19th Project Access Call – Awarded Projects

Results of the 19th Call for Proposals for Project Access.

Projects from the following research areas:

Biochemistry, Bioinformatics and Life sciences (8) Chemical Sciences and Materials (11) Earth System Sciences (3)
Engineering (12) Fundamental Constituents of Matter (11) Universe Sciences (5)

Biochemistry, Bioinformatics and Life sciences (8)

DynPil – Functional dynamics of the bacterial type-IV pilus and type-2 secretion system pseudopilus

Project Title: DynPil – Functional dynamics of the bacterial type-IV pilus and type-2 secretion system pseudopilus

Project Leader: Dr Massimiliano Bonomi, Institut Pasteur, France

Resource Awarded

  • 93 200 000 core hours on Piz Daint hosted by CSCS, Switzerland
Details

Collaborators

  • Yasaman Karami, Institut Pasteur, France

Abstract
Many aspects of bacteria’s life involve proteins that are assembled into long and dynamic filaments. Understanding the mechanisms that ensure their structural integrity and at the same time confer the flexibility needed to perform their functions is of paramount importance. In this regard, molecular dynamics simulations are a powerful computational microscope that can assist structural biology techniques in elucidating the dynamic behaviour of complex systems. In this project, we will perform extensive atomistic simulations of two fibers that play a crucial role in bacteria: the type-4 pilus (T4P) and the type-2 secretion system pseudopilus (T2SS-Ps). Our simulations will enable characterizing at atomistic detail the network of interactions between the building blocks of these filaments and they will reveal the molecular mechanisms that control their dynamic behavior. The knowledge acquired in this project will be instrumental to design therapeutic strategies to interfere with the function of T4P and T2SS-Ps. Given the size of these systems and the time scales under investigation, the PRACE Tier-0 computing performance is required to execute the proposed simulations. The PRACE allocation will give us the unique opportunity to investigate 22 different types of T4P and T2SS-Ps for a total simulation time of ~1 millisecond.

DPACK – Detecting DNA damage in nucleosome core particles: packed versus free DNA

Project Title: DPACK – Detecting DNA damage in nucleosome core particles: packed versus free DNA

Project Leader: Prof. Dr. Ursula Roethlisberger, Ecole Polytechnique Federale de Lausanne, Switzerland

Resource Awarded

  • 39 000 000 core hours on JUWELS hosted by GCS at FZJ, Germany
Details

Collaborators

  • Polydefkis Diamantis, EPFL, Switzerland
  • Murat Kilic, EPFL, Switzerland
  • Francois Mouvet, EPFL, Switzerland

Abstract
So far, most computational studies of DNA have focused on free DNA in solution. However, in the cell, most of the DNA is not free, but packed within nucleosome core particles (NCPs). Packed DNA exhibits very different mechanical properties than free DNA. The relevance of free DNA studies for the cellular environment is therefore questionable. In particular, the formation and repair of DNA lesions can differ largely in the two cases. Characteristically, the efficacy in uracil and 8-oxoguanine detection and repair is decreased in chromatin. Here, the redox properties of native and defect DNA within NCP will be calculated through a protocol employing hybrid QM/MM MD simulations. 4 defects will be introduced to native DNA. This study is particularly interesting, given the recently proposed mechanism of detection of DNA defects via charge transfer, based on differences in redox properties. This work will allow assessing whether defect recognition and repair can occur in NCPs, or if it happens only when the DNA is in an unfolded state. Finally, in view of the involvement of many base defects in the genesis of various diseases, results from this study could trigger the design of specialized drugs targeting these lesions in their biologically most relevant environment.

hMBRT: hadron-MBRT, towards clinical trials

Project Title: hMBRT: hadron-MBRT, towards clinical trials

Project Leader: Dr Yolanda Prezado, Centre national de la recherche scientifique, France

Resource Awarded

  • 17 200 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France
Details

Collaborators

  • Rachel Delorme, CNRS, France
  • Tim Schneider, CNRS, France
  • Martinez Immaculada, ALBA Synchrotron, Spain
  • Consuelo Guardiola, CNRS, France
  • Ludovic De Marzi, Institut Curie, France

Abstract
Radiotherapy (RT) is one of the most frequently used methods for cancer treatment (above 50% of patients will receive RT). Despite remarkable advancements, the normal tissues tolerances continue being the main limitation in RT, still compromising an efficient treatment of radioresistant tumors (i.e. gliomas) or paediatric cancers. To overcome this limitation, we propose a new approach, hadron minibeam radiation therapy (hMBRT), which partners the normal tissue sparing of submillimetric, spatially fractionated beams with the improved dose deposition of ions. Along this line, proton minibeam radiation therapy has already shown a remarkable reduction of neurotoxicity and an important widening of the therapeutic window for high-grade gliomas in small animal experiments. To prepare phase I/II clinical trials, we need to use Monte Carlo (MC) simulations and HPC to optimise the minibeam generation and to develop a suitable dose calculation engine for patients. Our MC simulations of heavy ions MBRT, performed thanks to the prior access granted in PRACE, allowed us to show the advantages of that approach. Our goal is to perform an in depth dosimetric evaluation to guide and interpret the biological experiments. The small field sizes used require important calculation resources. Only HPC will allow us to perform these calculations.

GPCRBias – Understanding pathway selectivity in GPCR signalling

Project Title: GPCRBias – Understanding pathway selectivity in GPCR signalling

Project Leader: Prof. Francesco Gervasio, University College London, United Kingdom

Resource Awarded

  • 30 000 000 core hours on MareNostrum 4 hosted by BSC, Spain
Details

Collaborators

  • Silvia Acosta-Gutiérrez, University College London, United Kingdom

Abstract
The long-term goal of this project is to understand pathway selectivity in G protein-coupled receptors (GPCRs) by means of enhanced-sampling simulations. GPCRs are targeted by more than 30% of approved medicines, highlighting their role in health and disease. Notwithstanding recent breakthroughs in crystallography and the increasing number of GPCR structures, the mechanism of their selective activation is unclear. GPCR signalling involves allosteric effects, significant conformational changes and the recruitment of specific intracellular partners. Thus, static crystal structures need to be complemented by other techniques, such as simulations, to fully understand their activation mechanisms. We plan to use enhanced-sampling atomistic molecular dynamics simulations combined with state-of-the-art force fields to understand how ligands can selectively activate specific signalling pathways. We have selected two GPCR families, the adenosine receptor (a prototypical class A family) and the glucagon receptor (class B). The simulation of the complex in a realistic environment needs substantial HPC resources. The results will be validated with experimental data provided by our industrial collaborators at Sosei Heptares. Understanding selective pathway activation in these two GPCR families will have a significant impact on rational drug discovery by enabling selective targeting and aiding the design of more effective and less toxic GPCR drugs.

AntiDENV – In silico design of anti-idiotypic antibodies as broad-spectrum vaccines of Dengue and Zika

Project Title: AntiDENV – In silico design of anti-idiotypic antibodies as broad-spectrum vaccines of Dengue and Zika

Project Leader: Dr. Walter Rocchia, Istituto Italiano di Tecnologia, Italy

Resource Awarded

  • 19 700 000 core hours on MARCONI (Broadwell) hosted by CINECA, Italy
Details

Collaborators

  • Miguel Soler, Istituto Italiano di Tecnologia, Italy
  • Andrea Spitaleri, Ospedale San Raffaele, Italy

Abstract
The immune system provides extremely powerful weapons to defend organisms from external and internal threatens. Exploiting and increasing its potential is a big challenge, due to its fascinating but also daunting complexity. However, this may lead to incredible advances in both fundamental and applied sciences. A valuable contribution in this respect comes from experimental data and computer simulation trying to describe and understand processes at the molecular level. For instance, the availability of crystal structures of an antigenic determinant common to all Dengue virus (DENV) serotypes and Zika (ZIKV) bound to a human antibody allows the unprecedented approach of a rational, computationally driven, design of an antigen exposing the broad-spectrum, highly-neutralizing DENV and ZIKV epitopes. We propose a protocol of in silico mutagenesis coupled to affinity and selectivity screening in order to test the anti-idiotypic antibody hypothesis, i.e. the capability of an antibody directed against the binding regions of another antibody (anti-A) to mimic the molecular features of the original antigen A. Our collaboration with an experimental group of molecular immunologists will close the synergistic loop. This might pave the way towards a single vaccine against all of the DENV serotypes and ZIKV, with a very prominent societal impact.

Antarex for Zika

Project Title: Antarex for Zika

Project Leader: PhD Carmine Talarico, Dompé, Italy

Resource Awarded

  • 50 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Federico Ficarelli, CINECA, Italy
  • Carlo Cavazzoni, CINECA, Italy
  • Candida Manelfi, Dompé, Italy
  • Andrea Rosario Beccari, Dompé, Italy

Abstract
The main goal of this project is to offer to the medical and scientific community a tool for the immediate response to a public health emergency (PHE) like ZIKA viruses and other pathogens as well. Thanks to the HPC and Computer-aided drug design (CADD) techniques, will be possible to carry on virtual screening against all possible pharmacological targets relevant for the PHE. The platform has been evaluated on 12K launched and clinical phase drugs and an experiment was already carried out on a 1.2 Billion virtual molecules that can be fast synthetized by using validated in silico chemical reaction and commercial chemical reagents. Our goal will be screen 12 Billion of molecules, from our Tangible Chemical Space database, on ZIKA’s proteome, to identify and evaluate new potentially active molecules. Dompé’s mission will be to making available for the EU (and National) Healthcare System the exascale virtual screening platform providing the fastest response for virus infections and multidrug-resistant bacteria basing on a Virtual chemical space of 100 billion ready to screen molecules (up to 1 Trillion in 2020), an Ultra High Performance Virtual Screening Platform based on LiGen™ Tool and Multi target poly-pharmacological approach.

Dynamics of the spliceosome bound to anti-cancer agents

Project Title: Dynamics of the spliceosome bound to anti-cancer agents

Project Leader: Prof. Modesto Orozco, IRB Barcelona, Spain

Resource Awarded

  • 80 000 000 core hours on Piz Daint hosted by CSCS, Switzerland
Details

Collaborators

  • Federica Battistini, Institute for Research in Biomedicine (IRB-Barcelona), Spain

Abstract
Cancer is a leading cause of death worldwide, and new treatment approaches are needed for this deadly disease. Compounds targeting the spliceosome have shown significant antitumoral effects; however, their mechanisms of action are unclear. They bind at a similar site, but small modifications to their scaffolds can have profoundly different effects on alternative splicing (AS), suggesting that their activity can be tuned to favor certain splicing events. This project will utilize molecular dynamics simulations to investigate conformational changes promoted by various spliceosome modulators. This will be a part of a multi-disciplined approach integrating structural biology and genome-wide RNA sequencing approaches in collaboration with laboratories at the forefront of spliceosome research. High performance computing will be vital to the success of this project. The inhibitors bind a large subunit of the spliceosome. We propose to conduct simulations with several inhibitors, different pre-mRNA substrates, and in the presence or absence of cancer-associated spliceosome mutations. These experiments will be very informative but will require a lot of computing power. The major outcome of these studies will be an improved understanding of mechanisms driving AS by these modulators. This will provide the basis for design of compounds better able to target cancer-specific splicing events.

TCR_MD – Towards a molecular understanding of the T-Cell Receptor dynamical behavior

Project Title: TCR_MD – Towards a molecular understanding of the T-Cell Receptor dynamical behavior

Project Leader: Prof. Marco D’Abramo, Sapienza University of Rome, Italy

Resource Awarded

  • 34 000 000 core hours on Piz Daint hosted by CSCS, Switzerland
Details

Collaborators

  • Josephine Alba, Sapienza University of Rome, Italy

Abstract
T cells, central actors of adaptive immune responses, guard against foreign invasion of microbial pathogens by mechanisms that effectively distinguish foreign from self peptides. T cell receptors (TCRs) recognize antigenic peptide fragments only when presented by an appropriate major histocompatibility complex (MHC) molecule. Following the interaction with a T cells co-receptor CD8, the MHC of class I presents antigen to cytotoxic T cells, which kill the cells exposing the pMHCI complex on their surface. In our work we will use HPC infrastructures to characterize the TCR-MHC-peptide complex conformational behavior by means of all-atoms molecular dynamics simulations. In this project, we would understand if there is any structural or dynamical component playing a major role in the activations, if the recognition generate conformational changes on the receptor, peptide or on MHC I, and if these changes are affected by the peptide or TCR sequence. The main goal is to understand at molecular level of details the mechanism involved in the T cell recognition and in the response to a specific antigen.

top

Chemical Sciences and Materials (11)

ANCIENT_ROME – study of mANy body exCItations in dEfective titaNium dioxide maTeRials by ab-initiO MEthods

Project Title: ANCIENT_ROME – study of mANy body exCItations in dEfective titaNium dioxide maTeRials by ab-initiO MEthods

Project Leader: dr Claudia Cardoso, Italian Research Council, Italy

Resource Awarded

  • 30 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Andrea Ferretti, Italian Research Council, Italy
  • Maurizia Palummo, Tor Vergata University, Italy
  • Daniele Varsano, Italian Research Council, Italy
  • Annabella Selloni, University of Princeton, USA
  • Ivan Marri, Italian Research Council, Italy
  • Letizia Chiodo, Campus BioMedico University, Italy

Abstract
Due to the growing concerns for environmental and energy issues, interest in metal-oxide semiconductors has increased considerably in the last decades. TiO​2 has emerged as one of the most attractive materials for a number of technological applications, including photovoltaics and photocatalysis. TiO2 is cheap, non-toxic and can be easily synthesised through well-established procedures. However, despite their huge interest, the excited state properties of TiO2 are still largely unexplored. In particular, theoretical simulations have generally considered only ideal systems, neglecting the presence of defects in the material, due to the intrinsic complexity of including defects in the excited states description. We propose here a novel advanced analysis of the structural, electronic, and optical properties of defective bulk anatase and TiO2 nanocrystals based on state-of-the-art Density Functional Theory and Many Body Perturbation Theory (GW and Bethe Salpeter Equation -BSE-). These methods have been demonstrated to provide a detailed quantitative description of the defect levels and optical response of doped semiconductors, thereby qualifying as an ideal approach for understanding the PV and PC properties of TiO2. We expect our analysis, never applied before to these systems, to set a landmark in the study of defect energy levels and optical absorption of TiO2 structures and, in turn, to open new perspectives for future theoretical and experimental studies and to have an important impact in technological and industrial fields. Due to the complexity of the theoretical methods, this project can be realized only by exploiting resources offered by modern Tier0 systems.

