Dresden 2017 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

MM: Fachverband Metall- und Materialphysik

MM 65: Topical session: Data driven materials design - uncertainty approaches

MM 65.3: Talk

Thursday, March 23, 2017, 16:15–16:30, BAR 205

Sensitivity analyses for large sets of density functional theory calculations — •Jan Janßen, Tilmann Hickel, and Jörg Neugebauer — Max-Planck-Institut für Eisenforschung GmbH, Düsseldorf, Germany

Over the last years methodological and computational progress in atomistic simulations has substantially improved the predictive power in materials design. However to compare the simulation results with experimental data, it is necessary to quantify the various sources of uncertainty. We therefore leverage the capabilities of our recently developed Python based workbench PyIron, to implement stochastic sensitivity analyses with the aim to differentiate model errors, statistical errors and systematical errors.

For each error we estimate the convergence gradient based on our sensitivity analyses combine it with the individual cost function of the convergence parameters and derive an algorithm for automated convergence. This approach allows us to quantify the precision not only of the energy of an individual ab initio calculation but moreover for derived quantities of sets of ab initio calculations.