ABSTRACT
Two new restrainer design methods based on different performance criteria were evaluated. One method is based on a rational equivalent linear static analysis approach and the other is a very simple method (W/2 method) that provides for supporting the span after unseating. Whereas the criterion for the first method is to prevent unseating, the second method aims at preventing collapse. A nonlinear response history analysis computer program was used in conjunction with five different earthquake records. The results for a series of two- and five- span bridges obtained thus far indicate that, compared to other available restrainer design methods for simply- supported bridges, the proposed method leads to a more rational distribution of restrainers while it requires considerably fewer restrainers. The W/2 method led to a satisfactory result in most cases, although limited yielding was observed in some instances.
