Bridges are essential infrastructure constituents that have been studied for centuries. Typically, seismic bridge design and assessment utilize simplified modeling and analysis techniques based on one-dimensional spine elements and zero-length springs/hinges. The geometry of the elements and calibration of parameters are based on assumptions for the lateral load path and failure modes, e.g., sacrificial backwall and shear keys, neglecting wing walls, and strength based on backfill alone. These assumptions have led to observations of underestimated resistance, overestimated displacement demands, and unpredicted damage and failure mode. The focus of the study is on ordinary standard bridges with continuous reinforced concrete box girder superstructures and seattype abutments. A bridge component calibration study was conducted first using simplified (spine models with 1D elements and springs) and three-dimensional nonlinear continuum finite element models (FEM). Model responses were compared with experimental results to identify the drawbacks in the simplified models and verify the adequacy of the material nonlinearities and analysis procedures. The components include a T-girder, abutment backfill, abutment shear key, elastomeric bearing pad, and a bridge pier. Results show the simplified models do not capture damage propagation and failure mode in the shear key case, nonlinear behaviors in beams with high aspect ratios (or deep beam action), and underestimate the strength and overestimate the stiffness for the backfill case. The component models (both simplified and continuum) were then used in studying the nonlinear static behaviors of key bridge lateral-load resisting substructures, namely abutments and bents. For the abutment subsystem, cases with and without backfill and several back wall construction joint configurations for the longitudinal direction, with monolithic shear key and shear key with construction joint for the transverse direction, and boundary conditions in the transverse direction were considered. Abutment subsystem results showed simplified models underestimate the resistance by 10-60%, neglect back wall and wing wall structural contributions, and localize damage in the back fill relative to the continuum models. For the bent subsystem, a full bridge system that considers material nonlinearity and damage in the bent segment only was adopted to determine the effect of the finite bent cap or superstructure-to-column connection. Inelastic behavior and damage was included in the columns, bent cap, and a superstructure segment with a length that correspond to the dead load moment inflection point. The other superstructure segments and the pile cap were modeled as elastic. Bent subsystem results showed simplified models overestimate the stiffness, induce excessive flexibility and deformation in the cap beam, and overestimate columns' deformations. Due to the differences observed in the abutment subsystem, and the potential impact of the abutment behavior on the seismic response of the whole bridge system, dynamic studies on the bridge system were conducted using four abutment parameters: abutment stiffness and strength in each of the longitudinal and transverse directions. Two models were developed to conduct nonlinear time history analysis: an equivalent single-degree-of-freedom (SDOF) model for each of the longitudinal and transverse directions, and a 3D spine bridge model. Constant ductility analyses were conducted using the SDOF systems, while standard probabilistic seismic demand analysis was used on the spine systems. Results revealed that, besides the columns yielding, the abutment has an early and significant contribution to the behavior. The SDOF system results showed that increasing the abutment stiffness or strength reduces the system displacement demand and increases the system forces. The consequence of such increase in the forces is mobilizing significant amount of force in the abutments, causing inelastic response. The full bridge study also confirmed the SDOF results and showed that the abutment forces are more than 200% of the columns forces that would result in the same aftereffect observed in the SDOF system.

Graduation Date





Mackie, Kevin


Doctor of Philosophy (Ph.D.)


College of Engineering and Computer Science


Civil, Environmental and Construction Engineering

Degree Program

Civil Engineering









Release Date


Length of Campus-only Access


Access Status

Doctoral Dissertation (Open Access)