Determination of Bond Stress Distribution Coefficient by Maximum Likelihood Method

Abstract

Transfer length is a significant parameter in the design of pretensioned concrete members. The estimation of transfer length is typically achieved by measuring concrete strain or strand slip at the member ends. The end slip method is simple and requires little effort, but relies on an empirical formula with an undetermined coefficient (α), which relates to the bond stress distribution. Many studies have proposed appropriate α coefficients; however, the variability in estimating the α coefficient exists between studies depending on the chosen empirical fit. This study gives an estimation of the α coefficient using a probabilistic approach called the maximum likelihood (ML) method. Approximately 600 data points of transfer lengths and strand end slips were measured for 25 pretensioned concrete beams. Strand end slips were measured at release and at 3, 5, 7, 14, and 28 days. Concrete strains were measured at the same period of time using a detachable mechanical strain (DEMEC) gauge to determine the transfer lengths. An estimated α coefficient of 2.61 maximized the likelihood of the observed sample of the measured transfer lengths and end slips.