Structural Engineering Documents Jérg SCHLAICH Hartmut SCHEEF 4') [dita BOX-GIRDER BRIDGES ion for Bridge and Structural Engineering IABSE n Internationale des Ponts et Charpentes = AIPC ereinigung fiir Briickenbau und Hochbau = IVBHAbout the authors: Jorg SCHLAICH ‘Stuttgart. Studies of Archi ‘and Civil Engineering, Dipl.-Ing. of Technical University, Berlin, Master of Science of Case Tech, Cleveland/ USA, and Dr.-Ing. of University Stuttgart. Professor and Director of the Insti- ‘tut flr Massivbau, University of Stuttgart. Teaching and research of ‘reinforced and prestressed concrete structures and of lightweight-struc- ‘ures (cables, membranes). Consulting Engineer and authorized Check Consultant, Schiaich & Partner, Civil Engineering Consultants, Stuttgart, Fed. Rep. of Germany. Selected Structures: Television towers at Kiel and Kéin, Shel! roofs, Cable-net roof for Munich Olympics, Cable-net cooling tower at Schme- hausen, Cable-stayed bridge over river Hooghly in Calcutta, Solar Chimney at Manzanares. & Hartmut SCHEEF Born in Stuttgart in 1944. After completing his Civil Engineering studies at the University of Stuttgart with the Diplom-Ingenieur, he joined Silt chaps nner aetna sulting firm Leonhardt and André, ‘Stuttgart. Since 1975 he has been involved in teaching and research ‘at the Institut fir Massivbau, Univer- sity of Stuttgart. His research is de- voted mainly to computer-aided nonlinear analysis of statically in- ‘determinate reinforced and pre- stressed concrete beams.SAFE 1G Structural Engineering Documents DIPARTIMENTO DL INGEGNERIA, STRUTTURALE POLITFGNICO DI MILANO as ae on 10% e OG CAM - doe Jérg SCHLAICH Hartmut SCHEEF CONCRETE BOX-GIRDER BRIDGES International Association for Bridge and Structural Engineering IABSE Association Internationale des Ponts et Charpentes AIPC Internationale Vereinigung fiir Brackenbau und Hochbau = IVBHCopyright © 1982 by International Association for Bridge and Structural Engineering All rights raserved. No part of this book may be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the publisher. ISBN 385748 0319 Printed in Switzerland JABSE - AIPC - IVBH ETH - Hénggerberg (CH.8093 Zirich, Switzerland Tol.: 01/377 2647 Telex: 822 186 JABS CH Tolegr.: IABSE, CH-8093 ZirichFOREWORD The box girder is today the most widely used superstructure in concrete bridge construction. That fact justifies the suggestion made by the Commission Ill of the {ABSE that a comprehensive survey be wittan concerning this particular bridge type. The authors proceed from the assumption, however, that its contents wil first be drawn upon when all possible design alternatives for the particular bridge pro- ject have been thoroughly examined, and the box girder has been proven appropriate. Their aim is less that of encouraging the one-sided propagation of box. girder bridges but rather much more that of con- tributing to the improvement of the quality of such bridges. They hope to contribute to this by exten- sively relieving the engineer of the study of today’s hardly surveyable mass of literature on the subject 0 that he can better devote that time to the actual design of the bridge. That explains why this paper is kept short, why in particular cases the reader is referred to the literature, and why subjects not per- taining to the central theme are only touched upon and not handled exhaustively. For greater clearness, the survey follows the sequence of a practical bridge design process by dividing itself into three main parts, namely, “Design”, “Structural Analysis”, and “Dimensioning and Structural Detailing”; each section with its individual numbering and literature list. This survey directs itself especially to the design engineer, which manifests itself, for example, in the fact that the construction methods are hendied only briefly and in the section “Design”, because they decisively influence the design at the very beginning, ‘Major contributions to Section ll, “Structural Analysis”, ware made by Prof. Dr.-Ing. Kurt Schafer, a cal- league of the authors in the Institut ir Massivbau at the University of Stuttgart. In this section the attempt is made to portray the calculation of the box-girder sectional forces resulting from eccentric vohicle loads with consideration of the folded plate action or profile deformation so comprehensively that itis not only easily understood but also rapidly applicable in the design office. This thereby elimi- nates the often-discussed, controversial question as to whether the effort involved in the “exact” calcu- lation of this loading case is actually worthwhile or whether an estimation of the transverse load distrib- ution would not suffice. The authors would like to take this opportunity to thank Professors R. Favre, Lausanne, and C. Menn, Ziirich, for their critical examination of the paper. They are indebted to Mrs. I. Paechter and Mrs. E. ‘Schnee for their conscientious preparation of the manuscript; and Mr. E. Kluttz for his empathetical translation of their German original into English. ‘Stuttgart, January 1982TABLE OF CONTENTS Part! DESIGN i: 2 3 4a 5 istorical Development ... . ‘Over All Design 41 Design Principles 3 am a 4.1.1 The Role and Sequence of the Design Process 4.1.2 Remarks as to Form 413 Costs .. 42 Construction Methods 43° Superstructure... . 431° General : 43.2 Longitudinal Direction . 433 Transverse Direction 44 Complete System ond Supports... . . . 45 Substructure... Bm ; 48.1 Abutments se 452 Piers... .. veces 453. Foundations ncn a Literature soeeen Fn in 15 aN a a . Part Il STRUCTURAL ANALYSIS 1 2 2 5. Introduction... . . . engi awe Loads and External Influences... ooo e eee eseee Structural System. 3.1 Final State... . 32. During Construction - teen eee wee J Loadings and Sectional Forces. - feces 4:1 Longitudinal Direction . seeensTwawe VR He ES 42 Transverse Direction . : ‘Simplified Structural Analysis of the Superstructure. 51 General... ..... - - 52. Analysis Procedure. 53. Analysis in the Longitudinal Direction 5.2.1 Sectional Forces due to Loads plus Restraints. a 822 Time-Dependent Alteration of the Sections Forces by Creep... 3 54 Analysis in the Transverse Direction . = eee . 5.4.1 Sectional Forces Acting on the Flanges... . 542 Analysis as a Frame io 543. Transverse Bending Moments in the Haunch of Variebe- Depth Girders 6... 55. Folded Plate Action... . . ania meen ne * 55.1 Fundamental Concept 552 Solution by Means of the Anaoay of & Beam on an Elastic Foundation 5.6 Multiple-Cell Box Girders cnet a 57 Curved and Skew Box-Girder Bridges... . Computer-sided Analysis of the Superstructure 6.1 Theory of Prismatic Folded Plates... . . 6.1.1 Hinged Folded Plates nee 6.1.2 Folded Plates with Rigid Joints eens 62. Finite Strips : SEE 63. Finite Elements . SSBBSSar res saws an BEIRRSKLSLLSSSSSSSESEKSSESSS ER7. Abutment... . ae3 : n BP eee n B1 Loadings eee n 82 Effective Length... . : : B 83. Moments According to 2" Order Theory... - ™ 8. Foundation % 10. Literature aanecan 70 Part ill DIMENSIONING AND DETAILING 1. Introduction... . . wees vette cee ee 9 2. General Detailing Principlea . 13 3. Prostressing . . - - : - oa 31 The Level of Prestress.. 6-2... : 83 32 Tendon Profile in the Transverse Direction 5 321 TopFlenge..............2- : 5 322 Prestressing of the Webs... 6. eee eee 87 323 Bottom Flange .. . . 87 33. Tendon Profile in the Longitudinal Di etacee8 sn wa 87 33.1 Inthe Webs a 87 332 Inthe Top and Bottom Flanges : 0 34. Transfer of Concentrated Prestress Force: a1 35. Construction Joints and Coupling Joints ® 4. Dimensioning of the Top and Bottom Flanges . - 5 §. Dimensioning and Reinforcement of the Webs... . BRE 97 5.1 Dimensioning for Sheer, Torsion, and Transverse Bending. 37 2 WebReinforcement ... 2.66.0 e eee eee eee eee eee ee 98 6 Transverse Diaphragms ........ 2... seen cece 100 7. Abutments, Piers and Foundations : 102 & Bridge Bearings . . 103 81 Bearing Types 103 82 Installation and Maintenance . . . 103 83. Design of Bearings z ve 104 9 BridgeFinishes 6.0.0.2... . 106 8.1. Expansion Joints F 2 108 92 Bridge Railing ...... 107 93 Roadway Surface... ... : 107 94 Drainage . . oes : : : 2. 107 WO. Literature eee : ~ 108 Note: Parts 4, I, and fll form @ whole and are only divided for organisation reasons. Should the reader be referred to figure, @ section, or a reference in one of the other parts, he will find that the Roman numeral of the other part of the text is placed before the Arabic number; for example, Figure Il,7 or Section Il, 9.1Part! DESIGN 1, TERMS, SYMBOLS SECTION a-a Terms, Symbols. 1 Fig. 1 Sections through a typical simple box-girder bridge The Elements of a simple Bo: Foundation 1 plate 2 pile plate 3 hored pil 4 driven pil piers (with bearinga) edge beam end diaphragm bridge seat support walle bridge seat beam access chamber bearing (can be fixed or allow movement) expanaion joint Superatructure 19 tranaverse dlaphragn: (at abutments, within the span and over the piera} with opening box-girder web top slab (area between the webs) top stab (cantilever section) bottom slab faxcla beam guard rail railing sealing membrane wearing surface drain inlet crose drain Jongitudinal drain By and large the text and formulas use CEB or ISO symbols.2 Design 2, INTRODUCTION ‘Though box-girder bridges are indeed often not the only solution to a bridge project they are, however,seldom the absolutely false one and really only excluded in the case of very small spans or sharply skewed bridges, This universal applicability they owe, from the point of view of load-carrying, to their indifference as to whether the bending moments are positive or ne- gative and to their torsional stiffness; from the point of view of economy, to their suitability for a factory-like construction sequence; and finally, from the point of view of form, to their sleek lines with which they fit into every landscape and surroundings. He, though, who expects a structure to reflect the flow of forces within it through its outer shape will regret the above-mentioned neutral load-carrying bebaviour of the box-girder bridges, especially those of today, As a result of the need to construct the bridges economically, the development has tended towards bridges with constant depths, even for varying spans. Box-girder bridges therefore deserve special care and attention with respect to pleasing proportions and conscientious shaping of their details. ‘The above mentioned characteristics have made the concrete box-girder bridges the most widespread bridge type today. The fact that this develop- ment will continue for some time justifies this survey concerning a topic that is certainly much too specialised for he who does not like to see his design possibilities restricted, Therefore let one be reminded that the box girder is only a part of the entire bridge structure and that the directly supported box girder continuous over the supports is only the standard case. The box girder can also be found in portal frame bridges, arch and bow bridges, and cable-stayed and suspension bridges of all kinds (Fig. 2). Fig. 2. Bridges with the box girder as main structural elementHistorical Development 3 3. HISTORICAL DEVELOPMENT ‘The first bridges of reinforced concrete were built as were their predecessors of stone, They were arch bridges with a gravel fill for the road surface between the two bridge parapets over the arch, Later the gravel fill was replaced by a transverse roadway slab, and the transition to a box girder was achieved. ‘The world’s first reinforced concrete bridge, an arch bridge, was built in 1875 [1]. Probably the first box-girder bridge was Hennebique’ s Risorgimento Bridge in 1911, a 3-hinged arch (see Fig. 3). SECTIONG-al SECTION b-b 1 3,50 fef-13,00—+fef 3,50 Fig. 3. Risorgimento Bridge Due to improvements in the quality of concrete and steel as well as a better understanding of the material behaviour of reinforced concrete, the arch lost more and more on importance as a load-carrying system to that of the beam loaded in bending. The longest span simply reinforced bridge to-date was built in 1939 at Ville-Neuve-St, George, a three -celled, thin-walled, vary- ing-depth box girder of three spans with a 78 m middle span, It was only with the development of high-strength prestressing steel that it became possible to span longer distances. The first prestressed concrete bridges, most of 1 or I-cross-section, were built towards the end of the 1920's. The great breakthrough was achieved only after 1945. The Sclayn bridge over the river Maas, which was built by Magnel in 1948, was the first continuous prestressed box-girder bridge with 2 spans of 62, 70 m(see Fig. 4). In the following years the ratio of wages to material costs climbed sharply. This thereby shifted the emphasis of development to the construction method. section prestressing cable (without bond) \ SoS Fig. 4 Sclayn Bridge over the river Maas; Prestressing by means of straight, large bundles of parallel wires without bond4 Design Important development stages were the following: = cantilever construction in situ: the bridge over the river Rhine at Worms in 1953 with a main span of 114 m, the Hamana bridge in Japan in 1978 with a main span of 230 m, the cable-stayed Brotonne bridge in 1978 with a main span of 320 m - cantilever construction with precast elements: the Chillon Viaduct in 1969 with a main span of 104 m, the cable-stayed bridge over the Columbia River in the USA in 1980 with a main span of 300 m - construction with a travelling scaffolding: in situ (bridge on the Kettiger Hang in 1959, with equal spans of 39 m each) or precast (Rio Niteroi Bridge in 1974 with equal spans of 80 m) - incremental launching method: Rio Caroni Bridge in 1962 with equal spans of 96 m and the Taubertal Bridge in 1965 with spans of 54 and 60 m, ‘The box-girder cross-section evolved structurally from the hollow cell deck bridge or the T-beam bridge. The widening of the compression zone that be- gan as a structural requirement at the central piers was in the end extended throughout the entire length of the bridge because of the advantageous trans- verse load-carrying characteristics. The first box-girder cross-sections possessed deck slabs that cantilevered out only slightly from the box portion (see Fig. 5, a - e), With prestressed concrete, the length of the cantilever could be increased, The high form- work costs caused a reduction in the number of cells (see Fig. 5, f and g). In order to reduce the construction loads to the minimum possible or to require only one longitudinal girder in the working state even with multiple traffic-lanes, the one-celled built-up cross-section constructed in modular fashion emerged as the last development (see Fig. 5h ), This allowed the Semorile Viaduct to be built by first incrementally launching the U-shaped portion of the box girder and adding the deck slab afterwards by means of precast elements and insitu concrete infilling (see Fig. 6). Some precast elements were built in before launching to provide stiffness. In the case of the Eschachtal Bridge (Fig. 7), the closed box girder was first constructed with only short cantilever deck slabs using the balanced cantilever method with erection girder. In the follow-up stage of constructions, the main can- tilever deck portion was added in insitu concrete and supported by precast conerete struts. The Kochertal Bridge and the Liesertal Bridge were simi- Jarly constructed, In the case of the West Gate Bridge in Melbourne however, the cantilevered portion was built entirely of precast elements which were covered later with only a thin concrete topping. A further reduction in the construction loads was only possible by dividing the bridge longitudinally as done for the Hammersmith Flyover in London (Fig. 8).Historical Development § ae Qa sLOOOOO, jc at sce Lote Co ,) . Se » NOY Fig. 5 Development of the box-girder cross-section f-—s20m——4 Fig. 6 Semorile Viaduct [3] Fig. 7 Egchachtal Bridge {4] fongitudinal stressing strand Bitéer components insitu conerete joint ad ‘cantilever component pier Fig, 8 Hammersmith Flyover Fig. 9 Brotonne Cable-Stayed in London [1} Bridge with centerline cable attachment [5]8 Design The concrete box girder with streamlined cross-section has also been success- ful recently in the cage of cable-stayed bridges. Ite high dead load favorably influences the dynamic stress amplitudes of the cables and the necessary Iong- itudinal prestressing steel. A torsionally stiff box girder is required to handle the torsional moments incurred by attaching the cables to the bridge’s centerline (see Fig. 9). Attaching the cables to both sides allows a much- reduced section depth. A case in point was the Columbia River Bridge where the side cells of the approach span cross-section (Fig, 10 ) were only requi- red in the area near the cable-stayed portion of the bridge, and transverse diaphragms were added only where the cables were attached to the cross- section. A further area of application for box girders arose from noise and automobile emissions control for inner city elevated highways (Fig, 11). Fig. 10 Columbia River Bridge [6]: a) cable-stayed bridge, b) approach span glezed_ openings for natural. lighting Fig, 11 Closed box girder for inner city elevated highways [7] Even though the structural development of the box girder cross-section is prob- ably close to an end, it will continue to retain its firm place in bridge con- struction for some years to come,Design Principles 7 . OVER ALL DESIGN 4.1 Design Principles 4.1.1 The Role and Sequence of the Design Process ‘The over-all structural and architectural design is the most important and, for its designer, the most stimulating and beautiful stage in the creation of a structure, Problems during detailed analysis and design and during construc- tion as well as defects arising during its use can be traced in most cases to a faulty overall design, Design entails finding an optimum compromise among the particular objective and subjective boundary conditions of an individual structure, That is why no bridge, even a box-girder one, can be like another. Much latitude remains for creative fantasy and considerations of quality and responsibility. The design engineer must be conscious at all times of the fact that he is irrevocably changing an area’ s environment with his bridge. ‘Therefore, not only considerations of stability, serviceability, and economy count, but of equal importance are the bridge shape, the bridge’ s harmoni- zation with the surrounding landscape, and its impact on the preservation of the quality of life of man and nature, The optimum solution, always and exclusively a subjective evaluation, can only be found through the comparison of many alternative solutions with a different assessment of the individual boundary conditions in each case. The entire planning process from the sketched design to the planning of the con- struction is a cyclical process of increasing refinement (Fig. 12), It is to be emphasized that the development of the structural details should be inclu- ded in the earliest phase of the design process, whereas the computer calcu- lations range at its end and should confirm only whether the roughly deter- mined dimensions suffice. Construction is carried out according to drawings, not calculations, Above all, it would be a great mistake to believe that the computer could relieve the engineer of the design of a bridge. 