Engineering Imagination and Local Vision
When Sigfried Ciedion (1888-1968) made the third revision of his classic Space, Time and Architecture in 1954 he added an afterword to the chapter on Maillart, stimulated largely. by the Zurich Cement HalJ.l There he spoke of the great future for thin shell roofs, and in a revised introduction to the book itself he called shells "the starting point for the specific solution of the vaulting problem for our period."2 Ciedion had discovered Maillart for the world of modern art even be- fore most of his greatest works had been designed. The Swiss art histo- rian saw in Maillart's bridges and Rat slabs of the l 9Z0s how pure engi- neering works could be works of art. But, although he could see that the 1939 Cement Hall, "in the hands of a great engineer ... became at once a work of art,"3 Giedion's evaluation of works built after World 171
THE NEW AGE OF STEEL AND CONCRETE
War II was less incisive. In extolling such vault designs as the exorbi- tantly expensive Sydney Opera House, he missed the deeper implica- tions of Maillart's works and, in more general terms, the separate tradi- tion of structural art. Giedion came as close as anyone to appreciating the art of the engineer in his perceptive essays on Eiffel, Maillart, and others; but, with regard to vaults, Ciedion did not follow the deve1op- ment of structural forms that had arisen after World War I. These vaulting forms arose from the engineering imagination which took off from simple ideas grounded in the laws of nature. For example, gravity dictates the shape taken by a suspension bridge cable under load. Imag- ine the cable frozen and turned upside down; the result is an arch, and a series of such arches can form a dome surface.
Yet the history of twentieth.century structural engineering does not follow a linear progression, mainly because the laws of nature must interact with-locally based-aesthetics. It is not a question of one de- velopment leading directly ta another and so on, but rather of parallel ideas arising in different societies and then Bowering in ways that re- Hect the particular patterns of those societies as we11 as the general laws of nature. In spite of serious efforts to share information and ideas inter- nationally during the past century-indeed, since the 6rst great World's Fair of 1851-local vision still strongly inffuences structural technology.4 The importance of local biases was already made evident in our last chapter. And it was local biases, far more than any general theories or scienti6c discoveries, which set the direction of thin shell roof design following World War I.
We can identify at least three distinct, independent, and nearly simultaneous lines of development, each associated with a different cul- tural tradition. In Germany that tradition was mathematical and scien- ti6c, in Italy it was historical and artistic, and in Spain it was rooted in an artisan building tradition. All three lines of development were aimed at the goal of covering large areas with curved concrete surfaces, and of creating strong structures with thin, curved slabs rather than with thick, Rat ones. {This is the principle behind corrugated metal and Gothic vaulting.) The Germans tended to work with surfaces that they could study mathematically, like spherical domes and circular cyl- inders, and they made visually separate systems of beams, wal1s, and arches to support those surfaces. The Italian, Nervi, by contrast, de- signed ribbed surfaces that were reinterpretations of earlier masonry 172
Roof Vaults and National Styles
ribbed vaults in Italy and elsewhere, and he sought to integrate visually the ribbed surface and the supporting structure. The Spanish designers, meanwhile, stimulated by a local artisan tradition of laminated tile vaults, used reinforced concrete to create smooth riblcss surfaces, Which they also tried to integrate smoothly with the support structure.
Each of the three lines devc1oped between the wars, but did not reach its full potential until the 1950s.
