Caglar Oskay delivered an invited lecture at United Technologies Research Center
Lecture Title: Multiscale Simulation-Based Progressive Damage Modeling of Composite Materials.
Location: United Technologies Research Corporation, United Technologies Corporation, East Hartford, CT.
Abstract:
We will present a multiscale simulation-based failure and life prediction approach for composite materials and structures. This approach is a new alternative to the traditional model-based life prediction paradigm. The multiscale methodology we employ considers multiple time scales, to account for the size disparity between loading periods and characteristic times associated with damage accumulation (in case of fatigue loading), as well as multiple spatial scales, to account for the size disparity between the characteristic lengths of the composite structure and the underlying constituents. The methodology is a space-time generalization of the computational homogenization method based on the rigorous mathematical homogenization theory. The primary issue of computational complexity is addressed by reduced order modeling in both space and time. The multiscale approach is also coupled with a probabilistic (Bayesian) parameter calibration strategy, which allows for quantification of uncertainty associated with failure and life predictions. The uncertainty quantification framework also ties the uncertainty observed at the structural scale to the sources of uncertainty at the scale of the material microstructure.
We will discuss various applications of the proposed simulation-based failure and life prediction paradigm including blind predictions of the static and fatigue behavior of laminated carbon-fiber reinforced polymers, validation studies in composite materials subjected to impact and blast loads, as well as compression after impact response.