Paper: Slade/Playle

Paper

The British Museum: upgrading the floor over

the Kings Library
R.E. Slade, BSc (Eng), CEng, MIStructE
WSP Consulting Engineers

c.Playle, BSc, CEng, MICE

WSP Consulting Engineers

Synopsis
The planned move of the British Library to its new location at St
Pancras gave an opportunity to resolve a long-standing problem
at the British Museum. Thefloor overthe Kings Library,
constructed in the 1820s by Robert Smirke incorporates long-
span hogbacked iron castings made by John Rastrick. The
strength and soundness of Rastricks girdershave been the
subject of much debate over the years, and the permitted imposed
load on the floor has been severely restricted. The Museumwas
keen to apply higher loads, and this paper describes the historical
background, the investigative work, and the appraisal of the
existing structure, together with the design solution adopted.

Introduction
The floor overthe Kings Library in the British Museum was constructed
in the 1820s by Robert Smirke. It incorporates hogbacked iron castings Fig 2. Floor construction
made by John Rastrick, some of which span over 50 ft, and had been
subjected to a severe load restriction for many years. With the planned
Smirke, a Greekrevivalist, was attached to the O f i c e of Works and was
move of the British Libraryto its new location in St Pancras, andwith the
consequent replanning at the Museum, the Trustees felt that the doubts and known for his design of the rebuilt Covent Garden Theatre and the Royal
Mint. In February 1821 Smirke recommended the erectionof two parallel
uncertainties over the strengthof the floor shouldbe resolved once and for
wings to the north of Montagu House and his scheme for progressive
all. Consequently,they commissioned an investigation into theloadcanying
reconstruction was presented j n July 1823 to the Treasury and to the
capacity of the floor and required the floor to be justified for 5kN/m2 or,if
that was not possible,for 4kN/m2. parliamentary select committee which voted E40 000 for construction. Final
impetus for the redevelopment was given by the need to house a libraryof
Historical background 65 250 volumes and19 000 tracts which had been collectedby George 111.
George IV needed more space at Buckingham Palace and donated his
In 1754 the Trustees of the British Museum acquired a decaying French-
fathers collection to the Museum in January 1923. Work started later in
style 17th century mansion in Bloomsbury known as Montagu House,
1823 and was complete by 1827. The Kings Library is the main feature of
together with 7 acres of garden. In order to cope with the increasing
the eastwing; it is 100 yards long, 41ft wide and30 fthigh with four central
collections and with the disrepair of Montagu House the Trustees setup a
pillars of Aberdeen granite surmounted by Corinthian capitalsof Derbyshire
Buildings Committee in 1802 to plan the expansion of the Museum. The
Committee - Sir Joseph Banks, Sir William Hamilton, Thomas Astle and alabaster. It is consideredto be Smirkes finest
interior design, elaborately
decorated with wood and gold-leaf and one of F e most handsome rooms in
Charles Townley- adopted a solution to build three sides of a quadrangular
London (Fig 1). The first floor above the Library, constructed using long-
building northwards into the gardens, leaving the southern side until last
span castings made by John Rastrick, was originally intendedfor agallery
when the house could be demolished and replaced. The recommendation
of paintings but these were transferredto the new national institution. When
was made in 1803 and planning was begunin 1804 by the Trustees architect
the east wing was finished at acost of E l 30 000, the first floor was occupied
George Saunders, andwas taken over by Robert Smirke in 18 16. Robert
by the Natural History Collection.

