HISTORICAL DEVELOPMENT OF "MODERN SHELL ROOFS 1_ INTRODUCTION When we plan long span roof structures, our choice usually is between long span steel frames and reinforced concrete thin shells or folded plates. With steel frames, we need additional material like roof sheets to cover the space, Shells or folded plates have the dul property of spanning and also covering, the space. Reinforced concrete thin shells can be defined as curved slabs whose thicknesses are small compared to their other dimensions like radius of curvature. Even though shell construction with steel and concrete is used in the industry for pressure vessels, reactors, tc. we will restrict our study to reinforced conerete shells used for roof construction. Many types of thin shell roof constructions are of recent origin and became popular only after World War 1l, even though progress was made in reinforced concrete after World ‘War |. This modern material (reinforced concrete) can be cast in any shape. Like stee!, it thas strength. Besides, it has the body to caver space. Steel by itself is rarely used to cover space whereas reinforced concrete is used to construct slabs, shells, etc. which in addition to spanning lengths can also cover space. This important property of reinforced concrete ‘an be taken full advantage in shell construction: “The concept of shell construction can be considered as a slow evolution made from Inasonry arches and domes, which were in use from the very early days of human existence. Brickivork could take only compression and no tension. Reinforced concrete is homogeneous and continuous and its action is different from that of brickwork made of individual bricks with mortar joints. Thus, the action of a brick arch with mortar joint is different from that of a reinforced concrete arch. Brick arch cannot take bending whereas fan RC. arch can resist bending also. It is a very interesting study as how, without the ‘bility of modern materials, our ancient architects were able fo design and constructDesign of Reinforced Concrete Shells and Folded Plates structures like large domes and vaults to cover large space for monumental buildings, In this chapter, we focus on the evolution of modem reinforced concrete thin shell structures. 1.2 ADVENT OF BRICK VAULTS Brick vaults (three-dimensional arches) are the oldest form of curved roof structure. As explained in Appendix A, they can be considered as an extension of the masonry arch. Arches are believed to have been invented by the Greeks around 5th century BC [1]. The use of bricks for roof became a necessity in old days as most of the earlier buildings built with wooden beams and columns tended to be burnt down by fire. Hence, humans invented the vaults made of bricks, which were safe from fire. Though the invention of vaults is generally credited to the Greeks, there is evidence to show that the Mesopotamians and Egyptians had also experimented with it, though sparingly, before the Greeks. But it was the Romans who used arches, vaults and domes extensively. 13 EVOLUTION OF BRICK DOMES Domes which can be looked upon as rounded vaults with circular or polygonal bases were the next to be evolved. Vaults are easy to build, but domes require much more skill and improved materials. Domes are three-dimensional structures subjected to special forces. The brick domes were invented by the Romans. They were fond of the circle and Used circular domes very freely. ‘The dome is three-dimensional in nature and has two types of stresses (meridinal and hoop stresses) instead of one as in a vault, It is as different from a vault as a plate is from a beam when used to span an area. This improved structural action of domes was used by the Romans for large structures like the Pantheon. However, these brick domes were massive structures. The Pantheon brick dome, about 48 m (142 f€6 in) in diameter built around 150 AD by the Romans, had its central portion about 1.2 m (4 ff) thick. Most of these domes had a central thickness of 0,025 times its diameter whereas a modern 50 m diameter R.C, dome needs only 60 to 70 mm in thickness, (A minimum thickness of 100 mm is usually adopted for placement of steel in all RC. shells.) From modern analysis, we know that in a concrete dome only a fully semicircular dome has compatibility with support conditions and that they develop hoop tension at points below about 51°.48' (say, 52°) below the central half angle. They also develop horizontal thrust at the base when built as segments less than 96° central half angle as ‘shown in Figure 1.1. These principles were not fully known in early days. The lack of this Knowledge of forces acting on a dome caused a large number of failures in domes, However, itis interesting to study how old master builders solved the problems. A short statement of this aspect is given in Appendix A. Itis worth mentioning that the building material which brought forth the structural innovation of the Romans was ‘Roman concrete’ described in Section 1.11ww 2 Fique 1-1. Evolution of domes: a) Arc, (b) Vault supported on wal (c Cour Remsphircal dame indicating tension below about 38° as shown, (A) Circular segmental dome, and (e) Ribbed dome 14 RIBBED DOME CONSTRUCTION Because of liable tension in semicircular domes and horizontal thrust in segmental domes, the ancient and Middle Ages builders resorted to special methods of construction. One ‘of the methods was to build a number of arches spanning the diameter and the space, in between filled with suitable materials like brickwork. For example, the dome in St. Andrews Church in Chennai on Poonamallee High Road in Egmore built by the British in 1821 (before reinforced concrete became popular) was constructed with brick arches spanning. across the diameter, with the space in between the arches filled with the Sicilian cones. (These are clay cones like the modem ice cream cones fitted into each other and arranged in rows between the arches and plastered, thus reducing the weight of the filling between arches.) This structure is still being used for worship in Chennai (See Appendix A for description of this dome). There are many other structures in the world which were built as ribbed domes (described in Appendix A). Even as late as 1911, the dome of the Melbourne Public Library in Australia was built of ribbed dome construction. Many famous domes of the world in masonry were of this construction. 1.5 ADVENT OF REINFORCED CONCRETE SHELLS ‘Though Lombart of France built a boat out of steel mesh and concrete (reinforced concrete) as early as in 1850, it was only just before World War I that reinforced concrete came into use. (Itis interesting to note that prestressed concrete became popular after World War Il.) The aircraft hangar at Corley in France built by Fressinet around 1917 shown in Figure 1.2 can be considered as the first real reinforced concrete shell construction. It was in theform of arches spaced at small spacings, The catenary shells spanned botween the arches. The arches were constructed from the ground (Foundation) level so thatthe reactions are taken by the ground. These arches were connected together in the longitudinal direction. by corrugations (shells) as shown in Figure 12 [1] ‘The mathematical theory of shells was not known to engineers till the membrane theory of shells (usually credited to G. Lame and G. Clapeyron) was published in 1828 (19th century. However, it was only as late as in 1923, about 100 years later, that Dischinger and Baversfeld used! rigar1.2_ the aan nny a Oy, Fane, themembrane theory to build a reinforced concrete "u's itt acer te tie dome over the planetariim in Jena. It was also Dischinger who, in 1928, rough to the notice of design engineers the concept of membrane theory for the design of shells. Another dome that Dischinger and Baversfield built is a hemispherical dome, only 30 mm thick and 16m in diameter. There is no herlzontal reaction as the ends are in 90° but it develops hoop tension below 52° which is rested by steel reinforcement Engineers soon realized that reinforced concrete, unlike sleel, can be made continuous and used to cover space, Reinforced conerete has both strength and body ant is a muck Superior material to cover large space in building construction. Steel and wood con be used as beams and colurmns but not as slabs or shells. With prestressing, we can cover ety large spaces. This concept has encouraged the present use of shell for covering very large areas like factories, hangars, assembly halls, ele. With the rising cost of ated orc concrete, shell construction is bound to become popular for large span roots 1.6_ THEORETICAL ADVANCES ‘Mauch progress was made in mechanics during the period around 1800 AD. The use of iron and steel brought in the method of theoretical design of columns and beams by engineers trained in mechanics to replace artisans {rained in traditional design by thumb rules obtained from experience in wood and rai, brickivork construction. As already stated, the membrane theary was applied to design of concrete shells only as late as 1923 by F. Dischinger. He also wrote a special chapter on cylindrical concrete shells fm the German. pgwy 1.3 yma stat (1) Sa, Edge ton, Handbook Eisenbetonban with U. fia) amas Gans, oe oie: Finsterwalder in 1928 [1]. The membrane ied art 0 ae)1 Development of Modern Shell Rools theory is a statically determinate solution and it advanced very much the analysis and design of RC. shells. "At first, interest in Western countries was limited to the design of cylindrical shells as shown in Figure 13. This was initiated by F. Dischinger and U. Fisterwalder. As Bending is also predominant in cylindrical shells, the bending theory of shells was evolved. However, solution to the bending theory is possible only with a number of simplifications. These were made by H. Schorer in 1935 [2] and also by RS. Jenkins in 1947 [3], The ‘American Society of Civil Engineers published a manual on ‘Design of Concrete Shell Roofs’ in 1952 [3]. The theory for design of many other types of shells was proposed in later years, Here, we, however, discuss only the basics of the subject to understand the fundamentals, Today, many computer softwares are available for the design of shells (hich has made this construction very popular. The aim of this book is not to describe The mathematical theories but to give a conceptual idea of the forces that act on the Various types of shells so that we can understand and interpret the computer output meaningfully. 17 COMPARISON OF OLD BRICK DOMES WITH MODERN RC. DOMES ‘The use of reinforced concrete has considerably reduced the weight of materials per square metre of area used in the construction of domes as can be seen from Table 1-1. TABLE 1.1. Comparison of Woights of Material Used (1) SNe. Ballaing Construction Dale of Span, Weightsg. Percent Construction in fet yar 1 St Peter's Dome Masonry dome 1 1a7—«12,500—« 100 2 Centenary Hall, “Ribbed RC. dome 1912215 375030 Breslaw 3 Market hall, Reinforced concrete 1933160 4803 Algecixas shell dome 4 Schawezwaldhalle Prestressed concrete 1953-240 2020 (Karistuche) saddle shell 5 Le Toureaut ‘Aluminium steel dome 1953 300 * 410.83 hhomisphere, Texas 6 — CNLT. Exhibition Double reinforced 1958720 sad Hall, Paris concrete shell dome 1.8 DEVELOPMENT OF RULED SURFACES AS SHELL ROOFS ‘The high cost of curved formwork is a major objection to the use of curved shells like Gireular domes and cylindrical shells in general. This has led to the development of ruledsurfaces for shells, like hyperbolic paraboloids and conoids, shown in Figure 1.4, for roof construction. For these thin structures, linear forms of formwork can be used for construction. ‘The reduction of the quantities of steel and concrete needed for shells as compared with ordinary slabs and columns is sure to make shells popular for covering of large spaces in the future. Hence, their study is important for any structural engineer. > Figure 1.4. Ruled surfaces: (2) Hyperbole paraboloid (Goubly ruled surtace), and (b) Cooid (singly rules surtace), 9 DEVELOPMENT OF FOLDED PLATES FOR ROOFING During the early 1920s, coal bunkers and similar containers were built in Germany without beams at the junctions of the plates (Figure 1.5). The concrete slab is folded and the reinforcement steel is especially detailed for continuity. This led to the development of the conception of folded plates, as shown in Figure 1.5, for roof construction also. The folded plate behaves as a slab in the horizontal direction and as a plate or beam in the vertical plane. This combined slab-plate action makes it a very economical roof structural ‘element. The use of moment distribution method to analyze these plates was proposed by G. Winter and M. Pei in 1947 [4]. Further refinement of the analysis by Yitzhaki [5] enables us to arrive at a good analysis of these structures [5]. The ACI publication by a task committee has given us a very good method to analyze these structures [6]. This has ‘made their analysis and design easy. Such construction became very popular in use as the formwork of this type of consteuction is simple (Figure 1,4), Today, in many factories, instead of cylindrical shells, folded plates are used. Even the use of prefabricated folded pln cast on ground and ited pn postion i lbo popular in the construction of roots over large areas o Figure 1.8. Folded plate constuction; (e) Goal bunker, and (0) Folde plate root wit‘ ype folded platesHistorical Development of Modern 0 GROINED VAULTS _ ‘The term ‘groin’ in architecture means “edge formed by inkesecting arches” as hoon Ht Figure 1.6, Intersecting surfaces made of parabolic arches are very beautiful to look at serie very popular in use, These groined vaults are used to cover large spaces The are ifterence between conventional shells and groined vaults is the edge effect 'n froined yaulls, as one curve merges into the other, there ae very lle ede disturbances Front iin hells (Infact, some of the sea shells like the one used for advertisement of sees company occur naturally inthis shape), These shells are widely used at presen for overing lange spans and are also amenable to conceptual design tsing the theory of arches, o inne 4.6 omposte stl: () Grint set with varios types oanslaanal (eine or hyerbok) te Gampesiesc'gone campuse of nil cynical sels, and (2) Combiration of skew hyperbole paratolods ove 2 polygonal plan J PROGRESS TN MATERIALS OF CONSTRUCTION We must zemember that the building material which led to the great structures of the Hee rr jas the invention of Roman concrete [7]. Very few engineers are aware that it etn strong material that enabled the Romans to build waults, domes, etc. of BRE beauty and ense. The important ingredient was pozzolane, a voleanic earth found in the ae aig wand around Rome, Romans mixed this pozzolanic earth with lime whieh, ase rotted, produced a very strong mortar. (This morlar sets under water ike modern eee) What is now known as Roman concrete consisted of layers of these mortars {410 cement) yale was Girmly pressed in, It was different from what we do today when aenent mortar and rubble are mixed together in a mixer before being placed in posi pene this Roman concrete that the Romans used in many of their important constructions Twas ths Rlumes, the advancement made with reinforced concrete by introduction of prestressing after World War II fibre concrete ete, has made it possible (0 make great progress in shell construction. 12 MODERN COMPUTER METHODS OF ‘ANALYSIS OF SHELL STRUCTURES in modern times, a large mumber of software are available for easy analysis of al types de oe ve cluding shell structures. However, unless one has 2 full physicalhe output of bine ts magnitude, Alo, some rough methods tens we also Pectefotmation, Later, matrix methods based on atin concepts evolved. This has nov deeueghitced by computer softvare mainly baer nn sor element analysis. In our analyanr nt i8 this book, we depend iore on conocrats behaviour and deformation fate of se ttireteal deiivations io eaatig a eet the magnitude and nature of the computer output exhibition nant body and strength, it is the fist chance Co constructing very large siruchan hiu®, sles and other poss of assenuhe cae choice for large span Se apis is throe-dimensional steel trusses, Howees, any steel structure behaves oniy we aietnal material ke aluminium sheets are requea eon, ign of important types of reinforced concrete A Sheth the following structural forms: () Arch, (i) Vaul, (Gi) Dome, (i) Hyperbotic Paraboloid, and (v) Folded plate construction, ai : REFERENCES _ U1] Zonan, Hilo The Development of the Modem Shell Roof, Australian Building Technology, Nov. 1963, 1a Seyere H. Line Load Action on Thin Cylindrical shells, Proc ASCE, vol. 61, 1535,[3] Jenkins, RS, Theory and Design of Cylindrical Shell Structures, O.N. Arup, London, 1947. [4] Winter, G. and Pei, M., Hipped Plate Construction, Journal A.C.t,, vol. 18, 1947. [5] Yatzhaki, David and Max Reiss, Analysis of Folded Plates, J. Struct Div, ASCE, vol. 188, Oct. 1962. [6] Phase I Report of Task Committee on Folded Plate Construction, J. Struct Div, ASCE, vol. 190, December 1963. [7] Banister Fletcher, A History of Architecture (Revised by R.A. Cordingley), University of London, The Athlene Press, 1967. [8] 1S-2210-1988, Criteria for Design of Reinforced Concrete Shell Structures and Folded Plates, BLS,, New Delhi.COMMON TYPES OF SHELL ROOFS AND DESIGN BY WORKING STRESS METHOD 2.1 INTRODUCTION We have seen that the shell is a thin curved slab whose thickness is small compared to its other dimensions like span, and radius of curvature. Usually, the thickness /radius ratio (¢/7) will be 1/100 to 1/500. In nature, we can find shells of all shapes and dimensions, such as sea shell and, egg shells. Generally, they are curved in both axes and do not form a definite geometric pattern. It is said that “Perhaps one of the most beautiful structural shape in nature is the shell of a bird's egg”. However, in the mathematical theory of shell structures as applied to engineering design, we have to restrict our study only to shells that are formed by definite geometrical shapes such as circle, parabola, hyperbola, and ellipse. In this chapter, we briefly describe common types of shells used for roofs in civil engineering practice. IS 2210-1988 gives a summary of these shells [1]. Shells come under the category of stressed skin structures. The common shapes of shell roofs that we deal in this book are shown in Figure 2.1 to Figure 2.6 and Table 2.1 We will also review design of reinforced concrete members by working stress method 2.2 GENERAL CLASSIFICATION OF SHELLS We generally classify shells into the following classes (Table 2.1): 1. Shells of revolution 2. Translational shells, 3. Ruled surfaces 4, Composite shells. [To these, we will add folded plates. Even though these are not true shells, these are also thin slabs and are usually associated with shell roofs.] Each class of shells is discussed in the following sections. 10 ee Ee eeeCommon Types of Shett Roofs and Design by Working Stress Method ui TABLE 2.1 Classification of Stressed Skin Roof Structures eel Polygonal domes Shell roots Folded plates Pyramidal roots Shells of Shells of Ruled Composite shells -Funicular revolution translation surfaces (Groined vaults) shells Circular domes 1, Hyperbolic paraboloids 2, Conical domes (bounded by straight lines) 2. Conoids 1, Cylindrical domes (a) Circular cylindrical domes (b) Parabolic cylindrical domes {(€) Elliptic cylindrical domes (2) Catenary cylindrical domes 2. Hyperboloids (@) Hyperbolic paraboloids (bounded by parabola) (b) Hyperbolic paraboloids (ounded by straight lines) (©) Circular paraboloids (4) Elliptic paraboloids 221 Shells of Revolution | The middle surfaces of shells of revolution are formed by rotation of a plane curve K or a line about the vertical = axis in its plane as shown in Figure 2.1, The generating curve is called the meridian, If the meridian is a straight line rotating around a vertical axis, we get a cone. If it is a vertical circle rotating about a vertical axis, we get a sphere. If the meridian is a parabola, we get a paraboloid of revolution, ete. (We will study the analysis of circular domes and conical domes which come in this category.) 22.2 Translational Shells ‘The middle surface obtained by the translation of a moving plane curve Ky, called the generator (generatrix), which moves parallel to itself over another stationary curve Kz, ‘alled the directrix, (the function of these being interchangeable) gives us a translational12 pesign of Reinforced ¢ ete Shells and Folded Plates shell a8 shown in Fighite 2.2. The directions nnd he magnitudes of the curvatures 1/ Re se Re tend of curvatures Ry andl Ra) of Ki and XK, are very important. The properties a hece ahelle are explained below. Thus, the cunies Toy ‘point in the same or opposite Ginections, When they are in the same direction, Whey 1 ‘chilled clastic or synclastic shells. When they are in opposite directions, they are ried aticlastic shells. The following shells come in this category: m@Dat o © 124 ataoal sets (a) Spl orelar dome, (b} Dame wh st ah {6) Ribbod dome, (6) Conical dame find (e) Conical umbratia root Fi 1. Cylindrical shells: Thus [Figure 2.2(0)] or formed by a straight line generator coving along, a specified directrix. The common fYPSS ‘af cylindrical shells are the following with straight line generators: = Circular cylindrical shells—Circular arc directx + Cotgnary cylindrical shells—Catenary directris 1 Parabolic cylindrical shells—Parabola directrbs 2 Elliptic cylindrical shells—Semi-ellipse ‘The layout of the curves will be as shown in Figure 24a). >, Parabeloid shells: As shown in Figure 2.2(b) and {o),a parabola with downward earaeeare moving on another parabola, with CArdaet fn the same direction will produce a circular paraboloid if the parabolas are equal. It will produce an elliptic Varaboloid if they are tunequal. On the owner ‘hand, with two parabolas of opposite Poremures, one moving on me other will produce & hyperbolic paraboloid (Hypa) shell. 4, Toroidal shells; These special shells of evoiation. have the generatrix of a much Sharper curve than the directrix as shown in FBT" 2.2(d). These can be used as 2 shar Per ected ag multiple abells to proaiiee very SE ‘curved areas like aircraft Rangars, Spans up to 100 m have been built ‘with these shells. A ADEA A @ igure 2.2 Translational sels: () Grey cynical Se aie (or scela)paraboot sh, @) Hyperbole rane sel () orldal ste and () Dieert PES ot vap-elipse, ice, catenary, p2rabolCommon Types of Shelt Roofs ana Design by Worki stress Method 13 223 Ruled Surfaces (Hypars and Conoids) Ruled surfaces are surfaces which can be generated entirely by straight lines. They can be singly or doubly ruled surfaces. A surface is said to be singly ruled if at every point on its surface a single, straight line can be drawn along the surface. The surface is called a doubly ruled surface if at every point on the surface two straight lines can be ruled. It will be shown in the chapters dealing with hypar and conoid shells that though these surfaces are of negative Gaussian curvature, a hypar can also be obtained by a straight line moving over two inclined straight lines and that the conoid can be obtained bya straight line moving over two curves at its ends as shown in Figure 2,3. Conoids are singly ruled surfaces whereas hyperbolic paraboloid is a doubly ruled surface ruled in two directions. Cylindrical shells can also be classified as a singly ruled surface. @ ° Figure 2.3 Ruled surfaces: (a) Hyperbolic paraboloid with straight edges (doubly ruled surface), and (b} Canoids (singly rulea surface). 