Probabilistic Evacuation Assessment with Real-time Monitoring Information

Abstract

The present paper proposes a probabilistic modeling approach for assessment and decision support of tactical loss reduction for roadway tunnel systems subject to fire events. The proposed probabilistic modeling approach combines scenario-based risk models for the probabilistic representation of accidents and fires with agent-based probabilistic representations of the escape scenarios of persons. The Fehmarn-belt tunnel is used as case study and a real-time daily traffic (RTDT) curve is considered. The vehicle population and the number and categories of persons in individual vehicles are modelled probabilistically based on statistics from the European Union. The example focuses on the modelling and assessment of the effects of tunnel closure, as a risk reducing measure in case of fire, on the evacuation dynamics. The results show that fast hindrance of traffic entry into the tunnel system efficiently reduces the expected value of the number of exposed persons but at the same time increases the variance associated with the number of persons evacuated within a given time frame.