Response of Heritage Buildings to Excavation-Induced Vibrations

T.G. Carter
Golder Associates, Toronto, Canada.

G. Pernica
National Research Council, Institute for Research in Construction, Ottawa, Canada

M.V. van Bers

Golder-VME Limited, Toronto, Canada

ABSTRACT: Little data is available in the literature on the response of heritage masonry structures to
excavation-induced vibrations. This paper presents unique data and observations relating to deep rock
excavations performed on Parliament Hill, Ottawa, Canada, during the summer of 1997. Some of the
excavations were carried out directly beneath the 80-year old Centre Block building and some were
undertaken immediately adjacent to the 150-year old Library of Parliament. The works were part of an overall
scheme of renovations to renew and upgrade existing mechanical, electrical and telecommunication services
within the Parliamentary complex as part of the construction of a new Underground Services Building. In this
paper, data from the extensive array of monitoring instrumentation is evaluated as a basis for determining
building vibration response characteristics and attenuation rates (decoupling) at the rock-foundation interface.

1 INTRODUCTION

The Parliament Buildings of Canada are situated on

Parliament Hill in the City of Ottawa overlooking
the Rideau River. The original buildings, consisting
of the West, East and Centre Blocks, the latter with
an adjoining Library, were erected between 1859
and 1868. In 1916 the original Centre Block,
housing the main Parliamentary Chambers, was
destroyed by fire although the attached Library was
saved. As part of a major scheme to removate
utilities and services to the old Parliamentary
Buildings, construction of new excavations
adjacent to and below parts of the Centre Block
was completed during the 1997 summer
Parliamentary Recess.

Mechanical excavation methods, including use of a

TBM, were employed beneath the Centre Block
Building (Carter et al. 1999). Outside, blasting was
carried out to remove the majority of the 20,000 m3
of rock excavation required adjacent to the 150
year old Library of Parliament (Figure 1). All of
this work was completed without damage to the
existing heritage buildings or disruption to the Figure 1: Excavation underway adjacent to
proceedings of Parliament. the Library and Centre Block Towers

Carter et al - Response of Heritage Buildings 1

1.1 Description of Library grouted rubble core and an inner face of coursed
random rubble. The inner ring wall is approximately
The Library of Parliament comprises a circular 1500 mm thick throughout its height while the outer
masonry loadbearing wall structure consisting of a ring wall varies from 1200 mm near its base to about
stone masonry inner ring wall, a stone masonry 850 mm near its top. The condition of the masonry
outer ring wall, and 16 stone masonry buttresses walls and buttresses are largely unknown. However,
which are connected to the inner ring wall by stone exposed portions of the walls and buttresses show
masonry flying buttresses, (Figure 1). The 16 main signs of deterioration with mortar missing from
1800 mm thick stone buttresses, which form part of joints between stones and vertical cracks evident in
the outer ring wall (Figure 2), extend 2500 mm out some of the buttresses.
from the exterior face of the exterior ring wall.
1.2 Description of Centre Block

The Centre Block, a 6-storey building built between

1916 and 1927, is a multi-element structure. The
perimeter of the structure rests on concrete
foundation walls about 1 m thick. These walls were
poured on concrete foundations placed directly on
bedrock or where the depth to bedrock was
considerable (more than a metre or so) on a coursed
ungrouted masonry wall which filled these localized
depressions in the rock surface.