22.4 Composite Shells (Groined Vaults) Composite shells are combinations of two or more shells as shown in Figure 24. Individual hyperbolic paraboloid surfaces or cylindrical shells can be joined together to form many surfaces. @ ® @ Figure 2.4 Composite shells: (a) Groined shells formed by c/lindrcal stills or hyperbolic parabolods, (b) Polygonal ‘dome with evindrcal shell, and (c} Combination of skewed hyperbole paraboloid 22.5 Folded Plates (Hipped Plates) ‘We have already seen in Section 1.9 that folded plates (Figure 2.5) are a series of inclined slabs connected together and supported at their ends. Though folded plates are not curved and cannot strictly be called shells, they are thin structures and are usually studied underthe subject of shell structures. These are sometimes called “stressed skin” structures. Nowadays, more of these structures are used in common practice because of their simplicity in formwork. Peg ret @ Figure 2.5 Folded plates: (a) V type, (b) Trough type, and (c) 2 (northlight) type. 2.2.6 Pyramidal Roofs Pyramidal roofs, as shown in Figure 2.6, are not pure shell structures. As they are also made of thin slab construction, we will briefly study its design. Figure Pyramidal roots, 2.2.7 Catenaty and Funicular Shell Roofs ‘The word ‘funicular’ means “of a rope of its tension”. In most shells, the principal load is the dead weight. If we shape the shell like a citenary, the shape a rope will take under its own weight, there will be no bending, (However, this is true for only the dead weight loading.) Resistance of these catenary shells against lateral loads can be improved by corrugating them. Shells of single curvature of the shape of a reversed hanging rope are called catenary shells, Funicular shells are of double curvature. These can be obtained by stretching a piece of hessian on flexible fabric along its boundaries (scaled down to the plan of the building) and loading it with a wet layer of plaster of paris. A short description of these is given in Appendix B. (Details of design is described in References 2 and 4. Reference 2 gives a chapter on funicular shells.) 2.2.8 Gaussian Curvatures ‘The product of the two principal curvatures with their signs is called the Gaussian curvature (G = 1/R, x 1/R,) at any point on the surface of the shell. This can be positive, zero orcommon Types of Shell Roofs an Design by Working Stress Method 15 negative, The surface is termed synclastic developable or antielastic depending on whetnet Te Gauss curvature is positive zero or negative, respectively. We will see later on that tne of the imporiant properties of the shell is given by the Gaussian curvature and its Sitect on the propagation of edge disturbances in the shell. For shells with positive Gauss chevature like domes, the edge effects tend to dampen rapidly and are confined to 2 svat sone of the shell. This makes the membrane theory of these shells (such a5 domes) Taha throughout the shell, except for a small portion near the edge member of the shel) Yer.a shell of zero curvature like the cylindrical or Conical shells, the dampening is not tapid but extends further into the shel than in the case of shells of positive Gauss curvatire. Huwover, in shells of negative Gauss curvature like the hyperbolic paraboloids, the boundary artes due to edge members penetrate further into the shell. This information gives us cn sdea of the extent of reinforcing of the shell portions near edge members. (Circular Shindrical shells are produced by a straight line moving ever twe equal circles af its hae Its Gauss curvature is zero. This is the case with conical shells also.) 23 TYPES OF SHELLS DISCUSSED In the ensuing chapters, we will confine ourselves to the following types of shells. Though there are many other types, a study of these shells will give us an understanding of the ware cof shell structures so that further studies can be made by referring to advanced textbooks on the subject: 1. Circular domes 2. Conical shells 3. Cylindrical shells 4. Hyperbolic paraboloids 5. Elliptic paraboloids 6. Conoids 7. Groined vaults 8. Polygonal domes 9, Folded plates 10, Pyramidal roofs ‘The aim is to make the readers familiar with the behaviour of these shells. An cuborate analysis will require more exhaustive treatment of these subjects 24 SELECTION OF DIMENSIONS OF SHELLS 18 2218-1988, Section 7 gives us some guidance in choosing the preliminary dimensions tthe various types of shells [1]. We will review it in the various chapters which dealing with various types of shells. 25 GENERAL METHOD OF STRUCTURAL ANALYSIS OF SHELLS Inthe analysts of shes, we will study their behaviour under two heads, namely, membrane analysis and bending analysis, Such a procedure can be used to understand thels base Mthtieur, Today, with the availability of computer software, we can solve the problems16 Design of Reinforced Concrete Shells and Folded Plates much more easily. But the basic understanding is needed to interpret the results for detailing the steel reinforcement and also for designing boundary elements of the shells. 1. Use of membrane theory: As a first step, we assume that all the loads are taken by membrane forces or direct forces, namely, direct tension or compression along the radical directions and shear without bending and torsion. As there are only three forces, the three equations of equilibrium will give us all the forces. It is a determinate or primary system. 2. Determination of errors towards the boundary: The forces and also displacement got by membrane analysis need not be in equilibrium or compatible with the actual situation at the edges of the structure, where the shell is restrained by edge members, Hence, we find the errors af the boundaries due to membrane stress resultants. 3, Corrections: Next, we find the corrections that have to take place due to unit edge effects (forces and displacement) applied at the boundaries. For this, we used the bending theory of the shell. 4, Compatibility: Next, we achieve compatibility by computing the magnitude of the edge effects necessary to eliminate the errors. 5. Resultant forces: The sum of the membrane forces and the compatibi gives the resultant forces. ty conditions ‘The above method for a complete solution of shell structures is similar to the analysis of indeterminate linear structure, where we first make the system determinate and then find out the necessary force resultants to make the forces and displacements compatible with the real structure, We will use this method to study the bending analysis of cylindrical shells. Alternately, we can also derive the general expression for total analysis of shells by matrix methods. General expressions for the combined membrane analysis and bending, analysis of all types of shells will be quite involved [2], [3]. Alternately, we can use the modem finite element analysis using computer programs [3]. Instead we will take the analysis of different geometric shapes of shells and show how the membrane analysis of these shells can be made. The bending analysis will be discussed with the help of published Tables for cylindrical shells and hyperbolic conoids. 2.6 STRUCTURAL ANALYSIS Itis easy to make membrane analysis of shells as we are analyzing a determinate structure. Although it is difficult for average structural engineering students to master the mathematical intricacies of the bending theory of shells, itis not so for them to understand, the underlying static principles of shell action. By understanding and applying the principles of RC. design, it is possible for an average student to plan and undertake practical design of simple shell structures with as much ease as with many conventional types of structures, We must remember that in many branches of structures, as in the case of flat slab and shells, their construction took place well before the exact theories of shell design were developed. Hence, intuitive design should also be used in all design, detailing and construction of shells.Common Types of Shell Roofs and Design by Working Stress Method 17 2.7 STRUCTURAL DESIGN BY WORKING STRESS METHOD [As shells are thin structures, it is considered better to use the principles of elastic design. We also use M20 or M25 concrete to have good strength but not too much shrinkage which can produce cracking. The use of Fe 215 stee! is ideal for shells. However, we use the readily available Fe 415 (out not Fe 500) for economy. The usual 15456 (Appendix B) recommended design values as given below are to be used. |. Loads on shells (see also Section 5.5): Shells are to be designed for all loads (dead, imposed and wind loads) as specified in 1S 875, parts 1 to 3. Generally, we adopt the live load as that specified for “roof without access” equal to 75 kg/m’ for sloping roof less than 10°. (For roofs sloping more than 10°, it is 75 minus (0-10) kg/m’) Hence, for a shell surface we may use a maximum live load of 75 kg/m? plus a water proof 15 kg/m? ‘Accordingly, a total addition load 70 to 100 kg/m? in addition to dead load will he satisfactory (Pischer in his book uses a live load of 40 kg/m* and insulation of 12 kg/m? a total of 52 kg/m?, Ramaswamy uses a total live load of only 15 Ibs/sqft (= 70 kg/m? including waterproofing). [Note: We should note that the live load in always assumed to act on the projected area of the shell surface which is less than the surface area of the shell. Hence when the specified live load is combined with dead load, its value in theory need to be only less than the specified value] Regarding wind loads as shells are curved addition wind load is not generally taken for shells, suction loads will only reduce the stresses, INote: Examples in this books use various amounts of live loads.] 2, Allowable direct and bending compression in concrete for design by working stress method (Blastic design) 18 456 (Annexure B) gives the following values for elastic design of reinforced concrete. TABLE 2.2 Allowable Stresses in Conorete (Wimm?) Concrete grade Direct compression, Beniling compression, ‘Shear fe (o) (oa) in concrete M20 3 7 0.18 to O51 M25 6 85 0419 to 057 M 30 8 10 0.20 to 0.60 (From Tables 21 and 23 of 15 456) INotes: ‘The above direct compression is about 0.25 fk “Allowable compression should be less than the critical stress for buckling (Chapter 19). Value of modular ratio m = 280/30, (or 280/fek). Allowable shear depends on percentage of longitudinal steel (IS 456 Table 23). Values given above are for percentages < 0.15 and 2.0]forced Concrete Shells and Folded Plates 3, Allowable tension in steel (f.) (@) For Fe 250 steel up to 20. mm dia. 140 N/mm? and above 20 mm. 130 N/mm’, {) For Fe 415 steel for all diameters 25 N/m’. Note: As shells are thin structures in which comosiog of steel is very dangerous, it is errspte to use the modem TMT bars. Tor steel should be “avoided in shell constructions.) 4, Design constants for bending: For design for bending 256 the following formulae ere the depth of section and amount of steel required. M M= K(onbd?) and A= FG 5 (one?) an Gh (ow = allowable bending stress given above) (@) For Fe 250 steel and all grades of concrete, we Wee K = 0.174 and j = 04 (O) For Fe 415 steel and all grades of concrete, we use K = 0.131 and j = 0:90. With Fe 415 and fo: = 20; M = 0.917 bi? ‘We may use SP. 16 for this purpose. 5 Cleat cover, 15 2210-1988 specify a clear cover of AO Jess than 13 mm or diameter ctee bar whichever is more. (Generally, we provide 9 ‘lear cover of 15 mm for Orn” bs mm for beams, 40 mm for columns and 50 mm $0 for footings according to TS 456, Clause 26.4.) 6. Concrete mix: Moder specification always stipulates the grade of concrete to be conttfor shell construction. We specify a mix like 1 © 4 corresponding to grade M20 concrete for medium shells and 1: 1.5: 3 corresponding to grade M25 concrete for large shells. The maximum size of ageresates ‘should be 20 mm. It should be 12 ween for places itis difficult to place the concrete, I is ‘also better to work with a low river cement ratio using superplasticizer than use @ Very ‘rich mix for satisfactory workability. 28 DETAILING OF STEEL Concrete should easily flow to fil the full thickness of re shell at all points. This is very fmportant. As shells are thin structures, its very important to see that there is no congestion epPteel in any part of the shell which will prevent free flow of concrete during the ot astruction of the shell. There have been cases of colltpee af shells due to steel congestion werich prevented free flow of concrete, Crowded placement ‘of steel should be avoided. ee reel reinforcements in each type of shell should Be. eiailed according to standard. practices for that particular type of structure. ‘as continuity of concrete is very important Piinin concrete, especially in compression, the “onstruction joints should be properly planned. Similarly, removal of shutting i syells should be carefully carried out so that Fre shell action takes place for stability of the structure Mt may be pointed out that in the early days of Use ‘of cylindrical shells in India for large span warehouses, it was the practice 10 '2y ‘concrete in alternate layers to reduce shankage. Unless the joints are carefully laid, (he warious parts can act as very thin Shenk Be. oval of formwork. Infact, there was a case of failure in this type of constructionCommon Types of Shetl Roofs and Design by Working Stress Method 19 where the shell buckled as a thin arch, This practice to reduce shrinkage is not to be recommended as it is better to take care of shrinkage by extra reinforcement, if needed. 281 IS 2210-1988 General Recommendations The following general recommendations are given in IS 2210 for detailing of steel in shells and folded plates. The following rules apply to the general body of the shell: ‘The minimum steel diameter must be 8 m. ‘The maximum steel diameter should be 1/4 thickness of the shell or 16 mm, whichever is smaller. In the thickened portion, we may adopt the same rule as in slabs. The maximum spacing of steel should be limited to five times the thickness of the shell and the area of the unreinforced panel should not exceed 15 times the square of the thickness of the shell (more details are given in Chapter 9). References 4 and 5 gives more details of the common types of shells used. _ This chapter gives an introduction to some of the commonly used shapes of shell roofs and their classification. It is summarized in Table 2.1. This chapter also gives a summary of the allowable stresses and loads to be used in the design of shell structures. 1. Give a short account of the classification of shells. 2. Define Gauss curvature. What is its importance in shell design? 3. Give the sketches for at least eight types of reinforced concrete shells and three types of folded plates (refer IS) that are commonly used in practice. [1] 1S 2210-1988, Criteria for Design of Reinforced Concrete Shell Structures and Folded Plates, BIS, New Delhi. [2] Ramaswamy, G.S,, Design and Construction of Concrete Shell Roofs, McGraw, New York, 1968. [8] Bandyopadhyay, J.N., Thin Shell Structures Classical and Modern Analysis, New Age International Publishers, New Delhi, 1998. If] Fischer, L., Theory and Practice of Shell Structures, Wilhelm Ernst and Sohn, Munic 1968. [5] Chatterjee, N.K,, Theory and Design of Concrete Shells, Oxford and IBH, Calcutta, 1971.CLASSICAL METHOD OF ANALYSIS OF SHELLS 3.1_ INTRODUCTION There are many ways to study the theory and design of shell structures. A purely mathematical approach is difficult. Simplified mathematical study combined with a conceptual approach is the one most suited for beginners like students. Hence, we first take up the membrane theory which deals with the theory of shells whose bending rigidity can be neglected. This theory is applicable to a wide variety of shells. We then examine the shortcomings of this theory in the case of practical shell construction, and hence the need for the bending theory. Each type of shell gives a different expression for the membrane and bending theories. Thus, we will have different expressions for domes, cylinders, conical shells, hyperbolic paraboloid, etc. In this introductory chapter, we examine briefly the general mathematical principles of analysis which are the same for all types of shells. 3.2_STRESS RESULTANTS IN A SHELL Shells are different from slabs. We have shells like an egg shell which is closed in shape or a sea shell which is not closed as in a roof. We will consider only thin shells which are used as roofs. Other types of thin shells and thick shells as used for industrial purposes and in atomic reactors are also important, They are very much different from these shells used in roof construction and we will not deal with them. In general, there can be ten (Which can be reduced to eight) stress resultants acting on a shell. These can be classified into the tangential group and normal group as shown in Figure 3.1. 20Analysis 21 & o ® ee Figue 3.1 Eight stress resuitants acting on a shell elamant: (a) Coordinates, (b) Membrane terces (3 numbers), (¢) Bending moments (2 numbers}, (4) Normal shears (2 numbers), and (o) Twsting moment (one number) (Total 8 forces). Ty Ty Sry Sy Tangential group (4 nos.) reduces to 3 as S,, = Sy, [Two direct stresses and one shear, three in number.) Mr Qe May My, Qy My are two moments, two normal shears and one torsion.] ‘This reduces the total number of indeterminates to eight in number. [Notes: 1. ‘The tangential (N) group is direct stresses and M group, the bending group. My, is the twist group. Generally, tension is taken as +ve and compression as--ve. (But the opposite signs are also used when the forces are mainly compression as in spherical domes.) .-M, increases to M, + AM, along X-axis, Similarly, M, along Y-axis. . We may use the right hand screw rule (with palm facing up) for representing the direction of moments as shown in Figure 3.1. With the first finger moving forward, the ‘movement of the right hand thumb shows the direction of the moment. In shells, moments producing tension in the inner fibres is considered positive, 4, Shears are considered positive when they create tension in the direction of increasing the value of x] }: Normal group (6 nos.) reduces to 5 as M,y = M,, [The five components 33 STRUCTURAL ANALYSIS Case 1: In the membrane analysis, we take that the bending and twisting moments are zero. Thus, in the tangential group, there will be only three unknown stress resultantts Ty,22 Design of Reinforced Concrete Shells and Folded Plates T, and S. These can be determined by the three conditions of equilibrium of forces in the three directions. This is the membrane theory. For example, in the membrane analysis of a cylindrical shell (Figure 3.1) we have to find T,, Ty and S. We first sum up all the forces in the x (longitudinal) direction, then in (transverse) direction and also in the radial ditection (the loads always act in the vertical direction). From these three conditions we can solve for T,, Ty and S. This is further explained. in Chapter 5. Using stress-strain relationship, we can also derive the deformations due to these membrane forces in the various directions. Case 2: In the bending analysis, if we take the middle surface as reference, the bending stress resultants acting on the shell can be reduced to five stress resultants. This reduces the total stress resultants to 3 +5 = 8. This simplification will lead the general bending theory of shells to an eight degree partial differential equation with eight unknowns. This, bending theory of shells is given in the next section. 3.4 BENDING ANALYSIS OF SHELLS Edge disturbances in shells produce bending. In bending theory of shells, in addition to simple statics, we have to use deformation conditions also to solve the problems. The quantities we have in the case of a cylindrical shell are as follows [1]: 1, Strains: 4, 4—direct strain (2 nos.); #-shear strain (1 no.) 2. Displacements: u—in longitudinal direction; vin transverse direction; w—in radial direction 3. We also have K,, Ky—curvatures (2 nos.) K—twist (1 no.). We then proceed as follows for general bending analysis of shells: 1. Write down the force equilibrium equations. 2. Write down the stress-strain relations, 3. Write down the strain-displacement equations, 4, Write down the stress-displacement equations. The resultant simplified equations after the above steps will give three simultaneous partial differential equations in u, v and w. By successive differentiation and elimination, ‘we reduce the three partial differential equations to a single 8th order partial differential equation with only one unknown displacement, say, w. The resultant equation for circular ‘cylindrical shells is known as the Donnel-Karman Jenkin (DKJ) equation. If we take 1 and displacement was variable, the equation can be represented as follows [3]: Ow, Fw 6 dw 4 ow, dw] aw ant aoe astagt ‘aae ae | ae As in the case of membrane analysis, in bending analysis also we can derive not only the stress resultants but also the expressions for the deformations of the shell.Classicat method of Anatysis of Shells 25 3.5 DEFORMATIONS In many cases of analysis of statically determinate structures, deformations are not considered as important. But in statically indeterminate system, no structural problem will be completely solved unless one has determined the corresponding deformations also at important points of the structure. Both compatibility of stresses and compatibility of deformations at junctions of elements to which we cut the structure have to be achieved. Hence, in the highly indeterminate shells when using the membrane analysis for shells, we are interested not only in the membrane forces but also in calculating the membrane displacements. Formulas can be derived for these deformations also. Similarly, we should be able to calculate the displacements produced by other bending forces also in the shell. We will see in Chapter 8 on analysis of cylindrical shells with edge beams that, it is through both the compatibility of stresses and deformations that we analyze shells with edge beams. ‘The pure mathematical approach, for example, to analyze a cylindrical shell will be to treat the structure as a whole as in References [1] to [3]. On the other hand, we may also analyze the indeterminate structure by first making it determinate, (under the action of membrane forces only), find the stresses and deformations at the edges. Then we calculate the additional stress resultants that are formed by making the stresses and deformations at the boundaries of the shell compatible. The sum of the membrane forces and the stress resultants obtained by making the edge deformations compatible will be the required stress resultants for the design. This method is followed in References [4] and [5]. We use this second approach in our study of circular cylindrical shells. ASCE Manual No. 31 gives us the tables for such an analysis and we will use that publication for our references. It is also good to remember that the edge disturbances in shells of positive Gaussian ‘curvature (like spherical domes) die down fast and that in shells with negative Gaussian curvature, the edge disturbances extend to more lengths into the shell from their edges. This will give an empirical approach to the design of shells telling us where we should be careful and where we can make approximations. Be — ma ee This chapter deals with the classical method of analysis of shell roof structures. Both membrane and bending theories are necessary for a complete analysis of the shell. References [1] to (7] deal with the analysis of R.C. roof shells in great detail Hae _ REVIEW QUESTIONS 1. Distinguish between “membrane forces” and “bending forces” acting on a shell.4 Design of Reinforced Concrete Shetis and Folded Plates How do you find a solution for the membrane forces? Indicate how you derive the differential equation for solving stresses and displacements in a shell. Indicate the nature of the differential equation with reference to a circular cylindrical shell. Explain how equations for displacement of the shell for membrane analysis are important in the final analysis of the shell. What is the difference between the membrane analysis and the bending analysis of shells? How many forces are to be solved in the first and how many in the second? P) G1 i {51 (61 ia Chandrashekara, K., Analysis of Thin Concrete Shells, Tata McGraw Hill, New Delhi, 1986. Ramasamy, G.S., Design and Construction of Concrete Shell Roofs, McGraw Hill, New York. Jenkins, RC, Theory and Design of Cylindrical Shel! Structure, The O.N. Arup Group of Consulting Engineers, London, 1947. Design of Cylindrical Concrete Shell Roof, Manual No. 31, ASCE, New York, 1952. Billington, DP., Thin Shell Concrete Structures, McGraw Hill, New York, 1965. Bandyopadhyay, J'N., Thin Shell Siructures, New Age Publications, New Delhi, 1998. Chatterjee, N.K., Theory and Design of Concrete Shells, Oxford and 1BH, Calcutta, 1971.DESIGN OF SPHERICAL DOMES AND CONICAL ROOFS 41 INTRODUCTION Domes are shells of revolution. They can be of many types and shapes. Many of the old. domes were ribbed domes built from masonry. They are very different from the modern RC. domes, Appendix A gives some of the details of the construction of a few of these ancient domes. With reinforced concrete, we can design and build domes with much more ease. Figure 4.1 shows a few of the different shapes of reinforced concrete domes that are generally used. First, we have the simple domes. An Ogival dome shown in Figure 4.1(a) has the shape of a Gothic arch. Polygonal domes are made of sections separated by ribs > BA a 3 am & o Figure 4.1 Different ae Nba sve ain io ‘Opival dome, (b) Conoldal dame (c), (d Elevation and plan of regular polygonal dome, and (e}(?) Elevation and plan of polygonal dome with lrcular sectors, 25Design of Reinforced Concrete Shells and {also called hips) with a foot ring at the bottom. As the transfer of forces is through the ribs and not as hoop tension, the restriction of central angle of 52° to avoid tension (that we will study later in this chapter) does not apply. Conoidal domes are formed by the rotation of circular arcs not about the vertical axis but with axis parallel to it and some distance away from the central axis as shown in Figure 4.1(b). Simple domes are surfaces of revolution formed by a simple curve about its vertical axis. The curve can be a circle (producing a spherical dome), an ellipse (producing an elliptical dome), and a hyperbola (producing a hyperbotic dome). Conical domes are formed by an inclined straightline fixed at one point and moving over a circle at the other end. We will restrict our study in this chapter to simple spherical domes and conical domes. Conical domes are commonly used as roofs, especially when they have to support a heavy load at the top. Most of the domes built over monumental buildings are polygonal clomes as shown in Figure 4.1, This reduces hoop tension considerably. ‘We must be aware that as domes are shells of positive Gaussian curvatute, the edge disturbances do not travel large distances into the shell from the edges (Section 2.2.3) Hence, the membrane theory can be used for reasonable design of a large part of the shell. The edges should be thickened and specially reinforced to take care of the bending due to edge disturbances from membrane action. References [1] to [5] given at the end of this chapter can be used for a detailed study of the subject. INotes: 1, We should remember that angle @ in domes is measured from the crown. This is different fzom cylindrical shells where ¢ is measured from the edges. 