Title: Monte Carlo train derailment model for risk assessment

Authors: Weidong Ruan, Zongli Lin, Ted C. Giras

Addresses: Charles L. Brown Department of Electrical and Computer Engineering, University of Virginia, P.O. Box 400743, Charlottesville, VA 22904-4743, USA. ' Charles L. Brown Department of Electrical and Computer Engineering, University of Virginia, P.O. Box 400743, Charlottesville, VA 22904-4743, USA. ' Charles L. Brown Department of Electrical and Computer Engineering, University of Virginia, P.O. Box 400743, Charlottesville, VA 22904-4743, USA

Abstract: A Monte Carlo train over-speed derailment model for risk assessment is developed based on the derailment coefficient (wheel lateral and vertical force ratio). The model considers a one lump train negotiating on curved tracks or tangent tracks. The derailment coefficients for these two situations are calculated and the probabilities of train derailment are obtained to enhance a large scale Monte Carlo railroad risk assessment simulator called the axiomatic safety critical assessment process (ASCAP). A genetic algorithm based approach is taken to performing the validation of this Monte Carlo train derailment model. From the perspective of railroad risk assessment, the model is analysed and compared with an empirical formula, which is currently used in ASCAP. The Monte Carlo train over-speed derailment model will replace the current empirical formula in ASCAP simulator and will be able to provide a more realistic determination of the probability of train derailment.

Keywords: train derailment; derailment probability; derailment coefficient; genetic algorithms; ASCAP; axiomatic safety critical assessment process; railroad risk assessment; Monte Carlo simulation; wheel lateral force ratio; wheel vertical force ratio.

DOI: 10.1504/IJMIC.2008.021091

International Journal of Modelling, Identification and Control, 2008 Vol.4 No.2, pp.134 - 144

Available online: 03 Nov 2008 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article