Skip to main content »

Trinity College Dublin

Ongoing Research Projects supported by TCHPC

TCHPC allocates resources to assist research in many different fields. Below is a list of current projects been undertaken with the help of TCHPC.
  • Engineering
    Compute
    Gareth Bennett
    Trinity College Dublin
    I am a first year PhD research student in the department of Mechanical & Manufacturing Engineering, working in the field of Acoustics under the supervision and guidance of Dr. Gareth Bennett. My research is being conducted in conjunction with TEENI (Turboshaft Engine Exhaust Noise Identification), a 7th framework EU program aiming to identify noise sources in turboshaft engines for applications in helicopter engine design. The funding for my PhD research is provided by this. My research so far has involved working with DLR (German Aerospace Agency), one of the consortium partners, as they are running tests on a small-scale experimental rig which simulates the acoustic stages of a turboshaft engine. My work will involve analyzing the data from these tests and applying novel noise source identification techniques based on advanced coherence function based numerical techniques and modal decomposition. The Need for High Performance Computing The data from these tests has been acquired over 256 channels. Up to 64 channels of microphone data needs to be loaded into Matlab for modal decomposition to be applied, and each channel consists of 1200000 data points of 32-bit floating integers. As such the memory requirements are considerable. Furthermore, the coherence function based techniques used require these channels to be modally decomposed for each averaging block, of which there are 146, and running such a code would take a standard desktop PC around a day to run (estimate based on initial tests with a simplified version of the actual code used).
  • Life Sciences
    Storage
    C
    Declan McLoughlin
    Trinity College Dublin
    Our team consists of psychiatrists, nurses, psychologists and basic scientists. We use data from volunteer patients and animal research, along with meta-analytic methods, to examine brain changes associated with severe depressive illness and its treatment. We are currently performing a randomised controlled double-blind trial of standard bilateral ECT versus high-dose unilateral ECT to test the hypothesis that the latter will have the same clinical efficacy but less of the cognitive side-effects. Our research also involves the systematic search for molecular biomarkers of depression. Such biological markers will enable us to diagnose and better manage depression and will also be informative about the molecular neurobiology of depression. Together, these studies involve collection and formation of large databases that are being stored and back-up on the TCHPC storage facility.
  • Physics
    Compute
    Dr Claude Ederer
    Trinity College Dublin
    Materials that combine two or more functional properties, such as e.g. conductivity, optical refraction, capacitance, piezo-/ferroelectricity, or magnetism, offer great potential for highly integrated microelectronic devices. In particular, materials where magnetic and dielectric (or ferroelectric) properties coexist and are coupled offer the intriguing possibility to manipulate magnetic properties via electric fields or vice versa. This project aims at developing a good physical understanding of "magneto-electric" and "multi-ferroic" properties of complex oxides, both in bulk form and in nano-composites. To this effect, we are using computational techniques based on first principles density functional theory, which facilitates an accurate calculation of structural, electronic, and magnetic properties of specific materials based solely on fundamental quantum mechanical principles. These calculations allow us to analyze the dominant mechanisms determining the desired functionality, and to predict the properties of new materials prior to their experimental synthesis. The overall goal is to identify materials with optimal properties for future multifunctional devices. In addition, we are exploring novel methods for the quantitative prediction of functional properties from first principles, in order to extend the accuracy and reliability of currently used methods to a wider class of materials.
  • Physics
    Compute
    A
    Prof. Stefano Sanvito
    Trinity College Dublin
    The project is to design, analyze and approve magnetic devices based on spin-torque transfer mechanism. The theoretical/computational part is mainly based on Ab-initio Density Functional Theory and non-equilibrium Green's functions method, as implemented in Smeagol code. Some further development of the code will be also necessary. Due to the very high complexity of the method and the code, the project requires considerable computational resources. This code is parallelized and well suited for the parallel machines like Lonsdale.
  • Physics
    Visualisation Support
    B
    Dr. Peter Gallagher
    Trinity College Dublin
    Solar coronal mass ejections (CMEs) are the most significant drivers of adverse space weather on Earth, but the physics governing their propagation through the heliosphere is not well understood. Although stereoscopic imaging of CMEs with NASA's Solar Terrestrial Relations Observatory (STEREO) has provided some insight into their three-dimensional (3D) propagation, the mechanisms governing their evolution remain unclear because of difficulties in reconstructing their true 3D structure. In this paper, we use a new elliptical tie-pointing technique to reconstruct a full CME front in 3D, enabling us to quantify its deflected trajectory from high latitudes along the ecliptic, and measure its increasing angular width and propagation from 2 to 46 (~0.2 AU). Beyond 7 , we show that its motion is determined by an aerodynamic drag in the solar wind and, using our reconstruction as input for a 3D magnetohydrodynamic simulation, we determine an accurate arrival time at the Lagrangian L1 point near Earth.
  • Life Sciences
    Storage
    Compute
    C
    Prof. John O'Docherty
    Trinity College Dublin
    In this study we aim to test for the existence of learning signals underlying reward-learningthrough observation. Such signals would be expected to modify neural plasticity inregions of the brain involved in encoding stimulus-reward and stimulus-responserewardassociations, in order that associations learned through observation can beused to actively guide choice behaviour in future. We hypothesize that regions ofstriatum found to exhibit PE signaling during experiental learning, may also showsuch signals during OL. In addition, we will test for the presence of observational PEsin areas that do not exhibit simple experiental PE signaling, but that are specificallyinvolved in social cognitive processes. Given our previous results, a prime candidatearea for exhibiting such signals might be posterior STS.

Last updated 08 Jul 2014Contact TCHPC: info | support.