THE ARUP JOURNAL

APRIL 1980
 THEARUP
JOURNAL
Front and back covers : R u n n y m e d e Bridge across the T h a m e s alongside L u t y e n s ' Bridge ( P h o t o s : Harry S o w d e n )
Safety factors:
Some aspects of
CIRIA Research
Project 223
Poul Beckmann
W h a t is a safety factor ? In the old d a y s w h e n
strength of structural elements could usually
only be quantified by testing, it w a s taken
to mean the ratio between the failure load
under test and the load the element w a s
required to carry in service.
W h e n engineering s c i e n c e had progressed
s o far as to make engineers believe that
stresses could be accurately calculated,
the quantification of the safety of the
structure took the form of stating that the
stresses under service load should be less than
the relevant 'permissible working stress.' T h e
permissible working stress w a s defined as the
ultimate strength of the material ( w h e n
tested in a certain standardized w a y )
divided by a 'factor of safety' w h i c h took
a c c o u n t of the estimated c h a n c e of overload,
the variability of the strength of the materials
and past design tradition.
Experienced engineers were using this format
quite successfully for more than half a
century, but the new generation, brought up
to respect the w o r d s of codes of practice more
than the advice of their elders and betters, ran
into trouble because some codes had not
stated explicitly that certain loads vary
independently of others and that variations
of materials strengths are of more significance
in some cases than in others. T h i s led to the
development of 'the partial safety coefficient'
approach in w h i c h separate factors k n o w n
as partial coefficients' or partial safety
factors, were allocated to the loads and to the
2 resistance of the structure.
Vol. 15 No. 1 April 1980
Published by
Ove Arup Partnership
1 3 Fitzroy Street, London W1 P 6 B Q
Editor: Peter Hoggett
Art Editor: Desmond Wyeth F S I A D
Assistant Editor: David Brown
The first code of practice in Europe to adopt
fully this principle w a s the Danish Foundation
Code and generally the S c a n d i n a v i a n s
seemed to embrace this approach ahead of
everybody else.
T h e European concrete committee introduced
the principle in their ' B l u e B o o k ' in the 1 9 6 0 s ,
and, as it is the unwritten rule for any code
committee never to be seen to reduce
substantially the factor of safety of anything,
they set themselves the aim to introduce
partial safety factors in s u c h a w a y that
structures designed on the basis of these
w o u l d provide substantially the same safety
as w h e n designed to existing 'permissible
stress' codes. T h e y had certain statistical
information available w h i c h led them to
deduce that the factors on the material sought
to be in the order of 1.5 for the concrete
strength and 1.15 for the strength of the
reinforcement. T h e requirement for similar
overall safety then led to a partial safety
factor on the loads in the order of 1.5.
The result of this w a s , however, that the n e w
method w a s no more rational in terms of the
real safety provided than w a s the old, and
hence it w a s thought that 'something ought
to be done about it.' T h e thought led to
C I R I A , in the beginning of 1 9 7 5 , sponsoring
a research project aimed at deriving a method
by w h i c h more rational partial safety factors
could be calculated using the tool of the
'reliability theory.'
Reliability can in this connection be defined
as the reciprocal of the probability of failure,
and this will of course depend on both the
loads and the structural resistance and the
likely variability of both.
Structural failures do of course mostly occur
for reasons other than those w h i c h can be
connected directly with factors of safety a n d ,
whilst it is possible to calculate the risks
due to a certain set of factors of safety, the
actual risks of a failure of a structure designed
to present day codes are about 10 times as big
as the calculated risks, due to c a u s e s s u c h as
Contents
Safety f a c t o r s :
S o m e aspects of
C I R I A Research Project 2 2 3 ,
by P. B e c k m a n n
R u n n y m e a d Bridge, 5
by W . S m y t h
T h e south abutment 9
at Kessock Bridge,
by K. C o l e
Design and construction
at the Lukasrand m i c r o w a v e 12
tower, Pretoria,
by C . M c M i l l a n
Central B a n k , Dublin
18
gross errors in design and construction and
acts of G o d . ' ( T h e latter w a s , incidentally,
defined during the research project as s o m e -
thing the designer c a n use as a n e x c u s e once.
W h e n s u c h an event has occurred, future
designs should foresee the possibility).
Before getting alarmed about this loose talk
of risks, it is however useful to remind
oneself of h o w small the risk of death from
structural failure is, compared w i t h a lot of
risks w h i c h people quite happily accept, (see
Table 1) The research project, w h i c h w a s led
and co-ordinated by Dr. A . R. Flint, included
a lot of work on reliability theory by M. J .
Baker of Imperial College, and later on a
considerable amount of computer w o r k by
S t a n Feneron of Building Design Partnership
aimed at demonstrating the reliability of a
number of structural elements designed to
various codes as w e l l as some exercises
aimed at s h o w i n g h o w the n e w partial safety
factors, derived from probability theory, could
be calibrated to give substantially the same
overall reliability as w o u l d result from present-
day codes. Ian Potter of J e n k i n s & Potter pro-
d u c e d a very interesting section on legal and
economic aspects of safety.
I w a s given the task of carrying out a 'state of
the art' investigation in current c o d e s and
regulations in the U K and abroad. T h i s in-
volved me in studying not only a large number
of British standards and codes of practice,
but also some of the German D I N Norms,
A m e r i c a n , French and Danish codes and some
reports by the Nordic Committee for Building
Regulations. T h e purpose of this study w a s to
demonstrate the extent to w h i c h the partial
factors in these codes covered all the
variabilities they theoretically ought to a l l o w
for.The C E B Bulletin III quotes the general
design equation a s :
T f x Effects of ( r , , D + T £ ^fi L Wi_)
3 0
< Resistance from — w h e r e W d is dead
Ym
load and W , covers live loads, w i n d loads,
etc. and fk is the characteristic material
strength. Fig.1 •
Both loads and strengths in this equation are Typical relationship
entered with a 'characteristic' value, that is between cost
a value that has a certain agreed probability and total probability
of occurring. of failure
Y is a reduction factor on material strength
m

w h i c h is intended to take account o f :

(7) Material strengths occasionally falling
below the specified characteristic value:
T h e probability of this depends mainly on
w o r k m a n s h i p a n d / o r quality control. Apart
fromthe masonry code, British c o d e s a n d drafts
don't allow for this. T h e probability of a failure
depends, everything else being equal, not only
on the characteristic strength of the material
but also on the variability of the strength. High Consequential losses
mean strength and high variability g i v e s
greater risk for the same characteristic
strength. T h e only mention of this is found in
the N K B document.
( 2 ) Possible differences between strength of
the material in the structure and that deter-
mined from control test specimens: Initial cost = f,
T h i s difference depends of the frequency of Fig. 2 •
sampling related to the consistency of 'Naive' application
production, and on the different behaviour of of partial safety
the standard sample in the testing machine factors to the 10 2
10" 3

from that of the bulk of the material in a real problem of

structure. For example, the British Standard Annual probability of failure
stability defeats
Concrete Cube is compacted and cured in a the object
completely different w a y from the concrete
in the structure.
It f o l l o w s that w h e n one is assessing an
existing structure on the basis of tests of
actual strengths of material in the structure
Y can be reduced.
m

{3) Possible weakness in the structural

I
material resulting from the construction
process:
T h i s is a function of w o r k m a n s h i p and super-
vision. British codes and drafts, except the
masonry draft w h i c h varies Y with m a n u f a c -
m

turing quality control and site testing, f o l l o w

A m e r i c a n and C a n a d i a n practice in having
invariant Y s.
m Scandinavian documents
grade their materials factors according to
frequency of a c c e p t a n c e testing and site F = /iA/ = n G cos * x (A
control.
(4) Possible inaccuracy in the assessment of N • G cos x
the resistance of a structural element resulting TfQ x l - Y\ G cos x x / i x i
G

from modelling errors:

It is impractical in a code of practice to
formulate simple calculations w h i c h will
predict accurately failure for the w h o l e range
of e a c h type of member that has to be covered
by the code. Hence, some a l l o w a n c e has to T a b l e 1 : Comparative annual probability of death per 1 0 0 0 0 persons
be made for the resulting errors. Grossly
erroneous assumptions can however not Hours Annual Approximate
be covered by partial safety factors. exposure/ risk/10,000 annual
annum persons risk/person
( 5 ) The effects of poor dimensional accuracy
in the finished structure on the resistance of a Mountaineering (International) 100 27 10- 2

section:
I n a c c u r a c i e s in construction are a function of Distant water trawling ( 1 9 5 8 - 7 2 ) 2900 17
w o r k m a n s h i p and hence supervision. T h e Air travel ( c r e w ) 1000 12 10- 3

British draft masonry code and some 1600 3.3

Coal mining
S c a n d i n a v i a n codes graduate mY accord-
ingly, but they are the only ones to do so. Car travel 400 2.2 2x10- 4

T h e effect of i n a c c u r a c i e s is greater for small Construction site 2200 1.7

members, so Y should vary ideally with size,
m
Air travel (passenger) 100 1.2
but no code seems to cater for this.
Home accidents (all persons) 5500 1.1 10- 4

Yf, is a multiplier applied to loads or im- Home accidents (able bodied) 5500 0.4 4x10- 5

posed deformations to take account of the Manufacturing 2000 0.4

possibility of loads exceeding their character-
STRUCTURAL FAILURE 5500 0.001 10- 7

istic value:
There is here the difficulty that very f e w types
of load are sufficiently recorded to establish All causes ( E n g l a n d and Wales, 1 9 6 0 - 1 9 6 2 )
a mean value, let alone a characteristic.
T h e codes tend to apply factors to nominal Male age 3 0 8700 13 10- 3

statutory' loads traditionally stipulated in

Female age 3 0 8700 11
regulations.
Male age 5 0 8700 73
Certain loads, e g : water pressures, are
calculable w i t h less uncertainty than usual Female age 5 0 8700 44
live loads and therefore deserve a lower value Male age 5 3 8700 100 10- 2

of Yf. S o m e codes a l l o w this, others don't.