2. In this chapter, we have taken compression as positive and tension as negative as is used in many standard textbooks for domes. In chapters on other shells, we have taken compression as negative. The signs used are indicated in each chapter.) 4.2 SPHERICAL DOMES 4.21 Planning of Spherical Domes The following properties of spherical domes are important in their planning. When we consider domes under dead load, semicircular domes with full9Q° half central angle will have no incompatibility with the reaction from the support it rests on. There is no ring tension. However, below the central angle of about 52°, it will experience hoop tension (On the other hand, segmental domes with less than 52° central angle will be fully in compression (meridinal and hoop compression), but the ends will require ring beams to take the horizontal component of the meridinal compression. (This is further explained under Section 4.2.3.) Similarly, if we have a circular dome with an opening at its apex, there should be a ring beam at the top to balance the horizontal compression of the meridinal stress at the top (see Section 4.2.5). It is also important that the shell should be so designed that concreting is easy. It is very difficult to lay concrete on any slopes greater than 30°-40° with theDesign of Spherical Domes and Conical Roofs 27 horizontal without back formwork. Hence, we always plan for a half central angle equal to or less than 40°, With higher rise, the stresses will be low, but the slopes will be large. For a 40° slope, the rise will be 1/5.49 span. For a 35° slope, the rise will be 1/6.34 span. For a 30° slope, the rise will be 1/7.46 span. Accordingly, a rise-span ratio of 1/5-1/7 is usually used in practice. Even a rise of 1/8 span gives easy concreting and can give low stresses. The minimum thickness specified by IS 2210 is 50 mm but for practical purposes, we should adopt at least 75-100 mm. Spherical domes up to a span of 60 metres can be built with R.C. ring beams. Larger domes will require prestressing of ring beams. Large monumental domes over important buildings are usually built as ribbed or polygonal domes with ornamentations and not as simple domes. Simple R.C. domes are commonly used for commercial purposes such as cover for water tanks. 42.2_Membrane Analysis In shell design, dead load is taken as acting along the curved length of the shell, and live load is taken as acting on the plan area only. Tt can be shown mathematically for a uniformly distributed dead load that the forces developed in a spherical dome under symmetric loading will be only the meridinal stress N, (along the meridians or longitudes) and the hoop stress along the latitude or horizontal bands as shown in Figure 4.2. As uniform dead load on the shell will produce the same Ne along the latitude, we can infer that the shear stress in such a surface will be zero under such uniform loading. Hence, we assume that the only two stresses that act on the main body of the spherical shell are the following. (In spherical shells, the angle @ is measured from the crown. We will see in Chapter 5 dealing with cylindrical shells that the angle ¢ is measured from the edge or springing.) Ny—Meridinal stress Ng—Hoop stress jute 4.2 Membrane forees in a spherical shel: (a) Directions of membrane forces, and (b) Symbol # and + (ierent ‘rom cylindrical shel) There can be a number of cases of shell analysis for various loadings such as: 1, Uniform dead load (DL) on a full shell surface with a central load on the apex 2. Uniform line loading over surface on top of a shell with skylight 3. Uniform live load assumed constant over the projected area (plan area) of shell surfaces 4. Shell with thickening of the edges and consequent variation of dead load [Notes: As the slopes are generally below 40°, the effect of winds will be suction. Hence, the wind load analys is carried out only for large and important shells only.In this chapter, we deal only with the first two cases. It is a common practice to adjust live load and even wind load to be considered as an equivalent dead load. But we must remember that non-uniform loading can produce shear stresses also. Wind loads on surfaces which are horizontal or inclined up to 40° to the horizontal (depending on height/width ratio) can produce suction, which is usually neglected as it reduces the stresses in the shell. 4.2.3. Case 1: Membrane Analysis for Uniform Dead Load and a Central Crown Load Notes: 1. As we try to make most of the forces in domes as compression, mtost literature n domes take compression as positive and te jon as negative. (This is different from the convention used in the analysis of cylindrical shells as given in Chapter 5.) @) co) Derivation of Ny at @ with dead load w and a concentrated load W at the crown (Note @ is measured from the crown): From statics, the vertical gravity load should be equal to the vertical component of Ny Surface area (of a dome of radius r and of height 2 where f= R(1 ~ cos 9) Resin 6. Let w = Weight per unit area W = Concentrated load at crown Surface area of a dome = 22k (height) Height ‘h’ at ¢ from crown = R(1 — cos @) (Half span of dome = r= R sin 4) Consider a circle at angle @ from the crown. From Figure 44, resolving Ny vertically and horizontally, with r = radius of circle at @ Ng sin 92x R sin ¢) = Vertical load from top = w(2aR) x R x (1 — cos @) + W As sin? @ = (1 - cos* 9), we get wRO=cos @) ,_W sint@ "2a Rsin?™@ wR w Treos ” 2aRsin™ 9 This is the meridinal force per unit length and its value increases with @ values (depth). It is always a compressive force. Value of No: Now, the component Ny in the radial direction acting inwards and also the inward component of the dead weight produce Ny By equating forces in the radial direction, we can arrive at the following formula: No= wR cos 9 - Np (42a) A= 2nrh, and an Ne or (ta)29 wR Tvcos 9" 2eRsin® 9 This is the hoop force per unit length, and the value of function tends to decrease with @ (depth) and will change in ‘sign at a given value of ¢. At @ = 0, we have or Np =witeow 6 -[ 42) y= No = 28 (with W = 0) Nature of Vatiation of Ny and No All designers should have a clear idea of the following aspects. ‘The variation of N, and Np with uniform loading along the surface of shell is given in Figure 4.3. Figure 4.3 Distribution of dead and lve loads, ‘The following five important features can be deduced from the formula for Ng and Ng: 1. Both Ny and Ne due to self load only are = wR/2 at @ = 0. 2. The value of Ng slowly increases with the value of ¢. 3. The value of Ng decreases with @ and without any apex load W and it becomes zero when Se T+cos é Hence, beyond approximately 52° central half angle, My is tension when only self Jond is acting as shown in Figure 4.4. We can also imagine how, by keeping the dome light in weight in the apex or higher portions, we can reduce the thrust and also lower the height of the tension angle. This principle was one of the methods of construction of old circular masonry domes (see Appendix A). 4, When # = 90°, the direction of N, is normal to the support and hence there is no horizontal tension. 5. When @ < 90°, there will be vertical and horizontal components for Ny as shown. in Figure 4.4. {It is interesting to compare the above stress distribution with that of a conical roof “supported on its edges as described under conical domes) cos or @=51°48"(ou) wy, i NM @) ©), Figure 4.4 Citibution of membrane fores in aspherical sell due to dead and fe loads: (a) Ny and My due to dead load, end (0) %, and Ay due to lve toad. 4.2.4 Analysis for Live Load Live load is taken on the plan area as shown in Figure 4.3. Usually, as live loads are ‘small, an equivalent dead load can be taken in lieu of live load. It is, however, interesting to study the effect of live load. We can derive the following expression for Ny and No for live loads: Ne we (constant compression) (43) te Boos 26 (compression changes to tension at $= 45°) a4) This is also indicated in Figure 4.4 42.5 Analysis of Domes with Skylight Many domes are not closed at the vertex but are provided with a circulan skylight lantesn, ventilation device, etc, as shown in Figure 45. Let us assume the following: We take the load exerted by skylight = P per unit length ‘The half angle subtended by skylight = ¢ Radius of shell = R ‘Weight of dome above @° = w x 2aR x R(J— cos @) = 2aR? (1 — cos @) ‘At angle @ from Equation (4.1) and Figure 4.6, we get Nj = Due to [DL on full area ~ DL on skylight area] + Due to P ‘As we derived Equation (4.1), considering vertical loads, Nq2mk sin? ) = w(2aRV[(1 - cos 4) ~ (1 — cos a] + Px (2a sin oy) p= DRcooty cons) , Pein as) sin? sin’ 6 and Np = wR cos $~ Ny [as in Equation (4.2a)] 46) Note: When @ = 4 we get Ny = P?/sin # which is true by simple staties as shown in Figure 45. "The maximum value of Ny due to P occurs at upper edge-]besign of Spherical Domes and Conical Roofs 3 igure 4.5. Analysis of domes with slight 42.6 Design of Ring Beams (Edge Member) [As we have seen in Section 4.2.4, if Ny is not vertical (as in a segmental dome), we have to provide for the horizontal and vertical components of Ny as shown in Figure 4.6. Ring beams are provicled at the base in all cases and also at the skylight level for domes with skylight. @ o Figure 4.6 Design of ting beams af @ dome: (a) Bottom ring beam, and (b) Top ring beam of skylight Design of Ring Beam at Base From Figure 4.6, we have component of N, of shell) T as follows (symbol { is used as thickness F=Ny cos ¢ an where @ is the half base angle. This tension acts radially along the full base circle. ‘Total hoop tension on ring beam = Nyman) % C08 @ x R sin g (47a) where R sin @ = B/2 = One half span of shell [We may also derive T from V equal to the fofal load on dome per m length of circunsference gf ring beam as in Example 4.1.) V for a shell with only self weight t/m? Total load __ 2aRhw _(R = 48) Qaxhalfspan 2ar (Fw id F=Vcos¢ (48a) where Vis the total load of the shell per meter length of circumference near ring beam. Total tension in ring beam T = (F x span/2) (4.8b)32 Design of Reinforced Concrete Shells and Folded Plates [As it is a tensile member, we compute its size as follows: 1. First of all, find the area of steel 4,, for tension assuming allowable tension, As = T/fe (4.9) 2, The area of concrete should be such that the stress in concrete should not be high, a ——1 __-g f, (Allowable tension in concrete, say, fx/10 awa & Ny, fa! 10)) A=2-(m-DAy (410) 3. All dimensions of the ring beam should be not less than twice the thickness of the shell, Design of Ring Beam of Skylight Ring beam of skylight distributes the load from the skylight to the shell. Hence, it will be in ring compression. Taking, as vertical load per metre length, P sing Value of Ny and the horizontal thrust H = P cot Hence, compression in top ring beam = (Radius of ring of skylight)(P cot $s) (4.11) 4.2.7 Design for Shear between Bottom Ring Beam and Dome The vertical component of the thrust between the shell and ring beams at the bottom produces shear between the junction of the shell and the ring beam YY as shown in Figure 4.7. This shear at the junction should be checked and the extra tension steel provided should enable the concrete to carry the shear without vertical stirrups. This shear will also be the total weight of dome transferred at the junction between the ring beam and the dome. af — @ ) © Figure 4.7. Detaling of stel in spherical dame with sky Hight: (a) Plan, (b) Bottom ring beam, and (c) Top ring bear,Sune Design of Spherteal Domes and Conical Roofs 33 4.2.8 Detailing of Steel 1. Main steel on shell: Generally, the meridinal steel in the shell is placed at the bottom of the surface and the circumferential steel (hoop steel) on the top with equal cover on top and bottom. (The reverse is also allowable.) In most cases, the stresses will be compression and theoretically no steel will be needed, The minimum steel for Water tank domes is usually specified as 0.3% for mild steel and 0.2-0.24% for HYSD bars both ways. As the meridinal steel will crowd towards the crown, alternate bare are discontinued at one half distances. Thus, start with the number of bars as multiples of four at the outer edge, cut off alternate bars at 1/2 the surface length from the centre (and provide development length), and again cut off alternate bars at 1/4 the same surface length. It will be difficult to continue all steel through the centre, co Provide a rectangular grid at the centre (as in the case of conical shell shown in Figure Example 4.5). 2 Detailing the ring beam at the bottom: As this ring beam is in tension, provide the Tecessary steel ancl concrete size of member. Usually, as bending will be present at the junction because of the edge member, we increase the thickness of the steel at the junction. As additional steel, we provide extra steel (the same amount as meridinal steel) at the top and bottom of the shell (to take care of bending) and extend it into the shell for a distance equal to at least 10 times the thickness of the shell as shone in Figure 47. 3. Detailing of steel in top ring beam: This ring beam is in compression, The size of this ting beam should be not less than twice the thickness of the shell and detailed as shown in Figure 4.7. Steel equal to at least 0.6% of the area is provided as minimum steel, 43 CONICAL SHELLS panycal shells are used as roofs, bottom part of intze type of water tanks, bottom part of bunkers, etc. As roofs, they can be used in the following two ways (Figure 48): + Conical dome roof supported at the base with a ring beam at the base * Conical umbrella roof supported at the apex Figure 4.8 Analysis of conical domes34 Design of Reinforced Concrete Shells and Folded Plates As explained in Appendix A, the heavy cross of St, Paul's Cathedral in London is supported by a conical roof inside the circular dome. In this chapter, we examine the forces that act on these two types of shell roofs. 43.1 Conical Dome Roof with Ring Beams Conical dome roofs are the roofs supported at the base as shown in Figure 4.8. It is always provided with a ring beam. The main parts of the shell are the shell proper and the ring beam. The membrane forces acting on the shell are the meridinal N, forces and the hoops Np forces. Both are compressive and zero at the vertex and increase linearly with the depth as shown Figure 4.8. We take the origin at the vertex and measure downwards. Let the central axis be Z axis and the depth of the point from vertex defined as z. Let half the central angle be a. We generally keep the angle of the core to the horizontal to be such that we can concrete the shell without difficulty (to less than 35°). Consider a point at depth z. Let the superload fon shell be w/m?, 1. Value of Ny: Load on surface = tw x (art) = watz tan a(z sec @) = waz? tan asec a ‘Ng acts along the slope equating its vertical component to vertical load on shell. Ny (22 tan 0 cos a) = waz? tan a sec o we sec’ a for any value of z from vertex (4.12) 2. Value of Ne 2Ny = (Normal pressure) x 2 radius of action Component of weight normal to cone = w sin & Ng = (w sin @ x (Radius normal to shell = r;) Np = w sin x z tan @ sec a = w2 tan?a al any value of 2 (4.13) 3. Tension due to horizontal component of compression at the base of depth I T = (Horizontal component) x (Radius) wh? tan2ar sec a 2 The distribution of forces is shown in Figure 4.8. The stresses are compressive, This tension component at the base has to be taken up by using a ring beam as in the case of a dome. T = (Np at h) sin ax h tan a= (4.14) 4.3.2 Umbrella Roof Conical shell can also be used as an umbrella roof supported at its vertex as shown in Figure 4.9. We take the origin at the vertex. Ata point z below the vertex, the value of tension N, will be due to the drag of the weight of the shell below the depth 2, i.e. (i — 2). Where fh is the height of the coneDesign of Spherical Domes and Conical Roofs 35 2arN, cos = wah? — 24) tan a sec _wO? = 24sec? or Meridinal tension = Ny = 5 (4.15) The value of N, in tension is zero at z = h (at the lower edge of the umbrella roof) and increases non-linearly and becomes infinite at = = 0 at the crown, Similarly, N (@ sin op(z tan o sec @% = we tan*e (hoop compression) (4.16) 0 and increases linearly to a maximum value at 2 = h. The distribution of the stresses Ny and No will be as shown in Figure 4.9, (Normal component) x (Radius 71) Ne It is zero at 2 CFension) Figure 4.9 Analysis of conical umbrella rot E “SUMMARY The design of circular domes, conical domes and umbrella roof shells are explained in this chapter. All these shells are subjected to meridinal and hoop forces. The distribution of forces in circular domes is shown in Figure 4.4 and that of conical shells in Figures 4.8 and 4.9, Segmented circular domes and conical domes must have ring beams to take care of tension. In the umbrella roof, the value of hoop stress is compressive but the meridinal stresses are tension. It is better to provide a smalll ring beam at the bottom of umbrella roofs for appearance and also better performance at ultimate load (overload). As the tension due to My in domes with skylight at the vertex is infinite, the steel in that part should be properly detailed. Formulae for membrane forces in the various types of shells are summarized in Table 4.1.sign of Reinforced Concrete Shells and Folded Plates TABLE A Formulae for Foovs in Spel an Conical Shel 1, Spherical Dome Dead load = w/m?; Central load is indicated as -ve) (a) Full dome: Skylight canopy load = P/m (Tension force y= + [Fitter ao = REESE a aa) Trcos * 2nRsin™9 No = wR cos 9 ~Ne (42) () Dome ait skylight: ny, ~ Sheen a ek (43) sin? sin? N= wR cos Ny 46) v ae 4s) (ovhere ¥ is total load per metre of circumference near ring beam) T=Tx (=) (48a) 2, Conical Dome (412) 4.13) Ring beam, T = wi? tana sec (tension) (414) 3, Umbrella Roof wh? =2)se& engt Ng = Mle EEE (ension) (413) No = we tan’a (compression) 16) EXAMPLE 4.1. [Design of a simple spherical dome (segmental dome)] aedine for ¢ water tank is 12.5.m in span, Design the dome and ring beam. (Use elastie design with Fe 250 steel, Allowable stecl stress = 140 N/mm) [5] [Note: R Radius of dome and span is sometimes referred as 5 = 2r-] Figure E4.1 Detaling of reinforcements in the spherical dome,Design of Spherical Domes and Conical Roofs 37 Reference | Step Caleulations Eq. (4.1) Eq. (4.1) Find radius of dome = R. Let span = S Adopt a rise of 1/6 span, say, 2 m. 1 = Span/2 = 6.25 m; rise h = 2 m; # = 0.15 m (150 mm) (S27 +1? _ (6257 +27 2h 2x2 Radius of shell R = = 10.75 m 625 1075. [Note: As semicentral angle is 35.8%, all forces are compression] Find load = w Dead weight of shell = 0.15 x 1x 24 KN = 3.6 KN/m? ‘Assume equivalent wind and other live loads @15 kN/m? = 1.5 KN/m? (As = 6 is nearly 40°) Water proofing @ 0.9 KN/m? = 0.9 kN/m? Characteristic load = Total load = 6.0 KN/m? Find stresses in the shell (elastic design) Ng and No wR Treos 6 Total ¢ = ant( ) = 35°48'(35.8°) < 51°52 wR _ 610.75 Treosp 2 ‘Table E4.1 Calculation of Ny and No [w = 6 KN/m?; R = 10.75 m; Use Eqs. (4.1) and (4.2)] ioe 3 a 5 6 wR w L¥cos@ 2Rsin?d = 92.25 No=Col(2+3) wR cos Ny=Col(@-4) 03225 0 525 65 32.25 + Enter values for other values of @ = ae eee a ee oe Maximum compression due to Ng = 35600/150 x 1000 = 0.24 N/mm? ‘These stresses are low, and only nominal steel is necessary, We provide 0.12-0.3% steel both ways using high yield deformed bars. Assume 0.2% Ay = 2 x 150 x 1000 = 300 mm2/m Provide 8 mm @ 15 cm (335 mm’) as hoop and meridinal reinforcement.Reference Step Design of Keinforced Concrete Shells and Folded Plates Calculations Chapter 19 Eq. 48 Eq. 48a Check compression for buckling Oy assuming E = 25 x 10? N/mm? Et RV3 _25%10* «150 10750 x V3 With safety factor 4 = 50 N/mm. Hence very safe against buckling (see Chapter 19) 201 N/mm? Calculate tension T and design ring beam Method 1: From N, of Table E.4.1 (Step 2) Hoop tension in ring beam = (Ny cos @)(1/2 span) T= Ny cos = 35.6 x cos 35.8 = 28.8 kN T = (F x span/2) = 28.8 x 6.25 = 180.5 kN Method 2: From total load on dome [Eq. (4.8)] R 10.75% 2%6 V per m= (Ahan = O22 2%6 20.64 kN/m f = Vicot ¢ = 288 N/m T = (F x span/2) = 28.8 x 6.25 = 180.5 kN Design ring beam by elastic theory 180500 = 1290 mm? Mi Provide 8 nos. 16 mm (1608 mm?) (b) Area of concrete (Assume Ghension = 1.1 N/mm? = ,,) i 179000 A.-+(m—TA, A, + (17 1608) Assume m = 18 for A,. Ag = 127000 mm? (Approx.) Use 500 x 300 mm ring beam. (a) Area of steel = ou Design for shear between bottom ring beam and dome Shear = Weight of shell = (2aRh) x w/2z x (half span) 10.75x2%6 = 20.64 KN/m = 20640 N/m. oO 20640 _ 147 mm? Steel required = eee 140 m Provide § mm @ 30 cm (gives 167 mm?)Design of Spherical Domes and Conical Roofs 39 IXAMPLE 4.2 [Design of spherical segmental dome with a canopy, (see also Example 4.3)] Design a reinforced concrete dome resting on a circular 340 mm thick brick water tank, 11m in diameter. Assume the dome is 100 mm thick. At the top of the dome, a circulay canopy 1.75 m in diameter is provided for ventilation and its total weight is estimated as BUEN: Design the reinforcements for the main dome and also the top and base ring beams. (Use Fe 415 steel [5].] \h Figure E4.2- Detailing of dome with shy loht, Reference | Step Calculations 1. | Adopt usual size for the dome Adopt rise 1/5 span for the dome Span $ = 11.0 + 0.34 = 11.34 m; $/2 =r = 5.67 m h = 11.34/5 = 2.27 m; let us also provide a rise h of 2 m. We can express radius also as, Bx 41 S$ _1134 i half central angle; 9, = = = ES = 0.63, Hence gy = 39 Find $o of opening where shell starts. Half angle = ¢y a d 175 sin = 3e= 3-5 ig = RL = cos 6) = 0.043 m Adopt thickness of shell = 100 mm Buckling compression for f = 25 and E = 2.5 x 10° N/mm? 0.097, @ = 5.58° Et _ 25x10*x100 fetta = ee = 160 N/mm? Allowable comp. for M25 = 6 N/mm? (- ;: Je)— 40 Design of Reinforced Concrete Shells and Folded Plates Reference | Step Caleulations 2. | Estimate loads Dead weight = 0.1 x 25 = 2.5 KN/m? Sec. 2.7 Assume live load to be equivalent to dead load, = 1.5 kN/m? Total UDL on shell = 4.0 kN/m? Mo 20 Load P due to ventilator = p = 20 ond P due to ventilat aa 3. | Find Ng and No for different @ values Estimation of Ny and Ny (o = 4 KN/m; P = 3.64 KN/m; R = 9 m; tp = 558° and @ = 399) Eq. (43) i 2 3 4 5 = 3.64 KN/m Q) Eq. (44) ° se a) Saree Ny (243) wR cos @ Ng (6-4) 5.58 0 37.5 37.5 35.8 “17 Enter values for other values of 15, 15.6 53 10.9 349 14.0" 39 19.8 09 20.7 28.0 73 “Max. value of No will be = 15.5 kN/m | 4 | Design the shell for maximum meridinal and loop stresses (a) Max meridinal compression = 37.5 kN/m 37.5% 1000 Max. st Foo qoy = 237 N/mm? (less than allowable Me shes 7mm? (less lowable) Provide the minimum nominal steel of 0.12% to 0.3% of concrete area both ways _ 0.12 «1000 x 100 s 100 8 mm at 270 gives 186 mm?