You are on page 1of 116
Structural Engineering Documents bra SCHLAICH Born 1954 a steten A. near Stutgert. Stuer of Architecture and Chl Engineering, Ois-Ing ‘Teenriea! University Borin, ta ‘of Sonce of Cas Teen, Cleveland) Usa, end Dr-tng. of Unversity Serer. Profettor and Direct ofthe Ist Jérg SCHLAICH Hartmut SCHEEF {res (ables, mombranes. Conslting Engine Check Consultan, Soh Ghul Engnering Consultants, ‘Stuttgart, Fed Rep. of Germany. Selected Structures: Televison tomers at Kiel and Ki, Shel rf, Cabienet oot for Munich Olymois, (Catianet cooing toner at Sens ‘neem, Cabiestaved brine over fiver Hooghly in Caleuts, Sloe Chimney a Manzanares. Hertmut SCHEER ‘Born in Stutgert in 1048. Atte comaleting his Cl Engineering Studies atthe University of stuttoart rm Leanhorat and And, Stunort Sines 1975 he ber boon inv in teaching and esearch ithe ltitut fr Maio, Une Sty of Sug. Hie rear ed International Association for Bridge and Structural Engineering IABSE ‘ed mainly to computersises ‘Association Internationale des Ponts et Charpentes AIPC aie at trecnt Internationale Vereinigung fir Briickenbau und Hochbau _IVBH ‘trond concrete bears Structural Engineering Documents Jérg SCHLAICH Hartmut SCHEEF CONCRETE BOX-GIRDER BRIDGES International Association for Bridge and Structural Engineering IABSE Association Internationale des Ponts et Charpentes IPC Internationale Vereinigung fiir Briickenbau und HochbauIVBH Copyriaht © 1962 by Internation! Assocation for Grdge and Structural Engineering All sghtsreterved. No part ofthis book may be reproduce in any form or by any means, electronic o mechanical, Including photocopying, recording r by any Information strape and retrieal syst, without permission in ‘writing trom the pune. sn 2057480319 Printed in Stata Publier: TABSE “AIPC-1VBH ETH - Honogerbera (CH8008 Zirh, Switzerland ‘el: 01/377 2847 ‘Telex: 82 1861485 CH ‘ele TABSE, CH8003 Zien FOREWORD ‘The box girder is today the most widely used superstructurein concrete bridge construction. That fact, justifies the suggestion made bythe Commission Il of the BSE that a comprehensive survey be ‘writen concerning ths paricular bridge type. The authors proceed from the sesumption, however, that ‘ts contents wil frst be drawn upon when ll posibiedesigr alternatives fr the parla Bridge pro: ject have boon thoroughly examined, end the box girder hasbeen proven appropriate, Their aimless that of encouraging the one-sided propagation of box-gitder bridges but rather much mors that of con- tributing tothe improvement of the quality of such bridges. They hope to contribute to ths by exten ‘vay roving the engineer of the study of today’s hardly suveyable mass of erature onthe subject ‘50 that he ean better devote that me tthe actual design othe bridge. That expaine wy ths paper is kept shor, why in particular cases the reader is referred tothe larature, and why subjects not per taining tothe cntral theme are only touched upon and not handled exhaustively, For grester clases, the survey follows the sequence ofa practical bridge design process by dividing ‘tsa into three main parts, namly, "Design", "Structural Aravind “Dimensioning and Structural Detailing”; ach section with ts individual numbering and Merature list This survey direct itso especialy othe design engineer, which manifest its, for example, inthe fact that the constuction methods are handled only riety and in the section “Design because thoy ecisively influence the design at the very begioning. Major contributions to Section I, “Structural Analysis”, were made by Prof. Dr-Ing. Kurt Schafer, 9 co laogue ofthe author inthe Institut fr Massivoau the University of Stutgert. In this section the attempt is made to portray the calculation ofthe box girder sectional forces resuiting from eccentric ‘vehicle loads with consideration ofthe folded pate ection or profile deformation so comprehensivaly ‘that its not only easily understood but also rapialy appicbis inthe design office. This thereby elimi nates te often-discussod, controversial question as to whether te effort involved in the "exact calcu lation ofthis loading casei actualy worthwhile or whether zn estimation ofthe transverse load dstrb- ‘uion would not eufice, ‘Te outhors woul keto take this opportunity to thank Professors R. Favre, Lausanne, and C. Mann, Zirich, for their ertical examination of the paper. Thy ae indebted to Mrs. Paechtor and Mir. E. ‘Schnoe for ther conscientious preparation of tha manuscript and Mr , Kutz fr his empathetical translation of ther German orginal into English ‘Stuttgart, January 1982 ‘TABLE OF CONTENTS Part! DESIGN 1 2 a “ 8 ‘Terms, Symbols Introduction Historical Development Over All Design 41” Design Principies 41.1" Tho Role and Sequence ofthe Design Process 412 Remarks ae to Form 413 Coste ‘42 Construction Methods 43 Supersructure 431 Gener 432 Longitudinal Direction 433 Traneverse Direction 444 Complete System and Supports 45. Substructure 451, Abutments 452 Piers 453. Foundations terete Parti! STRUCTURAL ANALYSIS 1 2 a Introduction : Loads and External Influences ‘Structural System 31” Final State 382 During Construction Critical Loadings and Sectional Forces 441 Longitudinal Direction 42 Tranoverse Direction ‘Simplified Structural Analysis of the Superstructure 51" General 52 Analysis Procedure 153 Analysis inthe Longitudinal Direction 5531” Sectional Forces due to Loads plus Restraints 1532 Time-Dependent Alteration ofthe Sectional Forces by Creep 84 Analysis inthe Transverse Direction 541" Sectional Forces Acting onthe Flangos 542 Ansisis asa Frame 1543 Transverse Banding Moments inthe Haunch of Vorible-Depth Girdors 55 Folded Pate Action 551 Fundamental Concept '552 Solution by Means ofthe Analogy of @Baam on an Elastic Foundation 58. Multple-Coll Box Grdors ‘87 Curved and Skew Box-Girder Bridges Computer aed Ana of he Suprsouctue "Theor of Prismatic Folded Pitas 8.11" Hinged Folded Plates 1612 Folded Plates with Rigid Joints 82 Finite Strips 5 (63. Fine Blmonts REBRBRSEEESS sss ewe BeIRRaLRAIIELSSSSASHH HSS SEES 2. Abutment 8 Pore 1” Loadings 82 Effective Length 83 Moments According o 2° Order Theory 8. Foundetion 10. Literature Partill_ DIMENSIONING AND DETAILING 31” Tho Lovel of Prestress '82 Tendon Prof in the Transverse Direction 21 Top Flange 322. Prestressing of the Webs 323 Bottom Fangs 33. Tendon Profilin the Longltuinal Direction 331 Inthe Webs : 332 Inthe Top and Bottom Flanges 34 Transfer of Concentrated Presress Forces 35. Construction Joints ond Couping Joints 4 Dimensioning ofthe Top and Bottom Flanges ‘5 Dimensioning and Reinforcement ofthe Wobs ‘51 Dimensioning fr Shoar, Torsion, and Transverse Bending 52 Web Reinforcement 8 Bridge Bearings 81 Bossing Types 82 Installation and Maintenance 83 Design of Brings 8. Bridge Finishes 81" Expansion Joints 82 Bridge Raling 83 Rosdway Surface 94 Drainage 10, Uterature 107 107 107 108 Pars | I and Il form 8 whole and ae only divided for oranisation reasons. Should the reader be referred to a figure, @ section, oF 3 ference in one of the other parts, he wil find thatthe Roman ‘numeral ofthe other part ofthe text i placed before the Arabic number; for example, Figure I? or Section Il, 81 Port! DESIGN 1, TERMS, SYMBOLS ‘Teome, Symbola 1 SECTION a-a Fig. 1 Sections through a typical simple box-girder bridge | BAZ ee Sn i eieroe om ‘EEE oo 2 _—— i epee Sn i zo 2: ieee. § BE Fo 2 ER. By and large the text and formulas u CEB or 1SO symbols. 2 Design 2, INTRODUCTION ‘Though box-girder bridges are indeed often not the only solution to a bridge project they are, however,eeldom 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 Iandscape 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 behaviour 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 detail ‘The above mentioned characteristics have made the concrete box-girder bridges the most widespread bridge type today. The.fact that this develop- ‘ent wil continue for some time futile this warvey concerning a tapi tat 4s certainly much too specialised for he who does not like to see his de: poseibilltice restricted. ‘Therefore let one be reminded that the box girder is only a part of the entire bridge structure and that the directly supported ox girder continuous over the supports is only the standard case. ‘The box girder can also be found in portal frame bridges, arch and bow bridget eable-atayed and euspension bridges of all kinds (Fig. 2). TIT FTN OE eh, Ti pe oe port: trome, forch 1 bow ‘suspension | cable-stayed Fig. 2 Bridges with the box girder as main structural element istrical Dovpment 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 fil 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’ ¢ Risorgimento Bridge in 1911, a 3-hinged arch (see Fig. 3). SECTIONG-al/SECTION b-b asof}-1a.00—4.a50 D Due to improvements in the quality of concrete and steel as well as a better understanding of the material behaviour of reiaforced concrete, the arch lost ‘more and more on importance as a load-carrying system to that of the beam Joaded in bending. The longest span simply reinforced bridge to-date wa: 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, Fig. 8 Risorgimento Bridge It was only with the development of high-strength prestressing steel that it became possible to span longer distances. The first prestressed concre bridges, most of 1 or I-cross-section, were built towards the end of the 1920". 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 coste climbed sharply. This thereby shifted the emphasis of development to the construction method. ee sation ana pressing cable (vibeut bond! Fig, 4 Sclayn Bridge over the river Maas; Prestressing by means of straight, large bundles of parallel wires without bond 4 eign Important development stages were the following: ~ cantilever construction in situ: the bridge over the river Rhine at Worms in 1958 with a main span of 114m, the Hamana bridge in Japan in 1978 ‘with a main span of 280 m, the cabile-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 {n 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 $4 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 firet 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-lane, the one-celled built-up cross-section constructed in modular fashion emerged ae the last development (eee Fig. 5h ). This allowed the Semorile Viaduet to be built by firat incrementally launching the U-shaped portion of the box girder and adding the deck slab afterwards by means of precast elements and insitu conerete 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- lever deck portion was added in insitu concrete and supported by precast concrete struts, The Kochertal Bridge and the Liesertal Bridge were simi- larly 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). Hits! Devlopment 5 o EEE er WOT Case ooSer . SS TS » Nov Fig. 5 Development of the box-girder cross-section 09 —______4 | x Fig, 6 Semorite Viaduct [3] Fig. 7 Eachachtal Bridge (4) 2 engitudinal y restressing stron iver component? “Hons instu concrete int 2 tilever component pier Fig, 8 Hammersmith Flyover Fig. 9 Brotonne Cable-Stayed in London (1) Bridge with centerline cable attachment [ 8 eon ‘The concrete box girder with streamlined cross-section has also been succes ful recently in the case of cable-stayed bridges, Ite high dead load favorably influences the dynamic stress amplitudes of the cables and the necessary long- {tudinal prestressing steel. A torsionally stiff box girder is required to hhandle 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 dinphragms 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 (8) : a) cable-atayed bridge, b) approach span glozes openings {or 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 Pincipee 7 4, OVER ALL DESIGN 4.1 Design Princip 4.1.1 The Role and Sequence of the Di ign 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 ean be traced in most cases to 8 faulty overall design. Design entails finding an optimum compromise among the particular objective and subjective boundary condition® 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’ « 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 cage, 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 ia 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 ie 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 Remarke 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 ite 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 fon 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 this subject, That will even be achieved in the event that the reader finds the following guidelines or orienation aids to be false and in this ‘manner critically grapples with the questions of form (Compare especially [8 10, 11] as well as (8, 12, 13]) 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 signt- feantly 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 , 1/4, alone is no guaranty for a light and elegant appearance (Fig. 13). = Order in the lines: Restrict the characteristic ines to a few directions and avoid their intersecting. A straight line should merge smoothly into a F80 a — $4 “SCL” a ~ Simplicity and clearnes: the necessity of the material to be perceived, and avoid unneces: = Integration into the environment: Bither 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 st 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 epang which decrease in length in the direction of the abut- ments are found to be pleasing (Figs. 14a, ¢, and f), Very irregular span lengths produce a feeling of uneasiness (Fig. 14b). 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 144 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. 140, lS TT ye CY oo oe pS 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. Especially 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. Ea a oo 44 oS ETa \ ie 8 at ETS Fig, 15 Shadow effect caused by Fig. 16 Fascia Beam: the form (b) 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: - For a/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. 18). 