RELTRAMEDI-Role of the electron-electron and electron-phonon coupling in transition metal dichalcogenides

Project Title: RELTRAMEDI-Role of the electron-electron and electron-phonon coupling in transition metal dichalcogenides

Project Leader: Prof Sandro Sorella, SISSA, Italy

Resource Awarded

  • 65 900 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Kousuke Nakano, SISSA, Italy
  • Natanael de Carvalho Costa, SISSA, Italy

Abstract
In this project we investigate fundamental properties of transition metal dichalcogenides (TMD) using recently developed ab-initio methods, and strongly correlated techniques, both based on state of the art quantum Monte Carlo methods. This study may shed light over open questions about TMD, specially concerning the nature of their CDW phase and its superconducting properties. In the first stage of the project we plan to establish on one hand the first principle determination of the stable structures of TMDs by state of the art quantum Monte Carlo techniques and, on the other hand, the qualitative phase diagram of the Holstein-Hubbard model in two dimensional clusters. In particular we would like to address how the competition of charge density wave, magnetic or superconducting phases depends upon the parameters of the model and the lattice structures considered. In the second and final part of the project we plan to compare the predictions of the model calculations and the first principle calculations, in order to understand whether it is possible to model these materials, very important from the technological point of view, with simple correlated lattice models containing few parameters. The final aim of this project is to determine the effective parameters of the model fully ab-initio and open the way for very effective methods of simulations for material properties, especially in complex electronic systems, where standard DFT based approaches may fail.

Exciton Separation at Catalytic Interfaces from Many-Body Perturbation Theory

Project Title: Exciton Separation at Catalytic Interfaces from Many-Body Perturbation Theory

Project Leader: Dr. Sivan Refaely-Abramson, Weizmann Institute of Science, Israel

Resource Awarded

  • 30 000 000 core hours on MareNostrum 4 hosted by BSC, Spain
Details

Collaborators

  • Sara Barja, Centro de Física de Materiales, Spain
  • Zhen-Fei Liu, Wayne State University, USA

Abstract
We will develop and use advanced computational approaches to study energetically excited states and reaction dynamics in catalytic processes upon interfaces, processes of critical importance in energy conversion and storage. We will study charge separation and transport across heterogeneous photocatalytic interfaces, and examine the effects of the participating molecular and surface components, their binding, and their structural order while including interactions with lattice vibrations. To be able to fine tune the structural effects on the interactions at the interface, predictive methods are absolutely required. In particular, understanding the dielectric environment, which is very heterogeneous and complex upon such interfaces, is crucial. For this, we will use and develop advanced many-body perturbation theory approaches within the BerkeleyGW code, that allows the use of predictive ab initio electronic-structure approaches on highly complex systems. The BerkeleyGW code is massively scalable, and the proposed computations are feasible, but require a large number of nodes to be operated in parallel on large-scale HPC systems and using tens of millions of computing hours, as we request below.

QSL_SPECTRE: Characterization of Quantum Spin Liquids using Energy Spectroscopy

Project Title: QSL_SPECTRE: Characterization of Quantum Spin Liquids using Energy Spectroscopy

Project Leader: Prof. Dr. Andreas Läuchli, University of Innsbruck, Austria

Resource Awarded

  • 16 000 000 core hours on Joliot-Curie (AMD) hosted by GENCI at CEA, France
  • 25 000 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France
Details

Collaborators

  • Clemens Ganahl, University of Innsbruck, Austria
  • Alexander Wietek, Simons Foundation, USA
  • Sylvain Capponi, Universite Paul Sabatier Toulouse 3, France

Abstract
Strongly correlated quantum states of matter exist in numerous materials or synthetic quantum optical experiments. Novel exotic phases of matter, such as quantum spin liquids, are expected to emerge in many of these systems. Although many experimental candidate systems for quantum spin liquids exist, it is still unclear under which circumstances these states are realized and how they could reliably be detected. A theoretical understanding of these questions, however, is hard to achieve due to the strong correlation effects and often relies on numerical simulations. We propose to solve two key problems in the field of frustrated magnetism, using the Exact Diagonalization (ED) technique. The phase diagrams of extended Heisenberg model on the two-dimensional triangular and kagome geometry are fundamental challenges, that hitherto have not been fully resolved. We address, whether so-called Dirac or Z_2 spin liquids are realized in these models. We study systems with 48 spin 1/2 particles, which amounts to diagonalizations of matrices with a linear dimension of up to 5 * 10^11, systems so large that only HPC machines allow to study them. Our study will contribute to an understanding, how quantum spin liquids emerge and can be observed experimentally. As the ED method is unbiased, our data will also be a reference for other approximate numerical techniques.

PROVING-IL – PeROVskite Interface eNgineerinG with Ionic Liquids

Project Title: PROVING-IL – PeROVskite Interface eNgineerinG with Ionic Liquids

Project Leader: Dr Simone Meloni, Sapienza University of Rome, Italy

Resource Awarded

  • 41 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Lorenzo Gontrani, University of Bologna, Italy
  • Alessio Filippetti, University of Cagliari, Italy
  • Claudia Caddeo, CNR – Natl. Research Council, Italy
  • Diego Di Girolamo, University of Rome Sapienza, Italy
  • Alessandro Mattoni, CNR – Natl. Research Council, Italy
  • Marco Tortora, Sapienza University of Rome, Italy

Abstract
3rd generation photovoltaics will further revolutionize the field of energy enabling a distributed energy harvesting. To achieve this goal, new low-cost, high-efficiency and robust materials are needed. Halide perovskites are key materials that attracted significant attention over the last 5-10 years thanks to their high efficiency: > 23% of the harvested light is converted in electric current in the best perovskite solar cells. To further enhance the efficiency, reliability and stability of perovskite solar cells several shortcomings must be solved, one being the relatively poor electron extraction at the TiO2/perovskite contact in planar solar cells, which are simpler to fabricate and thus commercially more promising. One of the strategies recently proposed to address this issue is the introduction of a layer of ionic liquids between the TiO2 and perovskite films. The objective of the PROVING-IL project is to use advanced simulation techniques to understand the microstructure and electronic transport properties of the TiO2/ionic liquid/perovskite heterostructure that lead to the sizable enhancement of performance measured in experiments. The final objective of the PROVING-IL project is to identify design principles to optimize the chemical composition of ionic liquids that can bring to the fabrication of high-performance planar perovskite solar cells. In the PROVING-IL project we will combine classical and ab initio enhanced sampling molecular dynamics to the identify accurate structure of the TiO2/MAPI and TiO2/IL/MAPI interfaces, where titania – TiO2 – is the electron transport layer (ETL), MAPI stands for methyl ammonium lead iodide, the prototypical hybrid lead halide perovskite considered in perovskite solar cells, and IL stands for ionic liquid, here 1-butyl-3-methylimidazole tetrafluoroborate. We will also consider the effect of defects on the electron transport properties across TiO2/MAPI and TiO2/IL/MAPI interfaces, as well as the healing effect of IL on surface and defects electronic states. Finally, long enhanced classical simulations on very large samples of the TiO2/MAPI and TiO2/IL/MAPI interfaces will allow us to establish possible interdiffusion of the various species, and thus the stability of the engineered surface.

SHP2-ReM-MD – The regulatory mechanism of the SHP-2 protein: a Molecular Dynamics investigation.

Project Title: SHP2-ReM-MD – The regulatory mechanism of the SHP-2 protein: a Molecular Dynamics investigation.

Project Leader: Prof. Gianfranco Bocchinfuso, Rome Tor Vergata University, Italy

Resource Awarded

  • 30 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Paolo Calligari, Rome Tor Vergata University, Italy
  • Valerio Santucci, Rome Tor Vergata University, Italy
  • Antonio Palleschi, Rome Tor Vergata University, Italy
  • Lorenzo Stella, Rome Tor Vergata University, Italy
  • Giorgio Ripani, Rome Tor Vergata Universit, Italy

Abstract
The present proposal follows a previous project funded in the 16th PRACE Program, focused on the protein SHP2, which regulates different pathways. SHP2 is a multidomain phosphatase with two SH2 domains (N-SH2 and CSH2) followed by the catalytic protein tyrosine phosphatase (PTP) domain. In the basal state, SHP2 is auto-inhibited, as the N-SH2 domain blocks the PTP active site (“closed” conformation). SHP2 activation is mediated by interactions between its SH2 domains with partners containing short amino acid motifs comprising a phosphotyrosine residue. This event is coupled with a rearrangement of the domains, whose final effect is a greater accessibility of the active site on the PTP domain (“open” conformation). SHP2 mutations are involved in several forms of leukemia and other cancers. Most of the pathological mutations map at the N-SH2/PTP interface, and their functional effect is an increase in SHP2 activity. The finding that RTK-driven cancer cells depend on SHP2 for survival increased the relevance of this protein enormously, beyond the pathologies caused by its mutations, making it a major target in cancer therapy. The present project is divided in two work-packages (WP): WP-1 aims to clarify the allosteric mechanism of SHP2 regulation. Two different structures for the SHP2 open state have been proposed in 2018. Several lines of evidences suggest that these structures are not fully representative of the open state in solution. Therefore, the SHP2 active state and the pathway followed during activation remain elusive. In WP-1, we plan to use parallel-tempering well-tempered metadynamics to investigate the closed-to-open transition. The simulations will be carried out on the wild type protein and five mutants. These simulations will shed light of the SHP2 activation process and on its perturbation by pathogenic mutations. In WP-2, we will develop a peptide-based lead compound binding the N-SH2 domain of SHP2 into a drug candidate. The parent peptide has been designed in the previous project as possible drug to inhibit SHP2 protein-protein interactions. The idea that a peptide binding N-SH2 can be used as drug against SHP2 represents a novel pharmacological strategy. In the traditional approach, phosphatase activity is inhibited through a direct binding of the drug in the PTP active site or with drugs binding an allosteric site able to block SHP2 in the auto-inhibited state. In the project, we discuss the limitation of these approaches, and we explain why the inhibition of the SHP2 interactions can be used as strategy to limit the effects of SHP2 aberrant function. In WP-2 the Potential of Mean Force profile of 30 different sequences will be evaluated starting from Umbrella Sampling Molecular Dynamics. The sequences with the best binding energy will be synthesized and tested. Of note, we are preparing a patent on the parent peptide developed in the previous project. More in general, a strength of the project is a tight collaboration with experimentalists, which allows a prompt validation of the computational predictions.

EXTEND – EXcitonic instability in two dimensional tungsTEN Ditelluride

Project Title: EXTEND – EXcitonic instability in two dimensional tungsTEN Ditelluride

Project Leader: Dr Daniele Varsano, Italian Research Council, Italy

Resource Awarded

  • 45 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Maurizia Palummo, University of Tor Vergata, Italy
  • Elisa Molinari, Universita di Modena e Reggio Emilia, Italy
  • Seyedeh Samaneh Ataei, Italian Research Council, Italy
  • Massimo Rontani, Italian Research Council, Italy

Abstract
EXTEND focuses on a 2d system, tungsten ditelluride, which has recently attracted much interest for its fundamental physical properties, driven by electron interactions. WTe2 shows a very large positive magnetoresistance, pressure-driven superconductivity, and possible Weyl semimetal state in the bulk form, while it is a topological insulator or even a superconductor in the monolayer form. The origin of this behaviour is not understood. Here we explore the possibility that it arises from a purely electronic instability, with the spontaneous generation of excitons -–electron–hole (e-h) pairs bound together by Coulomb attraction— and the stabilization of an exotic phase of matter called excitonic insulator (EI). The existence of the EI could in principle be favoured in this system because of the peculiar combination of suppressed 2d screening and indirect energy gap: a quantitative understanding requires a delicate evaluation of the balance of the quasiparticle band gap and the e-h binding, which are interrelated, and a reliable estimate of the resulting phase diagram. To carry out the study of the electronic and excitonic properties of the monolayer form, in view of the required accuracy we use ab-initio Density Functional Theory and Many Body Perturbation Theory (MBPT), well established for this purpose. We take advantage of very recent code developments which enable MBPT calculations in a full spinorial base, quasiparticle corrections on top of hybrid functionals, and solution of the Bethe Salpeter equation at finite momentum. PRACE resources are essential to ensure the extreme accuracy and high sampling required for predictive estimates of exciton energies vs strain. Based on the MBPT results, we then solve a self-consistent gap equation leading to the EI phase diagram in a multidimensional space vs temperature and strain. Since WTe2 shows topological properties and superconductivity, it is of great fundamental interest to investigate the relation between topological, superconductive, and excitonic instabilities in this system. If successful, our proposed search of the EI phase in T’-WTe2 could lead to excitonic superfluidity, at critical temperatures much higher than those of the BCS state. Excitons in the EI phase would condense at equilibrium (at variance with the case of exciton-polaritons, where condensation occurs out of equilibrium under optical pumping), which makes any achievement in this field extremely appealing. Also, exciton condensation may break inversion symmetry in a nominally centrosymmetric material like T’-WTe2, possibly leading to a permanent macroscopic electric polarization of purely electronic origin associated with the dipole of condensed excitons: a ferroelectric excitonic insulator. This is expected to radiate coherently if excited, with a characteristic frequency that may be controlled through a static external electric field. It would thus provide a highly tunable THz coherent radiation source that could be controlled on ultrafast time scales, in contrast to the much slower ionic dynamics of conventional ferroelectrics. Combined with the coupling of spin and valley degrees of freedom, this would be extremely attractive for spintronics and optoelectronics.

