Examples of the problems and topics on which ABACUS 2.0 is playing a key role
RNA 3D structures
Principal investigator: Assistant Professor Jing Qin
Faculty of Science - Department of Mathematics and Computer Science (IMADA) - SDU
In this project we study the three-dimensional folding structure of RNA molecules. Understanding folding of RNA is important as wrongly folded molecules could lead to serious diseases. We use a computer simulation method called “molecular dynamics” for studying the physical movements of atoms and molecules. On ABACUS 2.0 molecules are represented as 3-dimensional geometrical data and a single simulation of one experiment requires about 7000 simulation steps. This method requires a tremendous amount of computer simulations and it can take many years on a normal computer to complete only one experiment.
Meeting current challenges in computational spectroscopy on a supercomputer
Principal investigator: Professor Jacob Kongsted
Faculty of Science - Department of Physics, Chemistry and Pharmacy (FKF) - SDU
We develop and apply quantum chemical methods to study biological systems. We are especially interested in defining rational design strategies for development of novel functional optical biological materials relevant for example for light-harvesting and biological sensing. For our research, use of supercomputers like ABACUS 2.0 is of utmost importance – such computing facilities represent our laboratory and this is where we do our experiments using the computational strategies we develop.
Membrane Transport and Shape Remodeling Using True Multi-scale Simulations
Principal investigator: Associate Professor Himanshu Khandelia
Faculty of Science - Department of Physics, Chemistry and Pharmacy (FKF) - SDU
We are interested in the molecular and biophysical processes occurring in proteins and membranes. We use highly parallel numerical integrators to study the self-assembly and dynamic behaviour of biological complexes that drive life. Such highly parallel computations are only possible on large scale HPC resources, because we study the dynamics of millions of particles, which cannot be performed on desktop computers. Using our computations, we have discovered new transport pathways across cellular membranes, have examined the dynamics and mechanical properties of the Zika virus, and describe the molecular processes involved in the formation of organelles that control fat metabolism.
Understanding the fundamental structure of the Universe
Principal investigator: Prof Claudio Pica
Faculty of Science - CP3-origins - Department of Mathematics and Computer Science (IMADA) - SDU
Researchers at the CP3-Origins centre of excellence use the ABACUS 2.0 supercomputer to study new theoretical models to explain how the Universe is made at its most fundamental level. Supercomputers help researchers where experiments and observation reach their limits.
With the Nobel-prize discovery of the Higgs boson in 2012 after decades of searches, the last missing piece of the Standard Model of particle physics was put in place. Nonetheless many are the mysteries that the Standard Model cannot explain: why is the Higgs so light? why is there more matter than anti-matter? what is dark matter and dark energy?
At the CP3-Origins centre for Cosmology and Particle Physics Phenomenology, researchers are exploring the idea that the Higgs particle is made up of smaller pieces, i.e. it is a composite particle that behaves almost identically to the Standard Model Higgs boson. The existence of these new smaller pieces could explain why the Higgs is so light and at the same time they could be re-arranged to form dark matter particles not present in the Standard Model.
Understanding the dynamics of these composite particles is no easy task, as the theoretical models are so complex that no exact solutions exists. To get an accurate picture, researchers at CP3-Origins rely on the use of state-of-the-art supercomputers, such as ABACUS 2.0.
Statistical Mechanics, Theoretical and Computational Physics of Soft-Condensed Matter
Principal investigator: Postdoc. Federica Lo Verso
Faculty of Science – Department of Physics, Chemistry and Pharmacy (FKF) – SDU
My research focuses on multi-scale modelling of complex systems with different architectures and microscopic interactions under various geometries and environmental conditions, in order to target new material properties and performance. The systems at hand are characterised by different length scales, and the system sampling usually involve a large number of particles (monomer/bead units) and of run necessary in order to get good statistics.
I was involved in three different projects running on ABACUS 2.0.
My main project involved the computational synthesis of microgels, obtained via inter-molecular cross-linking of polymers. We accurately characterised the morphology and structure as well as the kinetics of the cross-linking process. We also studied the swelling and deswelling behaviour.