The superstructure of the building comprises a

variety of structural types. Portions of the interior
superstructure are reinforced concrete frames, steel
frames or a combination of frames and loadbearing
walls (masonry or concrete). The floors are
reinforced concrete slabs supported on walls,
Figure 2: Outline of CBUS blast zones and columns or beams.
mechanical excavation area with relation to
Library and Centre Block. The exterior loadbearing walls of the building have
an exterior faced wythe of Nepean sandstone. Much
Freestanding stone pinnacles are sited above the of the wall along the north side of the building in the
level of the flying buttresses. These pinnacles add area closest to the required excavations for the
another 7.3 m to the overall height of the exterior Centre Block Underground Services Building
buttresses. The outer ring wall, which ends at (CBUS), was in relatively good condition with most
about the midheight of the inner wall, encloses of the mortar still in place. At some locations, such
several stories of office and storage space around as the 2-storey stairwell at the north-west corner of
the outer perimeter of the Library. Structurally, the the building, conditions were not as good, with
inner ring wall assists the outer ring wall and vertical cracks running the full height of the stairwell
buttresses in supporting a network of cast iron within the loadbearing masonry walls; however, the
trusses and a ring beam to carry the dome portion floors, topped with terrazzo, showed few cracks.
of the Library.
2 EXCAVATION APPROACH
The walls and buttresses of the Library are of
random rubble construction, consisting of an Two main types of excavation method were
exterior wythe of faced Nepean sandstone, (a high envisaged to be used for rock removal exterior to the
quality local building stone), within which is a Centre Block - blasting and mechanical methods.
Carter et al - Response of Heritage Buildings 2
2.1 Blasting Approach North-West Tower of the Centre Block (about 50 m
high) and for the Library, the pinnacles at the top of
Prior to undertaking the actual excavation, a series the buttresses facing the footprint and the
of test blasts were carried out to determine the ornamental chimney at the top of the inner ring wall,
response of the two buildings to blasts occurring (ref. photograph included as Figure 1).
within the CBUS footprint, (Subercaseaux et al.,
1999, Curic et al 1999). The test blasts were 2.2 Blast Design Methodology
devised to determine propagation rates within the
limestone bedrock and structural responses at The blast methodology and contract vibration limit
various levels and locations within the buildings. guidelines used throughout the CBUS excavation
works in 1997 were based on comprehensive
Throughout, both in designing the excavation analysis of the test blasts and comparison of their
sequencing, and in laying out the required vibration response with data from three-dimensional
monitoring for the construction activities, attention structural analysis of the behaviour of the buildings.
was focused on ensuring minimal damage to the As detailed in Subercaseaux et al., 1999, the key
existing buildings. Accordingly, arrays of vibration recommendations arising from the analysis of the
sensors, consisting of velocity transducers and trial blast results were:
accelerometers, were installed at strategically • Blasthole diameter to not exceed 50 mm.
selected locations on the exterior and interior of the [Limiting blasthole size was deemed the most
two building structures. Gauges were also placed effective way of controlling pattern size and
across existing masonry cracks and exterior mortar explosive loads.]
joints to monitor changes in the width of these • Only cartridged explosives and non-electric
elements. Monitoring of ground behaviour using detonators to be used. [Specifying only
extensometers, tiltmeters and VW strain gauges cartridged explosives ensured that total explosive
was also performed and the movements evaluated weights per hole could be controlled and
in conjunction with the other movements and determined quickly. Non-electric detonators
motions of the buildings, (Carter et al., 1999). were a necessity due to the high density of
wireless communication devices and the presence
Vibrations induced by blasting and mechanical of radio towers on adjacent structures. To
excavation methods (hoe ramming and line drilling) minimize the potential for cut-offs during the
were monitored and, where required, controls were blast due to pinched shock tubes from the mats,
imposed on the Contractor to ensure that vibration primacord was used to initiate the non-electric
levels remained below the established criteria so as down-the-hole delay detonators. The primacord
to protect the delicate heritage building fabrics. was covered with soil to minimize air-concussion
effects prior to laying the mats.]
As a means to determining the response behaviour • Three layers of blasting mats to be required for
of the building structures, an array of triaxial each blast. [These mats were then covered with
geophones were installed on the building at various a tear resistant filter fabric or felt type material.
locations ranging from the basement to the top of This felt material was used to capture any small
the gothic buttresses and finials, and strain gauges rock fragments that may have been contained
and crack plates were installed across identified within the mats themselves. The presence of
