Stochastic Multiscale Prediction of Failure Initiation in Polycrystalline Materials

Professor Caglar Oskay

Abstract:
Understanding and predicting failure initiation in structural systems subjected to fatigue loads is one of the key outstanding challenges in engineering, given that at least half of all mechanical failures are attributed to fatigue. Fatigue failure initiation also happens to be a quintessential multiscale problem. Both the macroscopic structural behavior as well as the material microstructure play prominent and interacting roles in determining when and where microstructural flaws nucleate that lead to formation of long cracks and ultimately structural failure. Because of the inherent randomness in the morphology and properties of just about every material class at fine scales, fatigue failure initiation is also a highly stochastic phenomenon. Failure initiation prediction problem therefore sits at the confluence of engineering mechanics, materials science and uncertainty quantification.

In this talk, we present a new stochastic multiscale computational framework to perform physics-based prediction of failure initiation in polycrystalline materials subjected to fatigue loading conditions. At the heart of this framework lies a reduced order modeling (ROM) approach that effectively bridges the fine and coarse scales at a very low computational cost to facilitate physics-based probabilistic analysis. We demonstrate the effectiveness of the proposed ROM approach in not only capturing failure nucleation in pristine polycrystalline microstructures but also those with microcracks – a key capability that allows us to distinguish between flaw nucleation and microstructurally short crack growth regimes. The proposed stochastic framework quantifies the critical structure-property connection between the variabilities in the features of the material microstructure and the probability of failure initiation. The predictive capabilities of the proposed framework is assessed in the context of a high-performance titanium alloy subjected to high cycle fatigue.

Biosketch:
Caglar Oskay is Professor of Civil and Environmental Engineering, and the Mechanical Engineering Departments at Vanderbilt University. Concurrently, he is serving as Program Director of the Engineering for Civil Infrastructure Program within the Division of Civil, Mechanical and Manufacturing Innovation at the National Science Foundation. Prof. Oskay’s research focuses on the nonlinear response prediction of heterogeneous materials and structures using computational modeling and simulation. His recent research includes characterization of the failure response of systems that involve multiple temporal and spatial scales, computational method development, and reduced order modeling for failure analysis of composites, polycrystalline materials and metamaterials subjected to fatigue, impact, blast and other extreme loading and environmental conditions. Prof. Oskay was named Chancellor Faculty Fellow at Vanderbilt University in 2016, Fellow of ASME in 2017 and Fellow of the Engineering Mechanics Institute in 2019. Prof. Oskay serves on the Executive Committees of the ASME Materials Division and US Association of Computational Mechanics, and as the Associate Editor of the Journal of Applied Mechanics and the International Journal for Multiscale Computational Engineering.