Damage assessment of ultra-high-performance concrete protective wall against gaseous explosion

Ultra-high-performance concrete (UHPC) exhibits excellent mechanical strength and ductility, rendering it an ideal material for blast-resistant structures. This study focuses on the damage assessment of UHPC protective walls subjected to gaseous explosions, which differ from high explosive loads by having lower peak overpressures and longer durations. By combining advanced numerical modeling and machine learning techniques, the Karagozian & Case Concrete (KCC) material model was adapted to UHPC with a compressive strength of 160 MPa and validated against quasi-static and dynamic experimental results. Numerical structural analysis revealed that shock waves caused more severe damage than pressure waves for the same impulse, underscoring the importance of distinguishing between blast types in design. A parametric study examined the effects of wall thickness, height, and reinforcement ratios on pressure-impulse (P-I) diagrams, offering practical insights for optimizing UHPC blast-resistant wall designs. To address computational challenges, an artificial neural network (ANN) model was trained on 3044 numerical results to predict the dynamic response of UHPC structures, achieving high accuracy ( R ² = 0.989). This comprehensive methodology enables efficient evaluation and optimization of UHPC protective walls under gaseous explosions, supporting safer designs in high-risk industries handling flammable gases. • Validated the KCC material model for UHPC with a compressive strength of 160 MPa under gaseous explosions. • Developed P-I diagrams for shock and pressure wave blast loadings from deflagration and detonation. • Proposed an ANN-based predictive model achieving high accuracy and efficiency for UHPC damage assessment.