pre-existing cracks within the 100 year old stained glass original windows within 11 m of the
buildings. In order to maximize control on the blasting and a high density of tourists around the
excavation activities, the key monitoring locations site necessitated complete flyrock control.]
were concentrated around the perimeters of the • All blasting to be designed with a minimum delay
two buildings, bordering the excavation area, with interval of 25 milliseconds. [The close proximity
a number of installations set on the fragile exterior of some of the subsurface monitoring sites during
masonry cantilevers overlooking the CBUS the trial blasting (4.5 m) and the use of wire
footprint. These masonry structures included the triggers and sampling at a rate of 16kHz using a
Carter et al - Response of Heritage Buildings 3
 Blastronics BMX datalogger allowed the delay both the interior and exterior of the Library of
timing within many of the blasts to be analyzed. Parliament, as well as the north and west exterior
The highest ground vibration levels recorded walls of the Centre Block. All interior rooms facing
during the 20 test blasts occurred with the one the excavation site and any additional areas specified
blast specifically designed to speed up the shot, by the Heritage Conservation Program, plus Access
during which the delay interval was as short as Tunnels, the Summer House Pavilion and the base of
9 ms. Delay timing scatter by as much as 10 ms the promontory slope were also inspected. In total,
was observed at least once in each of 5 test 52 hours of video time and 2600 photographs were
blasts analyzed. The use of a delay interval required to complete the pre-condition survey. In
shorter than 25 ms was therefore deemed places a crane was needed to access upper exterior
unacceptable as it would have introduced the walls, towers and roof areas. A complete post-
potential for out of sequence or simultaneous condition survey was also carried out following the
firing of sequential delays.] completion of the bulk excavation works.
• All blasts to be designed based on the vibration
attenuation data determined during the test 3.2 Ground Vibration Limits
blast phase, using the following regression
equation for a 95% confidence level: A review of internationally suggested vibration limits
PPV = 2104(SD)-1.488 to be adhered to for controlling potential damage to
where PPV = peak particle velocity (mm/s) heritage or historically significant structures
and SD = Scaled Distance (kg/m0.5) indicated considerable differences of opinion on what
was acceptable. The various international guidelines
At the Tender stage, in order to meet the vibration and standards which are plotted on Figure 3, also
limitations established for the Library of Parliament demonstrated a divergence on allowable vibration
and Centre Block, the approved blasting contractor dependency with frequency, even to the extent of
divided the CBUS footprint into three excavation disparity with respect to measured limits and
zones, two with differing blast parameters, as maxima. As is evident from the plot, the German
shown in Figure 3. Within these two blasting Institute Standard 4150 is amongst the lowest and
zones, a total of 280 blasts were detonated over the most stringent of the criteria, being frequency
seven-week period for the main CBUS excavation. dependent between 3 and 10 mm/s. Some of the
Each blast typically involved firing between 21 and previous USSR and several other European
23 x 50 mm diameter holes on a 1.2 m square countries, by contrast, are less restrictive, allowing
pattern. Successive lifts commenced with a small permissible vibration limits of 30 mm/s or higher for
sinking cut in Zone 1 incorporating two to four 75 occasional blasting.
to 100 mm diameter relief holes. The majority of
blasts were fired one hole per delay and consisted On the basis of the test blasts maximum ground
of maximum explosive loads per delay period vibration limits established for the Centre Block and
ranging from 0.5 to 3.0 kg. Only the four initial Library of Parliament were set appropriately for the
sinking cuts had explosive loads greater than 3.0 kg vibration behaviour of the buildings. As shown on
(5.0 to 6.0 kg. was typical), and decking was only Figure 3, these criteria are quite stringent at the low
required for seven of the blasts, with only one blast frequency end of the range, but fall somewhere in the
suffering a cut-off. middle of the spread of the international standards,
as follows:
3. CONTROLS
Dominant Peak Limit Vibration (mm/s)
3.1 Condition Surveys Frequency Range Centre Library of
(Hz) Block Parliament
Prior to the commencement of blasting for both the <10 5.0 3.0
trial blasts in 1996 and the main excavation in 10 to 40 5.0 to 40.0 3.0 to 17.5
1997, a pre-condition survey was carried out of >40 40.0 17.5
Carter et al - Response of Heritage Buildings 4
As is evident from the table and the plot on Figure gauges were used rather than pin measurements, so
3, two limits were set for the main buildings, the that real-time data could be gathered more
lower limit being applied to the Library due to the effectively. In total, 17 representative cracks on and
amplification observed in the exterior pinnacles on within the Centre Block and Library were strain
each of the flying buttresses. For any other ancillary gauged and an additional 16 interior cracks were
structures or services, ground vibration levels were also monitored using tape tell tales.
specified at normal, standard levels of 50 mm/s.

