Optimizing Structural Performance of Short Concrete-Filled Steel Tube Columns with Stiffening Methods Using LEA-HGCNN Technique

Super tall buildings widely utilize concrete-filled steel tube (CFST) columns due to their ability to withstand significant axial stresses. As these columns experience extreme loading conditions, high-strength concrete is increasingly used to reduce the cross-sectional area while maintaining structural integrity. However, optimizing their performance remains a challenge, especially regarding stability and load-bearing capacity. This research examines the performance of steel-reinforced concrete-filled steel tubular columns using various stiffening techniques and the many confinement models proposed for square CFST columns. A new approach combining the lotus effect optimization algorithm and hybrid graph convolutional neural network technique is proposed to improve the structural efficiency of short CFST columns. The technique integrates machine learning and graph-based modeling to predict and optimize structural behavior under different loading conditions. The proposed technique is implemented using the MATLAB platform and is benchmarked against existing optimization algorithms, including heap-based optimization, particle swarm optimization, and the salp swarm algorithm. Experimental results indicate that the proposed approach significantly improves prediction accuracy, achieving a low error value of 0.7%. The findings show the efficiency of the proposed technique in improving the strength, stability, and load-bearing capacity of CFST columns, making it a promising solution for high-rise building construction.