A5068044515- Ionosphere and magnetosphere dynamics
- Solar and Space Plasma Dynamics
- Earthquake Detection and Analysis
- Astro and Planetary Science
- Geomagnetism and Paleomagnetism Studies
- Advanced Optimization Algorithms Research
- Magnetic confinement fusion research
- Planetary Science and Exploration
- Radiative Heat Transfer Studies
- Combustion and flame dynamics
- Optimization and Variational Analysis
- Computational Physics and Python Applications
- Mining Techniques and Economics
- Advanced Data Storage Technologies
- Solar Radiation and Photovoltaics
- Belt Conveyor Systems Engineering
- Mineral Processing and Grinding
- Risk and Portfolio Optimization
- Anomaly Detection Techniques and Applications
- Distributed and Parallel Computing Systems
- Time Series Analysis and Forecasting
- Parallel Computing and Optimization Techniques
- Vehicle Routing Optimization Methods
- Geophysics and Gravity Measurements
- Complex Systems and Time Series Analysis
KU Leuven
2014-2024
European Space Agency
2024
University of Chile
2004-2023
European Space Operations Centre
2023
Center for Mathematical Modeling
2005-2018
Forschungszentrum Jülich
2018
Centre of Plasma Physics - Institute for Plasma Research
2012-2015
Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique
2009-2012
Laboratoire des Sciences pour la Conception, l'Optimisation et la Production
2009
DePaul University
2006
Poirot is a Web-based tool supporting traceability of distributed heterogeneous software artifacts. A probabilistic network model used to generate traces between requirements, design elements, code and other artifacts stored in 3 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">rd</sup> party case tools such as DOORS, rational rose, source repositories. The designed with extensibility mind, so that additional artifact types can be easily added....
Models based on neural networks and machine learning are seeing a rise in popularity space physics. In particular, the forecasting of geomagnetic indices with network models is becoming popular field study. These evaluated metrics such as root-mean-square error (RMSE) Pearson correlation coefficient. However, these classical sometimes fail to capture crucial behavior. To show where lacking, we trained network, using long short-term memory make forecast disturbance storm time index at origin...
We present the first comparison of multiple global simulations solar wind interaction with Mercury's dayside magnetosphere, conducted in framework international collaborative project SHOTS - Studies on Hermean magnetosphere Oriented Theories and Simulations. Two 3D magnetohydrodynamic two hybrid simulation codes are used to investigate response without its exosphere a northward-oriented interplanetary magnetic field. cross-compare results four for theoretical case MESSENGER orbit similar...
The spatial extension of active regions the Sun (hence their associated images) can strongly vary from one case to next. This inhomogeneity is a problem when using convolutional neural networks (CNNs) study solar flares, as they generally use input images fixed size. Different processes be performed retrieve database with homogeneous-sized images, such coarse resizing, cropping, or padding raw images. Unfortunately, key features lost distorted beyond recognition during these processes. lead...
The growing energy demands of High Performance Computing (HPC) systems have made efficiency a critical concern for system developers and operators. However, HPC users are generally less aware how these concerns influence the design, deployment, operation supercomputers even though they experience consequences. This paper examines implications HPC’s consumption, providing an overview current trends aimed at improving efficiency. We describe hardware innovations such as energy-efficient...
We demonstrate the improvements to an implicit Particle-in-Cell code, iPic3D, on example of dipolar magnetic field immersed in flow plasma and show formation a mag- netosphere. address problem modelling multi-scale phenomena during magnetosphere by implementing adaptive sub-cycling technique resolve motion particles located close dipole centre, where intensity is maximum. In addition, we implemented new open boundary conditions model inflow outflow plasma. present results global...
Abstract We present a method based on unsupervised machine learning to identify and characterize regions of interest using particle velocity distributions as signature pattern. An automatic density estimation technique is applied provided by particle-in-cell simulations study magnetic reconnection regions. Its application new. The key components the involve (i) Gaussian mixture model determining presence given number subpopulations within an overall population, (ii) selection with Bayesian...
One of the goals machine learning is to eliminate tedious and arduous repetitive work. The manual semi-automatic classification millions hours solar wind data from multiple missions can be replaced by automatic algorithms that discover, in mountains multi-dimensional data, real differences properties. In this paper we present how unsupervised clustering techniques used segregate different types wind. We propose use advanced reduction methods pre-process introduce Self-Organizing Maps...
Abstract. In magnetospheric missions, burst-mode data sampling should be triggered in the presence of processes scientific or operational interest. We present an unsupervised classification method for regions that could constitute first step a multistep automatic identification Our is based on self-organizing maps (SOMs), and we test it preliminarily points from global simulations obtained with OpenGGCM-CTIM-RCM code. The dimensionality reduced principal component analysis before...
Abstract Determining a set of nested pits to support the design an open pit mine that leads high economic value is crucial for strategic planning these operations; thus, practitioners rely on optimization methods finding high‐value solutions. However, current approaches are not sufficient as they lack at least one following features: fast computations optimal solutions, good geometric properties, and nestedness pits. In this work, we propose model address problem determining multiple by...
The DEEP projects have developed a variety of hardware and software technologies aiming at improving the efficiency usability next generation high-performance computers. They evolve around an innovative concept for heterogeneous systems: Cluster-Booster architecture. In it, general purpose cluster is tightly coupled to manycore system (the Booster). This modular way integrating components enables applications freely choose kind computing resources on which it runs most efficiently. Codes...
Abstract Relativistic Particle-in-Cell (PiC) methods are among the most reliable for investigation of plasma phenomena at particle scale. Standard explicit and semi-implicit PiC affected by numerical instabilities that restrain range admissible simulation parameters, prevent their application to large domains over long time scales. Here, we present a three-dimensional, fully-implicit algorithm relativistic simulations conserves energy exactly (to machine precision) eliminates instabilities,...
An accurate forecast of flare and coronal mass ejection (CME) initiation requires precise measurements the magnetic energy buildup release in active regions solar atmosphere. We designed a new space weather mission that performs such using optical instruments based on Hanle Zeeman effects. The consists two satellites, one orbiting L1 Lagrangian point (Spacecraft Earth, SCE) second heliocentric orbit at 1AU trailing Earth by 80° 80, SC80). Optical measure vector field multiple layers orbits...
Simulations of large-scale plasma systems are typically based on a fluid approximation approach. These models construct moment-based system equations that approximate the particle-based physics as fluid, but result lack small-scale physical processes available to fully kinetic models. Traditionally, empirical closure relations used close equations, which pressure tensor or heat flux. The more accurate relation, stronger simulation approaches kinetic-based results. In this paper, new terms...
We present a new web service (http://transplanet.irap.omp.eu/) dedicated to the modeling of planetary ionospheres. Thanks development made for IRAP ionospheric model IPIM, it uses unified description different ionized environments (presently Venus, Earth, Mars and Jupiter). The provides complete set parameters characterizing these environments, including concentration, velocities, temperatures, production rates ions electron heating rates. It is based on modular approach allowing selection...