Scholarlink

<p>The use of fiber reinforced polymers (FRP) is being practiced from past several years to strengthen the performance of reinforced concrete (RC) beams. The selection of a suitable FRP wrapping is of utmost importance in this case. The present study deals with experimental study of RC beams with carbon FRP (CFRP) sheets wrapped on three sides of the beam except the top. The RC beams of 150&#xa0;mm x 150&#xa0;mm cross-section with an overall span of 700&#xa0;mm were considered. The flexural strength of these beams was estimated under two-point symmetric loading and modulus of rupture was calculated by IS 456–2000, ACI 318 and Eurocode 2 recommendations. The results were validated using ANSYS APDL software package showing close agreement, with errors ranging between 4 and 11%. Machine learning models—Random Forest (RF) and Gradient Boosting Regressor (GBR)—were developed using experimental variables such as concrete strength, reinforcement details, FRP type, number of layers, and wrapping scheme. The dataset was split into 70% training and 30% testing, with additional validation through unseen data. Performance was assessed using R², MAE, and RMSE. GBR demonstrated the highest prediction accuracy (R² ≈ 0.96), outperforming conventional code-based formulas. The study concludes that ML models can reliably predict MOR of FRP-strengthened beams, offering a promising tool for structural design and assessment.</p>

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

Machine learning-assisted evaluation of flexural strength of FRP-confined concrete beams: experimental and ANSYS APDL study

  • 期刊论文

  • 发表日期：

    2026年1月8日

查看全文
  • Neelima Varma,
  • Swati Kulkarni,
  • Sumant Kulkarni

摘要

The use of fiber reinforced polymers (FRP) is being practiced from past several years to strengthen the performance of reinforced concrete (RC) beams. The selection of a suitable FRP wrapping is of utmost importance in this case. The present study deals with experimental study of RC beams with carbon FRP (CFRP) sheets wrapped on three sides of the beam except the top. The RC beams of 150 mm x 150 mm cross-section with an overall span of 700 mm were considered. The flexural strength of these beams was estimated under two-point symmetric loading and modulus of rupture was calculated by IS 456–2000, ACI 318 and Eurocode 2 recommendations. The results were validated using ANSYS APDL software package showing close agreement, with errors ranging between 4 and 11%. Machine learning models—Random Forest (RF) and Gradient Boosting Regressor (GBR)—were developed using experimental variables such as concrete strength, reinforcement details, FRP type, number of layers, and wrapping scheme. The dataset was split into 70% training and 30% testing, with additional validation through unseen data. Performance was assessed using R², MAE, and RMSE. GBR demonstrated the highest prediction accuracy (R² ≈ 0.96), outperforming conventional code-based formulas. The study concludes that ML models can reliably predict MOR of FRP-strengthened beams, offering a promising tool for structural design and assessment.