Flexural Reliability Analysis and Preliminary Resistance Factor Calibration of UHPC Beams Reinforced with GFRP Bars
Flexural Reliability Analysis and Preliminary Resistance Factor Calibration of UHPC Beams Reinforced with GFRP Bars
Pourya amiri Paryan1 Mehdi Shahsavari2 Hamidreza vaziri3
1) Ph.D. Candidate in Structural Engineering, Department of Civil Engineering, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran. Email:
2) Ph.D. Candidate in Structural Engineering, Department of Civil Engineering, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran. Email:
3) Assistant Professor, Department of Civil Engineering, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran. Email:
IOI Code
:
ABFG-ACAZGB
Publication :
12th.international Congress on Civil Engineering, Architecture, Art and Urban Development(12icsau.ir)
Abstract :
Abstract. Ultra-high-performance concrete (UHPC) and glass FRP (GFRP) bars are both attractive for durable structures, yet GFRP flexural provisions were calibrated for conventional concrete and no published calibration for UHPC–GFRP beams was located. This paper reanalyses two published beams (Alhoubi et al., ۲۰۲۲): ۲G۱۰U (ρf = ۰٫۵۰%), which failed by GFRP rupture at ۳۷٫۱۲ kN·m, and ۲G۲۰O (ρf = ۱٫۷۶%), which failed by concrete crushing at ۶۸٫۶۰ kN·m (۱۸۵ × ۲۵۰ mm section; cylinder strength ۱۱۰ MPa). The effective depth is not reported in the source; values of about ۲۰۳ and ۲۰۰ mm reproduce the reported reinforcement ratios and test-to-model ratios (۱٫۰۵, ۰٫۹۱) within ±۲% and replace the earlier ۲۱۰ mm. A sectional model following ACI CODE-۴۴۰٫۱۱-۲۲, a limit state with model error, dead and live load, and FORM were implemented and verified by ۴×۱۰⁶-sample Monte Carlo runs (β differences ≤ ۰٫۰۶). For a target β = ۳٫۵ and L/D = ۱٫۵, the required φ is about ۰٫۶۹ for both beams with an unbiased model and about ۰٫۷۳ (tension) and ۰٫۶۲ (compression) when model bias equals each specimen s test-to-model ratio. At the code factors (۰٫۵۵ and ۰٫۶۵), β exceeds ۴٫۲ for the tension-controlled beam and ranges from about ۳٫۱ to ۴٫۱ for the compression-controlled beam, depending on model bias. Live load (۵۲–۶۴%) and model error (۱۹–۲۵%) dominate the variance. A supplied sensitivity figure is a fixed-demand surrogate and is not equivalent to design-code calibration; it is reported separately. Because the statistical inputs are assumed and each mode has one test, the results are preliminary and not a basis for code change.