The deterioration of the nation's civil infrastructure has prompted the investigation of numerous solutions to offset the problem. Some of these solutions have come in the form of innovative materials for new construction, whereas others have considered rehabilitation techniques for repairing existing infrastructure. A relatively new system that appears capable of encompassing both of these solution methodologies is the Sandwich Plate System (SPS), a composite bridge deck system that can be used in both new construction or for rehabilitation applications. SPS consists of steel face plates bonded to a rigid polyurethane core; a typical bridge application utilizes SPS primarily as a bridge deck acting compositely with conventional support girders. As a result of this technology being relatively new to the bridge market, design methods have yet to be established. This research aims to close this gap by investigating some of the key design issues considered to be limiting factors in implementation of SPS. The key issues that will be studied include lateral load distribution, dynamic load allowance and deck design methodologies.

With SPS being new to the market, there has only been a single bridge application, limiting the investigations of in-service behavior. The Shenley Bridge was tested under live load conditions to determine in-service behavior with an emphasis on lateral load distribution and dynamic load allowance. Both static and dynamic testing were conducted. Results from the testing allowed for the determination of lateral load distribution factors and dynamic load allowance of an in-service SPS bridge. These results also provided a means to validate a finite element modeling approach which would could as the foundation for the remaining investigations on lateral load distribution and dynamic load allowance.

The limited population of SPS bridges required the use of analytical methods of analysis for this study. These analytical models included finite element models and a stiffened plate model. The models were intended to be simple, but capable of predicting global response such as lateral load distribution and dynamic load allowance. The finite element models are shown to provide accurate predictions of the global response, but the stiffened plate approach was not as accurate. A parametric investigation, using the finite element models, was initiated to determine if the lateral load distribution characteristics and vibration response of SPS varied significantly from conventional systems. Results from this study suggest that the behavior of SPS does differ somewhat from conventional systems, but the response can be accommodated with current AASHTO LRFD bridge design provisions as a result of their conservativeness.

In addition to characterizing global response, a deck design approach was developed. In this approach the SPS deck was represented as a plate structure, which allowed for the consideration of the key design limit states within the AASHTO LRFD specification. Based on the plate analyses, it was concluded that the design of SPS decks is stiffness-controlled as limited by the AASHTO LRFD specification deflection limits for lightweight metal decks. These limits allowed for the development of a method for sizing SPS decks to satisfy stiffness requirements.