Ab initio modelling of coloured centers in insulators and semiconductors: the NV center in diamond at high pressure, the bivacancy in silicon carbide, vacancies in boron and boron carbide.

Project Title: Ab initio modelling of coloured centers in insulators and semiconductors: the NV center in diamond at high pressure, the bivacancy in silicon carbide, vacancies in boron and boron carbide.

Project Leader: Dr Nathalie VAST, CEA -DRF – IRAMIS, France

Resource Awarded

  • 16 000 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France
Details

Collaborators

  • Michele Casula, CNRS, France
  • Romuald Bejaud, Ecole Polytechnique, France
  • Antoine Jay, CNRS, France
  • Maksim Markov, Université Catholique de Louvain la Neuve, Belgium

Abstract
Point defects (PDs) in otherwise perfect crystals play a role of outstanding importance in emergent quantum metrology such as magnetometry. Both diamond and moissanite (SiC) are used for anvils, and PDs are used for magnetometry inside the anvil cell. Knowledge of the maximal pressure at which the PD is still able to detect the magnetic field is of tremendous technological importance. The presence of PDs in some crystals also questions the origin of their (non) thermodynamic stability at finite temperature. However, it is experimentally difficult to access to the local atomic structure of PDs, so that they are mainly known through spectroscopic data. Ab initio calculations based on density functional theory allow to compute mono-determinant electronic states. Combined with extended Hubbard Hamiltonians, they allow to compute all of the multi-determinant levels, thus providing the missing link between spectroscopic data and the underlying atomic structure. In the present project, we will first concentrate on the NV center in diamond with various charge states, and study the triplet-triplet transition and singlet-singlet transition as a function of pressure and non hydrostatic strain. Adiabatic effect will be accounted for by including phonons in the Hubbard-Holstein model. We will extend the Hubbard model to the case of the bivacancy in silicon carbide. Finally, partial occupations of Wyckoff’s sites and vacancies will be studied in boron carbide and elemental boron. High performance computing resources are mandatory because the long range nature of the electrostatic/elastic interactions brought in by (charged) defects requires the use of supercells of several hundreds of atoms. Moreover, the use of hybrid functionals is mandatory to account properly for dynamical correlations, and we plan to use the HSE06 functional. The project first expected outcome is the spectroscopic characterization of some PDs of diamond, SiC and boron carbide, with, as an application, the knowledge of the maximal pressure at which these defects may act as magnetic sensors in diamond and moissanite anvil cells. The second expected outcome is a major contribution to the long-standing puzzle of the nature of the ground-state of -boron and boron carbide crystals.

DIMAB – DIslocations in bcc Metals: an AB initio study

Project Title: DIMAB – DIslocations in bcc Metals: an AB initio study

Project Leader: Dr Lisa Ventelon, CEA, France

Resource Awarded

  • 15 000 000 core hours on JUWELS hosted by GCS at FZJ, Germany
Details

Collaborators

  • Lucile Dezerald, Institut Jean Lamour, France
  • François Willaime, CEA, France
  • David Rodney, Université Claude Bernard Lyon 1, France
  • Bassem Ben Yahia, Institut Jean Lamour, France
  • Guillaume Hachet, CEA, France
  • Emmanuel Clouet, CEA, France

Abstract
The fundamental understanding of metallurgy constitutes an essential aspect in worldwide basic research. Thus it is essential to keep investing in the next generation of metallic alloys, thereby helping tackling some of the societal challenges related to energy, renewables and CO2-reducing technologies. This is the reason why it is of great importance to understand the mechanisms underlying deformation in metals and alloys as a first step for a better apprehension of current material aging as well as for the next step for development of advanced metal-based products. In this context, the understanding of the behavior of metallic structural materials requires the modeling of atomic-scale events and their consequences on mechanical properties, thereby involving coupling between mechanical behavior and atomic bonding down to the electronic structure level. In this framework, the goal of this proposal is to study the link between dislocation core properties and plasticity in metals and alloys using density functional theory (DFT) calculations. Such DFT calculations are extremely computationally demanding but they are necessary to build a complete and physical picture of plasticity in metals and alloys that can serve in turn as input for larger-scale plasticity models such as dislocation dynamics, as well as as an absolute necessary DFT database in order to better train on dislocation geometries promising approaches based on machine learning. The goal of this proposal is to study at the atomic scale the link between screw dislocation core properties and plasticity in metals and alloys using DFT calculations. The first subproject of the first year will focus on the poorly understood phenomenon of anomalous slip in BCC metals, constituting a striking violation of Schmid’s law, using DFT calculations. Comparison between BCC metals, namely Mo, Fe and Nb, as well as comparison with experimental data, will enable understand and predict dislocation anomalous glide in all BCC metals. The second subproject of the first year will focus on the origin at the atomic scale of interstitial solute effects on mechanical properties in metals and alloys, and particularly to study how helium and hydrogen interstitials form atmospheres around dislocations and enhance/impede flow through notably pipe diffusion phenomena. The proposed approach in this project couples ab initio calculations of interaction energies and energy barriers and thermodynamic modeling of solutes and kinks in order to describe the dislocation glide mechanisms in the presence of helium and hydrogen solutes. The second year will be devoted to very innovative calculations on non-conventional glide in W and W(Re) alloys. The proposed simulations are very computationally demanding and challenging as they will use the Activation Relaxation Technique (ART), which allow the automated search of migration paths in systems of complex configurations. The third year also involves cutting-edge DFT calculations on dislocation-solute interaction in the FeCCr ternary system, where a substantial number of configurations will have to be investigated with carbon interstitial and chromium substitutional solutes interacting together as well as with the dislocation core.

1D-2D CDW CROSSOVER – Crossover in the mechanism determining CDW from 1 to 2 dimensions: the crucial role of Quantum Anharmonicity

Project Title: 1D-2D CDW CROSSOVER – Crossover in the mechanism determining CDW from 1 to 2 dimensions: the crucial role of Quantum Anharmonicity

Project Leader: Prof. Matteo Calandra, CNRS, France

Resource Awarded

  • 30 000 000 core hours on Joliot-Curie (AMD) hosted by GENCI at CEA, France
Details

Collaborators

  • Ion Errea Lopez, University of the Basque Country, Spain
  • Francesco Mauri, Università di Roma La Sapienza, Italy
  • Marco Campetella, CNRS, France
  • Jianqianq Zhou, CNRS, France
  • Oliviero Bistoni, Sorbonne Université, France
  • Francesco Belli, University of the Basque Country, Spain
  • Antonella Mennino, University of the Basque Country, Spain
  • Paolo Barone, Università di Roma la Sapienza, Italy
  • Lorenzo Monacelli, Università di Roma La Sapienza, Italy

Abstract
Charge density waves (CDWs) are ubiquitous in condensed matter physics. They occur in very different systems ranging from inorganic molecules (carbyne, polyvyne), blue Bronzes, di and trichalcogenides, High Tc cuprates and many other systems. They often coexist or compete with other orders like superconductivity or magnetism. A complete understanding of CDWs requires an accurate description of the three key interactions at play:electron-electron, electron-phonon and quantum anharmonicity. Only recently, partly due to the developments of our consortium, it has become possible to treat non-perturbative quantum anharmonicity completely from first principles via the stochastic self-consistent harmonic approximation. With this approximation, in our previous PRACE allocation, we were able to show that in 3D and quasi 2D solids anharmonicity completely determines the charge density wave critical temperature. Thus, the well accepted theory of charge density wave instabilities based on the temperature dependence of the electronic charge response and on concepts like Fermi surface nesting, completely breaks down and it is totally inappropriate. However, this could be different in quasi 1D system as the Fermi surface are simple (sometimes point-like) and their topology could play an important role. In this proposal we plan to investigate most of the 1D, quasi 1D and some of the 2D charge density systems with full inclusion of quantum anharmonicity to address the mechanism at the heart of the 1D-2D CDW crossover. Our results, ideal continuation of our previous PRACE applications, well lead to a new description of CDW formation from 1D to 3D and a rewriting of the basic mechanisms determining T_CDW. Given the ubiquitous character of CDW our understanding of the interplay between the electron-electron, electron-phonon and quantum anharmonicity will lead to prominent result in inorganic chemistry, contensed matter physics, 2D materials and 1D systems. It is finally important to outline that Quantum anharmonic calculations are not possible without Tier0 infrastructures and HPC.

Electronic and optical properties of high-performance monolayer and multilayer materials

Project Title: Electronic and optical properties of high-performance monolayer and multilayer materials

Project Leader: Prof Nicola Marzari, EPFL, Switzerland

Multi-year Proposal: Year 3

Resource Awarded

  • 60 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Davide Campi, EPFL, Switzerland
  • Marco Gibertini, EPFL, Switzerland
  • Antimo Marrazzo, EPFL, Switzerland
  • Nicolas Mounet, EPFL, Switzerland
  • Thibault Sohier, EPFL, Switzerland
  • Deborah Prezzi, CNR-NANO, Italy
  • Paolo Umari, University of Padova, Italy

Abstract
There is major effort worldwide in exploring the capabilities of graphene and related materials for advanced technological applications; this is even more relevant for Europe, thanks to the 10-year effort spearheaded by the FET Graphene Flagship. While the effort has now focused mostly on a handful of materials – most notably graphene itself and transition-metal dichalcogenides – we believe that many more opportunities are available, and that computational screening and discovery can greatly accelerate the process of identifying the most promising materials for innovative applications. In fact, using high-throughput computational screening on more than 110, 000 experimentally-known inorganic materials, we have recently identified close to 2000 layered candidates that can be exfoliated into mono- and multi-layers. These cover all the materials that give rise to the known 2D monolayers (from graphene to boron nitride, transition-metal chalcogenides, black phosphorus…), but obviously many more. Here, we want to urgently exploit and expand this portfolio, to identify as quickly as possible the systems with the most promising electronic and optical properties. This effort will involve a combination of rapid screening with approximate methods followed by a high-accuracy study of the most promising candidates. Deployment and dissemination will greatly benefit from our extensive expertise in high-throughput methods, powered by the AiiDA materials informatics platform (http://aiida.net), the use of state-of-the-art high-performance open-source codes for electronic-structure simulations, supported by the H2020 MaX Centre of Excellence (http://max-centre.eu), and our commitment towards full dissemination of the data and calculations’ workflows through the Materials Cloud (http://materialscloud.org).

top

Earth System Sciences (3)

TOPSyCled – TOwards a Predictive System for atmospheric CO2 and Climate feedbacks

Project Title: TOPSyCled – TOwards a Predictive System for atmospheric CO2 and Climate feedbacks

Project Leader: Dr Raffaele Bernardello, Barcelona Supercomputing Center – Centro Nacional de Supercomputación, Spain

Resource Awarded

  • 34 600 000 core hours on MareNostrum 4 hosted by BSC, Spain
Details

Collaborators

  • Etienne Tourigny, Barcelona Supercomputing Center, Spain
  • Valentina Sicardi, Barcelona Supercomputing Center, Spain
  • Pierre Antoine Brettoniere, Barcelona Supercomputing Center, Spain
  • Mario Acosta, Barcelona Supercomputing Center, Spain

Abstract
The so-called climate warming hiatus over the first decade of the 21st century took the research community by surprise, prompting a wave of skepticism among the public. A similar situation could occur again, if for example, atmospheric CO2 concentration was changing at a rate that would seem inconsistent with reported trend in global emissions. Even on a year-to-year time scale, changes in atmospheric CO2 growth rate, primarily caused by natural fluctuations are often misinterpreted as reporting of emissions growth rates. Such misunderstandings could be addressed with the development and deployment of a near-term carbon cycle prediction system. TOPSyCled will assess the predictability of the carbon cycle system, via ESM control experiments and decadal hindcasts over the last 60 years. We will then explore predictions of the climate-carbon system in the near-future, assuming anthropogenic emissions follow the United Nations Framework Convention on Climate Change (UNFCCC) Nationally Determined Contributions (NDCs) and quantify the direct impact of emission reductions on CO2 concentrations, accounting for the natural variability of the climate system and the carbon sinks. This represents a major step towards the verification of near-term emission trends, and the timing of any emissions peak, providing policy-relevant analysis for the UNFCCC global stocktakes.

