Stress Increase of Unbonded Tendons in Continuous Posttensioned Members

Abstract

A database of 253 unbonded tendon-reinforced members was assembled and analyzed with a focus on Δfps. Large scatter was observed and gaps were identified in previous databases due to possibly unsuitable inclusion of certain testing programs. The influence of several geometric and material properties was analyzed using a covariance analysis, and significant differences between simple-span and continuous members were observed, although some prediction methods do not differentiate between the two. The assembled prediction models for Δfps had evaluation statistics ranging from 1.84 < λ < 4.11 and 0.06 < R2 < 0.16 and indicating relatively poor accuracy and precision. Modifications to the scaled plastic hinge length ψ for different subsets (simple span, continuous, internally unbonded, and externally unbonded) were suggested to provide the most accurate and precise prediction when compared with the available methods. The proposed method, a modified version of the bridge design code prediction, provides the most accurate (λ = 1.34) and precise (R2 = 0.27) prediction and does not increase complexity.