At head of title: National Cooperative Highway Research Program.

Analyses on catastrophic collapse of some adjacent precast concrete box bean bridges reveal the fact that the function loss of shear keys can result in the formation of the "Single Plate Load Effect"; as a result, loads cannot transfer between adjacent beams and damage to both the structure and vehicles can be triggered on the loaded beam by heave trucks. This leads to the concern that the system safety of adjacent precast concrete beam bridges becomes insufficient when the shear keys deteriorate to a certain degree. To understand the system performance of this type of bridges, this dissertation studied the system performance of a model bridge over the full load range through an experimental investigation and finite element simulation. The results confirmed that the assumption that the robustness of certain adjacent precast concrete beam bridges may be considered safe according to the traditional safely check standards, but may not be sufficient in terms of the system safety criteria. Thus, a system factor is recommended to be applied in the safety check equation of this type of bridges. Even though some codes have already incorporated the criteria for evaluating the system behavior of highway bridges in the design and evaluating procedure, little guidance has been provided on how to determine corresponding redundancy or system factors. On the basis of the experiment findings and theoretical analysis, a simplified framework was proposed to address the particular structural feature and topology of adjacent precast concrete bean bridges, which can reduce the computation complexity compared to existing approaches for redundancy and robustness assessment. Sensitive parameters for system reliability were investigated and system reliability indices were computed using a finite element response surface method. The rationality and calibration of system factors were analyzed. In the end, the procedure for system safety and reliability assessment of adjacent precast concrete box beam bridges was applied on a bridge in service with different transverse connections and damage scenarios. The case study verified the proposed approach and outlined the procedure that can be adopted for routine bridge evaluation practice.

Satisfactory performance of non-composite box-beam bridges depends on the effectiveness of the key way, waterproofing membrane and tie rods, and the related construction practices. Development of cracks at the longitudinal joints of such bridges is often a recurring problem that causes water leakage at the joints and corrosion of the embedded prestressing strands. The primary objective of this study was to identify the sources, causes and effects of inadequate waterproofing at the joints and to develop prevention measures. The performance of waterproofing membranes and the structural performance of key way joints with the existing and new grout materials were evaluated and correlated with field measurements recorded under traffic loading. Construction practices for new bridges, and the investigation of a bridge that was in service for 32 years at the time of its demolition were also documented. Mechanical tests on membranes revealed that they are able to accommodate at least one inch of tensile and shear deformations without losing their waterproofing properties. That kind of elongation allows membranes to bridge over any cracks that may develop at the longitudinal joints. Therefore, membrane deficiencies may not the primary cause of water leakage. Key way joints with a combination of the currently specified ODOT geometry and ODOT-approved grouts were found to be incapable of carrying any shear loads in conjunction with out-of-plane moments. From the limited site inspections done in this project, the practices followed at construction sites seem to be seriously flawed and may be largely contributing to water leakage problems in box-beam bridges. Recommendations on new key way geometries and the grouts that were developed and tested in this project are suggested for implementation.

Bridge decks supported by adjacent precast/prestressed-concrete beams have become increasingly popular in recent years due to their ease of construction, shallow superstructure, and aesthetic appeal. In New York, such structures are built by placing a number of precast beams alongside one another and connecting them through grouted keyways called "shear keys". After the grout hardens, the beams are transversely post-tensioned and a composite, cast-in-place deck is poured over them. Before 1992, field inspection personnel frequently reported the appearance of longitudinal deck cracking over these partial-depth shear keys soon after construction. In response, a new system using full-depth shear keys with more transverse tendons was adopted in 1992. Since then, more than 100 such bridges have been built statewide. To evaluate their performance, a followup inspection survey was conducted in 1996 on 91 such bridges.

Bridge designs frequently employ precast concrete box beams, with adjacent beams connected using grouted keyways. Failure at these joints leads to water leakage and corrosion in reinforcing bars/strands, resulting in severe damage to concrete elements and reducing the bridge's lifespan. This dissertation investigates the causes for the ineffectiveness of waterproofing membranes in preventing leakage at joints of adjacent box beams.Testing of five commercially available waterproofing membranes showed low tensile and shear strength and high deformability, with membranes exhibiting large elongation and no leakage. Adhesion tests showed limited peel strength. Different practices such as use of primer, sealant, and direct heat were also investigated. Sealant provided better bonding than primer, and heating yielded no improvement in bond strength. All membranes showed low punching resistance but could resist loads up to 1,200 lb in wheel load tests without being damaged. Prestress losses were calculated for construction scenarios for concrete box beams of ages varying from 7 to 180 days at time of deck placement. The difference in prestress losses have an inverse relationship with concrete age at the time of deck placement, while an inverse relationship exists between the age of concrete of prestressed box beams at the time of deck placement and the differential deflections. Spans having different lengths were found to have a proportional relationship to the difference in prestress losses and differential deflections. While differential prestress losses at different concrete ages contribute to joint cracks between adjacent box beams, losses may not be sufficient to rupture the membranes. The maximum differential deflection between adjacent box beams with different concrete ages may cause cracking in the keyways but is likely to be negligible compared to the ability of the membrane to stretch. In visits to an existing bridge and a new bridge underconstruction, deficiencies were noted in the implementation of waterproofing membranes: inadequate specifications and ineffective inspection, absence of membranes on portions of the existing bridge, and protrusions/debris on the concrete deck of the new bridge prior to membrane installation. This study recommends that waterproofing membranes continue to be used in bridges and provides suggestions to improve their implementation.