Forth Bridge
See also: Forth Road Bridge.
General
The iconic Forth Bridge is a balanced cantilever railway bridge over the Firth of Forth in the east of Scotland. It is often called the Forth Rail Bridge or Forth Railway Bridge to distinguish it from the Forth Road Bridge.
It was opened on 4 March 1890, and spans a total length of 2,529 metres (8,296 ft).
1873 A company was incorporated to build a suspension bridge across the Forth, designed by Thomas Bouch. This got as far as the laying of the foundation stone and the construction of part of one pier (which survives today, as the base for a warning light - just visible at the foot of the Inchgarvie tower in Photos A1 & A3 below, and seen more clearly in A21). However work was stopped after the failure of another of his works, the Tay Bridge, during a storm. The public inquiry into the Tay bridge disaster showed that he had under-designed that structure. He had made inadequate allowance for wind pressure. He had not exercised sufficient control over the construction or maintenance. Numerous deficiencies were found in the cast iron foundrywork.
1882 Fresh powers were obtained, and the project was handed over to John Fowler and Benjamin Baker, who were responsible for designing the bridge. On 21 December 1882 the contract for construction was let to Tancred, Arrol and Co (later William Arrol and Co). Baker and his colleague Allan Duncan Stewart received the major credit for design and overseeing building work.
During its construction, over 450 workers were injured and 98(?) lost their lives.[1]
1890 The leading engineering event of the year 1890 was the formal opening on March 5th of the Forth Bridge, 1.5 miles long,- the main feature of which is the two great central spans of 1710ft each. This great work was the subject of a supplement to The Engineer. The new railways that joined up the bridge were brought into use on June 2nd.
The 4 railway companies involved in the project were the North British Railway, the Midland Railway, the Great Northern Railway and the North-Eastern Railway[2]
Other openings in 1890 were :- Chester-Dee Bridge, on March 31st ; Edington-Bridgwater, July 17th ; Halifax, High Level, Railway, September 5th. In view of its recent reopening after reconstruction, it will be interesting to recall that the City and South London was brought into use, as between King William-street and Stockwell, on December 18th, 1890. It was the first electric railway in London. The London and South-Western ran over its own metals from Lydford to Plymouth from June 1st.[3]
1908 The line is 4.25 miles in length. [4]
Aspects of Design and Construction
A1, A2 etc refer to photos further down the page, mostly taken in 2016.
Seen from close quarters, the bridge is overwhelming in its size and complexity, and the structural principles are difficult to comprehend (Photos A1, A2). However, a broadside view, from the pedestrian walkway of the road bridge, makes the design principles much clearer (Photo A3). Viewed in this way, it is possible to consider which are the ‘primary’ members (whose absence or failure would cause immediate collapse), and which are the ‘secondary’ members which contribute to the strength and stiffness.
Note: A close-up view of the railway bridge is best obtained from North Queensferry, which is served by frequent trains. Alternatively, the road bridge is served by a bus route.
Close study of the railway bridge begs many questions, and fortunately most of the answers can be found in Wilhelm Westhofen’s superb account of the design and construction[5].
Westhofen was one of the engineers engaged in the construction, and his detailed account was directed more at engineers than laymen, but there are fascinating facets to be gleaned by all readers, with the aid of numerous drawings and photographs. This source provided the basis for later books aimed at a wider audience, for example 'The Forth Bridge – A Picture History' by Sheila Mackay.[6]
Westhofen’s account gives an insight into how construction work benefited from late 19th century advances, but we learn that, for example, lighting technology was clinging on to the dark ages. Night shift working was essential, but arc lights proved very unsatisfactory, producing a combination of dazzling light and impenetrable shadows, and having a tendency to go out without warning, while the alternative creosote-burning Lucigen lamps sprayed the area with unburnt oil, an obvious hazard for those clambering on the girders.
Moving on to the design of the bridge, the basic cantilever principle of the main spans was illustrated in a well-known contemporary photograph (Photo A4), taken from Westhofen’s book.
