Development of a Novel Hybrid Unified Viscoplastic Constitutive Model

Abstract

2015Dedicado a mi familia, a los de aquí y a los que desde arriba siguen aquí. v ACKNOWLEDGEMENTS I would like to thank my advisor Dr. Calvin M. Stewart, whose guidance and advice was fundamental to complete my master degree. It has been an honor to work and learn from such a professional advisor. I want to state my gratitude to the Mechanical Engineering Department for all the opportunities that have been provided to me which have tremendously contributed to my professional career. Finally I would like to thank my family for being my inspiration and for all the provided support. vi ABSTRACT Gas turbines are now days used in power plants for power generation and for propulsion in the aerospace industry. In these applications gas turbines are exposed to severe temperature and pressure variations during operating cycles. These severe operating conditions exposed the turbine's components to multiple deformation mechanisms which degrade the material and eventually lead to failure of the components. Nickel based and austenitic super alloys are candidate material used for these applications due to its high strength and corrosion resistance at elevated temperatures. At such temperature levels, candidate materials exhibit a rate-dependent or viscoplastic behavior which difficult the prediction or description of the material response due to deformation mechanisms. Unified viscoplastic constitutive models are used to describe this viscoplastic behavior of materials. In the present work Miller and Walker unified viscoplastic models are presented, described and exercised to model the creep of Hastelloy X and the low cycle fatigue behavior of stainless steel 304. The numerical simulation results are compared to an extensive database of experimental data to fully validate the capabilities and limitations of the considered models. Material constant heuristic optimizer (MACHO) software is explained and used to determine both models material constants and ensure a systematic calculation of them.This software uses the simulated annealing algorithm to determine the optimal material constants values in a global surface, by comparing numerical simulations to an extensive database of experimental data. A quantitative analysis on the performance of both models is conducted to determine the most suitable model to predict material's behavior. Based on the two exercised classical viscoplastic models, a novel hybrid unified model is introduced, to accurately describe the inelastic behavior caused by creep and fatigue effects at high temperature. The presented