UCY-COMPSCI · Transfer of knowledge for the creation of a multidisciplinary core of excellence in the computational sciences at the University of Cyprus (UCY)

Final Activity Report Summary - UCY-COMPSCI (Transfer of knowledge for the creation of a multidisciplinary core of excellence in the computational sciences at the University of Cyprus (UCY))

To achieve the research and training goals of the project, one of the first completed tasks was the upgrade of a high performance computing (HPC) system, through doubling its computational power, and the creation of a second HPC system for training in parallel computing. Turbulent fluid flow is one of the great challenges to science and engineering because it contains eddies, representing seemingly chaotic zigzagging or swirling motion. Because eddies extend over many orders of magnitude in size, faithfully representing them in turbulence simulations by far surpasses the currently available computational power. A notable achievement was the transition of the algebraic structure-based model (ASBM) from academic development to industrial adoption. The ASBM fell in the category of engineering approaches called Reynolds averaged Navier-Stokes (RANS) models, where the effect of the entire range of eddy sizes was modelled instead of computed. The ASBM offered promising results over current RANS models, since it retained key information regarding the morphology of modelled eddies. An important collaboration was established between the Centre for Computational Sciance of the University of Cyprus (UCY-CompSci), the Office National d' Etudes Aerospatiales (ONERA), the Norwegian defence research establishment (FFI) and Airbus, which led to the evaluation of the ASBM in realistic transonic flows over real airfoils. The main conclusion was that the ASBM, coupled with the 2008 version of the v2-f model, performed better than the leading engineering model used by the aerospace industry. Another important achievement was the development of a new methodology so that the immersed boundary (IB) simulation method, which had previously been developed for hydrodynamic flows, could be also applied on contacting flows of liquid metals. Another addressed area was 'multi-scale modelling and simulation'. In this area, UCY-CopSci developed the capability to carry out large scale simulations of fluids at the nanoscale using a range of approaches, such as molecular dynamics, ab initio MD, Monte Carlo and lattice gas models. A promising use of carbon nanotubes (CNTs) was in the development of sensors for pollutants dispersed in water and for the absorption of carbon dioxide. Through a series of large scale simulations, UCY-CompSci scientists analysed the interaction of water and carbon dioxide with CNTs. The quality of this work led to the publication of an important paper in the Chemical Reviews, one of the most prestigious journals in the field. UCY-CompSci also developed the capability to carry out large scale simulations of aerosol dispersion in atmospheric flows, for example dispersion around obstacles representing buildings. Moreover, many efforts were directed towards developing the capability to simulate aerosol deposition in the airways of the human respiratory system. This was an important achievement for both optimising inhaled drug delivery and understanding the effect of airborne pollution on disease. When we inhale, the flow quickly transitions to turbulence as it passes through the pharynx. The details of this transition and the flow downstream depend greatly on the individual geometry and vary significantly among individuals. In the past, most of the simulations focussed either on the upper or the lower airways, where the flow was either turbulent or laminar, in order to avoid the complexities of the transitional flow in-between. Through the use of state of the art simulation technology, UCY-CompSci scientists were able to simulate the entire flow from the mouth to the smaller bronchi, capturing the details of the flow transition that took place in the laryngopharynx. This provided great promise for our ability to predict with much better accuracy aerosol deposition in the human respiratory system. The central theme linking the various scientific activities was large scale parallel computing. A number of scientific codes were revamped to take advantage of the latest developments in parallel computing on mutli-core systems. A notable achievement was the development, in collaboration with one of the senior incoming fellows, of a solver with a gridless representation of the Navier-Stokes equations. This approach rendered computational technology more easily accessible to non-specialised scientists and engineers.

Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз