Material Modelling for Gas Manifolds in Marine Engines
Pysyvä osoite
Kuvaus
Opinnäytetyö kokotekstinä PDF-muodossa.
AISI 316L stainless steel is a popular choice for marine engine parts because of its excellent resistance to corrosion, good weldability, and reliable mechanical performance. However, this steel undergoes complex cyclic plastic deformation under repeated loading, including nonlinear hardening, cyclic stabilization, and strain-amplitude-dependent hardening, which are challenging features for constitutive modelling. These effects are not properly captured by simplified constitutive models that are typically employed in industrial finite element analyses, e.g., elastic-perfect plastic formulation. Therefore, the estimation of local stresses, plastic strains and fatigue-critical areas might not be trustworthy. This thesis is mainly aimed at enhancing the constitutive modelling ca-pability of cyclic plasticity of AISI 316L stainless steel and investigating the impact of sophisticated material modelling on the predicted response of marine engine gas manifold components. More specifically, the work explores whether a constitutive model with nonlinear kinematic hardening, isotropic hardening, viscoplasticity, and strain-amplitude-dependent hardening can reproduce the experimentally observed cyclic behavior of AISI 316L and provide finite element predictions that are more accurate and credible than simple material models. A one-dimensional implementation for the chosen constitutive model was created and tested versus a simple example from the Z-set materials library. The cyclic loading experiments that had been performed and published were digitized, analyzed, and used for parameter identification. To calibrate the model, a combination of global and local optimization methods was used to fit the parameters to the stress-strain hysteresis loops and the evolution of peak stress data. The trained model was later tested on different experimental cases that were not part of the training set. The parameterized constitutive model was able to match the cyclic hardening, hysteresis loop development, and stress stabilization data for different strain levels obtained from the experiments. In addition, the parameters related to isotropic hardening and memory effects were varied to study their impact on the cyclic response. At last, the calibrated material model was integrated into an Abaqus finite element submodel of a marine engine gas manifold. The findings revealed major changes in local stress redistribution, stress concentration phenomenon, and plastic strain evolution as compared to the traditional elastic-perfect plastic material model. The research indicates that the advanced material model-ling of cyclic plasticity, an AISI 316L behavior representation could be achieved that is closer to reality and therefore enhance the prediction of mechanical weak points in marine engine structures. This modelling approach, therefore, constitutes a practicable method for engineering anal-yses requiring both high fidelity of prediction and numerical robustness.