Tendon stress evaluation of unbonded post-tensioned concrete segmental bridges with two-variable response surfaces. (15th October 2021)

• Record Type: Journal Article
• Authors: Halder, Rudhra; Yuen, Terry Y.P.; Chen, Wang-Wen; Zhou, Xiangming; Deb, Trissa; Zhang, Hexin; Wen, Tzu-Han
• Is Part Of: Engineering structures. Volume 245(2021)
• Publication Date: 2021-10-15
• Subjects: Precast concrete segmental bridge -- Externally unbonded tendon -- Discrete finite element model -- Response surface methodology -- Prestress change; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2021.112984
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: Nonlinear behavior of unbonded post-tensioned precast concrete segmental bridges (PT-PCSBs) was investigated by DFEMs. Proposed DFEMs could well capture global responses and local tendon-stress, duct-slips, and joint-openings responses. Combined effects of aspect ratios and prestress on the decompression and the ultimate limit states of the bridges were thoroughly investigated. Third-order two-variable response surfaces could predict the tendon stress change at ULS with high goodness-of-fit R 2 values of 0.97. Abstract: This paper comprehensively studied the combined effects of prestress change and aspect ratios on the nonlinear structural performances of externally prestressed precast concrete segmental bridges (PCSB). An experimentally validated discrete-finite element model was adopted and various analytical cases were generated with variable span lengths. Furthermore, a simulation study is performed considering the change in prestress level to understand its effect on structural response, failure behaviour, and tendon stress at the ultimate limit state (ULS). The result showed that the stress in the unbonded tendon before the failure stage varies from 0.79 to 1.03 f py for the shorter tendon (T6L) and 0.66 to 0.94 f py for the longer tendon (T5L), and on comparing with the prediction of the existing codes, the stresses are highly underestimated. However, for the typical prestress level of around 0.6 - 0.7 f pu, the ACI318-19 code could quite well predict the ultimate tendon stress change. To establish the dependency of stress in the unbonded tendon at ULS to the normalized prestress factor ( α ) associated with the aspect ratio ( L / d ), the response surface methodology (RSM) was implemented. The mediocre prediction of one-variable linear regression analysis concludes the dependency of the combined effects of the two variables on the response variables. The 3rd order two-variable response surfaces were able to predict an increase in stress and total stress of tendon at the ULS with high goodness-of-fit values of 0.97 and 0.92 respectively.