The 10th Call for proposals for PRACE Project Access (Tier-0) was open from 10 September 2014, 12:00 (noon) CEST, until 22 October 2014, 12.00 (noon) CEST.
Five Tier-0 machines were available:
- “Curie”, Bull Bullx cluster (GENCI@CEA, France)
- “Fermi”, IBM Blue Gene/Q (CINECA, Italy)
- “Hornet” – Cray XC40 (GCS@HLRS, Germany)
- “MareNostrum”,IBM System X iDataplex (BSC, Spain)
- “SuperMUC”,IBM System X iDataplex (GCS@LRZ, Germany)
Projects from the following research areas:
Biochemistry, Bioinfo (8)
Transpore – Antibiotics transport through specific porins from Pseudomonas aeruginosa and rules for enhanced permeability
Project Title: Transpore – Antibiotics transport through specific porins from Pseudomonas aeruginosa and rules for enhanced permeability
Project Leader: Matteo Ceccarelli, University of Cagliari, IT
Resource Awarded: 8500000 core hours on Curie Thin Nodes (TN)
Tommaso D’Agostino, University of Cagliari / Department of Physics, IT
Susruta Samanta, University of Cagliari / Department of Physics, IT
The urgent need of new antibiotics to combat multidrug resistant bacteria and re-emerging pathogens demands a new method to develop the next generation of antibiotics. This is particularly true for Gram-negative bacteria, where a thick outer membrane with some channels can filter molecule. For that purpose, identification of physical/chemical properties of small molecules for an efficient permeability through porin channels is important to define new scaffolds. To achieve that goal, it is very important to have a clear idea of the permeation of molecules with different chemical properties through the porins and to understand the interaction of these molecules with the porins. This understanding will help us predict the optimal chemical and physical properties that give to antibiotic a better permeability. Since a reliable experimental technique does not exist yet, molecular simulations can be of great help.
In the project, we will study the translocation of two carbapenems (imipenem and meropenem) and some substrates (glycine, lysine and glutamic acid) through three porins – OpdP, OpdQ and OprE of Pseudomonas aeruginosa (PA). These are representative members of the Outer Membrane channels of PA. Once we have a clear idea of the behaviour of these channels, the understanding will be very helpful to understand the behaviour of multiple porins of the same families.
PA does not posses large and trimeric general channels, like OmpF/OmpC in enterobacteriaceae. The above selected channels are specific to some substrate and possess long internal loops that can modulate the eyelet region. Our initial studies have indicated that the relative motions of the internal loops in the porins are important for the translocation process. For that reason, it is of utmost importance to clearly understand the dynamics of the proteins in detail. Standard all atom molecular dynamics (MD) simulations are not very efficient in probing the whole phase space within a feasible timescale and hence, it demands implementation of advanced techniques.
We propose to use the ‘Replica Exchange Molecular Dynamics’ (REMD) and ‘Bias Exchange Molecular Dynamics’ (BEMD) simulation methods coupled with metadynamics. These methods demand substantial amount of computational power since they include running several parallel simulations using multiple bias factors and/or at different temperatures for the proper sampling of the protein dynamics. On the other hand it allows the sampling of the many degrees of freedom not explicitly declared in the metadynamics. This way, the system will be able to overcome energy barriers of the potential energy surface, providing a free energy description in a low-dimensional space, being all other degrees of freedom well explored with BEMD and REMD. This is supposed to provide us the understanding of the presence of affinity sites and the proper path for translocation for antibiotics and substrate molecules. Combining the outcomes, we will be able to predict the physical and chemical properties of the molecules responsible for better permeability.
drugs4FtsZ-Targeting FtsZ assembly for the development of new antibiotics
Project Title: drugs4FtsZ-Targeting FtsZ assembly for the development of new antibiotics.
Project Leader: Pablo Chacon, Institute of Physical Chemistry Rocasolano, ES
Resource Awarded: 10000000 core hours on SuperMUC
Jose Manuel Andreu, Centro de investigaciones Biologicas (CIB-CSIC) / Chemical and Physical Biology, ES
Cell division GTP-ase protein FtsZ is the main piece of the bacterial cytokinetic machinery and an emerging target for new antibiotics with which to fight the emergence of resistant pathogens. This cytoskeleton protein assembles forming filaments that are essential for cell division and whose dynamics is modulated by the binding and GTP hydrolysis. Only very recently the assembled sates of a straight FtsZ have been solved at atomic detail. Using this novel information, our goal is to perform all-atom simulations of long filaments bound to several biologically characterized modulators to better understand their inhibition mechanisms. In particular, we are interested in the relationship between the assembly molecular mechanism and the binding of modulators to help the rational design of new antibiotics.
ENZYMEFRICTION – The origin of internal friction in enzyme dynamics
Project Title: ENZYMEFRICTION – The origin of internal friction in enzyme dynamics
Project Leader: David De Sancho, University of Cambridge, UK
Resource Awarded: 17208576 core hours on Fermi
University of Helsinki – FI, University of Sussex – UK
Proteins are the workhorses of living organisms. Their dynamics determine their function, and hence it is essential to know how the solvent and internal interactions influence the time-scales for proteins to work. A recurring theme in the last two decades has been that intramolecular interactions set an upper limit for the dynamics of proteins. This is due to an effect known as “internal friction”, which was extensively used to explain experimental results, but lacked a molecular explanation. Last year, we have produced such explanation in the context of small peptides and proteins. From our results, it is clear that the simplest possible movements of protein chains, termed “torsional transitions”, actually determine internal friction. However, alternative mechanisms may be at play when we look at full enzymes and the dynamics of globular proteins.
NANOCARGO – design of self-assembling DNA nano-cages provided with pH and temperature controlled gates for the target-delivery of biomolecules
Project Title: NANOCARGO – design of self-assembling DNA nano-cages provided with pH and temperature controlled gates for the target-delivery of biomolecules
Project Leader: Mattia Falconi, University of Rome “Tor Vergata”, IT
Resource Awarded: 21000000 core hours on Fermi
Birgitta Knudsen, Aarhus University / Molecular Biology, DK
Francesca Cardamone, University of Rome “Tor Vergata” / Biology, IT
Federico Iacovelli, University of Rome “Tor Vergata” / Biology, IT
Andrea Idili, University of Rome Tor Vergata / Chemistry, IT
Francesco Ricci, University of Rome Tor Vergata / Chemistry, IT
The unique self-recognition properties of DNA, its high thermodynamic stability and the ease by which it can be synthesized, have in combination made of these molecules one of the most efficient building block for the creation of predesigned self-assembling nanostructures.
Aim of this project is to simulate innovative DNA nano-cage cargo designed for controlled and efficient encapsulation, release and cell target-delivery of biomolecules, assisted by a pH or temperature finely controlled “opening” and “closing” of the cage-structure. These capabilities will be accomplished by some stimuli responsive mechanisms specifically introduced into a earlier published DNA cage design, covalently closed and unable to encapsulate biomolecules, that has been simulated several times by our group. In this structure eight oligonucleotides form a truncated octahedron having an inner cavity with a diameter large enough to accommodate a medium size protein (~10-12 nm), but surrounded by a lattice plot small enough (~5.5-6.5 nm) to prevent the release of the encapsulated protein. The original DNA cage is composed by 12 B-DNA helices constituting the “edges” of the structure, interrupted by short single-stranded linkers composing the six “corners”.
Our intent is to insert in this nano-structure a programmable, pH triggered nano-switch substituting two of the four short singlestranded linker regions in one of the cage corners with a DNA triple helix gate (named pH-gate) whose pH-dependence can be finely tuned and modulated over more than 5 units of pH. For the temperature-dependence 2 hairpin including three base mismatches, able to unfold at a temperature around 40 °C, will be introduced in the nano-cage (named T-gate) as already described. The pH nano-switch has been designed as a on-off mechanism in which the triplex exists only at a well defined pH range (pH 5-6), while changing toward basic pH values the whole triplex completely unfolds. For this purpose two research groups, expert in the fields of DNA design and assembling and one molecular modeling and atomistic MD simulations group, will actively collaborate in this project. The pH-gate and T-gate nano-cages will be atomistically built and simulated using classical MD. The obtained simulative data will assess the conformational properties, the stability and the potentiality of the designed nano-switches giving important structural information about the designed mechanism that will allow to refine their experimental performance. With this simulative research we want to verify if chosen DNA oligos are suitable for the maintenance of the assembly and the exploitation of the opening/closing mechanism. The design and the assembling of an innovative DNA nano-cage capable to transport and release its content under specific pH/temperature conditions represents a major step forward in nano-bio-medicine.
Characterization of genetic risk variants in ASD families using a reference-free approach
Project Title: Characterization of genetic risk variants in ASD families using a reference-free approach
Project Leader: Daniel Geschwind, University of California Los Angeles, USA
Resource Awarded: 9011200 core hours on MareNostrum
Santiago Gonzalez, Barcelona Supercomputing Center / Life Sciences, ES
Josep M Mercader, Barcelona Supercomputing Center / Life Sciences, ES
David Torrents, Barcelona Supercomputing Center / Life Sciences, ES
Laura Perez Cano, University of California Los Angeles (UCLA) / Neurobiology, USA
Barcelona Supercomputing Center – ES, University of California Los Angeles (UCLA) – USA
Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder that affects 1% of the population. In recent years, the analysis of next generation sequencing data has advanced our understanding of the genetics of the disease, which is essential in order to develop effective treatments. However, given its large heterogeneity, most of the genes involved in ASD remain still unknown and the development of large-scale studies is strongly required. In order to overcome this challenge, the Hartwell Foundation, in collaboration with AGRE, has launched the largest worldwide initiative to characterize the genetic basis of ASD: the Autism Research and Technology Initiative (iHART), which has generated open-source whole genome sequencing data for 1000 ASD families. In order to get a more accurate detection of de novo and inherited genetic ASD risk variants in these families, we will apply a novel strategy by using a reference-free approach, SMuFin, to directly compare the genome between all affected and unaffected children and their corresponding parents. Considering both, the use of a large-scale whole genome sequencing data set and the effectiveness of the method proposed for detecting genetic variants in ASD families, this PRACE project aims to characterize disease-associated genetic variants, identify novel risk genes and gain comprehension of the biological pathways affected in ASD.
UNIDE: Understanding Insulin Dynamic Ensembles
Project Title: UNIDE: Understanding Insulin Dynamic Ensembles
Project Leader: Kresten Lindorff-Larsen, University of Copenhagen, DK
Resource Awarded: 7306240 core hours on Curie Thin Nodes (TN)
Fabio Doro, University of Copenhagen / Department of Biology, DK
Elena Papaleo, University of Copenhagen / Department of Biology, DK
Insulin is a peptide hormone with a central role in human metabolism. Autoimmune destruction of insulin-producing cells may lead to insulin deficiency and type 1 diabetes, a widespread disease that affects millions of people worldwide. Diabetes may be treated by injection of insulin, and the production of insulin analogues with tailor-made properties has substantially increased life quality of diabetics.
Insulin is stored as hexameric assemblies, which dissociate into monomers in the blood and activate cell-signalling pathways via binding to and activating the insulin receptor. While the structures of both hexameric and monomeric forms of insulin have been known for a while, structural models have very recently appeared for insulin in complex with its receptor. These structures revealed a surprising conformational change of insulin upon receptor binding, where a substantial region of the hormone opens up to expose a hydrophobic binding patch.
These results thus adds insulin to the growing list of proteins for which the ability to change structure is crucial for them to perform their biological function. A detailed and quantitative description of the conformational free energy landscape of insulin is thus not only necessary to understand insulin biology, but will also substantially aid in rationalizing sequence-function studies of insulin and the design of new insulin analogues for diabetes treatment.
Recent years have seen an explosion in our ability to characterize protein motions using both computations and experiments. From an experimental point of view, nuclear magnetic resonance (NMR) spectroscopy has emerged as the central most important technique that can be used to study protein dynamics. NMR spectroscopy has the unique ability to provide atomic level data that reports on the structure, dynamics and thermodynamics of the motions in proteins. Simultaneous with developments in experimental methods, the last few years have also seen important progress in our ability to use molecular dynamics simulations to study the structure and dynamics of proteins. This progress has been enabled in part by substantial improvements in both the quality and accuracy of the force fields used in simulations, and in methods that allow for increased sampling of the conformations of proteins.
We suggest utilizing new computational methods that integrate the strengths of NMR spectroscopy and molecular dynamics simulations to study the structural dynamics of insulin. Simulations restrained by experimental NMR data will result in a highly accurate description of insulin free in solution, and help provide insight into the extent to which the dynamics that underlies receptor binding is present in the absence of receptor also. To complement these, and to enable predictions of the properties of insulin variants, we will also perform simulations using state of the art physical models and enhanced sampling methods. Together, these computations will provide new and important insight into the structural dynamics of insulin and how this affects its function.
1/2M-GWIMP. Genotype imputation of 0.5 Million patients and controls suffering 44 genetic diseases using the 1000 Genomes whole genome sequences as reference panel
Project Title: 1/2M-GWIMP. Genotype imputation of 0.5 Million patients and controls suffering 44 genetic diseases using the 1000 Genomes whole genome sequences as reference panel
Project Leader: Josep Mercader, Barcelona Supercomputer Center, ES
Resource Awarded: 5632000 core hours on SuperMUC
Sílvia Bonàs, Barcelona Supercomputer Center / Life Sciences Department, ES
Marta Guindo, Barcelona Supercomputer Center / Life Sciences Department, ES
Friman Sanchez, Barcelona Supercomputer Center / Life Sciences Department, ES
Despite tremendous investments to identify causal genes for complex genetic diseases, such as diabetes, asthma, and others, through genome-wide association studies (GWAS), for the majority of the diseases, less than 10% of the variance attributable to genetic factors can be explained with the currently identified genetic variants. The goal of this project is to to make use of whole genome sequencing data from the 1000 Genomes project and UK10K Exome sequencing project to gain more information and statistical power in 69 GWAS datasets comprising at least 44 different diseases and more than 250.000 subjects. This will represent the analysis of around 1.750 billions of genotypes. In order to perform this analysis, supercomputing resources and advanced statistical and systems biology techniques are required. For this, we have developed a Genome-Wide IMPutation work-flow that makes use of COMPSs parallel programming framework, termed GWIMP-COMPSs. This tool will allow the identification of new genetic regions in the genome that increase the risk for several diseases, as well as fine mapping the already known susceptibility risk regions. We will then use manually curated disease specific biological pathways and networks reconstructed by our group. Also using novel in-house developed pathway and network analysis methods we will identify new key biological processes and genes perturbed in subjects suffering from a variety of complex diseases, ranging from metabolic diseases, to psychiatric or autoimmune diseases. We will experimentally validate the findings by replication of the associations in other cohorts and by analysing the functional effect of the discovered variants in independent cohorts for which DNA and tissue banks are available. We expect this project will allow the discovery of novel molecular mechanisms involved in a variety of complex diseases, opening new lines of research in several diseases, and will provide a novel framework to better exploit the existing available genetic data to better characterize the molecular bases of complex diseases.
