Keynote Speakers

Dr. Andrew Pollard

Dr. Andrew Pollard

Andrew Pollard is a Professor of Mechanical and Materials Engineering, and a Queen’s Research Chair in Fluid Dynamics and Multi-scale Phenomena. He received his education in Canada (Waterloo) and England (Imperial College, University of London) and subsequently has held visiting academic positions in Germany, France and New Zealand. He has consulted widely for industry in both Canada and abroad. He serves on the editorial boards of academic journals and scientific advisory committees of several international symposia, as well as a member of various international and national government, university and scientific review panels, councils, committees and fora. During his 30-year academic career, he has published or edited over 200 articles and books. He has successfully supervised 50 graduate and post-doctoral students. He is an Alexander von Humboldt Fellow, professional engineer and a member of various academic and professional societies, including Fellow of the American Society of Mechanical Engineers.

Pollard was a founding member and inaugural President of the Computational Fluid Dynamics Society of Canada, a founding member and Director of the Centre for Advanced Gas Combustion Technology; the project leader for the High Performance Computing Virtual Laboratory, which is a multi-university consortium in Eastern Ontario that now supports the research of over 300 computational scientists, engineers, social scientists and medical researchers; a founding member, Chair of the Board of Directors and the President of a national organisation, C3.ca Association, which promotes and supports national sharing of distributed high performance computing facilities, personnel and applications; it now serves well over 3000 researchers in Canada. He is the inaugural Director of the Queen’s M.Sc. Specialisation in Computational Science and Engineering, which has drawn together 17 departments within Queen’s to focus on CSE. He is also the inaugural Director of the Sustainable Bio-economy Centre at Queen’s. This multi-disciplined Centre focus is on all aspects of the bio-economy for the Great Lakes regions.

Dr. Marian Nemec

Dr. Marian Nemec

Marian Nemec is a Senior Research Scientist at ELORET Corporation. He works in the Advanced Supercomputing Division of NASA Ames Research Center.  His research interests include the development of adjoint methods for discretization error analysis and shape optimization, embedded-boundary Cartesian mesh methods, adaptive mesh refinement and the application of these methods to problems with complex geometry.  He obtained his PhD in 2003 from the University of Toronto Institute for Aerospace Studies and came to NASA Ames as a Post Doctoral Research Associate the same year.

Abstract:  Error Estimation and Adaptive Mesh Refinement in Complex Geometry Problems

In many flow simulations, most of the time is consumed by specifying and cleaning the surface geometry, hand-crafting a computational mesh, checking the results and refining the mesh to obtain improved results.  The actual flow solution time is relatively short.  To address the tricky task of constructing meshes that appropriately balance accuracy with computational cost, we extend the method of adjoint-weighted residuals to embedded-boundary Cartesian mesh solvers. The adjoint approach is used to drive an incremental strategy of h-refinement to control discretization errors that influence the accuracy of user-selected outputs, such as lift or drag, and to estimate error bounds for these outputs.  We discuss the salient features of the implementation on irregular meshes.  To demonstrate the accuracy, robustness and efficiency of this approach, we begin with verification studies of airfoil databases spanning subsonic, transonic and supersonic regimes. We compare the standard fixed-mesh approach, where all computations are performed on the same mesh, with an adaptive approach, where all cases are computed to a prescribed error tolerance in selected outputs. Thereafter, we present several databases that satisfy prescribed error bounds for parametric studies of launch-vehicle and aircraft configurations in non-smooth flowfields that involve shocks and jets.

 

 


CFDSC Website


Sponsors

 

COMSOL

SHARCNET

Western Engineering

Western Research

 

Bombardier

 

Newmercial


Ansys Website

Pratt & Whitney Canada Website

MAYA Heat Transfer Technologies Ltd. Website

Rolls-Royce Website

MITACS Website

Also of interest:

Western provides the best student experience among Canada's leading research-intensive universities.