Fig 1. The Kings Library

The floor over the Kings Library
In 18IO Smirke had usedbuilt-in castings over 30 ft long
at CirencesterPark,
but he was not confident tospan the Kings Librarywith beams castin one
section.,He preferred trussed girders composed of sections of cast and
wrought iron, and only changed his mind after meeting John Rastrickwho
was a leading exponentof cast iron.In Rastricks own words: A friend of
mine who was connected with me in the iron trade in London introduced
me to Mr Smirke. He showed me his design for making his trusses upon
which he meant to carry the floor. I found that he proposed constructing
them by certain piecesof cast iron andcertain ties of wrought iron.In fact,
it was a truss composedof cast iron and wrought iron. My observation to
him was this, that I should not recommend him to make use of the trussed
girder, as I considered they would be liable by the weight upon the floorto
vibrate andto give way, and that after they were put in the building he would
never be able to get to them to tighten themup again. He said, What would
you do?, I said, I would make themin castings of one piece. He seemed
rather surprised that I should venture uponso large an experiment, forthe
width of the room was 41 ft, and the bearinghe proposed giving them was

al The 12/20 73/No

193 June 1995
Paper: Slade/Playle
~-

-- I . in one solid piece) and described tests he had made on lin square bars in
connection with the beams at Buckingham Palace. Incidentally, this practice
had earlier misled Tredgold into seriously overestimating the tensile strength
of iron in large castings. Rastrick however relied on proving his beams by
load test and describes in the evidence his contrivance for carrying this out
(see Fig 3). By good fortune he illustrated his work usingdetails of the Kings
Library beams but this evidence in terms of proof load issomewhat confused
as at one point he mentions a test load of 40 tons, another describes the
carriage as carrying 30 tons, and he also states that the weight on the beams
was 15 to 20 tons. His method was to load the beam, to strike the beam all
the way along it with a 211b sledge hammer, to measure deflection and to
measure elastic recovery after the load was removed. Apparently only one of
the British Museum beams failed the proving tests. It was common practice
to proof load beams before they left the foundry (and sometimes when they
reached their destination); this in itself indicates the uncertainty that existed
about the quality of iron castings. John Hawkshaw was prepared to make
girders for spans 1OOft long but others were more circumspect. Joseph Locke
would never employ a flat girder of any kind unless obliged to do so. Rrunel
was called to the inquiry and it was well known that he considered cast iron
an uncertain material and had avoided using it in bridge beams and where
it could be subjected to tension. William Fairbairn was also cautious and
I . . I showed his understanding of secondary effects by his dislike of loading
flanges eccentrically and of having openings in the webs of girders.
In truth it was not only the Dee Bridge disaster that brought iron into
question. (The Dee Bridge was a compound girder of cast and wrought iron.
The bridge had been designed by George Parker Bidder although Robert
Stephenson was responsible for it as engineer-in-chief to the Chester &
Holyhead Railway.) There were a number of other failures, some where the
brittle nature of cast iron was a major factor:
-disastrous collapses oftwo mill buildings inthe Manchester area in I844
and1847throughfaultydesign,inappropriatelayout and local
overloading of floors;
2 ft at eachend I . Rastrick had been consulted on the cast-iron beams used
- collapse of the Inverythan bridge in 1882 due toa large honeycomb or
by Nash at Buckingham Palace and assured Smirke that in 1821 he had
draw-hole;
directed the use of 90 ft-long beams for the Stourport Bridge over the
- the failure of a cast-iron bridge girder at Norwood Junction in 1891
Severn.Smirkewaspersuadedandhisconceptofspacious,box-like
caused by a similar defect.
galleries, his vision of a modernised Grecian architecture, became more
and more reliant on cast iron. Onthepositiveside,castironwasplaying a significantpart in
The floor over the Kings Library comprises floor boards on joists encouraging British engineers to study the strength of materials including
spanning onto timber beams which in turn are supported by girders made safety factors, vibration, fatigue, buckling, creep, torsion, integrity, quality,
by Foster, Rastrick & Co. atStourbridge (see Fig 2). Most of the girders span secondary stresses and so on. Its drawbacks - mainly its lack of ductility and
around 40 ft but some in the central area span over 50 ft. They are beautifully lack of warning of failure - were also being recognised and may have led
shaped hogbacked castings with large perforations through the webs and to early disquiet concerning large castingsin buildings.
with lower chords which support arched wrought-iron fire plates. The floor At the Museum the first signs of doubt came on Easter Monday1837 when
has a timber carcased ornate fibrous plaster ceiling. Contemporary accounts the Museum opened on its first public holiday and filled with23 895 visitors.
suggest that the castings were of high qualityand produced with great care. The Trustees noted that the pressure upon the floors twoofor three rooms was
They were almost certainly transported smoothly to Londonby canal, but very great and called in Sir Robert Smirke for an opinion as to their
we have no knowledge of how they might have been moved from, say, condition3. It is not known whether Rastricks beams were considered suspec
Kings Cross basin to site or how they were hoisted into place. Many years later, in 1928-29, a Royal Commission on National Museums
& Galleries was convened.Following observations made in the final report
Cast iron and concerns overits safety on the structural condition of the British Museum, Herbert Gough carried
In May 1847 the bridge over the Dee near Chester collapsed under a out investigations on cast-iron girders removed from the Museum during the
passenger train killing five people and injuring many more. The accident reconstruction of the Egyptian rooms in 1931. His work is described in a
caused a furore in engineering circlesand the Government ordereda public paper published by the Institution of Civil Engineers in 1934.
inquiry (the first of its kind in the world) into the Application of Iron to In brief, Gough carried out tests to destruction using three girders similar,
Railway Structures. Manyof the great engineersof the time were called to but not identical, to Rastricks beams over the KingsLibrary. He tested the
give evidence and it is apparent that there were wide differences of opinion girders at the National Physical Laboratory by first jacking two of them
on the useand safety of cast iron. against each other in a back-to-back arrangement (Fig 4). The surviving
John Rastrick favoured the use of cast iron (because beams could be made beam was then used in a subsequent test with the third girder. Gough

