Session: 12-12-02: Modeling of the Fracture, Failure, and Fatigue in Solids

Paper Number: 145420

145420 - Investigation of Fatigue Crack Nucleation in Ti Alloys With Microtexture Regions Using Parametrically Upscaled Models 

Microtexture regions (MTR), essentially defined as clusters of grains with similar crystallographic orientations, play a significant role in the fatigue failure of dual-phase titanium alloys. This work discusses the development of a unique multiscale modeling approach based on parametric upscaling hence referred to as the Parametrically Upscaled Constitutive Model (PUCM) and Parametrically Upscaled Crack Nucleation Modeling (PUCNM) framework for the prediction of structural-scale fatigue crack nucleation in alpha/beta Ti-6Al-4V alloys, which contain micro-texture regions (MTRs). This approach to multiscale modeling provides a thermodynamically consistent upscaled constitutive framework to bridge spatial scales through the explicit representation of microstructural descriptors in the constitutive equations in the form of Representative Aggregated Microstructural Parameters (RAMPs). The coefficients of model parameters in the PUCM are functions of these RAMPs. As a result, the upscaled model has a direct dependence on microstructural features, while overcoming the computational resource limitations of a micromechanical model to simulate a component. A novel RAMP (k_MTR) is proposed for quantification of MTRs that in addition to morphological and crystallographic features of the underlying microstructure, considers the nature of MTR (hard/soft) with respect to the local stress state. A micromechanics-based Crystal Plasticity Finite Element (CPFE) is used to demonstrate the dependence of fatigue crack nucleation on MTRs and their characteristic features. A strong correlation between fatigue life with intensity of MTR is observed. A grain scale crack nucleation model is used to obtain crack nucleation probabilities based on the load-shedding phenomenon. Further, k_MTR is incorporated into the PUCNM framework to predict the fatigue life of a specimen. The CPFE model with the grain scale crack nucleation model that is calibrated for Ti64 and validated using experimental results is used to generate a database of crack nucleation probability evolution with state variables and RAMPs. Statistically equivalent representative volume elements are obtained from EBSD micrographs of Ti64 containing MTRs and simulated under dwell-fatigue conditions for generating this database. A cycle jump technique based on wavelet decomposition referred to as the Wavelet Transformation-Induced Multi-time Scaling (WATMUS) method is employed here to simulate a large number (>50,000) of load cycles. Genetic Programming Symbolic Regression (GPRS), a machine learning tool, is used to derive the functional dependence of crack nucleation probabilities on the state variable and RAMPs. The capabilities of the PUCM/PUCNM are demonstrated by simulating fatigue crack nucleation in a Double Edge Notched Tensile (DENT) specimen. With the proposed framework, RAMPs can be distributed across a specimen to replicate the varying microstructure in a component. The effect of microstructure as well as geometric features (such as holes and corners) in crack nucleation are investigated. The results show that the location and presence of MTRs in the specimen can significantly change the fatigue life as well as the location of crack nucleation in the specimen. The uncertainty in the fatigue life due to microstructural variability is also quantified in this work. While for lower loads, the effect of microstructure dominates fatigue crack nucleation, in case of higher loads, the geometric features are observed to have a more significant impact.

Presenting Author: Kishore Appunhi Nair Johns Hopkins University

Presenting Author Biography: The presenting author is a Ph.D. candidate at Johns Hopkins University's Computational Mechanics Research Lab, supervised by Prof. Somnath Ghosh. His research concentrates on investigating fatigue fracture in Titanium alloys. Prior to this, he specialized in developing multiscale models that integrate atomistic details to simulate crack propagation within Ti alloys. Presently, his focus lies in upscaling the micromechanical crack propagation model to simulate damage within structural scale specimens.

Recent publication:
Appunhi Nair K., Ghosh S. (2023), Crack tip enhanced phase-field model for crack evolution in crystalline Ti6Al from concurrent crystal plasticity FE-molecular dynamics simulations, European J. Mech. - A/Solids

Authors:

Kishore Appunhi Nair Johns Hopkins University
Tawqeer Tak Johns Hopkins University
Vasisht Venkatesh Pratt & Whitney
Adam Pilchak Pratt & Whitney
Somnath Ghosh Johns Hopkins University