CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Jobs > Job Record #15559

CFD Jobs Database - Job Record #15559

Job Record #15559
TitleMassively parallel adaptive LBM simulation for wind turbines
CategoryPhD Studentship
EmployerUniversity of Southampton
LocationUnited Kingdom, Southampton
InternationalYes, international applications are welcome
Closure DateFriday, February 01, 2019
Description:
Expressions of interest are sought from EU students (no UK citizens) to put in a bid with me for a fully funded PhD scholarship under the University of Southampton Presidential Scholarship scheme, sub-category "Best of EU students", results expected to be available by end of February. A first class MEng or MSc degree and citizenship in an EU country other than the UK are mandatory for this scheme.

With a fully funded scholarship you will have considerably freedom in your research within a group that is focussed on CFD development with modern Cartesian methods. In this project you should concentrate on developing modern parallel adaptive lattice Boltzmann methods and their application for simulating the turbulent flow fields created by full-scale wind turbines and related open rotor laboratory experiments within our software framework AMROC. An understanding of the large-scale wake structures generated by operating horizontal axis wind turbines is vital for optimizing wind farm layouts. However, the flow over turbine blades is generally not Reynolds number independent in the velocity ranges of interest for wind turbines, which makes it difficult to draw reliable conclusions from small-scale model experiments.

Numerical simulation of full-scale turbines is a promising avenue, but the difficulties in solving the incompressible or weakly-compressible Navier-Stokes equations on moving three-dimensional meshes are enormous. As an alternative to conventional CFD solvers for this problem class, a novel parallel and dynamically adaptive lattice Boltzmann method for large eddy simulation of turbulent weakly compressible flows with embedded moving structures is currently under development based on the AMROC framework. Using a Smagorinsky-type large eddy turbulence model, our present implementation is already able to predict dynamic loads on a full-scale wind turbine rotor including rotor-tower interaction phenomena within a few percent of manufacturer's specification, while downstream wake structures are exceptionally well preserved.

The advertised position could concentrate particularly on improving the parallel performance of the software and simulating well-documented laboratory experiments for related turbomachinery with high accuracy thereby providing unambiguous method validation. At present, our C++ adaptive mesh refinement system uses a rigorous domain decomposition strategy for dynamic load balancing, and some generalization of this methodology and the hierarchical mesh data structures could be attempted. An extension of the algorithms to hybrid MPI-OpenMP, possibly MPI-OpenACC, communication would allow scaling to several thousand cores. The improved performance could be demonstrated with massively parallel high-resolution simulations of the turbulent wake structures generated by the 4.5m diameter rotor used in the Mexnext experimental campaigns. These laboratory experiments achieve tip speed ratios like full-scale wind turbines and are extensively documented making them an ideal choice for demonstrating the capabilities of the new LBM simulation tool.

This project is suitable for a student with Aerospace, Mechanical or Computational Engineering degree with demonstrated skills in computer programming (essential). Substantial knowledge of fluid dynamics and engineering mathematics from undergraduate coursework is expected. Familiarity with the lattice Boltzmann method as well as extensive parallel programming experience are not necessary but will be provided as part of the project. Good communication skills are indispensable as you will become part of a team working with the same code base applying modern software development principles. Large-scale computations will be carried out on the Iridis compute clusters of the University of Southampton (12,000 and 18,000 cores) and national supercomputing facilities.

See previous project website and my gallery webpage for some images. If you wish to discuss any details of the project informally, please contact Dr. Ralf Deiterding, Aerodynamics and Flight Mechanics Research Group, Email: r.deiterding@soton.ac.uk, Tel: +44 (0) 2380 59 3384.

Eligibility: This offer is for applying together for a fully funded PhD scholarship by the University of Southampton and is open to all EU citizens, however, UK citizen are not eligible. If you are a UK citizen interested in work on a closely related topic, please see the fully funded PhD scholarship advertisement High Fidelity Simulation of Atmospheric Dispersion with the LBM.

How to Apply: Please send your CV and transcripts directly to r.deiterding@soton.ac.uk
Alternatively: Click here to send a formal application and select the programme - PhD in Engineering and the Environment. Please enter the title of the PhD Studentship in the application form.
Contact Information:
Please mention the CFD Jobs Database, record #15559 when responding to this ad.
NameRalf Deiterding
Emailr.deiterding@soton.ac.uk
Email ApplicationYes
Phone+44 (0) 2380 59 3384
URLhttp://www.ngcm.soton.ac.uk/projects/Massively-parallel-adaptive-lattice-Boltzmann-simulation-for-wind-turbine-applications.html
AddressSchool of Engineering
Highfield Campus
Southampton, SO17 1BJ
Record Data:
Last Modified14:05:06, Thursday, December 27, 2018

[Tell a Friend About this Job Advertisement]

Go to top Go to top