Scholarlink

This paper presents a novel approach to characterize the direct tensile behavior of ultra-high performance fibre-reinforced concrete (UHPC), accounting for the complex relationships between the mix design constituents and the mechanical properties of UHPC. An artificial neural network (ANN)-based predictive model was developed to determine the mechanical properties of UHPFRC such as cracking stress, peak tensile strength, and the corresponding strain. Subsequently, this ANN-based tensile model for UHPFRC was integrated into a hybrid rotating crack model incorporated into an established nonlinear finite element analysis (NLFEA) software. In this formulation, suited for macro-modelling, the fibres are represented as smeared within the material. The proposed procedure was validated against UHPFRC specimens tested in the literature, including 5 membrane elements subjected to pure shear and (34) shear-critical beams. The simulated behaviours were in good agreement with the experimentally measured responses, with an \({R}^{2}\) of 0.97 for all specimens. The model was then utilized to better understand the influence of the cross-sectional shape on the behavior of shear-critical UHPFRC beams. The analysis results facilitated the development of a simplified shape effect parameter that can be integrated into existing shear capacity prediction models for UHPFRC. The proposed shear capacity model showed reasonably accurate results in predicting the capacity of the beams with unconventional cross sections that were tested in the literature.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrating an ANN-Based Tensile Model with a Hybrid Rotating Crack Formulation to Simulate the Behavior of Shear-Critical UHPFRC Structural Elements with Unconventional Cross Sections

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Amjad Diab,
  • Anca C. Ferche

摘要

This paper presents a novel approach to characterize the direct tensile behavior of ultra-high performance fibre-reinforced concrete (UHPC), accounting for the complex relationships between the mix design constituents and the mechanical properties of UHPC. An artificial neural network (ANN)-based predictive model was developed to determine the mechanical properties of UHPFRC such as cracking stress, peak tensile strength, and the corresponding strain. Subsequently, this ANN-based tensile model for UHPFRC was integrated into a hybrid rotating crack model incorporated into an established nonlinear finite element analysis (NLFEA) software. In this formulation, suited for macro-modelling, the fibres are represented as smeared within the material. The proposed procedure was validated against UHPFRC specimens tested in the literature, including 5 membrane elements subjected to pure shear and (34) shear-critical beams. The simulated behaviours were in good agreement with the experimentally measured responses, with an \({R}^{2}\) of 0.97 for all specimens. The model was then utilized to better understand the influence of the cross-sectional shape on the behavior of shear-critical UHPFRC beams. The analysis results facilitated the development of a simplified shape effect parameter that can be integrated into existing shear capacity prediction models for UHPFRC. The proposed shear capacity model showed reasonably accurate results in predicting the capacity of the beams with unconventional cross sections that were tested in the literature.