Figure 3: International Standards for

Figure 4: Digital calipers being used for
Vibration Limits for Historic Buildings,
measurement of typical crack monitoring pins
showing guidelines for CBUS.
In order to monitor the vibration intensity in the 3.4 Monitoring Instrumentation
bedrock and foundations of the buildings, simple
triaxial geophone arrays were mounted on the Various accelerometers and triaxial sensors and data
bedrock surface adjacent to the foundations of the recorders were installed to capture the dynamic
Library, Summer Pavilion, and north and west walls response of those portions of the buildings deemed
of the Centre Block. Monitoring results were most prone to excavation-induced vibrations.
reviewed at all stations following every blast. It Because of proximity to the excavation zones, a
was also a requirement to record drill patterns, concentration of vibration sensors was placed in the
explosive loading and initiation sequencing details. north-west corner of the Centre Block and across
the western face of the Library.
3.3 Crack Monitoring
As part of this instrumentation array, several multi-
During the test blast program, 29 existing cracks on directional sensor stations were installed beside the
the interior and exterior of the Library and Centre exterior loadbearing walls of the buildings to capture
Block were monitored both prior to and during the the effects of line drilling and hoe ramming activities
blasting operations. Pins set across eleven of these along the southern perimeter of the CBUS footprint
cracks (see Figure 4) were measured for lateral and immediately adjacent to the buildings. The vertical
shear displacement using digital calipers. Readings masonry cantilevers on both the Centre Block and
were taken at various times throughout the day, Library overlooking the CBUS footprint that had
and the average of two sets of readings was been monitored for the Test Blasts were again
recorded during each visit. The remaining 18 instrumented for the CBUS excavation works. This
cracks, all interior, were monitored using tape and monitoring instrumentation included several Library
glass tell tales. In these locations, pins could not be pinnacles, as the response of these elements was
installed due to either the locations sensitivity or found to control the blast parameters for the eastern
the fragile nature of the material in which the portion of the CBUS footprint
cracks existed. During the CBUS excavation, strain

Carter et al - Response of Heritage Buildings 5

4. BEDROCK/BUILDING RESPONSE Examination of the vibration response waveforms for
those transducers mounted on the building walls as
4.1 Ground Vibrations – Blasting Influences compared with the waveforms from those
transducers mounted on the bedrock, (as typically
Based on the data collected from the 280 individual
blasts carried out for the CBUS works, peak
ground vibration levels recorded at the Library of
Parliament remained below the prescribed limit of
17.5 mm/s during the entire blasting operation, the
highest level recorded being 17.1 mm/s. Similarly,
all but one blast remained below the 40 mm/s
criteria for the Centre Block. This single high PPV
reading of 43.2 mm/s was attributed to one, over
confined, hole in the second row of the blast, based
on immediate evaluation of blast waveform records. Figure 5: Percentage drop in peak particle
velocity level from bedrock to library
The initial blast results for the CBUS excavations foundation wall.
indicated that peak ground vibration levels could
typically be expected to be 70 to 85% lower than recorded as shown in Figure 6), suggested that the
the levels recorded on the bedrock adjacent to the vertical component of the vibration energy was being
foundation walls. However, during the course of preferentially transmitted to the foundation walls. As
excavation, as the cut became deeper, less of a the depth of excavation increased, though, measured
difference was evident. The comparative results vibration levels became more isotropic, with similar
plotted on Figure 5 shows clearly that by the end of responses in both vertical and longitudinal directions
excavation down to almost 10m depth, typically (perpendicular to the foundation walls). Fourier
only a 30% drop in energy was being measured on analysis of the waveforms, both in the bedrock and
the foundation wall as a percentage of the adjacent foundation walls, however, remained similar,
bedrock velocity response. This suggests that the indicating dominant frequencies consistently higher
coupling between the building’s foundations and than 40 Hz and often in excess of 100 Hz (Figure 7).
the bedrock improved with depth of excavation,
and, as a result, higher responses were being
generated in the building walls for essentially the
same levels of blast-induced rock excitation.

Figure 7: Fast Fourier Transform (FFT) plot

for blast record shown in Figure 6.

4.1 Ground Vibrations – Mechanical Excavation

Figure 6: Typical bedrock vibration response
for close proximity blasting. Similar decoupling characteristics were recorded for
the mechanical excavation equipment, when
Carter et al - Response of Heritage Buildings 6
compared with the blasting responses, apart from near the perimeter building walls. To reduce
one significant difference in behaviour, that it was vibration levels within the foundation walls, relief
found that the ram-induced vibrations were much holes were drilled into areas of the bedrock that
more localized, decreasing rapidly with distance were required to be broken by the hoe-ram. This
from the hoe-ram. As a result, even when hoe- effectively reduced the energy needed during
ramming with a large, 10,000 ft-lb. breaker, (with, ramming to break up the rock. The holes, each
of the order of 300 to 400 impacts per minute), the about 2-m deep, which were regularly spaced at
excavation operations being carried out within the about 1m intervals parallel and perpendicular to the
CBUS footprint (such as are shown in the foundation walls greatly assisted break-up of the
photograph in Figure 1), had only a small effect on rock and as a result significantly decreased measured
the Library, less than 15m away laterally. vibration responses.

