Probabilistic Security Vulnerability Assessment Framework for Drone Attacks: A Case Study of a Petroleum Storage Terminal

Abstract

The risk assessment of intentional threats is credible given the significant increase in security incidents at chemical facilities over the past few years. The increasing accessibility of advanced drone systems to potential adversaries has worsened the security of chemical plants, with 78 security incidents reported at critical facilities since 2018. In this work, a probabilistic-based Bayesian network framework (PBBNF) is developed to assess the vulnerability of hazardous chemical facilities to drone attacks. The graphical-based BN framework incorporated various factors, such as drone and antidrone system capabilities, weather conditions, location, human factors, time, and other influencing factors, to assess intrusion risk and attack success. A case study of a petroleum distribution terminal is used to test the proposed framework, and further analysis is performed by developing 27 scenarios that vary the probability of different factors to examine the best- and worst-case scenarios. A sensitivity analysis of the model is also performed for the given case study, indicating that the performance of the autonomous missile system has a highly influential impact on drone intrusion, while the training of drone warriors has the least influence on the system’s intrusion risk. While the findings highlight that the use of highly efficient drones has a higher likelihood of successful intrusion, the level of vulnerability is varied due to adverse weather conditions, strategic location, and proper human vigilance, accompanied by the deployment of high-potential antidrone systems, all of which could lower the likelihood of intrusion attempts and attack success. This PBBNF benefits chemical and other critical facilities that handle hazardous materials, enabling effective decision-making to develop resilient infrastructure against aerial threats.