/m (nominal) (b) Maximum hoop tension = 15.5 KN/m 15.5% 1000 Allowable stress 120 mm? /m \ = per metre length | 40 N/mm’) 7 Gor 415, use allowable stress, 2817 <415 _ 155 x 100 2 =p mm? im Minimum steel 120 mm?/m (see above) Provide 8 mm @ 270 mm (nominal)Design of Spherical Domes and Conical Roofs aL Reference | Step Catcatations 5. | Design at base ring beam [We can use Eq, (4.7) as there is a canopy] (a) Design steel Eg. 47) Tension T = Ng c08 9x r = 201.7 x cos 39 x 5.67 = 91 kN Steel area = 21%1000 _ 379 mm? 240 Provide 4 nos. 12 mm bars gives 452. mm? area of steel ‘Add 8 mm stitrup at spacing 200 mm (d) (b) Section of beam for limiting concrete stress Assume mt = 9 for cracking, and cracking stress = 43. — 2.N/mm? Let the area of concrete required be Ay [Ac + (m1 = YA] x 2 = 90.75 x 1000(N)) A, = 41,759 mm?. Let breadth = 340 mun Provide 340 x 200 mm ring beam. 6. | Design top compression ring for canopy ‘The ring beam distributes the load as a line load and also takes care of the bending stresses due to discontinuity. r = 1.75/2 = 0.875 m Bg, (4.9) Compression H = P cot @ x (radius) = 2©4*0875 _ 39.6 un tan 5.58 [As the shell is 100 mm thick, we increase the depth to 200 mm and adopt 200 x 200 mm thickening. 32.6%1000 2 Stress = SEE%I00 = 08 N/muw Provide about 0.4% steel for compression 200 x 300 x 0.4 5 A, = 2002300204 — 240 me Provide 4 rods of 8 mm (201 mm2) with 6 nn ties at 200 mm spacing as reinforcement. IXAMPLE 4.3 (Design of a spherical dome with skylight) Design a spherical dome with the following data [4] (Figure 4.7): Radius of dome R = 50 m Thickness of dome = 60 mm Lower ring half angle = $x = 36° Lower ring radius r Upper ring half angle = ¢o = 50 sin 36 = 29.39 m42 Design of Reinforced Concrete Shells and Folded Plates Ring load of skylight = P = 8 kN/m Assume self wt. plus other loads Rise of the shell 2 KN/m? Span of dome = 27; = 58.78 m; 1/2 span =r, = 29.39 m Upper ring radius = 7, = 6.225 m Reference | Step Calculations 1. | Determine Ng and No 6 = 2 KN/my; R = 50 mi dy = 6%; Oy = 36°; P = 8 KN/m r 5 3 a 3 os Hele eg Sen toe Osi 5 eee wR0sg) Pin dy MAP as tmaint ps CHS), eReGY Ny GDa) Eq. 43) [Pe Rae weet Pesh5 kovdals Ee Eq. (44) sor ba 33 547866 319 36° 537 pase | go 24.8 2 | Design of set Eq. (43) Al forces are compression; Max value = 76.5 kN Max. comp. stress = —75500_ 1.5 ni /mnen? P 1000%60 This is very long, Allowable (4) Et__ 25x10 x60 5 eck for buckli nay = 17.3 N/mm Check for buckling = ip = SEI AO N/m Assume FS = 4, Hence a, (safe) = 43 N/mm? Provide 0.3% steel 3. | tension on bottom ring beam T = Nye 008 ge X R= 56.1 X c08 36 % 29.39 = 1393 kN Eq. (47) [Design steel and section 1333 x 103/230 = 5795 mm?. Use 20 nos. 20 mm = 6264 mm® and 500 % 900 ring beam] Check for stresses 4. | Compression on top ving beam Eq. (49) H =P cot ox r= S*8225 — 357 4 Comp. allowed in concrete = 7 N/mm? 3971000 Area needed = “SS = 56700 man? Provide 200 x 300 mm ring beamDesign of Spherical Domes and Conical Roofs 43 EXAMPLE 44 (Design of conical umbrella roof) A conical umbrella roof is supported at the centre by a circular column and has the following dimensions [4 “Height of conical roof =1 m. Total column height = 1 +3=4m (Assume height below = 3m) "Slope angle = 78°.41' with vertical (< 30 with horizontal) Shell thickness = 60 mm. Column diameter = 400 mm Radius of shell from the centre of column Assume total load = 210 kg/m? = 0.21 t/m? “(Let z = depth below the apex. Total z = 1 m) >} Figure E4.4 Detailing of conical umbrella root Caleulations 1. | Divide depth into 6 dimensions 0 to 6 from top as shown Results of calculation are as follows: Point aim) Nat/m) 0 0 1 O41 -27.07 2 02 13.12 1.05 3 04 5.74 2.10 4 06 -291 3.15 5 08 -1.23 4.20 6 10 -0.0 5.25 (Tension) (Compression) 0.0385; tan a = 5.0; tan?a= 25Design of Reinforced Cx Calcilations a4 Reference | Step zr a 4. Calculate Ny (tension of meridinal forces) i? = 22sec? 22 Point 6, Ny = 0 0.21% (1 -0.36)26 Pa NDS 91 t/m. Calculate Ny (compression, hoop forces) No = wa tanta = 0.21 x 25 x 2 = 5.252 As the hoop forces are compressive, only minimum steel is required. Meridian forces are tensile. As it is infinite at zero radius, we flatter it at the column head. The structural layout is shown in Fig. Design of column Height of column = 4m Adopt diameter not less than 500 mm Design the column EXAMPLE 4.5 (Design of conical dome supported on ring beam) A conical dome roof is 10 m in diameter its height is 3m and its semi vertical angle = 5% ‘Assume the slab thickness 100 mm, Analyze the shell [5) Figure £4.5 Detailing of conical dome supported as ring beam, Reference | Step 1 L Calculations ‘Assume loading Loading DL = 0.1 x 1 x 2400 = 240 kg/m? Insulation = 25 kg/m? Live load = 40 305 say, 310 kg/m?Design of Spherical Domes and Conical Roofs 45 Reereice Step Caleulations z. Find expressions for meridian and hoop forces 1. Meridian force = puma 310 (sec? 59) Ny = OREO) 2 = (684.5) kg/m 2. Hoop forces = wz tanto No = 310 x (tan? 59) x z = (858.6)z kg/m Tabulate the values at z= 0, 1, 2, 3m z My (kg/m) Mo (kg/m) 0 0 0 1 585 859 2 1169 a7 3 1754 2576 (Tension) (Compression) Check compression Popes, 2 2 Max, stress = ToT ia = 0.02 ke/mm* = 0.2 N/mm This is less than 2N/mm? allowed for M20 concrete. Provide normal 0.12% Fe 415 slab Tension in ring beam wh? tan? @ sec a Z ‘The value of T with h = 3m 310 x9 x (tan 59)° sec 59 2 Design ring beam for this force Design as given in Chapter 4 on spherical domes. ‘Tension in concrete should also not exceed fu./10 = 181 kN = 7500 kg46 Design of Reinforced Concrete Shells and Folded rates os ee 1 What are the advantages and disadvantages of a (a) semicircular dome and (b) segmental dome 2, What is the difference in the assumptions of distribution of dead load and live load i domes and cylindrical shells? Indicate how wind loads are not of much importance in the design of the main surface of a dome. 3: Sketch the variation of meridinal stress and hoop stress in a dome due to (a) dead load and (b) live load. 4. Indicate the nature of forces in the following parts of a dome: and (b) top ring beam for a canopy for the top opening. 5. Indicate the types of shells made from a conical shell. © Explain the difference between the forces acting on (a) conical roof and (b) conical umbrella roof. 7. What is the need to provide a ring beam at the bottom of a conical umbrella roof? nae 11] Billington, D.P., Thin Shel! Conrete Structures, McGraw Hill, New York, 1965. [2] Design of Circular Domes, Information pamphlet, Portland Cement Association, Chicago, Illinois. [3] Wilhelm Flugge, Stresses in Shells, Stanford University, 1989, [4] Fisher, L., Theory and Practice of Shell Structures, Wilhelm, Ernst and Sohn, Berlin [SLGambhir, ML, Design of Reinforced Concrete Structures, Prentice Hall of India, New Delhi, 2008. {) bottom ring beamANALYSIS OF CIRCULAR CYLINDRICAL SHELLS. $1 INTRODUCTION Cylindrical shells are produced by a straight line generator moving over a given curve (directrix) at its two ends. Accordingly, we can have a circular eylindrieal shell (with a sirele as directrix), hyperbolic cylindrical shell (with a hyperbola as directrix), elliptic Glindrical shell (with an ellipse as directrix), and catenary cylindrical shell (with a catenary as directrix). In this chapter, we deal only with circular cylindrical shells. Circular cylindrical shells are generally classified as long and short shells. The geometrical characteristics of these shells are shown in Figure 5.1. Long shells were very popular for roofs of factories and short shells for aircraft hangars, etc, As the formwork for these shells is rather expensive for factory roofs, long shells are now being replaced by folded slabs. Figure 6.1 Classification of circular cylindrical shell: (a) Long shell with vertical edge beam, and (b) Short shell with horizontal edge beam; (1) Shell, (2) Edge bear, and (3) Transverse stifners-beams or frames for long shells and arches for short shells, Membrane analysis of circular cylindrical shells is very easy. For bending analysis of these shells, various investigators have presented different methods and equations. Closed form solutions are very useful for deriving computer based solutions. In fact, a748 Design of Reinforced Concrete $1 and Folded Plates there are many pieces of software available now for the analysis of these shells by computers We have already dealt with the general theory of the differential equation used for the analysis of these shells in Section 3.4, However, in this chapter, our main aim is to make a fundamental study of the problem, to understand intuitively the nature of the forces that act on these shells. The ASCE, Manual 31 on design of Concrete Shell Roofs method advocates a method similar fo that we are familiar with and use for elementary analysis of indeterminate structures in theory of structures [JJ[2]. As it is easy to understand this method, we examine the method given in the ASCE Manual No. 31 in this chapter. It will give us a good idea of the nature of the forces in these shells so that when we use the modern computer software, we can have a better awareness of the output. It will also enable us to design ordinary Jeinforced concrete cylindrical shells not by using software but by using published tables. ‘An analysis by hand computation is tedious and unreliable. For an approximate analysis of long cylindrical shells (where the length is greater than 3 times its radius), the beam gnethod described in Chapter 6 can be used. Section 5.12 of this chapter gives the general recommendations for the layout of these shetls. References [3] and [4] give additional tables (in addition to Manual No. 31) for the design of cylindrical shells. (Regarding signs, we will follow the ASCE Manual No. 31 convention of tension as positive and compression as negative.) 52 CLASSIFICATION OF CYLINDRICAL SHELLS The following symbols describe the dimensions of the shells (Figure 5.2): L = Span (the length of the shell) B = Chord width Radius of shell (ACI manual notation) (In Chapter 4, for domes, radius = R and r = % span) f= Rise above the edges or from top of edge beam Thickness Figure 6.2. Gioular cylindrical Shall dimensions. (Note # 18 6 = Half central angle ‘measured from edge) = Angle of any point measured from the right edge (valley) of shell x = Distance from left support (diaphragm, also called traverse) Method 1: Considering the structural action of these shells as a beam, they are divided into three groups: 1. L/r > 3 (say, m)—Long barrel shells. These act as curved beams. They can be analyzed by the “beam method” described in Chapter 6. 2. L/r < 0.5—Short barrel shells. These act more like arches. 3. Intermediate types—L/r > 0.5 but < 3.Analysis of Circular Cylindrical Shells 49 Method 2: ASCE Manual No. 31 classifies cylindrical shells into the following two classes and separate tables have been derived for each. (We will use this criterion in our study.) 1, Shells with r/L < 0.6 ot L/r > 1.66 (Type I shells). This means that the length of the shell is at least more than approximately 1.7 times the radius of the shell. Tn such cases, we can use ASCE Manual No. 31 for its analysis. 2, Shells with r/L > 0.6 or L < 1.6r (Type TI shells or short shells) ‘This classification is made on the observation that in type 2 shells, the length of the arc is so large that the effect of forces applied on one edge will not act on the other edge- Gn the other hand, in type 1 shells, the effect of forces applied on one edge will be felt ‘on the other edge also. This classification is similar but simpler than the Jakobsen classification given below. ‘Method 3: ‘The classification based on Aas Jakobsen’s parameters for cylindrical shells, as stated in TS 2210-1988, is based on p and k, i.e, ep <7 and k < 0.12, the effect of forces from one side of the shell affects the other side. ‘These are called long shells or Type I shells. If p exceeds 10 and k exceeds 0.15, the effect of forces from one side does not affect the other side. These are short shells or Type TI shells Shells with p between 7 and 10, and k between 0.12 and 0.15 are relatively infrequent. 53 NOTATIONS USED FOR FORCES AND DISPLACEMENTS IN MANUAL NO. 31 Note: According to ASCF Maniial No. 31 sign convention, tension is taken as +ve and compression euve. Bending moment that produces tension in the inner fibres of the shell ts taken as ¢ve] Angle ¢s measured from the springing of tee stll and is the half angle subtended by the shell ‘We will adopt these conventions. (In spherical shells, Chapter 4, 9 was measured from the crown.) 1, Primary system (see Sec 3.2 also). The major forces acting on the cylindrical shell are as shown in Figures 5.2 and 5.3 and taken as follows: Note: Positive values of T, and T, direct tension and negative values denote compression.] Direct force (tension or compression) in the longitudinal direction Direct forces in the ¢ (transverse) direction ‘angential shear force considered +-ve when it produces tension in the direction of increasing value of x and @ Mg =Bending moment in radial face is taken as +ve when it produces tension inside the shell (We neglect all the other four stress resultants stated in Section 3.2.)o) ae © Figure 6.8 Forces considered in analysis and design of circular cynical shells: (a) T,, Ty- S and My, (b) Balancing of To forces in long shells by Hy ané ¥,, and (e) Balancing 7, in shart shel 2. Corrective line load. At the boundary of a cylindrical shell, we have to apply boundary forces to bring back equilibrium. The forces we apply are: V; = Radial line load applied along the longitudinal edge 1H; = Horizontal line load applied along the longitudinal edge ‘M; = Transverse moment applied along the longitudinal edge 5, = Shear load applied as line load T, = Tangential line load applied along the longitudinal edge Ry, = Radial line load 3. Displacements AV = Vertical displacement Horizontal displacement Jacement in the longitudinal direction Longitudinal displacement +ve in the increasing direction of x Tangential displacement ~ +ve in the increasing direction of @ Displacement in the radial direction ~ +ve in the outward direction Note: We can also reduce displacements v and w to AH and AV.) 5.4 ASCE MANUAL NO, 31 METHOD OF ANALYSIS We briefly study the method without going into the derivations of the formulae, w! are available in References [1] and [2]. The ASCE method is similar in principles to the classical method of analysis of indeterminate structures. We first make the structure determinate (membrane analysis) and then make corrections to satisfy the equilibrium of the original structure. In the case of cylindrical shells, we first find out the membrane forces (which are determinate) and then make corrections to satisfy the boundary conditions. In the correction analysis, welar Cylindrical Shotts 51 have to use the bending theory of shells as bending moments are also produced when we apply correction forces on the boundary of the shell. (Note: Additional tables to those in Manual No. 31 are available in References [3] and [4]. Interpolation of tables is not valid.) As already stated in the Introduction, additional tables for designers have been published in References [3] and [4]. 5.5 REPRESENTATION OF LOADS Dead load on the shell acts uniformly along the circular surface. Live load is usually assumed as UDL on the plan area. 1S 875 parts 2 to 4, 1987 (IS Code of Practice on design loads for buildings and structures) states that the minimum live load should be as follows. Taking fas the rise and B as the chord width, the live load is to be, Py = [75 - 345(h/ BY] kg/m? but not less than 40 kg/m? Dead load is taken as uniform along the circumference of the shell and live load as uniform along the horizontal projection of the shell. As membrane analysis is determinate problem any load representation can be used for its analysis. However, for the bending analysis of the shell, as we have to depend on differential equations, it is necessary to represent tie load by its Fourier representation in the longitudinal direction. This will make integrations and differentiation easy. Thus, load is represented as shown in Fig, 54 pain $2 aE Though taking together = 1, 3 and 5 will give accurate results, for all practical purposes 11 = 1 gives good results. Thus, to make matters simple, we will take only the first term of the loading and represent load as shown in Figure 5.4, (Note: we enhance to (4/m x p’) as given below.) 4y’)sin() Poe PCE UI UU) = Co AD yA no Oye Tim ns és Tet re 5.4 Representation of uniformly distributed load by Fourier load for eases in mathematical analysis: (a) UDL, and (b) Fourier loading wrth n= 1, 3 and 5 and resuling toad.sign of Reinforced Concrete Shetis and Folded Plates 5.5.1 Conversion of Live Load to Equivalent Dead Load Reference [3] recommends conversion of live load into equivalent dead load for ease of calculation. Such conversion gives results which are on the safer side, We will follow this recommendation. Unlike dead load which acts on the surface of the shell, the live load prescribed acts on the plan area of the shell as shown in Figure 5.5. In addition, dead load also produces greater stresses than live load. Comparisons made with different r/1. values indicate that the effect of uniform live load can be converted into an equivalent dead load by the following formula giving conservative results [3] ven S24) The value of LL is reduced by this expression.) “This will enable us to work with equivalent dead load and thus simplify our work, ‘As live loads are usually small, we may also simply add them on the dead load. Tn shell surfaces inclined at less than 40° to the horizontal, the wind load effects can also be suction, Hence, in many cases, we neglect the wind load in shell design. Figure 6.5 Conversation of live load to equivalent dead tnd for ease of calculations 5.6 TABLES IN ASCE MANUAL NO. 31 FOR MEMBRANE FORCES AND DISPLACEMENTS 5.6.1 Use of ASCE Manual for Calculation of Membrane Forces Membrane forces are only three in number: direct forces T, and T, (which can be tension or compression) and (shear force). These can be easily derived from the three equations of statics. Similarly, the formulae for membrane displacements can be derived from the theory of elasticity. References [1] and [2] may be consulted for the derivation of the formulae for these forces. However, these forces and displacements for uniform loading and sine loading can be easily calculated from Table 1 of ASCE Manual No. 31, reproduced in Appendix E of this book. Interpolation of values in Table 1 used for member are analysis is allowed. The tables in the Manual for membrane analysis are Table 1A—Membrane forces and displacements for live and dead loads uniformly distributed along the length of the shelt. Table 18—Membrane forces and displacements for live and dead loads with sine loading ( = 1) along the length of the shell. Table 1C—Membrane forces and displacements for live and dead loads with sine loading (n= 3) along the length of the shellAnatysis ir Cylindrical Shells 562. Distribution of Membrane Forces Across Cross-Section and Along the Length of a Small Circular Cylindrical Shell Figure 5.6 gives the nature of the distribution of the maximum membrane forces for dead fad and live load on a circular cylindrical shell. The maximum values for T, and Ty occur atthe centre (x = L/2) and shear at the ends (x = 0 and L). 1. With dead load, we can note the following: (a) With central angle @ = 90°, both T, and T, values will be nil and 5 value is maximum at the springing. (0) If we make the central angle less than 90° (we usually adopt = 30 to 45 degrees only), Ty, Ty and S will have definite values. (0) At the edges, T, can be taken by the shell, but Ty and S have to be balanced. If we provide an edge member, we can take care of the S forces, but T, has to be balanced. If we have no edge member, both Ty and $ have to be balanced. ASCE Manual No. 31, Tables 2 and 3, give us data for balancing these forces. 2. With live load, the values vary as shown in Figure 5.6. es Se ay WA Soe ayeceress pet ©) 5.8 Distibution of membrane forces: (@) Distrinution along cross section due to dead and tive toads, and {b) Distribution along length for sine toading 7. BENDING ANALYSIS OF CYLINDRICAL SHELLS (CORRECTION ANALYSIS) boundary conditions of a cylindrical shell require modification of the membrane due to the presence of boundary members which cannot be in equilibrium. Por ple, as shown in Figure 5.3, the membrane Ty forces got by membrane analysis exist at the boundary of a shell. Hence, we have to apply an equal and opposite force at the boundary to cancel its effect. We apply it as H, and V,, for long shells and for short shells. This will produce bending in the shell, and thus we have to depend ‘the bending theory for its solution. The results of bending theory for application of forces are tabulated in Tables 2A for long shells and 3 for short shells in Manual a1. © We have already seen in Chapter 3 that if we take all the eight unknown forces, we up with an eighth degree partial differential equation. Finsterwalder et al. [1] have ide many simplifications to arrive at a solution to the problem. To simplify the problem,54 Design of Reinforced Concrete $ Folded Plates [ASCE published Manual No. $1 in 1952 giving simplified tables, Though, today, we have computer software to analyze these problems, in this book we will use ‘Manual No. 31 to saretend the fundamental behaviour of cylindrical shells, As already stated, the procedure Sdopted in this manual is the same as that we use in the structural analysis for the solution of indeterminate structures. 5.8 APPLICATION OF CORRECTIVE LINE LOADS We have already seen that if gis less than 90°, we have to balance the edge forces Ty and S, as otherwise force equilibrium will not be maintained, ‘Alow, to make the determinate structure (membrane analysis) satisfy boundary conditions in a cylindrical shell of definite size without edge members, we have to apply equal and opposite line loads to T, and S. The edges cannot sustain Ty and § if there are no edge members. Aer Trave ceen that ASCE Manual 31 divides cylindrical shells into two types of shells— those with (7/1 < 06) and with (r/L> 0.6). Different correction Tables (2A and 2B for long tells and 3A and 3B for short shells) have been devised for these shells, (Table A refers a forces and Table B refers to displacements.) This has been done for the following reason. ‘When r/L is 0.6 or less, the line loads applied on one edge create significant stresses and displacements at the opposite edge. Hence, Tables 2A and 2B have been compiled by sae ulteneous application of line load at both longituclinal edges. The points on the shell siz denoted by 9, the angle from the edge of the shell. It varies from @, to 0. The line Toads are applied as Vz, Hy, $; and Mi, Wihen r/L te more than 0.6, the application of edge force on one side does not significantly affect the points on the opposite side. Hence, Tables 3A and 39 have been saved fod one application of the line load on the edge. The points on the shell are represented by s/ ri, where s is the distance of point from the edge of the shell. It is wep rjated from 3.2 to 0, In this case, the correction loads are applied as tangential load Ty radial load R,, sheat $, and moment M as shown in Figure 5.7, {@) Longshots (Vand #4) {Shon shots igure 5.7 Anolon of cretion ads along tengo he sel according to Tales im ASCE Manual 10,91 Cemponert of Te Tas Vand Afr Tong shel, (b} Ty = 7 fr short sal, (e Line Ina Re, (6) Une load $j, and (e) tne LoadAnalysis of Circular Cylindrical Shetis 55 5.9 DESCRIPTION OF CORRECTION TABLES As already stated, Tables 2A and 2B, corresponding to long shells, and Tables 34 and 3B, corresponding to short shells, have been derived from the bending analysis of the shells. ‘The correction application of T,, will produce T;, § and Mg. The correction application of $ will produce T,, Ty and M,, Tables 2A and 3A give the forces produced and 2B and 38, the displacements produced due to application of corrective loads. Table 2A is for correction analysis for T,, Ty S and My produced by various edge loads for, 100 and r/L = 0A, 0.2, 0.3, 0.4, 0.5 and 0.6 (L = 10r to 1.66r) 200 and r/Z. = 0.1 to 0.6 (as above) For various @ from 30 to 50. Table 2B gives the coefficient for displacements AV, AH and @ for r/t = 100 and 200 and r/L, = 0.1 to 0.6 due to application of Vi, Hy, $ and Mg- ‘Table 3A is to be used for correction analysis for T,, (corresponding to Tj), $i. (corresponding to 5) and radial load R,, for various r#/L?, Similarly, Table 3B gives the displacements, v, w and @, for various edge loads for short shells. 