Two cantilevered arms project from each of the three central towers. Photo A5 shows that these towers comprise tubular frames with diagonal bracing. Note that the middle tower (RHS in photo A3) is wider than the other two. It also contains additional open lattice girders. The extra width was specified to prevent any tendency of the tower to tip (i.e. to lift the rear feet off their piers) in the case of heaviest loading (when two trains are crossing a central girder at the same time). In the case of the other (landward) piers, any tendency to tip was addressed by having a counterweight at the ‘land’ ends of the cantilevers (see photos A6 & A7). This corresponds to the stack of bricks in Photo A4. The actual counterweights are steel boxes with arched openings through which trains pass, the boxes containing about 1000 tons of iron blocks and scrap, all set firmly in asphalt. These weighted ends rest on rollers within the pier, to accommodate expansion.
The extra length of the centre tower was accommodated by piers founded on the island of Inchgarvie, which, as Westhofen wrote ‘Providence has so kindly placed in the middle of the Firth.’
Photo A8 shows one of the cantilevered sections. Each of these is 680 ft (207 m) long. The arched LOWER member is tubular, to provide strength and stiffness for loading in compression. It may look slender in the photo, but it has a maximum diameter is 12 ft (3.66 m). The UPPER members of the cantilevers are of lighter construction (lattice girders), being loaded mainly in tension and not required to resist buckling.
Note that both the upper and lower members taper down in section from the tower to the end of the cantilever. The lower member, as well as tapering down, also changes from circular to rectangular section as it approaches the end of the cantilever. This can be seen in A7 and A9 (A9 also highlights how the girders come closer together towards the centre central girder).
Between the upper and lower cantilevered members there are large tubular diagonal STRUTS, and diagonal TIES of lighter construction. These are not parallel to each other, but are radially disposed. Each portion of the bridge defined by a strut, a tie, and a top member and bottom member is called a BAY.
Additional latticework members can also be seen in A8. These support the railway deck and provide cross bracing (wind bracing) between the main members. Note that the railway deck girder becomes shallower as it approaches the end of the cantilever, presumably because its supports could be placed closer together.
Photo A10 shows one of the centre span girders. Apparently small when seen from a distance, it is in fact 350 ft long, and quite tall (note the size of the train in the photo). Although the centre girders appear to be relatively simple structures, a considerable amount of thought went into the way they connect with the cantilevers, and into the method of construction.
The connection of the centre girders to the cantilevers is such as to allow appreciable relative movement to occur due to thermal expansion, and also to permit some sideways displacement of the cantilevers due to wind forces. However, although the end result was a central girder which is able to move relative to the cantilevers, this freedom of movement only came towards the end of the girder’s construction. The central girder was not built as a unit and then connected to the cantilevers, but instead each half of the girder was progressively built out with rigid connections to the adjacent cantilever. In other words, each half was built as though it was part of its neighbouring cantilever. The length of each half was gradually extended until they met in the middle, where they were made to support each other via temporary sliding joints. Then, in a critical operation, the halves were temporarily joined and the rigid connections to the cantilevers were severed. In fact one of the connections severed itself with such a shock that people on and near the bridge feared imminent failure, or, as Westhofen put it, causing ‘some little commotion among the men’.
Westhofen’s description of the centre girders’ expansion arrangements, and of the process of connection the halves and severing the links, a process involving critical timing, a careful watch on the ambient temperature, and heating of girders, makes interesting reading, but does not lend itself to paraphrasing here.
Close examination of the structure in the area of the towers (e.g. Photo A2) hints at the massive task which faced the designers and the builders, not to mention the surveyors. A moment’s reflection will suffice to speculate on the potential for things to go wrong.
