Loss Functions and Their Applications in Process Safety Assessment

Process deviations, along with failure of control systems and protection layers, result in safety and quality loss in plant operations. This article proposes an operational risk-based warning system design methodology based on overall system loss. Loss functions (LFs) are used to define the relationship between process deviations and system loss. For this purpose, properties associated with quadratic LF and a set of inverted probability LFs are investigated and compared. The results suggest that LFs can be used in a novel way to assess operational stability and system safety. The proposed consequence assessment methodology using LFs is then incorporated into a risk-based warning system design model to analyze warnings associated with process deviations. A simulated case study is presented to demonstrate potential application of the proposed methodology; the study examines the response to a temperature surge for a reactor system. (c) 2014 American Institute of Chemical Engineers