194 The Structural Engineer/Volume 73/No 1 2 / 2 0 June 1995

 Paper: Slade/Playle

carefully simulated the transverse stability provided by secondary girders get some ideaof the quality of the iron, but in situ tests were considered
in the Museum floor and took other measures to ensure his experiments were impracticable.
as accurate aspossible.
Goughs findings were worrying. He calculated that the girders had failed Appraisal
at an extremefibrestress of 3.9todin (60N/mm) and3.2 ton/in The imposed load restriction on the floor was in theory 2kN/mZand the
(SONlmm) compared to ultimate tensile stresses on samples cut from the Museum required the floor to be justified for 5kN/m2 or if that was not
girders of up to 13.8 ton/in (210N/mm). He concluded that ...the main possible, for 4kN/m2.(It isat this point that the authors wish to acknowledge
girders were in a very unsatisfactory and unsound condition due to the the invaluable help and assistance of James Sutherland, FEng, a leading
presence of defects; these defects included the presence of cracks in the expert onhistoric cast-iron.)
tension chords, blow holes, wrought-iron stirrups in the brackets used for Five wayso f appraising the strength of the beamswere considered:
housingthesecondarygirders,andnodules of iron solidifiedfrom
- by examining records of proof loads
splashings prior to being entrapped as shots in the molten stream during
- by in situ load testing
casting. Due to somecombination of these defects, theeffective resistance
- by testing material and simulated components
of each of the two main girders tested to destruction was very seriously
- by considering records of past use
reduced, the principal factors being the presence of cracks and blow holes.
- by calculation based on recommended and widely accepted design
In relation to thesafety of similar girders still existingin the museum floor,
criteria
these cracks are particularly disquieting, as experience gained in the present
tests has shown definitely that the detection of such cracks by careful visual
The proof loading record was at first sight very comforting. Rastrick
inspection cannot be relied on, even when the girder is subjected to con-
stated in his evidence that he had tested every single beam and therefore the
siderable load.
confusion about the magnitude of the test load may not in itself be very
The cracks hehad detected had extended overmore than one-half of the
significant. A more important point concerns the central load which proved
tension chords. The cracksurfaces were found to bein a very rusty state and
the beam for twice its value if it were uniformly distributed. This applies to
according to Gough had therefore been open to the weather for some
a beam of uniform section, but Rastricks girder equates more closely to a
considerable time. He was convinced that these cracks could not be revealed
Vierendeel truss,and the central pointload certainly does not give equivalent
by visual inspection either in the test girders orin girders still in position at
stresses to uniformly distributed load on all parts of such a girder. Rastrick
the Museum or elsewhere.
had made the openings probably to save dead-weight, to make handling
In a reports to the Museum dated December 1932 ( 2 years bejbre Goughs
easier, and possibly in an effortto control shrinkagecrackingduring
paper was published) Ralph Freeman of Sir Douglas Fox &r Partners called
production. He may also have considered that they looked rightand as such
into question the results from the tests at the National Physical Laboratory.
were an important part of his carefully integrated design. Whatever the
He argued that the girders failed at such a low stress level that it was
reason, they played an important part in the investigations described here.
remarkable that they had not given way in service under live load. His
In situ load testing was ruled out on the grounds that it would have been
contention was that each girder must have sustained some injury which
extremely disruptive and expensive,especially if one ormore of the girders
diminished its strengthduring its removalfrom the Museum. Freeman noted
did not stand up to the test.
that the test girders had been left in an open space for somemonths before
Testing material and components had its attractions - in particular this
being sent away for tests. He showed that rusting could occur on crack
might have concentrated on simulating the top chordwhere it formsa strut
surfaces within weeks and he also noted that the crystalline fracture surfaces
over the openings in the girder. The analysis showed this to be a critical
on Goughs beams were similar both for the rusty area and for the recent
element, as will be explained later.
