Scribd Logo
Upload

Welcome to Scribd!

  • Timber Lattice Roof For The Mannheim Bundersgartenschau

    Uploaded by

    董思辰
    46 views36 pages

    Original Title

    Timber lattice roof for the Mannheim Bundersgartenschau

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    46 views36 pages

    Timber Lattice Roof For The Mannheim Bundersgartenschau

    Uploaded by

    董思辰

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 36
    une 6240712 624.9163 Paper to be read before the insitution of Structural Engineers at 11 Upper Belgrave Sires, Landon SWIX @BH, on Thursday 10 Apel 1978 Timber lattice roof for the Mannheim Bundesgartenschau E, Happold, ssc, Eng, Fistucte, FICE, FIOB W. I. Liddell, ma pic, cng, Misteucte, MICE Ove Arup & Putnees ‘Me Edmund Happold is the Executive Parte of Structures 3, which in one of te stveturaldusions of Ova Arup & Partners. Educated {at Leeds Gramroer Schoo. Bootham School, Yerk, sad Loous Universey he warked wht Sever, Elst sod Krieger, New Yorkin the early doysofeeble roots and pneumatistuctures, The {iat cable vote worked on with Ove dup & Partners were 1967, {ar tho Conterence Centein Mocea, Seu Arabia Since on he hos ‘been engeged on evany ighewslght root schemes, belng «taster lof tho Lightweight Structure Laberatary whieh Ore Aru & Paros share wth Fre! Otto. He ies cestooporiding member of ‘SFB 64, he Geren Governmant sponsored grovpig of university departientsitrested i long span roots, the Institution representative on the organising eommetcs forthe Conference on Tension Sirictres which held ie meetings ta Condon in 1974, wads a epresantative on the BS Coto Commitee for Air Supported Stoetres He's @ member ofthe Council of tre Instesion ‘Me. tan Liddell joined Ove Arup 8 Partners as an assistant engineer working on the Sydney Opers House, after gradusting rome {Camtrtige in 1980. tn 1962 he aid 2 post graduate holes ot Inparial College and subsequently spent four years with ‘Hoist & Company Linited as design engineer, etinotor nd Ste engineer on» cooing tower contest Ho etumed to Ove Arig @ Pars in 1363 0 eh thr Ed ope ‘brofectenginea In charge of» prestige building to Seer Acoba, “nds now an astoslte partner In Suuctaree 3. responsi for 3 ‘numberof projects at home and ebrowd: Hs bntved in ‘Suucturel design couse at Cambeldge Univrsy sis » mambo ‘ofthe Instation's od hoe comme investigatiy “Ctl tho Stuetural adequacy of beildngs” Synopsis, ‘Shes constructed by iting a at square lattice into.» doubly ‘curved shape are 2 recent form of construction. Such » ‘hol of four times’ greater span than any previous examples had to be completed in 18 months for an exhibition This paper describes the engineering design. It attempts to ‘The Stuctura! Enpinoer/March 1975/Ho. 3/Volume $3 show how the loads were defined, how the structure was ‘modelled and tested both physicaly and mathematically, end ‘how these models were used to determine the construction datail. The design process had to be evolved to derive sufficient understanding 80 that the decisions could be made Dy the critical dacs, INTRODUCTION Background Linear Arches or ibs. ‘Conecsive 2 cord or chain 10 be ‘exactly inverted, so that the load appli to i, unchanged in dinection, amount and distibution, shall act invrards Instead of outwards; suppose, further, thet the cord of chain isin some manner stayed or sttfened, 80 a8 to enable it 10 preserve its figure and to resist a vwrust; it thon becomes a finear arch, or equilibrated rib; and for the pull at each point of the original cord ie now substituted an ‘exactly equal thrust slong the rib at the corresponding point’ Willism John MacOuorn Rankine, 1858" Since 1946 Professor Frei Ott of Stutigan University has bboen using hanging chain nets to detine possible structures in which, when inverted, the salf-weight produces direct force ‘oniyé (Fig 1). Fino chains cannot transmit moment and suspended chain net can easily be used, provided itis a shear. fos mechonism, to determine the statically most favourable dome shape under gravity lowding for any continuous boundary condition. Such a direct force stuucturo, can theoretical be extremely thin but ite thickness will be determined by the stiffness required to withstand buckling and acyramettcsl loading. With this technique, Professor Onto was using the same ‘method employed by the Spanish architect, Antoni Gaudi, at the end of the nineteenth contury. Mothods of graphical analysis wore popular with the enginsers at thet tine, and Gaudi developed their two-dimensionel modeling into three dimensions (Fig 2). For the Guell Colony chapel he made Wire models hung with appropriate woights 10 achiove in roverse a logical structure which he could amend until it satistio him architecturally (Fig 3(a) and (b)). This method of visualing more complex forms gave him @ buildable ‘sculptural freedom which conventional simulation methods ‘such 2 drawing, could never have allowed. Professor Otto went on to develop an erection method {rom the fact thet the shape of a hung quackanguler chain net can be recreated in tho initial shape by 9 flexural semi-rigid lattice of etvel or wooden rods in a uniform mesh provided ‘that the lattice is rotatable at the inter-section points, Tho lattice can be prefabricated a an equal grid. pulled or pushed 'p and then fixed against collapse by the edge forces. Such a lattice has no in-plane shear stfneee and thare are differing, angular displacemonts at various points, the laigest at the iggonal edges and the smallest near the principal axes. ‘Shoarstfiness con, to 2 limited extent, be provided by fixing 99 Fig 1. Eerly hanging chain model ‘the joints, though ifthe deformations are excessive diagonal ties must be provides. Professor Otto's fitst structure of this type was erected for the German Building Exhibition at Esson* in 1962 (Fig 4). It was a latice dome, on 2 super elliptic! base, 15% 15m with a height of 5m at the centre and a mosh size of 048m, It was made with Oregon pine lathe of 40 mum x 60.mm cross section and 2 length of 19 m echieved With finger joints. The Inthe were connocted by bolts, The shape of the dome and lengths of the mombere were deter. ‘mined by a suspended chain net though, notuely, it mas. found that the flexural doflection curves of the laths were not Fig 3(e). Hanging mode! for Guell Chapel 100 Fig 2(3). Practice! graphical method of finding ine of resistance of an arch Fig 2(b). Model of arch formed of radiused vousoirs which alfow the fine of thrust to move in respense to point loads. This model was made by Frei Otto to demonstrate the _rincipe of inverting a hanging chain, ‘exoctly those of catenaries. The edge beam was spiked into ‘he ground and erection caried out with a crane with spreader ‘beams, the base being pulled in with diagonal tes, Later in 1952, at 9 seminar at the University of California, Betkeley, USA. Professor Otto conducted « series of study projects on suspended catenaties, the definition of @ dome Dy use of a net and. finally, a full-scale erection of @ lattice ‘dome made of round steel bars (Fig. 6). Fig 3(b). Detol from stucture ‘The Structural Engineor/March 1976/No. 3/Volume 53 Fig 4a), (b) and (2), Lattice dome at Essen, 1962 {In 1965, in collaboration with Professor Rell Guibrod of Stuttgart_he won tho competition for the design of the ‘Gorman Federal Pavilion for Expo’ G7 at MontreaP. The main Structure was a large continuous cable net roof but. within tho rool. thore was en auditorium with i vestibule covered by a timber lattice dome (Fig 6). The plan shape was very lnregular, with @ reentrant angle and spans of 17 rm 13m {and 20m = 4:8. For this project « further rotinement in the rection method was used when the lattices wore pro. febvicated in Germany, collapsed diagonally into narrow Fig 6. Erection of latice domes et Mortreat “Tho StvcturalEnginser/Match 1975/Ne. 2/Youme 53 Fig 5(a) and (b). Letice dome at Berkeley Dbundies of strips side by side, wensported collapsed to the sit in Canada and expanded and erected ther ‘These three lattice domes are the only anes previously built ‘though several studies have been cared out including one for a banqueting hall for the Conference Centie in Mecco, Sauali Arabis. whoro Professor Oto ie again in partnorship ‘with Professor Gutbred and Ove Arup & Partnors. Within the last two years, however, a rigorous sores of shape studies hhas been carried out with ehain nets by @ group ot the Institut {fir Loichte Flichentragworke at the University of Stuttgart (Fig 7) Nature of the structure ‘The term lati shells used in this paper to describe a doubly curved surfece formed from 2 lattlee of timber laths bolted ‘together at uniform spacing in two directions. When tat. lattice is © mechanism with one dagree of trosdom. If it were {formed of rigid members with fitionless joins, movement of ‘one lath parle! to another would evoke sympathetic ‘movement of the whole trame causing all the squares to ‘become similar paraielograms. This movement causes {changes in length of diagonal lines through the nodes (Fig 8). Fig 7. Chain net shape study 101 r + Fig 8, Latice distortions It is this property which allows the lattice to be formed into the doubly-curved shape of tho shel “The shape for the shall is ectabliched by photogrammatic measurement of @ hanging chain model and is funicular. IT the shell is loaded with its own weight only, no bending forees result. This i an ideal condition, as in’ practice the imposed loads on the shell ro greater than the