Modeling Fatigue Behavior of Additively Manufactured Ni-Based Superalloys via Crystal Plasticity

Veerappan Prithivirajan, Purdue University

Abstract

Additive manufacturing (AM) introduces high variability in the microstructure and defect distributions, compared with conventional processing techniques, which introduces greater uncertainty in the resulting fatigue performance of manufactured parts. As a result, qualification of AM parts poses as a problem in continued adoption of these materials in safety-critical components for the aerospace industry. Hence, there is a need to develop precise and accurate, physics-based predictive models to quantify the fatigue performance, as a means to accelerate the qualification of AM parts. The fatigue performance is a critical requirement in the safe-life design philosophy used in the aerospace industry. Fatigue failure is governed by the loading conditions and the attributes of the material microstructure, namely, grain size distribution, texture, and defects. In this work, the crystal plasticity finite element (CPFE) method is employed to model the microstructure-based material response of an additively manufactured Ni-based superalloy, Inconel 718 (IN718). Using CPFE and associated experiments, methodologies were developed to assess multiple aspects of the fatigue behavior of IN718 using four studies. In the first study, a CPFE framework is developed to estimate the critical characteristics of porosity, namely the pore size and proximity that would cause a significant debit in the fatigue life. The second study is performed to evaluate multiple metrics based on plastic strain and local stress in their ability to predict both the modes of failure as seen in fractography experiments and estimate the scatter in fatigue life due to microstructural variability as obtained from fatigue testing. In the third study, a systematic analysis was performed to investigate the role of the simulation volume and the microstructural constraints on the fatigue life predictions to provide informed guidelines for simulation volume selection that is both computationally tractable and results in consistent scatter predictions. In the fourth study, validation of the CPFE results with the experiments were performed to build confidence in the model predictions. To this end, 3D realistic microstructures representative of the test specimen were created based on the multi-modal experimental data obtained from high-energy diffraction experiments and electron backscatter diffraction microscopy. Following this, the location of failure is predicted using the model, which resulted in an unambiguous one to one correlation with the experiment. In summary, the development of microstructure-sensitive predictive methods for fatigue assessment presents a tangible step towards the adoption of model-based approaches that can be used to compliment and reduce the overall number of physical tests necessary to qualify a material for use in application.

Degree

Ph.D.

Advisors

Sangid, Purdue University.

Subject Area

Energy|Aerospace engineering|Artificial intelligence|Industrial engineering|Materials science|Mathematics|Medical imaging

COinS