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): 4,/4™1/8 to 1/4 but always < 1/3 44/1 ™ 1/20 to 1/80 for tong bridges 4, = 200mm. ‘An inclined fascia beam appears even lighter because of the angle at which ‘the light falls on it (Fig. 17), In addition, the wind is also deflected to ad vantage. In the cage of bridges near residential areas, the fascia beams can be placed higher to act as noige barriers (Fig, 18), Varying inclina- tion cause light-dark effects that give the bridge a sleeker aj With an appropriate design such deep fascia beams can be used as load~ bearing members together with the box-girder. PO Fig. 17 An inclined fascia beam Fig, 18 Fascia beam noise barrier - Great care should be employed with the shape and structural detaile 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 crose- beam mounted onto slender vertical standards placed at intervals of approximately 150 mm. Design Principle 11 Egpecial 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 she 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 Mange which they support. Round plers 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 @ ‘picture” of it. This means making sketches st first and later drawings to scale. Along with a plan view, crots-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 cen 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 impress.on of the structure by choosing fan unrealistic standpoint from which to observe the bridge,but one should design 1 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 algo 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 upan the regional construction materials at hand, the qualifications of the wor- ers, 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 lan economical solution 1s only achieved today ‘f the labour-intensive work is Ii- 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 1s 0 often found in papers are questionable for the most art, 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 reinforee- ment anyway, do not change the total cost of the bridge to any great extent, Lar- ge savings are only achieved if an element of the design is changed, such as a Jess expensive foundation or a less labour intmsive construction method. 12 Design 8 0% 85% 100% rk 5% 5 services] finish work f % % sabete superstructure: a= 7) cay concrete armal FEntorcement 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 (from [14]}: - m® concrete per m” bridge surface area, expressed as an average thickness 045-1 [m] - 9/2 0,98 + SAS 20) 8m? ‘quantity of reinforcement (mass of steel/m* concrete) normal reinforcement ~ 110. ‘kg/m’) prestressing steel for the case of predominantly j Continuous prestressing tendons ~ ( 4,5+ 0, 51m) fxg/m'] ) = average cost of falsework and formwork ~ 60 % of the expenditure for concrete and steel. 4.2 Construction Method: ‘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 cage of a small 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 ‘medjum-length bridges between 200 m and 500 m the incremental launching method and classical balanced cantilever method or cantilever method with Taunching gantry will be predominantly used, Construction Methade 13 Fig. 21 Basic formwork for a box-girder cross. 8) classical balanced cantilever constriction (Rhine Bridge ) b) cantilever construction with auxiliary T Oto ©) cantilever construction with launching gantry ( Siegtal Bridge ) —SS 4) 4) launching girder (Krannenberg Bridge ) basi-4 a q ©) incremental launching method ( ‘Taubertal Bridge ) Fig. 22 Construction methods 14 Design CONSTRUCTION SPAN TENG! TROGE TENGTH TRICTON PROCESS MeTHOo Rowupwwore|| ma, ww imml| yD oo @ ani CANTILEVER. FREUSTEEN LAUNGANG GFOER INCREMENTAL LiGcniNg Fig. 28 Classification according to construction method, span length, total bridge length, and construction progress (from [18]) 4.3 Superatructus 4.8.1 General The box girder often is more advantageous than say a1 um due to = te high bending stifmess combined with a low dead load, yielding a favorable ratio of dead load to live load; = {ts 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 1 -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 epan 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 ength ratio of end span to its neighboring inner span should be Ijp/Iy =. 9.75 but not < 0,40 #0 as to avoid uplift at the end span supports, Most end spans with Ip/iy € 0.8 do not look good. Supereructore 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 yan a7 multiple spans 1/d™ 18 stressed: single spans /d~ 21 multiple spans 1/d= 25 = pre 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 apan lengths over 150m 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, dy, ‘The depth of the box girder should vary in the longitudinal cirection 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: A/a ~ 33 to 50 Vag ~ 12 to 20 Should the bridge be placed high above the valley, aesthetics dictate that, de. ending 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 aid 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 p= 15° (Fig. 24) can be designed a8 if they were orthogonal bridges. For larger 8 the angle of skew must be considered (e.g. (19, 20}). 4 Fig. 24° Skewed brid, Fig, 25 Curved brid, ge i se 18 Design For curved bridges, where in addition to longitudinal bending moments Ma, torsional momenta, Mp,are also necessary to satisfy equilibrium, the box girder cross-section {9 especially advantageous. The angle of curvature,a, (Fig, 25) is the governing criterium for the ratio of Mp/Mp. For c < 30° it is sufficfent to calculate My, as for a straight bridge and Mey for the curved bridge, i.e, to neglect the coupling effect of the two upon each other (e.g. (21, 229); 4.3.3 Transverse Direction Figure 26 portrays the factors influencing the cros portrays are all approximately of the same importance, 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, minimum width of 2.5 m between the hand raila 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 fas 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 600m Fig. 27 Railway clearance Fig, 28 Double-track railway: Place the webe as nearly as possible underneath the pathe of the loads, and neigh of the box girder as well (Fig, 29), The details ofthe slab carry- {ng the tracks depend in adition upon whether th ballast bed lw continued over the bridge or whether tho sleepers are placed on: vbretion-abeosbing pede sitting directly on the structure. me For highway bridges the bridge should maintain the normal highway cross-section (e. g. Fig. 30). Should later wicening of the highway be planned, the additional Ianes 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, Fig. 29. Design of the Metro in Lyon (from (24) ) of a 4-lane Autobahn (R Q 29) 18 Design ‘The box-girder cross-section shown in Fi proven itself with regard to ite form and structural characteristics. Many variations of this "standard ‘cross-section " are possible. 4 Hf eed Fig. 51 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 1/9 of the bridge width gay, ("" roadway slab width), a single-cell box girder is in order, If d/bgisy'< 1/6, a 2-coll or multiple-cell box girder is more sen- sible (Pig. 82f(38)- Por wider bridges the vehicle Zoads acting onthe cntt- lever slab can be distributed longitudinally by means of a pronounced edge ‘beam, enabling the cantilever length to be increased (Fig. 38). jell box girder with Fig, $3 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 emall 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 cells and beyond, For economic reasons, today more than 2 celle 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 Joad distribution (Fig. 35 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 atreases in this direction (Fig. 35 It is then better to separate the two box girders. Superstuctue 19 — 100 18 single foal at A = 100 KN uniformly astruted 180 Fig. 34 Longitudinal bending stresses g at mid-span in relation to the number (example for 1/4" 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, 26), 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 (Pig. 37), The center of gravity is thereby lowered towards the bottom flange, Fig. 36 Kochertal Bridge at Geislingen Fig. 87 Thicker bottom flange for a d-lane Autobahn (from (27) over the support (Pine Valley Creek Bridge, trom (28) 20 ation 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 times 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. 38). cross-section c2oss-secton ‘rer the supports ‘rine spon f-— 02m — Fig. 38 Felsenau Bridge, single- Fig, 39 Oléron Bridge with ver- celled box girder with tical webs and two sepa- inclined webs 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. 1). 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 influenc 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. a) symmetrical widening b) unsymmetrical widening: equal cantilevers achieved through a curved aligument of the box section Fig. 40 Widening of top 1 4) simple design but too ) extensive formwork but in ‘stubby in the normal region return a well-proportioned cantilever slab in the normal region Fig, 41 Extension of the cantilever slab 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 special form for the coup- ing of prestressing tendons Fig. 43 Design aids for the cross-sectional dimensions ~ Tngunen/Jetay © 9-2 (to simplify the formwork, a value of Taunch/leten 7 5 18 recommended) = tylty + from 1:2 to 1:8 2 desien ~ slab thickness t, > 1,/30, otherwise the slab must be stiffened in the compression zone by'means of transverse ribs at intervals of a * 15. (this also applies for t,). = minimum dimensions: deck slab ty = 200 mm; t, = 200 mm webs ty = 300 mm or (200 + 284,54) bottom slab ty # 150 mm (with the utmost of care in construction, e.g. pre= ‘cast elements, these values can be decreased by 5 much as 50 mm). ‘The rato 1}1p shown in Fig. 43 is very much dependent upon the transverse ending stiffncés 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; Kes between 2.0 m and 3.5m. 1; should be so selected that no negative moments in the span occur in the dedk 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 radjus 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 f 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 {n Section 11,9. Complete System and Suppons 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 Ht 1) most common form; _b) for a central eupport _¢) for very stiff with access opening box girders locally thickened webs suffice Fig. 45 Diaphragms over the piers ‘2) small longitudinal displacement __b) very large longitudinal displacement Fig. 48 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 suprorts or piers, and earthquake forces. For very long bridges, longitudinal wind forces should algo be borne in mind as well as the dead-load components cf bridges with longitudinal gra- dients, as long as the bridge bearings are positioned with the same inclination ‘and not horizontally as shown in Fig, $8, 1 . on Shat ated rent Bae 7 snail ym w Wome Reta ns he oe ~~ + feel Vase te Te > = - i | tomes (eauae ot hrsart'nemeaay on™ 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 cage the bridge also needs only one expansion joint (Figs. 47 and 48 a). 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 fleity should lie at a pler 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 a 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 tail-piers, If the final fixed points are not located at the abutments or the bridge construc- ton 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 ). Compute Sytem and Supports 25 tixed point point of tixity| 4, Tr a Fig. 48 Support possibilities in the longitudinal direction tieed wt Ab AL eg ca oe Mo =O WetO5L ort et Fig. 49 Influence of end r+ Fig. 50 Variation of support effective Euler buckling length J conditions during « construction For extremely long bridges it can be more economical to provide intermediate joints in the superstructure and therefore several fixed points, aa the bearings become more complex and costly the heavier the loads they are required to hhandle and the larger the movements they are required to accomodate imul- taneously, i.e, the higher and wider the "trees! of Fig, 51 for the particular bridge. Oo 20 300 0 scm et fa without joint «fixed point i et : eee ——————— p eeeinees soeeeennes ements > load carried by movable support uf tixed support . Imorlithic cnnectontointenance tee) | seperstructure deod loads Tenperature- induced movement Fig. 51 Load vs movement depending upon the support condition in the lon~ gitudinal direction (from [29}) ‘To increase the number of joints because of an unsubstantiated fear of differential settlements to the point that the effect of continuity is ost (Fig. 52) should, however, be rejected, The best detailed expan- sion joint will always be much worse than the continuous superstructure: In addition, the structural reserv =r = i of the continuous system are lost by ‘the multiple-span statically deter- minate 6; Pig. 52. Superfluous arrangement of joints Complete System and Supports 27 ‘The torsionally stiff box girder enables the skewed abutment and pier walls to bbe avoided for the case that the bridge must cross over an obstruction at a skew (Fig, 53). 53 Skewed crossing with a torsionally suff bridge constructed at a right angle ‘Torsionally stiff box-girder bridges built on sharp horizontal curves and sup- ported by at least 9 individual columns located under the middle of the cros: Section are stable even without the bridge ends being fixed (Fig. 64). This permits very transparent support systems, which are especially welcomed in the inner city. Nevertheless, in order to be able to uee the normal expansion Joints, the bridge ends are fixed toreionally stiff attemative ‘alternative fC = Y o I Fig. $4 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 is not to be utilized (Fig. 856), Of course a fixed end as shown in Fig, 53 auffices for short bridges To" tem for long lightly curved box- girder bridges Fig. 85 Support 93 2 bevgn 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 bearing: ‘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). YS Fig, 56 Longitudinal movement in the direction of the bridge axis can be brought about by horizontal restraint forces (trom [9]) For bridges with tranaverse 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. 