4.1.2 Remarks as to form In contrast somewhat to arch, suspension, or cable-stayed bridges, the box- girder bridge fits into almost every surrounding, be it varied or monotonous [8]. Just because of its simplicity, the observer finds every imbalance in its proportions and uncleanness in its lines to be disturbing. Realizing well enough that, fortunately, it is not possible and perhaps even detrimental to put forth generally valid rules for aesthetics, a few remarks on the subject will nevertheless be hazarded here, It has served its purpose well if the fact that a separate section in this paper is devoted to architectural design and form promotes the sharpening of the consciousness among engineers with respect to thie subject, That will even be achieved in the event that the reader finds the following guidelines or orientation aids to be false and in this manner critically grapples with the questions of form (Compare especially (9, 10, 11] as well as [8, 12, 13]):8 Design ECONOMY |-material -tobour = machines. constuction method [DIMENSIONING COMPUTATION GF QUANTITIES Fig. 12 Planning: a continuous refining process - Order in the structural system: Retain the selected structural system (beam, arch, frame, suspension or cable-stayed) and only combine it with another system if the topografical boundary conditions change signi- ficantly along the length of the bridge (e.g. widened river bed or approach). - Harmony: Strive for balanced proportions among the length of the spans, the construction depth, and the depth of the valley; and between the supporting and supported structural elements, The span/depth ratio , 1/d, alone is no guaranty for a light and elegant appearance (Fig, 13), - Order in the lines: Restrict the characteristic lines to a few directions and avoid their intersecting. A straight line should merge smoothly into a curve.Design Principles 9 o) ry Fig. 13 Although 1/4, is greater than Lfa,. (b) appears lighter than (a) - Simplicity and clearness: Allow the function of the structural element and the necessity of the material to be perceived, and avoid unnecessary frills. - Integration into the environment: Either subordinate the bridge to its surroundings; or, if the landscape is monotonous or the surroundings are disorderly built up, make a feature of the bridge. - Pay attention to the order of scale between the bridge and its surroundings and the bridge and the individual person. A long bridge, especially if it passes through hilly terrain, should never be set out in a straight alignment but harmonically pick up the movement of the terrain and the lay-out of the curves of the highway. An odd number of spans which decrease in length in the direction of the abut- ments are found to be pleasing (Figs. 14 a, e, and 1). Very irregular span lengths produce a feeling of uneasiness (Fig, 14 b). Many spans of equal length produce a boring effect, and the valley appears to be walled in (Figs. 14¢ and d), although a design in the form of Figure 14d with very slender in- dividual piers can be a viable solution, The harmonic division of a line accord- ing to the "golden section" commonly used in antiquity is limited by various hindrances and the soil conditions and therefore can only seldom be realised (Fig. 149). Lo “SOT b) = = e) ~ ? ome re ea Fig. 14 On the selection of span lengths10 Design Should the bridge gradient follow a trough or a crest, i.e. the upper charac- teristic line is curved, then a satisfying appearance can be best achieved by means of a constant depth, d, of the box-girder for the entire bridge. specially for inner city elevated highways, the depth should be as small as possible and the corresponding reduction in the lengths of the spans be accepted. + a =} a Fig, 15 Shadow effect caused by Fig. 16 Fascia Beam: the cantilevered deck slab causes the bridge to appear sleeker than (a) ‘The length,a,of the cantilevered deck slab and the form given to the fascia beam and railing determine the side view of the bridge as follows: - Fora/d< 1 no shadow effect is created, but for a/d> 3 the shadow is very noticeable, and for a/d > 2 it can be usefully employed as an element of form (Fig. 15). For varying bridge depths, however, the shadow should not be allowed to extend beneath the soffit of the box girder. - A deep, if necessary white-painted, fascia beam reduces the perceived depth of the box girder (Fig. 16). Suitable values are the following [ 8, 10]: aja 1/5 to 1/4 but always < 1/3 d,/1 * 1/20 to 1/80 for long bridges d, = 200mm. An inclined fascia beam appears even lighter because of the angle at which the light falJs on it (Fig. 17), In addition, the wind is also deflected to ad- vantage, In the case of bridges near residential areas, the fascia beams can be placed higher to act as noise barriers (Fig. 18). Varying inclina- tion cause light-dark effects that give the bridge a sleeker appearance. With an appropriate design such deep fascia beams can be used as load- bearing members together with the box-girder. en Fig. 17 An inclined fascia beam Fig. 18 Fascia beam as noise barrier - Great care should be employed with the shape and structural details of the guard rail. Sufficient protection for the pedestrian plus an unhindered view for the driver of the motor vehicle are achieved with a hefty cross- beam mounted onto slender vertical standards placed at intervals of approximately 150 mm,Design Principles 11 Especial care should also be taken with the proportions of the piers with respect to the bridge superstructure. If the piers are too slender, the super- structure will appear too heavy, especially if the bridge cross-section is increased in depth from a minimum at mid-span to a maximum over the piers. ‘There the piers should never be any narrower than the box-girder bottom flange which they support, Round piers appear boring compared with prismatic ones with at least six sides, The abutments should be placed high in the embankment and be only marginally exposed. In order to assess the appearance of the bridge, it is important to compose a “picture” of it, This means making sketches at first and later drawings to scale, Along with a plan view, cross-section, and elevation, at least one per~ spective drawing is necessary in order to obtain an impression of the bridge in its surroundings. With an inclined view one can best determine if the super- structure and substructure are in harmony with each other, Of great help in this respect is the photomontage, which is really not so costly, or a model. The model must include a sufficient portion of the surroundings. Not only should one not attempt to evoke a favourable impression of the structure by choosing an unrealistic standpoint from which to observe the bridge,but one should design a bridge that is found to be harmonically proportioned even when viewed from the worst possible perspective. 4.1.3 Costs In letting the contract for a bridge one should not overlook the fact that the total costs include not only the costs of construction but also the follow-up costs of maintenance and the ensuring of proper functioning of the bridge for atleast 50 years, Unfortunately, even in the rich industrialized nations of the world, bridges are predominantly let on the basis of the lowest bid. One is seldom prepared to pay a bit more for a good architectural design and form and long-lasting quality. The growth of costs shown in Fig. 19 are valid for Central Europe, Depending upm the regional construction materials at hand, the qualifications of the wor- kers, the ratio of wages to material costs, the extent of mechanization of the construction industry, and the climate; the particular cost analysis yields dif- fering results. The course of costs over time portrayed in Fig, 20 shows that an economical solution is only achieved today if the labour-intensive work is li- mited through repetition of the same operations or construction under factory conditions. Cost optimizations with specification of the most favourable span length or the like, as is so often found in papers are questionable for the most part, because they can never take every possible parameter into consideration. ‘They often only consider material expenditure.Only if the ratio of labour costs to material costs is less than 1 do savings of material gain significance. One can observe from Fig. 19 that inflated calculations of savings of mild steel reinforcement, approximately half of which is required as minimum reinforce- ment anyway, do not change the total cost of the bridge to any great extent. Lar- ge savings are only achieved if an clement of the design is changed, such as a Jess expensive foundation or a less labour intensive construction method.12 Design 8% 50% 86% 100% ( a fa 30 “00% sorta sees | 400 -26%y SSog%e. 62% "~~100% 300] aij = (eel indations concrete + steel feeeRorr . fm Fv services ~ff J 18%. 1% concrete prestressing normal reinforcement Fig. 19 Cost breakdown:average values of Fig. 20 Growth of costs in the 17 typical examples (from [14] } industrialized nations For a bridge with average span lengths and a span/depth ratio, 1/d * 18, one can proceed from the following values for the superstructure (trom [14}): - m° concrete per m” bridge surface area, expressed as an average thickness 0,45 -1[m 3, 2. ty 0.35 SAE LLM 8 fen?y = quantity of reinforcement (mass of steel/m® concrete) normal reinforcement = 110 ‘kg/m?] prestressing steel for the case of predominantly a continuous presiressing tendons * ( 4,5+0,5-[m! lkg/m’] } - average cost of falsework and formwork ~ 60 % of the expenditure for concrete and steel. 4,2 Construction Methods The method of construction influences the design and its details in both the longitudinal and transverse direction. In the last 20 years construction me- thods have experienced a stage of rapid development, They appear today, however, to have reached the end of this stage. Box-girder bridges are con- structed today on stationary falsework (Fig. 21), if at all, only in the extreme case of a smal) number of spans or when the superstructure is not at a great height above the ground. For some of the large bridges highly mechanized constructions methods are used. By means of construction rhythms and many repetitive construction operations, these methods reduce construction time and formwork (Figs. 22 and 23). For the foreseeable future one can expect that long bridges will be construc- ted using either launching girders or segmental cantilever construction. For medium -length bridges between 200 m and 500 m the incremental launching method and classical balanced cantilever method or cantilever method with launching gantry will be predominantly used,Construction Methods 13 concreting position | formwork release position 5,5 200-4 950 —+12,00 5.15 { Fig. 21 Basic formwork for a box-girder cross-section f d) launching girder ( Krannenberg Bridge ) fost Ses a Alet9-250m! e) incremental launching method { Taubertal Bridge ) q ES Fig. 22 Construction methods14 Design CONSTRUCTION METHOD CANTILEVER dassieal Tear SFT ee [_—~seprenter PRECAST ELEMENTS| LAUNCHING GRDER INCREMENTAL LAUNCHING: ‘SPAN LENGTH BRIDGE LENGTH _] [CONSTRUCTION PROGRESS 20 40 60 go m0 U9 150-"0m]| 290 oO, KD " 19 2030 40 50 bomiaees Fig. 23. Classification according to construction method, span length, total bridge length, and construction progress (from [15}) 4.3 Superstructure 4.3.1 General ‘The box girder often is more advantageous than say aT -beam due to - its high bending stiffness combined with a low dead load, yielding a favorable ratio of dead load to live load; - its high torsional stiffness which allows freedom in the selection of both the supports and bridge alignment; and - the possibility of utilizing the space inside the box girder, Several of the following aspects apply to both the box girder and T-beam, however. The superstructure should always be designed as a complete entity. However, in order to provide a better overview, the longitudinal and transverse direc- tion are handled separately here. 4.3.2 Longitudinal Direction Because of excessive bending deformations even under constant loads and in order to avoid cracking under repeated loading, most box-girder bridges are prestressed, As short a span as 20 m is more economical and possesses more favorable load-carrying characteristics when partially prestressed than if it were simply reinforced, Under 20 m is a box-girder cross-section no longer sensible anyhow, The limit for mild steel reinforced bridges of a single span lies at approximately 35 m; for more than one span at a maximum of 60 m, Today practically all box-girder bridges are prestressed. ‘The selection of the span lengths should mainly follow from aesthetic aspects (compare with Section 4, 1,2), From the structural point of view, the span length ratio of end span to its neighboring inner span should be lp/ly * 0.75 but not < 0.40 so as to avoid uplift at the end span supports. Most end spans with Ip/ly = 0.5 do not look good,Superstructure 18 Especially important for the form and the dimensioning of the bridge is the selection of the bridge depth. Up to a span length of about 90 m, a constant bridge depth is sensible [16, 17], whereby beginning with a span length of approximately 50 m it is expedient to increase the thickness of the bottom flange over the piers on the inside of the box girder where this can not be seen from the outside. For span lengths, 1, in the middle range and constant bridge depth, d, the following ratios are normally used: - mild steel reinforced: single span Ade 17 multiple spans a= 18 - prestressed: single spans I/de 21 multiple spans fds 25 If not done earlier for aesthetic reasons, it is structurally and economically advantageous to vary the bridge depth in the longitudinal direction beginning with span lengths of about 60 m onwards, For span lengths over 150 m this cannot be avoided. According to [18], the depth, dg, over the piers should vary so that it is about 3 times as large as at mid-span, d,,. The depth of the box girder should vary in the longitudinal direction between the piers and mid-span in such a way that the forces in the tensile and compressive chords increase linearly and therefore the shear forces in the webs remain roughly constant throughout the span, Suitable ratios are: Idp = 33 to 50 I/dg ~ 12 to 20 Should the bridge be placed high above the valley, aesthetics dictate that, de- pending upon the width of the piers, the ratio dg/dp is better chosen some- what smaller than the structural analysis indicates as the optimum. Much larger slenderness ratios can be achieved in the case of suspension and cable-stayed bridges, as the depth of the main girder is not determined by the total span length but by the spacing of the hangers, which act as spring supports. For small hanger or cable intervals of from 6 m to 15 m, which also prove suitable for the free cantilever construction of cable-stayed bridges, the depth of the main girder when suspended from both sides is in principle determined only by its strength in the transverse direction, When the girder is suspen- ded from the center,the necessary torsional strength or torsional stiffness plays a decisive role as well. Skewed single celled box-girder bridges with g= 15° (Fig, 24) can be designed as if they were orthogonal bridges, For larger B the angle of skew must be considered (e, g, [ 19, 20), ey Vd Fig, 24 Skewed bridge Fig, 25 Curved bridge18 Design. For curved bridges, where in addition to longitudinal bending moments ,Mp torsional moments ,M.p,are also necessary to satisfy equilibrium, the bo: girder cross-section is especially advantageous, The angle of curvature,a, (Fig. 25) is the governing criterium for the ratio of Mp/Mg. For a < 30° it is sufficient to calculate Mg as for a straight bridge and My for the curved bridge, i.e. to neglect the coupling effect of the two upon each other (e.g. (21, 22}). 4.3.3 Transverse Direction Figure 26 portrays the factors influencing the cross-sectional form, They are all approximately of the same importance. PROTORTIONS POSSIBLE CHOSS-SF.CTIONS Caldened section Seth without aiapbre gone DbiIDER. IRISHES + vORM PORTS pier wall with rmttiple Trae bearing sreveral tnulvldial plera ingle mile pie uepended from bride Centerline Fig. 26 Influences on the design in the transverse direction ‘The use to which the cross-section is put determines its form, especially the width of the top slab. No standards exist for the necessary width of foot- bridges. That must be determined according to the expected number of pedestrians for the particular location in question [23]. In order to provide ample space for at least 2 baby carriages plus one pedestrian and to impart a feeling of safety and well-being to pedestrians, a minimum width of 3,5 m between the hand rails should be provided. Should the bridge possess light, open balustrades or carry pedestrians over major traffic arteries, the width should be chosen expecially liberally. For railway bridges, the railway clearance profile (in western Europe as in Fig. 27) determines the width of the upper flange and the distance bet- ween the webs, as the loads should act on the box girder as near to the webs as possible (Fig. 29). Should the rail traffic be transferred to the inside of the box girder , the railway clearance profile would then determine the widthFig. 