Seismic cable restrainer design method to control the large-displacement response for multi-span simply supported bridges crossing fault rupture zones

Highlights

• A new restrainer design method is proposed for the fault-crossing bridges.

• Faulting-induced permanent ground dislocation is considered for the design method.

• The design method was evaluated using parametric numerical analyses.

• Both elastic and superelastic restrainers can effectively control the large-displacement for fault-crossing bridges.

Abstract

Previous earthquakes have highlighted the seismic vulnerability of fault-crossing multi-span simply supported (FC-MSSS) bridges due to the large displacement of decks. Restraining devices, being of low cost and easy to install, can be a potential alternative to prevent the large displacement or falling of bridge spans for FC-MSSS bridges. However, the current restrainer design guidelines cannot provide an appropriate design method for such restraining devices in MSSS bridges accounting for the effect of faulting-induced permanent ground dislocation. To address this issue, this study aims to propose a restrainer design procedure for FC-MSSS bridges. In this proposed procedure, the restrainers are designed according to the combination of response spectrum analysis based on a linearized 2-degree-of-freedom (2-DOF) analytical model and quasi-static analysis of the bridge. A five-span simply supported bridge crossing Puqian-Qinglan fault, which is located in Puqian Bay in Hainan, China, is chosen as a case study. Two types of restrainers, i.e., elastic steel and superelastic shape memory alloy (SMA) cables, are considered for the fault-crossing bridge. Over 30 synthetic ground motions with increasing permanent ground dislocations are generated using a hybrid simulation approach. Numerical studies show that the restrainers designed by the proposed method could efficiently limit the relative displacement within a designer-specified value for the fault-crossing bridges. Using SMA cables as seismic restrainers could noticeably reduce the required length compared with elastic steel cables.

Introduction

During 1999, three major earthquakes with surface faulting (i.e., M_w 7.4 Kocaeli, M_w 7.6 Chi-Chi, and M_w 7.2 Ducze) demonstrated that the fault offset (also referred as “permanent ground dislocation”) was a severe threat to the transportation facilities. Several bridges crossing fault rupture directly suffered devastating damage resulting from faulting-induced permanent ground dislocation in these earthquakes [1]. Since then, the seismic behavior of bridges crossing active faults has attracted increasing attention in the earthquake engineering community. Because of the unawareness of the existence of an active fault during the design process of bridges and inevitable construction of the bridge across an already known active fault in a dense traffic network, a large number of bridges have been built across potentially active fault rupture zones around the world. Yang and Mavroeidis [2] listed 147 cases of bridges crossing potentially active faults around the world. Moreover, nearly ten large span bridges and long tunnels under design and construction in the Sichuan-Tibet railway cross active faults in China [3,4]. These infrastructures are susceptible to damage once the faults ruptured.

Multi-span simply supported (MSSS) bridges are widely found in the transportation networks due to their excellent features, such as low cost, straightforward mechanical properties, easiness of construction. The practical applications in bridge construction show that the isolated MSSS bridges are generally designed and constructed to cross an active fault due to their exceptional ability to accommodate the foundation dislocation and ease of repair after a large earthquake (e.g., Bolu Viaduct in Turkey [5] and Puqian Approach Bridge in China [6]). Nevertheless, a drawback of such a system is that the seismic displacement between the superstructure and substructure (hereafter referred to as “relative displacement”) can be very large, resulting in severe damage of bearings even unseating of the deck. To date, the seismic response of fault-crossing MSSS bridge was studied by several research groups experimentally and numerically. Yi et al. [7] conducted a 1/10 scaled shake table test for a two-span simply-supported bridge crossing a strike-slip fault rupture zone. In their study, it is assumed that the longitudinal axis of the bridge was perpendicular to the fault rupture. Therefore, large relative displacement in the transverse direction of the bridge was measured. In fact, the orientation of the longitudinal axis of bridges with respect to the fault rupture is not always perpendicular. Park et al. [8] analyzed the seismic performance of the Bolu Viaduct during the 1999 Duzce earthquake using nonlinear time-history analysis. The bridge crossed the fault rupture with an angle of about 25° and the faulting-induced ground dislocation in the longitudinal direction of the bridge was approximately 1.36 m. Hence, the relative displacement in the longitudinal direction of the bridge exceeded the capacity of the bearings at an early stage of the earthquake. Zhang et al. [9] investigated the effect of fault crossing angle on the seismic behaviors of a three-span simply-supported bridge crossing strike-slip fault rupture zones. The analysis result revealed that the risk of unseating or deck falling from the pier becomes larger when the bridge across the fault with a smaller angle. Therefore, it is meaningful and essential to study the method to mitigate the large relative displacement of the bridges crossing fault rupture zones.

A number of destructive earthquakes have confirmed the efficiency of hydraulic, friction, and hysteretic dampers in mitigating seismic response of bridges [[10], [11], [12], [13], [14], [15]]. However, the effectiveness of such dampers in controlling the seismic response of fault-crossing bridges is questionable considering the large permanent ground dislocation. Compared with dampers, the restraining device, as suggested by different seismic guidelines [[16], [17], [18]], maybe a suitable alternative to prevent the bridge spans from large displacement or unseating for the fault-crossing bridges (see Fig. 1). During the 1989 Loma Prieta and 1994 Northridge earthquakes, the reconnaissance of structural damage has shown that the cable restrainers could effectively mitigate the bridge damage [19,20]. Elastic steel cables acting in direct tension have been the primary forms for restraining the displacement against excessive movement in highway bridges [21]. DesRoches et al. [22] conducted a full-scale bridge model subjected to monotonic loading to investigate the capacity and modes of failure of the steel cable restrainer system. The steel cables had limited elastic strain range and poor energy dissipation capacity because they were designed to work in the elastic range. Besides elastic restrainers, a new type of self-centering restrainers, i.e. shape memory alloy (SMA) cable, was developed to retrofit the bridge. Numerous studies [[23], [24], [25], [26], [27]] have demonstrated that such restrainers could efficiently limit the displacement response and reduce the residual deformation of the bridge due to their superelastic behavior and superior energy dissipation capacity. Most of the above studies mainly utilized the restraining devices to decrease the probability of unseating for the bridge without crossing a fault. To the best knowledge of the authors, how to prevent the bridge spans from large displacement or unseating for a fault-crossing bridge is still not well investigated. The efficiency of both the elastic and superelastic restrainers for the fault-crossing bridge has not been well understood. Additionally, there is no appropriate design method for restraining devices in fault-crossing bridges.

The objective of this study is to develop a restrainer design procedure for FC-MSSS bridges. The design procedure estimates the displacement demand by superposing peak responses obtained from the response spectrum analysis of a linearized 2-degree-of-freedom (2-DOF) analytical model and quasi-static analysis of the bridge. The inclination of decks caused by permanent ground dislocation in the vertical direction is considered to calculate the required length of restrainers. A five-span simply supported bridge crossing active normal faults is taken as a case study. Elastic and superelastic restrainers (i.e. steel and SMA cables) are chosen as seismic restrainers. 30 pairs of ground motions with permanent ground dislocations (increasing from 300 mm to 1400 mm) generated by a physical model-based hybrid approach are used to describe the fault rupture. The effectiveness of the proposed design procedure is demonstrated using nonlinear time-history analysis. The effects of several key parameters, i.e., displacement ductility of the pier, allowable displacement, slack of restrainers, and stiffness ratio of pier and bearing, on the seismic performance of the fault-crossing bridge are investigated.