bright part - suggesting that the cracks which undoubtedly existed before
Considering thepast record of use is important. The beams havebeen in
the test took place were not the result of, say, shrinkage cracking during the
service for over 160 years and there has been no trouble with them except
casting process, but had occurred during removal or subsequent handling.
that load restrictions have been applied. One argument may be that they
Apart from some areasof the Museum where Freeman found seriously
would serve thenext 160 years equally well, but the pessimistic view is that
under-strength girders (now long since removed) his advice regarding the
something may happen tomorrow. Both arguments areequally unscientific
first floor galleries, which presumably included the floor over the Kings
- it isbetter to look at thepresent, the opportunities that may exist to dispel
Library, was not to dismiss thepossibility of weaknesses but to weigh this
worries and doubts, and to look at the likelihood of change of use inthis part
against a record of good service for overa century. Heproposed that mea-
of the Museum that may lead to sudden or significant change of loading.
sures should betaken to ensurethat future loads on thefloors would be less
Appraisal by calculation proved more fruitful. The beamswere modelled
than those sustained in the past (unfortunately Freeman does not quote
as Vierendeel trusses in order to take account of the holes in the web. Whilst
figures forpast or futureloadings).
this is probably closer to reality than treating the girder as a solid beam, it
More doubts of a different kind were expressed duringa discussion7 on
is still only an approximation and, because of the width of the individual
a paper6 by S. B. Hamilton published by The Structurul Engineer in 1949
members, does not properly model the overall stiffness. Nevertheless, it was
when it was suggested that old castings were rather weak in lateral buckling
felt that this approach was far easier tohandle and much less long-winded
and that the perforations in the web might have converted Rastricks deep
than a finite element analysis. Estimating the loading and calculating the
girders into twoshallow beams with quite inadequate chords,a point that
resulting stresses was relatively simple but selecting permissible stresses and
will be returned to.
appropriate safety factors was far more complex. The following documents
were considered;
Investigation
The initial brief from the British Museum was deceptively simple: to (1) The LCC (General Powers) Act of 1909. This limits working stresses
investigate and report on the loadcarrying capacity of the floor over the in cast iron to 1.5 tondin(23N/mmZ) in tension and to 8 tons/in2
Kings Library. With the impending moveof the British Library to its new (124N/mm) in compression and is thought to be quite restrictive. GLC
site in St Pancras, releasing 40% of the Bloomsbury space tothe Museum, Bulletin-No. 91 (2nd Series) No.7 (1 976) is based on the LCCAct of 1909.
it was felt by the Trustees tobe a once-in-a-life-time opportunity to inves- (2)
CIRIA Report No. The structurul
renovation of truditionul
tigate the floorand to remove theload restriction that had been inplace for buildings 1986. This quotes someultimate strength figures froma book by
many years. As well as searching the historical record, several inspections Twelvetrees of 1900 but doesnotaddresstheproblem of structural
were made on site. appraisal.
Site workbegan by lowering oneself througha manhole into the guts of (3) Historical structural steelwork handbookgives some guidance on cast
the floor, then crawling on all fours over arched fire plates and through iron but for struts recommends Goodmans formula with limits which
openings in the girders. Despite the dust, gloom and heat it was still preclude its use here.
possible to appreciate the beauty of these magnificent castings and to (4) ICE Paper No. S418 of 1944 (Chettoe et ul). Based on tests on bridges
obtain sufficient dimensional information for analytical purposes. The it recommends a working strength in bending of 2.5 ton/in (39N/mm).
tension chords, however, couldnot be inspected in this way because they ( S ) Ministry of Works internal instructions of 1952 permits a bending
lie below the arch-plate bearings. The only means of viewing them was strength of 2.75 ton/in (42.SN/mmZ)in beams reducing by 0.25 ton/in2 (3.9)
by using an endoscope - at that stage it was not considered appropriate to N/mm) each forholes in webs, stiffeners and brackets. All three reductions
open up the floor either by removing fire-plates from above orby removing apply to the beams over the Kings Library leaving an allowable stress of
small areas of ceiling from below. Onetiny sample was taken in order to 2 ton/in (3 1N/mm).