27. Railway clearance Fig, 28 Double-track railway: Place the webs as nearly as possible underneath the paths of the loads. and height of the box girder as well (Fig. 29), The details of the slab carry- ing the tracks depend in addition upon whether the ballast bed is continued over the bridge or whether the sleepers are placed on vibration-absorbing pads sitting directly on the structure. For highway bridges the bridge should maintain the normal highway cross-section (e.g. Fig. 30). Should later widening of the highway be planned, the additional lanes should be constructed for the initial bridge stage and the areas not planned for initial commissioning blocked off. This avoids expen- sive reworking of the bridge in the future, eee [200] 375! 375 || 4o0|| 375 | 375 |300) w 150 7 050 @50 1 Fig. 29 Design of the Metro in Lyon Fig. 30 Typical cross-section (from [24] ) of a 4-lane Autobahn (R Q 29)18 Design The box-girder cross-section shown in Fig. 31 has proven itself with regard to its form and structural characteristics. Many variations of this "standard cross-section " are possible, Fig. 31 Typical single-cell box-girder cross-section (1/d=20;see Fig. 43) If the available depth of the girder,d, is greater than from 1/6 to 1/5 of the bridge width bya), (= roadway slab width), a single-cell box girder is in order, If d/bgzap < 1/6, a 2-cell or multiple-cell box girder is more sen- sible (Fig, 32/728). For wider bridges the vehicle loads acting on the canti- lever slab can be distributed longitudinally by means of a pronounced edge beam, enabling the cantilever length to be increased (Fig, 38). + Fig. 32 Double-cell box girder with Fig, 33 Load-distributing edge support in the middle beam ‘The number of cells should be kept as small as possible even for wider bridges with a small depth in order to minimize problems in construction, As can be seen in Fig, 34, no substantial improvement in the transverse load distribu tion is achieved with 3 cells and beyond. For economic reasons, today more than 2 cells are rare. If 2 or more box girders are placed next to each other it is advantageous to connect their top and bottom flanges in order to achieve a better transverse load distribution (Fig. 36 b). If only the top flanges are connected, they will be hightly stressed due to bending moments in the transverse direction with- out being able to effectively distribute the stresses in this direction (Fig. 35 a). it is then better to separate the two box girders.Superstructure 19 — 100 iN single lood at A = 100 kN uniformly distributed over all webs 100 kN Fig. 34. Longitudinal bending stresses ¢ at mid-span in relation to the number of cells; (example for 1/d ¥18 from [ 26]) Fig, 35 The coupling of box girders The combination of extremely wide roadway decks with slender piers can be accomplished with a box girder by supporting the cantilevers with precast struts (Fig. 36), This is also advantageous for the construction. At the inner supports of continuous long span bridges it is usually necessary to increase the thickness of the bottom flange to take the compression stresses (Fig, 37). The center of gravity is thereby lowered towards the bottom flange, Fig. 36 Kochertal Bridge at Geislingen Fig. 37 Thicker bottom flange for a 4-lane Autobahn (from (27]) over the support (Pine Valley Creek Bridge, from (28})2 Design affecting favorably the moment arm of the prestressing force. However, the increased stiffness also causes an increase in moments over the supports, The bottom flange should be gradually thickened over a distance of 1/10 time: the span on both sides of the inner support. With varying bridge depth and inclined webs, the box-girder soffit is wider in the span than over the supports (Fig. 38). In order to simplify the formwork in this case, vertical webs are therefore often selected (Fig. 39). cross-section cross-section ‘over the supports in the span 10,92m —+ Fig, 38 Felsenau Bridge, single- Fig, 39 Oléron Bridge with ver- celled box girder with tical webs and two sepa- inclined webs rate box girders Especially in the case of inner city elevated highways, exit lanes often begin on the bridge, requiring a widening of the bridge top slab. If the widening is not too great, it can be handled by extending the length of the cantilever por- tion (Fig 40a). Its soffit must however be carefully designed (Fig. 41). If the widening of the roadway is not symmetric, an equal cantilever length for both sides of the box girder can be achieved by offsetting the center line of the box girder (Fig. 40b), The torsional moments are thereby reduced as well. The formwork costs are substantially increased however. In extreme cases (Fig, 42) the distance between the webs or the number of cells can be increased, The possible position of supports thereby influences the number of webs more in most cases than the maximum cantilever length of the top slab or the maximum span of the top slab between the webs. , potatoe, he =r , Se a) symmetrical widening b) unsymmetrical widening; equal cantilevers achieved through a curved alignment of the box section Fig. 40 Widening of top slaba) simple design but too stubby in the normal region b) extensive formwork but in return a well-proportioned cantilever slab in the normal region Fig. 41 Extension of the cantilever slab eh section_o-0 =o Fig. 42 Widening for exit lanes The following: - advantageous proportions - minimum dimensions - aesthetic and structural suggestions are intended to facilitate the preliminary design but should in no way restrict the scope of the design possibilities (Fig. 43). a al. Gr ei Pr special form for the coup- ling of prestressing tendons Fig. 43 Design aids for the cross-sectional dimensions 0.2 (to simplify the formwork, a value of Tnaunch/letay * 05 is recommended) Twinn we # ty/ty = from 1:2 to1:322 Design - slab thickness t, > 1,/30, otherwise the slab must be stiffened in the compression zone by means of transverse ribs at intervals of a= 1,. (This also applies for t,). - minimum dimensions: deck slab t, 2 200 mm; t, = 200 mm webs ty = 300 mm or (200 + 2-fauoy) bottom slab t; 2 150 mm (with the utmost of care in construction, e.g. pre~ cast element, these values can be decreased by as much as 50 mm). The ratio 1):1) shown in Fig, 43 is very much dependent upon the transverse bending atifingss of the slabs and the webs as well as the maximum possible eccentricity of the vehicle loading as ‘defined by the sidewalk. For normal cross-sectional dimensions, 80 % to 90 % of the fixed-end moment of-the can- tilever slab is transmitted into the web and 20 % to 10 % into the deck slab between the webs. The optimum cantilever length 1, lies between 2.0 m and 3.5m. 1, should be so selected that no negative moments in the span occur in the deck slab between the webs, In order to minimize the width of the pier and the span length of the bottom slab between webs, the webs should be given an inclination of from 4: 1 to 3: 1 for a constant bridge depth. The transverse tensile forces in the deck slab must be accepted thereby, In order to facilitate rainwater run-off, the roadway should be transversely inclined as follows: for a straight alignment: = 2.5% for curves, depending upon the radius of curvature 2,5 % to 6 %, For narrow bridges the transverse gradient can be achieved by varying the road surfacing thickness. In general though, the deck slab should be given a gradient itself. The possibilities to achieve this are shown in Figure 44 . a) rotation of the b) a rhombic cross- ©) differing web cross-section section with vertical heights and a about the axis webs of equal height horizontal bottom through point A slab Fig, 44 Transverse gradient of the roadway (from [14]) ‘The other elements and structural details of the bridge finishes are handled in Section 11,9,Complete System and Supports 23, Diaphragms hinder the construction of box-girder bridges in most cases, They are sensible in the span only for the case of very long spans, if then. Over the piers they can take the form shown in Figs. 45 and 46, depending upon how the cross-section is supported. They can be omitted if the bridge bearings are placed directly underneath the webs, This makes it more difficult to replace a bearing however. section 141 a} most common form; __b) for a central support _¢) for very stiff with access opening box girders locally thickened webs suffice Fig. 45 Diaphragms over the piers a) small longitudinal displacement b) very large longitudinal displacement Fig, 46 Design of the diaphragm dependent upon the movement allowed by the bridge bearing 4,4 Complete System and Supports ‘The superstructure is acted upon by both horizontal and vertical loads which must be carried into the foundation by way of the bearings, piers, and abut- ments (Fig, 47). ‘Transverse horizontal forces acting on the superstructure and piers include mainly wind but also earthquake forces. In exceptional cases, impact forces from vehicles or ships and ice pressure should be considered. Longitudinal horizontal forces include forces from vehicle acceleration and braking, resistance to deformation of the supports or piers, and earthquake forces. For very long bridges, longitudinal wind forces should also be borne in mind as well as the dead-load components of bridges with longitudinal gra- dients, as long as the bridge bearings are positioned with the same inclination and not horizontally as shown in Fig, 58,2% Design np nd ssructural system tor vertical ‘oes f pier butment abutment Structural system &r transverse horizontet loads (wha! ¢ ¢ ¢ “narrow box Se pers + ¢ $ + ; sie : Set : lire ° a es : aoe, TASES Sein oe tes Shans Sm ° eH ee : ena cous 2 ni -# hen columns necessary wery = 00m of Pe bridge because of forsionat moments) Fig. 47 Typical longitudinal and transverse support schematic for a medium- length (L* 300 m) box-girder bridge with short or medium high piers Up to a bridge of medium length, the fixed point should be at one of the abut- ments, as the abutment is heavy anyway and can take the horizontal forces more economically than the piers, In this case the bridge also needs only one expansion joint (Figs. 47 and 48a). For long bridges it can be more advantage- ous to place the point of fixity somewhere near the middle of the bridge in order to halve the horizontal movements to be accomodated by the bearings on the Piers and by the two expansion joints at the bridge ends (Figs. 48b and c). This point of fixity should lie at a pier with a large vertical load in the case of short piers, For a bridge with tall piers, several piers near the middle of the bridge may be _ monolithically connected with the superstructure, yielding @ floating support. The stability of the entire system should, however, be examined. Piers constructed integrally with the superstructure are not only advantageous because the vulnerable bearing can be done away with,but also because they then possess the shortest effective Euler buckling length (Fig. 49). One should therefore take advantage of the deformability of tall piers. If the final fixed points are not located at the abutments or the bridge construc- tion is advanced from both abutments, the fixed points must be moved during the construction (Fig, 50), This requires a careful determination of the initial set-up positons of the bearings (see Section III,8 }.Complete System and Supports 25 fixed point point of fixity | a a T fis 1 T ed Fig. 48 Support possibilities in the longitudinal direction (without friction) I free W201 2071 Isl ie at I I Fig. 49 Influence of end restraint on the Fig, 50 Variation of support effective Euler buckling length 1, conditions during construction For extremely long bridges it can be more economical to provide intermediate joints in the superstructure and therefore several fixed points, as the bearings become more complex and costly the heavier the loads they are required to handle and the larger the movements they are required to accomodate simul- taneously, i.e. the higher and wider the "trees" of Fig. 51 for the particular bridge.i ’A without joint sfc pall Ee i A without joint B 2 joints rr op load carried by movable support Lo A fixed support § . monolithic connectiontraintenance free) + = superstructure dead loads temperature- induced movement Fig. 51 Load vs movement depending upon the support condition in the lon- gitudinal direction (from [29]) SCION To increase the number of joints because of an unsubstantiated fear of differential settlements to the point that the effect of continuity is lost (Fig. 52) should, however, be rf rejected. The best detailed expan- sion joint will always be much worse than the continuous superstructure, In addition, the structural reserves a hinge — rT of the continuous system are lost by the multiple-span statically deter- minate system, Fig. 52 Superfluous arrangement of jointsComplete System and Supports 27 The torsionally stiff box girder enables the. skewed abutment and pier walle to be avoided for the case that the bridge must cross over an obstruction at a skew (Fig. 53). section g-a section b-b Fig. 53 Skewed crossing with a torsionally stiff bridge constructed at a right angle Torsionally stiff box-girder bridges built on sharp horizontal curves and eup- ported by at least 3 individual columns located under the middle of the cross- section are stable even without the bridge ends being fixed (Fig. 54). This permits very transparent support systems, which are especially welcomed in the inner city, Nevertheless, in order to be able to use the normal expansion joints, the bridge ends are fixed torsionally stiff. altemative alternative Fig. 54 Support of sharply curved box-girder bridges In the case of long lightly curved bridges, the columns can be alternatingly placed off-center to the bridge axis (Fig. 55a) if the bending stiffness of one or all of the columns {s not to be utilized (Fig. $b). Of course a fixed end as shown in Fig, 53 suffices for short bridges. Fig. 55 Support system for long lightly curved box-girder bridges2B Design As curved bridges experience horizontal deformations due to prestressing and creep parallel to the bridge axis and deformations due to shrinkage and tempe- rature radially from the fixed point, bearings capable of accomodating move- ment in any direction are to be preferred over roller bearings. ‘This requirement of freedom of movement in any direction produces difficul- ties for the free bridge ends, Therefore in most cases the movements there are restricted to one direction, and the piers or abutments are designed for the unavoidable restraint stresses (Fig. 56). Fig. 56 Longitudinal movement in the direction of the bridge axis can be brought about by horizontal restraint forces (from [9]) For bridges with transverse or longitudinal gradients the bearings can be placed as shown in Figs. 57 and 58, The transverse diaphragms should be placed vertically, Beginning with gradients a greater than 3°, the gradient must be considered. Fig. 57 Placement of Fig. 58 Placement of bearings for a longitu- bearings for a dinal gradient avoiding horizontal transverse gra- forces on the piers dient (from [14])Substructure 29 4.5 Substructure The loads of the superstructure are carried into the soil by the following: - abutments ~ piers - foundations. 4.5.1 Abutments ‘The abutment provides the transition between the earth embankment and the bridge superstructure. Its wing walls secure the embankment and ita back wall holds a space free for displacements of the superstructure, The super- structure is supported here by bearings mounted on the bridge seat into which the loads are carried into the support walls and they in turn into the founda- tion and the soil. In Fig. 1, two abutment types are sketched. For small bridges the spill- through abutment is sufficient, A box abutment large enough for access inside is normally provided for large bridges. The earth fill behind the abutment should be well compacted. As the abutments even for small bridges possess large dimensions, they should be covered as much as possible by the slope or embankment, The sloped area underneath the superstructure ends should be covered with dark paving stones, as no plants will grow there, The view into the access chamber, where the bearings and any drainage lines are located, should be concealed by the end diaphragm. The end diaphragm should, how- ever, leave a gap of approximately 100 mm between it and the soffit of the box girder, 4.5.2 Piers For low bridges, especially those for elevated highways, two individual columns are more transparent and therefore more advantageous than pier walls. It can be more aesthetically pleasing to set the columns back somewhat from the outside edge of the bridge cross-section (Fig. 59b) and to forego the more efficient carrying of the loads directly from the webs into the columns (Fig, 59a).In any case, good uniform soil conditions are necessary in order to prevent differential settlement between the columns standing so close to each other, Otherwise a common foundation should be provided, on piles if necessary. Should it not be necessary to handle any torsion moments, a single column underneath the middle of the cross-section appears more ele- gant than two (Figs. 59c toe). However, this is only sensible for one or two- celled cross-sections (Fig. 594). By inclining the webs more a smaller bottom slab is thereby possible (Fig. 59e). The single columns should be de- signed sufficiently robust, as columns that are too slender do not convey @ feeling of safety underneath a large box-girder (Fig. 59), Round, elliptic, or octagonal cross-sections should be preferred above square or rectangular ones because of their better appearance (Fig. 60a). Even for extremely wide, multiple-celled cross-sections, not more than 3 individual columne together should be used, as the skewed view through them produce an unsettling effect, Massive pier walls (Fig. 60b and 61) should only be used for low bridges. By breaking up the walls or varying their width, the squat shape produced can be lightened.30 Design | a) cd e) bd) a) 0 Fig. 59 Pier arrangement : : eles Fig. 60 Cross-sectional shapes for piers (from [9] ) Fig, 61 Comparison between pier and twin columns for multiple-celled box girders For piers in a river, the top of the pier can be wider than the bottom slab of the box girder (Fig. 62a). In general, though, the top of the pier should either be set back underneath the cross-section (Fig. 62b) or set flush with the bottom sides of the bottom flange. If the pier should be protected from abra- sion, hard natural stone can be used as protection and as an addition to the bridge’ s appearance. fA IL Fig. 62 Pier shapesSubstructure 