Project Title: 3DBULB
Project Leader: Michele Migliore, National Research Council, IT
Resource Awarded: 10000000 core hours on Fermi
Carmen Lupascu, National Research Council / Institute of Biophysics, IT
Rosanna Migliore,National Research Council / Institute of Biophysics, IT
Francesco Cavarretta,University of Milan / Mathematics, IT
Michael Hines,Yale University / neurobiology , USA
The main aim of this project is to investigate the emergence of higher brain functions. In choosing a brain system to reveal these functions a few specific properties are highly desirable. The olfactory bulb seems to be exquisitely tailored to fulfill all these properties: in one synapse, an external signal goes from sensory neurons to olfactory cortex. The entire system is composed of approximately 100,000 principal cells and a few millions of interneurons; it is one of the most studied systems; similar experiments can be carried out and compared across species; and it is an essential stage for odor recognition processes. Most importantly, however, although it is intensively investigated experimentally in terms of odor discrimination and dynamics of cell responses, the functional effects of network self-organization and the consequences for odor learning and recognition remain difficult to understand and to explore at the system level.
To aid in solving this problem, in a previous PRACE project we have constructed a large-scale one-dimensional model of the olfactory bulb, and started to analyse how the spatio-temporal dynamics of lateral inhibition produce glomerular-related cell clusters during presentation of a series of relatively simple monomolecular odors. In this project, we extend the approach with a novel 3D model for processing natural odorants. This is particularly important, since complex natural odors activate a large portion of the olfactory bulb with dense representation. Under these conditions even a large network organization with canonical simplified neurons, abstract representations, or even our previous models limited to 1D, cannot solve the problem of understanding how natural odor discrimination is carried out efficiently by the actual neurons and their microcircuit connections. By exploring the system evolution in its natural full 3D architecture, a condition that is not possible to achieve in the experimental situation, the model will be able to make experimentally testable predictions on the relation between a natural odor input and the encoded output that will significantly advance of the state of the art. As with our previous model, this would be the first implementation of the olfactory bulb at this scale and realism, and we expect it to have a major impact in the field, promoting new experimental investigations and implementing a new framework to investigate the functions of a brain system.
This project is part of the intensive effort worldwide to attack this kind of problems using ICT methods, including computational modeling. One particularly important example is the recently funded Human Brain Project (HBP), one of the 2 EU flagships FET projects. The PI of this proposal is a member of the HBP Consortium, where he is involved in the implementation of brain system models. The project will be carried out in close collaboration with Dr. Michael Hines, at the Department of Neurobiology of Yale University (New Haven, CT, USA), who is another member of the HBP Consortium and responsible of implementing the cellular simulator.
SEDTRANS – Interface-resolved simulations of turbulent channel flow over a sediment bed
Project Title: SEDTRANS – Interface-resolved simulations of turbulent channel flow over a sediment bed
Project Leader: Wim-Paul Breugem, Delft University of Technology, NL
Resource Awarded: 10922667 core hours on Curie Thin Nodes (TN)
Bendiks Jan Boersma, Delft University of Technology, NL
Wim-Paul Breugem, Delft University of Technology, NL
Pedro Costa, Delft University of Technology, NL
Mathieu Pourquie, Delft University of Technology, NL
Turbulent flows laden with solid particles are found in many applications. Examples are the erosion of river bed, the enhanced mixing due to the presence of particles in a fluidized bed reactor, and the flocculation/sedimentation processes in the treatment of drinking water. The dynamics of such flows are poorly understood due to complex inter-particle and particle-turbulence interactions. With the recent improvement of numerical methods combined with the continuous growth in computing power, it is now possible to simulate the flow conforming an order of one million spherical particles. These simulations are commonly referred to as ‘interface-resolved’ and provide a very fine level of detail, because the fluid velocity and pressure are known everywhere in the flow, inclusively around the particle. The goal of this project is to study turbulent channel transport over a sediment bed using these interface-resolved numerical simulations. In particular, we will study in detail the physical mechanisms responsible for causing particles settling in a sediment bed to be re-suspended into the bulk of the flow.
(BINAPT) Application of biological nanopores for the detection of protein post-translational modification
Project Title: (BINAPT) Application of biological nanopores for the detection of protein post-translational modification.
Project Leader: Fabio Cecconi, Consiglio Nazionale delle Ricerche, IT
Resource Awarded: 35000000 core hours on Curie Thin Nodes (TN)
Mauro Chinappi, Istituto Italiano di Tecnlogia / Center for Life Nano Science , IT
Emma Letizia Bonome, Sapienza University of Rome / Dipartimento di Ingegneria Meccanica e Aerospaziale, IT
The project will provide a detailed understanding of the protein transport through a biological nanopore with the final purpose to shed light on a recent experimental data demonstrating the capability of nanopore-based sensor to detect post translational modification (PTM).
PTMs are covalent processing events that happen after the ribosomal translation changing the properties of a protein. PTMs occur on nearly all proteins and are thought to be associated to a variety of physiological/pathological processes. A specific class of PTMs involves the addition or modification to one or more amino acids (e.g. phosphorilation, adenylylation) and their characterization is expected to constitute a rapidly expanding area within biology.
Nanopores demonstrated a great versatility in the framework of bioanalytical applications, as they can work as single-molecule sensors able to detect, analyse and even manipulate nanoscale constructs. The working principle of nanopore sensing devices is very simple. A nanopore connects two electrolytic cells. A voltage applied across the two cells induces an electric (ionic) current. The single macromolecule capture, entry and subsequent translocation through the nanopore depend on the physico-chemical and geometrical properties of the analyte. Accordingly, the concentration, identity and certain microscopic features of the passing macromolecule (e.g., diffusion coefficient, volume, charge) can be inferred from the analysis of the variation in the ionic current trace. The key point of the nanopore sensing is hence to determine the properties of a macromolecule from the current “signature” associated to its passage. 123 In the last years nanopore sensors, either biological or solid state-based, have emerged as powerful, alternative tools for the detection and the analysis of various chemical compounds and biomolecules, such as RNA, DNA, peptides and proteins. Very recently, the possibility to employ nanopore technology to the detection of post translational modifications (PTMs) has been experimentally investigated (CB Rosen et al, Nature Biotechnology 2014). That work showed that the current signals allow to detect PTM the presence of PTM (phosphorilation in the particular case) on specific residues, paving the way to possible innovative protocols for PTM detection.
In this project, we will start analyzing the specific case reported in Rosen et al. Our results will constitute the first complete numerical dataset to be compared with accurate experimental results. MD simulations certainly offer a point of view on translocation that is complementary to experiments and that, together with them, will help to achieve a deeper understanding of the process with consequent impact on innovative sensing protocols. Moreover we will expand our analysis also to a specific situation relevant to biology where, at the present stage, no experimental result exist. Namely we will explore the possibility to detect lysine 120 acetylation of p53 protein (that is thought to be associated with cancer) using nanopore sensing.
The proposed computational pipeline was already successfully tested for a smaller system by the research team. Moreover, we have already performed scalability and performance tests on the actual set-up that will be employed in the project (using the resources obtained in a dedicated PRACE preparatory call).
PlasTitZir2: Plasticity in Titanium and Zirconium (2nd year)
Project Title: PlasTitZir2: Plasticity in Titanium and Zirconium (2nd year)
Project Leader: Emmanuel Clouet, CEA, FR
Resource Awarded: 12000000 core hours on Curie Thin Nodes (TN)
Nermine Chaari, CEA / Department of Materials for Nuclear Energy, FR
Olivier Mackain, CEA / Department of Materials for Nuclear Energy, FR
This proposal aims to build a physical sound model of plasticity in zirconium and titanium, by looking more precisely at the influence of interstitial solute atoms on the plastic behavior. Both metals have similar properties arising from their hexagonal compact (hcp) crystallography and from their alike electronic structure. In particular, their plastic behavior is strongly influenced by the interaction of screw dislocations with impurities, like oxygen, nitrogen, carbon, or sulfur which all lie in the interstitial sites of the crystal lattice. At a macroscopic scale, a small solute addition to either Zr or Ti leads to an important hardening, with a plastic flow controlled by thermal activation instead of an athermal regimes in pure metals. This hardening cannot be associated with a simple elastic interaction between dislocations and impurities, but seems to be induced by a change of dislocation properties trough a modification of their core structure by the interstitial atom. The purpose of this project is therefore to understand how interstitial impurities modify dislocation properties in pure Ti and Zr, so as to model then the hardening induced by their addition. The role of oxygen has been examined first and the modeling approach is now extended to carbon, nitrogen and sulfur.
Atomistic simulations are the good tool to study dislocation cores. In Ti and Zr, dislocations glide in the prism planes of the hcp crystal. The relative ease of prismatic compared to basal glide has been shown to be linked to the ratio of the corresponding stacking fault energies, which in turn is controlled by the filling of the valence d band which induces an angular dependence of the atomic bonding that cannot be neglected. An important consequence for the modeling at an atomic scale of plasticity in Ti and Zr is that one cannot rely on simple empirical potentials, like EAM potentials, but one has to take full account of the electronic structure. We therefore propose to use ab initio calculations based on the density functional theory to study the interaction of screw dislocations with interstitial impurities in both Zr and Ti. This project will provide quantitative data for the modeling then of dislocation mobility in presence of impurities. Ab initio calculations with both an interstitial solute atom and a screw dislocation will be performed to see how the impurity modifies the core structure of the dislocation. A result of these calculations will also be the interaction energies, which will be used then to build a thermodynamic modeling of the impurity segregation on the dislocation.
A deep understanding of the interaction between screw dislocations and the different interstitial solute atoms will be gained from these atomistic simulations. This will allow us to understand how solute modify plasticity in both Ti and Zr, and to build a physical model of their plastic behavior. Such a model could be then extended to industrial alloys by incorporating contributions of other alloying elements.
ISSDIL – Interfacial Structure and Solvation Dynamics in Ionic Liquids
Project Title: ISSDIL – Interfacial Structure and Solvation Dynamics in Ionic Liquids
Project Leader: Maxim Fedorov, University of Strathclyde, UK
Resource Awarded: 30000000 core hours on Hornet
Ari Paavo Seitsonen, Ecole Normale Supérieure de Paris / Département de Chimie, FR
Maksim Mišin, University of Strathclyde / Department of Physics, UK
Vladislav Ivanistsev, University of Tartu / Institute of Chemistry, EE
Rustam Z. Khaliullin, University of Zurich / Physikalisch-Chemisches Institut, CH
In the present project, we aim to apply Ab Initio Molecular Dynamics (AIMD) method to investigate the solvation phenomena at electrochemical interfaces in ionic liquids (ILs). High concentration of charges and (electro)chemical stability make ILs very attractive solvents from the application point of view. From the fundamental point of view, the interfacial solvation in ILs is strongly influenced by oscillations of ion charge density and electronic charge density that are specific for both ILs and electrode material, respectively. Accordingly, we will focus on the charge and electric potential distributions across IL–electrode interfaces, which are at heart of electrochemistry where ILs can have a number of applications (e.g. in supercapacitors, batteries, solar panels, electro-deposition, electro-wetting and micro-electroactuators). IL interfaces with nanocarbon materials are of particular importance due to the extreme growth of interest in using IL–nanocarbon composites for the development of new flexible energy storage devices and miniaturized electronics (e.g. electrolyte-gated transistors). It is known that molecular-scale interfacial effects determine operational performance of such systems, both at micro- and macro-level. However, experimental techniques cannot directly access many important effects in ILs at the molecular level due to (i) resolution limitations; (ii) non-crystalline nature of the systems; (iii) fast processes at the femto- or picosecond time scale. In turn, classical molecular dynamics (MD) simulations can not describe explicitly several important effects at the quantum level (e.g. charge transfer, polarization, image charges). The AIMD directly incorporates electronic degrees of freedom into the model and, thus, offers a unique possibility to access the whole range of micro-scale effects in ILs at neutral and charged interfaces. Until recently, use of AIMD simulations was limited due to its expensiveness (compared to classical MD), as well as high viscosity and complex molecular structure of ILs. Nevertheless, progress in supercomputer technologies and computational chemistry software makes it possible to apply this method to realistic IL systems. We intend to perform AIMD simulations of 1-ethyl-3-methylimidazolium tetrafluoroborate near graphene surface (EMImBF4–Gr) with a focus on the following: (i) the role of the semi-metal nature of graphene in the formation of the electrical double layer at the graphene–IL interface, (ii) electron density redistribution between graphene and IL, (iii) the balance between electrostatic and dispersion interactions in the IL interfacial layer and (iv) the fast solvation dynamics of IL near the graphene surface.
We plan to use the system size in the order of thousand of atoms at simulation time scales up to 20 picoseconds. We note that such large-scale AIMD simulations require computer resources unavailable from Tier1 and smaller computation platforms.