 * F is a multiplier
0 applied to load combina-
tions to take account of the reduced prob-
ability of all loads exceeding their character-
istic value simultaneously:
S o m e codes, e . g : the C a n a d i a n , apply f 2

this as a multiplier to the s u m of Yf, factored o

loads, others vary the T| for each type of II Y,on CP3I ve load
5
load depending on the combination.
T f is a multiplier
3 on load effects to take
account of errors in predicting load effects as a
result of neglecting dimensional inaccuracies:
to take account of non-linear behaviour
where there is a magnification of load effects:
Different accuracies of assumptions and , Ymlstee ) — on charade istic yield strenr. th
computations should qualify for c o r r e s p o n d -
J
ingly different values. In fact, the a c c u r a c y )
a s s u m e d in the implied Y f (as incorporated 1
3 > »-i
in the various c o d e s ' Y f - values) is n o w h e r e
defined. Non-linear behaviour is another \
matter and ought not to be introduced here,
Y T, (concrete) — o n a characteristic cube strength
as its effects vary so much from case to case.
Yn does not appear in the equation but is used 1 — on nominal c ead load
to modify Y | or 7m to take account of the
0 5 i 1• 1
nature of the structure and its behaviour and 4 6 8 10 12 14
the seriousness of attaining the limit state: Span(m)

Elements and structures that fail suddenly

should have a greater overall factor of safety
Fig. 3 A Fig.4T
than those w h i c h , due to their ductile behav-
Partial safety factors for Partial safety factors
iour give w a r n i n g of collapse. T h e G e r m a n ,
singly-reinforced beams—office loading for Grades 5 0 universal beams
A m e r i c a n and C a n a d i a n codes seem to give
the most explicit a l l o w a n c e for this d i s -
tinction. None deal with hidden ductile 3 a*.
structures, w h e r e the w a r n i n g signs cannot
be observed. ^""•s^^ Y f on CP3 office live loading

Structures, the collapse of w h i c h may kill or o

injure a large number of people, w o u l d s e e m II
to qualify for higher safety standards than a ? 2- Y i on CP3 retail premises (storage areas) live loading
lintel over a door in a transformer h o u s e . o
quired i

T h e S c a n d i n a v i a n drafts have three c l a s s e s

of structure according to the probability — * "
of death and injury and the level of damage
•* '
to property resulting in case of collapse.
T h e C a n a d i a n code has an importance S * ._ n n
" -..
o 1

factor' but only with t w o values, the lower nj

one applying to certain farm buildings.

Difficulties are sometimes experienced in
r 'm All in this zone /
Yt on nominal dead load
allocating some of the partial factors, notably
T f to one or the other side of the equation.
3 Yt on CP3 retails premises (sales areas) live loading
A factor of 1.4 to be applied to o w n w e i g h t is,
in the absence of any explanation, taken by ' i i i i i 1 1 1 r
0 2 4 6 8 10 12 14 16 18 20
some engineers as an insult to their standards
of calculation and the application in CP110 of Span(m)
1.4 and 1.0 to o w n weight in adjacent s p a n s ,
stretches the imagination of most of us. The
explanation lies of course in the positioning methods for calculating this value for a finite ful models. T h e same applies to shipbuilding.
number of tests. Whether these methods are (Probability theory w o u l d have predicted
of r .
adequate to deal with the testing frequency in B 7 0 7 s falling out of the skies by the dozens I)
f a

