Date of Award

12-2016

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Civil Engineering

Committee Member

Dr. Brandon E. Ross, Committee Chair

Committee Member

Dr. Thomas E. Cousins, Committee Co-Chair

Committee Member

Dr. Weichiang Pang

Committee Member

Dr. Amin Khademi

Abstract

Transfer of shear forces across concrete-to-concrete interfaces is critical to the strength of many reinforced concrete structures. One common example is the horizontal interface between precast concrete girders and cast-in-place concrete bridge decks. Composite action between the girder and deck, and thus bridge stiffness and strength, relies on the capacity of the interface to transfer shear forces. This concept is explained through Interface Shear Transfer (IST) theory. In this dissertation, trinary objectives are presented to describe and scrutinize this theory. These objectives include evaluating the current code-based IST models, creating a new IST model, and demonstrating a new method for experimentally testing IST. Firstly, a database of IST experiments on uncracked reinforced concrete specimens was created from published test results. A total of 774 tests were reviewed, with data coming from tests conducted between 1969 and 2014. Once compiled, the database was used to evaluate the accuracy of the interface shear transfer provisions from the AASHTO LRFD Bridge Design Specifications, Eurocode 2, and CSA A23.3. Through this evaluation it was determined that experimental capacities were an average of 1.49, 1.93, and 2.83 times greater than the code-calculated nominal capacities for the LRFD, Eurocode, and CSA codes, respectively. While each of the codes was conservative on average, the degree of conservatism was found to be dependent on design variables such as concrete compressive strength, amount of interface reinforcement, and member size. In the first phase of the dissertation, it was shown that current code-based IST models produce inconsistent levels of accuracy for different values of design variables. In the second phase, sensitivity analyses were performed to identify the variables having the greatest impact on the IST capacity, and to create a design model that produces consistent levels of accuracy. Using a database of experimental results, an Artificial Neural Network (ANN) model was created to estimate IST strength and to perform a sensitivity analysis of the design variables. The sensitivity analysis demonstrated that compressive strength was the most significant variable affecting IST capacity. A multiple linear-regression analysis was also performed to aide in development of a new design model. Based on the results of the sensitively analysis, the new model directly accounted for compressive strength of concrete as one of the model variables. The model was strongly correlated with the experimental data and produced consistent levels of accuracy for a range of design variables. Finally, IST test methods were scrutinized and a new IST test method was proposed. Traditionally, IST capacity has been tested using a push-off test method, in which direct shear is induced through compression loads placed at the ends of notched test specimens. In this research, the 4-point bending test method, as proposed by Iosipescu in 1967, was investigated to study IST. The 4-point bending test created direct shear by strategically placing the supports and loads on a beam. It had the advantages of using test specimens that were easier to assemble. Additionally, it produced a more uniform stress state at the interface compared to the interface stress distribution of the push-off test, making the 4-point bending test a better representation of the stress state at cast-in-place deck and bridge girder connections. In this study, these two test methods were compared and contrasted through an experimental program along with analytical modeling. The conditions in which the proposed test method was an acceptable alternative for the push-off method were identified.

Share

COinS