Quantifying Uncertainties and Enhancing the Speed of climate model Tuning

Project Title: Quantifying Uncertainties and Enhancing the Speed of climate model Tuning

Project Leader: Dr Julie Deshayes, CNRS-IPSL-Sorbonne Universite-IRD, France

Resource Awarded

  • 45 000 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France
Details

Collaborators

  • Marie-Alice Foujols, CNRS Sorbonne Universite, France
  • Laurent Fairhead, CNRS Sorbonne Universite, France
  • Marion Devilliers, CNRS, France
  • Martin Vancoppenolle, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Nicolas Lebas, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Claire Levy, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Renaud Person, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Guillaume Gastineau, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Josefine Ghattas, IPSL, France
  • Matthew Menary, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Didier Swingedouw, CNRS, France
  • Juliette Mignot, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Frederic Hourdin, CNRS, Sorbonne Universite, France
  • Simona Flavoni, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Arnaud Caubel, IPSL, France
  • Olivier Marti, CEA, France
  • Frederique Cheruy, CNRS Sorbonne Universite, France
  • Christian Ethe, IPSL, France
  • Amelie Simon, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Patricia Cadule, IPSL, France
  • Victor Estella-Perez, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Jerome Servonnat, CEA, France
  • Jean-Baptiste Madeleine, CNRS Sorbonne Universite, France
  • Ionela Musat, CNRS Sorbonne Universite, France
  • Eric Guilyardi, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Casimir deLavergne, CNRS-IPSL-Sorbonne Universite-IRD, France
  • Yohan Ruprich-Robert, Barcelona Supercomputing Center, Spain

Abstract
Climate change is a central problem for humanity with important ramifications for policy and decision making. Robust and cost-efficient policies on mitigation and adaptation require assessments of current and future risks for natural and human systems. Those assessments rely on numerical simulations performed with state-of-the-art global climate models. These simulations are coordinated at an international level within the Coupled Model Intercomparison Project (CMIP) which provides the bedrock for a large fraction of the publications synthesized in the Intergovernmental Panel on Climate Change (IPCC) reports. These exercises are fundamentally documenting the large uncertainty in the projections that come from the choices made by the ~30 teams that develop CMIP-class models. The latest version, CMIP6, is now in the production phase. The first analysis suggests that several models are more alarming in terms of global mean temperature increase than previous versions, as quantified by the Equilibrium Climate Sensitivity (ECS). Another outcome of CMIP exercises is that the improvements in the performances of climate models are slow, especially when facing the emergency of the climate change. The derivation of the IPSL coupled model for CMIP6 (IPSL-CM6) was an unprecedented coordinated effort during which key processes and parameters for climate sensitivity were identified both in the atmosphere and in the ocean. The present project proposes an ambitious plan to quantify the uncertainties associated with this choice of free parameters in a systematic way using the latest version of the IPSL climate model with two major targets: i) speed up and improve the calibration of the future higher resolution version of IPSL-CM and ii) estimate the associated errors in present-day climate representation (i.e. mean state and variability) and uncertainty in the ECS. The improvements in the future version will target in particular oceanic processes in high latitudes and the potential benefit associated with a significant increase of the grid resolution, the reduction of the classical large patterns of sea surface temperature biases, and an improvement of the rainfall variability in the tropics. For both tuning and uncertainty quantification purposes, the parameter space will be sampled by applying, for the first time in climate change simulations, state-of-the art machine learning approaches developed by the Uncertainty Quantification community. The idea is to replace a long (3 actual years for CMIP6) calibration process for which 15 model configurations were re-adjusted based on typically 20 to 40 sensitivity experiments done by varying one parameter at once, by an automatic random Latin Hypercube sampling of the full parameter domain. First promising tests done since the first submission of the proposal suggest that the approach requires 3 waves of typically 200 few-years-long simulations of the stand alone atmospheric component for 20 parameters. This project will take on the additional challenge of extending this to the calibration of slower ocean and coupled atmosphere-ocean processes.

From Resolved Convection to the Quasi-Biennial Oscillation (reconQBO)

Project Title: From Resolved Convection to the Quasi-Biennial Oscillation (reconQBO)

Project Leader: Dr Marco Giorgetta, Max Planck Institute for Meteorology, Germany

Resource Awarded

  • HLST Support
Details

Collaborators

  • Luis Kornblueh, Max Planck Institute for Meteorology, Germany

Abstract
The tropical quasi-biennial oscillation (QBO) is one of the most prominent dynamical phenomena of the stratosphere. The theory stipulates that wave-meanflow interaction between vertically propagating waves and zonal jets creates the downward propagating easterly and westerly jets of the QBO. Existing simulations of the QBO in general circulation models (GCMs) rely on the parametrized convective heating as a source for resolved tropical waves and gavity wave parameterizations for sub grid scale gravity wave drag. Recent studies showed that the uncertainty originating from the parameterizations and their tuning effectively hinders the understanding of the full QBO cycle in the current climate and consequently obstructs the assessment of climate change effects on the QBO. We therefore propose a first direct simulation of the QBO in a deep-convection resolving GCM that by construction is independent of parameterizations for convection and gravity waves. By comparison of analyses and the direct QBO simulations we expect to understand better the wave meanflow interaction that generates the QBO. On this base we want to address the following questions: • What is the contribution of different types of tropical waves to the progression of the westerly and easterly phase through a life cycle of the QBO? • How do QBO jets influence the tropical deep convection? • How and why does the QBO respond to a warming climate? For the realization of the experiments we want to make global atmospheric simulations at a horizontal grid resolution of ca. 2.5 km and a vertical resolution of ca. 300 m. The length of the experiments are 6 months for the shorter QBO forecast experiments and 2 to 3 years for the full QBO cycle experiment. Such atmospheric experiments have not been tried so far, owing to the tremendous computational size and costs. Concretely we plan to use the ICON general circulation model for this experiment. The ICON model is a joint development of the Max Planck Institute for Meteorology (MPI-M) and the German weather service (DWD), where the ICON model is used for operational numerical weather forecasting at a resolution of 13 km. For research at MPI-M, the ICON model has been employed in a variety of experiments, including a recent global experiment at 2.5 km resolution, albeit at coarser vertical resolution than needed for the QBO study, and only up to 40 days due to the limited available computational resources. In order to overcome this limit we are currently working towards a GPU enabled version of ICON that can be employed on the very large “Piz Daint” GPU-system at CSCS. With this new model version we will explore the formulated QBO problem in idealized and realistic experiments on the “Piz Daint” system.

top

Engineering (12)

DISTRAIN – DIrect numerical SimulaTion of high-RAyleigh number turbulent convection at Infinite Prandtl Number

Project Title: DISTRAIN – DIrect numerical SimulaTion of high-RAyleigh number turbulent convection at Infinite Prandtl Number

Project Leader: Prof. Sergio Pirozzoli, Sapienza, University of Rome, Italy

Resource Awarded

  • 65 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Roberto Verzicco, University of Twente, The Netherlands
  • Paolo Orlandi, Sapienza University of Rome, Italy
  • Davide Modesti, University of Melbourne, Australia
  • Simone Di Giorgio, Sapienza University of Rome, Italy

Abstract
Turbulent convection is an ubiquitous mechanism of heat transfer in natural phenomena and in practical applications, occurring in a wide variety of physical circumstances ranging from stellar activities in astrophysics to natural convection in the atmosphere. Most studies of natural convection have been carried out for Prandtl number (namely ratio of kinematic viscosity to thermal conductivity) around unity, by establishing the flow response to changes in the imposed temperature gradient, as expressed in nondimensional terms by the Rayleigh number. In the present research we focus on natural convection at Prandtl number much greater than unity, and in particular on the extreme regime of infinite Prandtl number, at which the momentum transport equation is dominated by diffusion. We aim at achieving a major step forward with respect to the current state of the art by conducting direct numerical simulations (DNS) at Rayleigh number 10^(12), with four order-of-magnitude increase. The occurrence of large-scale self-organization will also be investigated through a series of additional DNS at more moderate Rayleigh number, in computational domains of various widths. We expect that this research will translate into improved understanding of the complex geophysical processes occurring in the Earth’s mantle, as well as improved prediction of heat transfer in fluids with high value of the Prandtl number, as occurring in many engineering applications.

CornerLES – Corner separation predictions using Large Eddy Simulations

Project Title: CornerLES – Corner separation predictions using Large Eddy Simulations

Project Leader: Dr Dimitrios Papadogiannis, Safran SA, France

Resource Awarded

  • 20 400 000 core hours on Joliot-Curie (AMD) hosted by GENCI at CEA, France
Details

Collaborators

  • Sophie Mouriaux, Safran SA, France
  • Florent Duchaine, CERFACS, France
  • Laurent Gicquel, CERFACS, France

Abstract
The CornerLES project is dedicated to the prediction of the corner separation phenomenon in compressors using Large Eddy Simulations. Corner separation is an unsteady 3D flow separation occurring in the junction between the hub and a compressor blade at moderate or high incidence flow angles. It has various deleterious impacts: it is a source of aerodynamic losses and it affects the operability of the compressor as it can provoke rotating stall. Correctly predicting the phenomenon with numerical simulations is essential for compressor designs. However, achieving such predictions is not easy as corner separations are highly unsteady and turbulent. This has been illustrated well in a linear compressor cascade test case: the NACA65-009, tested in Ecole Centrale Lyon. This compressor blade test case exhibits a large unsteady corner separation which has been attempted to be reproduced numerically with both steady-state Reynolds Averaged Navier-Stokes (RANS) and Large Eddy Simulation (LES). The findings illustrate that the the turbulence models involved in RANS have difficulties in capturing this complex phenomenon. The LES approach, unsteady and resolving the most energetic turbulence length scales, significantly improved the predictions but discrepancies on the unsteady dynamics of the phenomenon persist. A common point between all the reported numerical predictions is that the simulation domain includes a single blade and pitch-wise periodicity is assumed; essentially that the flow is identical across each blade of the test bench/blade row. However, the experimental results challenge this assumption as the large corner separation and the associated blockage impacts the flow across the neighbouring blades and disturbances propagate between blades. The objective of the project is to remove this assumption and perform multi-blade LES of the NACA65 test case with very high mesh resolution in order to finely capture and characterize the corner separation phenomenon and its propagation to the neighboring blade passages. The findings can be directly employed to improve the current compressor design practices as well as improve the available turbulence models for RANS simulations using artificial intelligence techniques.

SPECTRA – Universality of Kolmogorov spectrum in non-ideal turbulent flows

Project Title: SPECTRA – Universality of Kolmogorov spectrum in non-ideal turbulent flows

Project Leader: Dr Alessandra Sabina Lanotte, Consiglio Nazionale delle Ricerche, Italy

Resource Awarded

  • 56 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Luca Biferale, University of Rome Tor Vergata, Italy
  • Guido Macorini, Consiglio Nazionale delle Ricerche, Italy
  • Yuri V Lvov, Rensselaer Polytechnic Institute, USA
  • Fabio Bonaccorso, University of Rome Tor Vergata, Italy
  • Federico Toschi, Technical University of Eindhoven, The Netherlands

Abstract
Most of our understanding of turbulent flows is embedded in the well-celebrated Kolmogorov spectrum of kinetic energy. Indeed the “-5/3” spectral law is assumed in any phenomenological model of three-dimensional turbulent flows, from engineering, to atmospheric modeling or medical applications. Decades of experimental and numerical research have been devoted to assess its validity in a wide range of phenomena. However it is now clear that deviations from it can arise due to different features such as injection mechanism, anisotropic forces, spatial confinement. and we are at the beginning of a new research effort, willing to challenge its universality. This calls on the one hand for opening the parameter space in many different directions, and on the other hand to clearly define the applicability limits of different approaches and modelling solutions. In this project, we want to systematically study the turbulent dynamics of thin layers of stably stratified fluid at changing the energy injection features, i.e. spectral/physical locality. In particular, we aim at disentangling the 3D/2D nature of these flows, together with the wave/turbulence interactions, and understand their impact on the spectral law of the flow. The scientific question is simple but highly non-trivial: depending on the characteristic on energy injection and external forces, which is the turbulent model replacing K41 isotropic spectrum to characterize energy distribution in the system? To answer these questions, we wish to perform Direct Numerical Simulations (DNS) that accurately resolve both in space and time different ranges of scales (larger/smaller that injection, larger/smaller than the confinement scale, etc.) and avoid possible spurious contamination due to finite-size effects. This call for extremely demanding resources, in terms of both high resolution and long-temporal integration. PRACE is the only framework to perform such a study. Our project focuses on basic and fundamental fluid dynamical problems, that do not have a clear answer and a plethora of turbulent regimes are often invoked. The results are expected to impact many fields of applications, and in particular geophysical fluid dynamics. Indeed, the questions asked above can be translated in the language of physical oceanography, where the role of forcing (e.g. wind forcing, or forcing by mesoscale eddies) wrt density stratification, the eddy/wave interactions and their spectral separability are still open matters of research.

PASTEC – Plasma ASistEd Combustion

Project Title: PASTEC – Plasma ASistEd Combustion

Project Leader: Prof. Benoit Fiorina, CentraleSupélec, France

Resource Awarded

  • 20 000 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France
Details

Collaborators

  • Yacine Bechane, CentraleSupélec, France:
  • Vincent Moureau, CORIA-CNRS, France
  • Jean Michel Dupays, CentraleSupélec, France

Abstract
The more and more severe environmental norms on pollutant emission impose a technological breakthrough to combustion-related industries. An efficient solution to reduce pollutant formation is to maintain a relatively low flame temperature. However low-temperature flames are subject to instabilities and extinction, causing safety issues. A promising technology is to stabilize the flame by Nanosecond Repetitively Pulsed (NRP) discharges. A plasma, generated at the flame basis, produces active species and local heating sufficient to sustain the combustion. Despite this proven efficiency demonstrated experimentally, the mechanisms of plasma assisted-combustion are not understood, highlighting the need of numerical simulations. The numerical simulations of interactions between electrical discharges and turbulent flames has yet never been performed and are the final objective of this project. These simulations will be possible by using a novel, recently published, semi-empirical plasma model, designed to capture the influence of NRP discharges on the flame properties. This strategy will be used to include for the first time plasma kinetics in LES of flames. These highly resolved simulations will give a formidable new insight in the understanding of plasma-assisted combustion and will serve a reference database for model validations.

TurEmu – The physics of (turbulent) emulsions

Project Title: TurEmu – The physics of (turbulent) emulsions

Project Leader: Prof Federico Toschi, Eindhoven University of Technology, The Netherlands

Resource Awarded

  • 66 000 000 core hours on MareNostrum 4 hosted by BSC, Spain
Details

Collaborators

  • Gianluca Di Staso, Eindhoven University of Technology, The Netherlands,
  • Roberto Benzi, Università di Roma Tor Vergata, Italy
  • Ivan Girotto, International Centre for Theoretical Physics, Italy
  • Sebastiano Fabio Schifano, Università degli Studi di Ferrara, Italy
  • Xiaowen Shan, Southern University of Science and Technology (SUSTech), China
  • Pinaki Kumar, Eindhoven University of Technology, The Netherlands
  • Arnab Ghosh, Eindhoven University of Technology, The Netherlands
  • Karun Datadien, Eindhoven University of Technology, The Netherlands
  • Andrea Scagliarini, CNR – IAC, Italy

Abstract
Multi-component fluids are extremely common in industrial as well as natural processes and, in particular, multi- component emulsions are an important ingredient of many foods and cosmetics. Emulsions are fascinating systems from the point of view of fundamental science and in partic-ular for fluid dynamic phenomenology. Ordinary substances, such as mayonnaise or ketchup, are good examples of complex fluids that can display a surprisingly rich phenomenology, typi-cal of soft-glassy materials. These complex fluids can behave either as solids (below the yield stress) or as fluids (above the yield stress). Above the yield stress, when these complex fluids flow, their rheological properties can be complex and non-Newtonian; below the yield stress, the statistical physics of localized topological rearrangements (plastic events) give rise to ava-lanches that closely resembles the behaviour of earthquakes. In this project we will employ highly optimized computational codes, based on the multicomponent Lattice Boltzmann model (LBM), to explore the physics of complex fluid emulsions: from their production, via turbulent stirring, to their (statistical) behaviour under flowing as well as resting conditions. The large produced database will allow to explore the process of turbulent emulsification in great details, including individual breakup and coalescence processes between emulsion droplets and the temporal changes in the macroscopic rheological properties of the complex fluid, as a function of the varying internal emulsion structure and of the build-up of yield stress. Once emulsions are formed, we will study their temporal dynamics both below as well as above the yield stress. Below yield our simulations will allow to explore, for the first time ever in 3d, the statistical physics of avalanches, thus paving the way to a deeper understanding of avalanche physics, including the causal-connection between different plastic events, and the relation between our model system and the physics of earthquakes aftershocks. Additionally, our study may allow to develop phenomenological macroscopic models for the behaviour of complex fluids, potentially of great relevance for the simulations of emulsification in industrial processes.

ComSUV-Combined Effects of Surfactant and Viscoelasticity on Turbulent Bubbly Flows

Project Title: ComSUV-Combined Effects of Surfactant and Viscoelasticity on Turbulent Bubbly Flows

Project Leader: Prof. Metin Muradoglu, Koc University, Mecahnical Engineering, Computational Fluid Dynamics Group, Turkey

Resource Awarded

  • 18 000 000 core hours on JUWELS hosted by GCS at FZJ, Germany
Details

Collaborators

  • Zaheer Ahmed, Koc University, Turkey
  • Outi Tammisola, KTH Mechanics, Sweden
  • Daulet Izbassarov, KTH Mechanics, Sweden

Abstract
Turbulent multiphase flows are ubiquitous in a wide range of natural processes and engineering applications. Surfactants and polymers additives are usually added to multiphase turbulent flows separately or together to manipulate the flow structure and dynamics for various purposes such as drag reduction and emulsion stability. The long chain molecules of surfactants make bulk fluid viscoelastic while drag-reducing polymer additives often act as a surfactant in multiphase flows. Despites its fundamental importance from scientific and applications point of views, the combined effects of soluble surfactant and viscoelasticity on turbulent bubbly flows have not been studied experimentally or computationally. The motive of this study is to perform extensive large-scale direct numerical simulations (DNS) of turbulent bubbly channel flows to examine combined effects of surfactant and viscoelasticity by using a fully parallelized 3D finite-difference/front-tracking method. In this method, the incompressible Navier-Stokes equations are solved using a very efficient FFT-based pressure projection method that allows for massively parallel simulations of turbulent flows. The Navier-Stokes equations are solved fully coupled with the governing equations of interfacial and bulk surfactant concentrations and the FENE-P viscoelastic model. The proposed research is expected to reveal exciting new insights and open the door for novel applications.

TJIR-ICE – Turbulent Jet Ignition Research applied to ultra lean premxied combustion in Internal Combustion Engine

Project Title: TJIR-ICE – Turbulent Jet Ignition Research applied to ultra lean premxied combustion in Internal Combustion Engine

Project Leader: Mr Frédéric RAVET, RENAULT SAS, France

Resource Awarded

  • 15 000 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France
Details

Collaborators

  • Thierry Poinsot, Institut de Mécanique des Fluides de Toulouse, France
  • Olivier Vermorel, CERFACS, France
  • Antony Misdariis, CERFACS, France
  • Quentin Malé, Renault/CERFACS, France

Abstract
While the global demand for transport energy is large and is increasing, improving Internal Combustion Engine (ICE) is of major importance because transport almost entirely relies on ICE burning petroleum-derived fuel. Lean combustion is sought in many modern designs for its low emissions and high energy efficiency (Dunn-Rankin et al., 2016). However, lower burning velocities and harder ignition prevent classical spark ignition engines to be operated in very lean regimes (Quader et al., 1976; Li et al., 2009): spark ignition of lean mixture causes erratic combustion, misfires and partial burns. Experiments (Jamrozik et al., 2013; Filho et al., 2016 ; Attard et al., 2012 ; Roethlisberger et al., 2002) show that pre-chamber ignition systems can pave the way towards lean combustion in ICE. These systems produce multiple high energy ignition sources and turbulence injection which result in a reliable ignition and a fast combustion of the main charge. Aerothermochemical processes need to be investigated and well understood in order to design this technology. Only Direct Numerical Simulations (DNS) and Large Eddy Simulations (LES) are able to simulate unsteady phenomena which are of primary importance in these systems. However, these simulation are very costly, which motivates the present proposal.

ODDC – Oceanic double diffusive convection in the diffusive regimes

Project Title: ODDC – Oceanic double diffusive convection in the diffusive regimes

Project Leader: Prof. Roberto Verzicco, Università degli studi di Roma Tor Vergata, Italy

Resource Awarded

  • 60 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Yantao Yang, Peking University, China
  • Alexander Blass, University of Twente, The Netherlands
  • Richard Stevens, University of Twente, The Netherlands
  • Francesco Sacco, Gran Sasso Science Institute, Italy
  • Francesco Viola, University of Twente, The Netherlands
  • Pieter Berghout, University of Twente, The Netherlands
  • Martin Assen, University of Twente, The Netherlands
  • Chong Ng, University of Twente, The Netherlands

Abstract
Double diffusive convection (DDC), which is the convection flow driven by both temperature and salinity stratification, widely exists and plays a key role in ocean mixing. In real ocean, DDC is inevitable in constant interaction with horizontal shearing, such as those induced by internal waves. Understanding DDC with external shearing is of great significance for evaluating the vertical heat and mass transport in ocean. Field observations and experiments are very challenging and detailed flow and scalar information is difficult to obtain. Numerical simulations, when properly carried out, have unique advantages in this context. This project will study diffusive and sheared DDC by using large scale direct numerical simulations with state-of-art numerical tools. Such simulations present huge challenges for computational techniques. Ocean DDC simulations need to deal with the salinity field which has a large Schmidt number and requires very fine meshes. Turbulence induced by shearing adds more complication to the problem. Here we will rely on our in-house massively parallel and highly efficient code AFiD (www.afid.eu), and systematic simulations will be conducted for sheared DDC flow. The nonlinear interaction between shearing and DDC, the evolution of flow morphology, and the transport properties of such flows are the key scientific questions to be investigated.

ARIATOM – Primary spray breakup modelling of prefilming AIRblast ATOMizers in aeronautical burners

Project Title: ARIATOM – Primary spray breakup modelling of prefilming AIRblast ATOMizers in aeronautical burners

Project Leader: Prof. Raul Payri, Universitat Politècnica de València, Spain

Resource Awarded

  • 35 500 000 core hours on Joliot-Curie (AMD) hosted by GENCI at CEA, France
Details

Collaborators

  • Francisco Javier Salvador, Universitat Politècnica de València, Spain
  • Oriol Lehmkuhl, Barcelona Supercomputing Center, Spain
  • Daniel Mira, Barcelona Supercomputing Center, Spain
  • José María García, Universitat Politècnica de València, Spain
  • Ricardo Novella, Universitat Politècnica de València, Spain
  • Marcos Carreres, Universitat Politècnica de València, Spain
  • José Manuel Pastor, Universitat Politècnica de València, Spain
  • Carlos Moreno, Universitat Politècnica de València, Spain
  • Lucas González, Universitat Politècnica de València, Spain

Abstract
The air traffic increase raises concerns on the impact of pollutants on public health and climate change. ACARE has defined challenging goals with 5% reduction in CO2 aircraft emissions per passenger-kilometre and a 90% reduction in NOx. Fuel injection plays a key role in this challenge. Liquid fuel is injected into a gas turbine combustion chamber through an atomizer. Great effort is placed to ensure the atomizers generate small droplets, promoting fast evaporation and homogeneous mixtures, which in turn increase combustion efficiency and reduce pollutant emissions. It is common to investigate the droplet size, distribution and kinematic properties of the dense phase. However, most experimental works can only address the secondary atomization (related to the Far-field, far from the disturbance that produces liquid breakup), whereas the majority of computational studies only provide empirical approaches to characterize primary atomization (related to the Near-field, close to the source of the breakup). HPC enables the use of sophisticated tools to provide additional information that experiments cannot offer. On the one hand, Direct Numerical Simulations may provide a significant amount of data that can be used to know the behaviour of the velocity field, the exact process related to primary breakup and finally provide a framework which may be used to correlate these two aspects. On the other hand, a phenomenological model for the primary breakup specific for airblast atomizers can be derived from the simulations. The interest of such model resides in its industrial application: due to the limitations in computational time, the usual approach in the industry is to rely on low-fidelity RANS or LES simulations coupled to a phenomenological model for the spray breakup. This kind of models exist in commercial CFD software, but they are generally based on a theoretical derivation from instability analysis (Kelvin-Helmholtz, Rayleigh-Taylor, etc.) requiring heavy calibration. Alternatively, some models for a particular application can be developed based on experimental visualization techniques, relying on the experiments to compensate for the lack of knowledge in the Near-Field region. A phenomenological model based on the high-fidelity DNS simulations linked to this project and their related understanding on the primary atomization phenomenon could address these issues for the industry, providing a reliable tool to couple to their low-fidelity CFD simulations. Thus, the methodology proposed in this project, presented in the frame of a Horizon2020-CleanSky2 project (H2020-CS2-CFP07-2017-02 GA 821418 “ESTiMatE”), pursues a double purpose: both theoretical and applied. In any case, the theoretical knowledge derived is not restricted to the gas turbine sector, but could rather be applied to any other multiphase flow.