Certain earth stability problems t h r o w up even
w o r s e predicaments: in the classical slip
circle calculation the calculated factor of
safety remains constant regardless of the
value c h o s e n for 7f.
Apart from the partial safety factors there are
certain other parameters that affect the
reliability of a structure:
D e s i g n life
T h e limit state philosophy usually includes
the phrase: the structure shall not become
unfit for its intended use during its design
life.' No code defines the design life and yet
it w o u l d seem essential to do so w h e n
dealing with materials w i t h strengths w h i c h
reduce with time s u c h as timber or forces
w h i c h increase w i t h age such as drag on
marine structures due to barnacle growth.
S o m e confusion also occasionally o c c u r s
between design life and return period of
natural phenomena, s u c h as w i n d forces.
Characteristic strength and materials
testing:
T h e 'characteristic' value of the strength of a
material is defined in British and C E B prac-
tice as being the one attained or exceeded
by 9 5 % of a theoretically infinitely large
4 number of tests, and there are statistical
BS 4360: one tensile test for every 4 0
tonnes, is another matter. (It corresponds to
one test for each km length of 1 2 i n . x 6 J i n .
U B ( 3 0 5 x 1 6 5 x 4 0 ) or one test per 2\
milesof4in. x 3in.angle (102 x 7 6 ) . )
D e f i n i t i o n of f a i l u r e :
Mention w a s made earlier of modelling
errors in calculation of resistance of a struc-
tural member, but occasionally our traditional
assumptions about the mode of failure lead
to errors not of plus or minus 1 5 % but several
hundred %. For instance, a beam spanning
between heavy c o l u m n s at 1st floor level of a
multi-storey building will carry, in a catenary
action, a load several times that w h i c h will
collapse an identical beam at the top of the
building.
Design practice in allied
engineering :
T h e treatment of safety in aeronautical,
marine and nuclear engineering w a s studied.
T h e information obtained on nuclear practice
w a s scanty and did not indicate any methods
w h i c h were more rational than current
structural practice. Aeronautical engineering
uses probability theory extensively to predict
design loads but has to apply severe modifi-
cation factors to make the results compatible
w i t h the safety records of previous s u c c e s -
In both instances w e found little that could
usefully be adopted in building and civil
engineering design because ( a ) w e do not
have enough data on performance, failures
and their c a u s e s , ( b ) w e do not have long
enough production runs to be able to benefit
from prototype testing, and ( c ) w e cannot
practically enforce a materials testing regime
a n y w h e r e as stringent as the aircraft industry.
T h i s article only deals w i t h a small part of the
entire research project w h i c h is fully described
in C I R I A report no. 6 3 .
I hope that none of the hardworking chairmen
and members of code committees take any
of the comments in the report or in this
article as a criticism of their work. Having been
involved on the sidelines of one of those
committees, I k n o w that it is difficult just
fields of to make the transition from permissible
stresses to partial safety factors, especially
w h e n it is also stipulated that design a c -
cording to the n e w principle must not result
in something significantly different from
the end result of the old code.
Acknowledgement
Figs 2, 3 and 4 are reprinted from :
C I R I A Report 6 3 , Rationalisation
of safety and serviceability factors
in structural c o d e s 1 9 7 6 .
Runnymede Bridge
Bill Smyth
T h e t w o bridges over the T h a m e s at R u n n y -
mede allow an interesting comparison to be
made. One w a s designed by an architect w i t h
help from an engineer and the other by
engineers w i t h help from architects.
Sir E d w i n Lutyens designed the first bridge
to carry the S t a i n e s B y p a s s . T h e c o n c e p t u a l
design w a s completed in 1 9 3 9 , he died in
1 9 4 4 and the bridge w a s built in 1 9 6 1 1 2 . It is
basically a thin arch of white concrete
spanning 5 6 m over the river and thrusting
against cellular abutments founded in L o n d o n
clay. In the space b e t w e e n arch and d e c k
are steel trusses transferring the loads to the
arch and stiffening it. T h e bridge w a s
constructed by cantilevering the trusses out
from the abutments, and suspending the
formwork for the arch from them. T h e
spandrels are closed by brick panels w h i c h
c o n c e a l the trusses, the abutments are clad
w i t h brick and the parapets are Palladian
balustrades of Portland stone. T h e r e are t w o
t o w p a t h arches through the abutments a n d
these are lined w i t h brickwork and have
Portland stone facings.
A30Egham Bypass
3T
LT
u u u u jrrrrmr'
T h e L u t y e n s bridge w a s designed for six lanes
of traffic and t w o f o o t w a y s . T h e M 2 5 is n o w
to cross the T h a m e s and requires four lanes
and t w o hard shoulders in e a c h direction and
these are s a n d w i c h e d between t w o lanes of
the A 3 0 and a footpath in e a c h direction, s o
that a bridge of t w i c e the original w i d t h is
required.
There w e r e t w o major problems, one aesthetic
and one technical w h i c h f a c e d the design
team of the Civil Engineering Division and
Arup A s s o c i a t e s .
T h e first is w h a t kind of bridge c a n you
possibly put alongside the L u t y e n s bridge?
The Department of the Environment ( n o w
Transport) had emphasized the importance
w h i c h they attached to the problem by
appointing Arup A s s o c i a t e s directly as
consulting architects. T h e s e c o n d is that
L u t y e n s ' bridge is founded in L o n d o n clay
w h i c h is a bad material for founding an arch in.
S o m e spreading of the abutments w a s
a l l o w e d for in design but not e n o u g h to cope
with ground movements c a u s e d by another
arch alongside, and w e had to be careful not
to break its back.
W e made a cardboard model of the L u t y e n s
bridge and w e used this in conjunction w i t h
rough models made from card and balsa to
examine the visual effects of the various
New bridge
possibilities. T h e modelscope w a s a great
help because it a l l o w e d u s to get the v i e w -
point of a s c a l e - s i z e d m a n . W e decided to try
four kinds of s c h e m e s . T h e first w a s extension
of the existing structure. T h i s w a s thoroughly
unsatisfactory and w e w e r e able to get the
roads realigned slightly so a s to create a gap
of 3 m b e t w e e n the older and newer bridges.
T h e other groups of s c h e m e s w e r e :
( a ) Imitations of L u t y e n s ' bridge in appear-
ance or form only
( b ) Neutral bridges ( a very thin flat plate for
instance)
( c ) Other kinds of bridges w i t h curved soffits
w h i c h w o u l d be compatible w i t h the
existing bridge.
T h e final result of the studies of structure,
cost, and appearance w a s our project design
w h i c h w a s completed in 1 9 7 2 .
T h e bridge deck is carried by four concrete
frames w h o s e soffits over the river lie on the
same surface as that of L u t y e n s ' river a r c h .
E a c h frame is made up of t w o balanced half-
frames w h i c h are c o n n e c t e d over the centre
of the river. T h e trapezoidal portal ring at the
heart of e a c h half-frame is supported by t w o
sets of laminated rubber bearings and a simple
truss cantilevers out on e a c h side of the
portal. T h e top members of the frames are
prestressed and the other members simply
reinforced. Foundation loads ( w h i c h are of
course predominantly vertical) are taken into
the clay by large diameter bored piles w i t h
underreams.
T h e bridge superstructure is made of w h i t e
concrete using Balidon aggregate and w h i t e
cement, and is b u s h - h a m m e r e d . A similar,
though less strong, concrete w a s used by
A r u p A s s o c i a t e s at S t . J o h n ' s College,
Oxford.
T h e bridge deck is made up to the correct
level w i t h lightweight concrete, w h i c h helps
to reduce the effect of the very onerous
current requirements for differential t e m p e r a -
ture, w i t h waterproofing and black top
surfacing. T h e parapets are precast w i t h the
same finish as the rest of the bridge.
Fig.1
Site plan
Fig. 2
L u t y e n s ' bridge
( P h o t o : R. B e n a i m )
Fig. 3
Longitudinal section through L u t y e n s ' bridge
mnnnnr
The superstructure w a s detailed so that it
c o u l d either be constructed in situ on false-
w o r k (on piles in the river) or by casting the
frames on the banks and sliding them to their
final positions. T h e bridge w a s built by
Fairclough w h o decided to use the sliding
method.
The contractor had t w o sets of forms, one for
e a c h side of the river, so he got four uses of
them. T h e form and falsework were arranged
so that after casting and the first stage of
stressing, the units could be j a c k e d clear of the
soffit shutter and slid s i d e w a y s out of it on
P T F E coated skids sliding on tracks w h i c h
consisted of a thin layer of stainless steel fixed
to mild steel plates on a concrete beam
supported by piles. After the s i d e w a y s move,
the unit w a s j a c k e d up and the skids and
track rearranged so that it could be slid
forward to the river bank, again jacked and
slid s i d e w a y s to its final position. T h e motive
force for sliding w a s produced by hydraulic Credits
j a c k s pushing between the frame and nuts on
Client:
Dividag bars or, in some c a s e s , using strands
and automatically gripping j a w s on the j a c k s . Department of Transport
W h e n the t w o half-frames of e a c h pair were South Eastern Road Construction Unit
finally positioned there w a s a 2 m gap Road design:
between them. T h e s e c o n d stage cables were
South Eastern Road Construction Unit
placed, the gap concreted and the s e c o n d
Surrey S u b Unit
stage stressing took place. T h e top member
of each frame had sockets cast into the sides, Bridge design:
and rails w e r e fixed to these w h i c h carried the Ove Arup & Partners, Civil Engineering
travelling forms for the deck. W h e n the deck Division w i t h Arup A s s o c i a t e s as consulting
had been concreted, the third stage cables architects
could be threaded and stressed. T h e contractor
chose to use the B B R V 5 5 - w i r e s y s t e m , so the Contractor:
wires forming those cables w h i c h w e r e Fairclough Civil Engineering ( S o u t h e r n ) L t d .
threaded had to be button-headed on site.
L u t y e n s ' bridge w a s designed in c o n c e p t by
At the time of writing, w o r k on deck finishes
him w i t h Mr. Fitzsimons as engineer. T h e
and parapets is going on. W h e n our bridge is
w o r k s design w a s carried out by C . W . G l o v e r
ready, traffic will be diverted onto it and w o r k
& Partners, consulting engineers a n d Mr.
on strengthening L u t y e n s ' bridge c a n start.
George Stewart, consulting architect.

Fig. 4
Longitudinal section and cross-section of
Expansion |Oint - \
— A r u p s ' bridge

, 1 2 561 S H W M
Fig. 5
_ 9 290 River bed -Bearings AHUVMYI Model prepared for submission to the R o y a l
Fine Art C o m m i s s i o n . ( P h o t o : Henk S n o c k )
1200 dia bored piles - London Clay

Fig. 6
AA A;! 000

Diagram of sliding
Longitudinal Section

Fig. 7
V i e w from S o u t h bank
( P h o t o : Handford Photography)
3000 • 2500_ 3300 __yaries^

M 25 A 30 Fig. 8
The deck shuttering is carried on rails
Ewsting
-5 Uo o OO OO oo
4 bolted to the sides of the frames
( P h o t o : Kodak Vision)
bridge

Fig. 9
Precast paving
' slabs
Bearings access
chambers
Aerial v i e w looking north
( P h o t o : Milligan)
tl C 0 0 T T
5m
F i g . 10
HI. u uu Lj a u
General v i e w of n e w bridge
( P h o t o : Harry S o w d e n )
Section A A

4 A 5T
 1

w
During construction
A -
5**

Casting area

" 'T£ II

During sliding Pile cap

Section A - A

Thames •

Plan on slide path

6 A

Mttflfl

9 A

s
 Consultation with
Consultation
engineers

Send letter of
Finishes Parapets acceptance,
press release,
Preliminary Notice of inform unsuccessful
Appoint report probable tenderers, etc.
Oct 1971 completed delay
architects
•
ssue
Appoint DUbllC
engineers construct
notice Construction
Degins
Preliminary design

Consider alternatives Request

Collect available information programme
Assess requirements: S E R C U Meet contractor
TWA Set start date. etc.
Gas Board
Appoint
checkers Check
calculations
Revised tech
approval

Simplified, time-scale precedence diagram

R u n n y m e d e Bridge preconstruction stages

I
0 D A J A O D J A 0 J A O D A J A O J A 0 D F A J A

1971 1972 1973 1974 1975 1976 1977

R u n n y m e d e Bridge preconstruction
graph s h o w i n g cumulative staff hours
w o r k e d with time ( k e y events arrowed)

A Tender

Draft
documents
submitted

Draft technical
approval submitted

Public construction

Preliminary
report
/
/

Time

Runnymede Bridge
Method study diagrams by Greg A d a m i w

There is an interesting sidelight on our w o r k on
R u n n y m e d e Bridge. Lancaster University runs
a post-graduate course in Civil Engineering
Production Studies. Part of it involves the
students in spending t w o w e e k s at each of a
number of different organizations investigat-
8 ing some aspect of the running of jobs. A r u p s
had t w o of the students in February 1 9 7 9 and
one of them, Greg A d a m i w , studied our
method of working on the R u n n y m e d e
Bridge project, talking to people and studying
files, time sheets, operating plans and other
documents. Greg's very interesting report
included a precedence diagram and a
diagram s h o w i n g the total number of man
hours Arup staff booked to the project plotted
against time. T h e s e diagrams are reproduced
here. T h e y s h o w quite clearly the consider-
able gap between project design and w o r k s
design w h i c h has been a feature of many of
our h i g h w a y and bridge jobs.
The south abutment Bridge deck
Fill on heel
at Kessock Bridge, load (maximum)
225kNm included in
deadload of

Scotland abutment

Ken Cole
INTRODUCTION
Lateral force
T h e Kessock Bridge w h e n completed will of surcharge
span the Beauly Firth at Inverness, S c o t l a n d . -47kNm
A s s h o w n in F i g . 2, the abutment to the Deadload
Lateral soil force
southern approach viaduct retains a selected 1573kN/m
( K a = 0.3)
granular fill embankment some 1 2 m high, 524kN/m
and supports one end of the first 6 4 m span.
T h e bearings of the viaduct are designed to
transmit vertical load only, longitudinal forces
being absorbed in the piers and at the
northern abutment.

