Civil and Environmental Engineering Department Seminar: Data-assisted high-fidelity modeling for systems design and monitoring
Seminar | March 4 | 10-11 a.m. | 542 Davis Hall
Increased availability of measured data has recently generated tremendous interest in the development of methods to learn from data. In parallel, engineers have a long history of building high-fidelity physics- based models that allow us to model the behavior of highly complex systems. This talk aims at presenting some of the exciting research opportunities that arise at the intersection of data analytics and high-fidelity modeling, and illustrate potential applications for civil systems.
System identification methods use measurements from a system to learn its equations and parameters, with possible applications to damage detection and structural health monitoring. Bayesian inference algorithms are attractive as they allow quantification of uncertainties, which can arise when the data is not informative or the inputs are stochastic. However, Bayesian techniques become computationally expensive for inference in large dimensional nonlinear systems, i.e., finite element models. We demonstrate the potential of Bayesian filtering techniques and algorithmic enhancements to reduce computational cost, and how to integrate these learning algorithms into complex frameworks of model selection and optimal design of experiments.
The combination of data-mining and physics-based modeling finds applications in various engineering fields. In the materials sciences, machine learning algorithms can be used to build efficient structure- property linkages. Applying machine learning algorithms to engineering problematics gives rise to challenging problematics, for example the need to accurately quantify uncertainties or handling small amounts of data. This talk thus aims at illustrating some of the challenges and opportunities related to the use of both model- and data-based learning algorithms for engineering applications.