ACTIRAF- An atomic level understanding of paradoxical activation in the B-Raf dimer: Implications for drug design
Project Title: ACTIRAF- An atomic level understanding of paradoxical activation in the B-Raf dimer: Implications for drug design
Project Leader: Francesco Gervasio, University College London, UK
Resource Awarded: 18198279 core hours on Hornet
Kristen Marino, University College London / Chemistry, UK
Giorgio Saladino, University College London / Chemistry, UK
Protein kinases play a key role in cell signaling and control a number of cellular processes including cell growth, proliferation and differentiation. It is therefore unsurprising that dysfunction of kinases has been implicated in a number of diseases, most notably cancer, but also diabetes and inflammation. A crucial part of their regulatory processes is the tightly controlled transition between a catalytically active conformation and one or more inactive conformations which means that the kinase is “on” only when needed. The aim of this project is to understand the interplay between dimerisation, drug-binding and activation dynamics in the case of the RAF kinases, a family of kinases that play a key role in many cancers, including melanoma.
NanoTox – Camouflaging nanoparticles for reduced toxicity through limited aggregation and membrane disruption
Project Title: NanoTox – Camouflaging nanoparticles for reduced toxicity through limited aggregation and membrane disruption
Project Leader: Paraskevi Gkeka, Biomedical Research Foundation, GR
Resource Awarded: 8533333 core hours on Curie Thin Nodes (TN)
Zoe Cournia, Biomedical Research Foundation / Clinical, Experimental Surgery, & Translational Research, GR,
Apostolos Klinakis, Biomedical Research Foundation / Clinical, Experimental Surgery, & Translational Research, GR
Klaus Liedl, Leopold-Franzens-University Innsbruck / Faculty of Chemistry and Pharmacy, AT
Ioanna Zergioti, National Technical University of Athens / Department of School of Applied Mathematics and Physical Sciences, GR
Lev Sarkisov, University of Edinburgh / Institute for Materials and Processes School of Engineering, GR
The increasing applications of nanoparticles (NPs) in a variety of products, ranging from drug and gene delivery materials to consumer goods like paints, set as a prerequisite that these engineered nanomaterials will come in contact with human cells and the environment without damaging essential tissues. Recently, NPs have attracted attention as a new promising class of drug delivery vectors and agents causing targeted tumour cell destruction. However, NPs may unintentionally affect healthy cells causing great damage. Therefore, before NPs can be safely used, their cytotoxicity needs to be further assessed and quantified. The first step into assessing the NP cytotoxicity requires a thorough understanding of the NP-membrane interaction mechanism. The main objective of this joint experimental/computational project is to systematize how NP surface attributes, including functional groups, like chitosan, and peptide coating affect their aggregation and their interaction and internalization mechanism across the cell membrane. A cell-penetrating peptide (CPP), namely Pep-1, capable of efficient translocation through the cell membrane as well as chitosan and polyethylene glycol (PEG) will be used as NP coating. This type of coatings will be explored as potential programmable drug delivery vectors. Extensive Coarse-Grained Molecular Dynamics simulations will be performed in order to gain insights into the mechanism of NP-membrane interactions at microscopic detail. The membrane considered in the present study will be made of two types of lipids and cholesterol. Such a multicomponent mixture exhibits a complex phase behavior with regions of structural and compositional heterogeneity, resembling lipid rafts found in physiological membranes. As there is currently no existing study that quantifies the role of lipid rafts in NP membrane translocation, this project will advance the state of the art by providing invaluable information about NP-membrane interactions using a realistic membrane description. NPs that exhibit optimal features like direct cellular entry will be subsequently assayed in vitro for toxicity to assess their properties in living cells. In this proposal, we aim to exceed state-of-the-art timescales of NP-membrane simulations by an order of magnitude in order to capture the intricate mechanisms of spontaneous NP translocation through a lipid bilayer. Simulation of the aforementioned problems requires access to large-sized partitions (1,024 cores) with a turn-around time of ~2,000,000 CPU hours per run. Results from this project will have an impact in the area of targeted drug delivery, providing much needed insight for design of non-toxic NP drug delivery agents.
COMPHOTOCAT – Computational design of TiO2 based nanoparticles for improved photocatalytic activity towards water splitting under visible sunlight
Project Title: COMPHOTOCAT – Computational design of TiO2 based nanoparticles for improved photocatalytic activity towards water splitting under visible sunlight
Project Leader: Francesc Illas, Universitat de Barcelona, ES
Resource Awarded: 50000000 core hours on MareNostrum
Volker Blum, Duke University / Mechanical Engineering and Materials Science, USA
Bjoern Lange, Duke University / Mechanical Engineering and Materials Science, USA
Stefan Bromley, Universitat de Barcelona / Quimica Fisica & IQTCUB, ES,
Kyong Chul Ko, Universitat de Barcelona / Quimica Fisica & IQTCUB, ES,
Oriol Lamiel Garcia, Universitat de Barcelona / Quimica Fisica & IQTCUB, ES,
Francesc Viñes, Universitat de Barcelona / Quimica Fisica & IQTCUB, ES,
The seminal work of Fujishima and Honda (1972) on water photoelectrochemical photolysis opened new, interesting and technologically relevant opportunities in photocatalysis, in general, and in photocatalytic water splitting, in particular. The possible technological implications are enormous since it offers a direct way to take advantage of solar energy to obtain hydrogen just from water, thus providing a clean and sustainable fuel. This is even more the case since it has been shown that water splitting by TiO2 nanopowders spontaneously occurs under ultraviolet irradiation which is appealing since the whole process can be carried out without needing to rely on an electrochemical cell and expensive electrodes using scarce and expensive noble metals (Pt). The problem of this approach is that H2 and O2 are simultaneously produced but reactors have already been designed that allow for direct separation of these gases.
Unfortunately, over 40 years of research and thousands of papers have not been enough to find photocatalysts with activity in the visible region of sunlight higher than that of TiO2 under ultraviolet radiation which severely limits practical applications since sunlight at the Earth surface contains 2-3% of ultraviolet radiation only. Inspiration-driven, trial and error research is not efficient calling for new directions. This is clear when analyzing the impressive recent experimental advances showing that tailored TiO2 nanoparticles, stoichiometric or conveniently chemically modified, can be synthesized with predefined size and shape and exhibiting distinct photocatalytic activity in various reactions including water splitting. Unfortunately, the necessary link between the properties of a given type of nanoparticle and its photocatalytic activity has not yet been established.
A deep analysis of the electronic structure of these photocatalytic materials is necessary but going beyond the common solid state or surface chemistry paradigms, which often focus on ground state properties only and with little insight into the chemistry of the excited states involved in the photocatalytic reactions. The present proposal provides an alternative roadmap and a different strategy. It starts by building realistic TiO2 nanoparticle models with different sizes, shapes, compositions, and environments. From a systematic study of their electronic structure, a database will be constructed and candidate nanoparticles with appropriate absorption spectra in the visible light window selected, thus fulfilling a first necessary but not sufficient condition for photocatalysis. Excited states of interest on the candidate tailored nanoparticles will be investigated to select those with spatially well-separated electron-hole pairs (minimal recombination) and exhibiting long enough mean lifetimes (maximum photon efficiency). Further, effects of environment (solvent), on the electronic structure and nature of excited states will also accounted for. Finally, analysis of the reactivity of the nanoparticles towards photocatalytic reactions, mostly water splitting for hydrogen generation, in the appropriate electronic state will complete the picture. The task is huge, not exempt of risk, and requires enough computational resources to effectively deliver the necessary new knowledge for an efficient rational design.
Development of novel fluorescent protein based neural sensors for brain imaging
Project Title: Development of novel fluorescent protein based neural sensors for brain imaging
Project Leader: Pavel Jungwirth, IOCB AS CR, CZ
Resource Awarded: 400000 core hours on Curie Hybrid Nodes
Except for a few primitive organisms all living multicellular animals have a complicated nervous system. Its centre, the brain, is the most intricate part. In order to understand how the brain works, we must be able to observe electrical signals that travel through the axons and dendrites of its neurons. Our collaborating laboratory takes part in the development of the technique of two-photon polarization microscopy (2PPM). This experimental technique will lead to development of novel fluorescent protein-based neural sensors of highly desirable properties suitable for brain imaging. In the theoretical part of this project, we plan to assess the various prospective candidates for a successful neural sensor suggested by the collaborating laboratory. The theoretical approach allows for rational, more systematic and faster development of the proposed neural sensors than possible in experiments.
Towards a modeling of realistic quantum dots/graphene-based photovoltaic devices using DFT
Project Title: Towards a modeling of realistic quantum dots/graphene-based photovoltaic devices using DFT
Project Leader: Frédéric Labat, Chimie ParisTech, FR
Resource Awarded: 20000000 core hours on Curie Thin Nodes (TN)
Ilaria Ciofini, Chimie ParisTech / Institut de Recherche de Chimie Paris, CNRS, FR
Bartolomeo Civalleri, Università di Torino / Dipartimento di Chimica, IT
Roberto Dovesi, Università di Torino / Dipartimento di Chimica, IT
Roberto Orlando, Università di Torino / Dipartimento di Chimica, IT
With the decrease of available fossil energy sources and regarding their damaging impact on the environment, an increasing effort is invested in the research of alternative, environmentally friendly and renewable energy sources. Among these, photovoltaic solar technology is expected to play a key role in future energy breakdown. In solar cells, the photovoltaic effect involves generation of electrons and holes in a semiconductor device under illumination, and subsequent charge collection at opposite electrodes. Since photon absorption usually leads to electron-hole pairs formation, charge collection therefore requires these pairs to be separated at the heterojunction interface between materials of different ionization potentials or electron affinities. This is typically the case in the well-known dye-sensitized solar cells (DSSCs), in which a monolayer of dyes grafted onto the surface of a highly porous semiconductor material (typically TiO2 ) is the photoactive part.
In the last few years, quantum dots (QDs) as light harvesters in solar cells have gained marked attention due to their size-tuned optical response and their potential in exceeding the Shockley-Queisser efficiency limit of p-n junction photovoltaic cells via multiple exciton generation.
Major challenges in quantum dot sensitized solar cell (QDSCs) however include efficient separation of electron-hole pairs formed upon photoexcitation of the QDs, and facile electron transfer to the electrode, which is typically achieved by combination with nanomaterials with suitable band energy as electron acceptors for the excited QDs: TiO2 for instance. Despite their encouraging properties, their efficiency unfortunately still remains around 5% at best, that is far from the 33% of the Shockley-Queisser limit, and from the 11,9% of the best performing DSSCs, for instance.
Carbon-based nanomaterials of suitable band energies, such as fullerenes, single-walled carbon nanotubes (SWNTs) or graphene, have been recently proposed as suitable electron acceptors in QDSCs. Although the performances of fullerene or SWNTs-based QDSCs are still too low to meet the requirements for commercialization, graphene-based QDSCs with a layered graphene/QD structure were recently proposed as a significant breakthrough in graphene-based QDSCs technology. By stacking up to ten QDs/graphene bilayers, this nanoscale thin layered structure was proposed to help to solve the two inherent problems associated to QDSCs mentioned above: charge collection and transport. In particular, it was shown that in contrast to SWNTs-based QDSCs, a uniform distribution of QDs on the graphene surface is achieved in this case, leading to a significant boost in photovoltaic performances with up to 20-fold enhancement compared to corresponding SWNTs-based systems.
In this proposal, we wish to theoretically investigate these graphene-based QDSCs using density functional theory techniques, in order to address several key points in the photovoltaic performances of these systems. More precisely, these include:
– functionalization of QDs and adsorption on the graphene surface, considering graphene/QDs linkers based on inorganic ligands such as OH-, which have already been been shown to facilitate charge transfer compared to long-chain organic ones in QDSCs application.
– variation of the number of QDs/graphene bilayers in the model.
The models considered here are of very large sizes, requiring Tier-0 computational resources.
Ab-initio simulation of transition metal dichalcogenide devices
Project Title: Ab-initio simulation of transition metal dichalcogenide devices
Project Leader: Mathieu Luisier, ETH Zurich, CH
Resource Awarded: 60600000 core hours on Fermi
Sascha Brueck, ETH Zurich / Department for Information Technology and Electrical Engineering, CH
Petr Khomyakov, ETH Zurich / Department for Information Technology and Electrical Engineering, CH
Aron Szabo, ETH Zurich / Department for Information Technology and Electrical Engineering, CH
The first experimental demonstration of a single-layer MoS2 transistor in 2011 has renewed the interest of the scientific community for layered transition metal dichalcogenides (TMDs), which are of the form MX2 where different transition metals M=Mo, W, Ti, Hf, Zr, Pd, Pt, … and chalcogens X=S, Se, or Te are combined together. It is believed that semiconducting TMDs might possibly replace Silicon as the building blocks of logic switches when their gate length will shrink below 15 nm. The excellent electrostatics properties of TMDs coming from their two-dimensional (2-D) nature as well as their relatively large band gaps support this idea. However, there are still wide uncertainties concerning the intrinsic electron and hole mobility of these materials that is not precisely known in most cases. This PRACE project intends to solve this issue by computing the mobility of 21 different TMDs made of a single up to five layers and determine what configuration can challenge Silicon. For that purpose, an ab-initio quantum transport solver called OMEN will be used. It relies on the Non-equilibrium Green’s Function (NEGF) formalism and the density-functional theory (DFT) through a transformation of plane-wave Hamiltonian matrices into a maximally localized Wannier function basis. A mobility database will be provided to experimental groups working with TMDs as an output of this work.
TRANSOM – Temperature PRoperties of VacANcies in TungSten using Ab-initiO Methods
Project Title: TRANSOM – Temperature PRoperties of VacANcies in TungSten using Ab-initiO Methods
Project Leader: Mihai-Cosmin Marinica, CEA, FR
Resource Awarded: 23000000 core hours on MareNostrum
Rebecca Alexander, CEA / Departement od Materials for Nuclear Energy, FR
Manuel Athènes, CEA / Departement od Materials for Nuclear Energy, FR
Tony LELIEVRE, Ecole des Ponts ParisTech / CERMICS (Centre d’Enseignement et de Recherche en Mathématiques et Calcul Scientifique), FR
Gabriel Stoltz, Ecole des Ponts ParisTech / CERMICS (Centre d’Enseignement et de Recherche en Mathématiques et Calcul Scientifique), FR
This project aims at improving our knowledge of the aging properties of tungsten under extreme conditions by resorting to atomistic simulations based on electronic structure calculations. Performing the calculations at the finest scale i.e. the electronic scale, is the only way to overcome the limitation of empirical methods currently employed in the field of material engineering. The Service de Recherches de Métallurigie Physique (SRMP) of CEA Saclay has developed an expertise on the application of electronic structure calculations methods of DFT (Density Functional Theory) in order to describe defects in materials.