Dischinger, Finsterwalder, and the German School
When Wilhelm Ritter defended load tests against the arguments ad- vanced by Franz Engesser, he was putting the Swiss concern with phys- ical observation above the German concern with mathematical calcula- tion. Thirty years later, one of Engesser's star pupils, Franz Dischinger (1887-1953), from that mathematical bias began a German develop- ment in shell roof design that has continued to the present. In 1913, Dischinger went to worlc for the German building firm of Dyclcerhoff and Widmann A. C., which by then was already established, along with Wayss and Freytag, as Germany's leading designer-builders of reinã
forced concrete structures. In 1922, Walter Bauersfeld (1879-1959) of the Zeiss Optical firm, in collaboration with Dischinger, designed
ãa thin shell hemispherical dome roof for a planetarium in Munich. To- gether Bauersfeld and Dischinger toolc out patents on what came to be called the Zeiss-Oywidag system of thin shell concrete roof strucã
tures. 5 They then sought more general applications for their system, and began to design factory roofs and market halls. But, being Germans in a scientific tradition, they felt uncomfortable with any structural form that they could not analyze mathematically, and hence immediã
ately set out to find a mathematical formulation for domes. In 1928, Dischinger was able to present a fu]) mathematical treatise on domes and to show numerous designs, either built by then or under construeã
tion, that were based upon these formulations.6
More important than the mathematics itself, however, is the way in which Dischinger used the formulas to malce the forms. His starting 173
THE NEW AGE OF STEEL AND CONCRETE
points, as a designer, were physical images developed from mathemati- cal formulas. Because the formulas assumed axial symmetry (a surface of rotation formed geometrically by rotating a curve about a vertical axis), the plan form had to be close to a circle. Because the designers assumed roof loads to be carried to columns by separate structural systems, either beams or arches, they expressed the supports between columns by such systems.7
None of these limitations prevented Dyckerhoff and Widmann from building shells; indeed, the confidence inspired by the mathemati- cal theory encouraged them to explore other forms. Dischinger tried to find ways to apply his thin shell ideas not only to domes but also to bui1dings rectangular in plan. For such plans, he tried to devise sheU forms by stretching domes in one direction so that he could use some variant of his axisymmetrical thin shell theory. He tried to make form follow formula. This attempt did not work; Dischinger needed another approach. He knew that a simple barrel shell form (a slab curved to a circular arc) would be easy to build, but he could not find a satisfac- tory mathematical theory for it. However, a younger colleague took up that challenge and soon had a formulation. That colleague was Ulrich Finsterwalder (b. 1897), who would become the most versatile designer of reinforced concrete structures of his generation.
Finsterwalder inherited the German scientific tradition directly from his father, a professor of mathematics in Munich and a pioneer in photogrammetry and the theory of glacial movements. During World War I, as a prisoner of the French, the young Finsterwalder studied mathematics, and, after his release, completed the course in civil engineering at the Munich Institute of Technology. He graduated in 1922 and immediately joined Dyckerhoff and Widmann, with whom he remained for his entire career.8 Starting with his diploma project at Munich, Finsterwalder worked on the mathematical theory of barrel shells until he published the first workable formulations in 1933.9 Again the form came from the formulas.
Finsterwalder's first mathematical theory (which is technically called the membrane theory and considers the shell to have no resis- tance to bending) showed that unless the edge slope was purely vertical, the longitudinal edges of the barrel needed tangential supports. There- fore, he designed an elliptical cross-section to create a veTtical edge.
When this vertical wall-like edge proved difficult to build, he tried an- other form suggested ~y his more general theory of 1933. This form
174
Roof Vaults and National Styles had a circular crosNection with longitudinal edge beams. But even this theory was restricted by the mathematical requirement that the barrel be supported vertically along its entire arc length at each end. This requirement from the mathematical theory led Finsterwalder to design a thin vertical wall at each end to give the calculated support.
The theory-and, for the Germans, therefore the designed form-was restricted to circular barrels with longitudinal edge beams and transverse wall supports. This new form was ecOnomically competi- tive, and showed how concrete shells could be substantially lighter than other types of concrete roof forms. Dischinger showed, for example, that each of his Leipzig shell domes of 1929 (figure JO.I) was only one-third as heavy as the 1913 Breslau arch-ring dome (also built by Oyckerhoff and Widmann), even though the Leipzig structures each covered an area about 30 percent greater than that covered by the pre- war dome.10
These domes and barrels did not show visually their thinness, how- ever, and they appeared more to be made up of separately functioning elements than to be a single integrated form. But, even if these Cerman forms did not achieve the highest qualities of structural art, they at least proved convincingly that thin surfaces in artificial stone could be economically built and would safely stand. This was a major achieve- ment in structural engineering, and it encouraged subsequent develop- ments elsewhere, though others did not take up these specifically Ger- man forms.
FIGURE 10.1
Tlte Marhl Hall. Leipzig. Germany. 1~9. by Franz Dischinger. Each of the!!e 76-meter-span polygonal domes wu only one-third lhe weight of the 67-meter-span Bres- lau dome of 1913. lhen the longest-spanning concrete dome. The lightness was possible because these Leipzig domes were de.signed lll thin-shell .surfaces.
