ON THE ADEQUACY OF METAMODELLING TECHNIQUES IN PROBABILISTIC DESIGN OF RAILWAY SLEEPERS

Abstract eng:
The traditional approach used in railway sleeper design is deterministic and based on static strength analyses with input data conditions according to design guidelines. In the future, a probabilistic approach is of interest. For this an accurate and efficient reliability analysis method is needed. This work investigates the possibility of using metamodelling techniques in conjunction with Monte Carlo simulation to estimate the probability of failure of a sleeper. Different response surface techniques and radial basis functions are included in the study. The influences of the choice of experimental design and the extent of the parameter space are also investigated. Adequacy of metamodelling and the computational saving using metamodel techniques are in focus. Numerical examples using a mixture of real statistical test data and fictious data are given.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket