City Research Online

Abstract

In view of its flexural stiffness and its torsional rigidity, the hollow box-girder has become today the common type of section for steel as well as reinforced and prestressed concrete bridges. In the case of long span steel bridges it has replaced the traditional type built up of main girders, cross beams and stringers as it forms a coherent entity capable of resisting forces, bending moments and torque without any bracings. Moreover, its uniform shape is more appealing from the aesthetic point of view. The box girder is suitable for use as a simple or continuous beam as well as the deck system for suspension and cable stayed bridges. In this latter case stiffening girders are not required. It is also suitable for the cross section of arch bridges.

For narrow roads a single cell box section is sufficient, on the other hand for wider bridges two and more cells are necessary.

Some designers, however, prefer separate simple cells to multicell sections. Needless to say, the height of the box girder can be varied if necessary, as in the case of the old types of plate girders and trusses. From the point of view of prefabrication the box section is very suitable as it possesses both flexural and torsional stiffness. Consequently it can be built of relatively large segments of adequate strength to resist their own weight during transportation and erection.

The box girder is eminently suited for bridges erected by the cantelever method which has thus become the common method of erection today due to the fact that it does not impede navigation of surface traffic and
does not require obstructive scaffolding.

The thesis is concerned with a theoretical and experimental investigation into the structural behaviour of box bridges. A review of existing feyestications indicated that though there was a wealth of information about the elastic behaviour of such bridges, the information on post elastic behaviour up to complete collapse was somewhat limited. Moreover, the majority of investigations had been restricted to single cell bridges which were in the main simply supported.

Apart from some preliminary investigations of single cell bridges this investigation is concerned with the structural behaviour of twin cell bridges both simply supported and continuous over two spans through the elastic range and up to complete collapse. As this work formed part of a more extensive investigation of multi cell box bridges in general being carried out at The City University by Professor J.E. Gibson, a standard size for the cell dimensions of the models was chosen so as to minimise cost and time of construction.

Part of the main investigation by Professor Gibson was to determine the accuracy with which micro concrete bridge models were able to predict the structural behaviour of the prototype. To this end a four cell box
bridge continuous over two spans that had been tested by Professor Scordelis at Berkeley was selected as the prototype as the experimental results and the range of test loads were fully detailed. The width of the Scordelis bridge was 12'0" and its total span was 72'0", the model being at quarter scale was thus 3'0" in width and 18'0" in total span. The geometry of the model cross section was identical to that of the prototype and consisted of four cells with side cantilevers. Steel templates to this cross sectional shape were constructed and formed the main framing for the shuttering for the model. This standard cross section is shown in Figure

This standard shuttering was used throughout this research as it could be used to form not only four cell cross sections but also single, double or treble over any reasonable span. The reinforcement consisted of weld mesh and could therefore be varied within the range of specifications supplied by the manufacturers.

In the present work the templates and shuttering were used to construct single and double cell boxes and the investigation served in a sense aS a run up to that for the four cell continuous bridge.

The theoretical investigation of the structural behaviour of single and double cell structures in the elastic range has been fully covered in the last fifteen years by many authors. The methods employed for linear elastic behaviour consisted in the main of beam, folded plate and shell theory in the continuum methods and more recently finite element, finite strip and finite difference. Reference to these will be given later.

It was thus decided to concentrate in the main on attempting to derive upper and lower bound solutions for the collapse of the structures and to correlate these with experimental values. The upper bound being defined as that load which would cause the structure continuing excessive deformation, i.e. it would be incapable of withstanding any further load. In this field the work of Spence and Morley was adapted and extended to cover the multi cell case.

The lower bound was defined for the purpose of this investigation as the load which initiated first yielding of the steel at some point in the structure. This proved a much more elusive quantity to determine, and this will
again be examined in more detail.

Publication Type:	Thesis (Doctoral)
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) T Technology > TG Bridge engineering T Technology > TH Building construction
Departments:	School of Science & Technology > Department of Engineering

Export

Downloads