Dynamic Risk Assessment of Inherently Safe Chemical Processes: Using Accident Precursor Data

点击翻译按钮获取中文摘要

ABSTRACTPI: Warren D. Seider and Ulku Oktem Institution: University of PennsylvaniaProposal Number: 0553941Title: Dynamic Risk Assessment of Inherently Safe Chemical Processes: Using Accident Precursor DataIntellectual Merit. Chemical engineers, researchers, the chemical process industries, and regulators have focused on improving the safety of chemical plants since the accidents at Flixborough, Seveso, and Bhopal. In addition, due to terrorism concerns since 9/11, additional security standards have been applied to the chemical and petrochemical industries. It is therefore desirable to have inherent safety and security, and dynamic risk assessment and reliability as vital ingredients in the planning, development, design, control, and operations of chemical plants.This research project aims to develop:(i) Inherently safer plant (reactor-separator-recycle) designs using game theory: Initially, design techniques using game theory will be extended to design inherently safer polymerization reaction processes. These techniques involve shifting the operating regimes from unstable, non-minimum phase behavior (with inverse response) towards stable, minimum phase behavior (without inverse response) at comparable profitability levels, thereby enhancing inherent safety. Designs are obtained that account for the tradeoffs between profitability, controllability, safety and/or product quality, and flexibility, by solving a multi-objective optimization problem using game theory. Previous work will be extended to include distillation columns, tubular reactors, and fluidized-catalytic reactors, among other process units.(ii) Plant-specific, dynamic risk assessment techniques using accident precursor data: The PIs have developed a mathematical model to estimate the failure probabilities of various critical accident scenarios, associated with a process unit given an abnormal event, using probability theory, including copulas and Bayesian analysis. The method will be extended for plant-wide analysis and tested with chemical industries, including Rohm & Haas (which participates in the Wharton Risk Management Centers Near-miss Management Project [NMMP]), and the incident databases RMP & NRC, with accidents being investigated by the Chemical Safety and Hazard Investigation Board (CSB). Broader Impacts. This approach provides a dynamic method to perform risk and vulnerability assessment of chemical plants considering the uncertainty as well as the variability in failure probabilities. The technique, using high-speed computers, will permit more thorough safety analyses, providing safer chemical plants. The models and software will be used by the chemical industries and in design courses at the University of Pennsylvania. These risk-assessment techniques should lead to more quantitative safety coverage in the PIs design textbook. Although the project focuses on failure probabilities of chemical plants, these techniques can be easily extended to other industries/organizations where precursors are important. The work is multidisciplinary in nature involving chemical engineers, risk analysts, epidemiologists, and statisticians.