Quantitative Modeling of Human and Process Safety Risks Using Real Operational Data in High Risk Industrial Systems

Abstract

High risk industrial systems such as chemical plants, oil and gas facilities, and large scale manufacturing operations continue to experience severe accidents despite the extensive use of safety management systems and conventional risk assessment tools. One of the main limitations of traditional approaches lies in the insufficient integration of human related factors and process safety risks within a unified quantitative framework, as well as the underutilization of operational safety data generated during daily industrial activities. This study proposes an integrated quantitative modeling framework that simultaneously evaluates human safety risks and process safety risks using real operational data collected from high risk industrial environments. The proposed framework combines data driven human reliability analysis with probabilistic process risk modeling to capture the dynamic interactions between human performance, technical systems, and organizational conditions. Operational data such as accident records, near miss reports, safety performance indicators, and abnormal event databases are systematically processed to estimate human error probabilities and process failure likelihoods. Bayesian based modeling techniques are employed to represent causal relationships and uncertainty propagation across human and process safety layers, enabling a comprehensive assessment of coupled risk scenarios. The methodology is applied to a representative high risk industrial system to demonstrate its practical applicability and analytical capabilities. Quantitative results are presented through multi parameter risk matrices, probability distributions, and risk prioritization tables, allowing decision makers to identify dominant risk contributors and critical safety barriers. The results indicate that neglecting human process interactions can lead to significant underestimation of overall system risk, while data informed integration provides a more realistic and actionable risk profile. This research contributes to the advancement of quantitative risk management in industrial engineering by offering a structured and data grounded approach for integrating human and process safety risks. The proposed model supports evidence based safety decision making and can be adapted to different industrial contexts where reliable operational data are available.