DESIGN OF FOUNDATIONS
Soil conditions Pile reactions vary
maximum 590 kN/m
T h e great depth of alluvial deposits at the site
have influenced the choice of f o u n d a t i o n s ;
a borehole for a pier adjacent to the navigation
channel reached 9 5 m below bed level, and
a borehole 1 km a w a y reached 8 9 m below Abutment foundations Fig.1
ground level, both encountering only alluvial Conventional analysis of the southern abut- Loading diagram of forces on abutment
deposits. T h e alluvial deposits increase in ment foundation indicated that three r o w s ( n o n - s e i s m i c loads)
density with depth and pad foundations have of piles raked at 1 in 4, and a rear row of
been adopted where maximum bearing vertical piles w e r e required to provide
pressures c a n be kept economically to within adequate support for the loads s h o w n in F i g . 1 .
2 5 0 k P a . W h e r e bearing pressures e x c e e d this, T h e driving record of piles is given in Fig. 2 ;
or the depth of construction below water test loading of a vertical pile gave 7 m m
exceeds 1 0 . 5 m , steel H-section bearing piles settlement at the design workrng load of Fig. 2
have been driven into the alluvial deposits. 1600kN. Location and general details

1 052m
m 2 A 64m , 72m 80m 80m { 72m ; 5 6-lm

North abutment boulh abutment

2.5m M H W S

Embankment
constructed
Longitudinal elevation of Kessock Bridge during previous
contract

17.6m Finished road level Fill surface on 28.6.79 Fill surface

on 1.5.79 15.2m
Bridge deck
13.5m 14.3

StudY Volume A
Volume selected
4m granular fi
StudX

Original ground surface

WfJWkWSO*' 100 150
4 s
200 blows/

Hammer B S P B 2 5
Rated energy 67kNm

Universal bearing piles Medium dense Majority of

3 5 6 > 358 174kg/m sand gravel readings lie
Grade 5 0 B steel cobbles & within this
To B S 4 3 6 0 boulders envelope

Gravel &
silty shelly clay Studs X & Y
on centreline

Dense sand
Gravel &
sand Plan of abutment
-36m Dense silty sand
(Bridge deck removed)
Section on Penetration
centreline record of piles
of road Dense fine sand
with layers of
silty sandy clay
 MOVEMENT OF THE FOUNDATIONS
Table 1
A n a l y s i s treating only f o r c e s on S A F E finite element input data of material parameters. All strength and modulus properties in
abutment MN/m . 2

A n a l y s i s of the movement of the abutment,

taking a c c o u n t of the forces s h o w n in Fig. 1 , Concrete
and treating the piled foundation as elastic Material Soil (All Piles Soils
springs, s h o w e d that abutment movements parameters (Analyses 3 & 4 ) analyses) (Analyses 3 & 5) (Analysis 5)
w o u l d tend to be d o w n w a r d s and forwards,
the base of the abutment moving some 4 0 m m Material number 1 2 3 4 5 6 7
t o w a r d s the north. H o w e v e r , it w a s evident
that the analysis of movement of the abutment Material type LE LE LE LE LE LE(s) LE(s)
could not be considered in isolation from the
ground movements that w o u l d be generated Drainage Drained Drained Drained Drained Drained Drained Drained
by the construction of the short length of
granular fill embankment between the already E, 50 30 30 000 50 50 50 30
placed embankment and the abutment. U,z 0.3 0.3 0.1 0.3 0.3 0.3 0.3
E 2
50 30 30 000 4 809 7 642 50 30
Loading sequence U12 0.3 0.3 0.1 0.3 0.3 0.3 0.3
T h e loading sequence used is s h o w n in Fig.4, G-I2 19.20 11.54 13 636 19.20 19.20 19.20 11.54
the construction process being simulated by CO 0 0 0 -14.036 0 0 0
four stages, A to D. T o c h e c k on the effect of c' — — — — — 0.017 0.017
the simplifying assumption that the embank- 0 ' — — — — — 0 0
ment load applied only a vertical load, Stage Z,
in w h i c h horizontal shear loads from the
embankment w e r e applied, w a s run in addition Note
to Stage B in A n a l y s i s 3 . LE linear elastic model
LE(s) linear elastic model w i t h strength limitation
A n a l y s e s of i n t e r a c t i o n o f s o i l a n d Drained e x c e s s pore water pressures assumed dissipated
structure E„E 2 principal Y o u n g ' s modulii in x, y plane, suffix 1 denoting horizontal (in p l a n e ) ,
A f i n i t e e l e m e n t p r o g r a m , S A F E . 1 incorporating 2 denoting vertical and 3 denoting horizontal (normal to plane) directions w h e n
facilities to take a c c o u n t of soil and structural oo = 0
material properties, w a s used to compute shear modulus between directions 1 and 2
ground movements and stresses in the anticlockwise inclination of direction 1 to horizontal (degrees)
structural elements. Previous computations cohesive resistance (limiting v a l u e )
using S A F E program, including modelling frictional resistance (limiting value)
of anchored retaining w a l l s , had s h o w n that
the program gave reasonably accurate predic-
tions of movements w h e n compared w i t h
Volume B Volume A
field measurements. Abutment loads applied as pressure
loads and shear loads to abutment base
A t w o - d i m e n s i o n a l finite element model of Fill loads applied as pressure loads
(Material 3 ) Nt
the abutment and the soil w a s established, A * i » + » . * * i
see Fig.3, and a series of analyses undertaken, \W///////M, f* 1 nl
m i
l rhnir<rniv\ fill

the a s s u m e d soil and structural material J T i d e range — J J jJJJ (Analysis3, S t a g e Z o n l y )

properties for A n a l y s e s 3, 4 and 5 being
s h o w n in T a b l e 1 .
Upper stratum
T h e soils w e r e given linear elastic properties (Material 1,6)
and a s s u m e d to be drained, since their rate Piles ^^^Ijljl
of consolidation w a s expected to be rapid.
T h e assumptions concerning the pile elements (Material4) llllll * (Material 5)
are stated in Fig. 5 ; for A n a l y s i s 4 the piles
w e r e a s s u m e d to be absent in order to
demonstrate the general validity of the
method. T h e main defect of the model w a s
that the pile elements had to be considered
as having the same unit depth' (of 1 m ) a s
Lower stratum
the soil and concrete elements, thus pro- (Material 2, 7 )
hibiting relative movement in the line of the
bridge between the piles and adjacent soil.

