Optimization of an embedded rail structure using a numerical technique

Abstract

This paper presents several steps of a procedure for design of a railway track aiming at the development of optimal track structures under various predefined service and environmental conditions. The structural behavior of the track is analyzed using a finite element model in which the track and a moving train are incorporated. Parameters of the optimum track are determined by applying a numerical optimization technique. The optimization method employed here uses Mutipoint Approximations based on Response Surface fitting (MARS). To demonstrate the robustness of the procedure, it is applied to a problem of optimal design of an innovative railway track for high-speed trains - a so-called Embedded Rail Structure. Requirements for the optimal design are related to the wear of the rails and wheels and the level of acoustic noise produced by a moving train. To obtain the optimal design, component dimensions and mechanica properties of the track are varied. Results of the optimization are presented and discussed.