The designers’ starting point was a very wide expanse of water (more than 1 mile), unlimited amounts of steel plate and rolled sections, and millions of rivets, with much of the work to be progressed in spirit-sapping weather, at great height or great depth. To add to the difficulties of constructing such long cantilevers, careful account had to be taken of factors such as the force of the wind, tending to push the cantilevers out of line, and the effect of the sun passing from one side to the other causing the cantilevers to alternately bend to the east or the west. The magnitude of the achievement is underlined by looking at Photo A1, and reflecting on the difficulty in ensuring that the ends of all those massive tapered arched tubes would arrive precisely at their intended place in space, noting that there were twelve such tubes, all having to end up in precise alignment with their neighbours and with the approach viaducts, and having to be built out at the same rate, to maintain balance.
Moving back to details, and referring now to Photo A11, this complicated structure, where the five largest tubular members and five lattice girders come together, is known as a ‘skewback’. Another complicated junction was required at the top (A12). Away from these junctions, construction of the tubular section is relatively straightforward, but the limited size of steel plates meant that numerous riveted connections were needed. Also, it was necessary to progressively reduce the diameter of the section as the bottom tubes projected outwards. Clearly a major challenge for the designers and builders was the method of connecting each group of five tubular members together at the skewbacks. Note the transition from round tubes to square, and from ‘round-cornered square’ to square. See also A13, noting the relatively slender strip riveted on the corners, starting with slight curvature at the top, and becoming right-angled further down. A wealth of complexity is hidden within the skewbacks, there being numerous internal stiffening members.
The lower (arch) tubes reduce from 12 ft diameter at the skewbacks to about 3 ft rectangular at the ends, while the plate thickness reduces from 1¼” to ¾”.
All the towers are seated on granite-clad piers, but in each case only one of the four legs is firmly fixed. The other three have limited scope to slide and rotate on lubricated steel plates. This is to provide for expansion, and also to allow for rotation (as viewed from above) under the action of wind forces and solar heating. This most rigid of bridges incorporates a degree of articulation.
Photo A14 & A15 show how the main compression member changes direction to give the ‘curvature’ of the arch, and how the diagonal struts and ties are connected to this member. A8 shows a similar junction, with the addition of a rail deck support, viewed from inboard, i.e. from between the girders.
A16 shows the connection between a support for rail deck and the ‘arch’.
In Photo A17 we see rows of rivets which do not connect the curved plates to each other, indicating the presence of internal components. Photo A18 shows that there are internal stiffening rings and longitudinal beams. These appear as ‘H’ section beams, but in fact they are riveted together from 'T' and angle sections.
It might be expected that the curvature of the plates would be obtained by rolling, but in fact they were hot pressed, and then cold pressed. Each circular tube was assembled from ten curved plates, partially overlapped and riveted together at the laps (A17). The curved plates were also riveted to the internal beams at the overlap positions. The edges of the plates were all machined to butt up neatly together. Where the ends of the plates butted together, cover plates were riveted over to make the joint (see A17).
The diagonal braces in the towers were of ‘flattened tubular’ section (A19). This shape was chosen to facilitate joining the tubes where they intersect (A20). Yet another difficult junction to construct.
An aspect of construction, the use of electro-magnetic machine tools, was presented in a Paper by F. J. Rowan, and Paper and discussion were summarised in Engineering 1887/08/05. Surprisingly, the numerous holes for rivets were drilled using flat drills, not twist drills. Answering a question from Mr Smith of Smith and Coventry, who advocated twist drills, Mr Rowan stated that they had tried them but they were too costly, and the speed of drilling with flat drills was satisfactory.
The production of the main tubes was described in Engineering by Andrew Stevenson Biggart [7] [8]. The largest individual plates were 1.25" thick and weighed 1.6 tons. After bending hot between curved blocks in an 800-ton hydraulic press, the plates would deform in an unpredictable way. the edges were planed. The tubes were built round about a mandrel, being supported by temporary connections, and
drilled through the various parts, while in the exact form
they were to be when finally erected. 'Many methods were suggested, and tried, to overcome
this warping of the plates : thus, for instance, the edges
were covered up, thereby allowing them to cool more from
the centre, another mode was to reheat and give them a
second squeeze; yet another was to allow them to cool
partly, lying on a series of iron rollers, set to the true
form the plate should take. These and others gave only
very varying success. The plan finally adopted was to
curve a quantity at a time, laying each plate, as it left the
press, on the top of the immediately preceding one, with
a layer of ashes between, [allowing them to cool in piles
of convenient size. When cold, each plate is again placed
in the press, and straightened by means of repeated
squeezes, strips of thin iron being placed above and under
the points necessary to be brought to the true form. This
answers the purpose admirably, and is the only method
now in vogue.
