Enhancing abnormal events management by the use of quantitative process hazards analysis results
Systematic methods and tools for managing the complexity
Safety & Risk Management Systems (T4-5)
Keywords: Abnormal events management, process hazards analysis
Over the past 15 years, abnormal events management (AEM) has become an important issue in safety related process operation research. The goal of this research is to increase the safety and efficiency of chemical plants by developing sophisticated AEM systems. In order to achieve this goal, monitoring and diagnosis tools have been developed for the on-line detection and identification of the abnormal situations in a chemical plant.
Although AEM is still an open field requiring further work, significant progress has been made in providing support to plant operators in identification of possible causes of the abnormal plant conditions. Nevertheless, automation of AEM not only requires accurate fault diagnosis but also a complete system response with corrective actions which must be derived from the knowledge of the causes and consequences of the diagnosed faults. Process hazard analysis (PHA) is typically done off-line and provides qualitative answers about these causes (fault sources) and consequences (process variables crossing their allowable bounds). However, it does not provide the time at which specific variables will cross the bounds after the occurrence of an abnormal event.
In this work, a technique for AEM has been demonstrated which combines dynamic simulation with PHA to take advantage of the HAZOP analysis (a popular PHA technique). In this method, the HAZOP analysis is extended to generate quantitative information as thresholds of key variables, as well as corrective actions (CA) to rectify the consequences of each abnormal event.
The information from PHA is used to establish a CA protocol for providing on-line support to a plant operator against an abnormal event. The integrated approach uses a dynamic model of the plant to simulate the effect of the current diagnosed fault under the current plant conditions. The dynamic simulation is crucial in identifying the variable(s) that will reach the specified safety threshold(s) the earliest thereby helping the system to decide which corrective actions to implement. By combining dynamic simulation and the CA obtained from the HAZOP analysis, a new protocol can be developed that will rank CAs depending on the current plant conditions. The system has been developed using Matlab and Aspen Dynamics. An industrial sour water stripper plant is used as case study.
Presented Thursday 20, 12:00 to 12:20, in session Safety & Risk Management Systems (T4-5).