87 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 f4)) Suberucture 29 4.5 Substructure ‘The loads of the superstructure are carried irto the soil by the following: ~ abutments 4.5.1 Abutmente ‘The abutment provides the transition between the earth embankment and the bridge superstructure, Its wing walls secure :he embankment and its 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 emall bridges posse: 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 mim between it and the soffit of the box girder. 4.5.2 Plere 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-aection (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 80 close to each other. Otherwise a common fourdation should be provided, on piles if necessary, Should it not be necessary to handle any torsion moments 4 single column underneath the middle of the cross-section appears more ele- gent than two (Figs. 59 to e). However, this is only sensible for one or two- celled cross-sections (Fig. 594). By inclining the webs more a emaller bottom slab is thereby possible (Fig. 59). The single columns should be di ‘signed sufficiently robust, as columns that ar¢ too lender do not convey a 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 columns 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, tie aquat shape produced can be Aghtened, —s _t Fig. 59 Pier arrangement -@§ceaf806 = GE a= GED -— ea > Fig. 60 Cross-sectional shapes for piers (from [9] } Fig. 61 Comparison between pier and twin columns for multiple-celled box girders oI 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, 62) or aet flush with the bottom aides 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 ridge’ s appearance. a Fig. 62 Pier shapes Suberverre 31 For high bridges over valleys the piers should also be designed with one or two-celled box cross-sections (Fig. 63). The walle should be at least 200 mm to 300 mm thick in order to permit the use of slipforme or climbing formwork. => aye 3 : : oc; , Lt ‘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 2.0m, 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 sotl conditions be found only ai lower levels, 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 whieh 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 @ pile cap on top of several small individual piles and its complicated reinforcing arrangement (Fig. 65), Ifunfavourable 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 caissont ‘The pier can then either be placed directly on Fig. 65 (a) ie more the bottom of the excavation or on top of the favourable than (b) excavation or caisson wall 32 Design Paleework foundations should be locally widened pier foundations (Fig. 66 b) in order to prevent one-sided settlement (Fig. 66.) and the resulting dange- rrous tilting of the falsework. The falsework should also be statically deter- minate in order to clearly define its structural behaviour, as its stiffnes: Mies between only rough limits, P py Sv me (PX XV] chherwise ona ey ‘os Fig. 66 Careful attention should be pald to the falsework foundation 5, LITERATURE, g cr hc rae ies emai, a 3 "Sheen rsa Snares Bea Inwodeson 33 Part ll STRUCTURAL ANALYSIS (Co-author: Kurt Schafer) 1, 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 ‘ruchires, 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 IIl 2, this manner of structural analysis {s the only method which can simultaneously con- ‘Sider all of the effects to be found in conneetion with box- girder bridges and their interactions with each other. Other more realistic and consistent methods of analysis and determination of the dimensiors and properties of the structure hhave 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 preparstory remarks should be made ‘concerning the structural analysis, especially the computer-aided analysis, a8 follows: = Garbage in---garbage out, A structural anslyais 1s only as good as the structure’ 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 analyais by hhand so as to know approximately what reault to expect. - The amount of computer print-out {8 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 ameunt of refinement in the analysis and dimensioning. ‘The engineer with experience in the analysis of box-girder bridges will per- hhaps see an inconsisteney 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 Str Anais 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 80 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: Fig. 1 Comparison of various loading assumptions (from [1]) 2 fy: normal reinforced cqperete yp" 26 KN/m?, Lightwelght reialoresd coneketd yy c= 161021 kN/me ~ Dead loads of the bridge finishes, go: roadway surfacing (asphalt) y~ 22 kN/m’, Iasela bean Wrelaforeed Concrete) y> 25 KN/m®, bridge railing ~ 0.22 kN/m, guard railing ~ 0.25 - 0.35 kN/m. = Live loads: In most codes a single-vehicle loading with a series of wheel foad8,0,;and 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 11 [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 bearings: Depending upon the maximum Gosificieat of Trfction of the bearing (See the bearing’ s performance approval specifications), horizontal loads act upon the box-girder bottom flange and the piers. Leds and Extra! ntuancee Fig. 2 Notation usd for the determination of wind toad from Figure 3 ee [ea are [es To [without live toad Fig, 3b ; eco Ea 3 fae eae Fea i eee Fa 36 So a ae asthe terol 3 rs ee 2 hee es cy *Fi9 3t & exgewanen) someon | lem ar oO Se oF IS Grecten of eee [region Ce A cantilever Patt Fig. 3 tt Tee. = yt tee [wens tuft = ise. = ire Betton ange gah 7 es 3042 25 ato s 0 Le Se oe om resem Fig. 3 Determination of wind loads (from []) 98 Stuctal Anais With the cross-sectional dimensions (Fig. 2) and the dynamic F 4, obtained from the wind velocity, v [m/sec], according to ‘the formula q (KN/m*] ~ 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 9 g toi, Information concerning the maximum wind velocity can be taken from the applicable Code or obtained from a local weather station. If constant contact if maintained with the weather station during construction the design wind loads for the erection phase ean be re duced, depending upon the expected duration of the particular ereetion 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]) % sanbtya[te(£ -45)/ ee ] where f,, * fundamental natural frequency of Nexural a = SECTION. 2 vibration for the rn tte! shape Hose ars-05b -ommeaat aaa Seeye [as Mibration forthe Critical shape "0.13 “T7101-0.2 ‘Ps Tome | oy rm + mass per unit of length Son > om f+) radius of gyration Fig. 4 Factor a derived from wind be bridge width fom tee a= factor from Fig. 4 terived from wind fone! tate 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, Fyy = 3/20. Lb, but Fy Is at least = 0.3 - ) roadway (#N] (L200 m), ‘heavy vehicle" ~ Temperature: A uniformly distributed heating or cooling of the super- a th 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 can be taken Loads and ExermalInfvenoes 37 from the climatic conditions of the region. In central Western Burope, for example, the temperature varies from? 20K to 50 K from a construction empernture of + 10°C (283 K), Heat of hydration (Pig 8) and shrinkage fn thice pare of the Grove-section, especially between thtek und thn paris ofthe structure and lero of Setwaen young and old concrete, cab ge Pee gnt Five fo large localised revidial stressee Fig. 8 Large tomperatere which exzoed the tension strength of oung gradient due to the comcrate, Cement’ s unfavourable Ag a good crack distribution in young heat of hydration conorete cannot be achieved, even with development and low a finely distributed reinforcement, crack- temperature of previous- ing can enly be reduced or prevenied by ly conereted section cooling the aggregates and/or the water (trom [5)) used oF by insulating or heating the forms, ‘The insulation of the conerete should be maintained in position from 2 to 4 day 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 tresses 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, ean 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 a'so [8] and (8] for more detailed information). As such a nonlinear temperature distribution is not very euitable for a practical analysis, a linear one is often chosen, For central western Europe AT = 10 to 15 K appears appropriate. ee ee — d by v0 Tastee aT awe A b Fig. 7 Magnitude of the restraint F Blase moments due to temperature, Fig, 6 Temperature gradients used Mf, in relation to the size of o perature gr in dimensioning from [7] the load acting at the same time m8 Suet Anse ‘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 II. That is why an additional restraint moment acting simultaneously with a large loading i much emaller 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 T condition, Tt 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 Il.) 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 aubstruc- ture in the form of restraint stresses, PLAN. ~ Tov ii aT Ee ayar g week e Fig. 8 Horizontal movements 2.6: Should one analyze and dimension a box- ble differential settlement in the State I condition, 1s 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 sottlement 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). H necessary, the bridge bearings should be able to be suitably adjusted. It is false, though, to select statically determinate single spans in order to avold these restraint stresses, as the increased number of joints means a ‘greater possibility of trouble with the durability of the bridge. - Earthquakes: Only for very small and rigid bridges does it make sense to ffact of an earthquake into account by means of a horizontal equi- valent force that is proportional to the bridge mass, For large bridges & realistic dynamic analysis is required. (See for example [11] to (16).) ~ Impact: Instead of dimensioning the piers for the very high impact loads {fori 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. Seuctral Sytem 39 3, STRUCTURAL SYSTEM 3.1 Final State ‘The box-girder superstructure can be part of the most varied of systems (Fig. 1, 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 monolithieally connected (see Fig. I, 48), ‘The stif mess of the box girder, Elg, is generally much larger ‘than that of the pier, Elp, When BIp/1# 30Eip/1p the superstructure can be analyzed separately from the substructure a8 @ continuous beam on freely rotating aupports (Fig. 9). fi} ror the structural analysis of the piers, onan the movement of te pros = ted touions oie spertcusare st & Bclconidereds evn if te eupersiractare {Stings or feet move ete neon Fig. 9 Sutnsce of the frame ithe ple lee Sen 9 omere ‘The type of support also affects the structural analysis. A beam with more than one span with inner supports along ita 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 @ simplified analysis of the system subjected to horizontal wind loade 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. 108), which can be additionally supported by cables at critical loading stages (Fig, 10), becomes a one-column or multiple column framework upon the closing “TT ““TIIt Fig. 10. The transformation of the structural system with the progr: cantilever construction 40 Structural Anais of the span in the middle (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 construetion can also alter the structural system (Fig, 1, 80), 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 frTfTyFt 1 Cantilever method with erection gantry ) Incremental launching method 5 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 posaible combinations in the final state and during construction should be considered. Except for the vehicle loadings on the superstructure, this requires no further explanation. 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. 126) at the 1/10 points and at other eritical positions (Fig, 126) for vehicle loads, For the calculation of the critical sectional bending moments, My, and thelr associated V,, and Mz: in view of the flexural dimensioning, the bridge is loaded over its ditire width 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,Mp, should be carried out both for the maximum shear, with the associated MZ’ and My for the section, and for the maximum torsional moment due to a dhe-sided loading with its associated My and Vy. ‘ical Leedings and Sectonl Force 41 in nee BH oe Ago ty Mt Twat, Biv, mon Mp a ee ey wom SHH 2) Longltudinat sectional ») The minimum of sectonal forces forces required for te dimensioning y TTT loo pxsion br LEE -p armani eee ‘aver son ey Ty longitudinal tocing (See Fi 15) coon ~ a em My load position forthe = = Pr tases shear force Wy flaca | ee Ine fr the eyes shear eee Vy tierce tine for he ‘rsociated manent B SSS hae Be — a “hissed minent ty Fe Bee Pa tte te te tp 28a lead poston tribe = EP beset monet BE a Set nto line or My cra 1 7 4 ae ©) Influence lines and loading positions Fig. 12 Determination of sectional fore for vehicle loadings 42 Sovctrl Anais 4.2 Tran jerse Direction In general it is sufficient to determine the critical 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 Ife in the longitudinal axis of the bridge and ngt over the webs or even parallel to them. (eer Moy t 12 3605 ee m6y 5 Lr may %% May May Mp, when te 19 bottom flange is odes 9% b) section 0) PLAN | Fig. 13 Sectional forces required for the dimensioning of the flanges, aS ee eS mor MTA My min me MON REIN Mp. MEX MyM MEX Np min mg mn me wn a eae Fig. 14 Critical loadings {in general 5. SIMPLIFIED STRUCTURAL ANALYSIS OF THE SUPERSTRUCTURE 5.1 General ‘The box girder representa 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 Simpled Srctual Anat of he Supaetcture 8 [As the last 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 8.5, With this approach, the box girder loaded symmetrically in the trans. verse direction can be analyzed longitudinally as a beam and transversely as ‘8 frame independent of each other. ‘This is valid if the box girder’ s slender- ness ratio, 1/4 © 4, if the span lengths, 1 1,5. b (Fig, 15), and if the disturbance zones af 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 rapidiy 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 also 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 ca 5.2 Analysis Procedure ‘The analysis presupposes that the statical system, the loading, and the design trove accent duvensions are mows, aeyzay op) © ape & 4 fe cp TE ae 5 5 4) The loading of the top flange so as to) 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 A Str! Anais Firat the box girder ie analyzed in the longitudinal direction asa beam, i.e. with the assumption of a rigid cross-section (Section 5.3). ‘All tranaverse 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 momenta of the entire cross-section are determined with the eccentricities, yx, of the load resultant, gp. Re Atterwards 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. 