Investigation of geopolymer-based ultra-high performance concrete slabs against contact explosions

• Resistance of geopolymer-based ultra-high performance concrete (G-UHPC) slabs to contact explosions is tested. • Numerical simulations are conducted on G-UHPC slabs to reproduce the local damage induced by contact explosions. • Parametric studies with the variables of slab thickness and TNT charge weight are performed on steel fibre reinforced G-UHPC slabs under contact explosions. • A support vector machine (SVM) method was used to classify the local damage of steel fibre reinforced G-UHPC slabs under contact explosions. Geopolymer-based ultra-high performance concrete (G-UHPC) is a new form of UHPC, which has been developed to meet the demand for ultra-high strength, cost-effective and eco-friendly construction materials. This paper preliminarily investigated local damage of fibre reinforced G-UHPC slabs subjected to contact explosions. Under 0.4 kg TNT, three 150 mm thick slabs including one control specimen made of steel rebar reinforced normal strength concrete (NSC) and two G-UHPC slabs, i.e. plain G-UHPC and 1.5% steel fibre reinforced G-UHPC, were tested. Under 1.0 kg TNT, two 200 mm thick slabs including one control specimen made of steel rebar reinforced NSC and one 2% basalt fibre reinforced G-UHPC slab were tested. Comparing with steel rebar reinforced NSC, inferior performance of plain and 2% basalt fibre reinforced G-UHPCs was observed owing to their own brittle characteristics, whereas superior performance of 1.5% steel fibre reinforced G-UHPC was achieved. Further, numerical investigations were conducted on the steel rebar reinforced NSC and 1.5% steel fibre reinforced G-UHPC slabs to reproduce their local damage induced by contact explosions through the explicit finite element code LS-DYNA. With the validated numerical model, parametric studies were performed to explore the effect of slab thickness and TNT charge weight on the local damage of 1.5% steel fibre reinforced G-UHPC slabs under contact explosions, and then the local damages were identified and classified via the support vector machine (SVM) method. Based on the machine learning results, empirical equations were derived for the fast assessment of local damage levels of 1.5% steel fibre reinforced G-UHPC slabs under contact explosions.