Torsional Analyses of Hollow RC Members and Solid PC Members

• Main Authors: Je-Sheng Chen, 陳哲昇
• Other Authors: Chyuan-Huan Jeng
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/52962589200235460899

碩士 === 國立暨南國際大學 === 土木工程學系 === 98 === A new rational theoretical model called the Softened Membrane Model for Torsion (SMMT) was developed to predict the entire torque-twist response of solid reinforced concrete (RC) members, and is referred to as SMMT-RC-S in this thesis. The SMMT-RC-S has recently been modified and extended to analyze the torsion of hollow RC members and solid prestressed concrete (PC) members. In this thesis, these modified models are referred to as SMMT-RC-H (SMMT for hollow RC) and SMMT-PC (SMMT for PC), respectively. This study applies the SMMT-RC-H and SMMT-PC to analyze tested hollow RC and solid PC beam specimens, quantifies the model parameters, and validates the SMMT-RC-H and SMMT-PC. This study collects a total of 53 hollow RC specimens and 65 solid PC specimens in the literature. Some of the collected hollow RC specimens with excessive wall thickness are excluded and the other 21 specimens are adopted for comparison. Some solid PC specimens are excluded based on three criteria: insufficient reinforcement, excessive stirrup spacing, and insufficient or excessive thickness of concrete cover; the rest of the other 42 solid PC specimens are adopted for comparison. The amplification factorsλ and μ for the pre-cracking modulus and the cracking strain, respectively, of the tensile stress-strain relation of concrete are quantified by systematically comparing the SMMT-RC-H analytical results of the 21 hollow RC specimens with test data. It is found that the values of λ=μ=1.2 give predicted cracking torques that best correlate with the test values. Comparison of the SMMT-PC analytical results of the 42 solid PC specimens with tests indicates that the prestressing generally causes the predicted ultimate torque to be larger than test value; and further, the greater the level of prstressing is, the greater the difference between predicted ultimate torque and test value may become. Hence, two reduction factors G and G^' (G=1-0.88∙σ⁄(f_c^' )；G^'=1-0.93∙σ^'⁄(f_c^' )；σ⁄(f_c^' ) and σ^'⁄(f_c^' ) both represent level of prestressing) are derived by regression of the comparison between predicted ultimate torques and test values. The predicted ultimate torques multiplied by G or G^' agree with the test values very well.