The properties of mono-vacancies and small vacancy-clusters underpin any multi-scale model predicting the evolution of materials under extreme conditions. In this project we propose to perform a comprehensive analysis of the high temperature properties of mono- and di- vacancies in tungsten by means of ab-initio calculations of the free energy landscape. These thermodynamic properties are hardly accessible in experiments.
The goal of this project is to shed light on the temperature impact over the energy landscape of mono- and di- vacancy in tungsten alloys and to quantify the dependence of the computed free energies on temperature up to melting temperature. In order to achieve this goal we will combine our expertise in point defects and DFT with the knowledge of researchers in mathematics and statistical physics / free energy field from Molecular and Multiscale Modelling group at CERMICS (Centre d’Enseignement et de Recherche en Mathématiques et Calcul Scientifique, École des Ponts ParisTech, France) and of the Matherials project-team at INRIA (France). Recent adaptive molecular dynamics methods will be used. These calculations require high accuracy and CPU time and are well adapted to massively parallel computers having a perfect linear scaling. Our results will be directly compared with the experiment and the proposed calculations will also provide to community a key input parameter for subsequent multi-scale modelling in the field of future energy systems.
Direct Numerical Simulation of Gravity-Driven Bubbly Flows
Project Title: Direct Numerical Simulation of Gravity-Driven Bubbly Flows
Project Leader: Assensio Oliva, Technical University of Catalonia, ES
Resource Awarded: 22000000 core hours on MareNostrum
Néstor Balcázar, Technical University of Catalonia / Thermal Engines, ES
Jesús Castro, Technical University of Catalonia / Thermal Engines, ES
Oriol Lehmkuhl, Technical University of Catalonia / Thermal Engines, ES
Except for a few primitive organisms all living multicellular animals have a complicated nervous system. Its centre, the brain, is the most intricate part. In order to understand how the brain works, we must be able to observe electrical signals that travel through the axons and dendrites of its neurons. Our collaborating laboratory takes part in the development of the technique of two-photon polarization microscopy (2PPM). This experimental technique will lead to development of novel fluorescent protein-based neural sensors of highly desirable properties suitable for brain imaging. In the theoretical part of this project, we plan to assess the various prospective candidates for a successful neural sensor suggested by the collaborating laboratory. The theoretical approach allows for rational, more systematic and faster development of the proposed neural sensors than possible in experiments.
EBAPAM – Emergent behavior and active patterns in active materials
Project Title: EBAPAM – Emergent behavior and active patterns in active materials
Project Leader: Ignacio Pagonabarraga, University of Barcelona, ES
Resource Awarded: 22000000 core hours on MareNostrum
Francisco Alarcon-Oseguera, University of Barcelona / Fundamental Physics, ES
Joan Codina, University of Barcelona / Fundamental Physics, ES
Eloy Navarro, University of Barcelona / Fundamental Physics, ES
Raul Cruz-Hidalgo, University of Navarra / Faculty of Science, ES
Andrea Scagliarini, University of Rome Tor Vergata / Physics, IT
Active suspensions constitute a class of heterogeneous, structured materials, characterized by the internal motion of their constituent particles. Microorganisms, microrobots or molecular motor-biofilament mixtures, are examples of such systems in which energy consumption (either through their internal metabolisms, appropriate catalytic reactions or ATP dissipation) leads to self propulsion and/or internal stress generation, which provide a means to induce dynamic interactions and impart these materials with unique properties. The internal activity of active suspensions provides a new venue for self-assembly and pattern formation in these systems which are intrinsically out of equilibrium, The coupling of active particles through the embedding solvent constitutes a basic source for long-range dynamic interactions.
The intrinsic non-equilbrium character of these materials confer them with qualitatively new properties with respect to their equilibrium counteparts. As a result, these systems open new venues to the development of materials with specific new functionalities. Their non-equilibrium nature makes them very responsive, sensitive to applied forces and malleable.
Despite this great potential, still little is known about their behaior, how to achieve specific morphologies or textures, or how to control them. A basic understanding is required to clarify procedures to manipulate them and to rationalize how to produce them or to ensure specific functionalities.
The study of these systems is challenging; their intrinsic non-equilibrium nature requires to simulate their temporal evolution including not only the basic constituents (the self-propelling particles or active elements) but also the environment they move in, which is affecetd by the presence of these elementary active constituents. Moreover, they are usually characterized by long-ranged correlations which imply in turn the need to simulate long system sizes over long periods of time.
In this proposal we will focus on the study of the fundamental properties of active materials. We will consider two classes of such materials and will identify their mechancial properties and their sensitivity to external forcings. Specifically, we will be interested in studying model swimmers ( a class of self-propelling particles which displace inside a fluid medium) trapped at liquid interfaces. These systems are relevant for potentially new materials, such as active emulsions or gels. There exists recent experimental realizations of active drops, hence analyzing the implication of active emulsions is a particularly timely subject with relevant potential appiciations. The second class of materials is constituted by model swimmers with ambivalent interactions (as accomplished experimentally with Janus colloids). These systems exhibit a rich phase diagram already in equilibrium. The competition between such interaction and active forces associated to particle activity opens new directions to manipulate matter and develop novel structures. We will exploit supercomputing environments to analyze systematically the new features that characterize these active materials and will analyze the different textures that activity and the hydrodynamic stresses that active colloids give rise to.
APAM, Aquatic Purification Assisted by Membranes
Project Title: APAM, Aquatic Purification Assisted by Membranes
Project Leader: Sabine Roller, University of Siegen, DE
Resource Awarded: 42310000 core hours on Hornet
Clean water is one of the most important resources for mankind. This project aims to provide a contribution towards the means of securing this resource by providing detailed simulations of the processes close to membranes in electro-dialysis. These processes are involving multiple scales and different physical phenomena, like electromagnetic fields, fluid dynamics and diffusive mass-transport. A rigorous assessment of all the interactions at hand is not feasible, and models, that enable improved technical devices are only possible through large scale coupled simulations. Recently, there was a development that allows to decrease the necessary effort for seawater desalination by electro-dialysis with the help of a new membrane, that selectively separates only salt ions from the fluid. A detailed simulation of this membrane and possible extensions are at the core of the APAM compute time project.
Transport properties and thermodynamic properties of ionic liquids.
Project Title: Transport properties and thermodynamic properties of ionic liquids.
Project Leader: Sten Sarman, Stockholm University, SE
Resource Awarded: 10000000 core hours on Hornet
Aatto Laaksonen, Stockholm University / Materials and Environmental Chemistry, SE
Yonglei Wang, Stockholm University / Materials and Environmental Chemistry, SE
The purpose of this project is to calculate the viscosity of ionic liquids consisting of an alkylated phosphonium ion and a borate ion. Such systems present a great potential for becoming the new generation of lubricants. Depending on subtle details of the actual structure of the phosphonium ions and the borate ions the rheological properties can vary considerably. This means that time-consuming synthesis and testing of many substances must be carried out in order to obtain lubricants with optimal properties. Therefore, it is necessary to perform molecular dynamics simulations to calculate thermodynamic data and transport coefficients, not least the viscosity, to guide the synthesis of new promising substances.
The most immediate way to calculate the last mentioned transport coefficient is to simulate a measurement in a real viscosimeter by performing shear flow simulations. This can be done by applying the SLLOD equations of motion, which are an exact description of adiabatic planar Couette flow. The viscosity is found by calculating the ratio of the shear stress and the shear rate a few different shear rates and extrapolating to zero shear rate. It is usually possible to find an interval where the shear rate is so low that the linear relation between the shear rate and shear stress still is valid and high enough to overwhelm the thermal fluctuations so that the signal-to-noise ratio remains nonzero. In the limit of zero shear rate, linear response theory can be used to derive a fluctuation relation or a Green-Kubo relation for the viscosity. This relation is a time integral of the time correlation function of the shear stress, which can be evaluated by ordinary equilibrium molecular dynamics simulations.
These two methods have successfully been applied to evaluate the viscosity of rather complex liquids such as various alkanes, which are components of lubricants. However, these systems consist of neutral molecules which is an advantage because all the molecular interactions are short ranged, so that there will be rather few interactions whereby the simulations become faster. Unfortunately, most molecules of practical interest are charged or display charge separation either because they include ions or there are differences between the electronegativities and electron affinities between different atoms in the molecule. Thus the atoms must be decorated by full or partial charges and Ewald summations must be undertaken in order to evaluate the electrostatic interactions correctly which slows down the simulation. Moreover, when irregular molecules such as the above mentioned phosphonium ions are present the relaxation times increase, so that longer simulations are needed. Thus it is not possible to let these simulations run on ordinary workstations or moderately large parallel processors available at many supercomputer centers. So far a force field has been developed for an atomistic model of the phosphonium ion and the borate ion. Preliminary simulations using available computational resources have shown that the model reproduces the thermodynamic properties and the viscosity very well compared to experimental measurements. However, in order to finish the project successfully much more computation time is needed.
Conformational switching of peptides in materials-binding recognition; elucidating the role of intrinsically disordered peptides via advanced conformational sampling.
Project Title: Conformational switching of peptides in materials-binding recognition; elucidating the role of intrinsically disordered peptides via advanced conformational sampling.
Project Leader: Tiffany Walsh, Deakin University, AU
Resource Awarded: 5820000 core hours on Hornet
Mark Rodger, University of Warwick, UK
Peptide sequences that can undergo predictable conformational transformation upon adsorption to a substrate of a given materials composition, under aqueous conditions, offer a promising route to inducing and stabilizing surface adsorbed structures that can deliver enhanced functionality, e.g. to initiate the cascade of reactions related to the osseointegration process associated with bioimplants. Such materials-binding peptides are known to be “intrinsically-disordered” (IDPs). IDPs defy the conventional “one structure, one function” paradigm – in the unbound state these typically feature a broad conformational ensemble (often referred to as ‘random coil’). In the adsorbed state, some IDPs are thought to either undergo surface-induced folding. Currently, we lack the structural knowledge required to fully harness these switchable properties for the benefit of e.g. design of responsive bioimplant surface coatings. It is challenging for experimental approaches to gain this structural details of the peptide in the surface-adsorbed state. Molecular simulation approaches can contribute valuable input in this regard, but these also face significant technical challenges. In this project, we will exploit access to Europe’s premier supercomputing facilities, in partnership with novel advanced structural prediction techniques, and alongside recent experimental advances, to establish the first generalizable rules for understanding and designing conformationally-switchable peptides for materials recognition applications.
Earth System Sciences (3)
THROL – Transient Holocene
Project Title: THROL – Transient Holocene
Project Leader: Pascale Braconnot, CEA-DMS/LSCE, FR
Resource Awarded: 20303712 core hours on Curie Fat Nodes (FN)
Arnaud Caubel, CEA / DSM/LSCE, FR
Olivier Marti, CEA / DSM/LSCE, FR
Didier Swingedouw, CNRS / EPOC, FR
Marie-Alice Foujols, IPSL / Pôle de Modélisation, FR
Past climates offer a wide range of natural climate experiences that allow inferring how the different components of the climate system and their response to different external perturbations enter into play to produce climate characteristics. Paleoclimate studies shades light on recent climate warming and future climate change. 123 This computing project is a first step in a multi-year scientific grand challenge in which we intend to simulate long term climate transitions with the same model version as the one used for future climate projections. Our objective is to simulate the transition between the mid-Holocene (6 000 years ago) and 1 500 years ago using a global state of the art climate model that includes an interactive carbon cycle, with an horizontal and vertical resolution that was up to now out of reach for such long simulations. This original set up will allow us to progress on the understanding of the hydrological changes in the tropical regions and on the relative role of forced versus internal variability in the variation of the Intertropical convergence zone, the Afro Asian monsoon and the El Niño Soutern Oscillation. 123 We will be able to fully analyse multidecadal variability in a context of a long term climatic trend over a long period with a sufficient number of events for statistical inference. In addition the fact that both physical and biogeochemical variables will be simultaneously simulated offer the possibility to use the wide range of independent paleoclimate reconstructions from marine, ice and terrestrial records to evaluate the model results and test the consistency of the different paleoclimate records. 123 A small ensemble will allow to test different aspects of the model physics, and thereby of climate mechanisms. The ensemble is designed to test systematic differences induce by atmosphere, ocean or land-surface model physics, which will help us to understand differences between different models. We will analyse how the representation of convection and clouds affect the changes in tropical variability and what are the implications for tropical and extratropical teleconnections, including possible links with drought or floods over land. We will also analyse the impact of vegetation feedback, and of the ocean mixed layer characteristics. 123 These simulations will be unique and their results will be of great value for lots of people: climate modellers, people studying paleoclimatic data, and outside the climate community. We plan the model outputs to the international community, following what is done for the PMIP simulations (through Earth System Grid Federation). Past simulations are also of great value for human or diversity studies. As an example, PMIP simulations have been widely used to run ecologic niches models, as it can be inferred from the citation list of Braconnot et al. (2007). More than 2/3 of the citations come from the biodiversity community and not from the climate community. We will then organize ourselves so that these people can have access to the variable they need for the niches model and provide them with a reference atlas of the results.
OCCIPUT — OceaniC Chaos: ImPacts, strUcture, predicTability
Project Title: OCCIPUT — OceaniC Chaos: ImPacts, strUcture, predicTability.
