter, iy, is reasonable.
An alternative estimate of trough width can be made
u s i n g the 'mining' approach proposed by Cording &
Hansmire (1975) in which the settlement trough is Magnitude of surface movements
simplified to a triangular form, the width being related
to the angle of draw, ~, where gt =45 - @/2 (Fig. 6). Initial maximum settlement. In cohesive soils it is
This approach has been further developed by Attewell assumed, generally, that the volume of ground loss per
(1978). unit length into the tunnel, (that is, in excess of the
As illustrated in Fig. 7 various methods of estimating notional excavated volume) is equal to the volume of
trough width generate similar values for a given depth
to tunnel radius (zo/r) ratio. However, the simplified
relationship/v = 0.5 z o gives a larger trough width and
is therefore more likely to encompass the real perfor- -x ~ Longitudinal distance Longitudinal distance > +x
mance. behind tunnel face ahead of tunnel face

face position
-!~ x.O !~
Half trough width to radius ratio ~ Y

OK0
~ % V " ~
4
i
8
l
12
i
16
i
20
1

,Half trough width = rsec~§

24
i
-

tan-r
I
_ _

(x ' ' i w" 0"841 w max)

~ ' Maximumcurvature (hoggingl
dZw
~ =
wmox
o.z4z 9
Ill
*Maximum horizontal strain
/
"%.">. o-,,~'- Maximum curvature sagging (tensile)
%)<.. o,o d2w Wmox d_uu = O. 242 Wmox
oi~ onO .on,.,., ,.,., ~ i = 0.242 ix= dx zo
o
I0- 9Maximum longitudinal strain Point of inflexion (x=O, w=O'Swmox)
(compressive) dw Wmax
d_uu , 0.242 Wmax Maximum slope ~ - 9 0.399 ix
dx zo *Maximum horizontal displacement
o 15- w max
u -0'399
2 Half trough w_idth = 3;y "------ " ~ \\ ~ _ zo
_. i_y.,Zo,~ N ~ x~-3iY ",'2zo
20- ( Schmidt, 1969)
\ \ ~".",2'
' x
Direction of drive
-~ i i ~ - ~--
> *Horizontal displacements and
strains assume radial
ground movements towards
Tunnel axis level z o tunnel face

FIG. 7. Comparison of various methods of assessing half FIG. 8. Idealized longitudinal surface settlement profile.
trough width.
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

GROUND MOVEMENTS RESULTING FROM URBAN TUNNELLING 83

the settlement trough per unit length. This may not be the best fit dashed lines representing typical losses of
valid for granular soils or stiff fissured clays subject to 1% and 2% which are similar to the predicted line
dilation (bulking). derived using Equation 3. However, in practice very
The volume of the transverse settlement trough per much greater values of both volume loss and settlement
unit length ( V ) is obtained by integration of the area can, and do, at times occur. It is unfortunate that these
bounded by the transverse trough, and is approximately more interesting extreme examples do not reach the
given by: technical literature as it is likely that they could provide
a basis for considerably more insight into the real ground
V ~ - 2.5iy. Wmax (2) behaviour. In many instances such examples have
become subject to lengthy litigation, and the results
where win,x is the maximum settlement above the centre often remain unavailable.
line of the tunnel.
However, it is more convenient to express the Initial s e t t l e m e n t profiles. The assumption of an
maximum settlement in terms of percentage volume inverted Gaussian distribution curve for the transverse
loss ( V 0, such that: settlement profile is convenient because the settlement
(w) at any point from the tunnel axis (y) can be calcu-
Wmax= 0.0125 V~.r2/is (3) lated from the following expression:-

Typical values for volume loss for a range of soil types w =Wm. x exp(--y2/2r (4)
and tunnelling methods, based upon U K experience are
given in Table 2. With experience, and using this table Thus, with the knowledge of, firstly, the trough width
as a guide, a volume loss appropriate to the soil condi- parameter, iy, which can be related to tunnel depth and,
tions, position of the water table and tunnelling method secondly, the value of maximum settlement, which can
can be selected and potential variations can be readily be estimated for a likely range of volume loss, it is
assessed. Values of initial maximum settlement for the possible to predict the complete shape of the settlement
data base case histories assembled by Lake et al. (1987) profile. A dimensionless representation of the form of
are shown in Fig. 9, settlements being generally in the the transverse settlement trough is given in Fig. 10 which
range of 10 to 80 mm. A n indication of the relationship indicates that at the point of inflexion, iy, the settlement
of maximum settlement with volume loss is shown by is approximately 60% of that above the centre line.