HIFiTurb – T161 – full span computations of the MTU T161 low pressure turbine cascade using turbulent inlet boundary layers

Project Title: HIFiTurb – T161 – full span computations of the MTU T161 low pressure turbine cascade using turbulent inlet boundary layers

Project Leader: Dr Koen Hillewaert, Cenaero, Belgium

Resource Awarded

  • 42 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Michel Rasquin, Cenaero, Belgium
  • Alessandro Colombo, Università degli Studi di Bergamo, Italy
  • Thomas Toulorge, Cenaero, Belgium
  • Francesco-Carlo Massa, Università degli Studi di Bergamo, Italy

Abstract
The project aims at performing high-fidelity LES/DNS of the full-span low-pressure turbine cascade MTU T161, including the diverging end-walls. The campaign is organized under the umbrella of the HiFi-Turb H2020 European project, focused on the development of innovative turbulence models for complex aerodynamic turbulent flows. This study considers an unreleased configuration, whose experimental measurements will be made public available with the simulation data. We plan to analyse the turbulence structures related to the horse-shoe vortex, initiated by the interaction of the incoming wall boundary-layer with the blade, the separation bubble on the suction side and the subsequent transition, typical for LP turbine blades. The outcomes are the understanding of low Reynolds turbine passage flows, and the development of a high-fidelity database following a strict protocol for DNS, which will be used for improving wall-function and RANS turbulence models, and published on the ERCOFTAC Knowledge Base (KB) Wiki in a new category, elaborated during the HiFi-Turb project. Through the confrontation of different computational methods and strict budget closure checks high quality corroborated data is targeted. Previous computations indicated a high sensitivity to the inlet turbulence, and therefore the sensitivity to boundary layer injection procedures will be studied.

wakehydroLES – Characterization of the wake of an axial-flow hydrokinetic turbine via LES

Project Title: wakehydroLES – Characterization of the wake of an axial-flow hydrokinetic turbine via LES

Project Leader: Dr. Riccardo Broglia, National Research Council of Italy, Institute of Marine Engineering, Italy

Resource Awarded

  • 39 000 000 core hours on Joliot-Curie (KNL) hosted by GENCI at CEA, France
Details

Collaborators

  • Antonio Posa, National Research Council of Italy-Institute of Marine Engineering, Italy

Abstract
In this project computations on an axial-flow hydrokinetic turbine will be performed using Large-Eddy Simulation (LES), coupled with an Immersed-Boundary (IB) method, resolving the flow on cylindrical grids consisting of about two billion nodes, thus an order of magnitude larger than in the most advanced studies in the literature to date. This will allow reproducing the wake flow up to several diameters downstream, providing accurate turbulence statistics and visualization of the dynamics of hub and tip vortices. Furthermore, the same project is targeted at studying the dependence of the wake features on tip speed ratio (TSR), simulating three different values of such parameter, which is defined as the ratio between the tangential velocity at the tip of the blades and the velocity of the incoming free-stream. It is worth noting that LES is especially well suited to compute flows featuring a high level of coherence, populated by large vortices. In LES the large, energy-carrying structures are resolved explicitly, as in a Direct Numerical Simulation (DNS), with no turbulence modelling and associated approximations, typical of Reynolds-averaged Navier-Stokes (RANS) methodologies, where the time-averaged equations of the flow are instead resolved. Obviously, the accuracy of the approach improves together with the resolution of the computational mesh, since more scales are resolved, having dimensions larger than the spacing of the computational grid, while only the smallest scales are modeled using a sub-grid scales (SGS) model. These features make the LES approach both very accurate and computationally demanding, especially in the framework of engineering flows, requiring very fine resolutions in both space and time and large computational resources on massively parallel supercomputers. The IFREMER geometry, whose wake is going to be investigated, is a three-bladed axial-flow hydrokinetic turbine studied via wake measurements in the framework of Round Robin tests funded by the EU within the MaRINET Initiative under the FP7. Therefore, several experimental results are already available for validation purposes on both global performance parameters, as thrust, torque and power, and wake development. Reference results on the same turbine are also provided by the numerical solutions generated via a hybrid solver, developed at the Institute of Marine Engineering in Rome, coupling the RANS approach with a Boundary Element Method (BEM). The three computations on the IFREMER turbine will be carried out for three different values of TSR = 2.0, 3.67 and 5.0. The case at TSR = 3.67 will be considered for validation. Most data on wake development are indeed available for that value of tip speed ratio. The additional lower and higher values of the same parameter will be considered for assessing the influence by TSR on coherence (vortices shed downstream by the turbine) and turbulent statistics of the wake flow.

VaPORE – Vapour-phase transport in nanopores

Project Title: VaPORE – Vapour-phase transport in nanopores

Project Leader: Dr. Antonio Tinti, Sapienza, University of Rome, Italy

Resource Awarded

  • 25 000 000 core hours on Joliot-Curie (KNL) hosted by GENCI at CEA, France
Details

Collaborators

  • Anna Battisti, Sapienza University of Rome, Italy
  • Alberto Giacomello, Sapienza University of Rome, Italy

Abstract
Vapour-phase transport (VPT) is the phenomenon through which a mass flux is established through the vapor pockets naturally occurring within the channels of nanoporous hydrophobic membranes. The flux can be driven by either hydrostatic/osmotic pressure or temperature gradients. Recent studies ([lee2010, deshmukh2018]) have suggested how VPT through nanoporous membranes promises a breakthrough as an innovative desalination and energy harvesting technology. The nanometric size of the pores, along with the stability of the vapour bubbles makes it possible to sieve small ions, effectively allowing nanoporous materials to be used as semi-permeable membranes for forward or reverse osmosis. Access to World-class HPC can potentially play a crucial role, enabling scientists to directly simulate the molecular origin of VPT. Molecular Dynamics, empowered with rare event sampling techniques, has the potential to clarify, by accessing time and length scales which cannot be investigated by experiments, the physical mechanisms of VPT and to provide a microscopic understanding of the effect of geometrical and chemical parameters. A quantitative, simulation enabled, paradigm shift in the understanding and design of nanoporous membranes may have strong impact on the key green technologies of desalination and thermal energy harvesting, e.g., by providing design criteria from in silico molecular-scale insight.

top

Fundamental Constituents of Matter (11)

NPiTwist – The Nπ system using twisted mass fermions at the physical point

Project Title: NPiTwist – The Nπ system using twisted mass fermions at the physical point

Project Leader: Prof. Constantia Alexandrou, University of Cyprus, Cyprus

Resource Awarded

  • 35 000 000 core hours on Hawk hosted by GCS at HLRS, Germany
  • 68 000 000 core hours on Piz Daint hosted by CSCS, Switzerland
Details

Collaborators

  • Karl Jansen, DESY, Germany
  • Marcus Petschlies, University of Bonn, Germany
  • Carsten Urbach, University of Bonn, Germany
  • Giannis Koutsou, The Cyprus Institute, Cyprus
  • Kyriakos Hadjiyiannakou, The Cyprus Institute, Cyprus
  • Jacob Finkenrath, The Cyprus Institute, Cyprus
  • Davide Nole, The Cyprus Institute, Cyprus
  • Simone Bacchio, University of Cyprus, Cyprus
  • Floriano Manigrasso, University of Cyprus, Cyprus
  • Antonino Todaro, University of Cyprus, Cyprus
  • Bartosz Kostrzewa, University of Bonn, Germany
  • Roberto Frezzotti, Universita di Roma Tor Vergata, Italy

Abstract
In parallel to the high-energy experiments performed at CERN searching for Beyond the Standard Model (BSM) physics, low-energy high-intensity experiments, such as those at Jefferson Laboratory, FermiLab, and MAMI in Mainz, are providing complementary precision results that may prove equally insightful in detecting new physics. Examples belonging to the latter class are experiments on the proton charge radius and the muon g-2, which revealed puzzling discrepancies that could hint for BSM physics. For this precision frontier of particle physics, a major challenge is to determine accurately the contributions due to the strong interaction component of the Standard Model, governed by the theory of Quantum Chromodynamics (QCD). Being non-perturbative over the energies of interest, these contributions can only be accessed from first principles via large scale simulations of the theory. The goal of this three-year project is to study the nucleon sector, its resonances, and scattering processes from first principles using lattice QCD. A two-fold program is proposed, for new simulations on large volumes with physical light, strange, and charm quarks, employing the CPU system HAWK, and for their analysis using the GPUs of Piz Daint, to study nucleon structure, the Δ-resonance, and N-π scattering. Key observables that will be computed are nucleon form-factors, the Δ-width, and N-π scattering lengths. The calculation of N-π scattering lengths from the lattice is especially timely, since it addresses the tension between the value extracted from phenomenology and that inferred from lattice QCD computations of the nucleon σ-term, impacting searches for dark matter candidates. Our simulation will generate one of few ensembles world-wide at the physical point with box length over 7 fm. Beyond the hadron structure program foreseen within this project, this large volume will enable a series of high-impact projects of the Extended Twisted Mass Collaboration for its multi-year physics program, which includes flavor physics, semi-leptonic decays, nucleon structure including the neutron electric dipole moment, muon g-2, and direct evaluation of parton distribution functions.

MITCH – Multiscale kIneTiC simulations of the inner Heliosphere

Project Title: MITCH – Multiscale kIneTiC simulations of the inner Heliosphere

Project Leader: Dr. Elisabetta Boella, Lancaster University, United Kingdom

Resource Awarded

  • 16 500 000 core hours on MARCONI (Broadwell) hosted by CINECA, Italy
Details

Collaborators

  • Maria Elena Innocenti, Jet Propulsion Laboratory, USA
  • Alfredo Micera, Royal Observatory of Belgium, Belgium

Abstract
An understanding of plasma physics promises to help addressing fundamental societal challenges, from protecting technological assets from the adverse effects of space weather, to producing clean energy, to a plethora of medical applications. The solar wind and the inner heliosphere, in turns, hold the key to plasma physics: the solar wind is essentially a gigantic plasma physics laboratory just above our heads, that is currently being probed by revolutionary missions such as Parker Solar Probe, a ‘mission to touch the Sun’. The impact of solar wind observations can be greatly enhanced by complementary numerical simulations. In this project, we aim at producing numerical support to anticipate and interpret the raw data coming from space missions, with focus on the microscopic kinetic scales and on the link between those and the macroscopic bulk solar wind parameters. The computational cost of the massive fully kinetic simulations we intend to perform is reduced by the use of a state-of-the-art semi-implicit Particle-In-Cell algorithm.

Isotope Mass Impact on Tokamak Performance Improvement (IMIToPI)

Project Title: Isotope Mass Impact on Tokamak Performance Improvement (IMIToPI)

Project Leader: Dr Fabien Widmer, CNRS – Aix-Marseille University, France

Resource Awarded

  • 42 300 000 core hours on MareNostrum 4 hosted by BSC, Spain
Details

Collaborators

  • Yann Camenen, CNRS – Aix-Marseille University, France
  • Virginie Grandgirard, Commissariat à l’énergie atomique, France
  • Guilhem Dif-Pradalier, Commissariat à l’Energie Atomique-CEA/IRFM, France
  • Chantal Passeron, Commissariat à l’Energie Atomique-CEA/IRFM, France
  • Neeraj Kumar, CNRS – Aix-Marseille University, France

Abstract
International progress on magnetic fusion depends on our capability to understand and predict plasma confinement properties. Reducing or controlling the thermal loss of energy is necessary in order to optimize the thermal energy confinement time and provide CO2-free energy for civil usage at an industrial level. Actual fusion experiments run hydrogen or deuterium plasmas but the next generation reactors, such as ITER or DEMO, will use a mixture of deuterium and tritium. Therefore, knowledge of the mass scaling is needed to extrapolate experimental results from actual to future reactors. This is especially important since an improvement of magnetic confinement was experimentally observed moving from an hydrogen to a deuterium plasma. This confinement improvement is known as the isotope effect. It is not yet clearly understood and is one of the greatest challenge of turbulence transport theory. Assuming that the plasma is electrostatic, collisionless, without background flows, that the electrons are adiabatic and that transport processes are local results in a turbulent transport level proportional to the square root of the isotope mass, in contradiction with the experimental observations. Relaxing any of these assumptions can alter the mass scaling above, a key point to elucidate the isotope effect. Significant efforts in the past thirty years have been done to investigate how relaxing these assumptions modifies the isotope mass scaling of turbulent transport. Amongst other phenomena, it was very recently shown that at high plasma pressure, electromagnetic fluctuations can reverse the mass scaling of the Ion-Temperature-Gradient (ITG) instability, one of the more robust source of anomalous transport in fusion devices. Additionally, it is known that the ratio of the ion Larmor radius to the system size impact heat transport, the turbulence correlation length and avalanches, but no study of the implications on the isotope effect has been performed yet. Further investigations of these phenomena are needed before a consistent and comprehensive picture of the isotope scaling of turbulent transport could emerge. This is the goal of the present proposal. To scrutinize the isotope mass impact on both electromagnetic stabilisation and global electrostatic effects we combine two 5D state-of-the-art gyro-kinetic codes: GKW in the local gradient-driven framework, and GYSELA, utilizing a global flux-driven approach. GYSELA is one of the few codes able to model both the plasma core and edge in an experimentally relevant flux-driven framework with a kinetic electrons and ions description. Such capabilities allow to determine what is the impact of the isotope effect on the edge-core turbulence interplay and the global organisation and regulation of turbulence. Electrostatic local gradient-driven GKW simulations, based on GYSELA global simulations time averaged profiles, will be used to separate the local and global impact of the isotope effect on turbulent transport. Finally, GKW local gradient-driven electromagnetic simulations will provide a test-bed for advanced quasi-linear models attempting to capture electromagnetic stabilisation of turbulence. The rare combination of local gradient-driven and global flux-driven simulations is expected to provide valuable new insights to the puzzling isotope effect.

Fluctuations of conserved charges on the cross-over line

Project Title: Fluctuations of conserved charges on the cross-over line

Project Leader: Prof Zoltan Fodor, Bergische Universitaet Wuppertal, Germany

Resource Awarded

  • 68 000 000 core hours on Piz Daint hosted by CSCS, Switzerland
Details

Collaborators

  • Szabolcs Borsanyi, Bergische Universitaet Wuppertal, Germany
  • Jana Günther, University of Regensburg, Germany
  • Lukas Varnhorst, Bergische Universitaet Wuppertal, Germany
  • Chik Him Wong, Bergische Universitaet Wuppertal, Germany

Abstract
Quark Gluon Plasma has been studied in collider experiments for more than a decade. The possibility of producing matter at extreme conditions, thereby recreating the early moments of the Universe offers a fantastic insight into the microscopic interactions of the subatomic particles we are made of. The most interesting setting is when not only the temperature, but also the proton density is driven to extremes. It is assumed, that at certain temperature and density the system becomes critical, that is, the correlation lengths become infinite. As the created matter cools down to about 1.8 trillion K it freezes into particles that we know as bound states of quarks, such as protons and neutrons, the building blocks of nuclei. The number of protons (adding anti-protons with a negative sign) produced in the central part of the system fluctuates from collision event to collision event. These fluctuations are tell-tale signals for the temperature and density of the system when it froze. Excessive fluctuations are a hint for criticality. To correctly interpret the results, we must know the theoretical prediction for these quantities. To calculate the thermodynamic features (entropy, pressure and density) from first principles is possible only in equilibrium, but even there it is very challenging. The underlying theory, Quantum Chromodynamics, is well established. To solve this theory, large supercomputers are needed because of the complexity of the quark and gluon interactions. But even these expensive algorithms work only under the assumption of an equal number of protons and anti-protons present. This limits computations to zero net proton density. Our approach to describe the Quark Gluon Plasma uses an extrapolation technique to intermediate densities. Our calculation extends the range where the theoretical predictions are known. These will be provide the background in the search for the elusive critical point in heavy ion experiments.