self-weight ‘and afe not uniformly distributed at the nodes. A funiculer shape is an advantage but is not essential ‘Whoa the lattice has been curved to the shape of the shell, it is fixed only by its connections to the boundaries. The funicular shape ie modified by the effect of bending of the lathe and with no loads applied it would be such that the strain energy is minimized. Ta this condition the lattice sholl resiete point loads by bonding of the laths. This is acompanied by large movements fof the shell and changes in the angles between the laths. ‘These movements indicate that the overall shape of the shell con be easily altered, and to resist these movements diagonal stifiness hes to be introduced. ‘A continuous shell mado from an isotopic sheot material thas eaval propertios in ell directions. An elemental square on the surface can take direct forces and out-of-plane bonding ‘on orthogonal directions (Fig 9). The force/displacement [propertoe are not affected by the oxientation of the element. Howover, an olemont of a lattice shell consists of a paralelo- ram of four laths. This element can only rosist direct forces: in the directions of the laths, t can also resist out-of-plane ending. In iis intiat pinned condition, it cennot resist diagonal forces. It cannot therofore tansmit forces directly ‘rom one fath to the next. Diagonal tifiness can be introduced in various ways: {@) by making the joints rigid so that shear foroes ar Certiod by bending moments around the element ring (b) by edding cross tes of cross sectional ares considerably dese than the laths: (c) by adding rigid cross bracing of equal area to the Iaths. ; = ner 9| conrinvous 3 | sueue —— armeesie~ ‘DEFLECTION Fig 10 102 Fig 9. Continuous shell and latice shel elements facing type (c) would produce a shell which was directly comparable with a continuous shell. Obvicusly the smount of diagonal stifiness Inttoduced by brecing type (b) can be vaiied by alteting the thickness or material of the ties. Loxd carrying bohaviour The loads on a shell can be divided into funicular toads which produce only direct forces in the laths, and disturbing toads which produce bending moments and large deflections. The deflections produced by the disturbing forees change the shape of the shell from its original funicular shape. The direct forces from the funicular toads then produce bending moments which inereace the bending moments produced by Uisturbing loeds, Ae the funicular loads aro increased, the Stlfness and resistance to disturbing loads is decreased. At a Critical funioular load there is no resistance to disturbing Toads: a smal deflection from the funicular shape couses collapse. This decroase in stfinass to disturbing toods Charactorizes compression structures. With a tension net the ‘opposite ie true; the deflections under load increase the stiffness and resistance to disturbing toads up to the point ‘at which the members break. With continuous shell there is an infinite number of funiculer loads. Any distribution of toad without dis continuities can be carried by direct forces within the shell without the necessity for primary bending moments. This ‘meens that deflections will bo much smaller than fora lattice Shell where disturbing loeds are comied by bending of tho Taths, So the effect of funicular loads in reducing the resistance to disturbing loads is fess and the collspse load tends to “Typical loed deflection curves of a tension net. a lattice shell and © continuous ehall fer disturbing loads aro shown In Fig 10. Lattice shells with diegondl stifiness ‘A pinned lattice shell can only cary disturbing toads by bonding of the laths. Under small deflections the structure behaves as & series of inter-connected flexible arches. The ending offects ate spread along the whole length of the ‘rch. With lorge distortions 8 dimple is formed. and cerain of the laths outside the dimple are stretched into tension, preventing further deflection in that zone and causing some Increase in sitfess. This behaviour fs only possible with low jovele of funicular load, With a large funicular load the decroase in stlfnoss ie 20 great that collapse will occur before the deflections become large. “These lage deflections ero accompanied by changes in tength of diagonal lines through the nodes. In the rool “structure these changes in length would eupture the covering ‘mombrane unless it were mado strong enough to resist them. Wit is made strong enough then in effect diagonal tes are introduced, the laige deflections are controlled end the collapse load is increased. Clearly It is necessary to introduce ‘agonal stifiness by using the membrane or tie. Double layer grids ‘White ties can be used to increase the in-plane stitnass and prevent