Report No.: CCEER-93-6

Title: Response of Bridge Hinge Restrainers During Earthquakes-Field Performance, Analysis, and Design

Authors: M. Saiidi, E. Maragakis, S. Abdel-Ghaffar, S. Feng, and D. O'Connor

Date: May 1993

A Report to the:

California Department of Transportation
National Science Foundation
Nevada Department of Transportation

Performing Organization:
Department of Civil Engineering/258
University of Nevada, Reno
Reno, NV 89557

Executive Summary:

This report presents a summary of the important findings of a study aimed at several aspects of the behavior of hinge restrainers used as a seismic retrofit measure. Details of the study are described in five other reports [15 to 19]. The study included field investigations, extensive analytical studies, and an evaluation of the restrainer design method. The objectives of the study were:

  1. To review the actual performance of bridge hinge restrainers during the 1989 Loma Prieta earthquake
  2. To simulate the earthquake effect on the analytical models of several selected bridges and study their responses
  3. To carry out a parametric study of these bridges to determine the effect of stronger earthquakes and the effect of changes in the restrainer gaps
  4. To review the restrainer design procedure and recommend any needed refinements

A data base of the bridges with hinge restrainers which had been damaged by the 1989 ~Loma Prieta earthquake was formed. Twenty-three bridges were in the data base. The damage reports prepared by Caltrans maintenance Division were reviewed. Three bridges, namely, the Central Viaduct, the Route 580/24/980 Separation, and the Route 92/101 Separation were investigated in the field. Measurements were made of crack widths and patterns, and the condition of the restrainers was examined. An analysis of locally damaged components was subsequently made. The field investigations pointed out the need to consider the performance of the restrainer system and not merely the restrainers. In addition to the restrainers, the system includes (a) the connection between the restrainers and the superstructure including any diaphragms, and (b) the superstructure adjacent to the hinge.

The data base of the damaged bridges was also used to select four bridges for detailed nonlinear response history analyses using computer program NEABS-86. The four structures ranged from three to eleven spans. They had different number of hinges and different substructure characteristics. The earthquake intensity also varied considerably from one bridge to another. Two groups of earthquake analyses were carried out: in one analysis the input acceleration records collected at sites near the bridges were used, and in the other a series of parametric studies with larger peak ground acceleration (PGA) was conducted. The field investigations and the analyses showed that the Loma Prieta earthquake activated the hinge restrainers in the majority of the bridges investigated in this study. Except for a few instances, the restrainers and their supporting systems performed well. It was also noted that bridges with a small ratio of number of hinges to the number of spans and in which the substructure is relatively stiff, are less likely to be susceptible to support loss.

The evaluation of the current Caltrans restrainer design method consisted of two parts: (1) a study of the effect of refinement in the current methods, and (2) a large number of nonlinear analyses of the Caltrans example bridge for different earthquakes, hinge gaps, and the number of restrainers. Based on these studies, a new method for the computation of the relative hinge movement was proposed, and demonstrated for one of the four bridges which had been the subject of detailed nonlinear analyses. It was found that the current Caltrans method for restrainer design leads to a conservative and safe design in terms of the number of restrainers. However, the degree of conservatism for different hinges is not uniform. It was also determined that a more refined method to compute relative hinge displacements can lead to fewer restrainers even in hinges with a nominal seat width of 6 in. The refined method would explicitly incorporate the nonlinearity of soil at the footings and abutments, plastic hinging of the columns, and the nonlinearity of the hinges (Summary by authors).