Computed and measured movements

The results of the computations of movement Boundary of mode
are presented in diagrammatic form in Fig. 5,
w i t h the measured results to date s h o w n on &
A n a l y s i s 3. Studs X and Y , on w h i c h level
readings have been made, are located on the
road centreline at the positions s h o w n in Fig. 3
Fig. 2. Site measurements of movement of Finite element model.
the abutment in the line of the road w e r e For properties of materials 1 to 7 see Table 1 Stage A
attempted, but, s h o w i n g little if any movement Build abutment 185
. i I
and being difficult to make with sufficient
'116
accuracy, they w e r e discontinued. The
abutment movements have therefore been Stage B
presented on the assumption that only Place volume A 185 25 5 5 230 230
vertical movement took place. embankment i
Fig.4 fill
116
Movements were time dependent Loading diagram for finite element model 25 40 45 38 20
T h e diagram giving settlement against time, St.iqe C Stage Z Add lateral shear from fill
Fig. 6, s h o w s that realization of full settlement Place Volume B 185 240 240 240 230 2 3 0
at e a c h stage of loading w a s delayed by embankment fill i > i .. * * 1 i l
material against the stem of the retaining w a l l
classical consolidation involving the dissi- 16 277
as the base took up the major part of its
pation of e x c e s s pore water pressures from
b a c k w a r d rotation during the placing of
the fine grained soils, complete dissipation Stage D
Volume A, see Fig. 5. In the event, the fill Place bridge 195 240 240 240 230 230
for e a c h stage taking about 6 months. T h e
surface w a s raised uniformly over the entire deck
specified sequence of embankment filling 44.
fill length from about 5m above datum to 13m y-Z
w a s that Volume A should be completely 281
above datum in one continuous lift, followed
placed in Stage B ( F i g . 4 ) before the placing nb All load and shears in kPa units
by further filling to road base level at 1 7 m
of Volume B w a s started in Stage C , the
above datum after a delay of t w o months.
10 object being to eliminate the pressure of fill
EFFECTOF MOVEMENTSON
STRUCTURE
Restraint on b a c k w a r d rotation
A s a c o n s e q u e n c e of following the described
s e q u e n c e of filling, the lateral pressure of
retained soil restrained the tendency of the
base to tilt b a c k w a r d s by developing the
bending strength of the connection between
stem and base, the measured average angular
rotation between the stem and base being
about 1 in 9 0 0 after the filling to 13m above Stud
datum. During the subsequent filling this 120
rotation increased to about 1 in 5 2 5 . It is
thought that the average vertical movements
Dates on which settlement
of base will have not been affected by the 160 readings were made
difference b e t w e e n the intended and the ') 17 January 1979
actual s e q u e n c e s of filling, but the b a c k - ii) 23 April 1979
w a r d tilt of the base may have been reduced. in) 18June1979
iv) 28 September 1979
v) 1 9 November 1979
S t i f f e n i n g e f f e c t of p i l e d f o u n d a t i o n
ANALYSIS 3 ANALYSIS 4 ANALYSIS 5
From Figs. 5 and 6, reasonably good cor-
Assumptions Assumptions Assumptions
relations are apparent between the predicted Pile elements have correct Pile elements deleted and Pile elements have correct
settlement from A n a l y s i s 3 and the actual axial stiffness, are replaced by soil. axial stiffness, are
settlement on 17 J a n u a r y 1 9 7 9 , w h e n the anisotropic and are connected anisotropic and connections to
to adjacent soil elements. adjacent soil elements have
loading w a s the abutment base and stem
shear strength limitation
( S t a g e A ) and on 1 8 J u n e 1 9 7 9 w h e n all cohesion - 170 kPa, f> = 0
loads except the bridge deck had been placed
and consolidation under those loads w a s
virtually complete ( S t a g e C ) . It is evident that
the piles moved together w i t h the surrounding
soil to give m u c h greater settlements than Fig. 5
could have been predicted from the results Computed and measured movements
of the pile test loading and considerations
of group a c t i o n ; the movements w e r e of the
same form but smaller than those that w o u l d Fig. 6
have occured if the foundation had not had Loading and settlement w i t h time
piles. T h e smaller actual b a c k w a r d tilt than
predicted by A n a l y s i s 5 indicates that the soil
adjacent to the rear r o w of piles did not move
differentially to the piles, and that shear
failure played little part in the overall be- Final fill level - 17.2m
Height of 20 100%
haviour. fillm completion „
"^onstrucrj^n
above of structure
datum 10 complete
Credits 50%
S Fill Level
T h e client is the Scottish Development
Department w h o kindly gave their consent to
the publication of this paper.
T h e contractor is a consortium of Cleveland
20
Bridge and Engineering Company and
Stud 'X
Redpath Dorman Long ( C o n t r a c t i n g ) Ltd.
T h e bridge w a s designed by Dr. Ing H. 40
Homberg in association w i t h C l e v e l a n d , and
the substructure by Trafalgar H o u s e Engineer- Settlement
ing S e r v i c e s in association w i t h R D L . mm
T h e joint engineers for the client are C r o u c h
and Hogg of G l a s g o w and Ove Arup and 20
Partners.
A c k n o w l e d g e m e n t is made of the dedication 40
of the site staff in obtaining the readings
w h i c h made possible this contribution.
6.)

Reference
80
S I M P S O N , B., O ' R I O R D A N , N. J . , and
C R O F T , D. D. A computer model for the
analysis of ground movements in L o n d o n
Clay, Geotechnique, 29 ( 2 ) , pp. 1 4 9 - 1 7 5 ,
1979.
 Design and
construction of the
Lukasrand
Microwave Tower,
Pretoria
Cliff McMillan
Architects:
South Africa Public W o r k s Department,
Architecture Department

Introduction
In 1 9 7 3 the Public W o r k s Department
appointed Ove Arup and Partners as consult-
ing civil and structural engineers for the
design, documentation and supervision of a
n e w m i c r o w a v e tower at Lukasrand, Pretoria.
Initial design meetings were held immediately
to obtain the brief from the user client, the
Department of Posts & Telegraphs. T h i s
included the f o l l o w i n g main requirements: Fig.1 A Fig.2T
(a) A s an important link in South A f r i c a ' s Alternative forms considered General arrangement plan and elevation
m i c r o w a v e telecommunications s y s t e m , defining levels and key dimensions
the tower mainly serves a technical
function. Functional requirements in-
clude the provision of three antenna 176 950
platforms at the top of the concrete Steel
portion of the tower, with equipment mast
and technical rooms immediately below
them and a steel mast on top. LEVEL HEIGHT
N°
(b) B e c a u s e of the dominant location of the
tower in relation to Pretoria's skyline and 18 156 950
Radio
the numerous important buildings in the 16A
level
vicinity, particularly the U N I S A c a m p u s 16 149 250 SECTION 1-1
nearby on the same ridge, the tower had
to be an architectural statement of
quality appropriate for the capital city.
(c) There w a s to be public access to a
v i e w i n g platform but this should not
interfere with the technical functioning 11 A
of the tower. Restaurant facilities w e r e Technical
not required. level

V*
(d) There w a s to be a t w o - l e v e l building
serving the tower at the base, w i t h 112 900
related roads, parking, stormwater drains
and other civil engineering services.
(e) T h e height of the tower w a s limited by
considerations relating to the flight path SECTION 2-2
to Waterkloof Airport.
W o r k c o m m e n c e d immediately on the design, Public

r:iO
but in 1 9 7 5 the project w a s delayed for nearly observation
6A
t w o years due to lack of available funds and level
the general economic climate prevailing. 79 300
Work w a s h o w e v e r resumed in March 1 9 7 7 ,
enabling construction of the tower to start
t o w a r d s the end of 1 9 7 7 w i t h a target \

completion date of the end of 1 9 7 9 . <

Design concept
The design c o n c e p t evolved through very
close collaboration between the P W D A r c h i -
tectural Department and the consulting
engineers. S i x different alternatives providing
the s p a c e requirements called for in the brief,
w e r e finally developed and c h e c k e d for
SECTION 3-3

n
structural stability and cost. S m a l l - s c a l e
models were made so that their aesthetic
merits could be assessed ( F i g . 1 ) .
0
These were: SECTION 5- 5
•'
(a) A circular cylindrical tower with a
circular multi-level turret on top
(b) A rectangular shaft with rectangular
platforms
SECTION 4-4
(c) A cruciform-shaped shaft with