The Structural Engineer/Volume 73/No 12/20 June 1995 195

Paper: Slade/Playle

(6) Department of Transport Standurd BD21/84 of 198413. Uses a sliding dependent on theirend fixity. Unfortunately, the box seatings which support
scale from 1 .S todin (23N/mm2)to 3 ton/in2 (46N/mm2)depending on the the secondary timber beams provide only limited fixity. An attempt was
proportion of live to dead load. It quotes the Rankine-Gordon formula for made to show that the web connection between the top beam and the tension
columns usinga factor of safety of 5 (the Ministryof Works document uses chord was torsionally stiff enough to provideend fixity tothe top chord strut
a factor of safety of 3). (see Fig 6). This demonstrated, however, that the strut was imperfectly
A decision was made toproceedwith the recommendations of the restrained in direction at each end and that its effective length should be
Department of Transports Standard BD 21/84 mainly because it is the estimated at only slightly less than unity. Regrettably, and much to the
most up-to-date and most widely accepted. It was felt, however, that the authors surprise, the top chord could not be justified for a live load of
safety factor for struts, although no doubt appropriate for bridges, was greater than 2kN/m2.
perhaps unnecessarily restrictive in this situation. To summarise, the minimum acceptable live load of 4kN/m2 would
Fig 5. is a flowchart showing the investigation and appraisal process. produce high stressesin the tension chordsand unacceptably high stresses
The bottom tension chord just about metthe stress limit given
in BD2 1/84 in the top chords. These facts together with the prospectof a definite change
with a live loadof 4kN/m. However, the top chord is extremely narrow and in the use of the floor, and the still unresolved possibility of flaws in the
as a result of the long slotsin the webs is effectively
a strut witha very high girders, meant that there was little option but to recommend that this once
slenderness ratio. The buckling characteristics of these struts are highly in a lifetime opportunity of strengthening the floor should be accepted.

LOOR USE STRATEGY ?

