ABSTRACT
In a grillage analysis of Super-T girder bridge decks, linear-elastic analysis considering full torsion stiffness may not be appropriate for ultimate limit states. For members subjected to compatibility torsion, typically in Super-T girder bridge decks, the onset of cracking at the torsion cracking moment (Tcr) results in reduced stiffness as loads increase towards the ultimate limit case. Consequently, the section experiences lower torsion moments (non-linear behavior) that redistribute through greater bending moments. Therefore, as an alternative, a linear-elastic grillage analysis model considering a reduced proportion of torsion stiffness can help avoid this phenomenon of overestimating torsion and underestimating moments in grillage models. This study evaluates the effect of varying torsion considerations in grillage analysis. To represent the complete range of designs possible for a Super-T PSC member, it is necessary to analyze bridge decks with feasible lower and upper spans of a standard Super-T section. For the range of feasible spans, the section cracking torsion moment is first quantified, noting the amount of beneficial prestress force present. Following, grillage analysis is performed - each with increasing proportions of torsion stiffness from 0 to 100% of torsion constant, J for longitudinal members (Super-T girders). Torsion cracking resistance is then quantified for each feasible span. The results showed that most proportions of torsion stiffness did not lead to the torsion moment exceeding the torsion cracking moment. Overall, the results show that a range of 10 - 40 % of torsion constant J would allow bridge designers a good balance between the need for torsion design - minimizing conservatism in longitudinal bending moments.