(d)
octagonal platforms
Four separate circular shafts, one in the
0 000
v m-m
middle and three spaced at 1 2 0 ° around
( the perimeter, with circular platforms
(e) A configuration comprising four separate
rectangular shafts arranged in a c r u c i -
form pattern w i t h rectangular platforms
(t) A three-cornered 'star - s h a p e d shaft
(g) A configuration comprising a central
hexagonal shaft and three rectangular
shafts located at the a p e x e s of an
equilateral triangle, with hexagonal
platforms.
T h e simplest and cheapest solution from a
structural and construction viewpoint w a s the
circular shaft w i t h circular platforms. There
were valid aesthetic reservations about this
solution for a number of reasons. T h e tower is
not very tall, being about 1 5 7 m to the top
of the concrete portion and 187 m to the top
of the steel mast. T h e functional space
requirements for lifts, staircase and ducts
dictate the size of shaft, making it relatively
large in relation to its height and thereby
making it impossible to achieve the visual
appearance of a tall slender structure. More
important, the form did not relate w e l l to the
surrounding buildings or topography and
w o u l d not have created a distinctive a r c h i -
tectural statement for Pretoria.
The notion of separating the four shafts arose
from the functional requirements. Each shaft
serves a separate function, namely three
separate lift shafts plus the central core to
accommodate the staircase and service ducts
for cables, ventilation, plumbing and other
services. Moreover the brief called for
separation between the a c c e s s to the public
spaces and the technical areas. At the same
time it w a s decided that the public v i e w i n g
platform could be a c c o m m o d a t e d well
below the technical platforms on the grounds
that this w o u l d provide an adequate v i e w
over Pretoria.
A s a result of these factors it became feasible
to stop off each of the lift shafts just above the
level served, thereby giving a more distinctive
unsymmetrical form to the tower. After due
consideration of all the factors involved,
alternative ( g ) w a s c h o s e n . It w a s decided
to terminate each lift shaft just above the
highest level it served, namely the public
v i e w i n g platform, the technical service floors
and the top respectively. T h e result is a
distinctive sculptural form w h i c h reflects the
functional requirements of its various
elements, and is in harmony w i t h the pre-
dominantly linear appearance of the nearby
U N I S A buildings.
For structural reasons the three outer shafts
and the central core had to be inter-linked
at various levels to ensure shear transference
between the various vertical elements in
resisting the horizontal w i n d loads and
eccentric vertical loads. This inter-connection
is provided by a series of w a l l - b e a m s and
platforms at various levels w h i c h are inte-
grated into the functional requirements and
aesthetics of the tower.
Major horizontal cantilever platforms are
required at levels 6 and 9 ( F i g . 2 ) . Level 6
supports the public v i e w i n g platform and is a
tapering hexagonal slab w h i c h supports the
columns from the roof slab above on its
perimeter. Level 9 is a more complex structure,
being supported unsymmetrically on the
central core and the t w o remaining lift shafts
at that level and cantilevering some 8 m
beyond the central core. It carries the three
floors and roof of the technical service
building on c o l u m n s supported on its
perimeter. It is a concrete slab with a maximum
depth of 1.5 m tapering to 4 5 0 mm at the
perimeter.
Structural analysis
B e c a u s e of the unsymmetrical form of the
structure and the importance of the interaction
between the vertical elements through the w a l l
beams and platforms in resisting w i n d and
eccentric vertical loads, a s p a c e frame analysis
w a s necessary. T h e frame, ( F i g . 3 ) , had 1 8 6
Fig. 3 4
S p a c e frame configuration
Fig.4T
W i n d orientations and strain gauge positions
Medium
North
Low ^ S t r a i n gauge
positions
members and 84 joints and w a s analyzed for
various load cases, including many different
w i n d directions, using our i n - h o u s e G A 1 8 / 3 0
computer. T h e frame analysis w a s also used
during the conceptual design stage for a s s e s -
sing the required stiffness and strength of w a l l
beams and for determining the natural
frequency of vibration for the w i n d tests using
the R a y l e i g h - R e i t z method.
C P 1 1 0 w a s used as the basis for design and
load combinations.
The main cantilever platforms were analyzed
elastically using quadrilateral finite elements
with the program O V E F I N E to determine the
1
magnitude and directions of the principal
moments and to obtain the desirable arrange-
ment of reinforcement. The c h o s e n reinforce-
ment arrangement w a s then c h e c k e d by yield
line analysis.
W i n d loading
B e c a u s e of the tower's unusual shape,
information on w i n d loading co-efficients
and structural response to w i n d could not be
obtained from available s o u r c e s . It w a s
therefore decided to carry out w i n d tunnel
tests on an aeroelastic model of the structure
so that the force co-efficients as w e l l as the
aerodynamic excitation could be investigated
for various w i n d directions. T h i s testing
w a s carried out at the University of Pretoria,
under the direction of Mr. V. C h a s t e a u . 2
A model scale of 1 :250 w a s c h o s e n so that the
model w o u l d fit satisfactorily into the available
w i n d tunnel c r o s s - s e c t i o n of 1.3 m x 0.9 m.
T h e model w a s made out of 'jelugtong'
w o o d . All external dimensions were to
scale w h i l e the thicknesses of the concrete
in the shafts w e r e modelled as 1.25 times the
actual scaled thicknesses in order to give a
model stiffness w h i c h w o u l d lead to a
convenient model natural frequency and test
w i n d speed. Discreet metal masses were
introduced to model the mass distribution of
the tower correctly.
The f u l l - s c a l e natural frequency of oscillation
of the tower w a s determined from the
computer s p a c e frame analysis to be 0.22 Hz
and the torsional frequency as 0.39 Hz.
The damping of the model should be the
same as for the full-scale structure, and a value
of logarithmic decrement of 0.06 w a s
chosen as reasonable for good quality
uncracked concrete. T h e model w a s instru-
mented with three strain gauges on each lift
shaft at ground level to measure the vertical
strains at various points on the shafts. From
these strains the response of the structure for
the 12 different w i n d directions could be
determined. T h e positions of the strain gauges
are s h o w n in F i g . 4.
T h e natural frequencies of the model were
measured by t w o independent methods. T h e
bending oscillation frequency w a s found to
vary between 2 5 Hz and 3 3 Hz depending on
1
the direction of oscillation. T h e torsional
oscillation frequency w a s approximately
6 4 Hz. T h e logarithmic decrement for bending
oscillation w a s an average of 0 . 0 5 5 for all
directions.
T h e results of the w i n d tunnel tests w e r e used
to a s s e s s the d y n a m i c response of the
tower to w i n d as w e l l as to determine the
force c o - e f f i c i e n t s for the various w i n d
directions for use in the design. In addition,
pressure measurements w e r e taken to deter-
mine local w i n d pressures on elements of the
tower s u c h as the w i n d o w w a l l panels at the
public v i e w i n g platform and technical levels.
T o obtain the equivalent drag and lift c o -
efficients from the test measurements w a s not
straightforward as they had to be derived from
six independent strain gauge readings for
e a c h w i n d direction. T h i s w a s achieved by
a s s u m i n g that the w i n d velocity profile w a s
k n o w n and that the force co-efficient, Cp,
w a s constant w i t h height. A space frame
analysis of the model w a s carried out w i t h this
pressure distribution in the given direction to
determine unit strains at those points w h e r e
measurements w e r e made. B y fitting these
results the required co-efficients could be
obtained. T h e resulting force co-efficients
for the various w i n d directions are given in
F i g . 5. T h e average value of C F w a s found
to be 0.82 and this value w a s used in the
design. T h e s e results w e r e later verified by
direct measurement of the resultant w i n d
force by means of a w i n d tunnel balance for
T w o alternative foundation solutions w e r e
considered, namely a deep shaft under each of
the main shafts and a raft foundation. T h e
raft foundation w a s preferred on the grounds
of its lower cost, shorter construction time
and reduced construction risk. H o w e v e r there
w a s concern about possible differential
settlement and rotation of a raft because of the
weathered strata. Calculations based on the
percentages of weathered material observed
in the boreholes and on the a s s u m e d elastic
modulus of 4 0 0 M P a for the weathered
material, indicated a total vertical elastic
settlement of the order of 2 5 mm and negli-
gible consolidation settlement.
Differential settlements arising from this and
w i n d loading proved to be within acceptable
limits, the m a x i m u m calculated rotation being
less than 0.001 5 radians ( 0 . 0 9 ° ) .
It w a s desirable to check the elastic modulus
of the weathered material as this governs the
deformations. Laboratory tests on small
1,0
0,9
LL
<_>
c 0,8
samples were not feasible because of the
difficulty of sampling the weathered material
undisturbed. It w a s therefore decided to
carry out d o w n - t h e - h o l e tests using a
pressure meter.
Four NX boreholes w e r e drilled to depths of
b e t w e e n 26.4 and 29.1 m. T h e y w e r e
immediately filled with cement grout, left to
set overnight and then re-drilled the following
day. T h e pressure meter w a s a pneumatically
operated instrument w i t h a 7 5 mm diameter
probe w h i c h w a s buried into the borehole.
Tests w e r e carried out in the weathered
material only. Although the majority of the
tests failed due to breakages of the pressure
cells, the results obtained w e r e sufficient to
confirm that the a s s u m e d elastic modulus w a s
conservative and that the raft solution could
be adopted without c o n c e r n about differen-
tial settlement. T h e results of the tests
produced values of elastic modulus between
5 8 0 and 7 9 0 0 M P a .

various w i n d directions.
T h e s e results are also s h o w n in Fig. 5. T h e
0,7
centroid of the w i n d force w a s found to be
at about 7 9 m above ground level for all
w i n d directions.
For one particular w i n d direction ( 0 = 3 0 ° ) LEGEND
0,6
noticeable oscillation of the model occurred / / Most probable rang* of Cp
w i t h i n a small range of model w i n d velocity Cp determined with a wind tunnel balance (without mast)
around 1 3 . 4 m / s e c (approximately 8 6 k m / h r Weighted average Cp = 0,82.
full scale at the height from w h i c h the o s c i l l a -
tions w e r e found to emanate or 5 8 k m / h r at 4-
30 60 120 150 180 210 24 0 270 300 330 360
10 m reference height).
T h e amplitude of these oscillations w a s Wind direction 8*
halved for w i n d s p e e d 5 % higher or lower than
this value, and no oscillations w e r e evident Fig. 5 A
for w i n d directions other than 5 ° above or Variation of w i n d force
below 0 = 3 0 ° . T h e oscillations w e r e approxi- co-efficient with
mately a c r o s s - w i n d and had a frequency of w i n d direction
3 0 . 3 Hz w i t h an amplitude w h i c h varied with
time. Their cause w a s found to be regular I5tm
vortex shedding from the portion of the tower
above the public v i e w i n g gallery and below
A
the underside of Level 9. It w a s found these
oscillations could be controlled by the
introduction of aerodynamic fins within this
portion of the model.
H o w e v e r , in v i e w of the following it w a s \
decided that no s u c h precautions w e r e
necessary:
(a) T h e frequency of occurrence of the
\
disturbing w i n d velocity and direction is
less than once in 10 years.
(b) T h e m a x i m u m resulting acceleration on — 68m
the tower w o u l d be 0 . 8 2 % g, w h e r e a s
1 % g is considered to be the limit of
\
s
acceptability and 1.5% g disturbing.
(c) T h e amplitudes measured w e r e w e l l
w i t h i n the acceptable limits for the
structure. \\,\
No other oscillations w e r e detected up to a \\
N
m a x i m u m test speed equivalent to a full \
scale w i n d of 1 4 4 km/hr.

