122 



THE CIVIL ENGINEER AND ARCHITECTS JOURNAL. 



[April, 



*287 lb of coke to the ton of iron. Before the adoption of the hot blast, 

 the consumption of coke was 411 lb. to the ton of iron. The same quan- 

 tity of iron is melted in oue-half of the time that was required before the 

 adoption of this process." — Franklin Journal. 



AUST I'ASSAGK BRIDGE. 



(With an engTaning, Plate V.) 



The proposition to build a bridRe across the old passage of the Bristol 

 Channel, at Chepstow, has arisen from the intended improved connection 

 of South M ales with I'.ngland by means of railways, and whatever may 

 be the objects of crossing the Bristol Channel elsewhere, the old passage is 

 unquestionably the point where the nearest junction of South M ales with 

 Bristol, the West of England, and London, can be effected. 



In the session of 184.i I recommended, in my evidence upon the South 

 "Wales Railway, that this bridge should be constructed with a span of 

 1,000 feet from pier to pier, and height of 120 feet above high water mark 

 of spring tides. Since then, I have had a correct plan and sections made 

 of the Channel, from which I find it practicable to place the piers on rock 

 foundations, accessible at low water, but at distances of 1,100 feet apart. 

 The bridge will therefore require four spans of this length, and one at each 

 end of nearly 5o0 feet. In addition to which, another bridge should be 

 built across the Wye, with one span of about 500 feet in length, and one 

 at each end of about -.250 feet each. I expressed ray opinion in the ses- 

 sion of 1815, that the suspension principle adopted at Meuai Bridge and 

 elsewhere, would not be sufficiently steady for the Aust Bridge without 

 material improvements in it. I have accordingly, in conjunction with Mr. 

 Francis Bashforth, Fellow of St. John's, Cambridge, designed a bridge 

 for the Aust I'assage, which the accompanying engraving represents (see 

 plate V.) ; the calculations for which are subjoined, for the scrutiny of 

 scientific men, and I have great pleasure in associating my name with Mr. 

 Bashforlh's in this work. Francis Giles. 



Experience has shown that the instability of suspension bridges is their 

 great defect, whilst they may be made of sufficient strength to bear any 

 load that can ever be placed upon them. To guard against their liability 

 to undulation an arrangement is proposed, so that every part may be 

 always in its proper position to support any heavy weight plared on the 

 roadway. The bars radiating from the tops of the piers are of variable 

 thickness, so that each is capable of sustaining the same weight as when 

 placed at its extremity on the platform (supposed to be a rigid lever move- 

 able about the end). These radiating bars are kept straight and connected 

 together by transverse rods, but the former alone are calculated to sustain 

 the whole of the load, the latter being employed to keep them in their 

 proper position and to ensure the assumed rigidity of the platform. The 

 supporting bars are attached to the platform at equal distances and are 

 passed over the top of the pier (side by side or otherwise). On account of 

 their number the space covered would be considerable, but this would be an 

 advantage as the tendency would be to keep the platform steady. In 

 addition, the main chains are proposed to be connected by rods overhead 

 to within about 20 feet of the roadway. So as to form an immense trussed 

 beam. Let ( denote the number of Ions that a bar of iron of 1 square inch 



E. 01 * B 



section would bear without injury. A B =A ,EB =s, BC = C D=: 

 D E = &1C. = 5. Let A F be the r the bar from A B. Then B F = rS 

 •»-, the weight supported by A F at F on the platform, /i^ weight of a bar 

 of irou 1 foot long and 1 square inch section, n = number of bars between 

 B and .1. 



A F A F 



'AB = "' h 



The tension of A F : 



Weight of A F =K tensio" "''■*■ ^ v A F -1."^' _«C|W. +r= 8» 

 t ^ ~t h t K'- k--> 



and giving to r successively the values 1, 2, 3 n we get the necessary 



•veight of each bar, and four times their sum will be equal the weight of 

 the supporting bars between the piers = W suppose 



W =4) 



A a 



•■'I 



= ..A« f, .!ill"(l.^l^_L.U 



But if W be the weight to be supported on the platform when nDiformly 



g 

 distributed , W = 4 « u and n J =— 



•W x h 



t ( ' ■" 4A-' "-S " 2n 

 It is manifest that W' cannot be made a minimum by the variation of n, 

 but as n increases W decreases, and the least possible value is given by 

 W«A ( S' 



W'= 



/ S' , 1 1 , 1 , \ 



»=B ID which case W' 



{•*iS^} 



Id this case we should 



have the bars replaced by a very great number of wires, but there would 

 be no rigidity for preventing the platform being raised in the middle. A 

 bridge so constructed in Scotland was soon destroyed by the wind and 

 replaced by one of the common form. Hence the greatest number of bars 

 must be made use of, consistent with the rigidity of the whole. W is also 

 a function of A, the height of the platform above the roadway. Making 



dw s /"i i r 



dh 



we find 



d2_W' 

 rfA= 



A = + — A / —.—. — J. The positive value 



a positive quantity and . ■ . gives the minimum vakie 



of A makes 



of W. 



But the expense of raising the piers to a height necessary to ensure the 

 minimum quantity ol iron being required must be taken into account. Let 

 11 denote the height of the roadway above the ground, and x the height of 

 a course of stones on the pier above the platform. Suppose now the cost 

 of any course of stones varies as (height) s, where s is integral or fractional, 

 and also that n is the expense of a course at height unity from the ground, 

 expressed in terms of a weight of iron of equivalent value. 



Expense of the pier above the platform, 



=^J (H + x j \ln =g^ I H + A 

 must have such a valu 



s+i_jjS+n, 



i 



: such a value as makes 

 WkA f 



S- 1 1 



Hence A must have 



_ J_ 



t \ 4A= ^3 ■'"2n"*'CB2 



It must be remarked that the extremities of the radiating bars would re- 

 main in the same straight line for all variations of temperature, provided 

 each bar expanded in proportion to its length. Thus the equation to a 

 straight line is f cos e=a (1) If we suppose p to receive an increment ^ p 

 proportional to its lengtli, the radius vector then becomes p + ynp = p' (2). 

 (suppose) and eliminating p between (1) and (2) we get p' cos 6 = 

 a (l+t«) • • • (3), which is the equation to a straight line, paralUI to (1) and 

 at a distance jl a. For a span of 900 feet ; a ^ 00 feet suppose, and for 

 ordinary changes of the atmospheric temperature 

 1 90 ,.. 12.. 



. Iia = . 



feet = — inches : 

 20 



•0 inches. 



It has been objected that there would be a tendency in the network to 

 " buckle ;" this would be perfectly correct in the case of a girder bridge, 

 but there would be no occasion to fear that defect if the proposed plan were 

 properly carried out. For it must be remarked that everything is made 

 subordinate to the radiating bars. The straight horizontal rods are in- 

 tended to insure the rigidity of the platform. \t hen a weight passed over 

 the bridge, the tendency of the supporting bars would be to rotate about 

 the top of the pier, but these straight bars acting directly by tension on one 

 side and compression on the other, prevent this, and cannot buckle unless 

 the rods stretch. The bars must be coimected so as to admit of adjust- 

 ments without injury to the strength of the material. 



The weight of the material required for the supporting bars on this plan 

 would be from half to two-thirds of that required for the main chains of a 

 catenary of equal strength, but on account of the numerous cross pieces, 

 the saving effected by this plan would not be important, but it is satisfac- 

 tory to know that suspousiou bridges may be made much firmer and free 

 from undulation without increasing the cost and adding to the weight to be 

 supported. 