I
Survey Desk study
Defects Corrosion Dimensions p
Materials
behaviour
-
4 Materials tests

MP1 Dead load

I
Load Ukrason.ics test 1
Mathematical analysis
Kentledge
X
I l
I
Not practical
X
]-[ <--------- 3
I
Expensive, disruptive
X
I

Calculatestresses
permissible Calculate
stresses
applied
I I 1

I Permissible compressive
stress + Permissible
stress
tensile
rl
Geometry, loads

LCC (Gen.Powers) 1009 X Vierendeel web Solid

ClRlA report 111 X

I
GLC Bulletin No. 91 X
I X
ICE paper No. 5418,1944 X
restraints
MOW insructions 1952 X
Handbookhistorical
of X
sections - Euler ? *
Goodman ?
Gordon - Rankine
Long winded

Critical

DOT B2 1/ 8 4

v
Reassess F o f S

Permissible stress
I

1
Actual stresses

Key:
Reserve strength MP1 = Magnetic particle test for cracking
from end fixity, = Compression strut formula
arching etc. X = Discontinued line of investigation
I )= Todecision making process I

Fig 5. Investigution and appraisalflow-churl

196 The Structural EngineerNolurne 73/No 12/20 June 1995

 Paper: Slade/Playle

Original
arrangement 1 arrangement
First
4 m
(symmetrical about c) T (symmetricat
about
I

I i
l
$ New truss

Solution
Several options were considered for strengthening the floor. The intention
throughout was to leave Rastricks superb castings intact and untouched. Fig
7(a) showsan early scheme where the girders are strengthened by reducing
the slenderness of the top chord struts. This is achieved by clamping the
struts betweennew timbers. Fig 7(b) shows a similar scheme where, instead
of using timber clamps, the slenderness is reduced by inserting new timber
struts midway between eachof the existing beams. Bothof these schemes
had the disadvantage of tightening new material against the cast iron with
the inherent risk of inducing additional stresses into the castings. Neither I

scheme addressedthe problem of high stresses in the bottom chords. Fig 7(c) Fig 8. Twin trusses,first scheme
shows a series of the rodswhich are tightened against the weight of the floor
turning the castings into latter-day compound girders. This scheme had
potential for reducing stresses in both the top and bottom chords but was that the best approach was to provide new spanning members placed one
quickly abandoned because of the difficulty of providing suitable anchor- either sideof each casting. Fig 7(d) shows universal beams used in this way
ages for the tie bars, and because the scheme might produce high local but they would have been heavy and clumsy and would have interrupted
stresses in the old castings. service routes through the floor. Instead, the universal beams evolved into
All options involved opening up the floor, and eventually it was decided light Warren trusses pickingup the timber floor beams at node points.At
first it was believed the best solution wouldbe to insert these piece-small
and then transfer the weight of the floor onto the trusses
by cutting the ends
of the timber beams. Thisvery clearly reduced the stressesin the cast iron;
the trusses would be detailed to provide lateral restraint to the girders,
leaving the girders to support only the ceiling load (see Fig 8). However,
detailing the trusses around the girders so that they provided only lateral
restraint, but allowed no load transfer, presented practical difficulties. Slip
(a) Top chord clamps joints would work in theory but might easily malfunction in service. Apart
from these problems it became apparent, during the process of obtaining
listed building consent, that English Heritage disliked theof idea separating
the timber beams from the castings.In its view the timber beams were an

Original arrangement 1 Final arrangement

4
(symmetricat about $ 1

I
(symmetrical aboutQ )
(b) Top chord struts
($ New truss
I
r beam i
(C) Cable supports

(d) Twin universal beams

(4 Twin trusses

Fig 7. Strengthening options Fig 9. Twin trusses,final scheme

The Structural Engineer/Volume 73/No 12/20 June 1995 197

Paper: Slade/Playle

adjusting node stiffnesses,it was shown that the ratio of total truss stiffness
to girder stiffness was 1:l. A prototype truss (Fig 11) was tested with
simulated bearings and practical node showed that the stiffness was some
50% higher, and this was probably also true for the girders. Another
important feature of the strengthening works is the bolted connection
between the timber beams, the new trusses and the girders. If, due to
overload or accidental loading, a girder failed catastrophically, its weakened
or broken parts would be safely supportedby this interlinking.