162 1). Fb 4 th "ep tonge 22) Only edge loads se es py ; : UES 353 prep b2) Antisymmetical port of load of"oos ‘engitudinal bending: beam Wronsverse bending: frome omy tarbuld Sa a . Qe Fag % 7 “Eh oP Dae or Ser Or Fabs yal getoomy or mi cilfire torsion 2) aul on GE ° Selon rcs i “EPs 411 Stttened frome ‘only transverse bending longitudinal “ond transverse bending Fig, 16 Splitting up of the structural system and the loading of @ box girder, All arrows drawn represent external loads, The cantilevers are partially omitted for greater clarity. ‘Sinplied Smuctural Anaya of he Suparrucie 45 ‘The resultant reactions, qq and gg, and the fixed-end moments, mg and mp, t at the edges at A or B, respectively (Fig. 16.22), They are divided into & 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 beading 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 longitudinel 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 @ transverse strip, one would obtain results far removed from reality, for the idealized frame with edge 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 analysie of a frame strip whose edges are held fixed by the diagonal force, S (Fig, 13d 1), The resulting transverse bending moments, after their superposition with those from Figures 16a 1 and D1, generally Iie 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 antlsymmetrical loading part.This is achieved by no longer foreing the shape of the eros 1642), 5. 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 6.3 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 case 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 vieinity 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 cro section (Section 5.3; also Fig, 12), yielding M,, V,, Mp and o, 4 Sect Anche ~ 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 6.4; Fig, 27), yielding (m, m, “symm, ~ rigid frame analysis of the antisymmetrical edge loads (Fig. 28; Section 5.4), yielding (m, m, Wantisymm, a8 Well as S ~ superposition of the sectional forces in the transve: a, and v ¢ direction, yielding m, ~ analysis of the elastically bedded web panel loaded with qg due to the diagonal force, 8, (Section 5.5) yielding longitudinal stresses, oy and transverse bending moments, mg, (Figs. 39 and 28e, respectively) ~ seperpostion of gp with og and mag with m,yllding he asta ongitidioat tive. and transverse atrésses, rZspect ‘Frame analysis yields only transverse bending. "The edges do not displace with respect to their positions from the beam analysis ten core Tore eas Frame analysis yields edge displacements to- gether with the trans- verse bending (shown for unrestrained edg b) Eccentric loading (torsional deformation not shown) Fig. 17 Connected beam and frame action ‘Simpliied Structural Analysis of he Superevuctoe 47 5.8 Analysis in the Longitudinal Direction 5.8.1 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 Joads, and to about $5 % by dead loads, With increasing span lengths the in- ‘fluence of the vehicle loads and thereby the necessity of their exact analye! 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) ean 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 bbe 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 rotaticnal capacity, the restraint has no influence on the ultimate load-carrying capacity (see Sections 1 and III, 2). 5.8.2 Time-Dependent Alteration of the Sectional Forces by Creep Sectional forces due to external loads are only altered by ereep if components Acting together within the structure contain concrete of varying ages oF varying composition. This is indeed the case for segmental construction (ig. 18), For 2 sufiefenty exact ealculation of the final time-dependent moments, the follow Ing equation taken from [18] can be used: (ees algo [10] and [20)) : =I My, (tp EM): ot ggg 9 Me EMG #y ZG) Tepe ———— > _ where My 12,3, moments at the time of stripping of the == fatsework of the differ- Mes ent segments, 1.2, moments due to self- weight plus prestres: « eam moments assuming that — the entire bridge i 3 constructed at te same point in time KAN Ges ereep factor at times e. relaxation factor taken from Figure 19 ¢ Oe Fig. 18 Development of the bending moments from the construetion stages 4 strc Anse If the time of loading, tg, {.e. the time from the concreting of the section until it {s preatreased, ‘ie lees than 10 days and the construction rythm used {in producing the individual segments, At, is less than 21 days, then the valué produced by thie approximation do not vary from the exact values by more than 15 %, In consideration of the unavoidable inaccuracies of the analysis jeumptions and the fact that the results do not influence the safety of the structure, this is perfectly sufficient for the normal cas Should t, or at, respectively, deviate grossly from the above values, the Limiting values of the moment redistribution at time t = ©© can be estimated a Tite, 8 FESS = hg hehe le ~ 20h + HK ahs yes for Mgeit,) oF Ky ond P, tor 2 tft fr the reget o oan f consiyetion rere reapechivlyafe teh ftom Figure 18 with the standard gonsrgtion seen eeprom Fi tgs ; a 12] 10° ost yy sce] g “4 a SEP re CS Sa LORE RT Fig. 19 Factors for determining the effect of creep in the analysis (from{i8)) ‘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 prestress as a result of shrinkage and creep can be calculated with sufficient accuracy a a uniform reduction in prestress,AP_,.(t),up to time t 2 p. Saltitel Es + XP, toy + Spo) DR alt) BE ck Sipe [1* PlELIZZ] * Ble and gg (t,t,)... strain due to shrinkage within the time span (t - t,) {Tue values are dependent upon the humiaity of thai, ete, and lie between ~ 0,1 % and - 0.5 % (see [66]) modulus of elasticity for prestressing s! respectively land concrete, ‘Simpliec Structural Anais ofthe Supestuctre 48 P04,4,) «+ creep factor (the values are dependent upon the time - span (t~t,), the humidity of the air, etc, and lie bet- ween i andl 4, see (66]) Cogs ‘stress in the concrete at the position of the tendon due e to self-weight and other dead loads (compression negative) Topo «+ initial stress in the coxcrete at the position of the tendon due to pr {initial stre: tress alone (compression negative) {in the tendon due to prestre alone Sectional restraint forces due to differential ssttlement, temperature, and shrinkage are greatly altered by creep even without cracking. Suddenly occurring restraint stresses (0, g. due to differential settlement in eandy soils) are great- ly reduced (Fig. 20). According to [21] , the elastically calculated moment Me in State I is reduced to ea Yu ED MM) = Mae A~ 405 aay) peseeeees eecameeter ee Eases ee ey T ‘that of creep (e. g. settlement in for p= 0,85. clays, shrinkage) the moment that lusden] Gar 2005 ya produced instead of Mey ts only lt) orm, ME Mee FRA ‘That is why it is false to reverse the effects of a pier settlement that has occurred slowly by suddenly adjusting the bridge support to the full magnitude of the settlement, + 0 wo 0 >= t Fig. 20 Creep-induced change of restraint moments result- ing from support settle- ment 5.4 Analysis in the Tre 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 appreximately 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 Sect! Anais land then to redistribute the total moments acting at the to the frame's stiffness (see Section 5.4.2). White the sectional forces due to dead loads, prestress, temperature, and tniform vehicle loadings over large areas can be determined as acting on flange strips (or also acting directly on the frame), influence surfaces must Generally firet be evaluated for concentrated vehicie loads (Pig. 210). Wheel loads can be distributed under 45° to the middle plane of the flange (Fig. 21a) tnd the uniform loads can be integrated sufficiently accurately to concentrated Toads as shown in Figure 21b. In[ 29]. the influence wurfaces are supplied in table form for the majority of sectional forces, whereby the limiting moment dlisgram can be estimated ts shown in Figure 22. ‘The tranaverse 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: 029-8 |-0.20. oe ef -q 50h ven 4) Whee! load distribution ») Position of the distributed Toad resultants (from [22]) ei eens ©) An intuence surface (example rom [23] and it evaluation Fig, 21. Determination of the sectional forces due to vehicle loadings from influence surfac Simpliies Structural Anais of the Supereructre 51 35-034 035-4 Obl rar FA "hymn Lael + ™y bb tl ae he ee Lt wes: Sas tstoa mr at dat, ) M-limit envelopes tor concentrated lads focting on cantilevered partons of the ferge I M-tiit eveopes —c)M-lmitervelopes for fer cowentrted kads conceded loads ‘cating on flange sips acting on forge strips the tens. Verse chphrogin re- ion seed sport Fig. 22. Moment envelopes (from [24]) oat ® mt oe ps | 21, arn one Fig. 28. Influence of the cantilever design on the moment diagram (from [25] and [26]) $2 Stetina Anas 5.4.2 Analysis asa Frame Every arbitrary loading can be separated {nto @ symmetrical and an antisymmetrical part. On the element of length dx * 1 taken from ; the shear force differences, AT’, (and the longitudinal stress differences, ae-pivgenaTce trical stp tate Oy, Which are of no further interest here) maintain the external loading, q(x), in equi- a srnal Loading, q(x), in equ By $, __—=—sbrium (Fig, 24a), For eymmetrical loads =i (Fig, 248), where Phos Gam QerdK ; Te et + VS/Te the following holds true qr. S dvs. dx" dx * i 9e where with dx =I x te Fo ees aT FG Symmetrical moments atthe juncture of the ange with the box, mp /?2, (Fig. 16b1) produce no shear forces For antisymmetrical loads as in Figure 24, the following holds true ar Geb ein ZAK Me first moment of area of the partial half-cross-section about the centroidal x-axis ‘moment of inertia of the entire eros: kern area A, = d+ (by +b,)/2 (Fig. 28) yetion ‘The sectional forces acting on the frame must be calculated for the simultaneous ‘application of the external loads and dT” ne moments of inertia, comet tact transverse Tyrtelbs+2bJ72 L812 Get bo/e Este p Letubie Lets (continued next side ) Fig. 25 Notation for the longitudinal and trangverse direction Simpled Structural Anais of he Superrructre auxiliary values Ane bes nef gl RF a siz 4, It % % a ° >) Fig. 26 Redundant forces (a) for symmetry, (b) for antisymmetry 8 1 Leading For symmetrical loadings the frame is statically indeterminate to the 254 ogres (Fig. 26a), because no shear Is poseihe inthe plane of symmetry. Figures 279 and portray the moment end internal axial force diagrams enulting from the loads tad moments, respectively, acting at the rigidly fixed junctures of the box with ihe Mango, for aush & cross-section as shown in Figure 25.in addition, Figure 2Te portraye the moment and saial force diagrams resulting from an increase in the length of the top lange,®lg, over the bottom ange, This allows the possibilty of considering for example, a uniform increase in temperature of the top flange (T',) over the bottom flange (Ty) through Bl, = (T= Te Or cn orig of tay Zane ne ornare preset, E z i should be meno here fat, shod at bo maid wih a ree factor for the statically indetermingte analysis as long as the creep conditions are approximately the same throughout the entire cross-section. Aly Due to the effect of the forces resulting from the curvature of the longitudinal prestressing tendons (see [27}) upon the wets, shortening |A1,| of the webs occurs, which with the value 41, = 2 |a1,| - sin , can be converted to the case of Figure 27¢, This is only necessary tc 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, mp /2, acting on the joints of the lower flange with the webs such as those morfients arising from self-weight of the lower flange and webs, One need only turn the cross-section upside down and change the subseripts in the formulae from denoting the top flange to denoting the bottom flange. 