Figure 8: Typical bedrock vibration response Figure 9: FFT Plot for hoe-ram vibration
for hoe-ramming record shown in Figure 8.

The largest vibrations that were measured from the The other difference noted between the hoe-ramming
hoe-ramming operations occurred in the foundation responses and the blasting responses was that with
wall of the Centre Block when the ram was being increased depth of excavation, the hoe-ramming
operated to break fairly massive, unfractured operations were found to induce less and less of an
limestone bedrock at about the same elevation as effect on the Centre Block foundation wall vibration
the base of the wall and within a few metres of the levels, almost diametrically opposite to the effect
outside of the wall. Recorded velocities under noted with the blasting that showed better coupling
these conditions exceeded 25 mm/s in the wall with at the deeper excavation levels.
dominant frequencies at harmonics of the ram
frequency (but significantly >30 Hz). As is evident Examination of the records to check whether
from the waveform plot on Figure 8 and the FFT significant signatures could be identified associated
plot on Figure 9, dominant frequencies in the time with TBM operations or with the Line Drilling of the
record are a combination of the harmonics of the perimeter of the excavation footprint, yielded little
ram frequency (in this case about 6Hz) and the definitive data. Based on the measured responses, it
frequencies produced by each of the blows when was concluded that neither source had major
several seconds of hoe ram signal (containing many vibration effect on either the Library or Centre
blows) are analysed. Block, as vibration energy induced by these activities
failed to trigger the data acquisition systems installed
As a consequence of these high, localized velocities to monitor excavation-induced vibrations. Although
being measured in the foundation walls a change in known to be less than 1 mm/s, (the trigger level), no
hoe-ramming procedures was implemented for all actual magnitudes or frequencies were recorded.
areas where it was required that the ram operate
Carter et al - Response of Heritage Buildings 7
4.2 Crack Displacement responses in the upper portion of the Library were
less than those on the basement floor slab.
Crack monitoring both during the test blast and
main CBUS blasting programs demonstrated no Responses at most locations in the two structures
evidence of significant movement across any of the including the vertical cantilevers were composed
29 cracks monitored. 16 of the 17 strain gauges primarily of high frequencies (>20 Hz). However, as
installed across selected interior and exterior cracks the response levels of the cantilevers increased
on the Library and Centre Block for the main (because of either an increase in the size of the blast
CBUS excavation showed maximum displacements or a decrease in the distance to the blast),
not exceeding 0.1 mm between early July and early particularly the pinnacles and chimney of the Library,
November, 1997. Only one strain gauge showed so did the low-frequency content within these
movement in excess of this. This gauge, which was responses. This downward shift in frequency may be
sited at an interior location, directly adjacent to the indicative of a change in the overall response of the
west intervention, showed displacement of 0.2 mm cantilevers from one containing the contributions of
attributed to the actual intervention breakthrough. many modes to one dominated by the contributions
Apart from this movement, the only other of a few of the lower modes. On the other hand,
discernible trend was a consistent opening and building responses did not exhibit much low-
closing on the order of 0.05 to 0.10 mm noted in frequency or resonance-type behaviour as the
the exterior gauges. As this movement was not majority of the energy emanating from the blasts
evident at any of the interior strain gauge locations contained frequencies well above the low-frequency
and continued well into the period after external modes of the two buildings (probably less than 5
excavation activities had been completed it is Hz).
considered that the movements were related to Responses measured in both buildings remained
regular expansion and contraction of the cracks in below the recommended limits for all but a few
response to temperature and humidity changes. blasts indicating that the previously derived criteria
for the footprint, which was obtained from the test
6 CONCLUSIONS blast program (a 4-m excavation) applied to the full
depth of the 10-m excavation.
The comprehensive monitoring of the building
response to excavation by both blasting and Responses measured within the two structures
mechanical methods has revealed complex showed similar attenuation patterns within their
multimode behaviour of the more frequency confines for all excavation depths, decreasing with
dependent parts of the structure (the flying both building height and distance from the
buttresses and fragile pinnacles. foundation wall nearest the blast.

