I³MS - Kiendl Seminar
Prof. Dr. Josef Kiendl - Isogeometric Methods in Structural Analysis
Isogeometric analysis is a recent method of computational analysis where functions used to describe geometries in Computer Aided Design (CAD) are adopted as basis for analysis. Due to this unified geometric representation, the model transfer from design to analysis, called mesh generation, is omitted providing a better integration of design and analysis. NURBS are the most widespread technology in today’s CAD modeling tools and therefore are adopted as basis functions for analysis. Apart from the geometrical advantages, NURBS-based isogeometric analysis has proven superior approximation properties compared to standard finite element analysis for many different applications. Furthermore, the high continuity between elements also allows the discretization of higher order PDEs, which is especially useful in structural mechanics, where the classical plate and shell theories, based on Kirchhoff’s kinematic assumption, can be implemented in a straightforward way.
We show an isogeometric shell analysis framework with formulations ranging from linear, geometrically nonlinear, and fully nonlinear shell models. All formulations are based on the Kirchhoff-Love shell theory and are rotation-free, i.e., using only displacement degrees of freedom. These formulations are then employed for the simulation of various problems of structural mechanics, including large deformations, buckling, elastoplasticity, and brittle fracture as well as for fluid-structure-interaction problems including the simulations of offshore wind turbine blades and bioprosthetic heart valves.
Furthermore, we show how the high continuity provided by IGA can be used in order to develop innovative structural models. In particular, we show formulations for shear deformable beams and plates with only one unknown variable, which is a generalized displacement. Corresponding numerical formulations are characterized by having considerably less degrees of freedoms than the standard formulations and are also fully locking-free.
I³MS - McKenna Seminar
Dr. Sean McKenna - Modeling Ground Water Flow and Transport in Strongly Heterogeneous Formations
Senior Research Manager, IBM Research Dublin, Ireland
Transmissivity in heterogeneous and fractured media can range over many orders of magnitude. Predictive modeling within probabilistic risk assessment calculations requires numerical representations of these heterogeneous formations. Here, an example problem for a geologic repository is used to demonstrate the application of continuous and indicator geostatistics to create a set of seed fields conditioned to a complex geologic conceptual model and local measurements of transmissivity and storativity. Inverse parameter estimation with pilot points is used to modify these seed fields to condition them to observations of hydraulic pressure including the results of over 20 years of hydraulic testing. The resulting fields are used as input to an advective transport model. In order to quantify uncertainty in the transport predictions, multiple seed fields are run through the computationally expensive inverse calibration process. Several approaches to decreasing this computational load through decomposition of the parameters into a solution space and a null-space are examined and the impacts of each approach on the calibration and the advective travel time are quantified.
I³MS - Blocken Seminar
Prof. Dr. Bert Blocken - Computational Evaluation of Climate Adaptation Measures in Urban Areas
Even if actions for mitigation of climate change would be immediate, extensive and fully effective, a certain degree of climate change would be unavoidable due to lack of mitigation actions in the past. Its impact should be limited by climate change adaptation, i.e. adapting to its consequences. This presentation focuses on climate adaptation of cities and buildings to heat waves. It presents results of the computational evaluation of climate change adaptation measures, focused on the reduction of outdoor temperature and on the reduction of indoor temperature.
Concerning the outdoor environment, vegetation is often contemplated as a viable and effective adaptation measure against heat waves. To investigate its potential for reducing outdoor temperature during heat waves, a case study with Computational Fluid Dynamics (CFD) was conducted for a street canyon in the center of the Dutch city Arnhem. First, a double validation study was conducted based on available measurement data. Next, the case study was performed for the meteorological conditions of an afternoon hour on a hot summer day during a heat wave with wind of speed 5.1 m/s at 10 m above ground and direction along the canyon. Different scenarios were analyzed: no vegetation, avenue-trees, facade greening, roof greening and all trees, facade and roof greening combined. The results highlighted some important misconceptions, indicating that some increasingly popular adaptation measures might be much less effective than often assumed.
Concerning the indoor environment, we analyzed the effectiveness of six passive climate change adaptation measures applied at the level of building components using building energy simulations for three generic residential buildings as commonly built in - among others - the Netherlands: (1) a detached house; (2) a terraced house; (3) an apartment. The study involved both residential buildings that were built according to the regulations and common practice in 2012, and residential buildings that were constructed in the 1970s, with a lower thermal resistance of the opaque and transparent parts of the building envelope. The climate change adaptation measures investigated were: (i) increased thermal resistance; (ii) changed thermal capacity; (iii) increased short-wave reflectivity (albedo); (iv) vegetative roofs; (v) exterior solar shading; and (vi) additional natural ventilation. The performance indicators were the number of overheating hours during a year and the amount of energy needed to keep the indoor air temperature within acceptable limits. The results indicated that some of the most expensive measures are least effective, while the cheapest one has the largest beneficial effect.
Given the complexity of the heat and mass transfer processes involved, the above-mentioned studies demonstrate that computational evaluation is imperative to assess the potential of climate adaptation measures, both for the outdoor environment and the indoor environment.
I³MS - Haiat Seminar
Dr. Gauillaume Haiat - Multiscale Modeling and Characterization of the Bone-Implant Interface
, Universitè Paris-Est, Paris
Implants are often employed in orthopaedic and dental surgeries. However, risks of failure, which are difficult to anticipate, are still experienced and may have dramatic consequences. Failures are due to degraded bone remodeling at the bone-implant interface, a multiscale phenomenon of an interdisciplinary nature which remains poorly understood. The objective of this seminar is to provide a better understanding of the multiscale and multitime mechanisms at work at the bone- implant interface. To do so the evolution of the biomechanical properties of bone tissue around an implant will be studied during the remodeling process. A methodology involving combined in vivo, in vitro and in silico approaches is proposed. Different aspects related to the biomechanical behaviour of the bone-implant interface the static (contact problems) and in the dynamic (acoustics) regimes will be tackled.
CHARLEMAGNE DISTINGUISHED LECTURE SERIES-LeCun Seminar
Prof. Dr. Yann LeCun - Unsupervised Learning: the Next Frontier in AI
, Facebook, USA
Facebook AI Research & New York University
The rapid progress of AI in the last few years are largely the result of advances in deep learning and neural nets, combined with the availability of large datasets and fast GPUs. We now have systems that can recognize images with an accuracy that rivals that of humans. This will lead to revolutions in several domains such as autonomous transportation and medical image analysis. But all of these systems currently use supervised learning in which the machine is trained with inputs labeled by humans. The challenge of the next several years is to let machines learn from raw, unlabeled data, such as video or text. This is known as predictive (or unsupervised) learning. Intelligent systems today do not possess "common sense", which humans and animals acquire by observing the world, by acting in it, and by understanding the physical constraints of it. I will argue that the ability of machines to learn predictive models of the world is a key component of that will enable significant progress in AI. The main technical difficulty is that the world is only partially predictable. A general formulation of unsupervised learning that deals with partial predictability will be presented. The formulation connects many well-known approaches to unsupervised learning, as well as new and exciting ones such as adversarial training.