Title: Comparison of recent shear capacity models with experimental and simulation databases of prestressed ultra-high-performance concrete beams
Date Published: November - December 2024
Volume: 69
Issue: 6
Page Numbers: 37 - 57
Authors: Brandon Ross, Ahmad Tarawneh, M. Z. Naser, Anca C. Ferche, Amjad Diab, and Deanna Craig
https://doi.org/10.15554/pcij69.6-01
Click here to view the full article.
Abstract
This paper reports on an evaluation of two models for calculating the shear strength of prestressed ultra-high-performance concrete (UHPC) beams. Calculations from these models, referred to as the eCPCI and FHWA (Federal Highway Administration) models, were compared to a database of experimental test results compiled from the literature. Identifying comparable UHPC material properties from the published literature was a challenge because different testing programs used different material tests. Accordingly, UHPC localized tensile strength was estimated based on the available data and approximate ratios. In addition to the experimental database, a machine-learning prediction model was developed to generate a simulation database used for further evaluation of the eCPCI and FHWA models. On average, the FHWA model resulted in shear capacity estimates that were 17% lower than the eCPCI method. However, the analyses show that shear strength predictions from both the FHWA and eCPCI models were typically conservative relative to data from the experimental and simulated databases. In addition, both models were observed to produce less-accurate estimates for larger-sized specimens than for smaller specimens.