TABLE2. Summarised settlement trough data for a range of soils provided C/D> I, after O'Reilly
& New (1982) and Yeates (1985)

Ground Tunnelling Trough width Volume loss Remarks

Conditions method parameter Vl(%)
constant, K (at surface)

Stiff Shield or none 0.4-0.5 1/2- 3 Considerable data

fissured clay available; losses
normally 1-2%
Glacial Shield in free air 0.5 - 0.6 2 - 21/2
deposits Shield in 1 - 11/4 Compressed air
compressed air to assist control of
ground movements
Recent silty Shield in 0.6 -0.7 2-10
clay deposits compressed air
(cu= 10- 40 kN/m 2)
Granular 0.2-0.3 1-5
material above
the water table
Granular Compressed air/ 1-10
material below slurry/earth
the water table pressure balance
(EPB)
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

84 w . J . RANKIN

maximum settlement Wma x (mm) ~/;y

0 I0 20 50 40 50 60 70 80 0 2o
Ot i i i l i i ~ J 0
I x 0-2
x. . . . _ - - ~ 0"2
0.7 0~3
9 "-"
. / o . -4
; U ~ v_t~.~.~70
~nV I I P ' I
I / ZXl.4..---";~e ZX ~ --"~2.3 ~'6
0-4 /q
0"4X 2'1 AH
2/ ZX ~i'3 Q /
eex~ ~9 4.3 9 O~ /
/
3/ o o o ~9 / ~l~ / ~l~ O~l ~ ~0~ 7X0
0"8
]~ 0 0 l4 / " ''/)l~ X ' "~
| ~ o I, / // ,~<7 I'0- shape of settlement trough

4 I ~IB []

FIG. 10. Normal distribution function for transverse settle-

[]
ment trough.

the face of the tunnel on the longitudinal profile (pos-

s ition x = O on Fig. 8) is some 30 to 50% of the value
KEY
of the maximum settlement developed in the transverse
9 Granular soils
x Cohesive soils
trough at a distance behind the face of one to two times
zx Residual soils/silts and f i l l s the depth.
o Mixed soils
1.9 Volume loss Horizontal movements. During the initial phase of
-- P r e d i c t e d if w m a x " 0 ' 0 2 5 V I rS/zo development of the surface profile with the passage of
io-I ---Best fit to case history data the face of the tunnel, settlement is accompanied by
lateral movement, which at the surface is directed
towards the centre of the trough. An idealized relation-
ship between settlement and lateral movements and lat-
Fro. 9. Case history data for maximum settlement. eral strains for the transverse trough is shown on Fig.
11. The maximum lateral movements are developed
near the point of inflexion (iv) on the settlement profile
and on the basis of the field studies presented by Cording
Similar expressions are available for the longitudinal & Hansmire (1975) and Attewell (1978) the ratio of
profile based on the cumulative distribution. Attewell maximum lateral movement to maximum settlement
& Hurrell (1985) have indicated that case history data (%JWm,~) is commonly between 0.25 and 0.40.
tend to support the observation made by Craig (1975) The movements that are potentially the most damag-
that in firm to stiff clays the ratio of settlement above ing are the tensile strain and the hogging curvature,

limb of trough half trough width, T

< . . ~ Vma x ~
where vertical and hor,zontal -] |
displacements are approximatell~
equal. ~ "-'r-- ~ .....-

[~_.z ~ ~
TflJJ- ,Z ..otteos,,.s,7.i .--
3i~-------"~---- __ j i y ~ ~zone of compr'sssivestrain ~ -~iv_..~----~L---'-~---~
--~
,~ 0 2 - " -""/~..~1 iy ~ icompressive / / i y ~.--""" "

gE o.6
~0.8. typical
. horizontal
. . .strain profile ~ ~ .i- / ~ typical normal distribution form of
1.0- ~ " " transverse surface settlement profile
WinGx ~

FIG. 11. Idealization of transverse surface displacement and

strains shortly after tunnelling.
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