Foundations 0 0,1 0,2 0,3 OA 0,5 0,6 0,7 0,8

Boreholes drilled at the positions of each shaft Ratio R = F / F for
n 8 = 0*
indicated a s u c c e s s i o n of alternating layers
LEGEND
of intact and weathered shale d o w n to depths Fh Wind force on tower segment above h
of 3 0 m. Core recovery in the weathered F Total wind force on whole tower
shale w a s very poor, and it w a s estimated that Fig. 6 • —©— As measured in wind tunnel {without mast)
the percentage of weathered material in the Distribution of w i n d —H— Theoretical value (without mast)
upper 2 0 m below founding level w o u l d force with height o— Theoretical value (with mast)
vary b e t w e e n 2 0 % and 3 0 % for the four shaft
positions.
 m
The raft w a s 3.5 m deep and designed to
ensure that the centroid of the dead load
of the tower coincided with the centroid
of the raft w i t h a m a x i m u m bearing pressure
of 6 0 0 k P a under dead plus live load and
7 0 0 k P a w i t h w i n d loads included.

Construction
A number of special construction problems
existed, n a m e l y :
(a) T h e construction of the raft foundation.
(b) T h e method of construction of the i .1

vertical elements, their inter-connection 1

w i t h the w a l l - b e a m s and the control of
alignment and verticality.
(c) Control of the quality of concrete finish
on the vertical shaft elements and the
special off-shutter concrete finish
required on the low rise building at the
base of the t o w e r .
3

(d) T h e construction of the major horizontal

cantilever floors at Levels 6 and 9.
(e) Detailing a n d construction of the
window walls.
B e c a u s e of the special nature of the project,

l
the client adopted our recommended pro-
cedure of pre-qualification of tenderers.
In this w a y a short-list of invited tenderers,
fully qualified to handle the project and 1

briefed on the design intent, w a s obtained.

Raft construction
It w a s imperative that a high quality concrete . J
finish w a s obtained. T o achieve this, great
s
IB
care w a s taken w i t h the concrete mix design
and the choice of aggregates. T h i s w a s done 7
w i t h assistance from the Portland Cement =£•91
Institute. Test panels w e r e manufactured
prior to tender using various concrete mixes
to examine both the off-shutter finish and the
concrete colour. T o avoid c h a n g e s in the US"
colour of aggregates during the construction i I
period, the contractor w a s asked to stockpile
all the aggregate required for the project
prior to c o m m e n c e m e n t of construction.
Despite this precaution, the s a n d ran out
towards the end of the project and as the
supply at that quarry had changed colour
since the start of the contract it w a s necessary
to match the sand from elsewhere.
The 3.5 m deep raft involved some 1,400 m 3

of concrete w h i c h w a s cast in three separate

horizontally-layered pours.
The m a x i m u m depth of pour w a s 1.5 m.
Heavy reinforcement only occurred in the
bottom and top pours and could be fixed
only w h e n required, thereby avoiding the
need for e x c e s s i v e support for the top rein-
forcement. E a c h section w a s protected w i t h
polythene sheets for curing after casting.
Precautions w e r e taken w i t h the mix design
to minimize the heat of hydration by using a
5 0 / 5 0 combination of cement and slag and
large 4 0 mm aggregate to reduce the water
demand of the mix. Temperatures w e r e
measured by means of thermo-couples
through the depth of the raft after casting and
differential temperatures remained within
acceptable limits.
Sliding form work
The contractor proposed a sliding form of
construction for all the vertical elements
including the w a l l - b e a m s . T h e procedure w a s
to slide continuously to the underside of
each w a l l - b e a m , at w h i c h stage a steel soffit
girder w a s installed between the shafts and
fixed by means of welding to plates cast into
the w a l l s . T h e s e girders formed the soffit
of the w a l l - b e a m s and supported jacking
rods to enable the w a l l - b e a m s to be slid
concurrently w i t h the shafts after making
minor c h a n g e s to the shutter ( F i g . 7 ) .
T o a c h i e v e the required high standard of
finish on the vertical elements a steel shutter
w a s used w i t h j a c k i n g rods positioned to
miss the openings in the shafts. (Where
jacking rods had to go through openings.
Fig. 7
The partially completed tower s h o w i n g the sliding platform
temporary c o l u m n s were cast around the
j a c k i n g rods to give them the required
support). T h e sliding platform consisted of
three l e v e l s ; namely, the shutter level,
finishing level and the curing c o m p o u n d
application level, and w a s braced to ensure
rigidity during sliding. T h e sliding shutter
incorporated sections w h i c h linked the lift
shafts and w e r e used in sliding the w a l l
beams.
Before c o m m e n c e m e n t of the slide, t w o small
test slides were conducted using different
concrete mixes, and as a result the most
suitable mix satisfying both slidability and
finish w a s c h o s e n .
In early d i s c u s s i o n s with the contractor it w a s
decided to use reinforcement c a g e s in the
shafts tied together with the lighter horizontal
reinforcement. T h e height to w h i c h horizontal
reinforcement could be fixed w a s dictated by
the j a c k i n g frames w h i c h span the shutter.
The c a g e s were staggered vertically to enable
the reinforcement to be placed in a c o n -
tinuous sequence thereby avoiding a bottle-
neck in the fixing operations.
During the sliding of the deep w a l l - b e a m s it
w a s more difficult to control the surface finish.
A
The heavy w a l l - b e a m reinforcement w a s
placed prior to sliding, it being possible to fix
a maximum height of only about 9 0 0 mm at a
time due to the interference of the j a c k i n g
frames. T h i s portion w a s then slid before
fixing on the next could proceed. During this
stop-start procedure for sliding the w a l l -
beams, and in particular the 8 m deep beam
between levels 5 and 6 w h i c h took a total of
2 2 5 hours, care had to be taken w i t h the
concrete preparation at each stop to avoid
blemishes at the construction joints.
Where necessary, s c r e w couplers were used to
c o n n e c t the main reinforcement in the slid
concrete portion to the beams and slabs to be
constructed later. T h i s again meant a
s l o w i n g d o w n of the sliding speed and in
some cases a stop, and as a result special care
w a s necessary with the finishing.
The alignment of the slide w a s controlled
by means of optical plummets set in the
shaft corners at ground level. Readings
were taken on a target on the underside of the
sliding platform. External c h e c k s were carried
out using a s i n g l e - s e c o n d theodolite set up
over beacons near the base of the tower. In
the later stages a laser beam w a s also used, 1 5
 but this did not prove as s u c c e s s f u l as the 1

optical plummets because of the divergence,

shape and movement of the beam. A l s o , the
s h a p e of the t o w e r did not lend itself to the use
of a single laser. B e c a u s e of the rate of sliding
( u p to 2 0 0 mm per hour) the optical plummet
•
readings had to be taken every f e w hours.
E a c h j a c k i n g frame had a water level fed by a
central reservoir for levelling the shutter. B y
sliding w i t h the shutter slightly off level the
alignment could be adjusted. A lever device
w a s incorporated into the sliding shutter to
enable a force to be applied to adjust the
alignment of any shaft.
Difficulties w e r e experienced both in o b -
taining consistently reliable readings w i t h the
optical plummet and w i t h maintaining align-
ment, particularly because the balance and
response of the sliding shutter appeared to
be upset w h e n s u c c e s s i v e shafts stopped off,
resulting in a different shutter configuration.
Generally the alignment w a s controlled to

1
within 5 0 mm, but at one stage a deviation
of nearly 1 0 0 mm w a s detected on the
northern shaft. T h i s w a s gradually corrected
and is not visible and does not impair the
performance of the lift.
Fig. 8 A
Finishing of the concrete surface w a s done
The sliding platform
on the s e c o n d level of the sliding platform.
s h o w i n g girders,
Under normal sliding conditions the concrete
emerged from the underside of the sliding hangers, precast
forms relatively soft' (six hours o l d ) . T h e slabs and
surface w a s then rubbed d o w n using a reinforcement
w o o d e n float and brushed vertically to give it
an even appearance. After a stop in the sliding, i
the first concrete to emerge from the forms
w a s hard,-and to get an equivalent finish the