31 For high bridges over valleys the piers should also be designed with one or two-celled box cross-sections (Fig. 63). The walls should be at least 200 mm to 300 mm thick in order to permit the use of slipforms or climbing formwork, 7” 70 ol! a he ba Fig. 63 Box piers Fig. 64 Pier design The piers should be lightly tapered from the bottom upwards, at least for one dimension (Fig. 64a), Stiffening diaphragms are not normally required or desired because of thermal stresses. The top of the pier (Fig. 64b) should be designed compatible with the construction method of the superstructure. In the longitudinal direction of the bridge the width of the top of the pier depends upon the type of bearings selected, the span between the bearings, and the dimensions of the jacks necessary to replace the bearings. It varies between 1.20 m and 2,0 m, 4.5.3 Foundations The simplest type is the shallow spread foundation which is used if good soil conditions are found at a shallow depth, The foundation slab is constructed on top of an 0.1 m thick sub-base, The slab thickness should be go chosen that shear and punching reinforcement can be avoided, The bottom of the founda- tion must lie below the level of frost penetration. Should adequate soil conditions be found only at lower Jevels, an excavated pile foundation or one on drilled or driven piles should be selected. In gene- ral this type of foundation is provided with a pile cap with which the pier is connected. Depending upon the loads to be carried, the piles are either dri- ven (for loads up to © 1 000 kN) or drilled (up to ~ 9 000 kN), In the case of pile foundations for single columns, it is better to carry the forces directly into the foundation by means of large drilled piles or wells rather than for- cing them to travel a roundabout way through a pile cap on top of several small individual piles and its complicated reinforcing arrangement (Fig. 65). If unfavourable soil conditions are encountered or the foundations are into a hillside, it can be necessary to carry out open excavations down | to a suitable soil layer or to sink caissons. ‘The pier can then either be placed directly on Fig. 65 (a) is more the bottom of the excavation or on top of the favourable than (b) excavation or caisson wall.32 Design Falsework foundations should be locally widened pier foundations (Fig. 66 b) in order to prevent one-sided settlement (Fig. 66.) and the resulting dange- rous tilting of the falsework, The falsework should also be statically deter- minate in order to clearly define its structural behaviour, as its stiffness lies between only rough limits, —_ By a) Ms, bo) ‘only when A LX otherwise directly founded on rock Fig. 66 Careful attention should be paid to the falsework foundation 5, LITERATURE DEISHARDT. J.M.: Vom Cagimentum zum Spansheton, Rand 2; Bauveriag Wieubaden Berlin; 1964 HERBERG, W.: Spannbaton, Teil 2; Teobner Verlagagesellschaft Leiprig: (98? TTHUL, #.! Bntetekdungen tm Brickenbau; Weton- ur Stahlbetonbau, Heft 1, Verleg W. Emst & Soha, Bertin, 1978 MATHIVET, 1 ‘Part 2: FIP 6 Congress, London, 1878 LEONHARDT, F,. ‘The Columbia River Bridge at Pasco-Konnewick, Weohingion, USA! Proceedings, Part 2, FIP a% Congress, London: 1378 7 LRE, DJ, ‘The selection of box beam arrangements s2 bridge deaign; Developmente in bridge deaiyn and eonsteuc- ion, Helteg by ROCKEY, BANNISTER, EVANS, Croaby Lockwood [10., London, 1871 HOSHINO, K.: Gestaltung von Bichon: Witwer, Stategurt: 1072 TEONHAWDE, F._Voclenungen iber Masaivbas, Teul§, Grundlagen dew Mi tin» New York: 1970 10 LEONHARDT. F,. Aeathetee of bridge design: Journal of the Praaisessed Concrete Institute, Vol. 19. Nr. 1 February: Chieayo, iinole: 1978 11 SCHABCITENLE, K'LEONHARDT. Die Gestaltung der Brucker; Volk und Ketch Veriag. Bertin: 1997 12 POTYONDY, 4-6. Avmhatic probleme in contemporary concrete bridge design; AC! Publication SP - 23: 1869 12 TIEDIE, W.- Formprotlems fm Brdckenba, Vortrige Betontag 1965; Deutecher Beton Verein: 196) 1 MENM, ©." Brackerbaut, Vorlesungememumeipt: Ab f. Bauingoniodewesen, ETH Zarich: 079 15 FENZ, M,- Grotbelcken in Masuiebauweine, Wechaelalrkang von Koowtruktion und Bauducchfthrung: Zement leon, Halt 2, Wet 1 DEGERKOLS, O.i.: Conerote box girder bridg 1? MECMINGER, .: Dor absolute una Hefte 1 bis 3; Verlag 18 PAUSE A,, Waltgenpannte Rtceen io Betosbauwels 13 bla J.) PELLE, Rs Zur SohgittgraRerm(tiiang dee achiefen, einzellgon Hohe “Gerittendiehung der AUtedung Dnwweesn, Kenatrustiver Ingenieurbaa, Univerait® Dortewnd: 1077 20 PROFANTER, Hl’ Zur Berochnang schioler Kastentragerricken, Dauteciik, Heft 1, Verleg W. Ernet & Sohe Berlin -Manchen-Dileseidorf, 1978 21 LAPPOTH, W.. Vorlemmgenanuekript Berectaungaverfahron fur Mauetebeveken (ohIka edeken) Innttut for Maselvbac, Universita Stutigart, 1580 22° DECHENT. H.: Maseisheucken; Betonkadender: Verlag”, Kenai & Sohn, Berlin. tunchen-Di 28 NEUPERT, ©. Gasentwurfaieire, 90, Aufl Verleg F. Vieweg, Hraunechweig/Wieabaden: 1979 24 LACHOIX, H.: Métro de Lyon: La technique franealee av Pbion préconteaint ‘Congress, London: 1978 28 WITTRONT, H.” Detrachiungen sur Querachnltteauabiidung von Spannbetonbrucken; Belon- und Stahlbetorba Went 6 Variag W. Ernat & Sohn, Berlin, 1962 28 SULKE, BLM. Rerechoung denmmandiger priamet Diawertation, Universitat Stuttgart 187? 27 LANSE. A WOSSNER, Fe.” Kochertalbeicke-Fotworfe einer Grofbedcke; Bauingenteur 88, Heft 12; Springar Vor Deg, Hgelin = New York: 1978. 28 LaMUNY' JRL, PRREINS, ND.” Modern prostreated concrete highway bridge superatructurea: Grantville Publl- ‘thing Compagny Son Diego, Callforsia, 1976 20 SCHAMBECK, Ik Belcken sue Spannveton Wirklichkelten, Mbglichketten; Bauinganlaur 51, ef 8, Springer ‘Verlag Berlin - New York: 1926 vbeuekenbaue: Springer-Vacis ACT Monograph No. 1 0) 1 Cgekrtite orf: 1978 cher Peltwerke mit verfortbaramn maheselligem Queraehnit:Introduction 33 Part Il STRUCTURAL ANALYSIS (Co-author: Kurt Schiifer) INTRODUCTION ‘This part deals with the structural analysis i.e. the computation of the sectional forces or stress resultants of box-girder bridges, It is based upon the elastic theory. For statically indeterminate reinforced and prestressed concrete structures, however, the underlying assumptions of the elastic theory only agree with the real conditions for a low range of stresses. Therefore one may be of the opinion that a separation of the analysis from the dimensioning does not make much sense, Nevertheless, as explained more fully in Section III 2, this manner of structural analysis is the only method which can simultaneously con- sider all of the effects to be found in connection with box-girder bridges and their interactions with each other. Other more realistic and consistent methods of analysis and determination of the dimensiona and properties of the structure have not yet reached a satisfactory stage of development. Moreover, the method of calculating the sectional forces according to the elastic theory and thereafter of dimensioning of the critical sections for the limit state of failure satisfies the lower bound criteria of the theory of plasticity, thereby always supplying an ultimate load value that lies on the safe side, In the case of complicated static- ally indeterminate systems acted upon by high restraint stresses one should in addition to the elastic approach, always determine the real ultimate load capa- city of the entire structure by means of the theory of plasticity (Fig. Ill, 2). With these facts stated, a few critical preparatory remarks should be made concerning the structural analysis, especially the computer-aided analysis, as follows: - Garbage in---garbage out. A structural analysis is only as good as the structure’ s idealized model, the properties assumed, and the input data. - The results should always be drawn up so as to be able to visually check the variation of the sectional forces throughout the structure. - Computer results are not automatically correct results, One should therefore always check them by means of a simpler analysis, Still better is to plump for a computer analysis only after having first conducted a rough analysis by hand so as to know approximately what result to expect. - The amount of computer print-out is not proportional to the accuracy of the analysis. - Because of the ratio of labour costs to material costs, a straight-forward, clear-cut structural detailing and method of erection produce a greater effect on the total costs of the bridge than any amount of refinement in the analysis and dimensioning. The engineer with experience in the analysis of box-girder bridges will per- haps see an inconsistency in this report between the extent devoted to the cal- culation of sectional forces due to eccentric live loading and the effect that these have upon the total stresses. On the other hand, precisely and only this part of the analysis is particularly difficult for the less experienced, The literature dealing with this topic is also extensive and often not very easy to34 Structural Analysis, understand, The detailed explanation included herein should alleviate this difficulty. Formulae are derived which are not only suitable for hand cal- culation but also for small computer programs. With enough experience so gathered in this area, one can later on appropriately estimate the effect of this loading and forego the exact calculation, 2, LOADS and EXTERNAL INFLUENCES Those loads to which a bridge is subjected in the structural analysis are generally specified in the applicable code of practice for a particular country. As the design loads should not be given consideration independent of their respective safety concept and as the live loads vary greatly indeed from country to country (Fig. 1), only the general basis for the consideration of loadings is given here as follows: of Masog Mn Be anes shscoven ite" 8): normal reinforced eqnerete ¥¢ = 25 kN/m?, ete Yy = 16 to 21 KN/m ishes, go: roadway fuemeing’ (asphalt) y™ 22 kN/m? . guard railing * 0.25 - 0,35 kN/m, ~ Live loi In most codes a single-vehicle loading with a series of wheel Toads,Q;,ahd a uniformly distributed load , q; (main lane, secondary lanes), is given, Railway bridges are loaded with the appropriate loading configura- tion for each track (e.g. UIC 71 [2]). Impact is considered by increasing portions of the live loads by means of system-dependent impact factors, The machinery employed in constructing the bridge should also be included as live loads. In contrast to the assumed vehicle loadings however, these loads act with 100 % of their actual magnitude, a fact to be taken into account in the safety concept for the design, ~ Resistance to movement of the bridge bearin; coefficient of friction B specifications), horizontal loads act upon the box-girder bottom flange and the piers. 2. Depending upon the maximumLoads and External Influences 35, be L»,.1 Fig. 2 Notation used for the determination of wind load from Figure 3 without live load with live load = cya h vertical webs 2 hy cw" (h ohh without live load = with five (oad Fewash sXyewalheh) inclined webs cantilever webs bottom flange 3°96 9°12 5 buh "3 6 § 12 bib Fig. 3. Determination of wind loads (from [3])36 Structural Analysis - Wind loads: With the cross-sectional dimensions (Fig. 2) and the dynamic pressure, q, obtained from athe: wind velocity, v [m/sec], according to the formula q [kN/m? J} v?/1600, the wind loads, W and M, on the entire cross-section can be determined from Figures 3a to f and the localized wind loads, w, from Figures 3 g to i, Information concerning the maximum wind velocity can be taken from the applicable Code or obtained from a local weather station. If constant contact is maintained with the weather station during construction the design wind loads for the erection phase can be re- duced, depending upon the expected duration of the particular erection,any- where from 0.2. max W(< 1 day) to 0.7+ max W (> 3 days). ‘The critical wind velocity causing such vortex-shedding phenomena as flutter or galopping vibrations can be estimated with Figure 4 and the formula (from [4}) where fy = fundamental natural frequency of flexural ROSS SECTION ee vibration for the F405b critical shape Trio» CT Io3e Hogp [410,15-0,5b f= fundamental natural — . frequency of torsional F405 sone vibration for the critical shape 0,13 b T7718. 0,2b 10,5 - 0,36 m = mass per unit of length Se r+ radius of gyration b « Fig. 4 Factor a derived from wind = bridge width tunnel tests = factor from Pig. 4 derived from wind tunnel tests A more exact analysis requires wind tunnel tests in most cases, Bridges of normal dimensions continuous over fixed supports are generally not critical, Suspension or cable-stayed box girders could be, however. - Braking and acceleration forces, Fy: These forces act horizontally on the bridge bearings at the tixed points of the bridge and are degressively dependent upon the entire vehicle loading; i.e, the longer and wider the bridge the smaller is the chance that all of the vehicles simultaneously accelerate or brake, (According to the German Code, Fy = 3/20- Lb [KN] (L #200 m), roadway but Fy is at least = 0.3 - Q, heavy vehicle’) - Temperature: A uniformly distributed heating or cooling of the super- structure from its mean temperature at the time of construction only affects the movements to be accommodated by the bearings and possibly the sub- Structure, The magnitude of such temperature changes ean be takenLoads and External Influences 37 from the climatic conditions of the region. In central western Europe, for example, the temperature varies from + 20 K to T - 30K from a construction temperature B of + 10°C (283 K), Heat of hydration co (Fig. 5) and shrinkage in thick parts of SES the cross-section; especially between 4 thick and thin parts of the structure an 1 imperotre Of etween young and old concrete, can give inthe bex girder rise to large localized residual stresses which exceed the tension strength of young Fig. 5 Large temperature gradient due to the concrete, cement’s unfavourable Ag a good crack distribution in young heat of hydration concrete cannot be achieved, even with development and low a finely distributed reinforcement, erack- temperature of previous- ing can only be reduced or prevented by ly concreted section cooling the aggregates and/or the water {Exons [50 used or by insulating or heating the forms, The insulation of the concrete should be maintained in position from 2 to 4 days, depending upon the thickness of the structural component and particularly for construction joints. The temperature gradient between the top and bottom of the roadway slab as well as that between the top and bottom of the box girder produce deformations. When the system is statically indeterminate, restraint stresses are also pro- duced, During the course of a day, the temperature in the inside of the box girder remains within about 2 K because of the absence of air circulation, whereas the roadway slab, particularly underneath its black surfacing, can heat up very significantly, A linear temperature gradient can be assumed for the top and bottom slabs between the webs. The temperature gradient in Figure 6 comes the closest to describing the actual gradient in the cantilever portion of the top flange and the webs. (See also [8] and [9] for more detailed information). As such a nonlinear temperature distribution is not very suitable for a practical analysis, a linear one is often chosen. For central western Europe AT = 10 to 15 K appears appropriate. ance rae | _ web catlene, top sto 1700 y ad between the webs 5 L sa Thedon, ty25-0005h °C x food! Uterperate toad! “Tremperuture” Fig. 7 Magnitude of the restraint moments due to temperature, Fig. 6 Temperature gradients used AM, in relation to the size of in dimensioning from [7] the load acting at the same time38 Structural Analysis The magnitude of the restraint sectional forces is dependent upon the stiffness distribution, Those sectional forces determined for State I are maximum values which can be sharply reduced in State Il. That is why an additional restraint moment acting simultaneously with a large loading is much smaller than if it acted together with a smaller loading (Fig, 7), Alternatively, a reduced AT (e.g. only 5K according to DIN 4227) is occasionally used in the calculation of the State I condition, It is, however, strongly pointed out here that it does not make sense to calculate the temperature stresses with the elastic theory so exactly or to attempt to handle such stresses by increasing the prestress force and thereby forcing the structure to remain in its State I condition, It is more sensible to employ partial prestressing in combination with sufficient mild steel reinforcement and allow the reduction of restraint sectional forces to occur. Through the mild steel reinforcement one should ensure that the cracks are finely distributed. (See Part III.) Solar radiation on a web can also produce large horizontal deformations (Fig.8) which are either accomodated by the bridge bearings or taken by the substruc- ture in the form of restraint stresses. PLAN xt w sole 4 ; 2 kes +AT B _———— Raat 2 ue We aR Fig. 8 Horizontal movements - Differential settlements, 4s: Should one analyze and dimension a box- one would obtain meaningless results { 10]. As with the case of temperature, with poor foundation conditions one must design for more realistic State IL conditions, This is allowed with partial prestressing. Further, the fact that settlement occurs over a period of time and that the restraint stresses are reduced by creep over time should not be forgotten (see Section 5,3, 2). If necessary, the bridge bearings should be able to be suitably adjusted. It is false, though, to select statically determinate single spans in order to avoid these restraint stresses, ag the increased number of joints means a greater possibility of trouble with the durability of the bridge. Only for very small and rigid bridges does it make sense to 6 effect of an earthquake into account by means of a horizontal equi- valent force that is proportional to the bridge mass. For large bridges a realistic dynamic analysis is required, (See for example [11] to [16] .) - Impact: Instead of dimensioning the piers for the very high impact loads fFom vehicles (~ 1 000 kN concentrated load) or even ships, it is better to provide them with energy-absorbing barriers or fenders. See Section 7, See Section 8.Structural System 39 3, STRUCTURAL SYSTEM 3.1 