Experimental and analytical analysis of the stress-strain diagram of FRP-confined concrete with different loading rates

by Sutherland, Brandy

The accuracy and applicability of the existing stress-strain models for concrete confined by Carbon Fiber Reinforced Polymer (CFRP) were analytically and experimentally explored. This investigation includes major parameters affecting the stress-strain response of confined concrete, including the loading pattern and protocol. It has been observed and reported that the experimentally recorded stress-strain relationship of the same specimen will be different if the loading protocol of the test is switched from displacement control to load control.

In the experimental phase of this study, four standard 6" by 12" concrete cylinders were constructed using the same concrete batch for consistency. Three two-inch strain gauges were affixed equally spaced at mid-height on the surface of the specimen in the longitudinal direction, and two two-inch strain gauges were applied in the lateral direction at mid-height opposite each other. CFRP was then impregnated with a two-part epoxy and applied externally in two continuous layers, with an overlap.

During the first phase of the experimental program, the tests were conducted with a constant load rate or with a constant displacement rate. The data was collected from externally mounted strain gauges and potentiometers positioned on the opposite sides of the cylinder in the longitudinal direction.

Since the capacity of the existing actuator in the structural lab was less than the required failure level of the specimens, a nutcracker-like device was constructed to increase the mechanical advantage of the test frame in the second phase of the experimental program. In this phase, all tests were conducted in displacement control.

Various models were selected to be studied from a large number of existing models that propose to determine the stress-strain relationship of concrete. Analytical predictions of the models were compared against the experimental data. Results show that some of the models provide a reasonable prediction of the real performance of the specimen. However, in general, predictions are different from the real performance for most models.