One project-student I supervised performed simulation with LAMMPS in order to explore the mechanisms behind the anomalous diffusion of particles in a membrane from a numerical point of view in order to test theoretical previsions. This project was only preliminary, and gave insight in order to improve the properties of the model we used.
The second project-student I supervised simulated polymers in crowded environments. The student used an idealized model, and analysed the effects of crowding on polymers, characterising the response in terms of the dynamics, through the Mean Squared Displacement, and in terms of polymer structure via the calculation of asphericity, end-to-end distance and polymer pro lateness.
Epigenetic biomarkers as predictors of late-life mortality in the Elderly
Principal investigator: Professor Qihua Tan
Faculty of Health Science - Epidemiology, Biostatistics and Biodemography - Human Genetics - SDU
The research is about genetic epidemiology and medical bioinformatics on human complex diseases, aging and development. HPC provides an irreplaceable platform for increasing computational load and demanding data storage capacity big data in medical genomics. The data run on ABACUS 2.0 is Genome-wide SNPs genotype data and whole genome DNA methylation sequencing data.
We performed a genome-wide association study on cognition and an article have been published. The methylation sequencing data is still running.
Optimization of Deformable Object Manipulation Through Simulation
Principal investigator: Professor Norbert Krüger
Faculty of Engineering - The Maersk Mc-Kinney Moller Institute - SDU Robotics - SDU
In this project, we study simulation and optimization for robotic manipulation of deformable objects (e.g. meat). Simulation of deformable objects requires a substantial amount of computation, and when optimization is placed on top of this hundreds of deformable object simulations are required. Furthermore, to analyze the optimization several optimization runs also have to be evaluated.
Robotic task tends to contain some control parameters, some objective and some uncertainties. Numeric optimization is a powerful tool to pick a set of control parameters that maximize the objective score. But to ensure the uncertainties doesn't affect the solutions some extra steps must be made. When considering the importance of minimizing the impact of uncertainties, function fitting optimization techniques in particular “RBFopt” are powerful optimization tools.
Monitoring of invasive species using drones
Principal investigator: Assistant Professor Henrik Skov Midtiby
Faculty of Engineering - Mærsk Mc-Kinney Møller Institute - SDU Dronecenter - SDU
We work on recognizing objects of interest in images from natural environments. A case is to detect the presence of giant hogweed plants in an image. We use convolutional neural networks (CNNs) to detect these plants. The use of CNN requires access to suitable computing resources, especially graphics processing units GPU. We now have a CNN that reliably can detect giant hogweed in images, the next step is then to limit the number of false positive detections by the CNN.
THERMCYC – Advanced Thermodynamic Cycles
Principal investigator: Associate Professor Kim Sørensen
Faculty of Engineering - Department of Energy Technology - AAU
We research the adhesive nature of small micron-sized particles in a turbulent flow. Due to the small size of the particles, inter-molecular forces such as van der Waals attraction becomes important. Due to the complex interaction between the adhesive particles and a turbulent flow, experimental studies are almost impossible to control to a satisfactory degree.
In our work, we therefore rely on numerical simulations using Large Eddy Simulations (LES) to the Discrete Element Method (DEM) to simulate how a high number of particles agglomerate in a turbulent flow. As particles collide over time intervals much smaller than the overall agglomeration process, these simulations tend to be computational expensive. Based on preliminary results published in Powder Technology , we have used ABACUS 2.0 to parametrically investigate how particle and fluid properties affect particle agglomeration.
As the computational time scales almost linearly with number of cores for our simulations, we have been able to run simulations that would not have been possible without ABACUS 2.0.
Data Analytics on Occupancy Data
Principal investigator: Associate Professor Mikkel Baun Kjærgaard
Faculty of Engineering - The Mærsk Mc-Kinney Møller Institute - Center for Energy Informatics - SDU
My research is about Energy Informatics, Ubiquitous Computing and Data Science.HPC is important for my research because it allows us to scale our analysis to larger datasets and to explore the impact of more parameters. The data, I have run on ABACUS 2.0 is sensing data on the presence of people in buildings and associated energy data.
