BEGIN:VCALENDAR VERSION:2.0 PRODID:-//TU/e ME//Events//EN CALSCALE:GREGORIAN METHOD:PUBLISH BEGIN:VTIMEZONE TZID:Europe/Amsterdam BEGIN:STANDARD DTSTART:19701025T030000 RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU TZOFFSETFROM:+0200 TZOFFSETTO:+0100 TZNAME:CET END:STANDARD BEGIN:DAYLIGHT DTSTART:19700329T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU TZOFFSETFROM:+0100 TZOFFSETTO:+0200 TZNAME:CEST END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT UID:event-71@tuemeche.nl DTSTAMP:20261007T234622Z DTSTART;TZID=Europe/Amsterdam:20261208T133000 DTEND;TZID=Europe/Amsterdam:20261208T150000 SUMMARY:Bayesian Uncertainty Quantification for Nonlinear Constitutive Mo deling in Engineering DESCRIPTION:Speaker: Rodrigo Lima de Souza e Silva\nHost: Clemens Verhoos el\n\nConstitutive models play a central role in engineering analysis by mathematically describing material behavior. Combined with conservation l aws\, they provide a mathematical model for the behavior of engineering s ystems.\nUnlike conservation laws\, constitutive relations can be affecte d by significant uncertainties arising from imperfect knowledge of materi al behavior\, experimental variability\, and model simplifications. A goa l of this doctoral research is to develop Bayesian methodologies for the calibration\, selection\, and uncertainty quantification of nonlinear con stitutive models\, thereby improving the reliability and interpretability of engineering predictions.\n\nThis research is founded on Bayesian infe rence as a framework for solving inverse problems in constitutive modelin g. Within this framework\, unknown model parameters are represented by pr obability distributions that are updated using observations. This approac h enables the systematic combination of prior knowledge and measurement d ata while providing rigorous quantification of uncertainty in model param eters and predictions. Markov Chain Monte Carlo (MCMC) sampling technique s are employed to characterize posterior distributions and assess predict ive uncertainty in a statistically consistent manner.\n\nA first research line investigates the practical application of Bayesian inference to con stitutive modeling problems in engineering. Specifically\, it investigate s the performance and computational efficiency of MCMC algorithms used fo r posterior exploration. Through experimental case studies in heat conduc tion and rheology\, different sampling approaches are evaluated in terms of convergence\, accuracy\, computational effort\, and reliability. This work provides practical guidelines for the application of Bayesian calibr ation methodologies to constitutive models with varying levels of complex ity and computational cost.\n\nA second research line focuses on nonlinea r heat conduction\, where material properties depend on temperature. A Ba yesian calibration framework has been developed to infer temperature-depe ndent thermal conductivity from transient or regime measurements. The met hodology combines uncertainty quantification with adaptive refinement of both numerical discretization and constitutive model complexity. By linki ng model refinement strategies to measurement uncertainty and statistical model-selection criteria\, the framework achieves accurate parameter est imation while preventing overfitting and unnecessary computational expens e. The proposed methodology has been validated using both synthetic and e xperimental datasets and enables identification of nonlinear thermal cons titutive behavior together with credible intervals as uncertainty bounds. \n\nA third research line investigates the integration of goal-oriented f inite element methods with surrogate modeling techniques for efficient Ba yesian inference. The proposed framework exploits information contained i n the likelihood function to guide adaptive numerical discretization.\nTo further reduce computational cost\, multi-output Gaussian process surrog ate models are employed to approximate the forward problem while preservi ng predictive uncertainty. The combination of goal-oriented discretizatio n\, surrogate modeling\, and Bayesian inference aims to enable efficient estimation and uncertainty quantification of constitutive models in compu tationally demanding engineering applications.\n\nMore broadly\, this the sis investigates Bayesian uncertainty quantification for constitutive mod els across multiple engineering domains\, including heat transfer\, rheol ogy\, and electromagnetism. The developed domain-independent methodologie s enable probabilistic calibration\, model validation\, and uncertainty p ropagation for nonlinear and history-dependent material behavior. The res ults demonstrate that Bayesian methods provide a rigorous framework for c onstitutive modeling\, yielding both parameter estimates and quantified c onfidence in model predictions. By integrating experimental data\, physic al modeling\, and uncertainty quantification within a unified probabilist ic framework\, this research contributes methodologies that support more reliable and trustworthy engineering simulations of complex material syst ems.\n\nMore info: https://research.tue.nl/en/persons/rodrigo-ls-silva/ LOCATION:Atlas 0.710 URL:https://tuemeche.nl/peoplepages/event.php?id=71 CATEGORIES:PhD Defense END:VEVENT END:VCALENDAR