Project Leader: Thierry Penduff, CNRS, FR
Resource Awarded: 16000000 core hours on Curie Thin Nodes (TN)
Laurent Bessières, CERFACS / SUC-GLOBC, FR
Eric Maisonnave, CERFACS / SUC-GLOBC, FR
Marie-Pierre MOINE, CERFACS / SUC-GLOBC, FR
Laurent TERRAY, CERFACS / SUC-GLOBC, FR
Sophie VALCKE, CERFACS / SUC-GLOBC, FR
Bernard BARNIER, CNRS / LGGE, FR
Jean-Michel BRANKART, CNRS / LGGE, FR
Pierre BRASSEUR, CNRS / LGGE, FR
Jean-Marc MOLINES, CNRS / LGGE, FR
Guillaume SERAZIN, CNRS / LGGE, FR
The OCCIPUT ANR project, funded over the period 2014-2017, aims at separating and characterizing, through an ensemble of realistic, eddying ocean simulations, the intrinsic and atmospherically-forced components of the global ocean variability at low frequency (1-10 years). This self-sustained intrinsic component is unrealistically weak when mesoscale eddies are not explicitly simulated, especially in the laminar ocean models presently used in IPCC-class climate prediction systems. However, this intrinsic component contributes substantially to the low-frequency variability simulated by “eddy-permitting” ocean models, such as those being currently implemented into the coming generation of operational climate prediction systems. This intrinsic component is poorly known in the global ocean, despite [i] its acknowledged important contribution to the oceanic variability; [ii] its chaotic character (hence questioning our interpretation of low-frequency oceanic variability, which is still largely based on deterministic concepts); [iii] its expected interactions with the atmospherically-forced variability (on which most studies are focused); and [iv] its expected impacts on climate variability in the upcoming generation of coupled models including eddying oceans. One barely knows, in particular, the spatial and temporal structure of the intrinsic variability, its chaotic character, its precise footprint upon the surface and subsurface ocean (and on observations), and its weight on the variance of the main oceanic climate indices. Two research teams (MEOM-LGGE and GLOBC-SUC CERFACS/CNRS) propose to address these questions.
We are currently optimizing a NEMO-based novel modeling system requiring Tier-0 resources, to simultaneously integrate a 50-member ensemble of 57-year (1958-2014) global ocean/sea-ice simulations at 1⁄4° resolution. This experiment, the first of its kind, will allow a shift from deterministic to probabilistic climate-related oceanography. It is mandatory to robustly separate the deterministic (atmospherically-forced) and chaotic (intrinsic) parts of the low-frequency ocean variability, to study their spatiotemporal structures and possible interactions. Our results about the imprints of the intrinsic component on the upper ocean thermal variability will also help anticipate and interpret future coupled climate simulations with eddying oceans, and reassess the actual constraint exerted by the atmosphere on the ocean at interannual-to-decadal timescales. The ensemble experiment and ensemble statistics will be shared with the scientific community to foster collaborative investigations.
CLIMATE SPHINX – (CLIMATE Stochastic Physics HIgh resolutioN eXperiments)
Project Title: CLIMATE SPHINX – (CLIMATE Stochastic Physics HIgh resolutioN eXperiments)
Project Leader: Jost von Hardenberg, ISAC-CNR, IT
Resource Awarded: 20739000 core hours on SuperMUC
Susanna Corti, National Research Council (CNR) / Institute of Atmospheric Sciences and Climate (ISAC), ITALY
Paolo Davini, National Research Council (CNR) / Institute of Atmospheric Sciences and Climate (ISAC-CNR), IT
Antonello Provenzale, National Research Council (CNR) / Institute of Atmospheric Sciences and Climate (ISAC-CNR), IT
Hannah Christensen, University of Oxford / Department of Physics, UK
Timothy N. Palmer, University of Oxford / Department of Physics, UK
Antje Weisheimer, University of Oxford / Department of Physics, UK
Climate prediction is currently one of the most computationally challenging problems in science and yet also one of the most urgent problems for the future of society. It is well known that a typical climate model (with a resolution of ~120-km in the atmosphere and ~100-km in the ocean) is unable to represent many important climate features, in particular the Euro-Atlantic weather regimes. Recent studies have been shown that climate models at much higher resolutions (i~16km) simulate these patterns more realistically. Whilst few would doubt the desirability of being able to integrate climate models at such a high resolution, there are numerous other areas of climate model development which compete for the given computing resources: for example, the need to incorporate additional Earth System complexity. Instead of explicitly resolving small scale processes by increasing the resolution of climate models, a computationally cheaper alternative is to use stochastic parameterization schemes. The main motivation for including stochastic approaches in climate models is related with the observed upscale propagation of errors, whereby errors at very small scales (only resolved in high horizontal resolution models) can grow and ultimately contaminate the accuracy of larger scales in a finite time. A stochastic scheme includes a statistical representation of the small scales, so is able to represent this process. There is mounting evidence that stochastic parameterizations prove beneficial for climate simulations. These results highlight the importance of small-scale processes on large-scale climate variability, and indicate that although simulating variability at small scales is a necessity, it may not be necessary to represent the small-scales accurately, or even explicitly, in order to improve the simulation of large-scale climate. Climate SPHINX aims to investigate the sensitivity of climate simulations to model resolution and stochastic parameterisations, and to determine if very high resolution is truly necessary to facilitate the simulation of the main features of climate variability. The EC-Earth Earth-System model will be used to explore the impact of Stochastic Physics in long centennial climate integrations as a function both of model resolution (from 80km to 16km for the atmosphere and from 1° to 0.25° for the ocean) and in coupled and uncoupled configurations. The experiments will include historical and scenario projection following CMIP5 specifications. By comparing high and low resolutions integrations we will estimate the impact of the increased resolution on climate simulation. By comparing experiments with and without the implementation of stochastic physics we will estimate the impact of stochastic physics. By comparing experiments with stochastic physics with experiments carried out without stochastic physics, but at higher resolutions, we will assess to what extent the stochastic representation of the sub-grid processes can compare with the explicit representation of them. This project will provide a significant progress towards reliable climate predictions, exploring the respective role of numerical resolution and stochastic parameterisations in improving climate simulation quality. Indeed, to our knowledge, this will be the first time that the impact of resolution and stochastic parameterisations is investigated extensively and systematically for climate simulations.
STiMulUs – Lagrangian Space-Time Methods for Multi-Fluid Problems on Unstructured Meshes
Project Title: STiMulUs – Lagrangian Space-Time Methods for Multi-Fluid Problems on Unstructured Meshes
Project Leader: Walter Boscheri, University of Trento, IT
Resource Awarded: 8000000 core hours on SuperMUC
Abstract: This project is inserted in the framework of the STiMulUs project, which has begun in 2011, when the PI (Prof. Dr.-Ing. Dumbser) won an ERC Starting Grant of 60 months duration. STiMulUs main task is the development of new robust, efficient and high order accurate numerical algorithms for the solution of time dependent partial differential equations (PDE) in the context of non-ideal magnetized multi-fluid plasma flows with thermal radiation. It will consider both, high order unstructured Eulerian methods on fixed grids as well as high order unstructured Lagrangian schemes on moving meshes, to reduce numerical diffusion at material interfaces. This project and our request of computational resources will focus on the Lagrangian part of STiMulUs, whose growth started last year when the PI et al. were the first who developed a one-dimensional Lagrangian arbitrary high-order one step WENO finite volume scheme for stiff hyperbolic balance laws. The work is currently in progress and we have already ex-tended the previous work to two space dimensions, using unstructured triangular meshes . A very challenging field of application for non-ideal multi-fluid plasma flows with thermal radiation is nuclear fusion, in particular inertial confinement fusion (ICF).). The science is very interested in this topic because of the shortage of energy on the Earth. In fact the world human population is rapidly growing and as a natural consequence also its need for energy. For these reasons and due to their better instabilities resolution, our new Lagrangian schemes would be suitable to study those phenomena occurring during ICF. In this research project we therefore want to carry out very important basic research on that topic and develop completely new, very high order accurate numerical algorithms in space and time for the solution of time dependent partial differential equations (PDE) with stiff source terms on general unstructured meshes that govern non-ideal magnetized multi-fluid plasma flows with thermal radiation, occurring before the onset of the nuclear fusion process. We rely on high order schemes to resolve very well and with only little numerical diffusion also the fine details of the flow that are crucial for this kind of applications. The highly accurate next-generation mathematical tools emerging from the STiMulUs project may lead to completely new fluid-mechanical key insights in ICF flows that can subsequently be used by physicists and engineers to succeed with the next ICF experiments, thus providing modern civilization with clean energy in the future. The importance of our research topic for our society is underlined by the recent international discussions on the accelerating global climate change, mainly caused by modern civilization and its increasing need for energy.
SLIP – Salvinia-inspired surfaces in action: slip, cavitation, and drag reduction
Project Title: SLIP – Salvinia-inspired surfaces in action: slip, cavitation, and drag reduction
Project Leader: Alberto Giacomello, Sapienza University of Rome, IT
Resource Awarded: 50000000 core hours on Fermi
Matteo Amabili, Sapienza University of Rome / Department of Mechanical and Aerospace Engineering, IT
Carlo Massimo Casciola, Sapienza University of Rome / Department of Mechanical and Aerospace Engineering, IT
Emanuele Lisi, Sapienza University of Rome / Department of Mechanical and Aerospace Engineering, IT
Antonio Tinti, Sapienza University of Rome / Department of Mechanical and Aerospace Engineering, IT
Superhydrophobic coatings show promise for underwater applications: surfaces with drag reducing, anti-fouling, and anti-corrosion properties can have a huge impact in naval and marine engineering. However, superhydrophobicity is fragile because it relies on the trapping of gas pockets inside surface roughness. In a previous PRACE project, we demonstrated that the special geometrical and chemical properties of the water fern Salvinia enhance the durability of entrapped gas pockets over a broad range of pressures. This biological example can inspire a new generation of superhydrophobic surfaces for submerged use, where the resistance to liquid penetration is guaranteed by engineering the geometry and chemistry of surface roughness. In view of this objective, we plan to quantify with advanced molecular dynamics techniques the following points:
1. The stability of gas pockets on different surfaces decorated with three-dimensional nanopatterns inspired to the Salvinia leaves. In particular, we are interested in clarifying whether interconnected gas domains are more stable than independent ones. In these simulations, the nanopatterned surfaces are subject to very low to large hydrostatic pressures and the stability of the gas pockets assessed in terms of free-energy barriers. At the end of this part, we will select the geometries that guarantee durable gas trapping to be tested in the second part.
2. The friction properties of Salvinia-inspired surfaces are characterized during the second part of the project in which the submerged surfaces are subject to shear (Couette flow). The presence of liquid-gas interfaces in this case reduces the wall friction as compared to the simple liquid-solid one. The drag-reducing properties of a given surface are usually characterized in terms of the effective slip length, which describes the virtual position of the flow boundary “inside” the wall.
The final outcome of the present project is the identification of engineering criteria to build durable superhydrophobic coatings for submerged applications. The two main properties we will target are the stability of gas pockets at changing pressure and the reduction of friction under shear flow (large slip length).
TD-CCW Turbulence dynamics in the separation region of channels with lower curved walls
Project Title: TD-CCW Turbulence dynamics in the separation region of channels with lower curved walls
Project Leader: Paolo Gualtieri, Universita’ di Roma “La Sapienza”, IT
Resource Awarded: 50000000 core hours on Fermi
Francesco Battista, Universita’ di Roma “La Sapienza” / Dipartimento di Ingegneria Meccanica e Aerospaziale, IT
Jean-Paul Mollicone, Universita’ di Roma “La Sapienza” / Dipartimento di Ingegneria Meccanica e Aerospaziale, IT
Wall bounded turbulent flows and boundary layer separation has long been of interest to many scientists. In fact, the largest part of the energy spent to move vehicles through fluids is dissipated in the boundary layer close to the walls and/or in the turbulent wake behind the vehicle. Basically the fluid streaming around bodies, due to viscous effects, can not follow the actual body shape and becomes detached. This causes the flow separation, i.e. the fluid closest to the body boundary starts flowing in reverse or different directions, most often giving rise to intense turbulent fluctuation, recirculating regions and wakes.
The aim of the present research project is a deeper understanding of the turbulence dynamics in such highly anisotropic and non homogeneous conditions. Our study is intended to fill the gap between idealized conditions such as plane channel flows and the actual flow geometries where the effects of wall curvature and the presence of bluff bodies immediately generates a substantial separated flow region in the bulk of the flow and, more in general, produces turbulent wakes behind the bluff body.
To this purpose we will employ highly resolved Direct Numerical Simulation (DNS) data of a channel flow in presence of a curved wall (bump). Under a more fundamental point of view, the actual novelty of the present project is represented by the use in anisotropic and non-homogeneous conditions of powerful tool which characterize the scale-by-scale dynamics of the turbulent flow. In fact, we propose here to adopt a generalized form of the Karman-Howarth equation. The Karman-Howarth budget follows directly from the equations of motion and characterizes all the dynamical effects occurring at each scale in the turbulent flow. The budget successfully accounts for non-homogeneous effects, i.e. spatial energy fluxes which arise due to the presence of the bump and allows to characterized at each scale the mechanisms of turbulent kinetic energy production, the energy transfer and the spatial energy fluxes that arise in the different flow regions i.e. at different positions within the recirculating bubble or along the wake behind the separated flow. 123 The analysis is expected to highlight the basic physical mechanisms of turbulence regeneration and sustenance due to the presence of the bump. In fact, such understanding is fundamental to tune innovative boundary layer separation control technique or to design innovative turbulence models, which are able to capture the correct dynamics of the fluctuations in highly non homogeneous flows..
