Influence of Evacuation Policy on Clearance Time under Large-Scale Chemical Accident: An Agent-Based Modeling

Abstract

Large-scale chemical accidents that occur near areas with large populations can cause significant damage not only to employees in a workplace but also to residents near the accident site. Despite the increasing frequency and severity of chemical accidents, few researchers have argued for the necessity of developing scenarios and simulation models for these accidents. Combining the TRANSIMS (Transportation Analysis and Simulation System) agent-based model with the ALOHA (Areal Location of Hazardous Atmospheres) dispersion model, this study aims to develop a modeling framework for simulating emergency evacuations in response to large-scale chemical accidents. The baseline accident scenario assumed the simultaneous leakage of toxic chemicals from industrial complexes near residential areas. The ALOHA model results showed that approximately 60% of residents in the scenario's city were required to evacuate their homes. The majority of evacuees completed their evacuations within 5 h in the baseline scenario (evacuating maximum number of private vehicles without any intervention), while the distribution of the population and street network density caused geographical variability in clearance time. Clearance time can be significantly reduced by changing both the evacuees' behaviors and the evacuation policy, which suggests the necessity for proper public intervention when the mass evacuation of residents is required due to chemical accidents.