excessive shear distortion, the only way to increase ‘the out-of-plane bending stiffness is to increase the moment of inertia of the individual members. A double layer grid doot “The Structoral Engineor/March 1975/No, 3/Volume 63 Fig 14 Hanging chain model “Tho Suctural Enginen/March 1975/No. 3/¥otume 53 this most effectively but introduces other problems: during the bending into shape the two parallel lathe along each grid ‘must slide relstive to each other. To allow this to happen, one ‘of the layers just have slotted holes. When the final shape is achieved the laths must be prevented from slipping, and even ‘whan slipping is provented the shear stiffness ofthe composite ‘momboris fa from ideal Architectural design A federal garden show is hold every two years ia one of the principal chies in West Germany. The exhibition is open for six montis and about four milion people visit it. usually ‘consists of a large park which is landscaped with flowers ‘and shrubs, where growers and nurseries exhibit new special sticins of plants. Playgrounds are included in the landscaping both for children and adults. Special evente— ‘concerts, theatre, games and spons—are held on most days. “The gatcien show is popularamong ies because a depress ‘ed open area may bere-landscaped with gardens of every kind, paths. lighting, lakes, kiosks and restaurants, The exibition provides the stimulus fo local pride to do this and brings: considerable financial hop. ‘The cities of Mannhoim ond Ludwigshafen (joint popuia- tion 600 000) wore solected in 1970 as the home of the 1975 103 coxhitition. Two sites, one on either side of the Rivor Necker, ‘were planned end 6 cablo car transport eyetsm hae been rected to connect the two entry points. "A Manheim architectural practice, Mutschler & Partners, ‘wes one ofthe joint winners of a competition held in 1971 to design tho parks and was appointed 10 design the section ‘which included 2 rmuki-purpose hall and restaurant. In thoir competion design the main poth of one of the parks ed ‘through this complex. For the Multihalle they proposed to use 2 membrane roof supported by balloons. This scheme wes Lunaccoptable to the building authoritos, and after some more early thought they asked Profeseer Otto, who had designed ‘several tent structures for previous garden exhibitions, to help ‘them with a solution. “Togethet, they lnvestigated various tent and pneumatic soletions. They felt thats free form rounded structure would be most appropriate, as the hall was to be adjacent to the town's newly-created and only hill. Lattice shells wore then Considered and they evolved a scheme with two principal domes for the Mulfalle and the restaurant, connected by ‘tunnels over the walkways. This complex was adjacent t0 the hill and wos surrounded by an artificial canal. Fig 13 shows the final wire mesh model made ot Professor Otto's office, Atolisr Warmbronn, The latice was to be covered with o translucent PVC membrane reinforced with open-meshed polyester fabric. ‘After this model was agreed and accepted, Professor Otto prepared a hanging chain model which was to define the ‘Geomety of the roof, This model was made to a scale of 198-9, the node points being small rings connected by rigid links (Fig 72), each 15 mm long to represent every third grid fof @ 500 mm mesh. The whole mesh was put together by hand fend lowsred onto @ pe ronce anos ‘i Sessa t f over srs ye + Fig 28, Diagram to ilustate shear deformation ata joint Testun lhe MELATONAST ABA ASSUMED LEA eee Strength properties The Forest Products Research Bulatin 50 {quotes valuos of strength properties for a number of timbers Dazed on shor term tests on small clear specimens. Those for ‘wostein hemlock are takon from tests carted out by J. G. Sunley, who in his report” shows that they compare well with ‘American and Canadian tests. It therofore seemed reasonable {o teat these valvee ae representative and to assume that the timber for Mannheim eonformod to this distibution, "The values of bending strength and elastic modulus were as Tobie 1 TABLE 1 Bonding Modulus of Moisture | strength loti ‘Content | N/en™ vm meen | SD. | moon | SO Green ero | a0 | 100 | e700] 1430 Ait ry azs_ | #30 | 206 | 10400] 2040 EP.R, Bulletin 50 gives information from which curves of the change of modulus of rupture and modulus of elasticity with molature content ean be derived. It can be assur that wostein hemlock has timilar curves. These are plotted Fig 23(6) and (b). Duration of toad. Booth and Rooce™ give examples of tong term Bending tests cated out at the US forest products laboratory on small Douglas fi bears. These results wore used to dorve the time to failure against stress level curve as Fig 24. This cuve ingicates 9 minimum percentage stress level ‘95 56 por cant for 2 60-year lo2d duration and has been used for establishing permissible fong-term stresses in CP112. creep. Booth and Reece also quote daformation/time curves for mountain ash established by Kingston in 1962, These give ‘coop depth asa fraction of ial dellection (Fig 28). 