Quantitative risk assessment using Monte Carlo and dynamic process simulation

Abstract

Currently, the concern about the industrial risk is a key issue to implement any technology process or to improve the industry competitiveness. In this sense, the risk concept may be considered as the main tool to anticipate behaviors that can lead to further problems. Considering the process industry, different risk analysis techniques are employed to identify hazardous events, to estimate their frequencies and severities, and to characterize the risk, being such tools the best ones to improve the industrial safety. Knowing that, the present Thesis discusses these risk topics to propose four main contributions: (i) new procedure to identify hazardous events; (ii) new procedures to quantify frequency; (iii) new risk definition and representation; and (iv) a method to integrate the proposed procedures to manage a complete risk assessment management. The idea behind the contributions is to use computational tools in new techniques with improved results about the operational risk, helping its obtainment and understanding. Thus, based on a new risk definition that allow better relation between the developed analysis, process simulations are employed to identify hazardous events and Monte Carlo simulations are employed to estimate frequency and to generate a new risk representation characterized by a severity x time x frequency surface. Despite all contributions has its particularity and importance for the risk analyses development, as final contribution, the presented Thesis apply all developed techniques in a case study, proposing an innovative risk assessment procedure.