Final State The box-girder superstructure can be part of the most varied of systems (Fig. I, 2). Here the standard case of a directly supported box-girder bridge continuous over the columns or piers is considered, The superstructure and the piers form a framework in case they are monolithically connected (see Fig. I, 48), The stiffness of the box girder, Ely, is generally much larger than that of the pier, Bip. When Elg/1 # 30 EI,/1p the superstructure can be analyzed separately from the substructure as a continuous beam on freely rotating supports (Fig. 9). t | — For the structural analysis of the piers, tL, EL however, the movements of the supports e 8 e 8 and rotations of the superstructure must Elp be considered, even if the superstructure is hinged or free to move at the junction Fig. 9 Stiffnesses of the frame with the piers (see Section 8), members ‘The type of support also affects the structural analysis. A beam with more than one span with inner supports along its centerline behaves as a continuous beam in bending and shear but as a single-span beam in torsion. Should the box girder be reduced to a single-span beam between the abutments for a simplified analysis of the system subjected to horizontal wind loads and not analyzed more exactly as a continuous beam on elastic supports (Fig. I, 47), then the piers should most definitely be investigated for stresses in the trans- verse direction, 3.2 During Construction The structural system is changed many times over during construction, depend- ing upon the particular construction method employed. For example with the cantilever method (Fig. 10), the statically determinate cantilever (Fig. 10a), which can be additionally supported by cables at critical loading stages (Fig. 10 b), becomes a one-column or multiple column framework upon the closing oN © Mone | ay fd a) >) “TT “TIT Fig. 10 The transformation of the structural system with the progress of cantilever construction40 Structural Analysis: of the span in the middie (Fig. 10 c) or a continuous beam with the appropriate support or pier stiffness, respectively. Temporary supports, cable truss arrangements, or auxiliary cable hangers can in special cases increase the system diversity during construction even further (Fig, 11), Time-dependent influences (shrinkage and creep) should not be forgotten here, A variation of support conditions during construction can also alter the structural system (Fig. 1, 50). Further, post-tensioned tendons are often still unbonded in the construction phase, which reduces the ultimate load and makes the system internally statically indeterminate, “7 f[ -)TT fT 0) Cantilever method with erection gantry Oe ee b] Incremental {aunching method Fig. 11 The number of spans, and if the occasion arises, the stiffness, varies with the progress of construction 4, CRITICAL LOADINGS AND SECTIONAL FORCES In order to determine the critical sectional forces or sectional force combi- nations, all of the loads listed in Section 2 in their possible combinations in the final state and during construction should be considered, Except for the vehicle loadings on the superstructure, this requires no further explanation. 4,1 Longitudinal Direction ‘Those sectional forces necessary for the dimensioning of the bridge in the longitudinal direction (Fig. 12a) should be obtained from the influence lines (Fig, 12) at the 1/10 points and at other critical positions (Fig. 12b) for vehicle loads, For the calculation of the critical sectional bending moments, M,, and their associated V,, and My; in view of the flexural dimensioning, the bridge is loaded over its éntire widih but with the main traffic lane at the side. The most unfavourable position of the main traffic lane with respect to the folded plate action can be handled as explained in Section 5,5. The di- mensioning for shear, V,, and torsion,M,p, should be carried out both for the maximum shear, with the associated M, and Mp for the section, and for the maximum torsional moment due to a dne-sided loading with its associated M, and V,.Critical Loadings and Sectional Forces 41 min My oy my, a2 d2d x Se max Vg YM, max Me ey CeMp Ce My 4 + " em, ey, a) Longitudinal sectional b) The minimum of sectional forces forces required for the dimensioning ELE toad position for the ~ lee min lone ~ B targest moment, My. RR ja the non heovy lorry cocorett ere longitudinal loading (See Fig 15) SL" influence line for max My YAN ee KEELES et ee line for the largest shear force Vy. ee influence line for the associated moment SST paz] load position for the ~*D— largest torsional moment Mr 4 pa tea line for My _ LIL cernezzeeczel _ - __load position for the AZOLE torsional mom IRE largest mement Mr for torsionally rigid conditions at the columns influence line for My for central support in # 1 i iol c) Influence lines and loading positions Fig, 12 Determination of sectional forces for vehicle loadings42 Structural Analysis: 4.2 Transverse Direction In general it is sufficient to determine the critica? sectional forces for the dimensioning of the flanges (Fig, 13) from the loading positions of Figure 14, This holds true in spite of the fact that due to the folded plate action(Section 5.5) the zero lines of the influence surfaces do not lie in the longitudinal axis of the bridge and not over the webs or even parallel to them, My ™: t t 12 36 5 MH May | = adh dk Tey TS m7yi07 8 i 1 MyM Moy Mygy ON Mp, when the Ng bottom flange 1s loaded 4g b) SECTION a) PLAN | Fig. 13 Sectional forces required for the dimensioning of the flanges max me , min mM min mg , Max Ag, min My MAX Me , min Ne ,mMaX @ M5 max Og. 'S 5 10 or —S SSS sHosonhh min / ‘min a, min mg , max my, Fig. 14 Critical loadings, in general 5, SIMPLIFIED STRUCTURAL ANALYSIS OF THE SUPERSTRUCTURE 1 General The box girder represents a special form of a folded plate structure, The loads are always carried three dimensionally, The analysis of the sectional forces to which the structure is subjected can be handled with the aid of the following: (see Section 6), ~ analogy of a beam on an elastic foundation ~ folded plate theory with the use of series expansions - finite strip method ~ finite element method‘Simplified Structural Analysis of the Superstructure 43 As the Jast three approaches produce voluminous data and therefore are great- ly susceptible to error in the evaluation of their results, even today more reliable and clearer and still sufficiently accurate analyses for box-girder bridges are to be expected from the first approach as described in Sections 5.2 to 5.5. With this approach, the box girder loaded symmetrically in the trans- verse direction can be analyzed longitudinally as a beam and transversely as a frame independent of each other, This is valid if the box girder’s slender- ness ratio, 1/d® 4, if the span lengths, 1% 1,5 - b (Fig, 15), and if the disturbance zones at the introduction of the applied forces are separately handled. ‘The particular problematic nature of the simple analysis approach to the behaviour of a box girder under load arises with unsymmetrical loading in the transverse direction. In this case the longitudinal and transverse stresses are connected with each other, a connection which shall be designated in what follows here as the folded plate action. This connection is taken account of through the formulation of the condition of compatibility between the separately analyzed longitudinal and transverse directions. Warp resisting supports and diaphragms or discontinuous eccentric loading produce warping torsion along with St. Vénant torsion, The additional stresses due to warping torsion die away rapidly or are small in comparison with other stresses, however, and therefore can be neglected according to [17]. ‘The analysis procedure suggested here is so selected that the results are increasingly refined the further one proceeds. If desired, one can break off the analysis without considering the folded plate action. One can aiso include that action sufficiently accurately for a critical loading case and extend the results to the other loading cases by means of a uniform factor. Lastly, one can systematically follow the method through for every case. 5.2 Analysis Procedure The analysis presupposes that the statical system, the loading, and the design cross-sectional dimensions are known, el ha & 2 est Te cal eg om LONGITUDINAL LOADING. SECTION 2-2 SECTION 1-1 SECTION 2-2 a) The loading of the top flange so as to _b) The loading of the top flange so as produce the largest torsional moment to produce the largest moment, M, or the largest shear,V, respectively Fig. 15 Positioning of vehicle loading in the longitudinal direction44 Structural Analysis First the box girder is analyzed in the longitudinal direction asa beam, i.e. with the assumption of a rigid cross-section (Section 5.3). All transverse loadings are integrated together to form point loads or line loads (Fig. 15), With these loads the longitudinal bending moments and shears are determined, and the accompanying or least favourable torsional moments of the entire cross-section are determined with the eccentricities, yp, of the load resultant, a... Afterwards the box girder is analyzed in the transverse direction with the actual load distribution (Section 5.4), whereby the top flange is first assumed to be rigidly fixed at the webs (Fig, 16a 1). q is b= —4 "6 a —> Robot 01) Top flonge a2) Only edge loads rigidly fixed mR RAH RAW mR FMT MR TMH MB bil symmetrical part 2) Antisymmetrical part of load f, load longitudinal bending: beam transverse bending: frome aM. {mobule Sor. oe A odes My Mas Sr = Tacs tS atom, ei) re tersion c2 aslo ay by See Fig. 24 no resultant TAK sectiond forces oF ™ "8. = * di} Stiffened frame: d2) Folded piate action: only transverse bending longitudinal and transverse bending Fig. 16 Splitting up of the structural system and the loading of a box girder, All arrows drawn represent external loads. The cantilevers are partially omitted for greater clarity.‘Simpiified Structurat Analysis of the Superstructure 45, The resultant reactions, qq and qp, and the fixed-end moments, mg and mp, act at the edges at A or B, respectively (Fig, 162), They are divided into a symmetrical part (Fig. 16b 1) and an antisymmetrical part (Fig. 16b 2), A strip of length dx = 1 is taken from the bridge and analyzed as a frame sub- jected to the symmetrical part of the load (Fig. 17a). The strip is supported in this analysis by the shear force differences from flexure in the longitudinal direction (Fig. 24), Adding the transverse bending moments calculated in this manner to the symmetrical part of the moments (Fig, 16a 1) yields the final transverse moments due to symmetrical loading. The torsional part (Fig. 16c1) of the antisymmetrical load effect (Fig. 16 b2), which was taken into account in the longitudinal analysia, having been split off from the rest, a state of equilibrium with the applied loading as shown in Fig. 16 c2 is left in every cross-section.Were one to investigate this state in the same manner as the symmetrical loading by analy sing a transverse strip, one would obtain results far removed from reality, for the idealized frame with edges free to displace(Fig. 17b) does not consider the carrying of the loads longitudinal- ly in the plane of the plate. Nevertheless, neglecting the carrying of the anti- symmetrical loads in the plane of the plate presents a clear limiting case. Another limiting case is provided by the analysis of a frame strip whose edges are held fixed by the diagonal force, S (Fig. 16d 1), The resulting transverse bending moments, after their superposition with those from Figures 16a 1 and b 1, generally lie on the safe side in comparison with the actual transverse bending moments. In the analysis of the folded plate action,which follows in Section 5.5, the longitudinal beam and transverse frame action are connected together for the antisymmetrical loading part.This is achieved by no longer forcing the shape of the cross-section to be maintained by means of the diagonal force, S, (Fig. 1642). The longitudinal and transverse sectional forces due to this force, S, are obtained from the analysis of the elastically bedded web panels, The super- Position of these sectional forces due to S with those from Section 5.3 or 5.4, respectively, yields the total sectional forces. ‘This satisfaction of the conditions of compatibility increases the longitudinal stresses and deformations of the web immediate to the loading and decreases them for the web away from the loading, The frame deformations and moments lie between the two limit cases mentioned above. They lie nearer to the limiting cage of a freely distorting frame, if the loading changes little in the longitudinal direction and the transverse stiffeners are far removed (see Sec- tion 5.5.2). On the other hand, for the case of the heavily loaded vehicle or in the vicinity of a transverse diaphragm, the limiting case of a fully rigid frame is closer to reality, Formally the analysis takes the following steps: - determination of the loading in the longitudinal and transverse direction, respectively (Figs, 15 and 21) ~ analysis in the longitudinal direction as for a beam with a rigid cross- section (Section 5.3; also Fig, 12), yielding M,, V,, My and o,46 Structural Analysis - division of the eccentric fixed-end reactions into symmetrical and anti- symmetrical parts (Figs, 16b 1 and b2) - frame analysis in the transverse direction for the symmetrical edge loads (Section 5.4; Fig, 27), yielding (m,n, V)gymm - rigid frame analysis of the antisymmetrical edge loads (Fig, 28; Section 5.4), yielding (m, n, Vantisymi. as well as S - superposition of the sectional forces in the transverse direction, yielding m, n, and v - analysis of the elastically bedded web panel loaded with q, due to the diagonal force, S, (Section 5.5) yielding longitudinal stresses, 0, and transverse bending moments, mg, (Figs. 33 and 28e, respectively) - superposition of Of, with o, and my, with m, yielding the actual longitudinal and transverse stresses, réspectively. deformations de to —Apeam anclysis Frame analysis yields only tranverse bending. The edges do not displace with respect to their positions from the beam analysis delormations due 0 _-—beam onolysis _etrome analysis # “unrestrained edges) Frame analysis yields edge displacements to- gether with the trans- verse bending (shown for unrestrained edges) a) Symmetric loading b) Eccentric loading (torsional deformation not shown) Fig. 17 Connected beam and frame action‘Simplified Steuctural Analysis of the Superstructure 47 Analysis in the Longitudinal Direction 5.3.1 Sectional Forces due to Loads plus Restraints For the vehicle loads as prescribed by the German code DIN 1072 (Fig. 1) and medium span lengths of approximately 40 m the total sectional forces not con- sidering prestress and restraint are produced to about 5% by the concentated vehicle loads (characteristic 600kN heavy lorry),to about 40% by uniform vehicle loads, and to about 55 % by dead loads. With increasing span lengths the in- fluence of the vehicle loads and thereby the necessity of their exact analysis diminishes. The sectional forces of the beam (bending moments, shears, normal forces and torsion) due to external loads and imposed deformations (as described in Section 2) can be obtained with known methods of the theory of structures. Normally a small to medium-sized computer suffices. Should areas of the box girder be normally in the cracked State II, even under working loads the sectional forces determined using the elastic theory can be redistributed and the sectional restraint forces thereby reduced. This redistribution of forces is dependent upon many parameters and can only be obtained iteratively, whereby superposition is no longer possible. The extent of reduction of the sectional restraint forces depends upon the magnitude of the sectional forces due to external loads that act upon the section at the same time(see Fig. 7). With sufficient rotational capacity, the restraint has no influence on the ultimate load-carrying capacity (see Sections | and III, 2), 5,3,2 Time-Dependent Alteration of the Sectional Forces by Creep Sectional forces due to external loads are only altered by creep if components acting together within the structure contain concrete of varying ages or varying composition, This is indeed the case for segmental construction (Fig. 18). For a sufficiently exact calculation of the final time-dependent moments, the follow- ing equation taken from [18] can be used: (see also [19] and [20]}) ; Mat ZMyy +e -E Mos) Se cppatyuetion 2 = Stage as —_—— > _ where Ms 1,2, 3, moments at the time of stripping of the falsework of the differ- ent segments, i.e. moments due to self- weight plus prestress moments assuming that the entire bridge is constructed at the same point in time creep factor at timeso relaxation factor taken from Figure 19 ¢ Fig. 18 Development of the bending moments from the construction stages48° Structural Analysis: If the time of loading, ty, i.e. the time from the concreting of the section until it is prestreased, is less than 10 days and the construction rythm used in producing the individual segments, At, is less than 21 days, then the values produced by this approximation do not vary from the exact values by more than 15 %, In consideration of the unavoidable inaccuracies of the analysis assumptions and the fact that the results do not influence the safety of the structure, this is perfectly sufficient for the normal case. Should t, or At, respectively, deviate grossly from the above values, the limiting values of the moment redistribution at time t = 2 can be estimated from Tk = 1s =# i 7 rege, The values for ky*K(t) or k, and P, for t,= totAt for the youngest or oldest construction segment, respectively, are taken from Figure 19 with the standard creep factor,my, being taken at time t)* 28 days. k, =klt,lor kiekltget tI) or F(t) 20h 8 ) sist 1.6: Of 12 S10 | 08) BY 0.8; 04 on - a days 0 0 4 ae 137 m9 30 | 37u8 300 1000 | cage at time of kading tear tet tor tty 1803 2880380" tye d tne of tadng begin a) b) Wor tel eto oS Ere Fig. 19 Factors for determining the effect of creep in the analysis (from(18}) As long as individual construction segments are not prestressed very early or to extremely high values of prestress, the changes in the sectional forces due to prestresa as a result of shrinkage and creep can be calculated with sufficient accuracy as a uniform reduction in prestress,AP., .