Xiang Zhang and Rudraprasad Bhattacharrya presented their research at EMI 2017 Conference in San Diego
The title of the Xiang’s first presentation was: “Microscale modeling of creep deformation and rupture using cohesive zone-crystal plasticity finite element analysis”.
Abstract:
Nickel-based alloys are widely used as structural materials for high temperature applications due to their exceptional combination of high temperature strength and creep resistance. Inconel 617, a solid-solution strengthened Nickel-based alloy, is being considered as one of the leading structural materials for the intermediate heat exchangers (IHX) in very high temperature reactor (VHTR). In this work, a computational model considering isothermal and large deformation conditions at the microstructural scale has been developed to accurately describe the creep deformation and consequent rupture of the alloy 617 operating in high temperature.
In order to capture the creep strains that accumulate particularly at relatively low stress levels, a dislocation climb model has been incorporated into a dislocation glide-based crystal plasticity finite element analysis (CPFE) framework. The dislocation climb model describes creep induced mechanical deformations using kinematical arguments consistent with the climb mechanism that dominates creep deformation at the range temperatures considered in this study for the given alloy. In addition, a cohesive zone model has been fully implemented in the context of the crystal plasticity finite element model to capture the intergranular creep damage. Cohesive elements are placed at all grain boundaries to understand the progressive accumulation of grain boundary cavitation and consequent rupture under creep loading. The material parameters defining the creep deformation are calibrated and verified through experiments performed at 950°C. The parameters for cohesive zone model have been calibrated using available experimental rupture data. The numerical simulations illustrate the capability of the proposed model in capturing damage initiation and growth under creep loads compared to experimental observations.
The title of the Xiang’s second presentation was: “Sparse and Scalable Eigenstrain-based Reduced Order Homogenization Models for Polycrystal Plasticity”.
Abstract:
We propose a highly scalable reduced order multiscale computational framework for modeling the plastic deformation in polycrystalline materials. The key idea of this model is to concurrently link the response of the structure to the response at the scale of the grains, and track deformation and stress state of each grain throughout the domain of the structure and throughout the history of loading.
The proposed approach is based on the Eigenstrain-based reduced order homogenization model (EHM) for polycrystalline materials, which has been recently developed by the authors. EHM operates in a computational homogenization settings, which takes the concept of transformation field theory that pre-computes certain microscale information (e.g. localization tensors, concentration tensors) by evaluating linear elastic microscale problems and considers piece-wise constant inelastic response within partitions (e.g., grains) of the microstructure. By this approach, a significant reduction in computational cost is achieved, compared with classical computational homogenization approaches that employ crystal plasticity finite element (CPFE) simulation to describe the microscale response. While EHM provides approximately two orders of magnitude efficiency compared with CPFE for middle-sized microstructure, its efficiency degrades as microstructure size increases. A grain-cluster accelerated, sparse and scalable reduced order homogenization model (sparse EHM) has been developed to address this issue for computationally efficient multiscale analysis of complex polycrystalline microstructure. The acceleration is achieved by introducing sparsity into the linearized reduced order system through selectively considering the interactions between grains and consistently adapting the localization and coefficient tensors to account for the neglected long-range inter-grain interactions. The proposed approach results in a hierarchy of reduced models that recovers EHM, when full range of interactions is considered, and recovers the Taylor model, when all inter-grain interactions are neglected. The resulting sparse system is solved efficiently using a direct sparse solver. A layer-by-layer neighbor-clustering scheme is proposed and implemented to provide a robust solution for generating different ranges of inter-grain interactions. Performance of the sparse EHM is evaluated by comparing the results against the full EHM and CPFE simulations.
The title of the Rudra’s presentation was: “Mesh-size independent multiscale damage modeling of fiber-reinforced composites subjected to multi-axial loading”.
Abstract:
The progression of damage accumulation in structural components made of fiber-reinforced composites is complicated due to the presence of and interactions between complex microstructural damage mechanisms. In particular, even when subjected to unidirectional loading, the internal stress states within a laminated composite are multidimensional. In this study, we present a multiscale continuum damage model for fiber composites to predict failure under different stress triaxiality. The particular focus of this work is on capturing the ductile failure behavior under shear dominated load states along with the brittle behavior observed under axial loading, and relating the observed behavior at the lamina scale to the constitutive response of the composite constituents, i.e., the resin and the fiber.
A thermodynamically consistent continuum damage mechanics model is developed for the polymeric resin material, and deployed within the eigendeformation-based reduced order homogenization (EHM) framework. EHM is a reduced order multiscale modeling approach that has been demonstrated to capture the behavior of composite structures of various microstructural configurations. The proposed damage model directly addresses the presence of nonlinearity observed under the shear dominated load states. The continuum damage mechanics based model is regularized using a multiscale version of crack band modeling to alleviate the issues of mesh size and microstructure size effects. The proposed framework was verified using numerical simulations, and applied to investigate the static tensile and compressive behavior of IM7/977-3, a graphite fiber reinforced epoxy composite.