510 DESCRIPTION OF ANALYSIS OF SIMPLY SUPPORTED LONG SHELL WITHOUT EDGE BEAM (1/L < 0.6) Case (I) Long shells r/L < 0.6 or L > 1.67r. Step 1. Find membrane forces: By calculation or by use of Table 1, calculate the membrane fortes T,, Ty § for various points @= gy to @= 0 for x= L/2. It will be maximum at 6= and decreases with value of (6 ~ 4). Step 2. Determine and apply corrective tine loads: Membrane force T. can exist at $= 0. But jmembrane forces T, and § present at ¢=0 from membrane analysis if @ < 90° cannot nist. This has to be cancelled to satisfy edge condition, where T, = 0 and $ = 0. For long shells, apply line loads equal and opposite to T, and $ as Vi = Ty sin Va = Te 008 Se g=s ‘The resulting correction values are obtained by Table 2A. Details of the method of correction are shown in Example 5.2. Case 2) Short shells r/L > 0.6 or L < 1.67r. Step 1. Find membrane forces as in Case (1). Step 2. Determine and apply correction analysis.36 Design of Reinforced Concrete Shells and Folded Plates Membrane forces Ty and § at 9 = 0 have to be cancelled to satisfy the edge condition, Apply: T,, = Ty (tangential edge load) § = 5 (shear edge load) ‘The corrections are applied by means of Table 3A of Manual No. 31. Details of the correction are shown in Example 53. [Note: The table for radial edge load R,, is used when we consider horizontal edge beams in short shells] Step 3 (for both Cases 1 and 2)—Find resultant forces ‘The sum of Steps 1, 2 and 3 gives the final forces, Ty, Ty, and $ per unit length. The value of My is moment per unit length. The design of reinforcement and checking of stresses can be carried out in the usual ways (see Chapter 9). 5.11 ANALYSIS OF VARIOUS TYPES OF SHELLS The different types of commonly used cylindrical shells are the following (as shown in Figure 5.8): Figure 5.8 Types of oylndical shells commoniy used: (a) Single shell without edge beam, (b) Single shel with edge beam, {e) Muftple shel (@) Continuous shells and vertical edge beam, (e) Combined shell formed by Combination of shes and horizntal edge beam, and (I) Northight shelAnalysis of Circular Cylindrical Shells 37 1. Simple shells without edge beams (Figure 5.82) 2. Shells with R.C. edge beams (Figure 5.8b) 3. Shells with prestressed edge beams 4, Multiple shells without edge beams (Figure 5.8) 5. Multiple shells with edge beams 6. Continuous shells (Figure 5.84) We will deal with some of these types of shells separately in other chapters. It will ‘be suificient at this stage to know that, in these special cases, it will be necessary to satisfy displacement conditions in addition to equilibrium of forces. As an example, the special displacement conditions to be satisfied when we deal with shells with edge beams are explained in Chapter 8 (One can omit Sections 5.12 and 5.13 on first reading and study the example worked out on the basis of the above descriptions.) 512 GENERAL PLANNING LAYOUT OF CYLINDRICAL SHELLS Clause 7.2 of 15 2210—1988 “Criteria for Design of R.C. shell structures and folded plates” ‘gives the following recommendations for planning, of cylindrical shells: 1. The span should be less than 30 m unless prestressing is also planned. 2. The width of edge members should generally be limited to three times the thickness of the shell. 3. The radius of the shell should suit the acoustic requirements. 4, Single long cylindrical shells 1 > 38 (length more than three times chord width) shall have a total depth between L/6 for small span and £/12 for larger spans, The rise of shells without edge beam should not be less than L/10. 5, For short shells (chord widths larger than three times the span), the rise of the shell shall not be less than the 1/8 the chord width. The chord width should not be more than six times the span as then it will act as an arch, 6. The semicentral angle should be between 30 and 40 degrees to facilitate easy concreting (without back forms). 7. The minimum cover shall be 13 mm or nominal size of reinforcement, whichever is greater 8. The thickening at the junction of the beam should be of the order of 30% only, as cylindrical shells are singly curved shells and thickening is carried for a distance equal to 0.38 to 0.76Vrt, where r is the radius and f is the thickness of the shell. 9. The edges of the shell adjacent to the transverse are also thickened by increasing the thickness by 25 mm for a distance equal to 1/10th the length of the shell58 Design of Keinforced Conerete 5.13 DIMENSIONING SHELL TO SUIT ANALYSIS BY ASCE MANUAL NO. 31 ‘The following discussion is to show that the ASCE Manual No. 31 tables are restricted to be used for limited shell dimensions only. (Additional tables are available in References [3] and [4],) (This section can be omitted in the first reading). Type I shells (Long barrel shells) Tables are available only for r/# = 100 and 200 and for r/L = 0.4, 0.2, 0.3, 0.4, 0.5 and 0.6. ‘Type II shells (Short shells) Tables are available only for ri/L? to 0.010, 0.015 to 0.030 and 0.040 to 0.100. 0,002 to 0.004, 0.006 (a) Considerations for Type I shells. We should first fix the half central angle ¢ between 30 and 40 so that the necessary curvature for shell action and for easiness of placement of concrete. For example if we fix the angle as 35°, we get the relation as in Table 5.1 between radius r, chord width B, and rise of shell from springing / as follows. ‘TABLE 5.1 Relation between L, 8, rand h for f= 95° 7 Lit Bir 1/8 hr (Available in Table) (Constant) (Constant) o4 10 1447 372 0.181 02 5 1147 430 0.181 03 3.33 1147 2.90 0.181 o4 2.50 1.447 248 0.181 05 2.00 147 474 0.181 0.6" 1.67 a7 1.45" 0.181 ‘Type H Shells Type Il Shells ‘Type II Shells “h = rise of shell from springing. Another consideration is the thickness of the shell. It should not be less than 75 mm. For 75 mm for r/t = 100; r = 100 x 0.075 = 7.5 m For 75 mm for r/t = 200; r = 200 x 0.075 = 15 m For 100 mm for r/ = 100; r= 100 x 0.10 = 10 m For 100 mm for r/t = 200; r = 200 x 0.10 = 20 m We have to choose specified parameters for using ASCE Manual No. 31 table. For example the chord width for gy = 35° is (1.147 x the radius). It is also advisable for Type I shells with edge beam to make the value of the total depth of the shell measured from the crown of the shell to the bottom of the edge beam as L/10. However, as the width of the shell is increased, the depth of the edge beam can be reduced, In the limit for a short shell, the edge beam can be made horizontal, Example for choosing shell dimensions of long shell to suit Manual No. 31 ‘As an example, let us take a 75 mm thick shell with 7/t = 100 and = 35°. Then r= 100% 0.075 = 7.5 m.Analysis of Cirewar Cylinarieal Shetts 59 Let us choose table for r/L = 0.3. Then if you choose r/L. = 0.3, L = 7.5/0.3 = 25m ‘The shell will have L = 25 m; r= 7.5 m, t= 75 mm and @% = 35° The chord width B = 1.147 x 7.5 = 8.6m Value of k = 0.181r = 1.35 m. We choose a shell with total depth 1/10 span (see Section 5.13). ‘Total depth = 25/10 = 2.5 m Depth of edge beam = 2.5 ~ 1.35 = 1.15 m Width of edge beam 3t = 3 x 75 = 225 mm Preliminary dimensions can be as follows: Span 25 m; radius 7.5 m; thickness 75 mm; edge beam 1150 x 225 mm (b) Considerations for Type If shells (Short shells). The parameter for the analysis of short shells is r#/L” and tables are given for fixed values 0.002, 0.003, 0.004, 0.006, 0.008, 0.0010 to 0.10. If dy = 352, these are meant for the analysis of shells with r/L greater than 0.6 or L/B less than 1.45. Let f = 100 mm, then rt = 0.1r Foe 1 017 _ 9.902 the least value we have For we have L= vr Thus, the higher values of r!/1? are for smaller spans and shells of larger radius. [Note: References [3] and [4] give additional tables for design of cylindrical shells by the Manual ‘No. 31 method.) Cylindrical shells are classified into long shells or barrel shells and short shells. ASCE Mantal No. 31 classifies them into Type I and Type IT shells. This chapter explains the “use of ASCE Manual No. 31 for the analysis of these shells. Reference [5] also gives a good account of the use of ASCE Manual. More theoretical methods are discussed in ces [6] and [7]. It should be clearly understood that even though interpolation of values of Table 1 of Manual No. 31 is allowed we are not allowed to interpolate the values given in the \SCE Manual tables 2 and 3 for shells that we may arbitrarily plan. We should always shells in such a way that their dimensions suit the table in Manual No. 31. References and [4] give additional tables which supplement ASCE Manual No. 31.a 12a in breadth and 20 m in length, Assume it load of 25 KN/m? and live load of 0.5 kN /m?, fuocedure: From inspection of Table IA of Appendix E and Fi, for a vault with $= 90°, the velue of Ty and T, for dead load i (The shell is a semicircular vault T, at edge due to DL'= 0. Hence, no correction of Ts the major forces at the edge is Required, we need not use correction tables Tabla 2A 20m; L/r > 1.6, Long shell DL = 2.5 kN/m? = 0.5 KN/m? Membrane analysis with non-Fourier loading Calculate max. value of Ty at crown (G ~ = Oatx= 1/2 "=H Tx due to dead load (DL) = $ (66.5 x25x1 Appendix E coe Table 1A 41.6 KN (comp,) T, due to live load (LL) = 766.6) x 05515 Total compression = 54.1 kN/m (thickness 100 mm) 54.1% 1000 hii Be 2 ax. stress = Seog = 2-54 N/mm 5 kN (comp,) (Thickness of shell = 100 mm) Calculate Ty at (6, — 9) = 90° at x = 1/2 Due to DL = 0 Due to LL = 66.6 * 05 x15) = 125 KN (tension) 12.5 x10? 2 Stress = J25x10" _ 9. 2 ‘a Max. tension = 0.12 N/mm? (which is small) Calculate $ at crown for x = 0 (max. shear) Due to dead load + live load = pt. x (Coott )Analysis of Circular Cylindrical Sholis 61 Reference | Step Calculations Appendix E] 4. | Calculate maximum value of S at (6, — @) = 90° at edge (Figure 5.6) Table IA Due to DL = 2.5 x 20 x 1.0 = 50 KN (max. shear) Due to LL = 0, Max. stress = 0.50 N/mm? [Note: Maximum shear due to LL occurs at (gj — 9) = 45°, but its value is less than the above value.) 5. | Calculate maximum value of T, at crown at x = L/2 = pr x (coett.) Appendix E UNote: This does not depend on x value and is same over the length] Table 1A Due to DL = 2.5 x 6 x 1 = 15 kN/m (comp.) Due to LL = 0.5 x 6 x 1 = 3 KN/m (comp,) Total Ty = 18 KN/m; Stress = 0.18 kN/mm? comp.) 6. | Calculate minimum T. value at (gy — $) = 90° Due to DL; Due to LL = 0 7. | Summary of results (stress in N/mm?) Ee Al crown At edge Stress T, 0.51 (comp.) 0.12 (tension) N/mm? Stress Ty 0.18 (comp.) 0 Stress $ 0 0.50 pe ee ed ee As T, is zero at edge (valley), we need not undertake correction analysis. At the edge we will provide an edge beam to take care of the shear and small tension T, produced by live load only. 8. | Design of shell All the stresses are low. Provide noi both ways in the shell 9. | Design of edge for shear varying from 0 at x = L/2 to 50 KN at x = 0 and L (ends of the shell), Assume tension produced is equal to the shear. Total tension = g x Average shear = 3.8 = 250 KN al steel of 0.2 to 0.3% steel 250 x 1000 240 Provide a beam with the above steel at the ends to take care of this tension. If the beam is supported on walls, we need not provide for bending as in the case of shells with unsupported edge beams, (Provide 500 deep 200 mm wide beams with 6 nos. 16 mm bars giving 1206 mm’. Check for stresses in concrete for tension.) = 1042 mm* Ay0 62 ign of Reinforced Concrete Shelis and Folded Plates BRAMPLE 52 (Analysis of ASCE:Type 1 (Long) shell without edge eam] (& < 907) A simply supported cylindrical shell has the following dimensions. Radius of cizcle(7) = 7.5 m; Length of shell (L) = 15.0 m f= 050.< 06 cype 1 sno, £2 2 106 Half central angle (¢,) 40° (0.6977 radians) HAN L-Membrane Analysis: We first make four separate tables each for Ty, 8, Ty and THe gy, Out a membrane analysis. The operations ate carried pet inte tabular form. Tees 1 to 4 given at the end of this example are used to recoes the results of the caleulations of T,, S, T, and M, respectively, ceettnemzestion analysis: As forces Ty and S obtained by membrane analysis cannot Fie tar aBes. We have to make correction analysis using Table 2A, We apply correction line loads at the edges. That for T, will be, Vi = Ty sin d and Hy, = Ty cos @ Data for analysis is entered in the Tables 1 to 4 pages 65, 66 as we proceed with the calculations. For example, Table 1 of this example is tabulation of values for Ty forces which is Of Manual No. Seo ane first line represents the T, membrane force Chlodicesd by use of Manual No. 31 for various angles 9 (corresponding to Table 18 of kannst No. 31). jee accond line of Table 1 is correction due to V, (vertical correcting component of 7p) Here the values will be a constant x coefficient taken from Table oe Manual No. 31 The fourth line is the correction for shear § which has to be compensated. Here also, membrane archer ant % coefficient from Table 2A of Manual No. 31. The suse eh ne ‘membrane analysis and the corrections give us the final values for design. ‘itnilatly, we will have Table 2 of this example for shear(8) valves ed Table 3 for Ty values. qhere is no moment from membrane analysis, But as the correction analysis induces moment Me Te tae ae 8 separate table for My Table 4 of this example ie fey transverse prod and Tn this table, there Will be no membrane forces but we tabulate mee values produced by T,, Ty and S.Analysis of Circular Cylindrical Shetts 63 [Notes: for cortNe Ot (Appendix E) for membrane analysis and use Table ee of ee ee 31 for correction analysis 2 Are rll use g/cm? instead of N/mm for convenience in manual calculation (owe will be dealing with larger numbers with kg/cms & We Rept tet a8 ve and, contpnestion ye im Manual Ne. St] Reference | Step Caleulations 1. | Catcutate rise, chord length, etc. B = Chord length = 2 x r sin 40 = 9,65 m ht = Rise of shell = r(1 ~ cos 40) = 1.755 m stvB Ges 9.65 55 1s = xs os z (see Section 5.12) 2. | Caleutate tonds on shells Dead load = 0.075 x 2000 = 150 kg/m? Sec. 2.7 Add LL plus weather proofing = 15 kg/m? Total = 165 kg/m? Rise/breadth ratio = Rise span ratio = 2 See. 55 Live load = 75 — 345(4) kg/m? and not less than 40 kg/m? 2 = 75 ~ 345 es) ] = 62.3 (say, 65 kg/m?) sin 40) _ 65>sin 40 55. se equivalent = 65x (940) _65xsingo _ 2 Sec. 5.5.1 Use equivalent DL = 65 ( - ) an 65 kg/m’ (@pprox.) Total load = 165 + 65 = 230 kg/m? Fonaane loading increment factor = 4/7 [We use the first term of Fourier loading. | 4 Load = () X 230 = 295 kg/m® (say, 295 kg/m?) 3 Part I Membrane Analysis 3G). | Tabulate membrane force T, at x = L/2 [see Table Ex, 5.20) rf fon 0-2] contiant coal orp| 64 Deslgn of Keinforced Concrete Shelts and Potted Piates Reference | Step Caleutations Coefficient refer to Table IB of Manual No. 31 Constant = 295 x 7.5 4 = 8850 and costo = o(3) for ( - 6) = 0 equals 0.2026 as in Table 1B, Appendix F. Coeff. from Hence, T, at (6 ~ 9) = 8850 x (Coeff. for (g, ~ ” Appendix E The coefficient can be calculated manually from formulae or taken from Table (Appendix E), Table 18 INote: The values are tabulated in Table 1 for T, given below] Coeff. from: Hb). | Tabulate membrane force S at x = 0 [see Table Ex. 5.22) Appendix E S= prt sing, = od] can be read off from Table in Manual No. 31 or calculated. Table 1B S at edge 9 = 0 or (@ ~ 9) = 40* is 4425 x 0.4002 = 1810 [Coetf. in Table 2} Coeff. from 3(c). | Tabulate membrane force Ty at x = 1/2 [see Table Ex. 5,2(3)} Appendix E Ty = pricos(, ~ 9] Table 1B Constant = pr = 295 x 7.5 = 213 Coeff. = cos ( - 9) = 1 for (§ - 4) = 0 This can be read off from Table of Manual No. 31 or calculated manually, 34). | Vatue of My There is no bending in membrane analysis, but this will be produced when we apply correction edge forces. These calculations are made for (¢ ~ ‘) values 0, 10, 30 and 40. (See Tables at the end of this example.) Part II Correction Analysis using Table 2A of Manual No. 31 4. | [Note: Membrane analysis results in edge forces Tgand S. Now we apply G28e covrection to comet the 7, and § force, They shovla te sae ther a no edge members to take thee forces. 7 dose ra Seca correction a the shell edge can take. force (a). | Calculate edge correction loads Value of Ty at = 0 oF (6 ~ §) = 40; Ty = 2213 x cos 40 = 1695 This Ty cannot exist as the edge is discontinuous. Hence, we apply *qual and opposite correction forces as V;, and IH; as for long shells, Vi = Ty sin @, = 1695 x sin 40 = 1090 Hi = Ty cos % = 1695 x cos 40 = 1290Analysis of Circular Cylindrical Shells 65 Reference | Step Catcutations 4(b). | We tabulate T, at x = 1/2 in Table Ex. 5.2(1); 8 in 6.2 for S at x =0 in Table Bx. 5.2(2), Tg al x= L/2 in Table Ex. 5.2(3) and Mg at x= L/2 produced by Vi, Hy and S. INote: (1) Moment producing tension on the inside the shell is taken as ‘tve and tension on the surface as ve. @) As there is no edge beam, T, and § should be zero at the edges of the shell.) Table 24 (1) For T, caused by application of V;, Hy and S. Appendix E T, due to corrective load V; = V; x (L/r)? x Coeff. = 1090 x 4 x Coeff. T, due to corrective load H, = 1290 x 4 x Coeff. T, due to corrective load § = 1810 x 4 x Coeff. (2) For $ caused by application of Vi, Hi and S. S due to corrective load Vi = 1090 x 2 x Coeff. S due to corrective load Hy = 1290 x 2 x Coeff. $ due to corrective load $ = 1810 x 2 x Coeff. (8) For M, (There is no moment in membrane analysis.) due to Vi, Hy, and S. My due to corrective load V, = V, x7 x Coeff. = 1090 7.5 x Coeff. My due to corrective load H, = 1290 x 7.5 Coeff. M, due to corrective load § = 1810 x 7.5 x Coeff. Tabulation is as follows: Table 1 (Example 5.2) Distribution of T, at x = L/2 (in kg/m) for various values (Using Table 2A of Manual No. 31; r/t = 100; r/L = 0.5; & = 40°) pery 0 10 a) 40 No. T. 6 4030. 2010 _0 (Bdge) 1 Membrane load 793-1766 1685-1552 1373 8850 x Coeff, from Table IB Line loads 2 Vz = 1090, 423858 +6121 ~29042 25400 +113880 1090 x 4 x Coeff, 3 Hy = 1290, 1290 x 4x Coeff. 22110-9497 417227 +21668 ~62846 4 S=180x4x Coeff, 848 1019 617 +3054 413813, Total 893 6161 14117 2230 _+63579 INote: Value of T, at edge is not zero.] Example, Calculation for V;. Table 2A of Manual No. 31, for applying Vi we get coelf. for T, @ @ = 40° = + 5.472. Hence, Ty = 1090 x 4x 5.472 = +23858 (as in Table above)66 Reference Step Design of Keinforced Concrete Shells and Folded Plates Calculations (Comments. As a digression let us apply the beam theory to this shell using Table 6.2 page 78. We get; 1.= Ex x Coetts Cot for = 40" for 9=0 at valleys 3.928 sy 75% 295 X 3.928 = 34672 kg/m This value is much different from the theoretical value showing that beam theor However, if we which we will y is strictly applicable to long shells L/r > 3 only. consider total tension below the neutral axis for Provide steel, the result will not be significantly different. Table 2 (Example 5.2) Distribution of Shear S at x = 0 (in kg/m) for various @ values (Using Table 2A of Manual No. 31) o- 6 0 10 20 «30 40 No 5 Se a @ 4030 2010 ° ae es SOE eee cao Sie 1 Membrane 0-489 694-1408 i810 4425 x Coeff, 2 Vi = 1090, 0 +4866 +1650 7637 0 1090 x 2 x Coeff, 3 Hy = 1290, 0 4877-3882 42626 o 1290 4 x Coeff, 4S, = 1810, 0-303 -618 -33341810 1810 x 4 x Coeff. Total 0 =1308_~3774_~6752 0 [Note: Value of shear $ at 9 = 0 (edge) is zero] Table 3 (Example 5.2) Distribution of T, at x = L/2 (in kg/m) for various 9 values (Using Table 2 of Manual No. 31) 10 Wao 40 te 0 E a ee ee ¢ 4030 20 0 ° 1 Membrane “2218-2179 2079 1916 1690 2213 x Coeff, 2 Vi= 1090, -309 1165-2248 1323 +700 1090 x Coe 3 Hy = 1290, 284 +469 41828 +1973 +990 1290 = Coeff. 4S, = 1810, 140-106 +5 aL a 1810 x Coeff. Total 2946 -2981__-2494_ 1125 (Note: Value of T, at @ = 0 (edge) is zer0.]Analysis of Cirewlar Cylindrical Shetis 67 Reference | Step Galeutations Table 4 (Example 5.2) Distribution of M, at x = 1/2 (in mkg/m) for various @ values (Using Table 2A of Manual No. 31) ae 0 10 2 30 40 No. My a 4030. 20 10 0 1 Membrane 0 o o ° 0 2 ¥, = 1090, 1082-1107 1025-14 0 1090 x 7.8 x Coeff. 3 Hy = 1291 4660 473047934574 ° 1290 x 7.5 x Coett, 4 $= 1810, 46 46 tak o 1810 x 7.5 * Coeft. Total 465 3251336 o 1. Variation of T, along length is as sin x/L. 2. Variation of S along length is as cos mx/L. 3. Variation of Ty along length is as sin mx/L. 4. Variation of My along length is as sin mx/L FXAMPLE 5.3. [Analysis of type 2 (short) shell without edge beam] ‘Analyze the forces in the shell of the following dimensions for the live load specified by IS: Radius r=22m ] r/L = 2.09 > 06 (type 2 shell; use Table 3A) Length L= 105 m | t/t? = 0.015; L/r = 0.47 (< 1.66) Thickness #= 75 mm | (riL?)/ = 3,673 Chort width B= 27m | L/r = 105/22 = 0.477 ar sin = Ze 1 Hence, = 37.85" sis a 2 i Assume equivalent total dead load = 230 kg/m, ‘Assume sinusoidal load as 4 x 230 = 295 kg/m? as in Example 5.2. a Note: Inthe case of short shells the fist two terms ofthe load = 4p, 5: Lsin ie. = 1 and should be taken into account for an accurate analysis. In the present ease, we will take only the fist term as the object is only to illustrate the method. Table 3A gives correction values for f= 1 “and 1 = 3 in separate tables. We will use table for n = 1 only.] 220.075 ‘The characteristic of the shell is defined by the expression aoe = 0.015 (Use Table 3A and rt/1? = .015) given on page 322.68 of Reinforced Concrete Shells and Folded Plates Procedure Step I: As in Example 5.2, carry out membrane analysis to find the forces in the determinate structure. Use Table IB of Manual No. 31 given in Appendix E. (Values of T,, Ty and § can also be found using the equations for membrane analysis. Interpolation of values in Table I of Manual No. 31 is allowed) Step 2: Along the sides, the T, and $ values should be zero. Hence, we apply corrective forces T;, and §;, This results in bending of the shell. Using Table 3A, find out the corrections to be made to satisfy the boundary condition. Thus, we will get corrected T,, Ty, $ and My values. Step 3: The addition of the membrane forces from Step 1 and corrections from Step 2 give us the final results The procedure is the same as in Example 5.2, except that we use a different table (Table 3A) of Manual No, 31 for short shells. Reference | Step Calculations 1. | Calculation of (6 — 9) values for given short shell Table 1 for membrane forces gives coefficients for (@, - ¢) values. However, Table 3, applicable to short shells, gives coefficients for s/(RtL)", Hence, we have to find the correlation between these when we are solving problems on short shells. ‘Thus, use Table 1B (b) of Manual No. 31 ot by calculation to find membrane forces T,, Ty and $ in terms of (% ~ 9) in degrees. We use f, L and r to identify type of shells. In long shells, we use r/L and f/L values. In short shells, we have one quantity combining them equal to rt/12. Tables have been compiled in Manta No. 31 for bending analysis for 14 values of rt/L* ranging from 0.002 to 0.100. We will use the values of rt/L? = 0.015 for Fourier loading for = 1 for our analysis (ie. sine loading). In bending analysis of short shells, the points on the shell are represented by their distance from the edge of the shell. To represent tne” membrane analysis by Table I of Manual No. 31, these portions have to be converted to degrees. it in non-dimensional form, we use the expression For 1(a) | Find corresponding @ for various s=0. values. 1 x 3.673 = rp = 226 Hence @ = 0.0167 radians = 0.957 radians ‘This means for every s = 0.1, o = 0.957 degrees. We convert other values also to degrees as follows (see Table 1) page 69. For s=Analysis of Circular Cylindrical Shetts Reference Step Calculations 2(a). TABLE 1 Calculation of 9 = (64) for s values for given short shell Point No, stm) pdegrees 1 ° 0 ° 37.85 2 oa 03673 0.957 36.89 3 02 0.7346 1.9124 35.94 4 os 1.4692 3.8248 34.03 5 08 2.9384 7.6496 30.20 6 16 5.8768 15.2992 22.85 7 32 11.7536 30,5984 7.25 8 3.96 1453933785 - 0.0 We use these (¢, - ¢) values for membrane analysis using Table 1 of Manual No. 31 ‘TABLE 2 (Example 5.3) Interpolated Membrane Coefficients from Table IB (This interpolation is allowed when using Manual No. 31.) sR ao Factors n= 1 (degrees) By s Ty 0 37.85 0.1598 0.3903 0.7888 on 36.89 0.1619 0.3818 0.7991 02 35.94 0.1640 0.3730 0.8091 o4 34.02 0.1678 0.3560 0.8280 08 30.20 0.1750 0.3203 0.8641 16 22.55 0.1870 0.2439 0.9227 32 7.25 0.2009 0.0805 0.9911 3.96 0 0.2026 0 11.0000 (Crown) Part I—Membrane Analysis Tables to be prepared INote: The procedure is the same as in long shells Ex. 5.2. We first make separate tables for T,, §, Ty and My forces as in Ex. 5.2 and fill in the values as the calculation proceeds.) ‘We provide separate tables for maximum edge values of Ty, $, Ty and ‘My as follows: Tables 3, 4, 5 and 6) as shown in pages 71 and 72. Tabulate membrane forces, at springing at $= 0, T, at x = L/2 Ty = pr(L/1¥? [coelf. for (gy - 9) from Table 2 above] = 295 x 22 x (0.477)%x 0.1598 = 235.97 for s = 0 Enter this value in Table 3 shown on page 71 for Ty70 Design of Reinforced Concrete Shelis and Folded Plates Reference Step Calculations Page 69 Table 2 page 69 2(b). 