A somewhat striking incident happened during these
preliminary trials. It arose out of an attempt to bend one
of the 1 1/8 in. thick plates while cold. During this process
the plate cracked in several places, although the curve
was only equal to that of a circle with a 6ft. radius.
Samples for bending and tensile tests were cut off, and
showed the plate to be of remarkably good material, and
quite up to the specified quality.
Mr. Arrol attributed the failure to unequal cooling at
the steel works, and this is borne out by the fact that different parts of the same plate are not uniformly easy or
difficult to cut, but both these experiences are often found
in a single plate.
Mr. Baker thought the failure of the plate to stand the
bending was due to the fact that its edges and ends were
not planed, but in the state they were in when they left
the shears at the works. He had made a series of experiments with sheared and planed plates, and from the
results obtained arrived at this conclusion.
Annealing removes satisfactorily both these objections,
and in this lies the great benefit of bending the plates
while hot, and allowing them to cool as described. .... After bending one of the first plates, it was kept between the upper and lower blocks for a few minutes, while it was yet hot. The consequence was that the side of the
block next the plate heated much more rapidly than the
other, or remote side. This induced a very heavy strain
on the metal, so much so that it broke the upper one completely through, at the same time giving a report somewhat resembling that of the discharge of a pistol.'
The edges were planed using a planing machine having a toolholder guided by the machine's vertical columns. Another machine was used to plane the curved ends. The machine, which may have been adapted from a normal plate edge planing machine, had a column at one end, from which was pivoted a pendulum with the planing tool at the lower end. The pendulum was oscillated by a connecting rod attached to the reciprocating toolholder. The tool was fed and reversed manually. After machining one end, the plate was turned round to machine the other end, the machined end being located against a cast iron faceplate to ensure parallelism of the ends.
See also The Engineer 1885/01/16
People
The names of some of the prominent peopl involved in the design and construction have been mentioned above, and under the heading 'See Also'. Wilhelm Westhofen has provided us with the names of some lesser-known figures. These include:-
Reginald Empson Middleton: In 1883 he was appointed engineer-in-charge of the setting-out and measurements of the Forth Bridge, with general superintendence of the survey
William Newton Bakewell: Westhofen wrote: 'The duties of the survey department in connection with the erection of the cantilevers were of the heaviest kind. The work had to be carried out in the most exposed positions and in all weathers. To Mr. W. N. Bakewell belongs the credit of a great achievement, and it is not too much to say that to his courage and decision and promptitude in fixing points, is due the saving of much time and much expenditure.'
The names and occupations of 73 men and boys who died during construction have been recorded on a bronze memorial at North Queensferry (see Photos P1 & P2).
Allan Duncan Stewart: From 1881 to 1890 he acted as Chief Assistant Engineer for Sir John Fowler and Sir Benjamin Baker, on the design and construction of bridge.
See Also
Sources of Information
- ↑ http://en.wikipedia.org/wiki/Forth_Bridge
- ↑ The Times Mar. 5, 1890
- ↑ The Engineer 1924/12/12
- ↑ The Stock Exchange Year Book 1908
- ↑ [1] Online copy of 'The Forth Bridge' by W. Westhofen, reprinted from 'Engineering', 28 February 1890, made available by archive.com
- ↑ ‘The Forth Bridge – A Picture History’ by Sheila Mackay, HMSO, 1993
- ↑ Engineering 1885/01/09
- ↑ Engineering 1885/01/16