CApTURE – Control of wAll TUrbulence at high REynolds number
Project Title: CApTURE – Control of wAll TUrbulence at high REynolds number
Project Leader: Sergio Pirozzoli, Sapienza, University of Rome, IT
Resource Awarded: 50000000 core hours on Fermi
Maurizio Quadrio, Politecnico di Milano / Dipartimento di Ingegneria Aerospaziale, IT
Matteo Bernardini, Sapienza, University of Rome / Mechanical and Aerospace Engineering, IT
Paolo Orlandi, Sapienza, University of Rome / Mechanical and Aerospace Engineering, IT
Pierre Ricco,University of Sheffield / Department of Mechanical Engineering, UK
This project deals with the direct numerical simulation of turbulent channel flow with control devices based on wall movement. The main scientific goal will be to verify that novel drag reduction strategies based either on the use of traveling waves and rotating wall flush-mounted discs retain their effectiveness when made to perform at Reynolds numbers close to realistic engineering applications. For that purpose a series of targeted high-fidelity simulations will be performed whereby the friction Reynolds number will be varied in the range Reτ =1000-4000 (at the very high-end of the currently reachable envelope), for selected values of the control parameters (i.e. wavelength and speed of the traveling waves, diameter and angular speed of the rotating discs). Comparison with available uncontrolled flow cases computed in the framework of previous PRACE projects will allow to get insight into the physical mechanisms responsible for drag reduction. From a more practical standpoint, if the estimates based on low-Reynolds-number simulations are confirmed, flow control devices bed on wall movement could be qualified for practical drag reduction purposes, with potential substantial benefits in terms of energy saving and reduction of emissions, especially in the aeronautical industry..
SCAROLES – SCAling strategy in a solid ROcket motor using two-phase reactive Large Eddy Simulation
Project Title: SCAROLES – SCAling strategy in a solid ROcket motor using two-phase reactive Large Eddy Simulation
Project Leader: Eleonore Riber, CERFACS, FR
Resource Awarded: 23500000 core hours on Curie Thin Nodes (TN)
Cuenot Bénédicte, CERFACS / Computational Fluid Dynamics, FR
Laura Lacassagne, CERFACS / Computational Fluid Dynamics, FR
Vermorel Olivier, CERFACS / Computational Fluid Dynamics, FR
Franck Nicoud, University Montpellier 2 / I3M, FR
Pressure oscillations are a major issue in solid rocket motor design, as very small pressure oscillations induce strong thrust oscillations, involving vibrations detrimental to carrying load. Designing solid rocket motors that produce as small pressure oscillations as possible then represents an important industrial issue. One solution considered today is to use monolithic solid rocket motors for the two first launcher stages operating with an adapted propellant. However during the second operating phase, the significantly smaller launcher inertia induces very high thrusts and as a consequence increased pressure oscillation constraints. A high fidelity prediction of pressure oscillations is then crucial to design future propeller.
Due to this monolithic geometry, one source of intability becomes more important than in the past and needs to be studied in detail: the corner vortex shedding. This instability appears when vortices are produced by a propellant chamfered edge. Due to the coupling effect between the dispersed reactive phase present in the flow and the shedding, the study of the overall stability of this kind of motor requires the consideration of the two-phase reactive flow.
The huge dimensions of the motor (usually around 10 meters long) and the unfriendly flow environment make experiments and numerical simulations very costly and almost impossible in full-scale configurations. As a consequence, experiments and numerical simulations are usually performed on reduced-scale geometries, typically 1/15th or 1/35th of the real configuration, and results on the reduced-scale geometry are then analysed to predict the behaviour of the real engine. However, scaling-up laws of such complex two-phase turbulent reactive flows is not straightforward.
The first objective of this project is to develop a scaling-up strategy allowing to predict the overall stability of a solid rocket motor engine from the study of a reduced-scale motor. To achieve this aim, two two-phase reactive Large Eddy Simluations (LES) will be performed in one sector (1/8th) of the geometry using two scaling factors: the first one corresponds to the real-scale geometry, the second one to a reduced scale geometry (1/6th).
However, an important limitation of single-sector computations is that the velocity field is necessarily aligned with the symmetry axis, preventing large scale turbulent fluctuations to develop throughout the centerline. This may have strong effect on the flow organization and eventually on its stability. Thus, in order to investigate potential 3D effects that may be neglected when considering only a sector of the geometry, a two-phase reactive LES will be performed in the full annular geometry at 1/6th reduced scale. This annular LES will be compared to the simulation on the reduced scale geometry. This study on the impact of considering only one sector of the geometry instead of the full annular geometry represents the second objective of this project.
TopWing – Topology optimization of aircraft wing
Project Title: TopWing – Topology optimization of aircraft wing
Project Leader: Ole Sigmund, Technical University of Denmark, DK
Resource Awarded: 12000000 core hours on Curie Thin Nodes (TN)
Niels Aage, Technical University of Denmark / Department of Mechanical Engineering, DK
Erik Andreassen, Technical University of Denmark / Department of Mechanical Engineering, DK
Boyan Lazarov, Technical University of Denmark / Department of Mechanical Engineering, DK
Topology optimization is a widely used computational tool for saving weight and optimizing structural response of cars, airplanes and other mechanical devices. Based on pixel (2d) or voxel (3d) discretizations of the design domain, the method has huge design freedom with no restrictions on allowable geometries. Currently, fairly coarse design discretizations (up to 10 million elements) limit applications to part design or full system design with limited design freedom. The TopOpt Group at the Technical University of Denmark (DTU) has recently developed a PETSc-based research code that has been tested on up to 120 million elements. Access to PRACE resources allows the TopOpt Group to run simulations with +1 billion elements, paving the way for full-scale airplane wing design with hitherto unprecedented resolution and huge potential for improved design and weight savings for the aerospace industry..
Surfing for Thermals in Turbulent Flows
Project Title: Surfing for Thermals in Turbulent Flows
Project Leader: Roberto Verzicco, Physics of Fluids, University of Twente, NL
Resource Awarded: 68550000 core hours on Fermi
Vamsi Spandan Arza, University of Twente / Department of Science & Engineering, NL
Rodolfo Ostilla Monico, University of Twente / Department of Science & Engineering, NL
Erwin van der Poel, University of Twente / Department of Science & Engineering, NL
Yantao Yang, University of Twente / Department of Science & Engineering, NL
Turbulent flow is abundant in nature and technology. In contrast to a decade-old paradigm, even highly turbulent flow is strongly influenced by the boundaries. The strength of turbulence is characterized by the Reynolds number Re, which gives the ratio between inertial and viscous forces. Reynolds numbers in typical process technology applications are of the order of 10^8. In the atmosphere Re=10^10 is achieved, and in an open ocean one has Re=10^11 and in stars even much higher values.We focus our study on one of the paradigmatic systems used in fluid dynamics, namely Rayleigh Benard convection (flow in a closed box heated from below and cooled from above). Such a system is mathematically well defined by the Navier-Stokes equations with appropriate boundary conditions and exact global balance relations between the respective driving and the dissipation can be derived. They are also experimentally accessible with high precision, thanks to simple geometries and high symmetries.
Most RB simulations and experiments focus on very tall and thin cells, where the aspect ratio, i.e the ratio of the cell width to cell height is lower than one. This reduces the cost of the simulation or experiment while retaining some of the essential physics. However, many natural instances of convection have very large, almost infinite aspect ratios. Large structures much wider than the distance between the two plates develop on top of the small turbulent fluctuations. These structures are analogous to the thermals found in the Earth’s atmosphere. These structures are of course absent in low aspect ratio setups because of lateral confinement. Given their importance, within the scope of this project we want to go beyond the geometric confinement towards large aspect ratios in order to study these structures. A unique point of this project is that this database will have both Eulerian data in the form of complete snapshots of the flow fields and also the Lagrangian data, in the form of detailed time evolution of tracer particles injected numerically into the flow. Very little RB Lagrangian data exists in the literature, so this will considerably extend the available data, while unveiling new physics.
We expect these questions to have direct implication on our understanding of transport properties in buoyant flows with applications to engineering, marine and atmospheric science, opening also the way to the design of smart particles with an optimized strategy to surf thermal plumes.
Modeling of Multi-Scale Interfacial Flows
Project Name: Modeling of Multi-Scale Interfacial Flows
Project Leader: Stephane Zaleski, UPMC Université Paris 6, FR
Resource Awarded: 19000000 core hours on Fermi, 10000000 core hours on SuperMUC
Ruben Scardovelli, Universita Di Bologna / DIENCA, IT
Tomas Arrufat Jackson, UPMC Université Paris 6 / Institut Jean Le Rond d’Alembert, FR
Daniel Fuster Salamero, UPMC Université Paris 6 / Institut Jean Le Rond d’Alembert, FR
Yue Ling, UPMC Université Paris 6 / Institut Jean Le Rond d’Alembert, FR
We propose to investigate several two-phase flow problems with a wide range of scales with Direct Numerical Simulation (DNS), such as atomisation and drop impact in three dimensions (3D). Both problems are extremely challenging.
Atomisation is used to quickly turn a liquid mass into a large number of droplets. It controls the quality of combustion in diesel and cryogenic rocket engines. Often, the shear at the liquid-gas interface causes the 2D Kelvin-Helmholtz instability (KHI) which leads by subsequent 3D instabilities to ligaments and droplets. The wavelength of the KHI determines the final droplet size, an essential parameter for modelling combustion. The team has developed a linear theory of atomisation, which was compared to 2D simulations with the Volume-of-Fluid (VOF) method. Excellent agreement on the longitudinal disturbances was achieved [1,2]. Despite the progress made by those and other 3D simulations, two issues remain unresolved: the simulations do not provide insight into the transverse instabilities leading to 3D flows and the size distribution of droplets does not converge when the mesh refines.
The formation of a liquid “crown” or “corolla” in drop impact is another fascinating phenomenon. As in atomisation, the first step is to understand the mechanisms in axisymmetric geometry. The theory developed by the team  predicts the splash transition. This transition marks the boundary between the “no splash” regime and the “corolla” regime, the axisymetric flow related to the 3D crown. Several mechanisms have been proposed to explain the 3D drop formation, such Ricthmyer-Meshkov and Rayleigh-Plateau-Savart instabilities. However, 3D simulations with the same level of detail are missing in the literature and not yet available to investigate how these mechanisms operate.
The project will lead to very important applications in engineering and also open the way to other two-phase flow simulations such as two-phase flow in porous media, breaking waves, Taylor bubbles or blood flow in arteries.
Fundamental Conistituents of Matter (5)
CAPITOL – Computing Accelerated Particles, Intense Terahertz and Optical radiation by Lasers
Project Title: CAPITOL – Computing Accelerated Particles, Intense Terahertz and Optical radiation by Lasers
Project Leader:Luc Bergé, CEA, FR
Resource Awarded: 24115200 core hours on Curie Thin Nodes (TN)
Antoine Compant La Fontaine, CEA / Theoretical and Applied Physics Dept, FR
Xavier Davoine, CEA / Theoretical and Applied Physics Dept, FR
Arnaud Debayle, CEA / Theoretical and Applied Physics Dept, FR
Julien Ferri, CEA / Theoretical and Applied Physics Dept, FR
Pedro Gonzalez de Alaiza Martinez, CEA / Theoretical and Applied Physics Dept, FR
Laurent Gremillet, CEA / Theoretical and Applied Physics Dept, FR
Mathieu Lobet, CEA / Theoretical and Applied Physics Dept, FR
Gilles Riazuelo, CEA / Theoretical and Applied Physics Dept, FR
Agustin Lifschitz, ENSTA-ParisTech-CNRS-Ecole Polytechnique / Laboratoire d’Optique Appliquée, FR
Ihar Babushkin, Humboldt University Berlin / Department of Mathematics, DE
Stefan Skupin, Université Bordeaux 1 / CELIA, FR
Illia Thiele, Université Bordeaux 1 / CELIA, FR
Modern laser sources deliver optical pulses reaching unprecedented intensity levels by concentrating near-kiloJoule energy in ultrashort wave-packets of light. Such extreme optical objects question the fundamentals of wave propagation in all materials in general, and in plasmas in particular. They open broad application areas, which embrace the main recommendations on Key Enabling Technologies initiative of the European Photonics Strategic Roadmap 2014-2020, for which monitoring high-power, extreme light and proposing new far-infrared emitters are a priority in the European Research Area and overseas. 123 Fully involved in this framework, our project aims at modelling numerically a large panel of photon and particle sources created from the interaction between an intense ultrashort pulse and an ionized medium. Attention will be paid to biological applications and medical uses of these sources, which will rank from weakly energetic (non-invasive), terahertz radiation for tumor detection to highly energetic radiations and accelerated ions. Those can indeed supply low-cost, portable sources for deep-body imagery and cancer surgery. Our team is already involved in the SAPHIR consortium, an academic-industrial partnership devoted to laser-driven protontherapy, in LASERLAB/CHARPAC, a European FP7 project on laser-plasma accelerators, and in an ERC-granted X-5 project dedicated to novel laser-plasma radiation sources. Through numerous national Tier-1 projects, our team has performed predictive numerics and reproduced many experimental results on particle acceleration, X-ray generation and production of broadband THz sources. 123 The CAPITOL project is devoted to the numerical study of laser-driven sources of photons and charged particles over a wide range of laser and plasma parameters. Through 2D and 3D simulations, we wish to determine generic scaling laws characterizing the above sources and potential optimization strategies. Three tasks are planned.
A first task will focus on the production of X-ray or gamma-ray sources by ultra-intense lasers. Several radiation mechanisms will be explored, mainly in the frame of the wakefield accelerator. Prospective scenarios based on solid targets irradiated at extreme laser intensities, for which radiative quantum electrodynamic processes must be described along with collective mechanisms, will also be considered.
A second task will concern laser-induced generation of intense ion beams. We will address the potential of thin foil targets, thus extending, within the SAPHIR project, the work undertaken as part of our first project PRACE SOULAC (2011) on protontherapy. A more innovative interaction setup employing gas targets will also be investigated. These first two actions will make extensive use of the particle-in-cell (PIC) kinetic code CALDER in order to quantify the relativistic laser-matter interaction and the related generation of secondary radiation and particles.
A third task is the numerical control of broad optical spectra induced by multi-colored femtosecond laser pulses interacting with initially neutral gases, in tightly or loosely focused propagation geometries. The goal will be to optimize the terahertz part of the spectra for tumor spectroscopy applications. Direct comparisons of results provided by CALDER and those obtained from the unidirectional propagation code UPPE3D will cover a broad range of laser-plasma configurations and will open new perspectives in terahertz emission produced over remote distances.
