Computational & Technology Resources
an online resource for computational,
engineering & technology publications
International Journal of Railway Technology
ISSN 2049-5358
IJRT, Volume 6, Issue 4, 2017
Multi-Objective Optimisation of Railway Vehicle Seating Design for Passive Safety Improvement
M. Carvalho1, J. Milho2,3 and J. Ambrosio2

1UNIDEMI, Department of Mechanical and Industrial Engineering, Faculty of Science and Technology, Universidade NOVA de Lisboa, Portugal
2IDMEC, Instituto Superior Técnico, Universidade de Lisboa Lisboa, Portugal
3CIMOSM, ISEL - Instituto Superior de Engenharia de Lisboa Instituto Politécnico de Lisboa, Lisboa, Portugal

Full Bibliographic Reference for this paper
M. Carvalho, J. Milho, J. Ambrosio, "Multi-Objective Optimisation of Railway Vehicle Seating Design for Passive Safety Improvement", International Journal of Railway Technology, 6(4), 69-88, 2017. doi:10.4203/ijrt.6.4.4
Keywords: injury biomechanics, railway crashworthiness, simulation, optimisation.

Abstract
This paper presents a multiobjective optimisation problem formulation for the identification of the seating layout design parameters for the coach interior safety during the crash of railway vehicles. The optimal problem is defined by the minimization of the relevant injury criteria for railway passengers that leads to a set of design solutions that are searched using a multiobjective genetic algorithm. The evaluation of the objective functions is based on the use of surrogate response surface functions instead of using directly the dynamic analysis. The only real costs involved in the methodology presented reside on the number of analysis with the detailed model required to identify the quadratic surface parameters, this number always being smaller than the number of iterations taken in any optimisation process. The results show that the optimal designs of the interior seating layout are obtained with relevant decreases of the injury indices and it is also observed that the difference between the injury indices evaluated using the surrogate model are in weighted average less than 5% different from those calculated with the realistic detailed model.

purchase the full-text of this paper (price £25)

go to the previous paper
return to the table of contents
return to IJRT
purchase this issue (price £80 +P&P)