GROUND MOVEMENTS RESULTING FROM URBAN TUNNELLING 85

both of which reach a maximum at a point about X/--3-.iy

from the centre of the trough. The maximum compres-
sive strain occurs over the tunnel centre line and is two
or three times the magnitude of the maximum tensile ,.o. :~,
strain. \'~i', o
0'8- ~/'~ * ~Wmax/Wc = I-0"0527 (overconsolidated
Towards the edges of the transverse trough the mag- ; 'i\",~J.. r clay)
nitude of lateral movement may be similar to, or even ' i ~ "~e! j~Craig(1975)
0-6- ~,~,.. ~.~- ~,\
greater than, the vertical settlement, and this factor is Wmax
of considerable significance in assessing risk to damage o.+
.\ -~\.~ x 9
of structures or services. Discernible lateral movements "--.....~...."--..... , ./Clough and Schmidt (le$O
have been recorded beyond the limit of discernible ver- 0.2- \ '"-~.."~-'~.~7 . . . . ~ _ r \
',,o .... c====~-Zo',~ ~co,di.g
tical movements at the edges of the trough. 0 ,\ , , , C.Zo=0 j e t ol
5 I0 15 20
I-0"1670 z~
Sub-surface movements W mox/Wc= r r
(Atkinson ond Ports sond)
Reliable observations of ground movements at depth,
KEY
particularly close to the tunnel, are difficult to achieve Cloy
and are sensitive to tunnelling construction processes. ground level ~Wmax
[] Silt
The lack of reliable, high quality data makes even ~ I z zx Sand
generalized, empirical predictions of sub-surface move- o o Gravel
ments somewhat unreliable. A number of authors, 9 Chalk
(Ward & Pender 1981, Clough & Schmidt 1981, Cording
et al. 1976) have attempted to relate maximum surface
settlement (Wmax) to the maximum settlement at the
crown of the tunnel (w~). These approaches are com-
pared in Fig. 12, the chain dotted lines being drawn Fit. 12. Comparison of empirical relationships between
using the assumption iy = 0.5Zo. The lower dashed line, crown settlement, surface settlement, depth and radius of
which forms a linear limit to the data points, is derived tunnel with case histories (after Ward & Pender 1981).
from model data by Atkinson & Potts (1977) and rep-
resents plane deformation of sand. The upper dashed
line relates to model test results using drained overcon-
solidated kaolin, and to case records for overconsoli-
dated London Clay. For overconsolidated conditions in data and generally simplified relationships. Provided
clay Atkinson & Potts (1977) suggest that the ratio of judgement is exercised and appropriate allowances are
w,,~,/w~ does not fall below 0.40 for values of Zo/~greater made for variations in workmanship, construction
than 10. method, soil type and other practical considerations,
Although the well established correlation between they are reasonably reliable.
surface settlement profiles and the inverted normal dis- Finite and boundary element analyses offer the pos-
tribution curve is apparently independent of the soil sibility of calculating ground deformation and the
type and tunnelling method, this characteristic surface interactive effects on structures and services. However,
shape does not reveal any of the considerable differ- the application of these techniques to predictions of
ences in pattern of sub-surface displacements that occur deformation has not yet reached the stage where they
in reality (Hansmire & Cording 1985). In practice, the can be applied generally.
magnitude and direction of sub-surface movements Finite element modelling using an elastic soil model
appear to be closely related to the tunnelling method can be used in conditions where elastic behaviour is
and soil type and are likely to be linked to the stress approached but the number of situations where this
history of the soil. A summary of observed data from applies are likely to be few and probably restricted to
reported case histories is given in Table 3, which indi- stiff clays. However, most soils exhibit distinctly non-
cates the variable response and highlights the lack of linear mechanical properties and are anisotropic requir-
data. For instance, the case history reported by Clough ing more complex elastic and elasto-plastic models for
et al. (1983) for an earth pressure balance machine used analysis. Further, strain softening and plastic flow,
in San Francisco Bay, Mud shows initial movements of which occur in very stiff overconsolidated clays and very
soil away from the ~'hield followed by inwards move- soft or greatly overstressed cohesive soils respectively,
ments after passage of the shield. are not readily modelled.
Predictions are very sensitive to the assumed variation
in the soil stiffness with depth and to the thickness of
Analytical techniques compressible strata assumed to lie beneath the tunnel.
Thus, contrary to experience, computations with seem-
The previously described methods of prediction are ingly reasonable values of these parameters can gener-
essentially empirical methods based on observational ate misleading results. For instance, they may indicate
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

86 w.J. RANKIN

TABLE 3. Observed data on sub-surface horizontal m o v e m e n t s at tunnel axis level.