From Plasma Microphysics to Global Dynamics in Collisionless Accretion Discs: Magnetorotational Turbulence, Plasma Instabilities, and Collisionless Reconnection
Project Title: From Plasma Microphysics to Global Dynamics in Collisionless Accretion Discs: Magnetorotational Turbulence, Plasma Instabilities, and Collisionless Reconnection
Project Leader: Matthew Kunz, Princeton University, USA
Resource Awarded: 26000000 core hours on Curie Thin Nodes (TN)
Steve Cowley, Culham Science Centre / Culham Centre for Fusion Energy, UK
Nuno Loureiro, Instituto Superior Técnico / Institute for Plasmas and Nuclear Fusion, PT
Benoît Cerutti, Princeton University / Department of Astrophysical Sciences, USA
James Stone, Princeton University / Department of Astrophysical Sciences, USA
Eliot Quataert, University of California, Berkeley / Department of Astronomy, USA
Alexander Schekochihin, University of Oxford / Rudolf Peierls Centre for Theoretical Physics, UK
The magnetorotational instability (MRI) is the most promising means of transporting angular momentum and thereby driving the observed mass accretion rates in a wide variety of accretion discs. However, some of these discs — most notably the one feeding the black hole at our Galactic Center, Sgr A* — are too hot and diffuse to rigorously be treated as magnetohydrodynamic fluids. Instead, a kinetic approach is necessary, one which is able to simultaneously follow both the detailed plasma microphysics at Larmor scales and the global dynamics at the fluid-like macroscales. We plan to use the massively parallel hybrid-kinetic particle-in-cell code, Pegasus, to perform the first kinetic 6D numerical simulations of the MRI in a collisionless plasma. This work will be supplemented by dedicated kinetic simulations of (1) the firehose and mirror instabilities — velocity-space instabilities that act as parasites on the MRI and thereby regulate the MRI-driven temperature anisotropy, affect the efficiency of angular-momentum transport, and set the rate of particle scattering and heating; and (2) collisionless reconnection in non-Maxwellian plasmas. The latter simulations — which make use of the massively parallel, fully kinetic, electromagnetic particle-in-cell code Zeltron — will address how the reconnection rate and the amount of ion/electron heating depend upon temperature anisotropy and, thus, will aid our understanding of how the MRI channel modes break down into fully developed turbulence. While primarily focused on hot accretion discs, the results of these calculations will impact our understanding of other weakly collisional plasmas, such as the solar wind and the intracluster medium of galaxy clusters.
QUASINO – QUAntum SImulation of ultimate NanO-devices
Project Title: QUASINO – QUAntum SImulation of ultimate NanO-devices
Project Leader: Yann-Michel Niquet, CEA-INAC, FR
Resource Awarded: 6121000 core hours on Curie Fat Nodes (FN), 96000 core hours on Curie Hybrid Nodes
Ivan Duchemin, CEA / INAC, FR
Jing Li, CEA / INAC, FR
François Triozon, CEA / LETI-MINATEC, FR
Christophe Delerue, IEMN / ISEN, FR
Gabriel Mugny, STMicroelectronics / STD, FR
Denis Rideau, STMicroelectronics / STD, FR
The characteristic dimensions of the transistors on a processor chip have steadily decreased over the last fifty years, allowing for ever more performances and functionalities. The International Technology Roadmap for Semiconductors (ITRS), which sets targets for the microelectronics community, is now discussing options beyond the “10 nm node”. The physics of such devices is, however, very complex and goes beyond semi-classical understanding. Modeling and simulation are therefore expected to play an increasing role in the exploration of original and innovative designs. This calls for the development of advanced simulators able to account for quantum effects, such as tunneling and confinement, prevailing at the nanometer scale. The Non-Equilibrium Green’s Functions (NEGF) method is one of the most versatile approaches to quantum transport. It describes all important scattering mechanisms (scattering of electrons by impurities, lattice vibrations, etc…) in an unified framework. We have recently developed a new NEGF solver based on effective mass and atomistic tight-binding models of the electronic structure of the materials, carefully optimized for high-performance computing infrastructures and hybrid CPU/GPU machines. This code is able to simulate devices with realistic geometries and sizes in the 2-30 nm range, representative of the next generation of transistors being developed at present. The objective of this project is to model the latest “Fully-Depeleted Silicon-on-Insulator” technologies developed at STMicroelectronics in Europe, and the next “Trigate” transistors being prepared at CEA/LETI. We will, in particular, focus on the physics and performances at high electric field where the charge carriers are driven far out of equilibrium – a complex but extremely important operation regime. We will also look at the perspectives for III-V integration in the next technology nodes, as well as to the effects of Coulomb interactions on “ultra-scaled” devices with characteristic dimensions below 5 nm. Our aim is to introduce quantum simulation in the design of the next generations of transistors, in order to reduce the number of costly development batches..
QED corrections to meson decay rates in Lattice QCD
Project Title: QED corrections to meson decay rates in Lattice QCD
Project Leader: Silvano Simula, INFN, IT
Resource Awarded: 18000000 core hours on Fermi
Nazario Tantalo, CERN / Theoretical Division, CH
Nuria Carrasco Vela, INFN / Sezione di Roma Tre, IT
Guido Martinelli, SISSA / Direttore, IT
Daniele Belardinelli, University of Rome “Tor Vergata” / Physics Department, IT
Petros Dimopoulos, University of Rome “Tor Vergata” / Physics Department, IT
Roberto Frezzotti, University of Rome “Tor Vergata” / Physics Department, IT
Giancarlo Rossi, University of Rome “Tor Vergata” / Physics Department, IT
Vittorio Lubicz, University of Rome III / Physics Department, IT
Eleonora Picca, University of Rome III / Physics Department, IT
Cecilia Tarantino, University of Rome III / Physics Department, IT
Francesco Sanfilippo, University of Southampton / School of Physics and Astronomy, UK
Precision flavour physics is particularly powerful for exploring the limits of the Standard Model (SM) and in searching for inconsistencies which would signal new physics. An important component is the over-determination of the elements of the Cabibbo-Kobayashi-Maskawa (CKM) matrix from a wide range of weak processes. The precision in extracting CKM matrix elements is generally limited by our ability to quantify hadronic effects and the main goal of large-scale simulations using the lattice formulation of QCD is the ab-initio evaluation of the non-perturbative QCD effects in physical processes. The recent, very impressive, improvement in lattice computations has led to a precision approaching the percent level for a number of quantities and therefore in order to make further progress electromagnetic effects (and other isospin-breaking contributions) have to be considered.
Electromagnetic effects in the spectrum and in the determination of quark masses have begun to be included using a variety of approaches. In calculation of the spectrum there is a very significant simplification in that there are no infrared divergences. For many relevant physical quantities, however, infrared divergences in intermediate steps of the calculation require new strategies. This is the case, for example, of the leptonic Pil2 and Kl2 and of the semileptonic Kl3 decay rates. We stress indeed that for these quantities it is not sufficient simply to add the electromagnetic interaction to the quark action because amplitudes with different numbers of real photons must be evaluated separately, before being combined in the inclusive rate for a given process.
A new method to include electromagnetic effects in processes for which infrared divergences are present in the intermediate steps but which cancel between diagrams containing different numbers of real and virtual photons, has been developed recently by our group. The aim of this project is to apply the new method to the first complete calculations of Pil2, Kl2 and Kl3 rates within full QCD+QED.
GWEUPT – Gravitational waves from early universe phase transitions
Project Title: GWEUPT – Gravitational waves from early universe phase transitions
Project Leader: David Weir, University of Stavanger, NO
Resource Awarded: 17600000 core hours on Hornet
Abstract: Understanding the development of the very early stages of the Big Bang is tightly linked to the understanding of the fundamental nature of matter and interactions. With the Large Hadron Collider delineating the nature of the Higgs, we have the opportunity to make predictions using the new knowledge, enabling a sharper picture of the very early universe. In this project we will use state-of-the art high performance computing to calculate the gravitational wave signal from a phase transition in the Higgs field in at around a tenth of a nanosecond after the Big Band, including the hydrodynamics of the hot plasma. This radiation may be observable in planned space-based gravitational wave detectors such as eLISA.
Mathematics and Computer Sciences (1)
HP-Feel++ – High Performance Finite Element Embedded Library in C++
Project Title: HP-Feel++ – High Performance Finite Element Embedded Library in C++
Project Leader: Prud’homme Christophe, Universite de Strasbourg, FR
Resource Awarded: 6000000 core hours on Curie Thin Nodes (TN)
Silvia Bertoluzza, CNR / Istituto di Matematica Applicata e Tecnologie Informatiche “E. Magenes”, IT
Micol Pennacchio, CNR / Istituto di Matematica Applicata e Tecnologie Informatiche “E. Magenes”, IT
Christophe Trophime, Laboratoire National des Champs Magnetiques Intenses / Magnet Development, FR
Alexandre Ancel, Universite de Strasbourg / Institut de Recherche Mathematique Avancee, FR
Cécile Daversin, Universite de Strasbourg / Institut de Recherche Mathematique Avancee, FR
Guillaume Dollé, Universite de Strasbourg / Institut de Recherche Mathematique Avancee, FR
Romain Hild, Universite de Strasbourg / Institut de Recherche Mathematique Avancee, FR
Vincent HUBER, Universite de Strasbourg / Institut de Recherche Mathematique Avancee, FR
Marcela Szopos, Universite de Strasbourg / Institut de Recherche Mathematique Avancee, FR
Ranine TARABAY, Universite de Strasbourg / Institut de Recherche Mathematique Avancee, FR
Jean-Baptiste Wahl, Universite de Strasbourg / Institut de Recherche Mathematique Avancee, FR
Mourad ISMAIL, Universite Joseph Fourier Grenoble 1 / Laboratoire Interdisciplinaire de Physique, FR
Jérémy Veysset, Universite Joseph Fourier Grenoble 1 / Laboratoire Interdisciplinaire de Physique, FR
Vincent Chabannes, Universite Joseph Fourier Grenoble 1 / Laboratoire Jean Kuntzmann, FR
Abdoulaye SAMAKE, Universite Joseph Fourier Grenoble 1 / Laboratoire Jean Kuntzmann, FR
Pierre Jolivet, Universite Pierre et Marie Curie / Laboratoire Jacques-Louis Lions, FR
The HP-Feel++ project aims are twofold: (i) developing scalable solution strategies and (ii) enabling them in research applications. Two applications domains have been selected : high field magnets (HiFiMagnet) and diffuse optical tomography (DOT).
HP-FEEL++ and the associated applications require the use of a wide range of numerical methods for partial differential equation (PDE). These numerical methods are implemented in Feel++ (http://www.feelpp.org) and parallelized using MPI.
The size of our simulations range from several million of degrees of freedom to billions of degrees of freedom and therefore, they require access of TIER-0 computing resources. Moreover, the numerical methods used in both applications require to solve a linear system, and get an optimal scalability in this part is an active research topic.
In order to obtain a large scale scalability, we have developed theoretically and numerically non-conforming numerical methods in the context of domain decomposition methods, i.e. the mortar method (MM). Recent advances have been made to also equip Feel++ with powerful conforming numerical methods, such as the Finite Element Tearing and Interconnecting method (FETI) and Balanced Neumann-Neumann.The design of efficient preconditioners for these systems is a fundamental task, in particular in three dimensions, where the number of unknowns increases very rapidly. The objective is to verify the theoretical results obtained for both methodologies on standard large scale architectures ( from hundreds to tens of thousands of core) and to apply our novel scalable solution strategies on real applications such as the design of high field magnets and optical tomographs.
For high field magnets, the obtained results will be readily used to finalize the design of the Hybrid magnet. Regarding diffuse optical tomography our objective within the next three years is to develop a complete numerical model associated to the tomograph allowing for fast diagnostics (within minutes) of the presence of tumors.
Universe Sciences (7)
GRSimStar – General Relativistic Simulations of binary neutron Star mergers
Project Title: GRSimStar – General Relativistic Simulations of binary neutron Star mergers
Project Leader: Bruno Giacomazzo, University of Trento, IT
Resource Awarded: 15728640 core hours on SuperMUC
Wolfgang Kastaun, Max Planck Institute for Gravitational Physics / Astrophysical Relativity, DE
Daniel Siegel, Max Planck Institute for Gravitational Physics / Astrophysical Relativity, DE
Luca Baiotti, Osaka University / Institute of Laser Engineering, JP
Rosalba Perna, SUNY Stony Brook / Physics and Astronomy, USA
Riccardo Ciolfi, University of Trento / Department of Physics, IT
Takumu Kawamura, University of Trento / Department of Physics, IT
This proposal requests computational resources necessary in order to continue our successful investigations of binary neutron star mergers. By using our fully general-relativistic magnetohydrodynamic code WhiskyMHD, we will simulate the merger of magnetized neutron stars in order to study the gravitational wave (GW) and electromagnetic (EM) signals that these objects may emit. In particular, GW signals from binary neutron stars are expected to be observed in the next few years by the advanced LIGO and Virgo detectors. Binary neutron stars are also thought to be behind the engine of short gamma-ray bursts (SGRBs) and our simulations will investigate how magnetic fields and the neutron star equation of state affect the dynamics of such systems.
PHOENIX/3D – 3D NLTE radiation transport with PHOENIX/3D
Project Title: PHOENIX/3D – 3D NLTE radiation transport with PHOENIX/3D
Project Leader: Peter Hauschildt, Universitaet Hamburg, DE
Resource Awarded: 28800000 core hours on HornetDetails
Edward Baron, University of Oklahoma / Physics and Astronomy, USA
The PHOENIX code is a general-purpose model atmosphere simulation package. It is designed to model the structures and spectra of a wide range of astrophysical ob-jects, from extrasolar planets (both terrestrial and gas giants) to brown dwarfs and all classes of stars, extending to novae and supernovae. The main results from the cal-culations are synthetic images and spectra (and derived quantities, such as colors), these can be directly compared to observed spectra and, in 3D simulations, images. By adjusting the simulation parameters and comparing the results to the observed data, the physical parameters of the stars or planets can be determined.
