Constitutive Concrete Material Model Comparison to Tested Reinforced Slabs Subjected to Blast Loads

In a series of experiments, reinforced concrete slabs were subjected to simulated blast loads from a compressed air-driven shock tube. These tests utilized well-controlled test specimen designs. Applied blast load pressure histories were recorded, as were peak and residual displacements.

This paper evaluates three current concrete material models within LS-DYNA (Version 971) and DYNA3D by comparing their numerical results to two of these tests, as well as to one other. The concrete material models investigated include the Concrete Damage Model (Material 72, Release 3), the Continuous Surface Cap Model (Material 159), and the Applied Engineering Cap model with Three Invariants (AEC-3I).

The results indicate some variation in predicted damage and failure modes between the three concrete materials, but overall, all three models produced satisfactory results (i.e., peak positive deflection within a factor of 2 of the measured response). A surprising outcome is that the inclusion of rate-dependent material properties actually increased the error in the predicted responses. In terms of predicting crack patterns, the AEC-3I model appears to be preferable, whereas in predicting peak deflection, MAT72R3 is preferable. Overall, MAT159 was the most consistent predictor and the least sensitive to variations in representing the rate dependent properties of concrete.