Gjesteforelesninger og seminarer - Side 9

Abstract: What links a baby’s first breath to adhesive debonding, enhanced oil recovery, filtration or multiphase microfluidics? These processes all involve two-phase flows in rigid or elastic confined vessels and are often prone to interfacial instabilities. The canonical viscous fingering instability, which occurs when air displaces a viscous fluid in the narrow gap between two parallel plates, offers a versatile testbed for such phenomena. In this talk, I will use both experiments and numerical simulations of depth-averaged models to explore several aspects of bubble dynamics in Hele-Shaw cells. I will first show how the onset of fingering can be suppressed when replacing the upper plate of the vessel with an elastic sheet. Interfacial flows in narrow gaps can also exhibit considerable disorder, but they are rarely investigated from a dynamical systems’ perspective. I will show how compliance can promote rich multiplicity of front propagation modes in a channel before turning to bubble propagation in a rigid channel with a depth perturbation. There I will explore how the bubble’s organised transient dynamics is orchestrated by weakly-unstable steady propagation modes, and how its long-term behaviour may be practically unpredictable.

This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions. Hybrid format via Zoom possible on demand (contact timokoch at uio.no)

Abstract: A random, labyrinthine pattern emerges during slow drainage of a granular-fluid system in two- dimensional confinement. Compacted grains are pushed ahead of the fluid-air interface, which becomes unstable due to a competition between capillary forces and the frictional stress mobilized by grain-grain contact networks. We reproduce the pattern formation process in numerical simulations and present an analytical treatment that predicts the characteristic length scale of the labyrinth structure. The pattern length scale decreases with increasing volume fraction of grains in the system and increases with the system thickness. By tilting the model, aligned finger structures, with a characteristic width, emerge. A transition from vertical to horizontal alignment of the finger structures is observed as the tilting angle and the granular density are varied. The dynamics is reproduced in simulations. We also show how the system may explain patterns observed in nature, created during the early stages of a dike formation.

This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions. Hybrid format via Zoom possible on demand (contact timokoch at uio.no)

C*-algebra seminar by Alexander Stolin

C*-algebra seminar by Alexander Mang (Saarland University)

Abstract: Because of their huge compressibility difference with their surrounding media, air bubbles in water have a special relationship with acoustic waves: they are sub-wavelength resonators. In this presentation, I will show that this characteristic has great implications for both the surrounding fluid, because of the steady streaming effect, but also for the acoustic waves.

This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.

We introduce SMARTboost (boosting of symmetric smooth additive regression trees), a machine learning model capable of fitting complex functions in high dimensions, yet designed for good performance in small n and low signal-to-noise environments. SMARTboost inherits many of the qualities that have made boosted trees the most widely used machine learning tool for tabular data; it automatically adjusts model complexity, handles continuous and discrete features, can capture nonlinear functions in high dimensions without overfitting, performs variable selection, and can handle highly non-Gaussian features. The combination of smooth symmetric trees and of carefully designed Bayesian priors gives SMARTboost an edge (in comparison with a state-of-the-art tool like XGBoost) in most settings with continuous and mixed discrete-continuous features. Unlike other tree-based methods, it can also compute marginal effects.

This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.

Abstract: This talk is about the non-integer (fractional) derivative, its mathematical formulation by Abel in 1823, and present-day applications in modeling power-law behavior. These applications are in acoustics of complex media like tissue and sediments as well as in rheology, turbulence, and dielectrics. It will build on my book “Waves with Power-Law Attenuation”, Springer, 2019.

The talk will be streamed online. Please contact "timokoch at uio.no" for the Zoom link. This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.

C*-algebra seminar talk by Johannes Christensen (KU Leuven)

Abstract: Gravitational settling of a droplet in air onto a soft substrate is a ubiquitous event relevant to many natural processes and engineering applications. We study this phenomenon by developing a three-phase lubrication model of droplet settling onto a solid substrate coated by a thin, soft compressible layer of elastic material. By combining scaling analysis, analytical methods and numerical simulations we elucidate how the resulting droplet dynamics is affected by the soft layer. We discuss extensions to droplet settlings onto thin viscous liquid films and elastic sheets. Our results provide new insight into the coupled interactions between droplets and solids coated by a thin film of a soft material.

Talk can be followed online on Zoom as well as in person. Please contact "timokoch at uio.no" for the Zoom link. This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.

Maintenance plays a crucial role in ships and especially in the vital electric propulsion system. Intelligent predictive maintenance idealistically aims at preventing system failures and minimizing needless repairs, i.e., predicting failure likelihood and time to failure while providing the crew explainable predictions and recommending the best action for timely intervention. This presentation will cover a relevant work in collaboration with Sensor Systems in BigInsight, particularly a paper published under https://doi.org/10.1109/TII.2022.3144177. The failure prediction approach is driven by event logs, which include warnings, alarms, and operational information that describe all the happenings onboard the ship. The failure prediction objective is turned into classification and regression tasks; however, the training data pose three challenges. The events are irregular textual messages. The training data samples are not labelled. The datasets are extremely imbalanced, due to sparse failure events and multiple failure modes. The problem is casted into a weakly supervised machine learning framework. In a multiple instance learning process, the ungiven data labels are learned recursively while fitting the model parameters using deterministic annealing. The overall approach was tested on real ship data, and it successively forecasted few propulsion failures with explainable causes.

Abstract: We investigate the simulation of a rising bubble and a stationary droplet interaction to gain a better understanding of the rising dynamics and the morphology changing of bubble-droplet aggregate. A detailed study is conducted on the interaction process under different-size bubbles with various combinations of spreading factors. The current simulation framework consists of the conservative phase-field Lattice Boltzmann equation (LBE) for interface tracking and the velocity-pressure LBE for hydrodynamics. We simulate the contact line dynamics to confirm the method's accuracy. We further investigate the morphology changing of two contact droplets under different combinations of spreading factors and depict the final morphologies in a diagram. The separated, partially engulfed and complete engulfed morphologies can be replicated by systematically altering the sign of the spreading factors. The rising bubble and droplet interaction is simulated based on different final morphologies by adding a body force. The results show that the aggregate with double emulsion morphology can avoid distortion and maintain a greater terminal velocity than the aggregate with partially engulfed morphology.

Talk is online on Zoom. Please contact "timokoch at uio.no" for the Zoom link. This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.

Deep learning (DL) has had unprecedented success and is now entering scientific computing with full force. However, current DL methods typically suffer from instability, even when universal approximation properties guarantee the existence of stable neural networks (NNs). In this talk we will show that there are basic well-conditioned problems in scientific computing where NNs with great approximation qualities are proven to exist, however, there does not exist any algorithm, even randomised, that can train (or compute) such a NN to even 1-digit of accuracy with a probability greater than 1/2. These results provide basic foundations for Smale’s 18th problem ("What are the limits of AI?") and imply a potentially vast classification theory describing conditions under which (stable) NNs with a given accuracy can be computed by an algorithm. We begin this theory by initiating a unified theory for compressed sensing and DL, leading to sufficient conditions for the existence of algorithms that compute stable NNs in inverse problems. We introduce Fast Iterative REstarted NETworks (FIRENETs), which we prove and numerically check (via suitable stability tests) are stable. The reference for this talk is: https://arxiv.org/abs/2101.08286 (to appear in Proc. Natl. Acad. Sci. USA).

Abstract: The fungal kingdom is one of the most species-rich organismal groups, containing up to 6 million species worldwide with a large diversity of ecosystem functions. Multicellularity has evolved independently in fungi, and over time many different growth forms and structures have originated. I will present some basics on fungal growth and the formation of complex multicellular structures.

Online participation is possible too. Please contact "timokoch at uio.no" for the Zoom link. This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.

Super-resolution is a hot topic in current day Machine Learning.  The origin of the methodology dates back to applications in seismic imaging. I discuss the evolution from the early days and highlight some papers which have given new theoretical insights along the way. I illustrate the bridge between traditional convex optimization and current day convolutional neural nets. Along the way I show some examples where we have used this for current day applications in seismic imaging.

Abstract: Transport and mixing in multiphase flow through porous media plays a key role in a range of biological, geological, and engineered systems. Here, we use numerical simulations to investigate the effect of intermittent multiphase flow on fluid stretching and folding, a fundamental mechanism driving solute mixing and reaction in porous media. We show that, in contrast to steady single-phase flows, the concurrent flow of two immiscible phases induces chaotic mixing, characterized by exponential stretching in the pore space. The stretching rate is found to decay with increasing capillary number, implying that the increasing flow intermittency observed at lower capillary numbers enhances the mixing efficiency. We propose a mechanistic model to link the basic multiphase flow properties to the chaotic mixing rate, opening new perspectives to understand mixing and reaction in multiphase porous media flows. The results presented here form part of the background for the recently started RCN-funded project M4: Mixing in Multiphase flow through Microporous Media, which will also be introduced.

This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.

Abstract:  The cerebral circulation must ensure continuous blood perfusion of the brain which is the biggest oxygen consumer in the human body. It must also provide mechanisms for adaptability to changing oxygen demand as well as resilience to local blockages. We will look at such mechanisms at the level of the microcirculation where the mechanics of blood flow is dominated by red blood cells. We will find that red blood cells do not only play the role of oxygen carriers, but that they are an important element of blood flow regulation itself. To this end, we will compare results from in vitro studies in microfluidic chips to theoretical and computational models and to in vivo data from mice. We will derive local auto-regulation mechanisms for blood flow and will study how local modifications in the vascular network can modify the global hematocrit distribution. These results will emphasize the relevance of red blood cell mechanics and microvascular network geometry in cerebral blood perfusion.

This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.