Tunnel Horizontal Distance from Proportion of

Case History diameter Depth Soiltype cu movement side of tunnel distance to
D (m) zo (m) (kPa) (mm) (m) tunnel diameter

Stiff Clays
Barratt and Tyler (1976) 4.15 20 London Clay 200 5 1.0 0.2
London 32 London Clay 200 3-7 1.0 0.2
Anon. (1972) London 4.15 30 London Clay 200 7 1.72 0.4
6.5 2.17 0.4
Eisenstein etal. (1981) 2.56 24 Glacial Till 400 25" 2.0 0.8
Edmonton, Canada
Ward and Thomas (1965) 3.85 25 London Clay 5 3.0 0.8
London
Ward (1970) 4.15 24.4 London Clay I1 0.5 0.1
London 7.5 2.0 0.5
2.5 3.5 0.8

Soft Clays
Clough et al. (1983) 3.7 9.25 Soft silty clay 25 0.3 3.0 0.40
San Francisco 36 0.5 0.14
Giossop and O'Reilly 3 6 Soft silty clay 12 20-25** 3.0 1.0
(1982) Grimsby
GIossop etal. (1979) 2.7 4.5 Soft alluvial clay 10 5 3.0 1.0
Belfast

* Results unusually high.

** Assumes ratio of horizontal to vertical is in range 0.3 to 0.4.

general heave resulting from the stress relief caused by

tunnel excavation. For soft soils, the trough shape pre- type of movement;
dicted using an elastic analysis is generally wider than race of movement;
that found in practice. Applied to cohesive soils, two- magnitude and distribution of movement;
dimensional elasto-plastic analysis can give a more type, construction and condition of structure;
realistic trough shape, though rather narrower than has interactive soil/structure effects (either relieving or
been observed, (Rowe et al. 1983). concentrating effects).
The more sophisticated methods of analysis can be
useful in parametric studies where the sensitivity to the Recognizing that ground movements are an inevitable
likely variations in the different variables can be asses- consequence of underground construction, the allow-
sed and compared with case records. However, little able movements within existing structures can be consi-
progress can be made without both realistic and repre- dered under the following headings:
sentative ground parameters. At the present time, pre-
dictions made by analytical methods do not provide any safety;
greater accuracy or dependability than the simpler architectural or aesthetic damage;
empirical methods. functional damage;
structural damage;
prevention or repair.
Effects on existing structures
The assessment of some of these factors can be subjec-
Factors that have to be considered when assessing the tive, depending, for example, upon personal percep-
allowable additional movement of existing structures tion, the geographical/geological setting, or the use and
resulting from tunnelling include: status of the structure.
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

