Predicting the Ultimate Axial Capacity of Uniaxially Loaded CFST Columns Using Multiphysics Artificial Intelligence

Area of Science:

• Structural Engineering
• Computational Mechanics
• Artificial Intelligence in Civil Engineering

Background:

• Concrete-filled steel tubes (CFST) are crucial structural elements.
• Accurate prediction of CFST column bearing capacity is essential for structural integrity and lifecycle assessment.
• Existing prediction methods may be time-consuming or lack precision.

Purpose of the Study:

• To develop a Multiphysics prediction model for circular CFST columns.
• To evaluate the efficacy of Artificial Neural Network (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Gene Expression Programming (GEP) for this prediction task.
• To compare the performance of these AI techniques in predicting CFST column bearing capacity.

Main Methods:

• Utilized a database of 1667 CFST column data points (702 short, 965 long).
• Input parameters included geometric dimensions and material mechanical properties.
• Developed and statistically validated Multiphysics models using ANN, ANFIS, and GEP; performed parametric and sensitivity analyses.

Main Results:

• The Gene Expression Programming (GEP) model demonstrated superior performance compared to ANN and ANFIS.
• GEP predictions closely matched actual bearing capacity values, indicated by lower PI and OF error metrics.
• GEP's ability to derive solutions from experimental data without prior assumptions highlights its robustness.

Conclusions:

• Gene Expression Programming (GEP) is a highly effective tool for predicting the bearing capacity of circular CFST columns.
• The GEP model offers a reliable alternative to laborious and time-consuming experimental testing.
• Further research exploring other AI techniques like Random Forest Regression and Multi Expression Program is recommended.