COMPUTER SUPPORT FOR RESPONDING TO RAILWAY EMERGENCIES

Abstract

This study focuses on the automation of decision-making in analyzing emergency situations on railway transport involving the carriage of dangerous goods (DG). The relevance of the research is determined by the growing volume of DG transportation and the high likelihood of accidents, which can have serious consequences for the environment, economy, and public safety. The main objective of the study is to develop mathematical models that provide the basis for an intelligent decision support system (DSS) for the localization of accidents and minimizing their consequences. The tasks of the study include analyzing existing models and software for predicting pollutant dispersion, formalizing the DG transportation system as a directed graph of states, developing mathematical models for the probability of safe railway operation and estimating the duration of accident elimination, and studying the impact of organizational and technological measures on the efficiency of emergency response units. As a result, models were developed that allow forecasting the development of emergency situations, assessing the required resources for accident response, determining optimal deployment and concentration times for response units, and considering the influence of DG properties and external conditions on the incident dynamics. Statistical analysis of railway accidents involving DG in EU countries over the past decade revealed a dependency of accident numbers on cargo traffic. The mathematical models were implemented in a software prototype to evaluate the duration of liquidation operations and potential environmental and economic consequences. The study concludes that the application of the developed models and DSS increases the objectivity of decision-making, reduces delays in emergency response, and contributes to minimizing environmental damage. Future work includes expanding DSS functionality, integrating data from unmanned aerial vehicles and other sensors, and applying models for real-time planning of emergency operations.