GROUND MOVEMENTS RESULTING FROM URBAN TUNNELLING 87

The judgement of likely effects and the protective maximum allowable total foundation settlements for a
measures to be adopted, depends upon the level of range of structures (Skempton & MacDonald 1956, Pol-
confidence in the predictions and the consequences if shin & Tokar 1957 and Wilun & Starzewski 1972). In
they are wrong. There may be considerable uncertainty, all cases it is necessary to take account of project specific
which cannot be readily clarified in advance of construc- details, including type of structure, type or sensitivity
tion, with regard to the response of some buildings or of contained machinery and the actual ground condi-
utilities to ground movements. In these circumstances tions. These deformation criteria are not detailed in this
it may be necessary to adopt a conservative approach. paper but a full comparative study is given by Lake et
al. (1987). A combination of these criteria forms a useful
Damage criteria for structures basis from which to develop a tentative framework for
relating settlement and maximum ground slopes to
C r a c k width. A comprehensive classification for relating
potential structural damage and hence risk assessment
the width of visible cracks to the damage to non-struc- for tunnelling projects in urban areas.
tural elements with respect to ease of repair was
developed by the Institution of Structural Engineers
(1978). A summary of this is given in Table 4. Risk assessment
The assessment of aesthetic damage can be a highly
subjective matter, but there may be situations where Initial appraisal
relatively slight damage is intolerable either to the user/
owner or for the proper functioning of the structure. In the development and design of a tunnelling project
in an urban area it is necessary to assess the likely dam-
D e f o r m a t i o n s . A number of authors have used field age that would be caused to existing and sometimes
observations to correlate structural deformation with planned structures and services. A simple means of
tangible damage, (Skempton & MacDonald 1956, PoP assessment is required in the initial design stage when
shin & Tokar 1957, Bjerrum 1963, Burland & Wroth it may be possible to alter planned vertical alignments
1974 and O ' R o u k e et al. 1976). O ' R o u k e et al. (1976) to minimize the risk to structures. At this stage it is
relate damage to deformations caused by rapid move- convenient to assume that the near-surface foundations
ments, as opposed to self weight movements considered of structures follow the slope of the settlement trough,
by the other authors. The description of deformations appreciating that this makes no allowance for interactive
of structures in terms of relative rotation or angular soil-structure restraints. This simplification permits the
distortion is assisted by the use of an unambiguous and assessment of the critical deformation parameters based
consistent set of definitions. Those provided by Burland on the slope of the surface trough and the magnitude
& Wroth (1974) are widely accepted and recommended. of maximum settlement within it. Whilst overall uniform
In order to adequately quantify t h e distortional settlement may not damage structures, actual settle-
parameters for a particular structure, a large number ments vary from the predicted value due to inherent
of observation points together with detailed information variations in ground properties and workmanship. This
on the foundations and superstructure are required. irregular settlement also causes structure deformation
However, in practice such information is seldom avail- and may be assumed to be proportional to the predicted
able and interpretation may be difficult. Similarly, a magnitude of settlement at any given point. Typical
number of authors have suggested limits for the critical values of building slope and settlement that have

TABLE4. Classification o f visible damage to walls

Category of damage Degree of damage Approximate crack width

(mm)

Aesthetic 0. Negligible :~0.1

1. Very slight :~1.0
2. Slight <5.0
Functional 3. Moderate 5-15, or a number of cracks
>3 mm
Serviceability 4. Severe 15-25, depends upon number of
cracks
Structural 5. Very severe usually >25 mm but depends upon
number of cracks
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

88 W.J. RANKIN

been used for planning and design purposes, as well as Detailed a s s e s s m e n t

optimizing alignments, are presented in Table 5.
There are, as yet, insufficient case records to confirm Using the same risk classification for near-surface build-
the general validity of these values and therefore they ings or pipelines a suggested course of action is given
can only be considered as a guide to be applied with in Table 6.
experience and judgement. The above criteria, or the Particular structures or services which are identified
derived zones of risks, are readily marked up upon a as being in risk category 3 or 4, and in some cir-
plan of the tunnel route in tramline fashion. For a simple cumstances even category 2, should be examined in
tunnel the predicted width of the trough for the critical detail. This will require adequate knowledge of the soil
values of settlement and slope can be calculated from and groundwater conditions, access to the 'as-built'
the mathematical relationship for the assumed Gaussian drawings, knowledge of the structural loading and
distribution. The trough width for a given settlement details of the major elements of the structure. It may
has an algebraic solution but the width for a value of require an assessment of the likely deformations already
slope has to be solved iteratively. These values can be experienced by the structure and hence its tolerance to
computed for various tunnel depths and repeated for further movements. From this basis predictions of the
different values of volume loss in order to test the sen- likely structural damage and potential risk can be made
sitivity of the zoning to the assumed value. for the range of ground movements anticipated as a
An example of risk zoning for a 4 m internal diameter result of tunnel construction.
sewer tunnel to be excavated within silts and sands
below the water table is given in Fig. 13. Having iden- P r e v e n t i o n or repair
tified the limits for the various risk categories and assum-
ing the structures straddling zones, or connected to At the planning stage it is possible to assess the
other structures, to be within the more critical zone, it economics of various options such as purchase or reloc-
may be possible to adjust the tunnel route or depth to ation of the property/service and weigh these against
minimize the amount of damage, reduce the number of either positive, strengthening or supporting, protective
buildings affected or avoid sensitive structures. For measures carried out in advance of construction or
structures identified as being at risk once the tunnel allowances for costs of repair at a later date. A diverse
layout has been optimized a course of action for more range of underground protective solutions is available
detailed appraisal is required.