Volume dependence of fluctuations of the electric charge in QCD

Project Title: Volume dependence of fluctuations of the electric charge in QCD

Project Leader: Dr. Christian Schmidt, Universität Bielefeld, Germany

Resource Awarded

  • 49 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Frithjof Karsch, Universität Bielefeld, Germany
  • Lorenzo Dini, Universität Bielefeld, Germany
  • Guido Nicotra, Universität Bielefeld, Germany
  • Swagato Mukherjee, Brookhaven National Laboratory, USA
  • Patrick Steinbrecher, Brookhaven National Laboratory, USA

Abstract
The details of the QCD phase diagram at finite temperature and density are still – to a large extent – unknown. Lattice QCD calculations provide some inside to the phase boundary between the hadron gas phase and the quark gluon plasma by a leading and next-to-leading order Taylor expansion approach. Much effort is undertaken to map out this phase boundary by heavy ion experiments at the Relativistic Heavy Ion Collider (RHIC) at Brookhaven National Laboratory and the Large Hadron Collider (LHC) at CERN. A sensitive measure for the phase boundary, as well as for possible critical behavior indicating a second order phase transition point, are event-by-event fluctuations of conserved charges. A special role is denoted to the cumulants of the electric charge fluctuations as those can be readily compared to lattice QCD calculations. We propose to study here the volume dependence of the 4th and 6th order cumulants of the electric charge fluctuations in (2+1)-flavor QCD. All previous studies have been carried out on lattice with spatial extent N_σ⁄N_τ=4, here we will go to aspect ratios of N_σ⁄N_τ=6. With this study we will be able to judge finite volume effects, which will bring us closer to our long term goal of a controlled continuum extrapolation of the 4th and 6th order cumulants.

THALMEN – ThermoHydraulics of Advanced Liquid MEtal Nuclear reactors

Project Title: THALMEN – ThermoHydraulics of Advanced Liquid MEtal Nuclear reactors

Project Leader: Dr Lilla Koloszar, von Karman Institute for Fluid Dynamics, Belgium

Resource Awarded

  • 30 000 000 core hours on Joliot-Curie (AMD) hosted by GENCI at CEA, France
Details

Collaborators

  • Silvania Lopes, von Karman Institute, Belgium
  • Alessi Giacomo, von Karman Institute, Belgium
  • Raimondo Giammanco, von Karman Institute, Belgium;
  • Steven Keijers, SCK-CEN, Belgium

Abstract
Our aim is to analyze the thermohydraulics of new liquid metal cooled nuclear reactor systems with high resolution in space and time and provide useful information to the nuclear researchers and engineers still in the design phase. The goal of the project is to perform the validation of the myrrhaFoam solver family, a customized solver group aims to simulate the thermohydraulics of liquid metal cooled nuclear reactors. These codes were developed in the framework of several years of collaboration between the von Karman Institute for Fluid Dynamics and SCK•CEN, the Belgian Nuclear Research Center. It is targeting to simulate the flow and thermal field of the primary cooling loop of the Generation IV nuclear research reactor MYRRHA (http://sckcen.be/en/Technology_future/MYRRHA) currently under design by SCK•CEN and all the background research including pre-testing of the sub-systems, concepts and prototypes. The solver is based on the OpenFOAM simulation platform, it is a finite volume based computational fluid dynamics solver including multiphase modeling, thermal equation with variable density formulation. Conjugate heat transfer is considered, as well. The direct aim of the project is to simulate transient scenarios and compared with measurements perfomed in the scaled prototype liquid metal nuclear facility E-SCAPE. The European Scaled Pool Experiment (E-SCAPE) is a thermal hydraulic 1/6th-scale model of MYRRHA, which is also cooled with LBE but employs an electrical core simulator. These types of simulations are involving large dimensions, usually in the order of few meters, with several physical phenomena involved simultaneously. The time frame of the transients are several minutes, making it very challenging to accurately resolve them both in space and time. The usual practice nowadays to perform these studies by lumped parameter system codes, though their accuracy and the information they can provide is limited. In order to target these kind of applications by computational fluid dynamics (CFD), one need to demonstrate not only that it is feasible to do the simulations, but also do it efficiently in time and with high accuracy. In order to be able to perform these simulations efficiently, we apply for the current grant. In order to ensure accuracy, we propose to start with a validation procedure. Our first targeted case is a numerical counterpart of measurements done in the scale down prototype of the MYRRHA reactor. Currently we are limited in our computational system of few hundred processors, that we would like to extend, such that we can: first, increase the transient scenarios that we try to reproduce, second, by increasing the mesh resolution and with that the accuracy of our simulation results. The final goal of our project is to simulate operational and transient cases of the future research nuclear reactor MYRRHA in order to increase the understanding of the system’s primary coolant loop, provide an insight to the momentum and thermal field of the primary coolant loop, as well as, target safety issues due to accidental scenarios.

Abelian-Higgs cosmic strings: network evolution

Project Title: Abelian-Higgs cosmic strings: network evolution

Project Leader: Dr. Carlos Martins, Centro de Astrofísica da Universidade do Porto, Portugal

Resource Awarded

  • 68 000 000 core hours on Piz Daint hosted by CSCS, Switzerland
Details

Collaborators

  • José Ricardo Correia, Centro de Astrofísica da Universidade do Porto, Portugal

Abstract
Cosmic strings are the best-known example of topological defects: fossil remanants of earlier phases of the Universe which may survive up to the present day, encoding crucial information of fundamental physics regimes that would otherwise be inaccessible. Searching for therir observational imprints is a key goal of forthcoming observational facilities, including LISA, COrE and the SKA. Being non-linear objects, the study of their dynamcs, evolution and consequences heavilty relies on HPC, complemented by analytic phenomenological modelling. Current observational searches are bottlenecked by the lack of numerical simulations with sufficient spatial resolution and dynamical range to calibrate existing analytic models. The goal of this proposal is to use the output of a large set of high-resolution GPU-based simulations with sufficiently large dynamic range (only possible with state-of-the-art high GPU resources) to calibrate and improve analytic models, enabling significantly more robust observational searchers and constraints on the underlying physics.

PULSAR PIC – Pushing Ultrafast Laser Material processing into a new regime of plasma-controlled ablation -Particle in Cell simulations

Project Title: PULSAR PIC – Pushing Ultrafast Laser Material processing into a new regime of plasma-controlled ablation -Particle in Cell simulations

Project Leader: Dr Francois Courvoisier, FEMTO-ST/ University of Franche Comte, France

Multi-year Proposal: Year 2

Resource Awarded

  • 30 000 000 core hours on MARCONI (KNL) hosted by CINECA, Italy
Details

Collaborators

  • Kazem Ardaneh, FEMTO-ST/ CNRS and University of Franche Comte, France

Abstract
PULSAR PIC aims at developing novel ultrafast laser materials processing technologies. Its context is mass fabrication (solar panels, consumer electronics, microelectronics) where new technologies are needed to process materials that are always new, on extremely large surfaces, with sub-micron resolution. Ultrafast laser processing is ideally positioned and is also much greener than conventional lithography which uses numerous harmful toxic chemicals. However, even with high pulse energy, laser processing remains limited to point by point removal of ultra-thin nanometric layers from the material surface. This is because the uncontrolled laser-generated free-electron plasma shields out the incoming light and prevents reaching extreme internal temperatures at very precise nanometric scale. PULSAR aims at breaking this barrier by developing a radically different concept of laser material ablation regime based on controlling the generation of nanoplasmas which create intense micro-explosions inside the bulk of materials. We use Particle-In-Cell (PIC) to simulate in 3D the interaction between spatially shaped femtosecond laser pulses and nanometric plasmas generated in the bulk of materials. High performance computing is therefore necessary to run the high-resolution codes. Preliminary results with relatively low resolution indicate that we have identified basic phenomena for controlling the plasmas. However, reaching our objectives require much more computational power

OptiMom – Optical angular momentum in laser-matter interactions at ultra-high intensities

Project Title: OptiMom – Optical angular momentum in laser-matter interactions at ultra-high intensities

Project Leader: Dr. Jorge Vieira, Instituto Superior Tecnico – Instituto de Plasmas e Fusão Nuclear – Grupo de Lasers e Plasmas, Portugal

Resource Awarded

  • 60 000 000 core hours on Piz Daint hosted by CSCS, Switzerland
Details

Collaborators

  • Thales Silva, Instituto Superior Técnico, Portugal
  • Mariana Moreira, Instituto Superior Técnico, Portugal
  • Ricardo Fonseca, Instituto Superior Técnico, Portugal
  • Miguel Pardal, Instituto Superior Técnico, Portugal
  • Bernardo Malaca, Instituto Superior Técnico, Portugal
  • Joana Martins, Instituto Superior Técnico, Portugal
  • Anton Helm, Instituto Superior Técnico, Portugal
  • Camilla Willim, Instituto Superior Técnico, Portugal

Abstract
Particle accelerators and light sources are pillars of modern science. The most powerful are incredibly complex, expensive, and take a long time to deploy. Material breakdown, which severely damages conventional devices, is the major factor that currently limits the accelerating fields in conventional accelerators. Plasma appears exactly when this material breakdown occurs. Thus, a concept relying on the incredibly high accelerating fields sustained in plasmas could hold the key towards miniaturized accelerators and light sources, ready for breakthrough scientific discoveries and societal applications, for example in medicine. Plasma accelerators demonstrate great promise but a key challenge remains: how to deliver the required beams for the applications. Numerical simulations are vital to pursue this major goal, as the relevant processes can only be captured by following the individual motion of many particles from first principles. These simulations are very computationally expensive, but they are crucial as no purely analytical models are currently available and experiments are also expensive. In OptiMom, we use large-scale computer simulations to explore a new direction in laser-plasma interactions, plasma accelerators and light sources: instead of focusing on the traditional drivers of laser-plasma accelerators and light sources, we propose to enhance plasma accelerators and light sources by exploring the interaction between structured lasers with matter at ultra-high intensities. This project explores some of the unknown and exciting new phenomena associated with the new dimensions opened by structured light in connection with the most recent advances joining optics, photonics and plasma physics.

ElectroPulse – Electric field and Pulse propagation in Magnetised Plasma Turbulence

Project Title: ElectroPulse – Electric field and Pulse propagation in Magnetised Plasma Turbulence

Project Leader: Dr Guilhem Dif-Pradalier, Commissariat à l’Energie Atomique, CEA/IRFM, France

Resource Awarded

  • 45 500 000 core hours on Joliot-Curie (AMD) hosted by GENCI at CEA, France
Details

Collaborators

  • Virginie Grandgirard, CEA/IRFM, France
  • Chantal Passeron, CEA/IRFM, France
  • Laure Vermare, CEA/IRFM, France
  • Philippe Ghendrih, CEA/IRFM, France
  • Yanick Sarazin, CEA/IRFM, France

Abstract
With the development of large machines like ITER, controlled fusion makes a huge step forward towards mastering the energy of the stars for civil usage. Steady international progress regarding the achieved fusion performance relies on our ability to predict the confinement properties of the plasma. Turbulent transport is the key player in this matter, modelled through the 5-dimensional gyrokinetic description, as required by the low collisionality of hot and dilute plasmas. Aspects of magnetised plasma turbulence are multiscale yet despite orders of magnitude in spatial and temporal scales between injection/dissipation and transport, the confined plasmas feel their material boundaries. This is common knowledge experimentally yet a vastly unexplored territory computationally. Predicting the interplay between spatially distinct regions of the plasma (core and edge), on disparate timescales is one of the major current issues. This proposal wishes to tackle aspects of this important problem, and doing so address two nagging problems: (i) assess whereby the radial electric field builds up and endures and (ii) probe edge-core interplay, especially through the investigation of transient dynamics, i.e. cold and hot bursts. The PI has strong expertise in the field and will be fully committed to this activity.

