Jakob T. Ostien

Prediction of Final Material State in Multi‐Stage Forging Processes

Multi‐stage forging processes are used to manufacture reservoirs for high pressure hydrogen and tritium storage. The warm‐forging process is required to produce required macro and microscale forged material properties of 304 and 21‐6‐9 stainless steel. Strict requirements on the forged material strength, grain size and grain flow are necessitated to inhibit the diffusion of gas which inevitably leads to material embrittlement. Accurate prediction of the final material state requires modeling of each of the forging stages and tracking the material state evolution through each deformation and reheating stage. An internal state variable constitutive model, capable of predicting the high strain rate, temperature dependent material behavior, is developed to predict final material strength and microstructure. History dependent, internal state variables are used to model the isotropic and kinematic hardening, grain size and recrystallization. Numerical methodologies were developed to track and remap material sta...

Exploring emerging manycore architectures for uncertainty quantification through embedded stochastic Galerkin methods

We explore approaches for improving the performance of intrusive or embedded stochastic Galerkin uncertainty quantification methods on emerging computational architectures. Our work is motivated by the trend of increasing disparity between floating-point throughput and memory access speed on emerging architectures, thus requiring the design of new algorithms with memory access patterns more commensurate with computer architecture capabilities. We first compare the traditional approach for implementing stochastic Galerkin methods to non-intrusive spectral projection methods employing high-dimensional sparse quadratures on relevant problems from computational mechanics, and demonstrate the performance of stochastic Galerkin is reasonable. Several reorganizations of the algorithm with improved memory access patterns are described and their performance measured on contemporary manycore architectures. We demonstrate these reorganizations can lead to improved performance for matrix–vector products needed by iterative linear system solvers, and highlight further algorithm research that might lead to even greater performance.

Automatic Differentiation for Numerically Exact Computation of Tangent Operators in Small‐ and Large‐Deformation Computational Inelasticity

Advances in computer software tools and technologies have transformed the way in which finite element codes and associated material models are developed. In this work, we propose a numerically exact approach for computing the sensitivites required to construct local consistent tangent operators in computational inelasticity applications. The tangent operators that come from the derivatives of constitutive equations are necessary for achieving quadratic convergence in integrating material models at the integration point level. Unlike finite difference-based numerical methods, the approach proposed in this work is based on an exact differentiation technique called automatic differentiation (AD). The method is efficient, robust and easy to incorporate. Numerical examples in both small- and large-deformation inelasticity problems with complicated material models are presented to illustrate the efficiency and applicability of the proposed method.

Ductile failure X-prize.

Fracture or tearing of ductile metals is a pervasive engineering concern, yet accurate prediction of the critical conditions of fracture remains elusive. Sandia National Laboratories has been developing and implementing several new modeling methodologies to address problems in fracture, including both new physical models and new numerical schemes. The present study provides a double-blind quantitative assessment of several computational capabilities including tearing parameters embedded in a conventional finite element code, localization elements, extended finite elements (XFEM), and peridynamics. For this assessment, each of four teams reported blind predictions for three challenge problems spanning crack initiation and crack propagation. After predictions had been reported, the predictions were compared to experimentally observed behavior. The metal alloys for these three problems were aluminum alloy 2024-T3 and precipitation hardened stainless steel PH13-8Mo H950. The predictive accuracies of the various methods are demonstrated, and the potential sources of error are discussed.

Finite Element Analysis of Hydro-mechanical Coupling Effects on Shear Failures of Fully Saturated Collapsible Geomaterials

The fully coupled diffusion-deformation processes occurring within porous geomaterials, such as sand, clay and rock, are of interest to numerous geotechnical engineering applications. In this work, a stabilized enhanced strain finite element procedure for poromechanics is integrated with an elasto-plastic cap model to simulate the associative and non-associative hydro-mechanical responses of fluid- infiltrating biphasic collapsible porous geomaterials. We present a quantitative analysis on how macroscopic plastic response caused by pore collapse and grain rearrangement affects the seepage of pore fluid, and vice versa. Finite element simulations of shear failure problems will be presented to study the effect of pore pressure dissipation on the stress path and plastic response of the porous geomaterials.

NEAMS VLTS project

The objective of the U.S. Department of Energy Office of Nuclear Energy Advanced Modeling and Simulation (NEAMS) Very Long Term Storage (VLTS) Project is to develop a simple, benchmark model that describes the performance of Zry4 d-hydrides in cladding, under conditions of long-term storage of used fuel. This model will be used to further explore the requirements of hydride modeling for used fuel storage and transport. It is expected that this model will be further developed as its weaknesses are understood, and as a basis of comparison as the Used Fuel Disposition (UFD) Campaign explores more comprehensive, multiscale approaches. Cladding hydride processes, a thermal model, a hydride model API, and the initial implementation of the J2Fiber hydride model is documented in this report.