THE BEHAVIOR OF A CURVED BOX BEAM MODEL BRIDGE. THE DESIGN OF CURVED VIADUCTS. PROGRESS REPORT

PREVIOUS STUDIES OF THE BEHAVIOR OF CURVED BRIDGES HAVE CONSIDERED ONLY STRUCTURAL MEMBERS WITH OPEN CROSS SECTIONS, I. E., WIDE FLANGE OR I BEAMS. A MORE EFFICIENT SHAPE WOULD BE A BOX BEAM, WHICH RESISTS THE TORSIONAL FORCES DEVELOPED IN CURVED GIRDER BRIDGES. A SMALL PLEXIGLAS CURVED BOX BEAM MODEL BRIDGE WAS TESTED WHEN SUBJECTED TO VARIOUS CONCENTRATED LOADINGS. THIS MODEL, CONSISTING OF A THREE CELL UNIT WITH A COMMON TOP PLATE ELEMENT, WAS INSTRUMENTED WITH STRAIN, DEFLECTION, AND ROTATION GAGES AND TESTED AS A SINGLE-SAPN AND TWO-SAPN STRUCTURE. A SERIES OF STIFFNESS MODELS WERE ALSO FABRICATED AND TESTED IN ORDER TO EVALUATE STIFFNESS PROPERTIES. THE RESULTING EXPERIMENTAL DATA WERE THEN COMPARED TO THE ANALYTICAL DATA OBTAINED FROM A SERIES OF COMPUTER PROGRAMS INCORPORATING THE "SLOPE DEFLECTION THEORY" PREVIOUSLY USED FOR THE ANALYSIS OF OPEN CROSS- SECTIONAL GIRDERS. THE REQUIRED STIFFNESSES OF THE BOX BEAMS WERE OBTAINED BY CONSIDERING BORH SINGLE CELL BEHAVIOR AND MULTICELL BEHAVIOR. AFTER EVALUATION OF THE RESULTING ANALYTICAL DEFLECTIONS AND STRAINS THROUGHOUT THE MODEL, THESE DATA WERE THEN COMPARED WITH THE EXPERIMENTAL DATA. IN GENERAL, SINGLE-GIRDER CELL PROPERTIES PROVIDE CONSERVATIVE DEFLECTION AND ROTATIONAL DATA IN COMPARISON WITH THE EXPERIMENTAL DATA BUT EXHIBIT SMALLER VARIATION THAN THE MULTICELL DATA. THERE IS LITTLE DIFFERENCE BETWEEN THE TWO ANALYTICAL RESULTS FOR STRAIN DATA BOTH TECHNIQUES YIELDING SLIGHTLY UNCONSERVATIVE DATA. EXAMINATION OF THE WARPING STRAINS SHOWS THAT THE PRIMARY INFLUENCE IS BENDING, WITH NEGLIGIBLE WARPING EFFECTS. THUS THE TORSIONAL MODE IS RESISTED PRIMARILY BY THE ST. VENANT EFFECT, AS WOULD BE EXPECTED FOR A BOX SECTION. THE SLOPE DEFLECTION METHOD HAS PROVED TO BE A USEFUL TECHNIQUE IN THE ANALYSIS OF BOX BEAM MODELS. /AUTHOR/