THE NEW AGE OF STEEL AND CONCRETE
Nervi and the Italian Tradition
Turning from the Germans to the Italian Pier Luigi Nervi (1891-1979), we come to an engineer who centered his entire career on aesthetics. There is no doubt whatsoever that Nervi saw himself as an artist whose mission was to create beautiful objects. Beginning to design on his own at the time Maillart's greatest war.ks were appear- ing, Nervi saw that structure could be art when it arose out of correct form, careful construction practice, and a conscious.aesthetic inten- tion. During the 1930s, Nervi was designing and building large con- crete structures, mostly as a result of winning cost competitions, al- though his intellectual bent was for reflection and for aesthetics.
\¥hereas Dischinger and Finsterwalder were writing about domes and barrels designed on the basis of the theories they had developed, Nervi was writing such articles as "The Art and Technique of Building,"
"Thoughts on Engineering," "Problems of Architectural Achieve- ment," and "Technology and the New Aesthetic Direction."ll He wrote no treatises on scientific analyses. Indeed, even a more technical paper written in 1939, "Considerations on the Cracks in the Dome of Sta. Maria del Fiore and on Their Probable Cause," serves to indi- cate something of the role that the ancient Italian monuments played for Nervi. His first book appeared in 1945 with the title ls Building an Art or a Science?l2 By this title he meant to imply, as he stated in his next book Structures (1955, English edition 1956), that struc- tures "can be solved correctly only through a superior and purely intu- itive re-elaboration of the mathematical results. "B To illustrate this intuitive approach, he told of his teacher at Bologna who in 1913 read his students "the alarmed letters of his German colleagues, who proved mathematically that the Risorgimento Bridge in Rome was in immedi- ate danger of failing-and in fact should have failed already-although the bridge had been built and was then in full use."14 In fact, the bridge still stood, and Nervi's point was that the designer, Fran~ois Hennebique, had not needed a complicated mathematical theory to create a beautiful and safe bridge.
Nervi began practicing after graduation from Bologna in civil en- gineering in 1913, but it was not until I9n, at the age of forty-one, that he completed a major structure on his own. During those nearly twenty years of practice, however, he gained design and field experience 176
FIGURE 10.l
Tiie Pantheon, R.ome, 124 A.u. 'I'his Roman de. . meters and remained the wideslãSllannin sign 111
unreinforced concrete spans 4-3.5 Br~lau dome. g conerete dome until surpassed in 19l3 by lhe
THE NEW AGE OF STEEL AND CONCRETE
with reinforced concrete, so that when he first began to design and build his own works they were the product of considerable maturity.
The most spectacular of these are certainly the domes and barrel shells he built between 1935 and 1959. These are also the works that best illustrate Nervi's preoccupation with very simple overall shapes made up of an interplay of individual elements. In his domes, these elements are ribs which make the overall dome both stable and light. This is precisely the tradition that one sees in earlier Italian domes. The two earliest shells of great size that still stand are the Pantheon (figure 10.2) and part of the Basilica of Constantine. In both cases, the heavy ma- sonry surface is lightened by a coffering, which results in an interior structure of two-way ribs, the type of system used by Nervi in his Little Sports Palace (see figure 10.3, p.181).lS Moreover, the two great domes of the Italian Renaissance, although differently made, are both ribbed.
St. Peter's has meridional ribs outside and decorations inside that re- semble coffering. The Brunelleschi dome in Florence has, between inner and outer shells, two-way ribs which cannot be seen but which are evident in the well-known drawings of the construction. Significant dome ribbing also characterizes the works of the Turin designer Guarini (1624-1683), whose crossways ribbing for St. Lorenzo in Turin appear as small-scale precursors to Nervi's modern works. Some of Nervi's most spectacular early works were in fact in Turin.
With this historical background, Nervi approached reinforced concrete in just the way Maillart had: both as a builder of competitive structures and as a designer of new forms. As Nervi put it, his .early experiences "had formed in me a habit of searching for solutions that were intrinsically and constructionally the most economic, a habit which the many succeeding competition tenders (almost the totality of my projects) have only succeeded in strengthening."16_Nervi's whole outlook was, therefore, inAuenced by the search for economy. He would have had almost no chance to build had his designs not been the cheap- est. At the same time, this economy was, for Nervi, intimately con- nected with finding "the method of bringing dead and live loads down to the foundations ... with the minimum use of materials." Economy of cost and efficiency of materials were, however, never enough, for as he continued, "I still remember the long and patient work to find an agreement between the static necessities. . and the desire to obtain something which for me would have a satisfying appearance." Nervi 178