2(€). 2a). INote: ‘, is max. at crown as shown in Fig. 5.6 and decreases to the edges] Tabulate membrane forces, S maximum at x = 0 Se »(2) (Coeff, Table 2 above) S, = 295 x 2x 4) % 0.3903 = 1270 kg/m (membrane force) This should be made zero at edge at $= 0, as there is no member to carry it. As this will not be zero in membrane analysis, the edge value will have to be corrected by S, line load. Tabulate values as in Table 4 (on page 72). Tabulate membrane force Ty at x = L/2 Tp = pr cos — @) = pr x Coeff. from Table 2 above. T, at s = 0 = 295 x 22 x 0.7888 = 5120 kg/m At edge, $ = 0 Ty should be made zero by correction analysis. Tabulate these values in Table 5, page 72. There is no moment My by membrane analysis. My for membrane analysis = 0 correction analysis produces My. Part I1—Correction Analysis Apply correction loads—(Table 3A ~ rt/.2 = 0.015) [Note: As the shell has free edges at the springing Ty and S are the forces, to be compensated, Hence, we apply correction loads T;, and S. This will produce further additional T,, S and T, forces and in addition My in the shell, The results of membrane analysis and corrections are tabulated in Tables 3 to 6 shown on pages 71, 721] We proceed as follows for correction analysis (a) Apply T,, opposite to T, at = 0 Tabulate T,, $, Ty and M, produced by use of coefficients. From Table 3A. (b) Apply $1 opposite to $ at @ Tabulate T,, 5, Ty and M, produced by use of Table 3A coefficients We get the following resulting values (a) For 7, value (Table 3, Example 5.3 given below) for T, forces. 1. T, from membrane analysis 2. T, from T, = T, x Coetf. = 5120 x Coeff. 3. T, from S = T, x Coeff. = 1220 x Coeff.Analysis Circular Cylindrical Shetis 71 Reference Step Calculations (b) For § values (Table 4, Example 5.3) page 72 1. $ from membrane analysis 2. S from T,, = T, x Coeff. = 5120 x Coeff. 3, § from $ = 1210 x Coeff. (©) For T, values (Table 5, Example 5.3) 1. T, from membrane analysis 2. Ty from T, = Ty, x Coeff. = 5120 x Coeff. 3. Ty from § = 1210 x Coeff. (d) For Mg values (Table 6, Example 5.3) 1. My from membrane analysis = 0; (t = 0.075) 2. My from T,, = 5120 x £ x Coeff, = 384 x Coeff. 3. My from S = 1210 x t x Coeff. = 91 x Coeff. (Ref: We have used Table IB for (7 = 1) and Table 3A for (1 = 1) and rt/L? = 0.015. For more accurate results, repeat the operations for 1 = 3 of Fourier load using Table IC and Tables 3A for n = 3 (page 323). TABLE 3 (Example 5.3) Distribution of T, Forces Table Values of T, forces for different points (Values of 0-1 and 0.4 have been omitted.) kaa 0 02 08 «16 32 3.96 1, int e O11 7655S 0697.85 (Membrane —236~—~—AO=SDST BAS -297 forces, 1477 x (Coeff) @) Due to Ty 455514 418830 14004 1283-4560 ~ 5120 x (Coeff) Table 3A @) Due to, 45821 49022-34326 436 = 1210 x (Coeff) Total e100 21612-14694 1743-203 -297 Comments: In this short shell as compared to the long shell of Example 52, the T, forces are low. Hence, an edge beam can be ‘omitted or we can provide a horizontal cantilever for short shells. twill be very useful when short shells are used as storage godowns or as assembly shells.22 Design of Reinforced Concrete Shells and Folded Plates Reference | Step Calculations s a @ @ o @ a a Q eo TABLE 4 (Example 5.3) Distribution of S forces at ends He 0 02 08 16 32 396 @ 0 191° 7.65" 15.3" 30.6" 37.85" Membrane A210 597 0 forces, 3098 x (Coeff.) Due to T,, 0 7828-4874 42580-5890 5120 x (Coeft.) Due to $, +210 4260-31471 Total 08724 6185 +1805 8580 TABLE 5 (Example 5.3) Distribution of 7, forces at 1/2 ie 0 0208 16 32 96 a O igi 765° 153° 30.6 37.85 Membrane 5120-5251 5608-5988 6432-6490 forces, 3098 x (Coefi.) Due to Ty, 45120 +4111 1208-1231 4281 5120 x (Coeft.) Due to Si, 0 150 Hd 8048 1210 x (Coeft.) Total 09905600 7299 6133 64% TABLE 6 (Example 53) Distribution of My at L/2 0 02 08 16 32 336 6 0 191e 765° 153° 306° 37.85" ‘Membrane 0 0 o 0 ar) Due to Ty, 0 +49 130-178-4300 384 x (Coeff) Due to S,, 0 at “ 91 x Coefe Total 0 32 1a The final values of T,, 5, T, and M, at various points of the shell are total values in the above tables.ysis of Circuiar Cylindrical Shetts 1. What is meant by “membrane analysis” and “bending analysis” of shells? 2, Explain why bending analysis of shells is necessary. In what types of shell can we expect the effect of bending analysis to be less? 3. Explain how ASCE Manual No. 31 gives the tables for bending analysis of shells. "4. Are we allowed to interpolate the values of Manual No, 31? What are other publications that can supplement the values given in ASCE Manual No. 31? SEUEFERENGEO eee Il] ASCE Manual No. 31, American Society of Civil Engineers, USA, 1952. (Reference 2 below also gives Tables 2A and 2B of this manual.) [| Billington, D.P., Thin Shell Structures, McGraw-Hill, New York, 1965. {3] Design Constants for Interior Cylindrical Concrete Shells, Advanced Engineering Bulletin, Portland Cement Association, Llinois, USA, 1960. [4] Coefficients for Design of Cylindrical Concrete Shells Roofs, Portland Cement Association, Illinois, USA, 1959. (This is an extension of Tables 2A and 2B of Manual No. 31 for various @, r/ and r/L values). [5] Design of Barrel Shel! Roofs, Concrete Information Series, Portland Cement Association, Mlinois, USA, 1954. 16) Ramaswamy, GS., Design and Construction of Shell Roofs, McGraw Hill, New York, 1968. [7] Chandrashekara, Analysis of Thin Concrete Shelts, Tata McGraw Hitt, New Delhi, 1986.15.1 INTRODUCTION Reinforced concrete folded plate roofs (as shown in Figure 15.1) are composed of a seri of rectangular reinforced concrete slabs (whose length is more than three times its width) placed inclined to one another and joined monolithically, one after the other, along theit longitudinal edges. The unit as a whole is supported rigidly at its ends, by transverse diaphragms, as in the case of cylindrical shells. These structures are also called prismatic structures or hipped plates. These can be made into various shapes as shown in Figure 15.1 SAA I] @) © UNINNKAINANANI © @ Figure 18.1. Folded plate root construction: (a) V type, (b) Tiangh type, (c) 2 © northlight type, (d) Cylindrical type (prismatic), and (e) Non-prismatic folded plate (umbrella roo), 202ded Plates—Preliminary Analysis 203 Uke simplest folded plate is the V-shaped unit; but it may not provide enough area of te at the top and bottom to resist the compressive forces due to bending and also Paccommodate the needed reinforcement. Hence, the trough shaped or the trapezoidal units are more popular than the V-shaped units. Asymmetrical sections of Z-shapes can erve as northlight roof for factories [1]. ‘The important reasons for the popularity of folded plates are: (i) They can span ree distances like cylindrical long shells, (ii) They are cheaper than curved cylindrical fs because of their easiness in formwork (iii) They are more aesthetic in appearance, ind (iv) The thickness of the slabs is small as in the case of shells (of the order of 1/200 the span). Folded plates are studied along with shells under the heading “stress skin. ‘structures’. ihe three main parts of a folded plate are: (i) inclined plate, (ji) rigid joints between plates, and (iii) transverse support or diaphragms at the ends. It is important to note cre, joints are to be made rigid and should not undergo any change in their angles inder the action of loads. It is due to this fact (that there should be no change in their Bint angles) that we need correction analysis described in Chapter 16 to the preliminary malysis described in this chapter. : These folded plates cannot be classified as pure shell structures which take their th from their curvatures. In folded plates, in addition to the transverse bending as {tframe, longitudinal bending as a beam produces stresses similar to membrane stresses n shells. The reinforcements for these stresses are placed at mid-depth of the plates in he longitudinal direction. In this chapter, we examine the historical development of folded plates and their preliminary analysis. In Chapter 16, we examine the corrections to be made in the iyainary analysis for angle changes between plates that will be present in the preliminary Shown in Figure 1.5 (Chapter 1), where the bottom plates were built without any ms at the joints by just folding them. But it was only after World War II that American 'gineers applied the principles of folded plates for the construction of roofs. In coal inkers, we deal with plates which are similar to deep beams. However, the American ly Was restricted to the study of plates which are nat deep. In 1947, Winter and Pei published their paper on “stress distribution procedure” to folded plates in which the pan of the plate is more than three times its breadth so that we can apply the ordinary theory of bending to these plates. Their approach was to make only stresses in the pies at the junction where the plates meet compatible. This we study in this chapter as liminary analysis [2]. The above analysis proposed by Winter and Pei ignored the compatibility of formation or rotation at the joints (change in angles at the junction of plates) and the mrrections to be made, to keep the geometry of the joints. A number of methods for the204 Design of Reinforced Concrete Shells and Folded Plates full analysis of folded plates, including the joint corrections mentioned above, are available. In 1963, the American Society of Civil Engineers Task Committee published review of these methods [3]. A full history of the study of the development of correction analysis can be made from this publication. Among, the various methods of correction the preliminary analysis. Simpson's method (1958) and Whitney’s method (1959) are more popular than the others. We will deal with the correction analysis in Chapter 16 [3] [4} 15.2.1 General Dimensions Folded plates can span large lengths with little materials and simple formwork. thickness can be as little as 1/200th of the span. Their appearance is also good. Hi they are becoming very popular for covering large spans. ‘As already stated in Section 15.1, there are many types of folded plates of which most popular are the V-type for ordinary spans and the trough type for large sp: Northlight type and cylindrical shell type are used for special purposes. The V-type folded plates is not quife suited to large spans as the thin slabs may not always provide necessary concrete section to resist the compressive stresses and also for the pl of steel for bending in the transverse direction. Hence, we add horizontal slabs at top a bottom (usually about 40 cm in the middle and 20 cm at the ends) to make it into a tro type. Thus, trough type and cylindrical type are more practical type for use for la spans. As shown, Z-sections can be used as northlight roofs. It also has an inclined with horizontal slab at top and bottom. A series of simple units precast on the gro and then lifted up can also be used without use of elaborate formwork. The folded plates are planned by taking into consideration the following points: 1. The inclination of the plates to the horizontal should not be more than 40 de for easiness in placing concrete. A slope below 30 degrees is considered too for action as a folded plate. 2. The width of the plate, in order to avoid deep beam action, should not ex 1/3 the span. This width usually depends on the type of plates. For a northlight Z, it may be as much as 1/3.3 span. For a V-type, we adopt a wid! of 1/5 to 1/6.5 of the span. For a trough type, it is usually only 1/10 to 1/12 of the sp the trough type being used for larger span than others. The thickness of the plates is taken as not less than 1/200 the span, the minim thickness being 100 mm. The horizontal parts of the trough and Z-types are usually thicker than the inclined part. Thus, when the inclined part is 100 mm, we may take horizontal part as 110 to 120 mm. ‘The best method to dimension folded plates is to adopt the dimensions of fol plates which have already been built successfully. 15.2.2 Methods of Analysis Folded plates can be analyzed by the following methods: ‘+ Beam method ‘+ Elasticity method * Slab beam analysis with correction analysis for stresses and rotationid Plates—Preliminary Analysis 205 Modern methods of finite strip method (FSM) can be used for the analysis of these , but we will deal only with the classical method in this chapter. References [4] to [9] a good account of the design of folded plates. In the beam method, the folded plate is treated as a beam of irregular cross section the longitudinal direction and as a continuous slab in the transverse direction. The its of rigorous analysis show that their results are not satisfactory. The second elasticity method for the calculation of stresses in folded plates is difficult. carry out by ordinary calculations. The slab-beam method, popularly known as the ied plate theory”, is the most convenient and satisfactory method for engineering, ign. We will study this method in this book. 3 BRIEF DESCRIPTION OF SLAB-BEAM ANALYSIS OF FOLDED PLATES : This section gives an overview of the method of analysis and should be clearly understood. ‘good reference for study is Reference [5}1 The modern slab beam analysis of folded plates can be divided into the following lures. Details of each procedure will be dealt with in this and the next chapters. In chapter, we study the preliminary analysis in steps 1 to 8. In Chapter 16, we study correction analysis necessary to get correct final results. Preliminary analysis can be jded into the following procedures. (There are various steps in each procedure.) edure 1. Transverse slab analysis (Figure 15.2): We first take the transverse section of folded plate and assume that it forms a frame supported by external supports at the joints. We then assume that all the loads acting on the slab produce bending. sting the system as a continuous beam and by using ordinary Hardy Cross moment ibution method, we can find the bending moment in the transverse direction and also vertical reactions at the imaginary supports. (The thickness of the slab is generally mined from the maximum moment thus obtained, usually increased by about 15% ‘by the results of the correction analysis that will be explained in Chapter 16.) The cements in this direction should ensure the rigidity of the joints and the proper of folded plates. lure 2. Longitudinal beam analysis: In the next part, we remove the imaginary support had placed at the joints and replace it with corresponding equal and opposite vertical at the joints 1, 2, 3... by Ry, Ro, Ry ... as shown in Figure 15.3. These loads are then resolved as in plane loads acting on the two slabs meeting at joint. We assume these loads are carried longitudinally by the slab by acting as between their supports. As the beams are shallow (length more than three times ir depth), we treat them as ordinary beams and calculate the stresses at the “top and \" of the slabs; considering each slab as in bending under these in plane loads in Jongitudinal direction. (They are usually in tension at the bottom and compression at top of each plate.)206 Design of Reinforced Concrete Shells and Folded Plates Figure 18.3 Resolution of balancing reactions into pate loads for beam action Procedure 3. Bringing stress compatibility: The stresses calculated at the junctions as des above for the two adjacent plates at a joint can be entirely different. In most cases, Stresses thus calculated in the plate on opposite sides of a joint will be different in si ‘One may be tension on one side of the joint and the other compression on the other of the joint. But in reality, they must be equal both in magnitude and sign. This ‘compatibility should physically exist. Winter and Pei showed that this equalization of stresses can be carried out by a si method similar to Hardy Cross moment distribution in structural analysis (2. The derivati of this method is summarized in Section 15.5 and illustrated by Example 15.1. The f steps described below called correction analysis are detailed in the next cha be summarized as follows. Note: Preliminary analysis is made by procedures 1 to 3. Each procedure we carry out thr various steps.) Procedure 4, Correction analysis (Calculation of plate deflections and rotations): As well establishing compatibiity of rotation at the joints. This is described in Chapter 16 approximate design for limited cases can be obtained by the end of procedure 3, Foe Seer solution, we have to satisfy not only compatibility of stresses but also compatibi Gf deformations of joints, ie. the joint angles should not change. It is obvious thatided Plates—Preliminary Analysis 207 preliminary analysis where we have two plates joined together, the deflections calculated Pretwo sides of the same joint can be very different from each other. These differences ‘Rhections will greatly affect the stress distribution in the plates. Hence, correction ysis for preserontion ofthe angle at the rigid joints becomes necessary. This is carried out Se se ction analysis. There are many methods to carry out this correction analysis. fe will study the one proposed by Simpson {3}, [8] in Chapter 16. ‘The procedure is to calculate the plate deflections and consequent rotation of joints ge of angle at the joint) that will happen with the results of our approximate analysis. sR ould be eliminated as, in reality, the angles at the joints in the structure do not ergo any changes. There will also be minor changes in the transverse moment obtained Step 1 due to this correction. These minor corrections in the transverse moments, ee tially very small and may be neglected in most cases. The various steps for this correction lysis are also explained in Chapter 16. Procedure 5. Design for shear: In the folded plate which consists of different plates, it is the shear at the joints which makes the independent planks to work together as one unit thn example to show the complete analysis is given in Chapter 17. Chapter 18 give & detailed discussion of the shear at the joints and details of placement of ste! in folded cee he reinforcements at the joints should be designed properly so that no angle nge occurs. is4 DESCRIPTION OF PRELIMINARY ANALYSIS (PROCEDURES 1 TO 3—STEPS 1 TO 8) this chapter, we study only preliminary analysis (procedures 1 to 3 above) by Steps ‘to 8 as explained in Section 15.4.2. The following symbols are used for the analysis (Figure 15:4): (a) Span of folded plate structure, L. (b) Length of slabs L, width d and thickness t. (© Joints are numbered 0, 1, 2 ... clockwise as shown in Figure 15.3. (a) Plate between joints 0-1 is plate 1 and they are numbered clockwise: (6) Inclination of plate 1 with horizontal is denoted by 61, Inclination of place 7 is fb: It is measured from the plate to the horizontal in the clockwise direction. (H Angle between plate 1 and 2 is az, that between mand 1 + 1, Gane, and so set ig measured by extending the direction of the plate and measuring it clockwise. (@) Reaction got by transverse analysis at joint 1 is Ry and at joint » is Rw ‘These are applied as loads on the joints. (hy Reaction Rj, -... Ry are resolved along the plane of the slabs and designed as P faces, The P forces are designated with members indicating its direction joint by208 Design of Reinforced Concrete Shelis and Folded Plates joint. The component from junction 2 (ie. Rz) on plate 2 to junction 1 is called Pi, The component on plate 3 from junction 2 to junction 3 is called Pas. The rapaltant force on a plate is the sum of the forces from the junctions (one above and one below) of the plate. The load on plate 2 is P, (due to component of loads from joint 1 and joint 2). Hence, P; = P21 + Piz + Component of dead load along the plate. (i) Stresses calculated on plates are designated joint to joint. Thus in plate 2, the Stress at the top Gunction 2 to junction 1) of plate 2 is compression fay and the srres at the bettom of the plate 2 which is in tension is fra [ie. fiz = fal 15.4.2. Description of Preliminary Analysis (in the following descriptions, we must distinguish between the terms procedures and steps in each procedure.) The following are the first three procedures in preliminary analysis described in Section 15.3: 1, Transverse slab analysis 2. Longitudinal beam analysis 3. Making compatibility of stresses We will deal with the above three procedures for preliminary analysis in more detail in the following eight steps. (Gee in Example 15.3.) The following are the steps in preliminary analysis: Step 1; Tabulate the dimensions of the folded plate. Step 2: Tabulate the geometric properties of slab such as areas, etc. Step 3: For transverse analysis assume the system as a frame with 1 m of the span a width and supported at each joint by vertical supports. Calculate loads and support moments for transverse analysis. Step 4: Analyze the transverse frame by conventional moment distribution at support Ge carry out transverse analysis of the frame). In moment distribution, the unbalanced ie nent is distributed in proportion to the stiffness 1/L so that the moments on the two Tales will be equal in magnitude but opposite in sign. The carry over factor to the other end of beam will be + 1/2. ‘We take a unit width of the slab in the transverse direction and support it by imaginary support at the joints for the analysis of this structure for UDL by conventional moment i ibution. The spans theoretically should be the effective span between the supports but not much error is introduced if the spans are taken centre to centre of supports Complete the moment distribution and find the transverse bending moments and also the support reactions. Step 5; Calculate the reactions R at the supports. 1 wd , AM Se Pit 15.1 2 dcose org Support reaction where d is the width of the plate.209 ry Analysis fed Plates—Pretimina) & Resolve joint reactions R into P loads acting along the plane of the slabs as m in Fig. 15.4. (Consider joint 2~ Let P, denote load on plate from joint 2 to joint 1.) Next, we remove the support by applying equal and ‘opposite reactions at the joints sa nettion should be shared by the plates meeting 2 ithe joint. Hence, resolve the Sree forces acting in the planes of the plate 25 shown in Figure 154. Then for iat 2, we get (Pai means force from joint 2 to joint v R,_. —& Post Find, O51 08H ae ae Ss ae SRE (angles @ and @ as in Sec. 15.4.1) (5.2) wr plate 2.1) Fy = REE - fie. cos 9 of next platel (45.3) sin or plate 2.3) R,€03¢ te. cos ¢ of previous platel 54) sin Gy », (oad on plate) Z Plate (1) Joint 0 Figure 15.4 Resolution ‘of Ap at joint 2 to plate loads Pp, and Pro: ep 7: Longitudinal beam analysis—Determination of edge stresses. ‘The net inplane loads P2 that we get BY Step 6 on plates produce bending along the th of the folded plate by acting as 2 beam. Bending moment on, say, plate 2 is given by, Pi Mase int 2) and BY ” ads of plate 2 from joint above plate ? (joi re P3 is the resultant of inplane lo: int 1 below Goint 2).oO 210 Design of Reinforced Concrete Shells and Folded Plat The stresses at the extreme fibres of plate 2 are, mips teak faa ina where Z =! Similarly, the stresses on plate 3 will be, tonto (2) Step 8: Bringing stress compatibility by stress distribution procedure of Winter and In theory, the stress on the same edge, say, fx and fps at joint 2 must be the s in magnitude and sign. But in the foregoing calculations (Step 2), they will not be eq We correct this and equalize them as follows. ‘As explained in Section 15.3, Winter and Pei showed this can be carried out by method similar to conventional moment distribution. In this method we assume that 1. The free edge stresses are considered as similar to fixed moments in mo distribution. 1 2. The difference in stresses is then distributed in proportion Cee 1 difference in stresses is then distributed in proportion to (ae ie. ———_!_. The distribution should also resu (Depth = Thickness) of plate at the two sides of the joint to be equal in magnitude and sign. 3. In the distribution, the carry over factor to the adjacent joint (the opposite en of plate) will be -1/2 (instead of +1/2 in Hardy Cross moment distributi The above procedures (1-3) carried out by steps 1 to 8 will give us the transv moments as well as the final longitudinal stresses at junctions of the plate. [These res need further correction for compatibility of joint displacement, which we will examine: the next chapter.] [This completes the approximate solution called preliminary analysis by Winter Pei. But we know corrections are needed for compatibility of deformations, namely, change in joint angles. The error depends on the configuration of the folded plate and. restraints of one plate over the deflection of the other. In general, the results of Corrections are not negligible and a complete analysis requires this correction which will deal with in the next chapter.) 