TABLE5. Typical values of maximum building slope and settlement for damage risk assessment.

Risk Maximum Slope Maximum Settlement Description of Risk

Category of Building of Building (mm)

Less than %oo Less than 10 Negligible: superficial

damage unlikely.
1/500to a/zoo 10 to 50 Slight: Possible superficial
damage which is unlikely
to have structural
significance.
1/2ooto 1/% 50 to 75 Moderate: Expected
superficial damage and
possible structural damage
to buildings, possible
damage to relatively rigid
pipelines
Greater than 1Vs0 Greater than 75 High: Expected structural
damage to buildings.
Expected damage to rigid
pipelines, possible damage
to other pipelines.
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

GROUND MOVEMENTS RESULTING FROM URBAN TUNNELLING 89

~- ~
'- ::.-. o

/
m wm am ;,~
.~E _~
"G E
:)
\ ~G o~
| ~] ~
-J
o. m

~o
~ ~.- ,,. ~ E~ EE
3~

f
| -o DI
~E
o
~ 1 i"1 ~1 ; ~ ,,, I x l i . . ' , ~,,.~
"1 I' '~' ~ ' ' ' ' ' I'~'' ' t, '1 .x ,
m

- ~ ~.

I=~,,
~ /~
|
|

~l--r"
--qL
I1
@
~~ ~
| -~ __

|
"i _g

:72~ ~ ~:;~ ,
. Ii ..=
..I.-

ii I i', I-- r

~D

E
V x
o

,,,,,,,,i
E E
~2
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

90 w . J . RANKIN

TABLE 6. Action suggested for various risk categories.

Risk Description Description of Action

Category of Risk Required

1. Negligible: Superficial No action,

damage unlikely, except for any buildings
identified as particularly sensitive
for which an individual assessment
should be made.
Slight: Possible superficial Crack survey and schedule of
damage which is unlikely to have defects, so that any resulting
structural significance. damage can be fairly assessed and
compensated.
Identify any buildings and
pipelines that may be particularly
vulnerable to structural damage
and assess separately.
Moderate: Expected superficial 1 Crack survey, a schedule of
damage and possible structural defects, and a structural
damage to buildings. Possible assessment.
damage to relatively rigid
Predict extent of structural
pipelines.
damage, assess safety risk,
choose whether to accept damage
and repair, take precautions to
control damage or, in extreme
cases, demolish.
Buried pipelines at risk: identify
vulnerable services, and decide
whether to repair, replace with a
type less likely to suffer damage,
or divert.
High: Expected structural
damage to buildings.
Expected damage to rigid
pipelines. Possible damage to
other pipelines.

Note: The above table relates to near surface foundations or pipelines.

including underpinning, grouting, diaphragm or piled excavation and support appropriate to the ground con-
walls, and sheet piling. However the designer should ditions and the quality of workmanship are of
be aware that most of these methods will not prevent paramount importance in attaining the minimum
movements of the surrounding soil, and that some pro- ground displacement.
cesses can induce movements of a magnitude compara-
ble to those anticipated by tunnelling. Additionally, Monitoring
measures designed to avoid settlement may not afford
protection against lateral movements. The combination of observations and records of con-
The magnitude of ground movements resulting from struction procedure, together with measurements of
tunnelling is influenced by the method of construction movement of the ground and the nearby structures and
adopted and associated factors such as the rate of services, can, in appropriate circumstances, provide a
advance or pauses, the support system, the watertight- valuable means of assessing performance. The need for
ness of lining and effectiveness of grouting. Experience location and selection of instrumentation for monitoring
indicates that the correct application of a method of movements requires considerable understanding and
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

