Forecasting Strength of CFRP Confined Concrete Using Multi Expression Programming.

Area of Science:

• Civil Engineering
• Materials Science
• Computational Intelligence

Background:

• Carbon fiber-reinforced polymer (CFRP) confined concrete is increasingly used in structural retrofitting.
• Accurate prediction of compressive strength is crucial for designing and implementing CFRP confined concrete structures.
• Existing models may lack the precision required for complex material behaviors.

Purpose of the Study:

• To apply Multi Expression Programming (MEP), a machine learning algorithm, for predicting the compressive strength of concrete confined with CFRP.
• To develop and validate a computational multiphysics model based on experimental data.
• To assess the performance and reliability of the proposed MEP model against existing methods.

Main Methods:

• Utilized a machine learning approach, specifically Multi Expression Programming (MEP).
• Developed a computational multiphysics model using historical experimental data.
• Incorporated critical parameters: specimen geometry (height, diameter), CFRP elastic modulus, unconfined concrete strength, and CFRP layer thickness.
• Performed detailed statistical analysis and external validation against experimental results.

Main Results:

• The MEP-based model demonstrated high accuracy and predictability in forecasting the compressive strength of CFRP confined concrete.
• The proposed model outperformed other existing strength models in the literature.
• Parametric and statistical analyses confirmed the model's robustness and effective training.

Conclusions:

• The developed MEP model provides a reliable and accurate tool for predicting the compressive strength of CFRP confined concrete.
• The study highlights the potential of MEP in structural engineering applications, particularly in rehabilitation and retrofitting.
• The findings contribute to the advancement of sustainable construction materials and practices.