Strength performance of hexagonal, hollow, and perforated GFRP-Reinforced concrete units: experimental, numerical, and AI-based investigations

This study investigates the structural performance of hexagonal, hollow, perforated SEAHIVE® concrete units reinforced with GFRP bars under transverse compression and flexural loading. An integrated experimental–numerical–artificial intelligence framework was developed to establish structural capacity and derive design-oriented predictive equations. Nonlinear finite element models were implemented in ABAQUS using the Concrete Damage Plasticity (CDP) formulation and validated against experimental load–displacement behavior and crack patterns. Parametric analysis demonstrated that element thickness is the dominant factor governing compressive capacity, with increases from 100 mm to 180 mm resulting in over 280% improvement. Increasing concrete compressive strength from 20 MPa to 40 MPa enhanced transverse compression and flexural capacities by up to 60% and 66%, respectively, while larger perforation diameters reduced capacity due to stress concentration effects. To improve practical applicability, Differential Evolution Chromosomal Gene Expression Programming (DEC-GEP) was employed to develop explicit predictive equations. The symbolic regression models achieved coefficients of determination (R²) of 0.948 for transverse compression and 0.921 for flexure, providing accurate and interpretable tools for preliminary design. The proposed framework offers a validated approach for evaluating perforated coastal concrete systems and supports their optimized application in sustainable shoreline infrastructure.