GROUND MOVEMENTS RESULTING FROM URBAN TUNNELLING 91

judgement with a rigorous assessment of purpose, Road Research Laboratory Report PWA/IWF/TRRL/I972/
together with a well conceived plan for management of 1, prepared by the University of Durham, Engineering Geol-
the data. Strategies are necessary for implementing ogy Laboratory.
appropriate contingency measures if adverse conditions A~,~INSON,J. H. & Porrs, D. M. 1977. Subsidence above shal-
low tunnels in soft ground. Journal o f the Geotechnical
are encountered.
Engineering Division, American Society of Civil Engineers,
There is a growing body of opinion in the United 103, GT4, 307-25.
States that measurement of sub-surface displacement ATrEWELL, P. B. 1978. Ground movements caused by tunnel-
should be mandatory for monitoring of tunnel construc- ling in soil. In: GEDDES, J. D. fed.) Large Ground Move-
tion. Certainly there is a need for well-documented com- ments and Structures, Pentech Press, Plymouth, 812-948.
prehensive records of both ground and structure/service - -& HURRELL,M. R. 1985. Settlement development caused
displacements/deformations and strains coupled with by tunnelling in soil. Ground Engineering, 18, 8, 17-20.
information on the levels and consequences of distress BARRATT,D. A. & TYLER,R. G. 1976. Measurements ofground
or damage if prediction is to improve. movements and lining behaviour on the London Under-
ground at Regents Park. Transport and Road Research Lab-
oratory Report 684.
BJERRUM, L. 1963. Evaluation of allowable settlements. In:
Conclusions Proceedings o f the 3rd European Conference on Soil
Mechanics and Foundation Engineering, Wiesbaden, 2,135-
A review of extensive case history data indicates that 7
the use of simple empirical relationships to predict the BURLAND, J. P. • WROTH, C. P. 1974. Settlement of buildings
magnitude and nature of ground movements resulting and associated damage. In: Proceedings of a Conference on
the Settlement o f Structures, Cambridge, 611-54 and 764-
from tunnel construction can provide a rational basis
810.
for the assessment of risk and damage to nearby struc- CLOtJGH, G. W. & SCHMIDT,B. 1981. Design and performance
tures and services. However, the possibility of catas- of excavations and tunnels in soft clay. In: BRAND, E. W.
trophic failure of the ground has to be separately asses- d~ BRENNER, R. P., (eds), Soft Clay Engineering. Elsevier
sed. Scientific Publication Company, Amsterdam, Chapter 8.
The case history data suggest that the overall trans- , SWEENY, B. P. & FINNO, R. J. 1983. Measured soil
verse settlement trough width to the detectable limits response to EPB shield tunnelling. Journal o f Geotechnical
of surface settlement is approximately equal to three Engineering, American Society of Civil Engineers 109, 2,
times the depth of the tunnel. 131-49.
CORDING, E. J. & HANSMmE, W. H. 1975. Displacements
The data also indicate that, for cases where percen-
around soft ground tunnels. In: Proceedings of the 5th
tage volume loss per unit length is relatively small, there PanAmerican Conference on Soil Mechanics and Founda-
is a reasonably consistent relationship between this tion Engineering, Buenos Aires, 4, 571-633.
parameter and the maximum settlement and the radius - - , MACPHERSON,H. H., LENZINI,P. A. & VONDEROHE,
and depth of the tunnel. However, published data may A '9 D. 1976. Displacement around tunnels in soil. Final
not adequately reflect the ground behaviour in situa- report by the University of Illinois to the Department of
tions where greater ground losses could occur. Transportation, Washington D.C. (NTIS, PB, 267356).
Analytical methods at present have limited general CRAIG, R. 1975. Discussion on Instrumentation and Monitoring
application and are unlikely to provide more reliable (Tunnels and Tunnelling 1975, 7, 5), Tunnels and Tunnel-
ling, 7, 6, 61-5.
results than empirical methods, but can be useful for
EISENSTE1N, Z., EL-NAHHAS, F. & THOMPSON, S. 1981. Strain
parametric studies in some situations. field around a tunnel in stiff soil. In: Proceedings of the 10th
The tentative risk classifications proposed have been International Conference on Soil Mechanics and Foundation
used to optimize tunnel alignments during the design Engineering, Stockholm, l, 283-8.
of recent tunnelling projects in major urban areas. It is GLOSSOP, N. H. 1978. Soil deformation caused by soft ground
hoped that observations of performance from these and tunnelling. PhD Thesis. University of Durham.
other projects will permit further refinement of this --, SAVILLE,D. R., MOORE, J. S., BENSON,A. P. & FARMER,
assessment. I. W. 1979. Geotechnical aspects of shallow tunnel construc-
tion in Belfast estuarine deposits. In: Proceedings of Tunnel-
ling 79, Institution of Mining and Metallurgy, London, 45-
ACKNOWLEDGEMENTS:The author is indebted to the Construc- 50.
tion Industry Research and Information Association, as much -- & O'REILLY, M. P. 1982. Settlement caused by tunnelling
of the information presented is drawn from a research project through soft marine silty clay. Tunnels and Tunnelling, 14,
undertaken on their behalf by Mott, Hay & Anderson. The 9, 13-16.
author gratefully acknowledges the major contribution to the HANSMIRE, W. H. & CORDING, E. J. 1985. Soil tunnel test
research project by L. M. Lake and J. Hawley and the guidance section: case history summary. Journal of Geotechnical
provided by R. E. Williams. Engineering, American Society of Civil Engineers, 111, 11,
1301-22.
THE INSTITUTIONOF STRtJCa'URALENGINEERS1978. State of the
art report. Structure - Soil Interaction. The Institution of
References Structural Engineers, London.
K1MURA, T. & MAre, R. J. 1981. Centrifugal testing of model
ANON 1972. Determination and interpretation o f surface and
tunnels in soft clay. In: Proceedings of the lOth International
subsurface ground movements during construction of a tun-
Conference on Soil Mechanics and Foundation Engineering,
nel in London clay at Green Park, London. Transport and
Stockholm, 1,319-22.
 Downloaded from http://egsp.lyellcollection.org/ by guest on February 11, 2019