Here, we propose to apply PHOENIX/3D to 3 different science cases: (i) transmission & reflection spectra of extrasolar giant planets, (ii) NLTE line formation of CO molecules in the spots of cool stars and (iii) NLTE radiation transfer simulations of solar and M dwarf chromosphere & photosperes.
FRIG – From intermediate galactic scales to cores
Project Title: FRIG – From intermediate galactic scales to cores
Project Leader: Patrick Hennebelle, CEA-Saclay, FR
Resource Awarded: 15000000 core hours on Curie Thin Nodes (TN)Details
Philippe André, CEA / DSM/SAp, FR
Alexander Mechshikov, CEA / DSM/SAp, FR
Damien Chapon, CEA/Saclay / DSM, FR
Samuel Geen, CEA/Saclay / DSM/SAp, FR
Olivier Iffrig, CEA/Saclay / DSM/SAp, FR
Evangelia Ntormousi, CEA/Saclay / DSM/SAp, FR
Romain Teyssier, Zurich university / Institute of computational science, CH
Understanding star formation is one of the major challenges of modern astrophysics. Determinant aspects of star formation haven’t been yet fully unveiled because star formation is governed by the interstellar cycle, which entails: a huge range of spatial and temporal scales and a broad diversity of coupled physical phenomena such as turbulence, magnetic field, gravity and radiation. In particular, if it is well established that stars form in molecular clouds through the collapse of molecular dense cores, the origin of these cores remain controversial as well as the shape and the universality of their mass distribution. This is a major question since the core mass function may be at the origin of the initial mass function of stars, which plays a fundamental role in many areas of astrophysics.
Since the dense cores form in molecular clouds and are strongly affected by their characteristics, it is necessary to understand the core formation and to get reliable statistics, not only to describe the scale at which dense cores form but also the scale at which molecular clouds themselves form. This makes it a very difficult problem to model and investigate. It is thus necessary to find adequate strategies, which allow to capture the essence of the phenomena. Here we propose a strategy, which is based on the existence of 2 typical scales seemingly fundamental in this problem.
First of all the size of molecular clouds is on the order of 50-100 pc, which means that to address the question of their formation, it is necessary to simulate a spatial scale, which is typically 10 times larger and of the order of 1 kpc. Second of all, the size of the structures within molecular clouds is typically on the order of 0.1 pc. This implies that a simulation, which captures self-consistently the formation of molecular clouds and their internal structure, must cover at least 5 decades in scales (from 1 kpc to ~0.01 pc). To tackle such broad scale dynamics, we propose the following strategy. First, we will run a uniform grid simulation at a scale of 1 kpc. This simulation will include supernovae explosions, which drive the interstellar turbulence and stratification by gravity. In these simulations, dense structures develop and collapse under the influence of large scale turbulence and gravity. One region of the computational box will then be selected (about 1/10 of the computational domain). This new region will then be refined using a uniform grid strategy to warrant that the properties of turbulence, and therefore the statistics are well handled and not too affected by the refinement strategy. Finally, we will add a few more AMR levels to follow the dense structure specifically.
Using the simulation results, we will produce synthetic observations from which we will identify the dense structures. The statistics of these structures such as their aspects ratio distributions and the mass function will be compared directly to the high resolution data obtained recently with the Herschel satellite.
Magnetic Field Generation in Rotating, Stably Stratified Turbulence
Project Title: Magnetic Field Generation in Rotating, Stably Stratified Turbulence
Project Leader: Joanne Mason, University of Exete, UK
Resource Awarded: 23000000 core hours on FermiDetails
Steven Tobias, University of Leeds / Applied Mathematics, UK
The eleven-year solar cycle is one of the most intriguing unsolved problems in astrophysics. Sunspots (which are sites of intense magnetic field) appear to migrate from mid-latitudes to the equator over a period of eleven years, at which point the magnetic field reverses direction and the whole process starts afresh. This fascinating behaviour of the Sun’s large-scale magnetic field is due to the operation of a hydromagnetic dynamo, in which inductive motions within the electrically conducting fluid act to generate and sustain the magnetic field against the action of Ohmic dissipation.
The location of the large-scale solar dynamo is believed to be the solar tachocline — a thin layer of strong differential rotation (mean velocity shear) at the base of the solar convection zone. While the upper part of the tachocline is strongly coupled to the convection zone above, the lower part of the tachocline is believed to be very stably stratified and the properties of the magnetic field there are poorly understood. Indeed, there are many fundamental questions regarding tachocline dynamics, not least an explanation for why the region exists at all. At the heart of the matter resides a detailed understanding of the interaction of magnetic fields with turbulent, rotating, stratified flows.
Herein we propose a series of high-resolution direct numerical simulations of forced, incompressible, rotating, stably stratified, magnetohydrodynamic (MHD) turbulence. Our aim is to explore the dynamics of rotating, stratified turbulence in the statistically steady regime. We will allow the magnetic field to be self-consistently generated by the flow and one of the main goals of the study is to understand how the properties of dynamo action change as the rotation rate and the degree of stratification increase. Strongly stratified hydrodynamic flows are known to be approximately two-dimensional, and two-dimensional systems are known to form large-scales flows via an inverse cascade. However, it is known that rotation can hinder this process and the effects of magnetic fields are key. Whether turbulent, rotating, stably stratified MHD flows can self-consistently generate large-scale magnetic fields is a topic of great importance for understanding the dynamics of stellar and planetary interiors. It is precisely the topic that we shall investigate through a series of high-resolution numerical simulations.
Global kinetic modelling of space weather – with extreme scalability (VLASIATOR)
Project Title: Global kinetic modelling of space weather – with extreme scalability (VLASIATOR)
Project Leader: Minna Palmroth, Finnish Meteorological Institute, FI
Resource Awarded: 24100000 core hours on HornetDetails
Sanni Hoilijoki, Finnish Meteorological Institute / Earth Observation, FI
Yann Kempf, Finnish Meteorological Institute / Earth Observation, FI
Arto Sandroos, Finnish Meteorological Institute / Earth Observation, FI
Sebastian von Alfthan, Finnish Meteorological Institute / Earth Observation, FI
Urs Ganse, University of Helsinki / Department of Physics, FI
Space weather is a term used to describe the variable environmental effects within near-Earth space, caused by the Sun emitting solar wind, a stream of charged particles carrying the solar electromagnetic field. Space weather can be caused by solar high-energy particles or by dynamic variations of the solar wind that can cause extended periods of major disturbances on ground and space, affecting technological systems (e.g., telecommunication and weather spacecraft at geostationary orbit, and ground-based power grids). This proposal concerns the scientific basis of the dynamic space weather.
The Finnish Meteorological Institute developed a 6-dimensional Vlasov theory-based simulation called Vlasiator, in a Starting Grant project by the European Research Council awarded in 2007. In Vlasiator, ions are distribution functions, while electrons are magnetohydrodynamic fluid, enabling a self-consistent global plasma simulation that can describe multi-temperature plasmas to resolve non-MHD processes that currently cannot be self-consistently described by the existing global space weather simulations. The novelty is that by modelling ions as distribution functions the outcome will be numerically noiseless. Due to the multi-dimensional approach in ion scales, Vlasiator’s computational challenges are immense. We use advanced high performance computing techniques to allow massively parallel computations. The first runs enabled by a previous PRACE Tier-0 access have showed that Vlasiator produces the physics of space weather with unprecedented quality, revealing a rich playground of collective phenomena that cannot be interpreted by local in situ observations only.
This proposal concerns the scientific basis of space weather, in which three phenomena are above others in importance in explaining plasma behaviour: 1) magnetic reconnection enabling energy and mass transfer between different magnetic domains, 2) shocks forming due to supersonic relative flow speeds between plasma populations, and 3) particle acceleration associated with both reconnection and shocks. Synergetic advances are foreseen in simultaneous studies of the three phenomena, which are also important in laboratory plasmas, fusion, and in astrophysical domains. Combining Vlasiator with newest space-craft data, we aim at breakthroughs in explaining the local physics globally and self-consistently in
1) Magnetopause and tail reconnection for a variety of interplanetary magnetic field (IMF) orientations to quantify reconnection rates and instabilities.
2) Ion-scale phenomena in the foreshock-magnetosheath-magnetopause system and in shock – shock interactions to evaluate the interplay between the shock evolution and its consequences to magnetosheath processes in the global shock-sheath system.
3) Particle acceleration due to electromagnetic fields at shocks and in reconnection regions. The runs from science questions 1) and 2) are used with added (test) particles to evaluate the particle paths and energisation in the global system.
These runs present the world’s first self-consistent semi-global attempts to explain the near-Earth ion-scale plasma physics in a holistic manner including realistic boundary conditions. This project is extremely timely now that the supercomputing facilities are large enough. The science themes are crucially important in Cluster, van Allen Belt Probes and THEMIS missions to be supplemented by MMS data launched during the project.
Stars’ turbulent birth in galactic collisions
Project Title: Stars’ turbulent birth in galactic collisions
Project Leader: Florent Renaud, University of Surrey, UK
Resource Awarded: 21800000 core hours on Curie Thin Nodes (TN)Details
Frederic Bournaud, CEA-Saclay / Service d’Astrophysique, FR
Damien Chapon, CEA-Saclay / Service d’Astrophysique, FR
Pierre-Alain Duc, CEA-Saclay / Service d’Astrophysique, FR
Jeremy Fensch, CEA-Saclay / Service d’Astrophysique, FR
Patrick Hennebelle, CEA-Saclay / Service d’Astrophysique, FR
Eric Emsellem, ESO /Science , DE
Bruce Elmegreen, IBM / Research Center, USA
Chanda Jog, Indian Institute of Science / Department of Physics, IN
Francoise Combes, Observatoire de Paris / LERMA, FR
Paola DiMatteo, Observatoire de Paris / LERMA, FR
Philippe Amram, Universite de Marseille / LAM, FR
Curt Struck, University of Iowa / Department of Physics and Astronomy, USA
Oscar Agertz, University of Surrey / Physics, UK
Valentin Perret, University of Zurich / Astrophysics, CH
Romain Teyssier, University of Zurich / Astrophysics, CH
Galaxy mergers host fireworks of star formation, called starburst, during which the star formation rate can be multiplied by factors up to 500. The physical reason for such activity is not yet fully understood because the main trigger of the burst (namely the galactic collision) occurs at scales much larger than that of star formation. It is therefore numerically difficult to capture both processes in the same simulation. Our project aims at exploring how different collisions (mass ratio, impact parameter, amount of gas in galaxies etc) and different microphysics parameters (sub-grid recipes implemented in the code) couple to change the physical conditions leading to starburst. We will run several simulations at high resolution to test these effects and show how they affect the net product, in term of turbulence, star formation rate, formation of star clusters etc. Hopefully, our simulations will help us to propose new theories of star formation, in these extreme environments..
POGO – The Physical Origins of Galactic Outflows
Project Title: POGO – The Physical Origins of Galactic Outflows
Project Leader: Orianne ROOS, CEA-Saclay, FR
Resource Awarded: 11000000 core hours on Curie Fat Nodes (FN)Details
Frédéric Bournaud, CEA-Saclay / Service of Astrophysics, FR
Jared Gabor, CEA-Saclay / Service of Astrophysics, FR
Yohan Dubois, Institut d’Astrophysique de Paris / Grandes Structures et Univers Profond, FR
Joakim Rosdahl, Leiden University / Leiden Observatory, NL
Florent Renaud, University of Surrey / Department of Physics, UK
Valentin Perret, University of Zürich / Institute for Computational Science, CH
Romain Teyssier, University of Zürich / Institute for Computational Science, CH
Abstract: In the Universe, less than 20 % of all baryons (e.g. usual matter: atoms and ions, in the form of gas, stars and dust) belong to galaxies. This “missing baryons problem” is in apparent contradiction with the universal laws of gravitation, according to which all baryons should have fallen in today’s galaxies. To explain this, efficient expelling mechanisms, such as gas outflows, are needed to remove gas.
Gas outflows are detected in both ionized and molecular gas, and have been observed to accelerate inside the halo, away from the disk. In simulations, though, properties of modeled winds are generally set arbitrarily to satisfy some observational constraints (e.g. mass expulsion rate and outflow velocity set by hand), rather than being modeled from basic physical principles. Such models are useful to study the consequences of outflows on the intergalactic medium, but are not able to constrain the physical parameters of the winds, nor the mechanisms creating them. Furthermore, even though they could in principle expel enough gas to prevent their host galaxy from forming stars, neither observed nor realistic simulated outflows reach the high level of mass-loading needed to explain the missing baryons problem, at the moment.
With the POGO project, we will perform high-resolution simulations of star-forming disk galaxies including stellar and AGN feedback, and study particularly the physical origins galactic outflows. Several mechanisms related to stars are able to generate outflows: young stars ionize the surrounding gas and create cavities (HII regions), which expand and shock because of radiative pressure, while supernovae (SNe) heat and accelerate gas around them. It has been shown that the combination of both SNe feedback and radiation pressure from young stars generate outflows that are more powerful than the sum of the models taken separately (non-linear coupling). Nonetheless, these stellar outflows are not powerful enough to explain the huge fraction of baryons located outside galaxies.
Beside stellar mechanisms, there are also several kinds of galactic nucleus-related mechanisms that are able to drive ultra-fast outflows. Active galactic nuclei (AGN) are able to heat and push the gas away from the central region, creating diffuse bubbles, while radiation from AGN can ionize and accelerate the surrounding gas. The major differences between AGN and stellar feedback are the location and number of sources (one at the center of the galaxy, versus plenty of sources distributed in the galactic disk), and, their power (AGNs are the most powerful objects in the Universe).
The main novelty of this project is to account for all kinds and sources of feedback listed above simultaneously, and follow the galaxy evolution at high resolution over an extended period of time, to really understand the physical origins of galactic outflows and what are their main driving mechanisms. Finally, if the coupling between stellar- and AGN-driven outflows is also non-linear, our simulations will generate outflows even more powerful, with a mass loading that might be sufficient to explain the missing baryons problem..