(t), up to time t: © p, baltitedEs +X Plt tog + Spo) ARasltl = RE ipo [1 # PlEL V2] where; @ = E/E, and ey (t, t,) ... strain due to shrinkage within the time span (t - t.)- The values are dependent upon the humidity of the air, ete, and lie between - 0,1 % and - 0,5 % (see [66}) EE sr Ee modulus of elasticity for prestressing steel and concrete, respectivelySimplified Structural Anal is of the Superstructure 49 Plt, t,) ..- creep factor (the values are dependent upon the time 2 span (t-t,), the humidity of the air, etc, and lie bet- ween 1 and 4, see [66]} oe stress in the concrete at the position of the tendon due oF to self-weight and other dead loads (compression negative) po « initial stress in the concrete at the position of the tendon due to prestress alone (compression negative) oe initial stress in the tendon due to prestress alone Sectional restraint forces due to differential settlement, temperature, and shrinkage are greatly altered by creep even without cracking. Suddenly occurring restraint stresses (e.g. due to differential settlement in sandy soils) are great- ly reduced (Fig, 20), According to [21] , the elastically calculated moment Mg) in State I is reduced to ko: Yu F(t) Mit) = My: (4~ RP ) M(t) For restraint stresses that occur a gradually at a rate approaching m T that of creep (e, g. settlement in for p= 0,85 clays, shrinkage) the moment that is produced instead of Mgy is welt) = Mey zee fll NU f+ Pk By FED ‘That is why it is false to reverse the effecta of a pier settlement that + © O 100 == ¢ has occurred slowly by suddenly adjusting the bridge support to the Fig. 20 Creep-induced change of ‘all magnitude of the settlement. restraint moments result- ing from support settle- ment 5. Analysis in the Transverse Direction For the vehicle loads as prescribed by the German code DIN 1072 (Fig. 1), the sectional forces in the transverse direction are a result of an approximately 10 % contribution from the uniform vehicle loadings of the main and secondary lanes, an approximately 70 % contribution from the concentrated vehicle loads of the characteristic 600 kN heavy lorry, and an approximately 20 % contribu- tion from dead loads, 5.4.1 Sectional Forces Acting on the Flanges For the practical analysis, the clearest method is first to analyse all flange sections as if they were rigidly fixed at their junctures with the webs (Fig. 16 al)50 Structural Analysis and then to redistribute the total moments acting at these junctures according to the frame's stiffness (see Section 5.4.2). While the sectional forces due to dead loads, prestress, temperature, and uniform vehicle loadings over large areas can be determined as acting on flange strips (or also acting directly on the frame), influence surfaces must generally first be evaluated for concentrated vehicle loads (Fig. 21¢). Wheel Joads can be distributed under 45° to the middle plane of the flange (Fig. 21a) and the uniform loads can be integrated sufficiently accurately to concentrated loads as shown in Figure 21b. In[23] the influence surfaces are supplied in table form for the majority of sectional forces, whereby the limiting moment diagram can be estimated as shown in Figure 22. The transverse distribution of the concentrated loads on the cantilever portions of the flange depends largely upon their structural details (Fig. 23). The moments occurring in the top slab due to a temperature gradient, AT, between the flange surface and its underside are constant between the two fixed ends and are given by the following: mam, =p OT Et2/ 12 (tor state 1) rocdway wearing surface Pee d 0,29 bi | 0,29: by of ‘ Oe fg hb} rep d-t-a 2! * ~0.29-b5 |*0.29-b5 Pi a) Wheel load distribution a: b) Position of the distributed qecbiren, load resultants (from [22]) wo 08 06 oe oz 902 gs 98 U8 yo 2 Me T —t Lance WaT FED OS 32 gdp 12 my 7 CE, c) An Influence surface (example from [ 23] ) and its evaluation Fig. 21 Determination of the sectional forces due to vehicle loadings from influence surfaces‘Simplified Structural Analysis of the Superstructure 61 pee ey ‘a ‘y.miing . 0 »h ds Re - rs be a eg : SiaEae - O21“ OS-1 0,251 0,254 zy Mx ymax Tx, max om 1" Gt bat tie i, cIM-limit envelopes for forconcentrated lads concentrated loads acting on flange strips acting on flange stops in the trans- verse dophragmn re gion; 3-sided support a} M- limit envelopes for concentrated loads —_b) M-limit envelopes: acting on cantilevered portions of the flange Fig. 22 Moment envelopes (from [24]) »y a o k= Pe @ af aad Wrae ° ® = a ° ® 4 Sas tonge bg 5 vi Le Cf Vorious cantilever designs ‘The area under the curve yields Ob, every cose m4 10 209% ©) Guolitotive diagram of the moments m,, at y- 0 Fig. 23. Influence of the cantilever design on the moment diagram (from [25] and [ 263)52 Structu 5.4.2 A ral Analysis nalysis as a Frame Every arbitrary loading can be separated into a symmetrical and an antisymmetrical part, On the element of length dx = 1 taken from the structure, the shear force differences, dT’, (and the longitudinal stress differences, do_, which are of no further interest here) maintain the external loading, q(x), in equi- librium (Fig, 24a), For symmetrical loads (Fig, 24b), where = * . Ve Oi f by Frome wih Symmes ores. equim Ga= Qixrdx 5 T= vt = VS/Ts % j . at ___ the following holds true: ‘eal a 7 ar Sav SL. 7 teeta dx i dx i 9" cl Frone with entsyimireat ces n eqilbmar where with dx = 1 1S, aT’ F'Op Symmetrical moments at the juncture of the flange with the box, mp/2, (Pig, 16b1) produce no shear forces, Fig. 24 Frame analysis For antisymmetricat loads as in Figure 24c, the following holds true 2 alr. Gab. +2, ats FA A where 5S... first moment of area of the partial half-cross-section about the centroidal x-axis I... moment of inertia of the entire cross-section A+. kern area Ad: (byt b,)/2 (Fig. 25) The sectional forces acting on the frame must be calculated for the simultaneous application of the external loads and dT" atx! 6 moments of inertia longitudinal transverse Tpte(bs+2b,)/M Betz Lt baz e812 Tete Ty=tubi/t2 (continued next side ) bal Fig, 25 Notation for the longitudinal and transverse directionSimplified Structural Analysis of the Superstructura 83 auxiliary values tb’ Xo tbe > is Os ot 1 f= pe 3 a bo elu, bbe . a/q?s (eat) Oe 5 EB 3 gafd*+(Pe%) 7 S12 2x, =| x3 b) Fig, 26 Redundant forces (a) for symmetry, (b) for antisymmetry Symmetrical Loading For symmetrical loadings the frame is statically indeterminate to the 294 degree (Fig, 26a), because no shear is possible in the plane of symmetry. Figures 27a and b portray the moment and internal axial force diagrams resulting from the loads and moments, respectively, acting at the rigidly fixed junctures of the box with the flange, for such a cross-section as shown in Figure 25. In addition, Figure 27 portrays the moment and axial force diagrama resulting from an increase in the length of the top flange,‘ 1g, over the bottom flange. This allows the possibility of considering, for example, a uniform increase in temperature of the top flange (T,) over the bottom flange (Ty) through Al2(T- Th) Ob, and the shortening of the top flange due to transverse prestress, P, Pb alee Ee It should be mentioned here that Al should not be multiplied with a creep factor for the statically indeterminate analysis as long as the creep conditions are approximately the same throughout the entire cross-section. Due to the effect of the forces resulting from the curvature of the longitudinal prestressing tendons (see [27]) upon the webs, shortening |Al,| of the webs occurs, which with the value 41, = 2|41,| - sin@ , can be converted to the case of Figure 27¢, This is only necessary to be considered if the radii of curvature of the tendons in the webs are small and the angle, ©, is large. Figure 27 also provides the sectional forces from moments at the rigidly fixed junctures, m, /2, acting on the joints of the lower flange with the webs such as those moments arising from self-weight of the lower flange and webs, One need only turn the cross-section upside down and change the subscripts in the formulae from denoting the top flange to denoting the bottom flange.54 Structural Analysis: LOADING TRANSVERSE MOMENTS m AXIAL FORCES me ™R " 4 2 pe f oe WT TT mm ego Blea a, « 220! led 07 ey Veg aby euxiliory value hy 116242 +3i—ty Formulae ter load ease a} 4 mya met Bong roc 8 Og + Mage pl ta? no + ph ty b2lo-2, seat ts, my + ge tubs Tenn! . 8 fe 22 4 Ms MOTT (rie Same} aD eS tac TL Os Fig, 27. Effect of the symmetrical part of the loading on a frame: a)Nodal Joads b)Nodal moments c)Temperature-induced increase in length of top flange or prestress-induced shortening Varying thicknesses of the flanges and webs can be approximated with constant thicknesses, providing the same cross-sectional areas, In extraordinary cases or when the separation of moments into those moments acting at rigidly fixed junctures and those on a frame is to be avoided, one can of course carry out a thorough frame analysis by the usual methods. The shear and normal force deformations can thereby be neglected, with the exception of the 6;, values from prestressing. A For antisymmetrical loading the moments and normal forces are zero in the plane of symmetry. The sectional forces acting on the transversely stiffened frame (Figs. 28a and b) can be determined on a system that is statically indeterminate to the 294 degree if the symmetry of the system is maintained by means of the antisymmetrical diagonal forces S/2 as shown in Fig. 26b . The sectional forces acting on the freely deformable frame, which is statically indeterminate to the first degree, are given in Figures 28 c and d for comparison. They provide a limiting case, which may be of interest now and then (see Section 5,2), If the folded plate action is considered as described in Section 5.5, the sectional force diagrams due to the diagonal force, 5, of Figure 28 e are needed,uonoe ae1d papror oy anp g seosog(o ‘sepou aiqeurosap f19043 YIM aes] B UO SHOWOUL TepON(P pUE SpROT TepON(> ‘sepou paxt} ATPra uitm eurery 2 Uo syuaWOW! TePON(A PUe sped] [epoN(e ssureds a4} uO BuTpLoT ayy Jo 74xed [woTzjoWUSTUE ay; jo 19ETIT Bz “Bla S'S voN.s 885 § 55 se on, 4 2 S84 gies, J+ Se deegarz= & Fane 5p 2 Zz non rE’ be (eeMy (zee Mg Srnp rT ] fA | FA « ¥, ws ideend- 2d ee ue MW Ba ‘Simplified Structural Analysis of the Superstructure Mu gas My GueMitisos OL = Wy Bue Pisesiz way. aE oy. OH Pideur . ov, Yb (Pae8a1 e ay,, ray My ges = oe fe ous™npso gt ogas bs MW] Kas gegen s a] Bue On sR, ev, 1 S $3940s IYNOSVIC W SLNSHON 3SHAASNVML ONIGVOT56 Structural Analysis: 5.4,3 Transverse Bending Moments in the Haunch of Variable-Depth Girders In the bottom flange of haunched box- —t girder bridges (Fig. 29), forces due to ov! a Dy! Th + curvature, q=o,: t,/R, directed towards \ 4 by F the inside of the haunch arise in the zone a of negative moments, These forces create Dil compre transverse bending moments which act force per against the moments due to self-weight of unit the flange. According to [28], the stresses, 0,, are not uniformly distributed across the box-girder width, b,, when the haunch is sharply curved but are concen- trated in the areas near the webs. The determination of the transverse bending ty Y ay ot moments therefore employs a mean stress bu i | Gym» instead of o,. Using Figure 30, the r following is then obtained: 1 -B} Dy) Fig, 29 Loading of the bottom ~ effective width, of the bottom flange flange of the haunch b= u “Y ui 2 4BY - 1, is the moment of inertia and 2, 1 the distance between the neutral axis of % 7 Pui Hi Bui the section and the mid-depth of the 2byj= Vb, flange, both for the idealized crogs- sectional values with by; - mean stress at mid-depth of the flange : z Ss Om 7 My alg (State 1) - transverse bending moments at the edge and in the middle respectively, due to non uniform distribution of q: be mt hn 10 1 ew 02 04 «06 «= 8 pF Ny bt 4am SP Oke Fig. 30 Factors®, y, ap, a as taken from ge (28] also handjes the influence on the moments of elastic fixed-end conditions in case of deep or thin webs as well as the influence of transverse haunches at the bottom flange.Simplified Structural Analysis of the Superstructure 57 5.5 Folded Plate Action Among the various methods of analysis (see in [291] and [30]), the one employ- ing the analogy of a beam on an elastic foundation is the most suitable for the actual practice, This method is therefore summarily portrayed here as taken from [17, 25, 32] and expanded by some own developments. 5.5.1 Fundamental Concept The analysis is based upon the following assumptions: - The lengths of the plates are large in comparison to their widths (1/b> 4) - The plate thicknesses are small in comparison to their widths - The box girder is symmetric about its vertical axis. A constant cross- section is maintained throughout the length of the girder (Fig. 25), - The thicknesses of the flanges and webs remain constant in the transverse direction, (The influence of haunches can be neglected according to (31]) - The principle of superposition is valid here Having handled all symmetrical loading effects plus the antisymmetrical ones as far as torsion and transverse bending on the diagonally stiffened frame are concerned in Section 5, only the partial loading of Figure 16d 2 remains to be investigated. The imaginary diagonal forces, S, which were determined as shown in Figures 28a and b as acting on the stiffened frame become the external loads with reversed signa on the box girder. qe = S-bu/g. Qs 2S bsg Yu qu #5 b/g Fig. 31 The resolution of S into forces acting in the planes of the plates (notation defined in Figure 25) Under the applied load, S, which can be mentally resolved into the components Gg: Gy and q, acting in the planes of the plates (Fig. 31), the webs and flangea deform not only in the longitudinal direction (flexural behaviour) but also in the transverse direction (frame behaviour), because the deflection of the plates are coupled with the displacements of the joints of the cross-section (distortional deformation) (see also Section 5.2), The load, S(x), is there- fore partly carried in longitudinal bending as in a simple beam, Sx), and partly by the resistance of the frame to distortional deformation of the cross- section, S (x): Say = Ske + Sex) Qs6X) = FER) + Fs K) (1)58 Structural Analysis. Rage shear forces, T, and Tl), contribute to the carrying of the load of the individual plates in the longitudinal direction, q”. ‘These forces, as the shear stresses in a girder of a standard section, are distributed in proportion to the shear, V(x), due to the load, q*, (Fig. 32). The condition that the longitudinal inplane stresses, %, of adjacent plates at their junctures must be of equal magnitude yields the idealized moment of inertia for the web as Taj = ZB fleerh) lotw 2) = 19. “I. si = 4+2B)> 3+ 3 B+ het Xu fh ‘S and places the neutral axis (Fig, 33) at 1+Z2B thu B . bs Yo= 343A +0, + ly B Yu = Bs ~ Yo therewith the longitudinal stresses in the web resulting from qt can be determined as for a regular girder: - eee “Yo Siu Yu (8.2) Loading of a wed panel in its plane fr My eq Fg, dagram Vv, «Ty diagram Fig. 32 Loading in the plane Fig. 33. Stresses o, due toS of the web Because of the antisymmetrical loading, all longitudinal stresses in the planes of the flanges are also known (Fig. 33). The deflections of the webs, v_, due to q* can be calculated from the differ- ential equation of an elastic béam as follws: Et, 2% a4 = - Tihs qs (53)ified Structural Analysis of the Superstructura 59. As the stresses, o,,, and O,,,, in the longitudinal direction, x, have distribut- ions proportional to one another (¢¢9/o¢4 = Vol Tus Fig, 33), the curvatures and deflections derived from them for all-of the plates must also be proportional to each other (e.g. ¥o/vg = - 2y,/b,). For that reason the diagonal deformat- ion of the edges, 6 (x), can be expressed by means of the geometrical relations of Section 6,1 in terms of the deflection of the web as follows: +, +, +. Fig Retin rere ye OB ee a Oe) 5 4) The resistance, 5, of the frame to such a diagonal deformation, 6, can thereby be determined by calculating 5due to a diagonal force, 5, acting on a frame element of length dx=1.The component of § in the plane of the web is of the form = bse - QEL 22842844 15 +1? Ferg S bebe he 5 wmere hex Breas] AT or, if from Eq. 5.4 is used: Jett =K- yon) whens Ke feo -kekeEls (5.5) From Eqs, 5.1, 5.3 and 5.5 the differential equation of a beam on an elastic foundation is obtained for the web E15: S42 + Kym = 4st) (5.6) whereby the first term describes the longitudinal bending, the second term the elastic foundation produced by the effect of the frame, and q, the load due to S. Similar equations can also be written for the flanges, These are however not necessary, because the distribution of the sectional forces in the planes of the flanges are proportional to those of the web. The solution, v, (x), yields the longitudinal moments 2, Mex) = - BIg; - cae (S.#) the stresses, (from Eq. 5.2), and the transverse moments, m (x),as shown in Figure 28 Yor the analysis of the deformable frame with the loading Sink Bo yw,60 Structural Analysis 5.5.2 Solution by Means of the Analogy of a Beam on an Elastic Foundation The solution, vg (x), sought for the web must not only be a function that satis- fies the differential equation (5, 6) but also one that fulfills the boundary con- ditions of the structure being analyzed. For example, a simply supported beam with transverse diaphragms at the supports can be represented as the system shown in Figure 34a, The basis for the solution of such a system is the infinitely long beam on an elastic foundation (Fig. 34b). It is subjected to all loads of the form, q¢(x) =S(x)-bg/g . which as shown in Figure 16 dl, arise from the resistance to the diagonal deformation, where S is derived as shown in Figures 28a and b. Moreover, the beam is loaded with unknown additional forces, Q) and Qy, and additional moments, M, and Mp, at the supports which are determined from the support conditions. 