42 Fig 23. Rig (or shear sitfress tests “These tests were carried out on sir dry specimens. Other teats have shown that the creep rate Is more or fess constant provided that the moisture content is constant but variations fccolerato the creep. “These basic proportion were used to establish and vority the swucture properties of the lattice. However, for some fefiects where thers was uncerainty about the behaviour Special tests were caried out. In the case of tho fattico jotnts those tests were cartied out in paalll with the buckling loed “cloulations end the development of the details, and was 2 Important part of the decign process. Testing programme Suess relgxation investigation. For significant sroas of the Tattice shel the radius of curvature is relatively smal, of the crder of 101m with some areas at ® m. This curvature results high bending stecses during erection with the possibilty ‘of overstiessing the mombors by the application of service toads Stresses in timber held to a constant detlection relax. and fa 8 ftst approximation the reciprocal rulo applies. TNs states ‘hati takes * days for deflection to increase by a factor of ¥ {for constant load-detlection testing, then in the same x é8ys ft constant deflection, the stress will reduce by the factor Y- 1 i.e, be + times the orginal stress. y “This tule is very approximate but does suggest en order of magnitude. A benefit of this etaep-relaxation phenomenon Te that the initial high bending stress will relax ond reduce (0 ‘Dovoptable levels by the tine service loads end stresses ave “pplied. To investigate thie further, several lengths of hemlock Wrote bent to 6 predetermined radius of curvature by mens ‘of an eccentric calibrated sping 26 shown in Fig 26. ‘This radius of curvature war eeloulated (0 give the seme ‘extreme fibre sess 98 would exist in SO mm square timber “The Stucwral Engineer/Masch 1975/No. 3/Voume $8 ‘eFLEcTION Fig 30. Typical toed dettoction curves recorded in shear stiiness toss ‘being bont to 6 oF 12 m radius of curvature. The curvature was ted as 6 or 12m equivalent radius. ‘The radius of curvature was maintained constant and the Fedlucing spring force was measured over a substantial potiod of time. The offect of compressive stress due 10 tho lood applicotion method is small and the stresses ‘con be onsidered to be purely bending forthe purpose of calculation of stress relaxation. The results of these teste ate presented in Fig 27, non- imensionaized nd plotted to @ logarithmic time. cesie ‘The tosis wore useful in that they demonsirted thot the high initial bending stosses relaxed quito quickly, akhough ‘thoy were caried out in intemai conedltions where the releioe humidity would not vary as much as outside, end this coua alter the eeap behaviour. ‘Shear tests on node joints, The development of tho stendetd ‘node is described in the section Development of dersle nt P42. eshte el edna Diora F's the force transfered between the upper and lower {aths at a joint and $ isthe relative movement (Fig 28). tie value of my controls the out of plane shear stiffness of the {attice. The problems associated with tho shear stiffness of she | tho section on Mathematice! ‘modeling on page 118. In order to justiy the values chee, Ax frst, it was thought thet the interface slip would be the main problem and so tests wore carried out to hwestioats ee A tost rig wes set up as Fig 28, a large spring of Taenn stifiness Being used to control the clamping force, This wey tested on an Avery Dennison Universal testing machine ana continuous load-detioetion plot was made. Fig 90 a “The Suuctural Engineer March 1975/No. 3/Volure 83 goer T E No. OF TURNS & sf iF } NUT UNDONE 20 = 2 S154 — _ & gv0-+ LL i ALL Tests: & EXCEPT 5 WeRE Loaoina omy os | ° 1 2 3 SLIP mm Fig 31. Typical toad deflection eurves for test using aise sorings 'ypical sot of results fora joint loaded several times cyclically; ‘the stip points can be clearly seen. Ifthe joint did net sip the piece of lath in the middle would start to roll round. The Jpint could then be taken up to a high lood without slipping ‘These tests indicated a shear stifiness at low lowds of Up te 150 kal et 1000 kgt/em, and this value was used for the computation, Aitor some computer runs had been made it became ppetent that the shear stltnss was more critical to the collapse load than envisaged. It alco oppeated possible thet the long spring could influence the sheer stiffness and so then the