We have developed an algorithm to process sensing data about presence of people in buildings to estimate occupation properties. In our evaluations we have demonstrated that it can process sensor data to accurately estimate occupation properties. The estimated data can among others be used to analyse the energy consumption in buildings in relation to occupation patterns.
Light-matter interaction at the nanoscale
Principal investigator: Professor Sergey I. Bozhevolnyi
Faculty of Engineering - The Mads Clausen Institute - SDU NanoOptics - SDU
At SDU Nano Optics, we utilize the strong interaction of light with nanostructured components to conduct research within the fields of quantum plasmonics and optical metasurfaces, where the accurate description of the systems requires a simultaneous knowledge of the distribution of the electromagnetic field at both the nano- and micrometer scale, which makes it a computationally demanding task. The calculations performed on ABACUS 2.0 are all depending on the finite element approach (FEA) implemented in the commercial software Comsol Multiphysics. Based on the simulation results, we have demonstrated various optical functionalities, such as focusing, beam deflection, integrated single-photon source and surface plasmon polariton couplers.
Fault Detection and Prediction in Offshore Wind Turbines
Principal investigator: Associate Professor Esmaeil S. Nadimi
Faculty of Engineering - Maersk Mc-Kinney Moller Institute - SDU Embodied Systems for Robotics and Learning- SDU
The research efforts we address with ABACUS 2.0 are related to data volume. The variety of sensor installed in wind turbines offer the unique opportunity to use large scale data to predict events altering the wind turbines’ performance. The problem at hand is the limited abilities of handling those data by local workstations. ABACUS 2.0 offers the platform to handle large scale data and perform analysis within reasonable time compared to human efforts. With ABACUS 2.0 we are turning various data types (times-series, logs, and other sensory data) into valuable insight, such as the remaining lifetime of turbines. Recent results have shown that learning from large scale wind turbine data can provide prediction horizons of several month prior to a failure of a wind turbine. In the future we want to utilize the parallel computing power to decrease the dimension of wind turbine sensor data into more informative data series. Ultimately obtaining knowledge about the underlying root cause for failures in wind turbines.
Mathematical Modeling of ultrasound propagation in multi-phase flow
Principal investigator: Associate Professor Jost Adam
Faculty of Engineering - The Mads Clausen Institute - SDU NanoSYD - SDU
Ultrasonic flow meters are used in a wide range of applications. This project specifically focuses on the particular problem of measuring flow in multi-phase flow condition, where the flow media consists of two or more substances. One industrially relevant example is the so-called bubbly flow scenario, where the flowing media is a gas/liquid mixture. In state-of-the art flow meters, this inhomogenity significantly affects the measuring accuracy, since ultrasonic signals get scattered from the second phase, adding distortions to the received signal. In practice, this leads to highly error-prone, unreliable measurement results. In this project in collaboration with Siemens Flow Instruments, we develop a numerical model that helps better understand and predict measurements in this screnario, with the ultimate goal of overcoming the described problem.
As it is nearly impossible to establish a measurement setup, where multiphase flow conditions would be sufficiently controlled to compare a single measurement to single simulation, we have to introduce a statistical modelling approach. Series of simulations hence need to performed for various phase distributions, flow speeds and flow profiles, in order to look for common received signal signatures. To this end, we implement our model in Matlab and run this vast number of necessary simulations in a highly parallel fashion on the ABACUS 2.0 supercomputer.
Evolutionary Robotics and Embodied Cognition
Principal investigator: Assistant Professor Andrés Faíña
Robotics, Evolution, and Art Lab- Department of Computer Science - IT University of Copenhagen
We work on designing robotic morphologies and controllers for locomotion tasks by using evolutionary algorithms. Specifically, we employ modular robots, simple and autonomous devices that can be assembled together to achieve a functional robot. In ABACUS 2.0, we run physical simulations to evaluate the suitability of a robotic morphology and a controller. As these simulations are very time consuming, they are carried out in parallel in a cluster. Our results show that certain module´s geometries and generative encodings can greatly speed up the optimization process. Additionally, the best robots can be assembled and tested in reality, which give us a measurement of their performance. Comparing real and simulated results, we can shed light on how to reduce the reality gap.