Construction
A detailed measured survey was undertakenby surveyors working within
the floor priorto the contractor being given full possession of site. the This
proved to be extremely valuable in planning the work and making sure the
new trusses would fit into the limited space available. Local variations
within the layout of the original construction could also be taken into
account. Twomodels of partsof the floor were builtby the design teamto
demonstrate the proposals to all interested parties. Studies were alsocamed
out into site access, security arrangements and site safety. Safety was of
particular concern because work was to take place over areas open to the
public. Dust and noise control were also very important because of the
proximity of the Library andits priceless possessions.
The issues of buildability, safety, and control of dust and noise, were
Fig IO. Detail offinal scheme and study model addressed in the contractors method statement and agreed with the design
team. Methods of installing new trusses were developed so that the risks of
sudden or heavy loads being applied to unstrengthened floor structure were
integral and fundamental part of the structural system which consisted of eliminated.Trialerectionswere camed outtoensurethatnoonsite
joists, timber beams, castings, floor arch plates and ceiling and, as such, modifications would be necessary. Debris netting was positioned under the
should remain unaltered. ceiling and dust and dirt was minimised by the use of industrial vacuum
After lengthy discussion the scheme was modified so that the new trusses cleaners. A protective tunnel was contemplated within the Library but, afte
and the castings were locked together. This changed the scheme into a careful inspection of the rigidity and stability of the floor arch plates, this
loadsharing or helping hand system and had the great advantage from a was deemed unnecessary.
conservationists point of view of turning it into a fully reversible proposal. Work was carried out in two phases, and very few problems were
In other words, if at some point in the distant future it was decided that theencountered. The south and central areasof the floor were strengthened first
1990sstrengthening should be removed, for whatever reason, this could be and the north area followed later. Phase I started in April 1993 and phase
done leaving the original structure in place and unaltered. The final scheme I1 was completed by July 1994. Higgs & Hill Ltd was the main contractor
is shown in Figs 9 and 10. The timbers are bolted down to the girder with and MCL the steelwork subcontractor. Eachbay between two girderswas
bolts passing through very convenient empty holes in the seatings. The opened up in turn and the timber beams were carefully prised out of their
new trusses also support the ends of the timber beams by means of bolted seatings to allow the new trusses to be installed. The truss sections were
connections, andso the new trusses and old castings are neatly tied together. manhandled into a position vertically above their final location and spliced
Stiffness compatibility between new and old members was examined in together at midpoint. The old timber beams, which now have a reduced span
considerable detail. Light new trusses were desirable for obvious build- (from new truss to new truss instead of from girder to girder) and are
ability reasons but there was concern, that, once installed, they would attract therefore capable of supporting the higher imposed load, were threaded
toolittle loadfromRastricksgirders.Infact, by improvingthe through the trusses and the whole assembly was then lowered into the floo
mathematical models of both the girder and the new trusses, primarily by (see Fig 12). Care was taken at all times to ensure that the castings were

Fig I I . Test rig

Fig 12. Installation of new trusses

198 The Structural EngineerIVolume73/No 12/20 June1995

 Paper: Slade/Playle Informal study groups

provided with adequate temporary lateral restraint. Verticaland horizontal

fire barriers were also built into the floor. This requirement was another
good reason for carrying out extensive works at this time, especially
in view
Informal study groups
of the fire that occurred at Windsor Castle in November 1992.

Conclusion The purposeof the study group scheme is to create opportunities

for
The floor over the Kings Library been has the subject of much debate since members of the Institution to exchange ideas and workon deepening
the day John Rastrick persuadedRobert Smirke touse his great castings. The and developing their knowledgeof structural engineering, thus
strengthening works now complete have also been discussed and debated of
stimulating a greater interest in and promoting the art and science
at great length in an effort to arrive at sound judgment and sympathetic structural engineering.
solutions. The authorswish to record their gratitude for the invaluable help Members wishing to take part in the workof a study group or who
and support of the British Museum Architectural & Building Services require further information about a study group should write to the
Department and for the supportof the Trustees of the British Museum. appropriate Convener.

