Simulation-based predicting the position of road tank explosions. Part II: a case study

Abstract

A case study describing a simulation-based prediction of geometric characteristics of a road tank accident is presented. The case study evaluates the performance of the accident prediction model proposed in the first part of this study (Vaidogas et al. 2012). The prediction of accident characteristics is decomposed into three tasks: (i) prediction of the longitudinal rest position of the tank vehicle within the road segment under analysis; (ii) prediction of the transverse rest position with respect to road centreline; and (iii) prediction of the departure angle of the tank. These tasks are performed by applying stochastic (Monte Carlo) simulation. The results of the prediction are simulated samples of the geometric accident characteristics. These samples are considered to be input information for the assessment of risk posed by a potential explosion of tank vehicle vessel. In this case study, the potential targets of the explosion are three reservoirs built in the roadside territory. The case study presents and discusses in detail probabilistic models used for the simulation. It is stated that a considerable part of these models are chosen subjectively due to scarcity of circumstantial data on road tank accidents. The predictive Bayesian approach to risk assessment is used as a methodological basis of the simulation. Results of the simulation are intended to be utilised for increasing the safety of transportation of hazardous materials by road tanks.