Computational & Technology Resources
an online resource for computational,
engineering & technology publications |
|
Civil-Comp Proceedings
ISSN 1759-3433 CCP: 104
PROCEEDINGS OF THE SECOND INTERNATIONAL CONFERENCE ON RAILWAY TECHNOLOGY: RESEARCH, DEVELOPMENT AND MAINTENANCE Edited by: J. Pombo
Paper 39
Aerodynamic Shape Optimisation of a High Speed Train Locomotive using Adjoint Solver M. Bujny, A.B. Jaworski and J. Rokicki
Institute of Aeronautics and Applied Mechanics, Warsaw University of Technology, Poland M. Bujny, A.B. Jaworski, J. Rokicki, "Aerodynamic Shape Optimisation of a High Speed Train Locomotive using Adjoint Solver", in J. Pombo, (Editor), "Proceedings of the Second International Conference on Railway Technology: Research, Development and Maintenance", Civil-Comp Press, Stirlingshire, UK, Paper 39, 2014. doi:10.4203/ccp.104.39
Keywords: adjoint methods, gradient optimisation, high-speed trains, shape optimisation, mesh morphing, radial basis functions.
Summary
Computational fluid dynamics tools are routinely used in the analysis of high speed
train aerodynamics. Shape optimisation methods can further improve the aerodynamic
performance although the major limitation remains the computational cost. One may
consider to limit the computational effort by reducing the accuracy of numerical model
but this may affect the optimisation result. Another solution is to use gradient based
optimisation together with a very efficient adjoint method. However, this requires
an additional adjoint solver and parametrisation. In this paper a three-dimensional
gradient based optimisation method is presented combined with a very efficient mesh
parametrisation. The optimisation cost is limited to a few flow calculations by using a
commercial adjoint solver for the gradient calculation. The performance of proposed
optimisation approach is analysed using a numerical test case. It consists of the shape
optimisation of a train locomotive with a pantograph using aerodynamic drag as a
objective function.
purchase the full-text of this paper (price £20)
go to the previous paper |
|