Experimental and computational characterization of glass microsphere-cementitious composites

Abstract

Mechanical characterization of Portland cement with hollow glass microspheres of two different crushing strengths, 500 (FM) and 10,000 psi (3 M), was carried out to determine both quasi-static and dynamic responses of the cementitious composites. In the dynamic tests, the time resolved high speed camera photos of the deforming specimens revealed the stress levels at which incipient damage occurred. The dynamic split-Hopkinson pressure bar experiments elucidated that statistically there is no difference in specific energy absorption when comparing cement with no aggregates to FM 5%, 3 M 5%, and 3 M 10% cements using the Anderson-Darling criterion. To validate these findings, both quasi-static and dynamic experiments were employed to inform a three-dimensional empirically-based homogenous continuum damage evolution concrete model (called the RHT model) in LS-DYNA. Simulations predicted the addition of glass inclusions delay the onset of the effective plastic strain, however the time between onset and failure was suggested to be lower than the Portland cement. Due to varying compressive strengths of the glass microspheres, both experimental and RHT model results found ideal volume fractions for the two manufacturers to be 5% for 3 M (10,000 psi glass microspheres) and 15% for FM (500 psi glass microspheres) to achieve the highest compressive strength and greatest energy absorbed for their respective glass microsphere specimen group. The authors believe 3 M 5% and FM 15% predictions are due to relative strength gain due to compaction and a decrease in porosity—signified by the plateauing of plastic strain which is not seen in any other volume fraction for each respective manufacturer. Computational simulations of this composite material agreed with experimental findings showing ‘Type II’ and ‘Type III’ failure mechanisms, discussed in experimental section, dominate when catastrophic crack growth occurs.