45.5 THEORY OF STRESS DISTRIBUTION METHOD FOR S' COMPATIBILITY (This portion may be omitted on the first reading of this chapter.) ‘Consider two adjacent plates as shown in Figure 15.4. Take joint 1. The stresses obtai by preliminary transverse analysis on the adjacent plates plate 1 and plate 2 are diffePlates~Preliminary Analysis 21 they should be the same. This becomes possible due to the presence of additional shear there joints which we will deal with later chapters. Hence, we introduce shear forces Ty, T; ... at junctions 0, 1,2 ... to make them equal. Let A = d x t and Z=I/y = td*/6, Taking the plate 1, we have fo: as the stress in the top of plate 1. Similarly, taking te 2, fiz is the stress at the bottom of plate 2. TB [end |M A f= 5 ai dx6_ 8 (ado- see Al M, fo o ilarly, for stress at bottom of plate 2, AT, +2T, _M, this should be equal in magnitude and sign, fo. = fin. we get, “(o(tealeahls Ay Ay Z; and Zz are constants and T forces vary as M. For UDL, M is parabolic. This equation is similar to the well-known three moment equation, from which the ‘dy Cross moment distribution is derived Ma (b)}+2Mp (0 +h)+Me(b)+ S454 S429 *e, we adopt the following “stress distribution method” for equalization of stresses [6]. 5.1 Winter and Pei Stress Distribution and Carryover Factors stress distribution, we have to make the stresses equal in magnitude and sign. We luce the following important concepts [1]: 1. The effect of the shear force T, which we introduce to equalize the stresses that meet at the joint is proportional to (4/A,) where A = depth of the plate. This is analogous to saying the stiffness factor for stress distribution is proportional to 1/A,, where A = (width x thickness) of plate. (Thus, the distribution factor will be inversely proportional to the areas and is to be carried out so that the stresses are equal in magnitude and sign.) 2, The effect of shear force T, acting at junction m on the stresses of the plate at junction n ~ 1 is (-2/A,). Compared to 4/A,, (as in above), its value is (1/2), is the carry over factor is 1/2. (Note the sign as different from Hardy Cross moment distribution where it is +1/2.)212 Design of Reinforced Concrete Shells and Folded Plates 3. The free edge stresses are similar to fixed end moment used for moment distribution in structural analysis. The difference in the calculated stresses at ‘each joint for the two plates on either side of the joint should be distributed so that they become equal in magnitude and sign. The distribution factors are in proportion to the inverse of the areas (width x thickness) of the plates, ae ae In this chapter, we examined the first part of the analysis of folded plates which we call preliminary analysis. This is carried out through eight steps explained in this chapter Further analysis is necessary for a complete solution to satisfy the condition that the angles between the plates should not change (compatibility of deformation). In some simple layouts, this simple solution can give a fair solution of the problem EXAMPLE 15.1 [Preliminary dimensioning of a trough shaped (trapezoidal) folded platel Give the preliminary layout of a trough shaped layout of a series of folded plates to cover a shed of span 18 m. oa + 09m fo" oa oe 15.1 Reference | Step Calculations Sec. 152.1 | 1. | Adopt a suitable plate width, say 1/12 of the span We will adopt a width = 1.5 m ‘ 1 Width/Span ratio of plate => =75 <3 2. | Adopt a suitable inclination of plate For concreting * 40°. Adopt 40 degrees. Horizontal projection of plate = 1.5 cos 40 = 1.15 m Sec. 152.1 | 3. | Adopt a suitable “horizontal length” between the inclined plates and for end plate ‘Adopt 0.2 m at ends and 0.4 m between plates. 4, | Adopt a suitable thickness for the plate We usually adopt a thicker plate for the horizontal part and a slightly thinner plate for the inclined portion. Horizontal plate thickness must not be less than 100 mm. Span/Thickness = 18000/100 = 180 (O.K.) (Inclined plate can be lesser, say 90 mm) Span/Thickness = 18000/90 = 200 (0.K.) We may also adopt a uniform thickness of 100 mm.rolded Plates—Pretiminary Analysis 213 EXAMPLE 152. [Proportioning a Z-type folded platel AA Nhe dimensioning of a Z-type northlight folded plate for a factory for a span of 18m. 20 em 40” ‘som Figure €15.2 Reference _| Step Calculations See 152.1 | 1 | Adopt a spanhwidth ratio and slope angle for Z type ‘Adopt span/width = 5.8 > 3; Angle = 40° Width of plate = 18/58 = 3.1 m Horizontal projection = 3.1 cos 40 = 2.4 m; height = 2m 2 | Adopt front and rear horizontal lengths ‘Adopt 1/3 x 2.4 = 0.8 m at front and rear. 3 | Adopt thickness ‘Adopt horizontal portion as 10 cm. Adopt inclined slabs as 9 cm. EXAMPLE 153. [Preliminary analysis without correction for rotation—Winter and Pei solution] The geometry of a trough type folded plate uniformly 100 mm thick and of span 18 m is Trev in in Figure Ex. 15.3. The total load (DL + LL) is 340 kg/m? on plate art ‘Compute the first part of the plate analysis (preliminary analysis) neglecting joint displacement [6], [7]. (Refer Section 15.4.2 for the eight steps.) (Span L = 18 m; width of plate = 14 m (L/B = 18/14 transverse span of assembly = Ex. 15.1] 128 greater than 3); total 88 m; thickness of plates is 100 mm.) [see also, Figure Figure E15. Tough tp folded pate: Joins ropesoed as 0, 1, 2, et, and plates as (1+ (2) at, [Note: Due to symmetry, we consider only the First 5 plates.]weed Concrete Shells and Folded Plates CALCULATION Step 1. Tabulate dimensions of plates (span 18 m) TABLE 1, EXAMPLE 16.3. Dimensions of Folded Plates (Span = 18 m™) Plate No Pate wid gy Inclination to Arle teen Thickness of ‘Horiz. projection horizontal (0) plates (o) plate him) i 2 0 319-58" oa As, 40°22" 40°-2" Oo. al 3 aS o 402 oa oa 4 a 319-58" 319-58" on 1072 5 a 0 319-58" on 04. [Note: ¢ is the angle the plate makes with the horizontal tis measured Koa) plate line (exte flockwise. (See Figure 15.4.) ais the angle the plate makes wi to the next joint) to horizontal, the next plate. It is measured by extending the plate and measuring the angle fo the next ph measured clockwise} Step 2. Tabulate geometric properties ‘TABLE 2, EXAMPLE 15.3. Geometric Properties (in meter units) Longitudinal analysis Plate No. Cross See. Transverse anilysis ‘Area (im?) 1104) (inclined width) (And (1 m along span) 1,00) Z(0%) i 002 0.838 067 0.07 2 ona 0.833 228.67 32.67 3 0.04 0.833 533 267 4 od 0.833 228.67 3267 5 0.04 0.833 53 267 Example fy = 1 « (0.1)'/12 = 0833 x 10 (Thickness of plate 0.1 m) Jy = 0.1 x (0.2)°/12 = 0.67 x 104 Ty/Md/2) = 0.67 % 104/01 = 0.67 x 10°fed Plates—Pretiminary Analysis 215 3, Calewlate load and support moments for transverse analysis taking 1 m of span as width TABLE 3, EXAMPLE 15.3. Loads and Support moments (Consider 1 m width) Self-weight + Other loads = 240 + 100 = 340 kg/m” [Note: Units used for moments are meter kg per meter = cm kg per cm) Plate No. Total load perm Horizontal ‘Support moment along. span (ss) span (a) makg/per m span 02 x 340 = 68 02 (8x02 649 1a x 340 = 476 1.072 Mere sp 0.4 x 340 = 136 oF 136x04 2 1.4 x 340 = 476 1.072 ae ae 04 x 340 = 136 04 pee ae Je: Plate 1—Area of plate for 1m span = 0.2 x1 = 0.2 m. Hence, total load on plate for sverse analysis = 0.2 m x 340 = 68 kg/m span. 4. Carry out transverse analysis ‘TABLE 4, EXAMPLE 15.3. Conventional Moment Distribution for transverse analysis Moments taken in m.kg/m units. (We distribute the difference in moments (with opposite ign) in proportion to 1/L so that the resultant moments are equal in magnitude but epposite sign. We then carry over moments with carry over factor +1/2)] 0.833 0.833 14 2 z }, Stiffness (I/L) ratio = ——— to 4 to 14 = —=ttos it 2, St (I/L) rat rr 04 0. ar G5 Joints 0 1 2 3 Org Plate 4 z a + 3 on 2/9)7/9 7/9) 2/9 2/9|7/9 { 7/9) Fixed end +68) 42.5, +42.5/-4.5, +4.5"| 42.5 4425/45 +45) ‘moment ' Dist. 4357 -ea|-296 296|4+04 84-296 | +296 co, faa vi7ajias-14s[42 +42] +148 | 148 Dist. +42 7.2|-25.4 414.7] 42 —4.2|-148 | 414.8) (Continued)216 Design of Reinforced Concrete Shells and Folded Plat CO. Be sala -127[-2aSCA 7a Dist -2al-74 —sane432 2a|-74_ | co. 15/58 =s7|40 s16|437 | Dist. 10 -t4|-60 —+37|s10 taf-4a_ Final moments |0 +eal-o8 408/448 +327|-32.7 +945|-345 | 134 INotes: ‘L. In simple moment distribution, we find the difference between + and — moments at joints and distribute the difference so that values at junctions are equal in magnitude opposite in sign. 2. The final moments are in m.kg/m span.) Example: In conventional Hardy Cross moment distribution, the moments at the joint from adjacent spans are to be of same magnitude but opposite sign. Hence in this we simply add the two with the sign and the result is distributed with opposite si Example, Joint 2, first line. Add 42.5 — 4.5 = +38, Distribute ~38 in ratio 2/9 and 7/9 84 and -29.6 = 38. The resulting stresses are +42.5 ~ 8.4 = +34.1 and ~4.5 ~ 29.6 = 34 Step 5. Calculate reaction (R) at supports (See Table 5, Example 15.3) load due to wt 4 (from left and right spans) (i.e. Items 1 and 2 in Table 5) i. M * 70s due to moments (It will be easy to arrive at the results by drawing a figure for each plate with the and moments and calculate the values.) (from left and right spans) (1c. Col, 3 and 4 of Table 5, Example 153) ie. ‘TABLE 5, EXAMPLE 15.3. Calculation of reactions A at supports (kg/m) Ttem Description Ro Ry Ry Ry Ry ee Se ee Firion oenmewet 0-0 gem cage Bes sien ee ee 4 eM o 35.49 30.28 =1.707 o dy C08 Oy ouienion —_0=R0 2705 AI 719-2 40=RS_wT7 =m Example. Item 1—Rz from VDL plate 1 on left = 0.2 x 340 = 68 as cantilever. Item 3—R, from moment on left span = (44.8 ~ 6.8)/1.072 = 35.49 kg.217 ided Plates—Preliminary Analysts 6 Calculation of P loads—Resolve R into P Toads in plane of slabs for longitudinal bending Jeo mel & co) (Eq. (15.2)1 sino, sin @4 is from Ryy and Red ‘Note: This expression for Py ae te St tS eS ae ee ee 0.7664 0.7664 sample: (For Table 6, Example 153 below). Using the above formula, 1. Loads on plate 1—Using Eq, (15.2) Rycos # _ Ry cos 40°.2" 0+ sin & x 270x0.7664 ge 4 2. Load on plate 2 £056, , Ro c08 by a sing, sin @ _ ~Ricos0_, Recos0 ‘sin 319°.58” sin 40°.2 -2705_, 3718 = eS, SS ee al xamy Canisay 06432 998.8 kg (See Table 6, Example 15.3) ‘TABLE 6, EXAMPLE 15.3, Calculation of P Values Value of P is taken as positive when it acts right to left Ckg/m span) Plate No. From left support From right support Net load P) Plate 1 0 3223 (0 to 1)* 3223 (0 to 1) Plate 2 4420 @ to 1) +578 (2 to 1) 4998 @ to 1) Plate 3 442 @ to 3) 4326 G to 2) =116 (2 to 3) Plate 4 425 @ to 4) -A78 (3 to 4) -903 (3 to 4) [Notes: “| Notation 0 to 1 indicates Joint 0 to 1. 2. Forces from higher joint to lower joint are +ve as they the plate. Forces from lower joint. 43, P forces are the results ofthe weight and live load from plates. Total P forces plates = Px (L/2) on each plate (see Figure Ex. 15:1)1 produce tension at the bottom of 1s at end of218 Design of Reinforced Concrete Shells and Folded Plates Step 7. Longitudinal beam analysis and determination of edge stresses [Note: We will work in kgsem units in further calculations] Maas = PORE «405 P gin = 050% P kgcm Find stresses in plate: fiz = far = Mayoao/Z TABLE 7, EXAMPLE 16.3. Calculation of stresses at joints In ko/em? Plate No. Max (kg €™) Zz Fralues at ends Joint Stresses at joints om 105) (em) (kg/cm?) ‘on opposite sides (kg/em’) 1 1305 670 ¥1956 1 4+1956/+123.8 2 4045 32,670 21238 2 123.8/-176.6 3 an 2,670 F1766 3 +176.6/-1120 4 3659 32,670 #1120 4 +112/0 5 ° 2,670 o [Note: M is in kg.cm units and Z is in cm units taken from Table 2, Example 15.3.] Example: Plate No. 1 - Mmax = ~1305 x 10° kg.cm; f = 1305/670 = 1.95 Step 8. Bring compatibility of stresses at joints by stress distribution by Winter and Pei method The stresses at opposite sides of the folds are not equal. We introduce shear stresses 0 provide compatibility which has been shown can be carried out by a procedure similar to moment distribution with: +. Distribution factors proportional to inverse of areas (depth x thickness of plates) de Dy: Dias 3 5- Present cases as thickness of plate dis same, it will be inverse of the width of plate. For joint, tot, ieZtot- yi. 2 to Z as shown. 9° 9 For joint 2 = x 14 04 2. The aim in this distribution is to make the stresses on either side equal in magnitude and sign unlike moment distribution where the moments are equal but opposite in sign. 3, Carry over factor = =1/2 (as against +1/2 in ordinary M.D.) [This is very important] Example: For stress distribution, we proceed as follows (Refer Table 8 below), Taking joint 1, first line where both moments are +ve, the difference in stresses is 1956 ~ 123 = 1833. After distribution, we want them to be equal in magnitude and of same sign. We divide it as -1833 x 7/8 = -1603 to moment which is to be reduced and +1833 x 1/8 = +230 to moment that has to be increased as shown in Table 8, given below. We add these to the moments to be balanced so that the values are of same magnitude and sign, i.e. 1956 — 1603 + 230 = +353.——————_ folded Plates—Pretiminary Analysis oe joie 3 where jonamncinend ts ave’ andl theigthes negece the difference is re der 12) — 288, We have to distribute itso that the values ore ‘equal in magnitude pra sign. Hence, we add ~288 x 7/9 = ~224 to 2176 ‘making it equal to ~38. We and sign [pi x2/9 = +64 6 —112.to make it 48, then making the stresses equal them genitude and sign. For joint 2 where both moments Swe the difference in magni the slgnasare the sme. Hence we distribute itvan shown: add ~12 and tA making the both stresses equal to —135. ramus 8, EXAMPLE 15. Wieter and PEI Stress dltrbuton for compat ot Srsses at joints (ka/em’) points 0 1 2 3 4 5 (Pilate 1 2 3 a 5 Dist. F 7/s\ 1/8 2/9779 773\2/9 279|779 Saag Vatoes [HOES Ase] Iza =1BB|=ATE +178 = ~—«+12|0 Dist. 1603) +230) =12| +41 22a) +64 =25| +87, ©O. 4801 +6 cais| +12 =20) +13 —32| 44 Dist. 4s[4 +50|-177, 426| -7 -3|+9, co. 25 25) +05| +13 cee 438/45 Dist. 422) 43 +05 68] 419.5 -2|+6.0 co. it a5 34 SG! =10|-3.0 Dist. 13]-02 “e0l275— a.0]-2 s15[55 phuat-ntionn, tiene ea] tae — 98) 198 -ai}-21 +46) +46 [Note: We call this operation as “stress distribution” and its different from moment distribution.] [Notes: 1. Free edges are considered similar to fixed edge of conventional moment distribution. Free eo Seon joint 4, we go to the symmetric centre tine of the folded plate. 2. Unlike moment distribution, the stresses are to Be the sttus oo magnitude and signe on ane de of the joint (unlike in regular bending moment distribution). 4, ‘hie solution proposed by Winter and Pei assumes that coe plate is free to deflect and ‘This solution cay neateiction. But the’ adjacent plate restricts ye freedom, Hence, the rota on will not be correct untess we take into account Whe ‘corrections described in the solution pler. The magnitude of correction will depen on ‘restriction of deformation next chert ge which will depend on the configuration of Me folded plates. We will between pistorrection calculations in Chapter 16. [This example i continued in Chapter 17 fas correction analysis-]220 Design of Reinforced Concrete Shells and Folded Plates EXAMPLE 15.4 [Preliminary analysis by Winter-Pei method] Carry out preliminary analysis of a V-type folded plate for a span of 18 m shown in Figure E15.4. The thickness of horizontal slabs is 120 mm and that of inclined slabs, 100 mm. Assume applied live loading is 150 kg/m’. The dead load varies with the thickness of concrete. (We will assume an easy slope of 31° and a width about 1/5 span, ie. 18/5 = say 3.5 m giving a rise of the system as L/10 = 1.8 m.) “i 2 oe © 6 LG ‘ 7 5 Flow £184 V type folded lt ois denoted by 0,1 ee and plates by (1), PART I—PRELIMINARY ANALYSIS TABLE 1, EXAMPLE 15.4 Dimensions of folded plates—Span—18 m (Step 1) (Consider to centre line of the system) Plate No. Width remit — Tica Mdegrees) ay? late No. fidth (mm) Horiz. projection ickness (m) On" (degrees) a,**(degrees) 24 . oh 229 1 210 a 0.2 2 298 2 3.50 00 31 oe 3 3.50 0.0 a 298 ‘Symmetric with respect to Junction 3 * = is the angle made by the plate to the horizontal. (See Step 1, Example 15.3.) “* ~a is the angle between the plate and the next plate, TABLE 2, EXAMPLE 15.4 Geometric properties of plates in metre units (Step 2) Plate No, Cross Sect. Area Transverse analysis Longitudinal analysis (mn?) Ip x 109 (mn) 7,004) 1 0.252 O44 0.0926 0.35 0.0833 0.3873 3 0.35 0.0833 0.3573 Example: 1, in plate 1 = 0.12 x (2.1)°/12 = 0.0926 Z is in m’, For conversion to cm units, multiply by 10° when we work in cm units (see Table 7)plates—Pretiminary Analysis (Consider unit metre along span) ) = 150 kg/m?. DL for thickness of TABLE 3, EXAMPLE 15.4 Loads and support moment for transverse 221 analysis (Step 3) slab @ 2400 kg/m? (nsulation + LL) Pate No. Total Load Horizontal span Support moment kg/m span (o») (kgm) 19.8 18 sasxts =8278 1365 30 1365%3.0 _ 341.3 2 1365 x 3.0 1365 30 13654300 368 a +3413 smple: Plate No. 1 - Load = ‘TNBLE 4, EXAMPLE 5.4 Transverse analysis [150 + (0.12 x 2400)] x 2.1 = 919.8 kg by moment dstnbution as supported at joints (Sten 4) (kg and m units) 0 1 2 3 Plate 1 2 3 Dist. factor oft 1/2 (1/2, Support moment “+827 |-341 yan |-341 4341 486 @ 243 +122 | 122 61 él Final values 827|-827 -+202|-202 +410 TABLE 5, EXAMPLE 18. Calculation of reactions at supports (Step 5) (R = W/2 + M/d, cos 4 from both sides) Description Ro Re Ro Rs UDL from left span 0 198 682.5, 682.5 ‘UDE from right span 0 682.0 682.5 682.5 Mic / dy 608 Oy 0 ° 208.4 694 Magna/ dy £08 Oy 0 208.4 09.4 coy Total Reaction 9 1810.7, 1087.2 1503.8 w, M Rin9* Frcs e {Eq. (15.1)]229 besign of Reinforced Concrete Shells and Folded Plates ‘TABLE 6(A), EXAMPLE 15.4 Calculation of P loads (In-plane toads in plates) (Step 6) ‘Trigonometrie values ‘Angle sine eos 329 0.5150 0.8572 31 298, R fq. (154 RB = Sibs 2B Be z "(ee Ne apie ae EXAMPLE 15.4 Calculation of P forces (In-plane loads in plates) In kg per metre units ; STOR 1 i Spa eh 3 2516 (Jt. 2 to Jt. 3) [Note: Loads from lower to higher joints are -ve; loads from higher to lower joints are + ‘TABLE 7, EXAMPLE 15.4. Stresses (f values) at ends (L = 18 m = 1800 cm) (Step 7) (Note: We work in kg and cm units and other dimensions in cm.) Plate 1: Load = 1758 kg/m = 17.58 kg/em PI? | 17.58%(1800)* Pe =4807 kg/em* fez ~*~ 8x 888200 e/a Plate No, P oat egFem) Zien) Fewlem) 1 1758 88200 807 2 28.14 2204200 3558 3 25316 204200 #499 + tension; — compression. 3 INote: We now calculate stresses in kg/cm?)Folded Plates—Preliminary Analysis 225 TABLE 8, EXAMPLE 15.4 Stross distribution for compatibility of stresses (Step 8) Distribution in proportion to inverse of areas, i.e. 1/A and carry over -1/2 Joint ° 1 2 3 Plate T 2 3 Distribution. o.ai9 [0.581 05/05 Stresses 80.7 480.7 [455.8 55.8 [49.9 449.9 =10.4| 414.5 -29|-3.0 $5.2 1S 8. 15] 0.6) +0.9 +37|-3.6 Final values |-74.8 468.8 |+68.8 -57.0|-57.0 453] Now joints which are not free (restrained) have to be corrected for rotation of joints by Simpson method given in Chapter 17. (The theory of the following calculations of part Ils explained in the next chapter (16) and is based on Equations (16.1) and (16.2) of Section 16.3.2.) Part I1—Calculation of Plate Deflection of Plates from Data got from Preliminary Analysis (Span = 18 m) (Theory given in next chapter.)(Refer Eg. 16.1) ={fe-v=fa (12) _[ 180%], int) fu _1-688(fn-s)— fu) 96xd, JL E)~|96x2~10° oe d, TABLE 9, EXAMPLE 15.4 Deflections in Preliminary Analysis Plate No. Fox te Toa fe a a 1 “748 688 =143.6 210 “154 2 +688 “570 4135.8 350 0.607 3 -570 453 -110.0 350 0.530 {This step is showit here to indicate that the deflections of the plate are usually extremely small ‘The procedure is detailed in the next chapter.) 1. Give a short account af the devel plates. What is meant by “transverse beam analysis”, “longitudinal beam analysis” and “correction analysis” as used in the analysis of folded plate structures? lopment of the theory of modern analysis of folded224 Design of Reinforced Concrete Shells and Folded Plates 3. What are the length/width ratios prescribed for folded plates and what is the necessity for these limitations? 4. Explain the following: (a) “moment distributions used for transverse analysis” and (b) “method of stress distribution between joints” used in folded plate analysis. 5. Explain the Winter and Pei method of analysis of folded plates and state why the method needs further corrections to give a complete analysis. 6. In what way does the design of R.C. plates for coal bunkers differ from ACT design procedure of folded plates for roofs? Ce Re [1] Criteria for Design of Reinforced Concrete Shell Structures and Folded Plates, Bureau of India Standards, New Delhi. [2] Winter, G. and Pei, M., Hipped Plate Construction, Journal of the ACI, January 1947. [3] Phase 1, Report of the Task Committee on Folded Plate Construction, Journal Street Eng Division, ASCE, December, 1963. [4] Whitney, C-S,, Andersen, B.G and Birnbaum, Reinforced Concrete Folded Plates, Journal Struct Division, ASCE, Vol. 85, 1959. {5] Belington, D.P., Thin Shell Concrete Structures, McGraw Hill, New York, 1965. [6] Ramaswamy, GS, Design and Construction of Concrete Shell Roofs, McGraw Hill, New York, 1968. 17] Chatterjee, N.K,, Theory and Design of Concrete Shells, Oxford and IBH, Calcutta, 1971. [8] Simpson, H., Design of Folded Plate Roofs, Journal Street Eng. Division, ASCE, January 1958. [9] Chandrasekara, K., Analysis of Thin Concrete Shells, Tata McGraw Hill, New Delhi, 1986. [10] Rao, P-S, and Sitapathi Rao, P., Notes on Design and Construction of Concrete Shells and Folded Plates, Structural Engineering Laboratory, IIT, Madras, 1972.FOLDED PLATES—CORRECTION ANALYSIS 161 INTRODUCTION In the analysis of folded plates by the method proposed by Winter and Pei, explained in Chapter 15, we stopped with compatiblity of stresses, But compatibility of deformation tif joints is also important and this has not been achieved i the preliminary analysis. The plate was allowed to bend freely without any constraint by the adjacent plate to The Plates ttached at the joints. Hence, this analysis in Chapter 15 can be taken only as { preliminary analysis and is incomplete. In this chapter, Wt ‘examine the corrections to aera the results of the preliminary analysis for the conditions that the compatibility oe Tefgemation of rotation of joints is also satisfied. We call this second analysis as correction analysis. {62 BRIEF SUMMARY OF CORRECTION ANALYSIS ‘The important condition to be satisfied in our analysis is that a8 the various joints of the plates of the folded plates are rigid, in actual field condition they do not undergo any Bunge of angle when they are loaded. However, the preliminty ‘analysis assumes that the plates are free to rotate which will cause changes in the ‘angles at the joints. Hence, we ine Pifrat find out what are the changes that have happened in the preliminary analysis rane ermact these changes. Correction analysis should ensure that tere is no change in the angles at the joints. As already stated, several methods are available for this analysis. The three popular methods are: 1, Whitney's method of analysis: This method reduces the problem *o 9 solution of vHnulteneous equations. This method is suitable for computation using ‘modern computers [1]. 225226 Design of Reinforced Concrete Shells and Folded Plates 2. The Simpson method: This is deserved in References [2], [3] and [4]. In this method, we apply arbitrary rotation of each plate successively. 3. Yitzhaki’s method: ‘This is a continuation of the Winter and Pei method, expla in Chapter 15. The method gives a physical picture of the calculations made f the correction analysis. We will study only this method in more detail [5I[6) 16.3 EQUATIONS FOR ROTATION OF PLATES AND ROTATION OF JOINTS We first examine the equations used in the analysis before we describe the method. We must first know that joint rotation correction has to be made by introduction of extra moments at junctions. First of all, we must differentiate between rotation of the plate or plate rotation denoted by D and rotation of the joint. We take the rotation of the plate positive when it rotates in the clockwise direction. The rotation of the joint can take place due to the difference between the rotations of the two plates joined together at the joint. Thus, the rotation of the plate n indicated as D, is caused by the deflection of the plate 7 indicated as y» as well as the deflection of plate m ~ 1 (indicated as y,-1) and plate m + 1 (indicated as Ynq1)- [Se Eq. (16.1) and (16.2) below]. As we will see later, the joint rotation is caused by the difference in the plate rotations meeting at the joint. 