92 w . J . RANKIN

LAKE, L. M., RANKIN,W. J. & HAWLEY,J. 1987. Assessment associated with tunnelling in soil. PhD Thesis, University of
o f effects o f underground construction upon structures and Illinois, Urbana.
services. CIRIA Report No. RP316. CIRIA, London. SKEMPTON, A. W. & MACDONALD,D. H. 1956. The allowable
O'REILLY, M. P. & NEW, B. M. 1982. Settlements above tunnels settlements of buildings. Proceedings of the Institution of
in the United Kingdom, their magnitude and prediction. In: Civil Engineers, 5, 2, (III), 727-68.
Proceedings o f Tunnelling '82, Brighton, 173-81. WARD, W. H. 1970. Discussion on the paper by FOLLENFANT
- - & -- 1983. Report of discussion on Tunnelling '82. et al. (No. 727OS, Proceedings of the Institution of Civil
Transactions o f the Institution of Mining and Metallurgy, Engineers Supplement 1969). Proceedings of the Institution
92, A35-A48. of Civil Engineering Supplement 1970.
O'ROUKE, T. D., CORDING, E. J. & BOSCARDrN,M. 1976. The - - & THOMAS,H. S. H. 1965. Development of earth loading
ground movements related to braced excavation and their and deformation in tunnel linings in London Clay. In: Pro-
influence on adjacent buildings. United States Department ceedings of the 6th International Conference on Soil
of Transportation, Report No. P.B. 267311. Mechanics and Foundation Engineering, Toronto, 2, 432-
PECK, R. B. 1969. Deep excavations and tunnelling in soft 36.
ground. In: Proceedings o f the 7th International Conference - - & PENDER, M. J. 1981. Tunnelling in soft ground - -
on Soft Mechanics and Foundation Engineering, Mexico General Report. In: Proceedings of the lOth International
City, 3, 225-90. Conference on Soil Mechanics and Foundation Engineering,
PoLsmN, D. E. & TO~:AR, R. A. 1957. Maximum allowable Stockholm, 4, 261-76.
non-uniform settlement of structures. In: Proceedings of the WmuN, Z. & STARZEWSK~,K. 1972. Soil Mechanics in Founda-
4th International Conference on Soil Mechanics and Foun- tion Engineering, 2. Theory and Practice. Intertext, London.
dation Engineering, London; 1,402-5. YEATES, J. 1985. The response of buried pipelines to ground
RowE, R. K., Lo, K. Y. & KACK, G. J. 1983. A method of movements caused by tunnelling in soil. In: GEDDES, J. D.
estimating surface settlement above tunnels constructed in fed.) Ground Movements and Structures. Pentech Press,
soft ground. Canadian Geotechnical Journal, 20, 1, 11-22. Plymouth, 145-60.
SCHMtDV, B. 1969. Settlement and ground movements

W. J. RANKIN, Mott, Hay & Anderson, St Anne House, 20/26 Wellesley Road, Croydon CR9 2UL, UK.