Acknowledgements
Client: The British Museum History of Structural Engineering
Project management and site control: British Museum Architectural & Convener: Frank Newby,MA(Cantab), FEng, FIStructE, HonFRIBA,
Building Services 27 Mayfield Avenue, London W4 IPN
Structural engineer: WSP Consulting Engineers vormerly Kenchington The Structural Engineer, March 1973, I p10
Ford)
Architect: The Devereux Partnership
Specialist consultant: Mr R.J. M. Sutherland Model Analysis as a Design Tool
Convener: F. K. Garas, PhD, CEng, FIStructE, MICE,
References Taylor Woodrow Construction Ltd., Taywood House, 345 Ruislip
1. Evidence of John Rastrick, Reportof The Commissioners Appointed Road, Southhall, Middlesex UB 1 2QX
to Inquire into the Applicationof Iron to Railway Structures, HMSO The Structural Engineer, February 1977, p63
1849, ~ ~ 2 8 5 - 1 9 1
2. Berridge, P S A: The Girder Bridge, Maxwell, 1969
3. Caygil1,M: The Story of the British Museum, BMP Ltd,1981
4. Gough, H J: Tests on cast-iron girders removed from the British Qualitative Analysis of Structural
Museum Institution of Civil Engineers selected engineering papers, Behaviour
No. 161, 1934 Convener: D. Johnson,BSc(Eng), PhD, CEng, FIStructE, MICE,
5. Report for the British Museum, Freeman, R: British Museum first Department of Civil & Structural Engineering, Nottingham Trent
floor galleries - strength of cast-iron girders, Sir Douglas Fox & University, Burton Street, Nottingham NG1 4BU
Partners, December 1932 The Structural Engineer, November 1978, p309
6. Hamilton, S B: Old cast-iron structures The Structural Engineer,
XXVII, April 1949, p 173.
7. Old cast-iron structures, discussion of Mr S. B. Hamiltons paper,
The Structural Engineer, March 1950, p79 Computing in Structural Engineering
8. The LCC (generalpowers) Act, I909 Convener: A. T. Humphrey,CEng, MIStructE, MIMechE,
CIRIAReport No. 1 11 Thestructuralrenovation of traditional
Analysis & Test Division, GEC Research, Marconi Research Centre,
9.
buildings, London, CIRIA, 1986 West Haningfield Road, Gt. Baddow, Essex
10. Bates, W: Historical structural steelwork handbook, BCSA Ltd, April The Structural Engineer, March 1987, p83
90 p57
11. Chettoe, Davey, Mitchell: lnstitution of Civil Engineers papers, No.
5418, 1944 Advanced Composite Materials and
12.
13.
Internal instructions, Ministry of Works, 1952
The assessmentof highway bridges and structures standard BD 21/84,
Structures
Convener: P. R. Head, BSc(Eng), ACGI, CEng, MIStructE, MICE,
Department of Transport, 1984 Maunsell Structural Plastics Ltd., Maunsell House, 160 Croydon
Road, Beckenham, Kent BR3 4DE
The Structural Engineer, Part A, June 1987, p22 1

Management and Maintenanceof

Bridges
Convener: G. Davison,BSc, CEng, MIStructE,
c/o The Institutionof Structural Engineers, 11 Upper Belgrave Street,
London SWI X 8BH
Structural news, 23 January 1990,p4

Arch Bridges
Convener: Professor c . Melbourne, BEng, PhD, CEng, MIStructE,
Bolton Institute, School of Civil Engineering and Building, Deane
Road, Bolton BL3 5AB
Structural news, 15 March 1994, p3

The Structural Engineer/Volume 73/No 12/20 June 1995 199