Human Reliability Under Physical Security Threats: Modeling and Experimental Design

Ensuring the safety and security of a safety-critical system requires a significant investment of financial and material resources. The operation and maintenance costs of a nuclear power plant have increased significantly and account for about 7% of the power generation cost. About half of the operation and maintenance costs are labor, such as the costs of security personnel responsible for physical security and the associated training of the operators. Indeed, the operators and the security personnel play an integral part in mitigating the threats incurred in safety-critical systems such as nuclear power plants in extreme situations. In the event of an attack on a safety-critical system, operators must take specific actions to safeguard the system, while security personnel takes actions to resist the attackers and ensure the operators perform their assigned duties. An efficient way to study the possible optimization of the security personnel workforce, understand their behavior as well as that of the operators, and evaluate their performance during an extreme situation, such as a physical attack, is to develop a simulated environment that can model such extreme situations. According to our previous investigations of the relevant literature in the field of human behavior under extreme conditions, the most common human behaviors during a natural disaster are identified as denying behavior, acceptance of loss behavior, practical behavior, and extreme behavior. In addition, the most common human behaviors during any extreme situation are identified as agitated behavior (flight, fight), depressive behavior (freeze, fright), pro-social activities, and affiliation (seeking the proximity of familiar persons and places). Various studies have been carried out to identify different factors that influence human behavior during extreme situations, and these factors can be classified into several major categories (such as Physical factors, Physiological factors, Psychological factors, and Human characteristics). In this study, experiments are designed to collect data on human behavior and performance during extreme situations, such as physical attacks on a nuclear power plant. The design of the experiment is based on those factors that influence human behavior during extreme situations, and the factors are measured to understand their impacts on human behavior and identify any particular behavior during an extreme situation. For example, the physical factors are measured by different physical parameters such as temperature, blood pressure etc., while psychological factors such as fear, performance stress etc., are evaluated through pre-designed questionnaires and facial expressions of the participants during the simulated experiment integrated with the virtual reality environment. The data acquired will be integrated into current human reliability methods, such as an extended version of THERP, to enhance our comprehension of human reliability and assess human performance in such critical scenarios. By doing so, we can optimize the allocation of operators and security personnel, ensuring adequate security and operational safety while reducing costs associated with training and maintaining security personnel and operators responsible for physical security.