9 Fig. 34 Systems for the beam on an elastic foundation: a) in reality b) basic system with equivalent loading For example, for the beam in Figure 34a at the supports, x0 and x-l, the deflections, v,,and moments, Mg, due to the loads, additional forces,Q,,Q2, and moments, M;,My, are required to be zero. For a continuous beam with diaphragms over the inner supports, the condition of vg = 0 at these supports requires an unknown additional foree, Q, but no moment, at each of the inner supports in the basic system, Diaphragms in the span also compel v, = 0 at that point and the corresponding unknown additional force Q, Should the warping of the cross- section be completely hindered at some point, the condition of dv, /dx = 0 applies there, which ean be achieved by means of an additional moment at that point. “The influence of the resistance lo warping by transverse diaphragms and frames can be neglected here, however, For a supporting condition with- out transverse stiffening, the antisymmetrical part of the bearing forces should be handled as external loads (Fig, 35) as correspondingly shown in Figure 16, This translates into a large concentrated load, Q,, in the plane of the webs for the beam on an elastic foundation and not a support condition, Vy 4 Fig, 35 Loading at a support without transverse stiffening‘Simplified Structural Analysis of the Superstructure 61 The solution to the differential equation (5,6) takes the form of strongly damped vibrations as follows: Atxy =e*. (cosAX + sin Ax) Buxyze*. Cx) =e** (cosdx - Snax) Dexyse**. cos rx 4 } FETgi and 1+ characteristic length The effect of a change in loading or in the support reactions can be seen to die away to only a few percent of their original values at a distance of 3 L to 4L so that loads or support conditions for distances greater than 4 I, from the point under consideration can be neglected, For the same reasons, transverse stiffening i.e. diaphragms to prevent distortional deformations in the span is only worthwhile, if at all, if such stiffening is placed at intervals S 2 L, sin Xx where The sectional forces and deformations acting on a beam on an clastic foundation are handled in detail in the literature such as [33]. Loadings as shown in Figure 34 produce the following: yore ie {@rAcd + QrAd-x) + Q¢AUK-G.D} + 2 +h [my Bar +M,Bu-0} + +e {ex[Dia,-x) ~ Dia, a,-x)] + 6,[2-Dx-a,) -Dra,ra,-xi] + +¢,[Dix-a,-a,) ~Dix-a}} where ced C,= 0530 for 08x a, e,=4 C47 0y20 a,=x 4 ata, Cy=7 C, =C,20 QeQtx SL Mynr= Fx [QrCend + Qe Ct-x) + Q,- Cex a.))} + + {tt Don +M;Du-x)} + +e (-c[Bea-x) ~ Bea.ra,-x] +c[ Bora.) + Bra,ea, -x)) - ~¢,[Bix-a,-a,) ~ Box-aa]}62 Structural Anshysi Ve fade * Gy Dil-x) + COs DiK-a.)} + +f-Me Aa + MyAUn} + eft fed[ Cea.-x) - Clay a,- xi) + 0, [Cex-a) ~ Cava, x)] -e,[Cex-a,-a,) ~ Cax- ani} t.=4 for x$Q, ; Ge=-4 for x® a, ‘The unknown additional forces and moments can be obtained by plugging in the boundary conditions given, or from the following formulae (see also (33): 0? Acar — Hr [Ba ~Ba.+a] My * ote + Pte [Bia,ea,) ~ Beas)] 2 Sh Aca. +A [Daa - Dearan] = SR au-an - HE Dara) -Diasy] xf card reas 1 ee2S ceanal = eat QAR LZ (Mot Myo) BU) - 4A" ETsgj (Yo ol [4 * DELI] } Q" 2-2 xtef{2 (Mio Mao) BU +40 EIgj (Veo > Veo ~ DULY } Mie 24 f-lMyo* Mao) [4#AUL)] + 2 Elsi (Vo * Volt + COL] } Ma res{(Moo~ Mao) [1 ~AU)]-20 Els; (tio ~ Veal ~ Cee] } Q,=@Q+Q” Q,= Q’-Q” M, = MM +M” M,=M'-™M"Simplified Steuctural Analysis of the Superstructure 63 The transverse moments obtained from the folded plate consideration are Proportional to the web deflection, vg. The influence lines for v, take the form of Figure 36 in the transverse direction and in the longitudinal direction are proportional to the elastic curve of the beam on an elastic foundation sub- jected to a point load at the section under investigation (64 * Sq) In order to determine the critical positions of the loadings, one, however, rather seeks orientation from the known shapes of the influence surfaces for the plate effect and the influence lines for the fixed frame effect. This yields sufficient accuracy as the transverse moments from the folded plate action play only a minor roll. ‘The influence line (Fig. 36) has its maximum generally between the webs, Only in the case of very deep box girders with rigid flanges can the maximum lie at the end of the cantilever arm, In order to avoid calculating the folded plate action for every individual load- ing position in the transverse direction, it normally suffices to calculate the longitudinal streases for the folded plate (fp) action due to one characteristic loading position only (Fig. 36). A ratio is then determined between these stresses and the longitudinal stresses for a symmetrical loading of the same magnitude, This factor, ff, is used to increase all eccentric loads for the analysis in the longitudinal direction. é - Sp" a. oR stress due to the force S from the antisymmetrical q/2 (folded plate action) ++ stress due to the sym- metrical q/2 Fig. 36 Influence line n for vg, where the position of max ‘1 is dependent upon the cross-section. For the normal case yp™ 0.4b, For normal box- girder cross-sections, f,, lies between 1,1 and 1.3 for uni- form loads (main lane; qi as in Fig. 15) ‘Bnd between 1.3 and 1,5 for concen- trated loads (characteristic heavy lorry). As long as the folded plate action is not neglected in calculating the transverse bending moments it should be separately considered for every eccentric loading.64 Structural Analysis: 5.6 Multiple-Cell Box Girders Basically the same is true for multiple-cell box girders as is for single-cell ones, With increasing degree of statical indeterminacy an analysis by hand becomes too time consuming. In Section I, 4,3,3 comments are made con- cerning the design of multiple-cell box girders. As multiple-cell box girders are only constructed in special cases, the design engineer is here referred to the literature [29][1, 26] as well as Section 6.2 for further information concerning their analysis. 5.7 Curved and Skew Box-Girder Bridges In horizontally curved bridges the longitudinal bending moments, Mg, and the torsional moments, Mp, mutually influence each other, Some sectional forces for a single-span girder are compiled in Figure 37. For special supporting systems and multiple-span girders, the reader is referred to [25] and [34 to36). According to [38], two-cell box girders can be roughly analyzed by using factors for the longitudinal and transverse bending moments, the shear forces, and the deformations taken from an idealized straight bridge. Skew box-girder bridges are in principle not sensible, as in most cases the box-girder can instead be supported perpendicular to its longitudinal axes by affording a minimal increase in length (Fig, 1,53). [ 37] contains a method of analyzing a skew box-girder bridge. a Se cea Mme sce"? [8] 40 ro ny Mg to Hy tt wow | [xg |e 8) lat and) q2imersee | ‘fice | [alge oes) | ert at Galore a fngesne a | eee o | 1 ton | ome (S22e# 4] PERT ng erie | mine - sn8 cence | 1 Pr SE ag atwa) | “Osmo Fig. 37 Support reactions and sectional forces acting on a single-span girder torsionally fixed at both ends (from [ 25, 34])Computer-aidad Analysis of the Superstructure 65 6, COMPUTER-AIDED ANALYSIS OF THE SUPERSTRUCTURE 6,1 Theory of Prismatic Folded Plates A folded plate is a 3-dimensional structure of thin plane plates joined together along their longitudinal edges. Normally an analysis of a continuous box- girder bridge by means of the folded plate theory is too involved to be carried out by hand, The following presentation is therefore not intended for practical implementation but rather serves the main purpose of calling to mind the basis upon which a computer program using this theory is built. ‘The loads to which the folded plate is subjected are carried by the plates both parallel and perpendicular to their planes to the end diaphragms serving as end supports, A simply supported prismatic folded plate (Fig. 38) is here used to present the fundamentals of the analysis as according to [39] . The following assumptions are made: - that ordinary bending theory is used - that the section under investigation is at a sufficient distance from the end diaphragms - that a maximum of two plates or diaphragms meet at an edge. ‘The loading is expanded in the form of a Fourier series (Fig. 39) as follows: qm 2£ Qtr sin ax where o¢ = Oi Loadng Case Sn 9 eq pommmmmm | A metas ee are E13 ec sme N28 pots EK, Fig. 38 Simple folded plate Fig, 39 Fourier coefficients for the most important loading cases In the following, the analysis is portrayed for the m™ loading term. For simplicity, however, the index, m, is dropped in the discussion that follows.66 Structural Analysis 6.1.1 Hinged Folded Plates The system is idealized here as a number of prismatic plates joined together by hinges which induce no moments into the statically indeterminate system (Fig. 40), The applied loads on the plates are reduced to loads, Q., acting at the edges of the structure, whereby the edges are at first assumed to be rigidly supported, The edge loads are resolved in the directions of the planes of the plates as shown in Figure 41, The dead load becomes, for example Qn = t Gas Daag r Qn Bn) Sina Qu’ sit te [sit Yor Spneg® Qn’ sin Fra fin Sn This results in a load in the plane of the plate, n, of * si - Gn Gr sin ax = Saneg ~ Snegn Q ® Fig. 41 Resolving the load Fig. 40 Hinged folded plate subjected into ite components to dead load in the directions of the planes of the plates As the members cannot deform independently of each other along their mutual edges, edge shear forces, T, and T,,,, are necessary, Requiring equi- librium on an element, dx, of the plate, n (Fig, 42) yields AVn = - GnAK ; ANy= (Ty - Trea) tk 5 (6.4) an AM,= Yd - $I, Tasa) bn OX ED edge Mn+ dMp Integrating, one obtains Nine dNn Va = VarCOS OK 5 Nn = Na-sin ox ex Wied y M=Misinax , t= Ttcosax * Rage net with the fixed values T,* and de = * ne ) Nar TT) Fig, 42 Sectional forces (6.3) in the plane of the member = Pe tet)Computer-aided Anoiysis of the Superstructura 67 ‘The condition that the strains and therefore also the resulting stresses of the adjacent plates along the edge, n, are equal is given by N, 6 — Nae 6M, a ae ld From this plus Eqs, 6,2 and 6,3 the "three-shear equation" for edge, n, is obtained as follows: Bien (oe tec) ate = oR, te) This yields one equation per edge: i.e. for a single-cell box girder, four equations for its four edge shear forces, and with Eqs. 6.2 and 6.3 the sectional forces, M,, N, and V,, for each plate. The deflections in the planes of the plates follow from the differential equation of the beam by substituting Vv, 7 VA sin ax as follows: PY, Elwdxt which yields 4 Ej . ie ye [Gn Oat Tate): 25] The edge deformations, w, perpendicular to the plane of the plate can be determined as shown in Figure 43. . -EI,Vfatsinax =-My 1, =f Fig. 43 Geometrical relationships of the deformations, e. g.on edge 2 6.1.2 Folded Plates with Rigid Joints In order to remove the rotations of the hinged plates in relation to one another, the redundant moments, Y_ = ¥* sina x, are applied at the joints (Fig. 44). nn ‘The associated support reactions of the plates AQ, = Yee Yo68 Structural Analysis: create the addittonal plate edge loads é a. AGnes 4 qn= ~AQn (cot y+ cot Yaa) — Ze SIN Yrnat From this one can determine the moments, M_, the normal forces, N., the shear forces,V_, the edge shear'forces, T,, and the deforinations, vj, and w,,as shown above. ‘They are all dependent upon Y,, (for n= 1 to 4), The 4 determining equations for Y, to ¥, are obtained from the condition that every angle between two adjoining plates does, Fig. 44 Edge loads, 4Qn, not change, due to the edge moments, Y) ‘The rotations in the hinges due to transverse bending of the plates are determined as for a beam. The load due to self-weight, gq, yields for example (Fig. 45a) bn , Onnee = Onetn = FEE: sin & ‘The redundant forces yield (Fig. 45b) 2b, Yaymet * EES (2Yn* Yous) and 2b, YWootin Ete (2 Yous * Yn) The angle produced by the deformation of the plate loaded in its plane (Fig. 45c) is =- = Wiretin = Wayne Boos => Voy = Liens es Fig. 45 Angle of rotation at the edges With the above, the compatibility condition for edge, n, takes the form ance nme * Yana + Mines * Ula * Wines =OComputer-aided Analysis of the Superstructura 69. By substituting the values determined above, all moments Y, all sectional forces, and all deformations can be determined. As a closed summation of the Fourier series is not possible, separate calculat- ions for several periods (for m * 1,2,3.....) must be carried out in order to achieve sufficient accuracy. With this can be seen the limits of the manual folded plate calculation, Influences such as diaphragm stiffness, plate wall shear deformations, or continuous folded plates are not even considered in the above presentation. For theoretical investigations, the reader is referred to the extensive literature (e.g. in {40}), Of practical significance are those methods prepared for the computer. The extended folded plate theory of [41] is transformed into a computer program in [42} and (43). 6.2 Finite Strips Analysis by means of finite strips lies somewhere between the rigorous folded plate theory and finite elements. In order to decrease the amount of calculation required and especially to be able to analyze folded plates with any sort of boundary conditions, the boundary conditions are satisfied here in the longitu- dina] direction by periodic functions (Fourier Series) and in the transverse direction by simple polynomial functions (44, 45] . For long girders, often only a few coordnate yet gy element strips are necessary. For short girders sities ima and for greater accuracy, the number of strips is increased (Fig. 46). With further refined strip elements {46}, any boundary conditions, multiple spans, and transverse diaphragms can also be considered (see also [47]). Fig. 46 Dividing the box girder into strip elements (from [ 46} ) Analysis by means of finite strips or strip elements is advantageous for multiple- cell [ 44, 45, 46] or curved [48] box-girder bridges, while for single-cell straight bridges an analysis as described in Sections 5.2 to 5.5 is sufficient in most cases. 6.3 Finite Elements With the finite element method, the two-dimensional partial differential equations are reduced by discretization to a two-dimensional system of algebraic linear equations, The structure is divided into elements that are only connected at their nodes (Fig. 47).70 Structural Analysis finite element nodal point fine mesh coarse mesh Fig. 47 Finite Element analytical model (from { 49] ) The forces and moments acting on each element (Fig. 48) are reduced to forces and moments acting at the nodes (Fig. 49), The conditions of equilibrium and compatibility supply the n equations in n unknowns, whereby n depends upon the number of elements and the number of degrees of freedom of each node, se ni Nagy SY Membrane Forces Slab Forces Fig. 48 Positive internal forces and Fig. 49 Nodal point forces, S, displacements in a finite and displacements, v, for element (from [49] } yi ‘inite element (from [{49}) i Fig. 50 Triangular elements (a)with 3 nodes, and (b)with 6 nodes for a refined analysis (from (50) ) Along with rectangular elements, predominantly triangular elements are used (Fig. 50), In order to avoid numerical difficulties, triangular elements should not be given too small an angle, Parallelogram-shaped elements have been developed for skew bridges. A survey is given in [50]. Recently the development of 2- or 3-dimensional finite elements capable of simulating the actual nonlinear behaviour of reinforced concrete after the formation of cracks has made considerable progress. [51] and [52] contain examples of this, Figure $1 portrays the results of such an analysis in which the large deviation of the actual behaviour from the elastic theory can be seen when the structure is heavily loaded,Computer-aided Analysis of the Superstructure 71 —A stiffened => B unstiffened a) Failure mechanism of an b) Loading and stiffening at mid-span unstiffened box girder Load P [MN] 0.8 T 1 10 web away from load _| ‘ 06 () i | Elastic "Theory tng > 20 5 0 5 a a O Of 02 03 04 05 06 O7 Deflection (mm| Load P [MN] — box girder A stiffened = box girder B unstiffened ¢) Relation of the deflection d) Relation of the angle of rotation to the to the load at mid-span load at mid-span Fig. 51 Influence of nonlinearity on the deformation at mid-span of a box girder (from [53}) A uniform gridwork for the closed analysis of an entire box girder should be avoided because of the associated effort involved, especially that for the evaluation of the computed data, It can be more sensible to analyze the structure using a rough grid in the area of no particular discontinuities and a refined grid only in those regions where there are abrupt changes in geometri- cal or statical conditions. For an introduction to the fundamentals of finite element analysis, the reader is referred to such literature as (54) or (55). In [56], [57] and (58) box- girder bridges are analyzed by means of finite elements. Because of the extensive input data required in using finite elemént programs, especial emphasis should be placed upon the ease of use of such programs, This trans- lates into the requirement that the gridwork is automatically generated for somie regions,72 Structural Analysis: 7, ABUTMENTS ‘The abutment is subjected to the following loads (Fig, 52): ~ dead load, g ~ support forces from the superstructure, F,, F\, F x By Fe - earth pressure from the backfill e,,; ‘The magnitude of the earth pressure depends upon many factors of the earth (settlement, subsequent compaction by the traffic) and the abutment walls (movement, surface roughness, type of foundation) and is therefore difficult to completely consider. It is therefore recommend- ed to use the earth pressure at rest in the calculations, ~ earth pressure due to the working load on the backfill, eg Fig, 52 Loads on an abutement braking forces on the backfill, eg working load on the abutment, e.g. footpaths, q lateral impact, Qy other loads {59] concerns itself with the analysis of abutments and contains tables for plates supported on two or three sides. 