54 Several Anais eal me : oa ee gs cuniiry vole bp25424-25% — famioe tread ease a no +4g' Sine erate onan dhawd + adie Ns coer, me «ah taon fev of eed % root Feetads) oe 8 nae suns Fig. 27 Effect of the symmetrical part of the loading on a frame: a)Nodal loads v)Nodal moments ¢)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 f 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. For antisymmetrical loading the moments and normal forces are zero in the planewf 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 meane of the antieymmetrical diagonal forces $/2 as shown in Fig. 266 . ‘The sectional forces acting on the freely deformable frame, which is statically indeterminate to the first degree, are given in Figures 28 ¢ 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 ¢ are needed. ‘Simpled Struc Analysis of the Supersructoe $5 otaeurs0zap 410033 1 1p uo syuaurour TePON(G owresy » uo syuowour TepoN(P puE SPU’ nme spBot TePON(e -OUre] otf UO SuNPYOT oxy Jo ied TeoRNOMUKSTUE OM Jo OHI ve “TL 5°§ uours 235 § 20+ Seg2 fe werd poptoy 0 onp g sa0u0.9(9 :sopou PON(® :sopou poxis Atprars wr sures] devs Sa Vie wrote goths s SoM ego A One oew ge Cestue ve os ture, a Sega enone Bye Pigee . “i © $3080a_WiKw S_S3040s TWNOOWIG W SINBAON ASUBNSNVEL 56 Suc! Anais 5.4,3 Transverse Bending Moments in the Haunch of Variable-Depth Girders In the bottom flange of haunched box- 4 girder bridges (Fig. 29), forces due to Py 22 Tis _9 curvature, a= 0, ty/R, directed towards Ba He rani oahu ai ene are T '1-proy Fig, 29 Loading of the bottom flange of the haunch Bay aby asco dae compression [omax 05: “10 oo Os opp tay Fig. 30 Pectora®., son: 6 real gj om tnegative moments, These forces ereste tranoverse bending moments which act Syuinst the moment due to self-welght of the flanges Aecording's (26) te trenses, 0, ara not union dietributed Acrove the Bore girder width, by, whan the hauneh ie sharply curved bu af concen trated in the areae near tho webe. The fctermination ofthe transverse bending moments therefore emplaye a mean stress Suan instead of Using Pigure 20, the 13iwing te then obained «~ effective width of the bottom flange Palisa = 1, is the moment of inertia and 2,, the distance between the neutral avis of the section and the mid-depth of the =yb, flange, both for the idealized cro sectional values with bus ~ mean stress at mid-depth of the flange Henge | Nopi-Alonoe oan * My 24; /yy (State 1) } _ - transverse bending moments at the edge and in the middle respectively, due to non uniform distribution of 4. ou [28] also handles the influence on the moments of elastic fixed-end conditions in case of deep or thin webs at the bottom flange. well as the influence of transverse haunches ‘Simolfed Sructural Anais of he Superstctue 57 5.5 Folded Plate Action ‘Among the various methods of analysis (see in(28] and (30]), the one employ- ing the analogy of a beam on an elastic foundation is the most auitable for the actual practice. ‘This method {s therefore surmarily 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 ascumptions: = The lengthe 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 eymmetric about ite vertical axis, A constant cro: section is maintained throughout the length of the girder (Fig, 28). = 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 effecte plus the antisymmetrical one! ‘a8 far ag 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, 8, which were determined ‘a8 shown in Figures 28a and b as acting on the stiffened frame become the external londs with reversed signs on the bor girder. ge S bal 9e 7 S bsg rar qu 5 bly Fig. 81 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 dg: and q, acting in the planes of the plates (Fig. 31), the webs and flanges deform not only in the longitudinal direction (Nexural 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 crose-section (distortional deformation) (see also Section 5.2), The load, S(x), is there- fore partly carried in longitudinal bending as in @ simple beam, SM(x), and partly by the resistance of the frame to distortional deformation of the cro section, 8 (x) Six) = Ste) + Sexy 9.0) = Ga + Fs) (1) ‘58 Seti Ansys ‘Baye shear forces, and 1, consibute to the carrying ofthe Toad of he individual plates in the longitidinal direction, q. These forces, as the shear Streose in girder of # planar section, aro diatriboted in proportion tothe Shean, WG)y eto the tongs (Pig. a8). he sonditon tht the longsudinl Inslec etreoses, of adjacnt plates their jnctares must be of egal mibgaitadeyiclas te idalicol moment of neva forthe wab an ZB Ll hot2Z (hur) ~ 1] Li = Heherspcdee ea) Ts and places the neutral axis (Fig. 33) at o> Sea Bs Y= Os ~Ye therewith the Longitudinal stré determined sses in the web resulting from gy can be for a regular girder: M(x Mo, : Ge == BOE ye Sous Rs Ya 18.2) BL 9 ‘ot fenetty fea, Tr Loading ef @ wed pone in its plone Ve Gly dearom Fig, 32 Loading in the plane Fig, 33. Stresses o, due to S 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, ¥, due to qt can be calculated from the differ- ential equation of an ie teh as folldwe: +, a, ety th = - hh = g? (53) ‘Simoliied Sructural Anahi ofthe Supestuctre 69 Gyo And Ogy, in the longitudinal direction, x, have distribut- fons proportional to one another (oxo/aqy * ~Yo/Yys Fig. 32), the curvature and deflections derived from them for allof the plates must alzo be proportional, to each other (0.8. vo/¥g = ~ 2¥g/bq). For that reason the diagonal deformat= ton of the edges, § (3), can be exbrodsed by means of the geometrical relations of Section 6,1 in terms of the deflection of the web as follows: ANeZA+Z Ar tor uf, YOK) where bee PANE a tas al) (4) ‘The resistance, 5, of the frame to such a diagonal deformation, 6, can thereby be determined by calculating Sdue to a diagonal force, 3, acting on a frame element of length dx=1.’The component of § in the plane of the web is of the form = bs 7 HGER |. 20202 84% +n hE Jorg Ss Glide be 6 where hes Rep Rae or, if 6 from Eq. 5.4 is used: Jatn =K- yon) wore K=PE be kekeEl (5.5) From Eqs. 5.1, 5.3 and 8.5 the differential equation of a beam on an elastic foundation is obtained for the web EI: A) + K yoo = goo 66) 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 hovever not necessary, because the distribution of the sectional forces in the planes of the flanges are proportional to those of the web. ‘The solution, vg (x), yields the ongitudinal moments Mex) 2 HEI yeh (37) the stresses,c, (from Eq, 5.2), and the transverse moments, m (x),as shown in Figure 28e for the analysis of the deformable frame with the loading Sak Be yoo. 0 stuctural Anshse 5.5.2 Solution by Means of the Analogy of a Beam on an Elastic Foundation ‘The solution, vp (x), sought for the web must not only be a function that satis- fies the differential equation (5..6) but algo one that fulfills the boundary con ditions of the structure being analyzed, For example, a simply supported beam with transverae dlaphragme at the supports can be represented as the system shown in Figure 34.8, ‘The basis for the solution of such a system ia the infinitely long beam on an elastic foundation (Fig. 340). It is subjected to al Loads of the form, g(x) =S(x)-bg/q » which as shown in Figure 16 1, arise from the resistance to the diagonal deformation, where $ ie derived as shown in Figures 28a and. Moreover, the beam is ionded with unknown Additional forces, Qy and@z, and additional moments, My and Mg, at the supports which are determisied from the support conditions. Fig. 34 Systems for the beam on an elastic foundation: a)in reality ) basic eyatem with equivalent loading For example, for the beam in Figure 34a at the supports, x*0 and x°1, the defections, v,,and moments, M,, due to the loads, additional forces, Q,,Q2,and moments, i, My, are required to be zero, For a continuous beam with diaphragms over the innér supports, the condition of vg = 0 at these supports requires an Unknown additional foree, @, but no moment, at each of the inner supports in the basic system, Diaphragms in the span also compel vg * 0 at that point and the corresponding unknown additional force @. Should the warping of the cross. section be completely hindered at some point, the condition of dvg/dx = 0 Applies there, which can be achieved by means of an additional mément at that point. ‘The influence of the resistance to warping by transverse diaphragms find 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, @,, in the plane of the webs for the beam on an elastic foundation and not a support condition, LT 4 Fig. 35 Loading at a support without transverse stiffening Simplfed Structural Anais of he Suparatroctie 61 ‘The solution to the differential equation (5,6) takes the form of strongly damped vibrations as follows: Aa) =€™. (cosAx + sindx) Ba=e™. sin Ax Cx) =e (cosax ~ sin dx) Done. cos rx where and L.= 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 thelr original values at a distance of 3 L to 4 L 0 that loads or support conditions for distances greater than 4 L from the point under consideration can be neglected. For the same reasons, transverse stiffening i.e, diaphragms to prevent distorticnal deformations in the span is. only worthwhile, if at all, if such stiffening is placed at intervals © 2 L. ‘The sectional forces and deformation: are handled in detail in the Literature such a8 [33], Loadinge as shown in Figure 34 produce the following: yor Pie [OrAa + QeAU-R) + QeAUK-aall} + oBeletBan + MBa-n} + +B fea [Dea,-x) ~ Dray? a,-x)]* ¢[2-Dx-a,) ~Dea,ea,-x)] + +6,[Dix-a,-a,) ~ Dix-ael]} where cnt y= C570 for 0#xe a, Gat HOO 2X5 ata, Gt =O 90x SL Mor Fe [OrCon) + Qe Ctx) * Oy Ceix-anl)} + +FDO + MDa} + +Biel-clBra-n -Biawa,-x] +[Box-a,) * Baya, -x)] - ~G[Btx-a,-a,) ~ Box-ay)]} 2 scl et Mors fa, Dar + ay Dix) +c QeDIK-asp} + eEMAm + MyAd-n} + fe tel Cea-n1- Ciay+a,-n] + ¢,[Cox-a,) ~ Ccayea,-x)] ~ ~e4[Cox-a4-04) ~Crn-a,)}} tnd for x#0, 5 Cye-4 for x*ay ‘The unknown additional forces and momé san be obtained by plugging in the boundary conditions given, or from the f ring formulae (see also (33) My = Be-Car — Bee [Bray ~Bea,ray] My = Be Cea.) +e [Blay+a) ~ Beay)] ve 2 Aca +-fie [Dan ~ Diaya,] Mp = BR Awan) - B- Draeay) ~Diay)] f a a XL canal casa) %* L cosnal = os = Ae{Z (Mot Meo) BU - 42° Elsi (Yo* eo) H + DIY] } Q=-Rreef2 (Mio Meo) BU) +42 Eli (eo Veol[t - Dtti] } Me ree fUMo* Mo) [A *AlN]*+ 22 Elsi Vw * Voll + CNT} MP=~2eef Mio” Mee) -AUy]- 20° Elsi (Uo ~ Yell 4 ~ Ce00]} Q = a +Q” = Q’-a” My =P em MeMt-M? ‘implied Stctural Anat of he Suparetctie 63 ‘The transverse moments obtained from the folded plate consideration are Proportional to the web deflection, vg. The influence lines for vg take the form of Figure 36 in the transverse direction and in the longitudfaal direction 1are proportional to the elastic curve of the besm on an elastic foundation sub Jected to a point load at the section under investigation (6, * b,j). In order to determine the critical positions of the loadings, one, however’: rather seeks orientation from the known shapes of the influence aurfaces 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, @ 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 stresses for the folded plate (fp) action due to one characteristic loading position only (Fig. 36). A ratio is ther determined between these stresses and the longitudinal stresses for a symmetrical loading of the same magnitude. This factor, fyp, 1s used to increase all eccentric loads for the ‘analysis in the longitudinal direction, where oj... stress due to the force S from the antisymmetrical 4/2 (folded plate action) = stress due to the sym- metrical g/2 | Fig. 36 Influence line 7 for vg, where the position of max) is dependent upon For normal box-girder cross-sections, f¢ Hes between 1.1 and 1.9 for uni- form loads (main lane; q) a8 in Fig. 15) nd between 1.9 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. ot Suuctu! Analisis 6 Multiple-Cell Box Gir 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, 4.3.8 comments are made con- cerning the design of multiple-cell box girders. As multiple-cell box girders are only constructed in special eases, the design engineer is here referred to the Isterature [29] [l, 26) as well as Section 6.2 {or further information concerning their analysis. 5,71 Curved and Skew Box-Girder Bridges In horizontally curved bridges the longitudinal bending moments, Mp, and the toreional moments, My, mutually influence each other, Some sectional forces gle-span girdér are compiled in Figure 37, For special supporting span girders, the reader is referred to (25) and [84 to36) . fora systema and multipl According to [38], two-cell box girders can be roughly analyzed by using factore for the longitudinal and transverse bending moments, the shear forces, ‘and the deformations taken from an {dealized straight bridge. Skew box-girder bridges are in principle not sensible, as in most cases the irder can instead be supported perpendicular to its longitudinal axes by affording a minimal increase in length (Fig. 1, 8), [87] contains a method of analyzing a skew box-girder bridge. (=aeen bse Jefe Ts aime b] fe | Be fee] fe | oe Eee] Ts ps ate she ity my 4 vA that the section under investigation is at a sufficient distance from the end diaphragms = that a maximum of two plat or diaphragms moet at an edge. ‘The loading is expanded in the form of a Fourier series (Fig. 39) as follows: gin =Z gmsinax where ae = A pnevaa pezaysn OE. 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 Suet Anais 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 eystem (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 asaumied to bt 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 Equa by ‘This results in Toad in the plane of the plate, n, of Jn GE Sin 0X = Sunes ~Snegn ] Fig. 41 Resolving the toad Fig. 40 Hinged folded plate subjected into its components Me, 40 Hinged flded pate su pasar As the members cannot deform independently of each other along their mutual edges, edge shear forces, T, and, ,, are necessary, Requiring equi- Librium on an element, dx, df the plate, n (Fig. 42) yields n= - Fn dx ; ANq=(In- Tree) AX a, My= Made LUT *Treallondx — " Integrating, one obtains Ya = Varcos ax ; Na=Na-sinax (6.2) My= Maesinax 5 Th = Tat cosax with the fixed values T," and fare Web-Gt 5) Niet (i-th) Fig. 42. Sectional forces : (63) Inthe plane of the . “meer Bameceee Eee = fate Ta) Rigen ‘Comptes Anahi of the Superstuctre 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 Nn-. 7 From this plus Eq obtained as follo Bicett bee alee) aes = ty cathe) This yields one equation per edge: {.e. for a single-cell box girder, four ‘equations for its four edge shear forces, and with Eqs. 6,2 and 6.9 the sectional forces, M,, N,, and V,, for each plate. The deflections in the planes of the plates follow ffom the ‘differential equation of the beam by substituting Yq" Wain @x a8 follows: ‘, . e145 My Tyne which yields PREY RARE ree ee er ae whe eptee— 19h Ci + Tots): 5%] ‘The edge deformations, w, perpendicular to the plane of the plate can be determined as shown in Figure 43. 