washers for tho roa joint became avollabe a further seslos of tests was made. These wore carted out with the ise springs in vatious stagos of compression. Fig 31 shows @ typical st of loac/dellection curvos. These indicate thet the curves are linear for a range of bol loeds irom 500. down to 350 kat (four turns undone) with an 7 of 1000 kgf/em, ‘This can only moan that the deflections are controled by the shear deformation of the middle piece of timber and local {deformations of the contact points. A chock eateulation on this basis gave shear stifiness of a similar order of maunitude Tests on boundary connection. The maximum lath foreo ie ‘round 760 kgf and is transferred t0 the boundary by bolts in ‘shoo, The minimum edge distance alowed by DIN 1052 is 3 d, ‘which means that the maximum permissible diameter of bolt should be 8mm. The llowable bolt loads from DIN 1082 {and from BS112 are shown in Table 2. TABLE 2 ‘Shor tere ane Toad sharing 108 2 In single shear would require seven 8 mm diameter bolts and the rules for spacing would not allow these to fit on the edge board. It wos ecessaty to go outside the DIN rules to make the boundary connections and so it was proposed to carry out fests to prove the strength of 10 mm and 12 mm bolted joints, 13 Fig 32. Assemble for boundary connection tests A test sssombly a6 Fig 22 was arranged using 12mm ard 20mm bolts in accurately drilled hols. The end clearance wes 50 min. The assembly wae tosted in tension and load detiec~ tion curves es Fig 33 were obtained, ‘on the basis of these tests, bolt loads of 200 kgf and {400 kof ware agreed forthe 10 mm and 12mm bolts. Rotation of late joint. The model testing wotk indicated thet it wos necessary 10 develop some in-plane shear stiffnoss in the lattice either by tias oF by making the node joints rigid ‘against rotation. The typical node joint, Fig 50, incorporates Spring washers to obtain a clamping force. This force gives the node joint some resistance to rotation. ‘in order to evalvste this shost resistance, simple rotation tests were caried out. joint was assembled as Fig 34. and Toads were applied to the cantilever part by hanging weight fat varying radius. Deflections were measured using @ cial ‘gauge. Tests were done with virgin surfaces and efter polishing by repeated rotations. A plot of the moment/ otetion curves for polished surfaces is given in Fig 35. Ettoct of tre-prooting salts ‘The hemlock laths had to be treated with flame proofing salts {0 setialy the requirements of the fire officer. A sample panel complete with covering mambrano was tested by lighting & fire beneath it to see Hf the fire spread. This test was con- sidered to be sstsfactory in the impregnation process 6 to 8 per cent of salts were absorbed into the timber in 9 water solution. The timber was {thon died to a moisture content of around 10 per cent for ‘losing, In this drying process the salts combine with the rile of the colllose fibres. This can teduce the stiength of m4 TITIEA Fig 93. Results of boundary connection tests the timber, especially i the drying schedule is too sovore, os is how admitted by some suppliers of salts. “As generally the strength and stiffness properties of timber very in proportion, this discovery caused considerable anxiety Tost the elastic modulus be reduced as well, To check on this @ Fig 36, Rig for tots on rotation of node foint “The Steututel Enginee/March 1975/No. 3/Vokume 53 —_— | 53 eh Be : 10" * Fig 38. Moment rotation curves for polished surfaces series of 4-point bending tests wat caidd out at the ‘contractor's factory on treated and untoated timber. In these {sts the central deflection was mesaured with » dal gouge ‘and used to calculate the elastic modulus. The tests indicated that thore wos 2 20 per cent drop in bending strength but it ‘=ppeared thatthe elastic modulus increased by 10 per cont, ‘As soon as structural design was started, it wes decided to ‘ake a Yo ssale mode! of the 15 m span Essen dome to help gain some understanding of the behaviour of latico shelle, It was also decided to make the lathe ae thin and flexible as possible so that the structural propertios resembled those of the Mannheim Muttinalle shell. The most suitable moter for the model lathe was perspex stip which had Young's ‘modulus % of that of timber. K was also easily worked and ipinted, ‘The bosic Intice for the model consisted of 34:7 mm laths at 50 mm contres. The dimensions of the model wore os Fig 36, the joints were made by diling small hotes in the members at 60mm centres and then passing small pine through the holes. The pins were then bent ovat to complete the joint. Tho laths were cut to length cing the ‘cutting Petter obtained from Atelier Warmbronn and were tod

    You might also like