Thank you very much to our guest panelists and to the ERC candidates.
Computational and Data Science Seminar. Igor Peterlik from Inria discusses data assimilation for surgical simulation.
The purpose of this project is to reduce the errors associated with numerical solution of wave propagation problems and their computational costs. In high-frequency regimes, the solution of conventional FEM sufferers from pollution error which is due to dispersion and can be visualized as a phase shift of the numerical solution. To maintain a desired level of pollution error in conventional FEM, it is necessary to increase discretization density faster than the wave number which rapidly increase the computational cost. On the other hand, the geometrical accuracy of the scattering surfaces play an important role in accuracy of the solution. The possibility of representing man-made objects exactly in IGA even with very coarse meshes and the convenience of its refinement makes it a desirable platform to perform scattering analysis. The convergence graphs obtained for scattering problems performed in IGA platform approves the anticipated properties and makes it possible to increase the solution accuracy even for very high-frequency analysis. [http://mechanical907.rssing.com/browser.php?indx=25030677&last=1&item=1], [http://hdl.handle.net/10993/28982].
It was great to have you in the Legato team and your work on composite delamination has been a pleasure to discuss with you. Excellent work! Have a safe trip back!
In this paper we demonstrate the ability of a derivative-driven Monte Carlo estimator to accelerate the propagation of uncertainty through two high-level non-linear finite element models. The use of derivative information amounts to a correction to the standard Monte Carlo estimation procedure that reduces the variance under certain conditions. We express the finite element models in variational form using the high-level Unified Form Language (UFL). We derive the tangent linear model automatically from this high-level description and use it to efficiently calculate the required derivative information. To study the effectiveness of the derivative-driven method we consider two stochastic PDEs; a one- dimensional Burgers equation with stochastic viscosity and a three-dimensional geometrically non-linear Mooney-Rivlin hyperelastic equation with stochastic density and volumetric material parameter. Our results show that for these problems the first-order derivative-driven Monte Carlo method is around one order of magnitude faster than the standard Monte Carlo method and at the cost of only one extra tangent linear solution per estimation problem. We find similar trends when comparing with a modern non-intrusive multi-level polynomial chaos expansion method. We parallelise the task of the repeated forward model evaluations across a cluster using the ipyparallel and mpi4py software tools. A complete working example showing the solution of the stochastic viscous Burgers equation is included as supplementary material.
My long-standing colleague Chris Pearce (University of Glasgow) just passed on the sad news of the passing of Professor Nenad Bićanić.
I will remember Nenad as a passionate, gentle and dedicated mentor and I am deeply saddened by his passing. I am sure he inspired many young engineers in the many years he taught the subject and as many researchers to follow the path of knowledge.
I would like to wish all the best to his family in these difficult times.