'As deflections are involved, we will first derive an expression for bending deflection of a beam with given stresses due to bending f, and fz at its extreme fibres (tension and compression) and then the expression for the rotation of the plate. 16.3.1 Derivation of Bending Deflection of a Beam from the Bending Stresses at Its Extreme Fibres The following discussions show how deflection can be calculated from the stresses. We have the relation ssw 9°34 EF for a UDL. But andPiates—Correction Analysis 227 in a folded plate n, 2 =| fazh\(Z) = © for (16.1) and = 7 for a loading which varies as a sine Aion, = 966 for a uniformly distributed load 2. Equations for Plate Rotation from Deflections and Joint Rotation from Plate Rotation ‘already stated in Section 16.3, pilates. 4 Plate rotation: Plate rotation is a result of beam deflections and hence it will Pe a function of deflection y- From the geometry of Figure 16.1, we can derive the equation for rotation of plates hand n+ 1 as follows. (For details, see Reference (2}.) we want the magnitude of joints rotation due to deflection Yoo fe Anin=1 cea ©) 46-4. Plate rotations due to detection of pats. (y. deletions along the plane ofthe plates edges perpondiclat to plane ot plates) Rotation of joint Dy. .s D, will be a function of deflections of plates ‘and Yqu1 and will be as follows [2]: ] (16.2) sin a, and a doflections of ‘The rotation of plate 1 designated a: n—1, mand n+ 1 given BY Yrs Yn Vp (Cot ys +60 +228 Design of Reinforced Concrete Shells and Folded Plates Thus, the rotation of plate n + 1 will be, 1 Ye Ynsz D, - i 16.2 [Clockwise rotation is taken as positive. The value of D, is a function of three deflections Yns, Yn aNd Yor] 2, Joint rotation: Joint rotation is given by the difference in plate rotation of plates joining at a joint. The notation we use for joint rotation is as follows. If we take joint n, the rotation due to preliminary analysis as 0, the rotation of that analysis is denoted by Dyo. (Later, we will be using unit moments at joints denoted by X; = 1 for joint 1, X for joint 2, etc. The joint rotations caused by these moments for joint ‘n’ will be denoted a8 Dy, Dray etc.) ‘As the joint rotation at joint 1 due to plate rotation Dyo = Difference in plate rotations at the joint, bi Dyo = Dy = Dror (163) For example, Day = Dp — Dy [indicated in Eq. (16.2)] For correction analysis, we first find the rotations that will result due to the stresses in the plates we arrived at by Winter and Pei method. As there should be no rotation, we correct it by unit moment method described below. 16.4. EFFECTS OF APPLICATION OF UNIT MOMENT AT JOINTS Correction analysis by unit moment method uses application of unit moment at joints taking one joint at a time. (We apply a moment of 1 m.kg/cm at a time which is equal to 100 emkg/cm). We will examine the effect of this moment on the plate rotations. ‘Let Day be the rotation to be corrected from our preliminary analysis. We correct by applying unit moments at the joints. Hence, we have to examine the effect of applica of unit moments at the joints. Tt will have two effects. ‘As can be seen from Figure 16.2, it will produce in-plane forces which will cause joint rotation D, the same type we examined in Section 16.3. In addition, the unit mom will create rotation of joints which we call @, which will be small and can be ev neglected. Thus, the application of unit moment will have the following two effects shown in Figure 16.2: 1. The major effect of application of unit moment will be joint rotation D prod by the inplane P forces. It will be as follows. As shown in Figure 16.2, the moment at joint m causes reaction (R forces) at that joint and also at the joints adjacent to it. These R forces at these joints (a ~ 1, m and + 1) wi Produce in-plane forces in the plates (P forces), stresses and rotation of plates as we found in the preliminary analysis. Let us designate these again letters D, corresponding to rotation of plate 1; D, rotation of joint due to applica of unit moment X; = 1 at joint 2. It can be calculated in the same way as calculated Dyg in Section 16.3.folded Plates~Correction Analysis 229 2. We will have another minor rotation due to beam action. Itis the rotation of the joint considering bending in the transverse direction. We will denote this by Eymbol @. The value of this rotation can be found by Eq. (16.4) given below: Cloually, this value is small and can even be neglected. (The equation for Os derived as shown below.) [Note: The following two sections give the theory which may be omitted in the frst reading of the text.) ‘two-fold effects of application of unit moment at joint (support reactions and joint rotation (-) which is usually smal) Figure 16. {644 Evaluation of Plate Rotation Due to Unit Moment When we apply unit moment at joint it produces reactions R at the supports as shown in Fig. 16.2. These reactions produce P forces in the plates. The rotation of these plates due 6 these forces can be determined as in the same way as we found it in our preliminary snalysis, From plate rotations we can determine joint rotation also for each joint This fotation is the major rotation for unit moment. 164.2 Equation for Joint Rotation @ Considering Beam Action Due 10 Unit Moment at Joint 7 {Note! This joint rotation isan additional rotation due to unit moment. Its not a major factor and roe omitted in the fitst reading of the text and students can proceed to Section 16.6) Let us take effect of unit moment as shown in Figure 16.2, where we apply unit moment at joint n. The joints at n— 1, n and x + 1 undergo rotation @. Its value can be Ghined by the second theorem of “moment area theorems” (formulated by Otto Mont stge8) which states that the tangent deviation is equal to moment of the bending ‘moment diagram. Tet d, and dy,x be the width of the plates and unit moment be applied at joint 1. We as the denote the rotation of joint n ~ 1 due to unit moment at 1 as @,-1,.. Assuming, hy thickness of the plate, Moment of BM diagram (ED (dy) 4 Ehip d,xha, |+| BB |x a | (22) a, Bie Eh,230 Design of Reinforced Concrete Shells and Folded Plates Hence, Riis (16.4a) = (16 4b) EW Eins) ses 24, and att ¥ ner. Edin) (16.4c) In general, the value of @ will be very small and usually need not be considered in our calculations except for very exact analysis. 16.5 TOTAL JOINT ROTATION DUE TO UNIT MOMENT The total joint rotation is due to D effect and @ effect, ie. Dix + Ope [For all. practical purposes, we may neglect 6,,.] We will first take the D effect or the rotation of the plates due to the P forces produced. by application of unit moment as stated in Section 16.3.2, item (2). We call it Djs. ‘According to Eq, (16.3), the rotation of plate produces the rotation of joint. Denoting. the rotation of joint n due to application of moment X = 1 and cause P forces a5 Di We get, Doe miDyilloes (165) For example, when X; = 1, we have Daa = Dy ~ Dnst (Of plates due to Xz = 1) Total rotation due to application of unit moment: For the total rotation of the joints, we have to add @ values also to the above joint rotation values. Total rotation, Dy = Dix + Onx ‘Thus, for joint rotation due to application of X = 1 at joint 3, we will have for joint 2 Day = Diy + Oy (166) [As already stated, as @ is usually small, we may delete it in our calculations.) 16.6 DESCRIPTION OF CORRECTION ANALYSIS Based on the theory explained above, we adopt the following procedure. As already stated, correction analysis is a continuation of the “preliminary analysis” described in Chapter 15. The preliminary analysis accomplishes only stress compatibility. Correction analysis brings in joint rotation compatibility also. The following procedure is commonly used for correction analysis.folded Plates~Correction Analysis 231 Part I—We carry out the preliminary analysis as described in Chapter 15 by steps 1 to 8. These eight steps in the Winter and Pei method are (see Section 15.3): Step 1: Tabulate details of plate. Step 2; Tabulate geometrical properties. Step 3: Tabulate loads and moments for transverse analysis. Step 4: Carty out moment distribution. Step 5: Tabulate R forces. Step 6: Tabulate P forces. Step 7: Tabulate stresses (f, and f:) at ends of plates. Step 8: Carry out stress compatibility. Part 1—Calculation of joint rotation from results of preliminary analysis (Dy values) Continue preliminary analysis to find the resulting rotation of joints which we have to cancel by correction analysis. This is carried out in the following three step: Step 9: Calculate and tabulate the deflection caused by the stresses obtained in primary analysis using Eq. (16.1). Step 10: | Calculate plate rotations, D,, using Eq. (16.2). (We designate the rotation to be tancelled at joint n as Dyy as calculated in the next step.) Step 11: Calculate the joint rotations using Eq. (16.3). These are the joint rotations we have to cancel D,o. (Note the rotation.) (This is the end of the primary analysis.) Part II—Unit moment method for correction analysis Apply unit moment Xz = 1, X; = 1, etc. at each joint successively one by one. (Usually, the first plate is free to rotate and hence the first joint we take for unit moment application is joint 2, Xz = 1.) Note: For convenience in calculation instead of X = 1m kg/m we work with X = 100 cm kg/em. The resulting values are multiplied by 100 to get m kg/m] item 12: (a) Calculate transverse beam rotation @ using Eq. (16.4). [As already stated, this will be a small quantity and may be omitted for a rough calculation.] (b) Calculate R forces due fo unit moment and repeat steps 5 to 11 of preliminary analysis (called items 5 to 11 indicated here) of Chapter 15 and find joint rotation, D values. item 13: Add @ and D to find total rotations of joints due to application of unit moments ata joint. Just as Dye denotes rotation to be corrected obtained from preliminary analysis, Dyx denotes rotation at n due to X,, equal to unity applied at joint m.) Part [V—Solution of simultaneous equations Dy is the rotation we get after Winter and Pei preliminary analysis. There should be no rotation of the joints as the joints are rigid. Hence, we write own the condition that the rotation of each joint including the rotation in primary analysis should reduce to zero. (Stated otherwise, the rotations produced by Xz, Xs, etc. should cancel the rotation of the preliminary analysis.)232 Design of Reinforced Concrete Shells and Folded Plates For joint 2, XpDnz + XaDaq + XqDay «+s Do = 0 67) For joint 3, X:Dy + XsDa3 + XeDaq --- Dao = 0 From these, we can solve X;, Xs, ete. and also correct the Winter and Pei stresses and get the real stresses for calculating the reinforcements required as follows. Part V—Find “corrected final design values” of transverse moments and f, and f; values for midspan In step 4, we carried out the transverse analysis in the Winter and Pei method. This transverse moments are used for the design of transverse steel and thickness of plate. To the results of the transverse moments obtained by preliminary analysis, we have to add the effects of the value of the joint moments X1, X2, etc. obtained by correction analysis to get exact values, These give the design moments at the joints. ‘Similarly, the stresses f, and f; are used for the design of longitudinal steel in the slabs, The final values of f; and f, for design are obtained by taking the results of the preliminary analysis and adding to them the effect of the values of X;, Xo, ete. obtained in the correction analysis. "These values are for the midspan. For other places, say, for quarter span, they can be obtained as follows. Part VI—Calculation of design moments and stresses for sections along the length of the folded plate other than midspan [4] ‘As already stated both the preliminary analysis and correction analysis are made for midspan. For the application of these results for the other sections we may assume following as shown in Tables 16.1 and 16.2 given below. 1. The transverse moment of the preliminary analysis is constant along the span. 2. The longitudinal stresses of the preliminary analysis vary parabolically. 3. The effects of the correction analysis. The moments and stresses produced by X values can be assumed to have sine variation along the span. TABLE 16.1 Design moments at any point along the span (Say at quarter plan) Joint Preliminary Analysis Effect of Correction Moment 1 Constant along span 2 Sine variation along span 3 ‘TABLE 16.2 Design longitudinal stresses along span (say at quarter span) Joint Preliminary Analysis Effect of Correction Moment 1 Parabolic variation along span 2 Sine variation along spanfolded Plates—Correction Analysts 233 16.7__ PROCEDURE OF COMPLETE ANALYSIS OF FOLDED PLATES We may summarise the complete analysis of folded plates as follows: As the complete analysis of folded plates will involve many repeated type of calculation, we divide the analysis into the following parts. The following is the summary of the operations. Part Preliminary analysis: This procedures consist of steps 1 to 8 described in Chapter 15. Part I—Continuation of the primary analysis to find the joint rotation to be corrected: With results from Part I, we continue steps 9, 10 and 11 as follows to find Dp values. Step 9: Calculation of plate deflection y due to stresses obtained in the preliminary analysis. Step 10: Calculation of plate rotation of the plate due to the deflections, Step 11: Calculation of joint rotation due to step 10. Part IIA pplication of unit moment at each joint for correction analysis: (We generally apply unit moment as 1 m.kg/cm.span or 100 cm.kg/cm.span.) Unit moment X = 1 is applied successively joint by joint. liem 12; Carry out transverse analysis for moment X rotation 9 (beam rotation) as described in Section 16.4.1. 1 at various joints to find joint liom 5: Find reactions R due to moment X = 1 and rotation D. This is done by repeating the preliminary analysis steps 5 to 11 to find rotation D. (We will designate them as items $10 T1 as they belong to correction analysis.) lem 13: Find total joint rotation 8+ D. (As 6 is small, we may neglect it for convenience.) Repeat the procedure part III for each joint for X;, Xs, ete. Part IV—Solve for X:_ Write down the condition that the joint rotation of preliminary analysis will be cancelled by the total rotation due to Xz, X, etc. for each joint. Past V—Make correction for the preliminary analysis data: (a) for the transverse moments used for the design of the thickness of plate an transverse steel and (b) the stresses in the plate used for design of the longitudinal steel. Part VI—Interpolate values: The above analysis is made for the centre of the span. We ‘ay interpolate the intermediate values along the length of the plate (values at quarter span) as indicated in Tables 16.1 and 16.2. The above procedure used for correction analysis is illustrated by the following example. EXAMPLE 16.1 [Correction analysis of trough type folded plates in Example 15.3] Indicate the correction analysis by the Simpson method for the data obtained by no ‘tation (preliminary analysis), steps 1 to 8 in Chapter 15, Example 15.3, trough type folded plate. Part I: (For steps 1 to 8 as given in Section 16.5, see Chapter 15.) Part 1: Calculation of joint rotation with data from preliminary analysis (continue part 1).| en 234 Design of Reinforced Concrete Shells and Folded Plates 20 9 £ par o4 ala, vo dete tales } | Figure 16.1 Application of unit moments at joints 2 and 9 simultaneously. Continuation of Preliminary Step 9. Calculate deflections from stresses calculated in Chapter 15 [Bq. (16.1)] yy = fp E)- "~0.6Nd,) (E 4, L = 1800 em, E = 2 x 10° kg/cm? and f in kg/cm’. TABLE 9* Calculation of plate deflections y from preliminary analysis (We use kg and cm units for easiness of calculation) Plate No. (kg/cm) , (em) Yn (mn) i i =1145, 4334 20 vy = 124.30 a 4334 “195 140 Ya = +638 3 195, 21 40. = 7.37 4 -21 +46 40 ve = 0.80 5 +46 +46 40 ¥s=0 “This problem is a continuation of Example 15.3 of Chapter 15 which ended in Table 8. Hence: mark continuation of Table 8 of Example 15.3 as Table 9; 4, = width of plate n; y~ deflection, Step 10. Calculate plate rotations (D,) LEq. (16.2)] 1f_ Yas Yoo p, =—2 {tea 7 tas Yu (cot ay. +0 a+ Hest Working in centimetres, 1 f_-1245 (737) pease oe 12)+ E257) | » sa 319.58 ~638(cot 319.58 + cot 40:2)+ 757) |__ 299 1f_638 0.80 Ds [2 Goa 797)(c0t 40.2.+ cot 402)+—2 a |--06folded Plates—Correction Analysis 235 1 [_-737 D=-a ral in 402 Ds = 0 (see Table 10) (-0.8032)(cot 40.2 + cot319.58. 0] .0818, Step 11, Calculate joint rotations (Dg) For joint ‘n’, it is Dy — Deas) Day = Dz — Dy = -1.299 ~ (0.656) = 0.643, Da = Ds ~ Ds = (0.656 — 0.082) = -0.738 Dy = Dy — Ds = 0.0818 - 0 = 40.0818 TABLE 19 Plate Rotations and Joint Rotations in Proliminary Anab Plate Plate Rotation Joint Joint Rotation z =1299 2 Dag = -0.6434 3 0.656 3 Dy = -0.7374 4 +0.0818 4 Dy = +0.0818 F the joint rotation is positive, it means enlargement of the angle at the joint. Part INA Correction Analysis Joint 2 Correction analysis for Xp = 1: Apply unit moment at joint 2 (X; also 100 cm.kg/cm or I mkg/cm for easiness of calculation) Notes: 1. As plate 1 is a cantilever, joint 1 is free to rotate. Hence, first apply unit moment to joint 2. Xz = 1 mkg/m or 100 em.kg/em. 2. We apply correction moments 1 m.kg/em span for convenience of talculation. Also, this we put as 100 em.kg/em span. This should be specially noted.| INote: As already stated, the steps corresponding to the preliminary of correction analysis are Sesignated as items.) lem 12: Calculate “beam” rotations (a) @ [Eq. (16.4)] and (b) D for X, = 100 cm.kg and rotation D. [We will work with kg and cm units.] dy = 140 cm; ds 130 em; ht 10 cm (a) Beam rotation @ Since joint 1 is free, no rotation is developed. With joint 2, as quantities are small, epply Xp = 100 em.kg/cm span length [From Eq. (16.4)] ag = 421OOM 421004; __400_(149 + 40) = 0.00036 = 0.036 x 10° Ely En 2% 10° x (oy Bq = 2a 20010 Lo. 904x 107 Eby 2x10" (10)°bt 236 Design of Reinforced Concrete Shells and Folded Piaict (These values are small compared to D values got from deflection as given below due to X_ = 100 em-kg/cmspan.) ‘TABLE 0 values for X- = 100 cm.kg/em Tem Joint 1 Joint 2 Jot 3 100 em.kg/em, 0 (0.036 x 107 0.004 x 107 x [Note: These values are small compared tothe joint rotation values ‘got from deflection as given below and Monee need not be taken into account for rough analysis:] [Note: Asthenextsteps correspond tosteps5to1! ofpreliminary snaly thenextcorresponding “steps” fis analysis will be referred as “items”. Thus, eee corresponds to step 5 of pretiminary analysis] Calculate beam rotation D for X; = 100 cm-kg due to P foress stem 5: Reactions R at joints are on application of moment Xs = 100m kg/cm [Eq (152) oe Be 1G, cos@, 140 cos 40.2 0.9322 kg/em run 100 =-25kg/em ru AO? = -2.5kg/om run [Note: { we were to use X; = 1, the resultant values would have Peon small] item 6 (same as step 6 in preliminary analysis)—Resolve K into P forces Resolve R,, Rp, Ra, etc. due to in-plane loads Py, Pa, Pay etc. and repeat stress determinat Riess compatibility as in Chapter 10. Taking ¢ a5 angle, non(Sa)a(Se) Ry will affect pates 1 and 2—Py>*Po-1 Ry will affect plates 2 and 3—P22*Ps Ry will affect plates 3 and 4—P3.2*Ps (These will be kg/cm run along span) Find stresses due to Py, (Pini+ Pa), (Poay+ Ppa) and Pe a.-( cits Joennsa( soem TABLE FOR ITEMS 5 AND 6 Val 1s of Rand P for X: = 100 em.kg/cm run R values (kg/cm run) BR values (ke/em run) 0.9322, 3.4322 (The P values are given in Table above.) _—————=— |folded Plates—Correction Analysis 237 Item 7 (same as step 7 in preliminary analysis)—From the table above, determine bending ‘moment and find stresses fio and fro 2 fu (2 i £ ) for UDL, where L, hand d are in cm [h = thickness} Baa, {The value of Z = hd?/6.] 2 2 'we adopt a sinusoidal loading, hence instead of (=) we use, (4}: (1800)? 6 < —* __ |xp, = 3 eae mt t90omo ‘TABLE ITEMS 6 AND 7° P forces and f values for X; = 100 cmkg/em along span Plate Value of Pikg/em.ran) Fikelem’) fk em’) 1 va +5461 546 & 3.88 439.0 39.0 3 7.07 869.4 4869.4 4 3.89 +39.0 39.0 “Items 6, 7 are similar to steps 6 and 7 of Chapter 15. lem 8 (same as step 8 in preliminary analysis)—Make stress compatibility (for X; = 100 cm.kg) (Distribution factor 1/Area) (Carry over factor -1/2) TABLE ITEM 8* Stress compatibility for X; = 100 kg.cmvem (Similar to step 8, Chapter 15) jonat 0 —e 2 3 4 DF. 7/8] 1/8 2/9779 7/9] 2/9 2/9] 2/9 Stress | +546 546 [+39 —39| 869 +869 | +39 39 Dist. +512|-73 +184 | +646 646 | +184 +8.7| -303 256 4923 +365 | +323 323-44 92] 152 32598 4286/4286 =13933 [1393 185 [+153 68 | 68 DF. = Distribution factor; Dist, = distribution; C.O. = Carry over. Item 9 (same as step 9 described in Section 16.6 in preliminary analysis, part 1— Calculate deflection of plate and tabulate [Eq. (16.1)]238 Design of Reinforced Concrete Shelis and Folded Plates TABLE ITEM 9 Calculation of detections X, = 100 kg.cm/em Plate No. a, (em) ies te ts y (em) 1 20 4259.8 +286 +18.95 2 140 +286 139.3 4197 3 40 139.3 +153.0 11.98 4 140 +153.0 68.0 42.59 5 40 68.0 68.0 o ¥ = 1.688 (fn ~ fy)/dy (See Part I—Step 9 at start of Example 16.1.) Items 10 and 11 (same as steps 10 and 11 in preliminary analysis, part II, Section 16.6)— Plate rotations Plate Plate Rotation Joint Joint Rotation 2 Dz = 0.3438 20 Dap =D, — Dy = 1.2342 3 Dy = -0.8904 3 Dz = Dy ~ Dy = -1.0234 4 Dg = 0.1330 4 Dy = Dy- Ds = 0.1330 (item 12 was carried out at the start of part IIL) Hem 12 from page 235 and 13 (same as step 12 in preliminary analysis) ~ Calculate total joint rotation D + 6 Total joint rotation of joint 2 = Sum of the (rotation of the beam + rotation of slab) (Steps 4 and 5) Dag = 0.036 x 107 + 1.2342 = 1.23456 Daz = 0.004 x 107 + (1.0234) = -1.02336 Dg = 0.1330 TABLE ITEM 13. Total rotation of joints for X, = 1 Joint rotation Transverse beam Rotation from Total rotation rotation @ from Step (1) ___deflection (D) from Table 10 (@+D) Dn 0.036 x 10 1.2342 1.23456 Dea 0.004 x 107 1.0232 Ppa Da ~ 0.1330 1330 [Note: We may neglect 6 as it is small.) Part IB Correction Analysis Joint 3 (Apply X3 = 1) Apply unit moment at junction 3 (X; = 100 cm.kg/cm and repeat steps) Item 12. Find beam rotation 033, 643 and 64, 2%100xd, __200%40 = = 0.004x107 Eh, 2x 10° «10 2 ‘aysFolded Plates—Correction Analysis 239 2x100xd4 __ 200x140 Eh. 210° x10° x = 2(0.004 + 0.014) x 10 = 0.036 x 107 (This is twice sum of 035 and 634) (These are also small and can be neglected.) Oy= =0.014%107 TABLE FOR X; = 1 (09 values for %) Os os On X=1 0004x107 0036x107 OIE x 107 em 5: Find adjacent reactions R;, Ry and Ry Item 6: Resolve into P forces and find stresses and carry out stress compatibility (Xs = 1) (Plates affected on either side of joint 3, ie. plates 2, 3 and 4, 5) Net values, -3.887 kg/cm run 7.058 kg/em run 3.887 kg/em run Ps = Due to symmetrical load, there will not be any load on Ps. TABLE FOR ITEMS § AND § Values of A and P for X= 1 Bales (kg/m) P walues (kg/em) ~3.887 Item 7: Find fy stresses due to P for X= 1 and use distribution method for stress compatibility TABLE FOR ITEMS 6 AND 7 P forces and f values for Xs Plate No. P values fi fh i -3.887 39 30 2 7.058 +868 868 3 3.887 39 +39) 4 Nil = 4240 Design of Reinforced Concrete Shells and Folded Plates Item 8: Make compatibility of stresses (Xq = 100 cm-kg/cm run) TABLE FOR ITEM 8 Stress compatibility for X = 100 kg.onvem ° 1 2 3 4 iz 7/8| 17/8 2/9 [7/9 7792/9 2/9] 7/9 39 +39 | +808 868 |-39 439] =34| +5, +184 | ~645 +645 |-184 -38| +30 47 92, =25 | -922 $322 | 444 y92|_—15 446.5 =91,9| 91.9 +165.1 | +165.1 156.7 | -156.7 +68] 468 Item 9; Find deflection of plates as in Step 11.1 (y values) for X = 100 kg.cm/cm ‘TABLE FOR ITEM 9 Deflection calculations for X = 100 kg.cm/cm eS a gee eee Plate No. fa) fo) ym) 1 +46 91 y= 1135 2 1 +165 ya = -3.01 a 165, “156 yy = 13.19 4 156 +68 va = 2.63 5 +68 +68 Ys=0 Items 10 and 11: Find rotation of plates (D values) and joint rotations ‘TABLE FOR ITEMS 10 AND 11° Plate and Joint Rotation 0 for Xj = 100 ka.cew/em Plate No. Plate Rotation Taint No, Joint Rotation (Slab) 2 D, = 0.0204 2 Dy =D; ~- Ds = -1.025 3 Dy = +1.0016 3 Dy = Dy- Dy = 1.1511 4 D, = 0.1465 4 Duy = Dy ~ Ds = -0.1465 *Similar to Table 10 for Chapter 16. Item 12 was carried out at the beginning of Part TILA. Item 13: Total rotation of joints (Beam + Plate) rotations (0 + D) for Xz = 1 kg.cm/cm Due to X; = 100 kg.cm (Note subscript 3) Day = 0.00004 ~ 1.025 = -1.025 Dsy = 0.00036 + 1.1511 = 1.1514 Djs = 0.00014 ~ 0.1465 = -0.1463