8. PIERS 8.1 Loadings Along with the sectional forces acting on the undeformed system duc to - dead load of the pier - support forces from the superstructure + movement and rotation of the bearings due to changes in length of the superstructure - wind - temperature differences - impact load from vehicle hitting a pier 2 - pressure of moving water for piers in water (Fp) = Cy + 5° v' + A, where the drag coefficient, Cp) ¥ 0,8 to 2 from tables, v > velocity of the water, and A = cross-sectional area of the pier to the flow) - ice pressure [60, 61} - impact from ship hitting a pier [ 62]Pers 73 arise moments due to = geometrical nontinearity (2 order theory) - physical nonlinearity (nonlinearity of the M-curvature-relationship for reinforced concrete) - horizontal force from the deformation of the system (Fig. 53) - unplanned eccentricity (construction inaccuracies) ~ foundation rotation At due to The influence of the theory of 24 order depends 4 ae temperature uyon the slenderness ratio of the pier and can Hl be neglected for A= Byg/i< 25 where Sopp... effective length (buckling length) Fig. 53 i... radius of gyration A horizontal force, Hl, arises from the inclination of the column hinged at both ends which acts on the fixed pier Effective Length In order to determine the slenderness ratio, X , the effective length, 8,45, the distance between the inflection points on the elastic curve, must be determined, In most cases, the pier is taken to be fixed at one end (Fig. 54). For the case of Sar¢/d,% 15, where d, is the width of the pier, and pinned supports the free movement at the top 6f the pier can be assumed to be limited [63] .In [64] , the influence on the pier moments of the manner in which the elastomeric bearings accomodate movement is portrayed. If the superstructure and piers together form a framework, the effective length can be taken from tables (e.g. [65]; see also [24, Vol. 11} ). ye T | ey | ! Li: —y S a i meron t 4 the foundation LL IN where2< M,, the calculation must be repeated. The values for M,, k,, and By, “depending upon the axial force, N, and the ratio of reinforcement,p,can be taken from tables (e.g, [65, 67] for rectangular and circular cross-sections, determined for ¢, Sim and ¢ - 3.5%), 's, Limit ‘¢, limit ~ ‘The tension stiffening provided to the reinforcement by the concrete between the cracks, was not considered above, This leads to an overestimation of the moments, MIT . Therefore for very high piers a more exact calculation should be made with consideration of the mean steel strain ¢,_ on the tensile side (68, 69]. 9, FOUNDATION Along with the ultimate strength of the foundation, its safety against over- turning should also be checked, whereby in most cases the freely standing pier during the construction phase is the critical case, Furthermore, the settlement due to constant load should be investigated. The analysis of the foundations depends upon the soil conditions. The foundation requirements are handled in detail in [70]. The foundation of the scaffolding requires exact knowledge of the soil conditions (see Figure I, 66), In order to be able to follow the flow of forces from the loads on the scaffolding into the soil a statically determinate model should be used in the calculation if possible tecause of the uncertain stiffness distribution.6 10, 2» Fy n 2 3 ” ” 2 0 a 2 ° “6 ‘Structural Analysis LITERATURE Danas experience, Serie R, No 129; ROSTAM, S,. PEDERSEN, F,5,: Partially prastroxred conarete bridge ‘Technical Univeraity of Danmark 1080 DBUTSCHE BUNDESBAHN: Voracheif {ar Rlsonbahnbedckon und gonstige ingenteurbauwerke (VU; Vorausgnbe (DV AO¢I: Bundesmatnaentralart, Manehon:. 1979 Windlaet het Bracken, Vorvcblag fr de Neufseaung der DIN 1055, Vell 4,Bundemanetalt {dr Stradenweven, Koln: 1980 HIRTZ, H,: Borteht dbur den Stand der Arbelton en Regeln aur Befassung der Waudwirking auf Hauwerke: Konsteuktt- ‘ver Ingenleurbas, Boriehte Hoft 38/36; Rubp-Univeretie Bochum, 1 PAUSER, A.: Grindsatze fOr dle bauliche Durchodung; Zement und Beton, Heft 3 Wien; 1976 HAMBLY, & C," Temporsture distributlon and wtrenwen {n concrate bridges; The Structural Hngincer; Vol. 56 A, and Nr, i, Vol. SPA: Londons 1978/70 PRLESTLEY, Wi,3,.N., aberaetet von W, FUCHS: Beressung von Detonbrdcken unter Tempersturbesnapruchan Bavingemeur, 5. 93; 1970 THURSTON, 8.3.) Thermal ate ring, Ualveralty of Canterbury LRONMARDT, F, Vorlesungen aber Ma Berlin - ‘New York, 1579 SCHAMBNCK, Tl, Bracken aus Speanbeton: Wirklichkelten, Mégichkelten: Bauingewlasr 81, 1 Berlin - New York, 1976 HUANG, 7.” Dynamite response of concrete highway bridges: ACI Publication SP ~ 23; 1989 CHU, KM, JONES, M_' Theory of dynamic analyale of box girder bridges Dynamic anslyals of a box girder bridge: ‘Abhandivngen 361i, VBR, 1976 Symposiucs on tynamie Behaviour of bridyew: RRL. Supp fealory; Crowthorne: Berkshire. 1977 Indign National Group of the IABSE, Dynamic response uf bridge decks in reinforced and prestressed concrete with ‘pectal reference to Vibration; Special Publication 3, IABSE: New Delhi. 1076 MATSUMOYO, ¥. NISHIOKA, T., SIIOUIRI, H., MATSUZAKI, K.” Dynainic dewign of footbridges: Proceetin 17/18; IARSB, 397m NARASIMHAN, S.V., PANT, C.P,: Free vibrations of urthe IARSK, Val 8, No. 2, 1979 STRINLE, A.: Torsion ina Profllverformang beim etnrelligon Kaatontrager; Weton und Stabibetonbau Het! 8/1970, (elohe wich Hott 10/1970 und Hott 8/1072) THOST, 31, WOLER, 11.J.: Zar wirklichkosterghen Ermittlany der Besnpruchungen in abschnitiswelse hergestell- tes Spanbetontragwerken, Hauingenieur 5, 1070 RUSCH, H., JUNGWIRTH, 9,, MILSDORE, i Kritiacho Sichtung der Verfahcen zur Hevicksichtigung der Binfvese vo Kelechon sid Selurindés dog Betong aut dae Veehalten der Tragwerke: Beton- und Stakivetonben 8 > 4 + 6/1977 DAZANT, Z.P.- Predicsion of conerete crecp effects using age adjusted effective modulus method, Journal of ACL, 63: fa concrete structure: Heaoarch Report 78 - 21, Depart ‘Christchurch, New Zealand, 1978 ivi, el 6; Grendagen Joe Ma vbedckenbaues: Springer Yering Springer Verlag tary Report 175; Traneport and Road Research Lano- Journ of the Indlan Watton Group w girder bey vp. 212; 1972 TROST, IL, MAINZ, B.: Zur Auewtekung vor ZaKegungen in Spanibetontragwerkon: Beton- und Stahlbetonbau 8. 1970 RIECKMANH, II. -P.- Nusvectacke integration von Einfubfluches fir Plattenechritlaaten: Bavlngenieur 11/970 HOMMERG, Hl,, ROPERS, W.” Pakehabsplatten mit vordnderlicher Dicke, Hd, 1 il Springer Veriay, Berlin - New Yorks 1965 MERK, C.1 Trdckonbay 140:Vorlequnganvanuakript; Abteilung fr Bauingenlaarwonsn, ETH Zartch:1372 LOPPOTH, W," Vorlesungemencakript Derechnungeverfahren {Ur Magsiverucken | Wtohikastenhrickend* I (gekrumsnte Briickent; Inatitut fOr Mawelvbay, Univerattat Stattgart UAKICT, B. Sinplitied analysie of cdye aiffened canitiever slabs; Tournel of rhe Structural Division, Vol, 107, No ‘ST-§. Murch, WR LiPPOTH, W.. Zur Besnapeactang mehescliger otlkastonguerachnitte quer 2r Lingsac! ‘Langevorapanning: Beton und Stanibetontou 12/1070 ENOL, 1, IVANYL. Go’ Ermittlang der Querbieging von Druckplatten gevouteter Honlkastentrayer; Bautechalk, Nett 4 Weriag W, Beast & Soha, Berlin - Munchen = Diweeldor®, 1971 CASTRILLON, 0, -4,- Bolteng sur Borechnung langer dtnnwandiger dreizetliger Trier unter Beeocksichtigung der Profilverformung: bigseriation, Untversicat Stuttgart; 1073, MAISEL, D-1., ROLL, F_: Method of unalyaia and deslun of concrete boxbeame with side ewntevers Report a2. 494; CCA; 1974 PRAKASH HAG, D.S.: Elnfluf dor Queractalitonbmegsungen auf dle Profliverformung van messiven Hohlks eto ud Stehlbetonban, Ileft 1, 1081 TUNG, Doi, Torsional snyeta of wingle thin-watled trapezoidal concrete box girder bridges: ACE Pubileation ‘SP = 75 1969 HETENYL, J,” Meas on vlagtie foundation; The Univeraity of Michigan Presa: 7. Auf., 1904 DABKOWSK, X." CGolersuante dannwnadige Tedger: Springer Veriag, Berlin - New York: 1988 MeMANUS, BF, NASIH, G,/,, CHLVBH, C.C.' Horizontaity cucved gindera'~ state of the ‘May 19 KONISIN, T,, KOMATSU, S,: Three dimentlonal analyela of curved girder with thin-watled cros fen 25, IVBH: 1965, kim, }., PELLE, K,. ZurSchithrattermttiing dew achisfen, sinzstligen Hohtkertenw im Betonbrickenbau; Verdifont- Tighiig. der Abtotiung Besweaon, Konstraktiver ingenteuries: Universitat Dortmund 1977 MUKHERJEE, D,, TRIKWA, D,'." Design coefficients for curved box girder concrete bridges: Indian Concrete Journal Ne 11, Vel. 84: 1980 GIRKMANN, K.> Fiacheptragwerke, 5, Auf; Springer Veriag, Wien: 1089 ‘Task Comitive: Phage 1 Repart on Folded Plate Construction: ASCE, Vol, 89, NoSt 8, S, 366 M2 GOLDAERG, 1, F,, LEVE, H.1.." Theorie of priamatte folded plate structures: IVBH, Vol. 17, 1957 SCORORLIS, 4.6): Anaiyale of atinply supported box girder briagew: Report No, SESM 66 - 17. Berkeley: 1956 MEYER, C.| SCORDELIS, A.C, Camper prayrem for prismatic folded plates with plate and beam elements; Report NoSESM 10-3, Berkeley: 1970 CHRUNG, ¥.X,: Flaite Strip Method In etracturs Paria - Franetuet, 1976 CHUNG, ¥.K," Analyala of box gitder bridges hy the Finite Stelp Method; Vol. 3, ACI Pubtieutions SP 26: 1971 we us Umiennkrarten der ‘echnical tenteagern: ASD, Vol 95 No ST 5: jection: Abhandlun- soalyeia: Pergamon Praga, Oxford - Now York - Toronto - Sydneye © s0 8 oo ° 60 6 Be 6 6 8 56 8 6 Literature 77 DU PREEZ, R.J.: Rerechnung prusmatischer Fitckentragwerie mit Hilfe von finiten Stralfenolementen; Beutechritk eft 5; Yering W. Eanat & Sonn, Berlin: 1972 00, ¥.-¢,, CUSENS, A.H,: Developments of the finite strip method in the snalyais of bridge decks; p. 53 CMEUNG, MI¥., CHEUNG, Y.K." Antlysla of curved box girder bridges by Fine Strip Method, Abnaudiungen 2. - Wall 1971 SCORDELIS, A.C,, DAVIS, FE.» Streames In continuous conerete box gitder bridges; Val. 1, ACI Publications SP - 26, 1871 SISODIYA, R.C., GHALI, A.: Analysis of bor girder bridges of arbitrary shape; Abbandlungen 39 - 1, 1VBH; 1973 ARGYRIS, J, i, ot al.. ‘Finite Elementberechnung von Spannteton-Reaktordruckehlltern; Deutecher Aveechu flr Stantbetin, Hem 279, 1877 IARSECollogstum. Advanced mechsales of reinforced concrete: Delft, 1981 MAURER, G., MEMLIORN, G,” Zum Tragverhalten des clnzelligen Voryeepannion Kaatentrigers onter Bericke!ehtl- ung wiridichieitenahen Werkstoftvorhaltena; Porachong, Strofenbau und Stralenverkehretechaik, Heft 270, Bont, 1970 WUCK, KR, SCHARPP, 0.W., STHIN. £., WUNDERLICH, W." Finite Elemente in der Statik; Verlag W, Ernat & Sohn, ortin " Manonen; 1873, GALLAGHER, 1 11,: Pinite-Blement-Analyaie; Springer Verlag, Berlin - New York 1976 VAN 2¥L, S¥.; Analyste of curved aegmentally erected prestressed concrete box girder bri ‘SESM 782. Derkeley: 1970 SCORDELIS, A.C. ‘Analyiical and experimental stadiew of multi-cell concrete box girder bridges: bulletin». 58, Vol XVI-2, of tne IASS; Madea, 1975 SINMODIVA, R.G., CHEUNG, 'V,K,, GHALL, A.: Finite Element analyete of skew, curved box girder bridge: Abhandlun- yen 30-11, (IH; 1970, SMHL. J. TVANY!, G,, SCHAMBECK, H,: Berechnung ksstenformiger Brdckenwiderlager: Werner-Veri ‘ons NEILL, €.R.: Dynamic ice forces on plera and piles: Can, Journal of Civil Engin FREDERKING, R.. SCHWARZ, J-- A todet Investigation of ice forces on feed and ovciiating con 1980 CSTENFRLD, C.- Ship collisions against bridye plers: IVBH Publications Vol, 26 1966 HIRKMANN, K. "Das Trag- und Verformungeverhalien van Stahibetonbrdckenpfeilern imit Rollelagern; Deutacher Aus lichot fr Stahibeton, Tet? 266; 1976 KANNING, W.” Elastomer-Lager fur Pendeletttzen ~ Einflul der Lager auf die Beanapruchung dur Statoen: Bauingenieur (2; 1ge0 KORDINA, K., QUAST, U.- Nachweia der Knickaleherheit in: Bemosmung vou Beton- snd Stahlbetonbautetlen; Deutacher Ausuehu far Stahlbeton, Mett 220, 2, Auf; 1978 CEBIFIP Model Code for Concrete Structuren; 1978 WALTHEW, Tl, HOUNTET, B. Remessuagetabellen fur Stattbetonquerechattt polytechniquoe romanies, Laueanne, 1880 NOAKOWSKI, P,” Vereinfachte Eemistvng dec Momente 2. Ordnung in Indusirloschorast des Inetiuts fur Baulngenieurw egen Il, Loheatutt fr Mesatvoau, "TD Munchen: 1980, SCHLAICH, J., SCHOBER, 1, KOCH, R." Versuche aur Mineirkung dee Betona in der Zvgzone von Stahlbetoardheen: Inatitu foe Mugsivbou, Universiee Stuttgart 1079 KLOCKNER, W,, SCHMIDT, H.-G. Grundungen, Betonkalender (877, 1 §. 737 ff Report No UC Dusseldorf etn. Vol. 3: 1926 [IMT 80-208 /01-08, and I; Honiqueraennitte: Pees wens Bericht Nr, 1392/0 /Introduction / General Detailing Principles 79 Part Ill DIMENSIONING AND DETAILING 1, INTRODUCTION In this part the dimensioning and detailing of box-girder bridges is handled. This includes also those areas whose stresses cannot adequately be described with the normal bending theory as assumed in part II, Further appropriate prestressing tendon profiles and reinforcement arrangements are suggested, and elements of detailing and bridge finishing are illuminated. 2, GENERAL DETAILING PRINCIPLES The general design principles of reinforced and prestressed concrete struc- tures apply of course as well to box-girder bridges but with the more stringent requirements necessary to ensure their serviceability and durability. In this publication intended for the international forum of IABSE,only the CB/FIP Model Code in its Jatest edition [1] can be recommended as a safe and practi- cal basis for ensuring the structure’ s ultimate strength and serviceability, Meeting the safety requirements of the national codes of practice must be left up to the individual reader, With today’s level of knowledge and state of the art, the structural analysis and dimensioning of box-girder bridges cannot dispense with any of the three methods named below or achieve a satisfactory result with the exclusive use of only one, The following methods are also allowed equal status in the CEB/ FIP Model Code: - the determination of the sectional forces as according to elastic theory; i.e, the assumption of the uncracked State I independent of the magnitude of the stresses (see Part Il of this paper) and the dimensioning of the cri- tical sections so that the behaviour of the member with regard to cracking, deformation, and vibration is satisfactory under working load or service load conditions and so that the ultimate resistance of the member at failure or collapse (Fig. 2a) should provide for sufficient safety, - the determination of the sectional forces when the nonlinear stress resul- tant - deformation relationships of the cracked State Il are considered, in particular the nonlinear moment - curvature relationships (Fig, 1) which are a function of the loading and the restraint deformations (e.g. due to temperature or settlement), Step-by-step iterative integration or appro- ximate methods are used [ 2, 3], The dimensioning of the sections is carried out as described above, fint- Healt leamL 5 1 Fig, 1 Moment - curvature relationship80 Dimensioning and Detailing - a closed check of the ultimate limit state based on the theory of plasticity #, 5, 6]. The static method, which supplies a safe lower bound for the ultimate load, ie especially suitable for beam members (Fig. 2b). The kinematic method or yield-line theory which supplies an unsafe upper bound, is especially suitable for plates (Fig. 2c). The sections must, however, provide a sufficiently large rotational capacity, Box girders with their ro- bust slabs satisfy this requirement in most cases with sufficient mild steel reinforcement, whereby overly reinforced box girders (p> 2%) are ex- cluded for economic and structural reasons anyhow. Although the ultimate load design methods for beam members and plates are loddy in principle so thoroughly developed that they can be used, sufficiently simple methods for box-girder bridges that can be comprehensively applied and that take every effect into account are still lacking. This is especially true for such cases ag eccentric loadings or bridges with curved alignment. Up until now this has only been met by a structural analysis according to the ela - stic theory and a dimensioning based upon the load-carrying capacity of the pet arm Oy (oo Te Pe a sents mi 2plast moments ang “Ta = equilbrian satisfied ee =kinemat_ plousitie mechanism. upper -plast. moments | bound 2w Mp tMy Age Ay tow a ru(yhsant) Lea f,4, as they are reduced at time t = © by the dead load, g9, plus the live load, q, and no aggressive road salt environment is to be expected in the inside of the box girder, Extremely large transverse tensile stresses occur in the top and bottom flanges of cable-stayed bridges (Fig, 8) for which the flanges must be nearly concen- trically prestressed. When the bridge section is suspended from both sides by cables however, this favourable influence reduces the need in most cases for a box-girder section to that for a deck slab stiffened by transverse and longitu- dinal ribs, hanger enforcement tor Zp recuired — tension <==+ compression WITH CENTERLINE CABLES — tronsverse tensle stresses Lehind! the anchorage are ‘usually negtigibie Ise aso Fig. 19) Fig. 8 Transverse tension in the flanges of cable-st ayed bridgesPrestressing 87 3,2,2 Prestressing of the webs Only in special cases for very heavily loaded webs such as when a parallel- girder bridge has a bottom flange that is extremely heavy or is even required to carry vehicle loads is a prestressing of the webs to be recommended (Fig. 9). To simplify the construction, the prestressing tendons in the webs should be placed vertically if possible. They must be grouted from below. 3.2.3 Bottom Flange ‘The bottom flange is only transversely prestressed when it must handle vehicle loads, concentrated transfer of forces (see Section 3.4), or large transverse tensile stresses due to lateral distor- tions induced by the longitudinal pre- stressing tendons. In the first case, the tendons are placed as for the case of a beam fixed at both ends (Fig. 9); in the other two, they are located con- Fig, 9 Prestressing tendon profile Centrically. in a heavily loaded flange endon Profile in the Longitudinal Direction Depending upon the construction method selected, quite special requirements emerge for the anchorages and couplings of the prestressing tendons, There- fore, no statements can be made here which are valid for all cases. ‘The longitudinal tendons should be placed in the top and hottom flanges if at ali possible {see Section 3.3.2), as they greatly weaken the slender webs (Section 3.3.1), 3.5.1 In the Webs ‘The longitudinal prestressing tendons in the webs are generally placed such that their profile follows the shape of the bending moment diagram due to dead loads (Fig. 10), In the case of bridges with a curved alignment, the outside web is given a higher prestress or the tendons are placed at a greater eccentricity in order to counteract the torsional moments. ‘The curvature of the tendons in plan should be given due consideration in this case. rs seecal cose ote diaty 71 a ° Qy. A= 05 uy eb wet aati ai anchoring. ond ot aupersteucture i Caupiing or overtap Fig. 10 Prestressing tendons in the web