Comparison of adjacent bedrock and basement or Library pinnacle responses stayed below levels
concrete foundation wall responses indicated that obtained during the test blast program partly because
the buildings were poorly coupled to the bedrock. of the conservative criteria selected for the Library
For most of the test blasts, basement/foundation and partly because the upper 1.5 m of the pinnacles
responses were considerably less than 50% of those were removed prior to the commencement of
measured on the adjacent bedrock. A further excavation operations.
reduction was noted within the buildings
themselves as the energy propagated upwards and It was found that responses due to blasting were
outwards. Responses overall decreased with more significant in terms of magnitude but contained
distance from the blast and with elevation within less low frequency harmonics than responses created
the buildings. Responses (PPV) in the Centre by mechanical hoe-ramming. Guidelines for future
Block including those on the towers were excavation-related response monitoring suggest
significantly less than those measured in the that….
foundation wall nearest the blasts. Similarly
Carter et al - Response of Heritage Buildings 8
As the lowest natural frequencies of the Library are Parliament Buildings, Ottawa. Proc. 37th U.S. Rock
below 20 Hz, most of the energy produced by hoe Mechanics Conf., Vail, pp. 585-592
ramming and blasting was contained in frequencies
well above its dominant lowest modes. As such, Curic, A., Van Bers, M.V. and McAnuff, A. L.,
the Library responded to these excitations primarily (1999) Blasting on Parliament Hill, Ottawa adjacent
as forced vibrations at frequencies well above the to Canada’s No. 1 Heritage Building. Proc. 25th
frequency range of its dominant modes. High- Annual Conf. On Explosives and Blasting
frequency components comprising the majority of Techniques, Nashville
individual responses were attenuated and low-
frequency components close to natural frequencies Chae, Y. S., Design of Excavation Blasts to Prevent
amplified. Responses in the Library, therefore, Damage, Jnl. Civil Engineering- ASCE, 1978.
varied from component to component with
amplifications in vertical cantilevers (such as Din 4150 Part 3, Deutsche Norm, Structural
pinnacles and chimney) producing responses which Vibration in Buildings, Effects on Structures, 1975.
were larger than those measured at the base of the
buttresses. Dowding, C. H., Blast Vibration Monitoring and
Control, 1985.
Based on the vibration data, the lowest natural
frequencies of the Library which control the seismic Konon, W., Schuring, J. R., Vibration Criteria for
response of the building are in the 8 to 12 Hz Historic Buildings, ASCE National Convention,
range. On the other hand, blasting and hoe 1983.
ramming contain very little energy at these low
frequencies. Response to these events will, Rainer, J.H., Effect of Vibrations on Historic
therefore, be primarily forced vibrations at the Buildings: An Overview 1982. Proc. Association for
dominant frequencies induced within the bedrock Presentation Technology Bulletin, Vol. XIV, No. 1,
adjacent to the building. Response within the pp. 2-10.
building itself will change as the waves propagate
upwards and outwards from the foundation. Since Sedovic, W., Assessing the Effects of Vibration on
there are few if any major components in the Historic Buildings, Association for Preservation
building with significant high-frequency modes, Technology Bulletin Vol. XVI No. 3 & 4, 1984.
higher frequencies in the waveform will tend to die
out or dampen at a faster rate than lower Subercaseaux, M. I., Pernica, G., van Bers, M. V.,
frequencies with increasing distance from the event. (1999) Designing a Test Blasting Program for an
Underground Building on Parliament Hill,
In comparing the tast blast program to the actual Association for Preservation Technology Bulletin
CBUS blasting the overall response of the Library Vol. XXX No. 2 & 3, pp. 67-73
increased either because of increased blast size or
reduced blast distance, the low-frequency response
of some of the vertical cantilevers closest to the
blast increased. One reason for this change may
have been an increase in energy in the
neighbourhood of the lower modes producing a
resonant-type of response.

7 REFERENCES

Carter, T. G., Webb, G.S., Van Bers, M. V., and

Culham, G. (1999) Design and Construction
Solutions for excavation under and adjacent to the
Carter et al - Response of Heritage Buildings 9