Date of Award


Document Type

Open Access Master's Thesis

Degree Name

Master of Science in Civil Engineering (MS)

Administrative Home Department

Department of Civil and Environmental Engineering

Advisor 1

Theresa M. Ahlborn

Committee Member 1

C. Kennan Crane

Committee Member 2

Stephen L. Kampe


Casting structural elements with ultra-high performance concrete (UHPC) tends to create preferential fiber alignment, which affects the strength and must be accounted for in design. Fiber orientation effects in tension have been studied extensively, but less work has been performed for compression. This work characterizes the fiber orientation occurring in a typical UHPC beam and how that orientation affects compressive behavior at high strain rates. Specimens (36 total) were cored from the beam and their fiber orientations were non-destructively evaluated using x-ray computed tomography (CT). Fibers showed flow-induced alignment along the length of the beam, with orientation angles in the x, y, and z-directions differing significantly. The perpendicular orientation number was used to describe orientation, as fibers perpendicular to the load were most effective in crack bridging. Cored specimens tested quasistatically achieved compressive strengths of 14.3–23.6 ksi, which appeared to increase with perpendicular orientation number. However, limited data makes this correlation uncertain. Quasistatic strengths were lower than expected due to the use of neoprene pads. Quasistatic strengths of cores tested without pads averaged 26.6 ksi. Dynamic tests at strain rates of 130–200 1/s were performed with a split-Hopkinson pressure bar (SHPB). Dynamic compressive strength ranged from 38.1 ksi to 58.5 ksi and was independent of orientation number, although results suggested that the distribution and orientation of fibers influenced crack formation. The strain at peak stress, a measure of ductility, ranged from 0.0105 to 0.0131 in dynamic tests. Strain at peak stress increased with perpendicular orientation number, but the correlation was weak. Sources of error, including stress non-equilibrium and radial confinement due to inertia, were assessed.