6.2 and 6,8 the "three-shear equation” for edge, n, is EI, veo? sin ax 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,, = YX sina x, are applied at the joints (Fig. 44), ‘The associated support redttion® of the plates AQ, = Mey Ya a Struct! Anais create the additional plate edge loads Gn =~ AQntcot far Cot Yoo) — From this one can determine the moments, M,, the normal forces, N,, the shear fofces,V,,, the edge shear'Yorces, T,, and the defornations, vq and wpyas shown above, They are all dependent upon ¥,, (or n= 1104), ‘The 4 determining equations for Y, to are obtained from the condition that eveity angle between two adjoining plates dos not change. ‘The rotations in the hinges due to transverse bending of the plates are determined as for a beam, The load due to self-weight, dn, yields for example (Fig. 45a) Uninet = Onstn = FEB —sin 2 Tue redundast forces yield (Fg. 48) Yaynee = EBB (2Y + Yoos) Ins and oS VWootn “EEE (2 Yous * Yn) ‘The angle produced by the deformation of the plate loaded in ite plane (Fig. 45¢) is bn Fig. 45 Angle of rotation at the edges With the above, the compatibility condition for edge, n, takes the form +h, Onna t Onna * Yang * May Computer sided Anais of the Superstuctre 69 ‘By substituting the values determined above, all moments Y, all sectional forees, and all deformations can be determined, Ag a closed summation of the Fourier series 's not possible, separate calculat- ions for several periods (for m= 1,2,3.....) must be carried out in order to Achieve gufficient 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 [48]. 6.2 Finite Stripe, Analysis by means of finite strips lies somewiere 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 longita- dinal direction by periodic functions (Fourier Series) and in the traneverse direction by simple polynomial functions (44, 45] ~ For long girders, often only a few strips are necessary. For short girders ‘and fer 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 te 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- coll (44, 45, 46] or curved [48] box-girder bridges, while for single-cell straight bridges an analysis as described in Seetions 5.2 to 5.5 is sufficient in most cases, 6.3 Finite Elemente With the finite clement method, the two-dimensional partial differential equations are reduced by discretization to a two-dimensional system of algebraic linear equations. The structure is civided into elements that are only connected at their nodes (Fig. 47). 70 Swuctr! Anais Fig. 47 Finite Element analytical model (trom [ 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, aR: Bes Ne May GO Bee: Membrone Forces Slab Forces Fig. 48 Positive internal forces and Fig. 49 Nodal point forees, 8, displacements in a finite ‘and displacements, v, for ‘clement (from [49] ) a finite element (trom (49) oa I. » 9 f 3 f > Fig, 50 Triangular elements (a)with $ nodes, and (b)with 6 nodes for refined analysis (from [50]) Along with rectangular elements, predominantly triangular elements are used (Fig. 50). In order to avoid numerical diffiulties, triangular elements should not be given too small an angle. Parallelogram-shaped elements have been. developed for skew bridges. A survey ie given in (50) Recently the development of 2- or 3-dimensional finite elements capable of simulating the actual noalinear behaviour of reinforced concrete after the formation of cracks has made considerable progress. [81] and (52] contain examples of this. Figure 51 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 1a heavily loaded, Comtuter sided Anahi ofthe Superstuctre 71 r co e--to- A stttenes 8 unstienes a) Failure mechanism of an») Loading and stiffening at md-epan untiffened box girder ‘Load P [MN] t M0) 08. 1 1 — FT ve oro ton od] se 7 ost—/ ; ; mee uncer the 6 ont isee ‘ lost @ A | Ete 20 0 S a = @ 01 02 G3 O4 OS 06 O7 Detiee tion tmmi tose PIM) — box girder A stttenee A= box girder Ursitened ©) Relation of the deflection _) Relation ofthe angle of rotation to the to the load at mid-span toad at mid-spen Fig. 61 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, Tt 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 condition: 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 element programe, especial emphasis should be placed upon the ease of use of auch programs. ‘This trans- lates into the requirement that the gridwork ie automatically generated for some regions 72 Sirs Anshse 7, ABUTMENTS ‘The abutment is subjected to the following loads (Fig. 52): = dead load, g ~ support forces from the superstructure, Fy, Fy, F = earth pressure from the backfill eyy: ‘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 4s therefore difficult to completely consider. It is therefore recommend fed to use the earth pressure at rest in the calculations ~ earth pressure due to the working load on the backfill, @g Pig. $2 Loads on an abutement = braking forces on the backfill, ey = working load on the abutment, e.g. footpaths, 4 = lateral impact, Qy = other loads [80] 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 due 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 Pe © pressure of moving water for piers in water (Fp) Cp 2+ v7 - A, where the drag coefficient, Cy ~ 0,8 to 2 from tables, v= velocity of the water, and A= cross-sectional area of the pler to the flow) ~ ice pressure [60, 61] = impact from ship hitting a pier [62] arise moments due to ~ geometrical nonlinearity (2 order theory) = physical nonlinearity (nonlinearity of the M-curvature-relationship for Feinforced concrete) = horizontal force from the deformation of the system (Fig. 58) = unplanned eccentricity (construction inaccuractes) 7 foundation rotation Al, que to | The influence of the theory of 2°" order depends Temperature upon the slendemness ratio of the pier and can be neglected for De agg lt< 25 where s.yg-++ effective length (buckling length) Fig. 68 i... radius of gyration ‘A horizontal force, H, arises from the inclination of the ‘column hinged at both ends which acts on the fixed pier 2 Bffective Length In order to determine the slenderness ratio, \, the effective length, Sep, 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 Sorr/ dy® 15, where dy is the width of the pier, and pinned supports the free movdment at the top df the pier can be assumed to be limited [63) .Im (64) , the influence on the pier moments of the manner in which the clasiomeric 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] ), me uy. ' Ft Tey OR if ‘ i a, kL \r MA LAG \ \! ‘ Fig. 54 Comparison of the deformations and effective lengths of a column rigidly fixed in the soil and one elastically fixed fan "8 whee2= Ba for practicel case 2A Seva Anais 9.8 Moments According to 2™4 Order Theory ‘The determination of the moments on the deformed system cannot be separated trom the dimensioning of the structure, ‘The ratio of reinforcement influences ‘the deformation by way of the stifiress and thereby influences the moments, If tables [65] cannot be used in which the ratio of reinforcement is already in- cluded, the moments must be determined by iteration, An approximate method from (65) i8 shown below aa follows: «determination ofthe moments according to the 1" order theory, mt a Gig. 58) with consideration of an initial eccentricity (e.g. according to SIA" 162 [60] 0,02m™ u, = 2/300) = rough dimensioning of the pier, choice of ratio of reinforcement = determination of the effective stiffness, By, from the M-curvature- relationship (Fig. 57), depending upon the Feduced cube compressive strength of the concrete, fy/ yyy, and the reduced yield strength of the Steel fyx/tgqa. 88 Well as the Liking strain, eq yyrqe And ‘which are based upon the particular safety concept uss ~ calculation of My due to the Y, and yp factored loads and the resulting deformations Bal ~ Bilal Na $y [lbw He -2e)] timit? a Mi = nr 14 Wile * eegCyNa FOBy where: N- internal axial force (compression negative) ye BiB safety factors, e.g. according to the CEB (EG) Yo tod with Y) = Y. for the unfavourable effect and vy foro fevourae eect of the ssial MM, factor fc def tion due to the - wt th actor for creep deformation due tothe creep cuxt t Producing moment, M,; for A My, the ealeulation must be repeated, The values for My, ky, and By deporting upon the axial force, N and the ratio of reinforedined pcan baltacen from tables (e.g. (69, 67] for rectangular and sireular cross-sections, determined for Gy aug s 8% aad eg iyeyes = 239) ‘The tension stiffening provided to the reinforcement by the concrete betwoen the cracks, was not considered above, ‘This ieade to an overestimation ofthe moments, MII. ‘Therefore for very high plers a more exact calculation should be made with consideration of the mean steel strain cy, 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 1, 66). In order to be able to follow the flow of forces from the Joadé on the scaffolding into the soil a statically determinate model should be used in the calculation if possible ncause of the uncertain stif ness distribution. 7 Strvctrl Ansise 10, LITERATURE, 2 neu SEY SunDESEAl! VoP0SCin Pamainbrastn ad song lngenibawere VE Yornaae (BY 84 4 HORRE"a' Berit inr dn Sta er Avtar ea Rapeln sar inanung er Windebng mu Deore Rostra 5 ppchtnhielaa ane itt cane nea at ams 2 RANy, "Peepers ce eres ene age ie Sucre Regn: Ve. #6, and 1 PRUE 3 Sesto, FUCHS: sobre ster Tampereereaennnict + Ri a wen gn ct, ea pr 3, Dine or a MasrthahesSpige Vong. 1 Suman naman rie: TAR, Semen Rap; repr an Rare ae WP 12/18; TARE! W978 - = = ve talk Bac: seer othr er oe nar rg ty aati “ vere ME GLE US" 2 Ea rng sr preg ni rt to: et Sabrina 8/90, aici Se ere rll dog, Zor Aumsng rom Erne Smnbttraweri: Btn: od Sabet SS Utara vorenmgusnscneigt berccmupereanies far Mestbcoccet oases cn (ners Wasa Goa Ci, C8 nay ces arena dt J8D, Yl eT 8 ee are ee re a eee en mney es epee ae eee ee eee ice acm Sa te PE, z rmpuet progres for prints flee lets with plate oa eum secant, Report oot 262 cin, SR. Deroy of he fae scp mt nthe ase edge dca: p. $3 ‘Sin ig. cain, Va at feeb te tae Pot ri Ma agen 35, ScOnDHL Ac, OAM, RSet nme oer a rt be Va, AC Pantene sancyni Ze ‘on Spent Reaiorarerboem,Desachr hase fa ee Pro w Seen, E WINDERLICA, WF Heme nS SatheSriagW, ot Se, i RemaAuayl Spring Varig, Dern «Ne Yr 8 Sanaa Re Rar econ po Bt be gr MoE SCORDEES 4." eprint mt aero reg: it a, "8: Dyan fares operand pn: Cn una Chl Bahra, Ya. 3: 1896 sn, Ri SOMA SR mods vegan to aa el ag on rw. 24/-8, as. Banorg fr Peele - ier Lae ae Benge er se; Bese Irodveton / Gener Detiing Prncinas 79 Part ill DIMENSIONING AND DETAILING 1, INTRODUCTION In this part the dimensioning and detailing of tox-ginder bridges is handled. ‘This includes also those areas whose stresses cannot adequately be described with the normal bending theory as assumed in part Il, Further appropriate prestressing tendon profiles and reinforcement arrangements are suggested, land 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 thelr serviceability and durability. In this publication intended for the international forum of IABSE,only the CEB/FIP ‘Model Code in its latest 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 ieft 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: {.e, the assumption of the uncracked State | independent of the magnitude of the stresses (see Part II of this paper) end the dimensioning of the cri- tical sections so that the behaviour of the member with regard to cracking, deformation, and vibration is satisfactory ander working load or service oad 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 II 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, 4 légl IE, M | Fe eg leoml. ae Fig. 1 Moment - curvature relationship 9 Dimensioning and Detaing = 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 expecially 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, 20), The sections must, however, provide a sufficiently large, rotational capacity. Box girders with their ro- bust slabe satisfy this requirement in most cases with sufficient mild steel reinforcement, whereby overly reinforced box girders (p> 2%) are ex- Cluded for economie and structural reasons anyhow. Although the ultimate load design methods for beam members and plates are today in principle so thoroughly developed that they can be used, sufficiently simple methods for box-girder bridges that can be comprehensively applied fand that take every effect into account are still lacking. This is especially true for such cases as eccentric loadings or bridges with curved alignment. Up until now this has only been met by a structural analysis according to the ela- tie theory and a dimensioning based upon the load-carrying capacity of the a8 oe co ey ae Mo ‘plas moments. autism cit owe, ry Tat MeM, |b 9 = =| ea 4 eat a md Bolan) a8) Fig. 2. Example of the determination of the ultimate load in bending with the ultimate load design method (failure in shear excluded) ) for comparison: taking the moments determined according to the elastic theory and dimensioning the critical sections satisfies the lower boundQ, 4, =Q, © state metnod "1 jecting the correct bending moment diagram €) kinematic method) 9 fay, as they are reduced at time t= © by the dead load, gp, plus the live load, 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 euspended from both sides by ables however, this favourable influence reduces the need in most cages for 8 box-girder section to that for s deck slab stiffened by transverse and longitu- dinal ribs, — tension === compression 4 COTERINE ALES EE, ttre SE Fig. 8 Transverse tension in the flanges of cable-st ayed bridges 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 {t must handle vehicle loads, concentrated transfer of forces (see Section 3.4), or large transverse tensile stresses due to lateral distor- tions irduced 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 ether two, they are located con- Fig. 9 Prestressing tendon profile centrically. in a heavily loaded flange 3.4 Tendon 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 bottom flanges if at all possible (see Section 3, 3,2), as they greatly weaken the slender webs (Section 8.3.1). 