surface w a s first rubbed w i t h carborundum
stones and then lightly brushed w i t h a grout
made by sieving the coarse aggregate out of
the concrete used in the slide.
Due to the difficulties involved in getting
uniform distribution of water for curing, and
the possibility of freezing during the winter
months w h e n the majority of sliding took
place, water spray curing w a s not used.
Instead, a resin-based curing compound
(Cormix CM 90) w a s sprayed onto the c o n -
crete by hand from the lowest level of
the sliding platform. During the early stages
of sliding, the spray equipment w a s not
adequate to ensure e v e n distribution of the
c o m p o u n d and over-application occurred in
certain areas, leaving y e l l o w marks. T h e s e
have almost entirely disappeared w i t h time.
Cantilever platforms
Detailed d i s c u s s i o n s w e r e held w i t h the
contractor on alternative proposals for the
construction of the main cantilever platforms
at levels 6 and 9 and w e w e r e appointed to
design the temporary w o r k s on behalf of the
contractor.
T h e final solution adopted w a s a launching
system using precast concrete panels approxi-
mately 1.5 m square and 1 0 0 mm thick a s
permanent soffit shutter panels. T h e pattern of
joints between panels w a s detailed to meet the
architects' requirements. T h e slab at Level 9 is
technically the most difficult, being supported
unsymmetrically by the central core and the
t w o remaining lift shafts at that level and
cantilevering s o m e 8 m beyond the central
core. ( S e e F i g . 2 section 2 - 2 ) . It tapers from
1.5 m deep at the core to 4 5 0 mm at the edge
w h e r e it supports the c o l u m n s carrying the
technical floors and roof above.
A concrete a n n u l u s about 1.5 m deep w a s
first cast around the shafts using conventional
shuttering supported on steel brackets off the
vertical shafts. Radial structural steel c a n t i -
lever beams w e r e then attached above this
concrete and supported a series of trans-
verse beams at their extremities. T h e precast
soffit panels w e r e supported on the c o m -
pleted concrete on the inside and hung by
adjustable hangers, from the transverse
16 beams on their extreme edge. T h e joints
\
Fig. 9 •
Ground leve
off-shutter concrete
between panels w e r e sealed using a sealing
cord and grout. T h e reinforcement w a s fixed
also using the steel cantilever s y s t e m for
support and the s t o p - e n d s w e r e positioned
before casting the concrete. T h e s t o p - e n d s
w e r e located so as to allow the hangers to be
re-used.
O n c e the concrete had reached the required
strength, the hangers w e r e loosened and the
w h o l e system moved out using the n e w
concrete for support. T h e next row of precast
panels were then hung and the process
repeated.
T o finish the edge strip, steel shutters w e r e
supported from the cantilever s y s t e m . T h e
same system w a s used for both levels 6 and
9, there being four r o w s of panels and the edge
strip at Level 6 and t w o r o w s of panels and
the edge strip at Level 9. T h e initial in situ
a n n u l u s w a s more difficult to construct at
Level 9 because of the a b s e n c e of a shaft on
one side. T h i s w a s replaced by a steel
cantilever bracket in that position.
Due to the critical nature of the hanging
details, a f u l l - s c a l e load test w a s carried out on
one of these precast panels at the S A B S . T h e
girders, hangers, precast slabs and rein-
forcement are s h o w n in Fig. 8.
A n interesting facet w a s the detail developed
for the stop-end between pours. This
consisted of a lightweight frame covered w i t h
e x p a n d e d metal, w h i c h enabled the heavy
cantilever reinforcement to be placed simply
by cutting holes through it w h e r e required.
A certain amount of shutter oil w a s left on the
precast panels and some grout had leaked
from the joints. It w a s therefore necessary to
clean the underside of the slabs after c o n -
struction, using a lightweight structural steel
platform hung under the slab. After this
treatment a very satisfactory soffit finish w a s
achieved.
Off-shutter concrete
on the l o w - r i s e building
The finish required for the l o w - r i s e building
at ground level w a s to express the grain
pattern of s a w n timber boards on the
surface. T o achieve this, tongued-and-
grooved boards w e r e sandblasted to e n h a n c e
the grain pattern and then sealed. J o i n t s
w e r e sealed by means of compriband strips
against the timber m o c k - m a s o n joints.
The design mix for the off-shutter concrete
w a s developed in consultation w i t h P C I .
Strict site controls w e r e set up to monitor the
stockpiling and use of the various mix
constituents.
The details of the formwork for e a c h panel
w e r e d r a w n up by the contractor and s u b -
mitted for approval. T o gain a s many r e - u s e s
from e a c h form a s possible, the w o o d e n
m o c k - m a s o n joints w e r e made movable
on the form. This w a s a c h i e v e d by the use
of compressible sealing strips to seal it
against the 'rough' timber grain. T h i s meant
that, w i t h i n the restriction of the board
pattern, one form could be used for any size of
panel.
All construction joints had to fall on mock-
mason joints and to prevent grout leakage the
forms w e r e sealed onto existing concrete
using a compressible sealing strip on the edge
of the completed concrete. All details of
ferrule holes, expressed joints, colour of
concrete, concrete finish, etc. w e r e resolved
between the architect and the contractor
priorto starting construction. ( F i g . 9 ) .
T o ensure that the required quality of finish
w o u l d be a c h i e v e d , a number of sample
panels w e r e prepared prior to c o m m e n c i n g
construction and compared w i t h a test panel
w h i c h had been produced prior to tendering
and w a s specified as representing the required
quality.
Fig.10
T o w e r during construction

A n interesting aspect of the shuttering w a s the

method used for construction of openings for
w i n d o w s , etc. T o avoid the problem of
concrete not reaching the centre of the base of
an opening, small temporary c o l u m n s were
constructed w i t h i n the opening. T h i s a l l o w e d
placing and vibration of concrete in the
inaccessible areas.
Forms were generally stripped in under 24
hours w h i c h meant that if some patching
became necessary, it c o u l d be undertaken
w i t h minimal effect on the final finish.
These areas were patched using grout and a
sponge.
Water curing w a s used, the w a l l s being
sprayed w i t h water t w i c e daily. A total of
2,000 m of off-shutter concrete w a s cast
2

in the ground level structure.

Window walls
T h e w i n d o w w a l l s w h i c h occur between
Levels 6 a and 6, 9 and 1 1 a and 16 and 1 6 a .
consist of a series of structural steel mullions
hung from the upper levels and supported
horizontally at e a c h level, into w h i c h the
w i n d o w frame is fixed so that all glazing is
done from the inside of the building. A n
additional feature is the use of single glazing
and external w a l k w a y s , enabling the glazing
to be cleaned from the outside as w e l l as the
inside of the building and thereby eliminating
expensive pivoting w i n d o w frames.
Special care w a s taken with the specification
for this steelwork.
A m o c k - u p w a s made to study the details
and test the system against rain penetration.
A s a result of the tests, modifications had to
be made to meet the architectural require-
ments and still ensure adequate corrosion
protection. T h e specification adopted w a s
1 5 0 microns of bronze metal spray over 1 0 0
microns of zinc metal spray, sealed with a
clear acrylic sealer.

Credits:
Architects:
S o u t h Africa Public W o r k s Department,
Architectural Department
Quantity surveyors:
Southby, B i h l , Detert & Slade
Consulting electrical engineers:
Bidermann, Finn, Beekhuizen & Partners
Mechanical services:
South Africa Public W o r k s Department,
M e c h a n i c a l Department
Main contractors:
S t o c k s - F u t u r u s ( P t y ) Ltd. 2 - a consortium
between S t o c k s and S t o c k s ( P t y ) Ltd. and
Futurus.
T h e author is indebted to the many individuals
w h o formed part of the team w h i c h made this
project a s u c c e s s : firstly, to the relationship
w h i c h existed throughout between the
architect responsible, Mrs H. Lenddorff, the
consulting engineers, the P W D engineers
and the quantity surveyors, out of w h i c h it
w a s possible to create a tower of quality;
secondly, to those members of Ove A r u p &
Partners w h o participated tirelessly on the
project, particularly J a n Heynike for his
contribution to the development of the basic
concept and to the engineers w h o executed it,
Vincent Diesel, R o n Finkelstein, H. Scott and
the late Fred H o o g e n d i j k ; thirdly, to Mr V.
Chasteau for his contribution in advising on
and carrying out the w i n d tunnel t e s t i n g ;
finally to the contractors, and particularly
Dr. W. V a n c e , for the excellent relationship
w h i c h prevailed on site and the enthusiastic
manner in w h i c h challenging problems were
met and overcome.
Central Bank, Dublin
Architects: Stephenson & Associates
T h e A s s o c i a t i o n of Consulting Engineers of
Ireland has given its 1 9 8 0 Excellence A w a r d
for a structural project to the Central Bank,
Dame St. Frank Lydon of Ove Arup &
Partners Ireland, consulting engineers for the
project, accepted the A w a r d at the annual
dinner of the Association on 7 M a r c h .
T h e main building is an eight storey, 4 5 m
high structure and measures 4 5 m • 4 5 m on
plan. It is set towards the north of the site
overlooking a landscaped piazza fronting
onto D a m e Street. T w o ancillary buildings
are also incorporated on the site. One of these
a c c o m m o d a t e s the staff dining facilities. T h e
other is a reconstruction of a listed 19th
century building w h i c h stood on the site
prior to this development. There is a t w o
storey underground car park covering most
of the site to accommodate 1 3 0 cars.
T h e principal feature of the building design
of engineering interest is the superstructure
w h i c h is suspended from the top of t w o
reinforced concrete service cores. Other
aspects of the building construction, the use
of slip forming to construct the concrete
service cores and the lifting into position of
complete floors assembled at ground level,
are also unusual features.
— :

I T

•
 w ,

Pi

i
:
o

£ •