PULSAR – Plasma physics of ultra high fields in neutron stars

Project Title: PULSAR – Plasma physics of ultra high fields in neutron stars

Project Leader: Prof. Luis Silva, Instituto Superior Técnico, Portugal

Resource Awarded

  • 36 000 000 core hours on MareNostrum 4 hosted by BSC, Spain
Details

Collaborators

  • Marija Vranic, Instituto Superior Técnico, Portugal
  • Thomas Grismayer, Instituto Superior Técnico, Portugal
  • Kevin Schoeffler, Instituto Superior Técnico, Portugal
  • Fabio Cruz, Instituto Superior Técnico, Portugal
  • Fabrizio Del Gaudio, Instituto Superior Técnico, Portugal

Abstract
How are electron-positron pair plasmas generated at the polar cap of pulsars? How does the produced pair contribute to feeding the pulsar magnetosphere? What are the expected radiation signatures? These are prominent scientific questions where plasma astrophysics is intimately connected with quantum electrodynamics. This proposal aims to exploit the outstanding computing facilities provided by PRACE to address these compelling challenges by leveraging on the pioneered advances in coupling quantum electrodynamics effects with particle-in-cell simulations. The main scientific route to answer some of these questions consists in understanding the coupling of the pair production region with the global field organization of the pulsar magnetosphere. In particular, it is crucial to assess self-consistently the dependence of the pair multiplicity at the polar cap on the global magnetic field topology and vice versa. This task requires a special computational effort given the separation of the scales involved, namely the macroscopic size of the magnetosphere down to the plasma kinetic and quantum electrodynamics scales. The first goal of this research project is to study the production of pair plasma via cascades that occur in a localized portion of the magnetosphere around the polar cap of pulsars. The second goal is to model the global pulsar magnetosphere, leveraging on the knowledge coming from the first task. The self-consistent study of the interdependence among the plasma dynamics pertinent to such different spatial and temporal scales is difficult and massively parallel kinetic particle-in-cell simulations are critical to establishing this bridge. The unique computational infrastructures provided by PRACE would be critical to explore some of the most exciting fundamental physics questions at the forefront of science identified in this proposal. This project will leverage on the massively parallel, fully relativistic particle-in-cell code OSIRIS, which includes a set of dedicated modules to study the development of pair cascades at the polar cap of pulsars and their global magnetosphere. These are (i) a QED module for photon emission and pair creation, (ii) a merging module for withstanding the exponential increase of macro-particles in the cascades, (iii) a Compton collisional module for the accounting of direct exchange of energy between photons and electron-positron pairs, and (iv) a spherical coordinates module for modeling the global magnetosphere. This research route embraces the spirit and the path outlined by the ERC advanced grant in Pairs, which aims at understanding the dynamics of electron-positron pair plasmas in extreme fields. Many steps were already taken in this direction, and our numerical infrastructure is now ready to be massively employed in studying some of the most puzzling but marvelous questions in plasma astrophysics

top

Universe Sciences (5)

SIMPLE – Six-dimensional Ion Modelling of PLasma near Earth

Project Title: SIMPLE – Six-dimensional Ion Modelling of PLasma near Earth

Project Leader: Prof Minna Palmroth, University Of Helsinki, Finland

Resource Awarded

  • 60 000 000 core hours on Hawk hosted by GCS at HLRS, Germany
Details

Collaborators

Abstract
Space weather refers to the potentially harmful environmental conditions within the near Earth space, driven by variations in the solar wind and affected by the dynamics of the Earth’s magnetic domain, the magnetosphere. Among the most unpredictable space weather phenomena are substorms, periods of magnetospheric energy loading and unloading, which can lead to spacecraft failure. Substorms have been investigated from 1960s with tens of spacecraft costing thousands of millions, however, even constellations have failed to understand the cause of substorms due to the inherently local nature of measurements. Decades-long debate exists between two competing models for the substorm onset, caused by either an instability-driven current disruption, or magnetic reconnection. This project carries out the world’s first 6-dimensional ion-kinetic global magnetospheric simulation, accurate both locally and globally. We will investigate unambiguously for the first time the timing of magnetospheric tail reconnection relative to the instability development. Our results help to understand the unpredictable substorms, with technological and societal impacts in interpreting spacecraft measurements, and devising new missions and instruments. A growing demand exists to understand spacecraft environments as the utilisation of space is in rapid increase, with markets worth of billions. Our efforts are a significant contribution to these ends. This project proposes to investigate substorm physics globally in a 6-dimensional ion-scale simulation, Vlasiator. Vlasiator is global, i.e., able to cover the entire solar wind – magnetosphere system spatially in 3 dimensions (3D) from dayside to the tail self-consistently. Vlasiator further solves the ion velocity space in 3D, providing a plasma description that includes the required ion-scale kinetic plasma physics. Altogether, this approach is called 6D global ion-scale simulation. The following science questions are the objectives of this proposal: 1. Which process, magnetic reconnection or a plasma instability is responsible for the substorm onset? 2. Is the process that initiates a substorm onset itself triggered, or does it occur spontaneously? Vlasiator has already assessed the substorm physics in 5D, that is, including a 2D spatial simulation space, and a 3D velocity space. Even this simulation required supercomputer resources and was carried out using PRACE Tier-0 grant at the HLRS, Stuttgart. Previous research found that reconnection is sufficient but not perhaps necessary for the substorm to develop, and that a 6D simulation is necessary to investigate the role of instabilities in disrupting the tail, and thus arrive to a conclusion about which process is behind the substorm onset. This project is directly related to two European Research Council awards, as Vlasiator was developed with a Starting grant (2007), and further used and developed to investigate long-standing space scientific questions with a Consolidator grant (2015).

COLIBRE – Simulating the evolution of galaxy populations at high resolution

Project Title: COLIBRE – Simulating the evolution of galaxy populations at high resolution

Project Leader: Prof. Joop Schaye, Leiden University, The Netherlands

Resource Awarded

  • 50 000 000 core hours on Hawk hosted by GCS at HLRS, Germany
Details

Collaborators

  • Matthieu Schaller, Leiden University, The Netherlands
  • Sylvia Ploeckinger, Leiden University, The Netherlands
  • Folkert Nobels, Leiden University, The Netherlands
  • Alejandro Benitez Llambay, Durham University, United Kingdom
  • Alexander Richings, Durham University, United Kingdom
  • Camila Correa, Leiden University, The Netherlands
  • Claudio Dalla Vecchia, Instituto de Astrofísica de Canarias, Spain
  • Simon White, Max Planck Institute for Astrophysics, Germany
  • Carlos Frenk, Durham University, United Kingdom

Abstract
Starting from small initial density perturbations in the early Universe, gravitational instability produces a wealth of emergent structure. This ranges from the formation of starbursting galaxies that are observed to drive supersonic shock waves into intergalactic space, to the assembly of spiral and elliptical galaxies that are thought to be embedded in a cosmic web of hot, intergalactic gas and to host supermassive black holes at their centres. Our understanding of this rich cosmic history is limited by the complexity of the non-linear physics governing the baryonic component (i.e. ordinary as opposed to dark matter). Cosmological hydrodynamical simulations, in which the initial conditions that are determined by the physics of the early Universe are evolved to late cosmic epochs, have therefore emerged as the leading technique for i) confronting theoretical models for cosmic evolution with observational data and ii) gaining insight into the relative importance of the physical processes germane to the formation and evolution of galaxies. Indeed, thanks to recent breakthroughs in the realism of their predictions, hydrodynamical simulations of galaxy formation have become critical for the interpretation of observations. Our previous Prace simulation EAGLE illustrates this success, already having resulted in 200+ refereed papers. However, current models suffer from two important limitations. First, simulations of representative volumes lack the resolution and physics to model cold, interstellar gas. This limitation implies that galaxy disks are too thick, star formation is insufficiently clustered, and observational diagnostics are sensitive to unresolved structure. Second, cosmological simulations suffer from spurious transfer of energy from dark matter to stars due to the use of more massive particles to model the former. We propose to address these limitations by a) simulating a volume large enough to include 100 galaxies with mass similar to that of the Milky Way, but small enough to increase the spatial (mass) resolution by 1 (2-3) orders of magnitude relative to EAGLE; b) using similar mass particles for dark matter and baryons. The proposed COLIBRE simulation is enabled by our new code SWIFT, which is an order of magnitude faster than the code used for EAGLE and includes new models for cold gas and star clusters. Even with SWIFT, COLIBRE is a demanding calculation requiring 50M core hours and 20TB of memory on a Tier-0 facility.

Nyx Hydrodynamical simulations for inference on DESI Ly-a forest data

Project Title: Nyx Hydrodynamical simulations for inference on DESI Ly-a forest data

Project Leader: Dr. Michael Walther, Commissariat à l’Énergie Atomique (CEA) Saclay, France

Resource Awarded

  • 30 000 000 core hours on Joliot-Curie (AMD) hosted by GENCI at CEA, France
Details

Collaborators

  • Eric Armengaud, Commissariat à l’Énergie Atomique, France
  • Nathalie Palanque-Delabrouille, Commissariat à l’Énergie Atomique, France
  • Zarija Lukic, Lawrence Berkeley National Laboratory, USA

Abstract
The Lyman alpha forest has proven to be a valuable tool for cosmology, providing a unique data set to constrain fundamental, yet poorly understood, properties of the Universe such as the epoch of inflation, dark matter or neutrinos. The interpretation of the data, however, requires grids of extensive hydrodynamical simulations that simultaneously probe small and large scales. Upcoming datasets, e.g. from the Dark Energy Spectroscopic Instrument (DESI), will improve observational precision to the 1% level demanding similarly accurate modelling. However, to obtain this accuracy from a single hydrodynamical simulation requires 60 billion fluid elements which has not been possible with existing simulation codes for a large number of input parameters. But recent breakthroughs in computational performance of hydrodynamical codes, e.g. in the proposed Nyx code, reduce the computational time needed per simulation, making grids of percent-level accuracy feasible within a typical PRACE allocation. In addition, new interpolation techniques allow running relatively sparse grids of input parameters without losing accuracy. The proposed project will enable us to run the largest grid of hydrodynamical simulations for the Lyman-alpha forest ever obtained. Combined with early data from the DESI survey and existing measurements of the cosmic microwave background from the Planck satellite, we will use this grid mainly to tighten constraints on inflation. But we will also use it as a starting point for future studies pinpointing the mass of neutrinos and determine the nature of dark matter which require additional simulations. Finally, we intend to allow usage of the simulation outputs by the general community which could induce exciting results in related fields, e.g. on reionization or tomographic mappings of the universe.

IMPACT – Retracing the formation of the Moon in the aftermath of the Giant Impact

Project Title: IMPACT – Retracing the formation of the Moon in the aftermath of the Giant Impact

Project Leader: Dr. Razvan Caracas, CNRS, France

Resource Awarded

  • 30 000 000 core hours on Joliot-Curie (AMD) hosted by GENCI at CEA, France
Details

Collaborators

  • Zhi Li, CNRS, France
  • Natalia Solomatova, CNRS, France
  • Renata Schuler Schaan, CNRS, France
  • Francois Soubiran, CNRS, France
  • Anais Kobsch, CNRS, France

Abstract
We aim at characterizing the evolution of the protolunar disk generated in the aftermath of the Giant Impact from its formation until its condensation. This project was awarded an ERC Consolidator Grant, IMPACT, for the duration 2016-2021. Very little is understood of the physics governing the Giant Impact and the subsequent formation of the Moon. According to this model an impactor hit the proto-Earth; the resulting energy was enough to melt and partially vaporize the two bodies generating a large protolunar disk, from which the Earth-Moon couple formed. Hydrodynamic simulations of the impact and the disk are currently based on unconstrained models of equations of state and phase diagrams: estimates of the positions of critical points for realistic geological materials, when available at all, vary by one order of magnitude in both temperature and density. Here we use large-scale ab initio molecular dynamics to determine vaporization curves, position the supercritical points, and characterize the sub-critical and supercritical regimes for the major rock-forming minerals. We use these results to simulate the thermal profile through the disk, the ratio between liquid and vapor, and the chemical speciation. Eventually we constrain the impactor, the proto-Earth and the plausible impact scenarios.

AIfor21CM – Artificial Intelligence for 21-cm Cosmology

Project Title: AIfor21CM – Artificial Intelligence for 21-cm Cosmology

Project Leader: Prof. Andrei Mesinger, Scuola Normale Superiore, Italy

Resource Awarded

  • 20 000 000 core hours on Piz Daint hosted by CSCS, Switzerland
Details

Collaborators

  • Nicolas Gillet, Scuola Normale Superiore, Italy
  • Brad Greig, University of Melbourne, Australia
  • Jaehong Park, Scuola Normale Superiore, Italy
  • Yuxiang Qin, Scuola Normale Superiore, Italy
  • David Prelogovic, Scuola Normale Superiore, Italy

Abstract
Our project aims to optimize the analysis of upcoming 21-cm images of the epoch of reionization and the cosmic dawn, expected from the Square Kilometre Array (SKA) telescope. This cosmic signal, which encodes the sought-after properties of the unseen first galaxies, is highly non-Gaussian. The common approach of compressing the images into a power spectrum (PS) summary statistic wastes potentially valuable information. To extract as much information as possible from the signal, we made use of Convolutional Neutral Networks (CNNs). CNNs are especially useful for this purpose because they can adaptively select the optimal summary statistic that maximizes their ability to recover astrophysics. However, our original work was severely limited by computational resources. Network tuning was done “by hand” using only a few configurations, since each training of the CNN took days on our local CPU cluster. Here we wish to optimize the performance of CNNs at astrophysical parameter recovery from 21-cm images, by using the efficiency of GPU clusters. Since the parameter space of adjustable hyper-parameters (governing the network architecture) is enormous, running an automatic optimization requires hundreds of thousands of CNN trainings; this can only be done on a tier-0 GPU cluster like PizDaint. The end goal of this research is an optimized artificial neural network trained to infer the properties of the unseen first galaxies from realistic 21-cm images of reionization and the cosmic dawn. Having such a tool will allow us to understand upcoming results from the SKA, thus maximizing Europe’s significant investment in the SKA.

top