3.3.1 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 cage of bridges with a curved alignment, the outside web is given @ 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. OG Na f NF Sebo ach tr iperete 1 coo Fig. 10 Prestressing tendons in the web {88 Dimonsionng and Detaing ‘The anchorages of the tendons should be spaced as widely apart as their anchor cones permit at the bridge ends (Fig. 10) and should be spread as far as possi- ble at the coupling points (Figs. 10b and 228), ‘The radius R of the prestressing tendons over the inner supports should on the one hand be as small as possible in order to convey the forces due to profile curvature directly into the support. On the other hand, this produces a rapid reduction in the longitudinal direction of the distance # from the compressive force in the conerete to the line of action of the tension foree, the inner lever arm. With a large radius however, tensile forces are produced in the lower fibre by the prestressing. ‘The draped longitudinal prestressing tendons should be complemented if possible by straight longitudinal tendons that amply overlap the points of zero bending ‘moment but in any case with mild steel reinforcement that extend either at least 1,5 d beyond the calculated points of zero bending moment or if possible over ‘the inner support. In case the tendons are deflected up at the end of the auper- structure, the end support reaction must be carefully and safely diverted into the inclined compressive strut of the web by mild steel reinforcement duly ex- tended over the support (see also Fig. 24a). For the design of the prestressing tendon profile, its influence on the statically indeterminate bending moment M', as well as on M,, = M°+M!, are shown in Fig. 11 for a 2-span beam, poe factors Mp em) tendon profile reaive| cmt or an] Mp [terest Rim! foes Joey fey | kN x04 130. P 267 |010| 04]08 | o70-P | +092 P a5 -P saa |o20| 04/06 | oss -P | | 0.98 | vere 10.0 [005 Joas Jos | 1,01 -P ~0,80-P eo | = | - | - |n8R-088-r, Fig. 11 Influence on M!, of the prestressing tendon profile for a 2-span beam with 1/d = 18; PP positives Py, P Prosvessng 89 For those tondons placed in the web, the minimum allowable spacing of Fig. 12 applies as long as no other conditions are made for the particular tendons employed, ‘The prestressing ducts must be properly secured euch that a sufficiently large space is left for the placing and compacting of the concrete. id it possibile, eaten words. sccur Underneath tw cts Lo Ha ‘min 60me ninsdaim or 6 min Somm oF ys: Detter 15-3,4, better 1584 Fig. 12 Minimum spacings for ducts (from [10], approx. as in [1]) As the grout possesses a smaller compressive strength than the concrete and as the duct is unfortunately often incompletely filled with the grout. the reduc- ton in the effective section due to the dicts should be considered tn order to avoid failure of the web in compression, ‘The reduced width of the web fe, for ungrouted ducts, Dyed = bo = Pquct 8nd, for fully grouted ducts, Dyag* be 0.5 Fauct [trom (1}}. With a duct diameter of fgyct = 0.2 web width, he higher ocdlized compressive stresses remain within Seceptable limits in the rest of the concrete cross-section according to] (Fig. 138). According to [1], Sauce should be lesa than 0,125 " boy The forced flow of the web compreasive stresses around the duct causes trShaverse tensile stresses which can lead (0 cracks in the concrete (Fig. 196), Especially ifthe tendons are sharply bent in addition, stirrups of reinforcement mate (Fig. 14) may be necessary to avoid splitting of the concrete, ungruted et ie fay rote tet fh rons ‘ese tele ‘SHresses ore Smoller) compression 8) compressive stresses (according to the) transverse tensile stresses elastic theory) in the rest of the cross: Gun to an ungrouted dice Section, aa related to the ratio of re~ dition of the concrete eection of the web a/b (from [9)) Fig. 18 Stresses in a web with duct 20 Dimensonng and Detailing ieene 7 Ui rr Bie — tension compression the’ encon Fig. 14 Danger of concrete splitting along tendons in the web In order to avoid greatly weakened cross-sections due to ducts, the prestres- ‘sing tendons can be concentrated along the sides of the webs and deflected at vertical ribs. This method will remain limited to very large bridges, however, It has the disadvantage that only after conereting the entire bridge can the brid~ ge be prestressed [10] . 3.3.2 In the Tbp and Bottom Flanges In very large box-girder bridges the longitudinal prestressing tendons are pre- ferably placed exclusively in the top and bottom flanges in order to keep the width of the webs small. This also applies to the construction phases of such construction methods as the cantilever method and the ineremental launching ‘method. In the cage of cantilever construction, the tendons run straight through (Fig. 16) and the cross-section is increased in thickness at the coupling points, (CROSS-SECTION LONGITUDINAL SECTION ih DeTAN. 1 PLAN ‘ep flonge Poy attention to the forces. veto Profile curvature (transverse tension)! Fig. 15 Prestressing tendon profiles in the top and bottom flanges for the cantilever method (from [11] Prsvesing 9 PLAN bottom tlange cerns conatrctifn joint Fig. 16 Prestressing tendons in the bottom flange of an incrementally laun- ched bridge; similar details for the top Mange (from (10)) If some tendons are left unbonded in the construction phase, this should be considered in the design. For this reason, it is important that the tendons be carefully grouted at the completion of construction, Without grouting there is no bond between the tendons and the concrete, and the load-carrying characteristics are altered, ‘The transfer of concentrated forces influences only a particular zone whose dimensions can be estimated according to the principle of de St. Vénant. It is to be recommended that a strut model with tension and compression member be developed for this zone which is oriented te correspond with the lines of principal stresses according to the elastic thesry [12,13] as determined for the particular statical and geometrical boundary conditions, Using this model and satisfying equilibrium, the magnitude of the forces in its members can be determined. The mild steel reinforcement for the tensile forces determined in this manner should be amply selected and finely distributed in order to limit tho strains to be expected. This is economically acceptable since this rein- forcement comprises only a small portion of tne entire reinforcement required. itis, however, important that the reinforcement be placed in the proper posi. tion. Figures 17 and 18 portray two examples, Details of this method and other examples can be found in [14] . tn the case of anchorages within the superstructure as in Figure 18, or a8 in Fig. 7, Detail forthe transverse prestressingtendont of fee ton Ranger te not necessary to design the hanger reinforcement forthe elaatieali eater Inted forces (o-g. Zp, 2g in Fg. 18). The rellsribution ofthe lnsoy forces {rom the not tensiod melabors to ine sift com preseion snomiere ney fe eaten Into account here, In[ 19} is recommended tats pr of Fesfoneesent be Added to the reinforcement atthe anchorage wih A, =0.26 P/allow oo, hie Gan be reduced forthe case f favourable compressive siressss oy préduces by contimous prestressing tendons o red A, = (0,25. P-5-a5 Aanchorage)/ How. og $82 Dimansioning and Dealing ta i pay Flag sre bet 3 SI" » koa eS sl “= senercene seco Fig. 18 Intermediate anchorages (from (14]) Prosvessing 89 Cracking can only be avoided if compressive sirgases of OG” 1 to 2 N/mm’ afe produced by continuous prestressing tendons Fig. 19 Inner anchorages of prestressing tendons 3.5 Construction Joints and Coupling Joints For bridges with a total length greater than about 30m, the majority of con- struction methods require construction joints, While they allow efficient sec- tion-by-seetion construction, they are, however, weak points, The concrete tension strength ie reduced at the joints, and the heat of hydration of the fresh concrete in the newer section causes additional stresses in the concrete. As most construction methods foresee the anchoring or coupling of prestressing, tendons at the joints, the concrete cross-section is further weakened by the anchorage. Therefore, reinforcement should be especially carefully arranged {in this region. In order to compensate for the reduced tensile strength at the Joint, closely-spaced longitudinal reinforcement (g~ 0.7 %, in thin flanges g~ 1%, bar spacing <100mm) should be placed as connection reinforcement, in the newer section, reinforcement is required parallel to the joint to take the tensile stresses due to shrinkage, The tensile stresses due to the heat of hy- ration, especially for thicker members,can oaly be kept small through mea sures that keep the heat of hydration within reasonable limits, Reinforcement is useless here because of the lack of tensile strength of fresh concrete. No tensile stresses should be allowed at the joints of segmentally constructed Superstructures in spite of the use of joint adhesives capable of taking tension. ‘This method is not proper concrete construction because of the lack of rein- forcement at the joints and should be limited to special ca If prestressing tendons are anchored at the construction joints, the con erete of the following adjoining section is hindered from following the defor- ‘mations due to creep of the previous section because of the high compressive stresses behind the anchorage [16]. Tensile stresses therefore arise adjacent to and behind the anchorage, which can lead to cracking if these stresses are not counteracted by the continuous prestresaing. Compressive stresses arise for equilibrium reasons, but somewhat removed from the anchorage (Fig. 20) ‘Therefore, reinforcement must be placed near the anchorage so as to produce ‘a small crack spacing. 91 Dimansioning and Dealing section old new pogdnd 4) Joint deformations due to ‘creep after the prestressing of the old section due to (a) Fig, 20 Construction joint with an anchored prestressing tendon ‘The coupling joint before the prestressing of the coupled tendon behaves of course exactly as the construction joint with an anchored tendon. After pre- Stressing the coupled tendon, though, the conditions are reversed, Near to the tendons, increased compressive stresses occur, and only at a distance can tension stresses arise for which reinforcement should be provided, These ten- sion stresses remain extremely small if the coupler is separated untouched by f gasket from the following section. The full prestress is then carried through the coupling joint and not simply half of it, as would occur if the gasket were not included (Fig. 21). The nonuniform distribution of compression stresses {in the coupling joint and in turn the boundary tension str ‘smaller the smaller the ratio b/a is, It is therefore advantageous to spread out the tendons 4n the region of the joint as shown in Fig. 22a, The longitudinal cracks due to the concentration of tendons is thereby also avoided (Fig. 22b, also [17]. Of course it is still better to avoid coupling all of the tendons at a single joint. _—<—> section gechon II concrete, section It Zo Prosressed fee of sttewes os” posiressed final Bate long as section I ol prestressad, ust ose hecose with onstrction ints a) Construction sequence b)Coupling concreted e)Coupling with compressible gasket Fig. 21. Longitudinal stresses in a coupling joint, portrayed on an end-loaded plate with one tendon Dimensoning ofthe Top and Bottom Fanges 85 Senet 1) Spreading the coupled tendons ») Tendon concentration leads over the entire depth of the web to cracking Fig. 22 Coupling joints with tendons in the webs LONGTUDINAL SECTION (CROSS-SECTION 6-8 ef SECTION AA ETI ~ od TREAT. Fig. 28 Coupling of tendons threaded through the web after hardening of the concrete for the incremental Inuncking method: eriontal Hesitto the thickness of the anchorage block because of the incr formwork traveler (from {103} ‘Should the tendons be required only at the end of the construction stage and should they therefore be threaded through at completion of the superstructure, then inside anchorage blocks are needed at the position of overlap of the ten- dons (Fig. 28), 4, DIMENSIONING OF THE TOP AND BOTTOM FLANGES ‘The principat reinforcement is shown in Fig. 7. Normally the transverse rein- forcement is placed orthagonal to the longitudinal reinforcement, In the case of skewed bridges or those sharply curved in the horizontal, it can be more econo mical to place the main reinforcement along the lines of the principal tensile stresses. The amount of reinforcement required when deviating from the direc- {ion of the principal tensile stresses can be determined according to [18] orf 19). In designing the flange between the webs for partial loading in the longitadinal direction, it can be advantageous to make use of the fact that the flange supports itself through arching action against the cantilevered portion of the flange, 98 Dimorsloning and Detng thereby creating compression stresses in the tranaverse direction [20]. Even if one does not exploit the arching effect, the cantilevered portions should be Provided with robust edge reinforcement at their ends (Fig. 7, Detail a) ‘The connection of the top or bottom flange with the web should be determined from equilibrium conditions by means of a complete strut model which includes the webs a well (Fig, 24) With the transverse bending moments,m normally occurring in the flange simultaneously, its design can be based upon Section 5. 1, Fig. 27. Fig. 24 Flow of forces (strut model near the support of a box-girder bridge: from [21] [14] ) Dimensoning and Reinforcement of the Webs 7 5, DIMENSIONING AND REINFORCEMENT OF THE WEBS 5.1 Dimensioning for Shear, Torsion and Transverse Bending ‘The webs of a box girder must not only handle shear but also torsion and trans- Yerse bending, According to[1],, the "refined method" should be used to cal. culate the tension and compression strut forces occurring in the webs due ¢© combined shear and torsion. The inclinations ©) (between 30° and 60°) of the web compression struts of the truss with para‘lel chords and the inclinations a (80° in most cases), of the web tension ties, i.e, the etirrups,are thereby sclec- ted and their forces determined (Fig. 25). In checking the web compression stresses, the cross-sectional reduction of the webs by the prestressing ducts ‘Should be considered, especially for slender webs (see Section 3.3. 1). iy aera \-* nee A fers we cl or singe ross spre enolate) | Seen “Weegee eompresson chrs [0 B= ieatB-cata! |lcarpesion cee tersion chars Heb ietg-ctay | 2 BL late-ctal a “Segre oar tension toces Taine “2 SeZateorOvatal || eomsoe ithe webs 7; 2; ’ _ Eines anal nee ata [beso ames or) Fig. 28 Definitions and forces in a parallel chord truss for bending and shear {In the case of a simultaneous combination with a torsional moment, Mp, the equations of Fig. 25 should contain instead of V the value V*= V+T, with T= My 2/(2 Ag) If, as in the case of bridges constructed with the cantilever method, the com- pression chord is inclined, smaller tensile forces are produced for the deter- ‘mination of the stirrups; but larger forces occur in the tension chord (Fig. 26). Up until now the stirrups were designed for Vert = V - M/z- tan and the larger tensile force in the chord was not considered, as it was safely covered by the usual staggering rules. To be consistent, however, the tension chord forces of Fig. 26 should be considered, Fig, 26 Forces for the case of an inclined compression chord (from (221) ‘The design of the web for transverse bending has up until now normally been carried out separately for shear and for torsion,and the reinforcement for each hhas been fully superimposed. However, by projecting the compression mem- bers out of the plane of the truss, the trusa can take a certain amount of trans- verge bending without additional reinforcement, This, however, leads to high conerete compression stres: begin. 8 in the contained compression member at web ‘)Boc- girder wed loaded bY ransverse bending Interaction lRetererce wi) W ond Inte wed. generat” diagram moment cose Mae Fig. 27 Design for shear V and transverse bending m (trom (23]) In [28] an interaction diagram with simplified assumptions 1s introduced (Figs, 27D and c) that allows a design for combined m and + . For the given shear stresst and ¢ = 0,5,(Bq = Bg,), the largest transverse bending moment that can be handled without additional reinforcement is given below. ‘This equa- ‘tion is also valid for ¢> 0.5 (the selected reinforcement distribution, where G= Bg/R and R « total force taken by the stirrups = By + Br, = Zstirrups)- BE L1-T/tmax +25 -1)-B/b] Dimansionng and Renforcoment ofthe Webs 98 If for large transverse bending moments (¢+1! the actual moment m, ig greater than m~ BGR/s , the web width must be increased Fatlure in compression is excluded if Ap) + fy< + (b= 2_Web Reinforcement ‘The principal reinforcement is shown in Fig, 7, For details see Fig. 28 below. a) Upper region of the web b) Position of bearings (see also Fig, 7, Detail c) and reinforcement in the lower region of the web Fig. 28 Web reinforcement details In order to avoid spalling at the support, the bottom flange should be designed as shown in Fig, 28 b. ‘The stirrups are mostly vertical. Only for extremely large shear or torsion are they inclined,or the webs are prestressed (Fig. 9). ELEVATION (oROSS-SECTION beorng inclined sterups oxean Foyer vetealstrapstl Zloyerigtu- byes sina rs rl Te. hocks placed logitudnal inclined Fig. 29 Special case of inclined stirrups and vertical reinforcement for large shear (from [10} ) 100 Dimansioning and Dealing 6, TRANSVERSE DIAPHRAGMS ‘The reactions at the supports of box-girder bridges cannot be directly intro- duced into the webs in most cases and can also not be carried by means of transverse bending of the bottom flange (Fig. 30). In addition, often room bet- ween the bridge bearings must be provided for hydraulic jacks needed to re- place the bearings at a later time, The transverse diaphragm, which is favou- able in stabilizing the cross-section under torsion anyway, can be used in solving this problem as well, The forces in the diaphragm can be simply de- termined with the truss analogy, whereby the transition between the truss of the diaphragm and that of the web (Fig, 24) must be maintained without a break (Figs. 81 to 95). With sufficient frame stiffness the toraional moment can be handled without @ diaphragm Fig. 30 Box girder without a transverse diaphragm with the bearings directly underneath the webs al Forces acting longitudinally Strut model 1 €) Strut modet 2 Fig. 31. Strut modele in the region of the transverse disphragm Fig, 82. Strut models for single columns Tronsvere iaphagms 101 =e eee Fig, 83. Forces at the abutment 11 f“ccen soe ! TA SS ee oe Fig. 34 Transverse dlaphragm reinforcement a) Side support b) Middle support, (Lahn-Bridge Limburg) Fig. 35 Detailing the hang-up reinforcement 102 Dimeasioning and Detaiing 7, ABUTMENTS, PIERS AND FOUNDATIONS ‘The structural design of these bridge elements poses no special problems as arule. They are reinforced with mild steel in most cases. By looking ahead while in the design phase, the abutment block-outs and connecting reinforce ment to the roadway deck slab and pedestrian slabs can be properly planned {in the construction sequence, For the piers and columns the following points require consideration: - Allow for sufficient space on the pier cap for the placement of jacks to replace the bridge bearings. = Carefully follow the transfer of forces from the superstructure into the pier cap (Fig. 36) and from the base of the pler into the foundation, - Provide for sufficient reinforcement, especially horizontal reinforcement, {in order to keep the cracks small that result from temperature restraint stresses. ~ In earthquake regions connect the piers, superstructure, and foundations together so that they act as a unit under tensile forees. - Check the consistence of the vertical reinforcement for the piers with the calculation assumptions (constant amounts continuous throughout the pier or stepped reinforcement). poe tension SS compression Ly 2 pe Fig. 36 Force transfer into the pier ‘The following applies to the foundation = Select ample foundation slab thicknesses in order to avold stirrups or bent-up reinforcement if posaible. Avoid requiring too many layers of reinforcement, With several layers of reinforcement and a large moment gradient the failure of the bond bet- ‘ween concrete and steel is a real danger. ~ For pile foundations strive to achieve a continuous force flow from the pier into the piles (see Fig. 1, 65 a). Bridge Bearnge 108 BRIDGE BEARINGS 8.1 Bearing Types Aside from the bridge expansion joints (see Section 9), the bridge bearings are the most sensitive elements of the entire bridge, Therefore only tested and approved bearings may be used. ‘They must be carefully selected according to the magnitude of the anticipated loads and the required rotation and movement to be accommodated by them. ‘They can be of steel (Fig. 37) in the form of a roller bearing or of an elastomeric polymer in the form of a plain elastomer pad or a steel and elastomer laminated pad bearing (Fig. 87) or in the form of a pot bearing (Fig. 37e), an extensively used form today. Not to be for- gotten is the concrete hinge (Fig. 37d), whick is completely maintenance-free ‘and long-lasting, Bridge bearings are comprehensively handled in { 24) . ft T 7 _ neon tang org nop omngh Fig, 37 Examples of bridge bearing typ 8.2 Installation and Maintenance: For the installation, inspection, and maintenance of the bearings the following points should be kept in mind: ~ Bearings function properly only if installed accurately as to elevation, incli- nation, and orientation (Fig. 38). ~ Plug instead of grout the gap between the bearing and the support = The bearings must be well protected from sorrosion. Use non-rusting bolts. = The bearings should be accessible and capable of being replaced, 104 Dimensining and Detaing clamping ut SS W ed ISS A] Willa bearing plate damp, moctar a es Ep— og. honmered eect g] to pace tev S tat sheath ttt get out Sih Shotts tha cocreted Fig. 38 Adjust the bearing on concreted bolts with the set nuts and then hammer in the damp mortar plug (from { 10) ) 8.3 Design of Bearings ‘The bridge bearings must handle vertical and horizontal forces, rotation, and movement. In addition to the vertical forces acting on the bearings from the superstructure and the horizontal forces due to wind and vehicle braking, con- sideration must especially be given to fixed bearings with regard to forces re~ sulting from inclined columns (Fig. 39). bya ue to femperoture Fig. 39 Forces on bearings due to inclined columns ‘The expected bearing movements and rotations must be determined very care- fully in conjunction with the construction sequence (Fig. 40) and should always be increased by an appropriate factor of safety. ‘The movements depend upon the distance between the bearing under consideration and the fixed point, The ‘causes of such movements can be summarized as follows: ~ shrinkage ¢,~ 0.2% ~ prestress; elastic portion: plastic portion: me. > temperature + 35°C, proceeding from a construction temperature of + 10°C (depen- ding upon the country) eqn ep: OT with aT op Bdge Bearings 106 @ © © © © Yixed point at ena of | ae —Q vor rome rca ese, —@ swore wa tinea ie ease ction - construction Secton 1°" construction Seehion van BO: Sar Fig. 49 Sequence of bearing movements at support C and the determination of the expected movement; Bridge constructed in sections with & ‘change in the position of the fixed point (from [25]) For a bridge with curved alignment, the direction of the movements due to the individual influences varies (Fig. 41). In order to simplify the bearing and expansion Joint details, the movements should only be allowed to occur tan- gentially to the bridge, whereby the free movements of Fig. 41 must be eance. Jed by the restraint stress resultants (see Fig. 1, 56). Fig. 41 es eRe i we icone ae Ne |8 ete trata “ Tap tal Te ae oa Curved bridge movements (from (26] } 105. dimensioning and Dtaing 9, BRIDGE FINISHES ‘The bridge expansion joints, the roadway surfacing, the guard rails, and the drainage system should be carefully designed and constructed, as they consi derably influence the quality and life of a bridge. Seemingly minor defect: such ag an improperly sealed roadway surface, can lead to extensive damages .@, corrosion of the reinforcement in this example. 9.1 Expansion Join Expansion joints are necessary where longitudinal and gradient differences bet- ween bridge sections or between the bridge and the abutment must be compen- sated. They are the most important wearing components of the bridge and the most perceptible mark of quality for the driver of a vehicle, That is why cost should not be a factor of consideration here. Because of the heavy loadings en- countered by them, the following points should be observed: = The joint should be built robust and as nearly watertight as possible, = The best and most durable should be selected. = The joint should be carefully mounted, = The joint should be installed exactly on the level with the roadway surface, Fig, 42 shows an expansion joint example for each of the cat of small, me- ‘jum, and large movement accomodation. tothe) gasp | Fmmne_obuiment Permanent AMEE FSGS we ake 8) Joint for small movements (up to 30 mm) abutment 350 min EBL ox ) Joint for medium movements (30 to 180 mm) ©) Joint for very large movements (up to 3.000 mm) Fig, 42 Expansion joints (from [ 10] ) ‘Bdge Fishes 107 9.2 Bridge Railing ‘The design of the bridge guard railing as well as the necessary distance bet- woon the roadway, guard rail, and railing are normally standardized and vary from country to country (Fig. 42). | eng opt Fig, 43. Bridge railing and roadway surfacing (German standard detail, from (10] or [27}) 8.8 Roadway Surface Particular problems in bridge construction, encountered mainly in countries ‘which frequently employ road galt, are the sealing of the top alab against cor- rosion and the roadway surfacing, On the one hand, the capillary water in the concrete must be able to escape, and on the olher hand, aggressive water ‘must not be allowed to reach the structural concrete, That is the reason for placing a sealing membrane on top of the structural concrete (Fig. 43). The roadway surfacing serving as protection for the sealant, should be as durable, strong, and impermeable as possible. The wearing surface must algo be non skidding and abrasion-resistant. Asphalt and bituminous concrete have pro- ven themselves as materials to be recommended (Fig. 43). Concrete surfa- cing is only suitable for countries not experiescing frost and only with @ cover to the top reinforcement of at least 40 mm and a reinforcement that limits the crack widths to 0.1 mm. 9.4 Drainage Sufficient transverse gradient should be provided (see Section I, 4.3.3) so that the rainwater flows naturally towards the sides of the roadway and is carried away at an appropriate distance from the piers and abutments (Fig, 44), In the case that vehicle traffic areas are located underneath the bridge, the water should be carried away through the piers or abutments into the drainage net- work as shown in Fig. 45 and 46 according to a carefully designed drainage plan, Under no circumstances may the water be allowed to collect in the box- girder, ‘The necessary number of drain inlets depends upon the longitudinal gradient of the bridge, The collection area for a normal drain inlet with a pipe of 150 mm diameter should be smaller than 400 m®, Intervals are thereby arrived at of from § m (0.2 % longitudinal gradient) to @ maximum of 25 m (1 % gra- dient), The longitudinal gradient of the drain must be at least 5%. 108 Dimensioning and Detaiing ‘> 330 mm Fig. 44 Bridge drain inlet with Fig. 45 Bridge drain inlet with vertical drain and disperser side drain Fig. 46 Drainage within the box girder 10, LITERATURE Scenjrp soa cate : af anarcedsncrae eam: TSE Cotten Ropes, tony Structural Engineering Documents Objective Information forthe precesing eng: nate through documents of Woh ‘cette and een iva, oF ine ‘engineering themes Fields covered ftroctural analyse construction methods = deion = preeution ster of construction eines and techoology interaction betwen stroctural on: Sreingand ober esa Readeshia Engines troughout the world interartadin elo

You might also like