Loss Functions and Their Applications In Loss-Based Process Safety Monitoring

Process variations along with failure of control system and layers of protection result in safety and quality loss. The purpose of this research is to provide a new loss function-based process safety monitoring method by exploring the relationship between process variations and system loss. For this purpose, properties associated with quadratic loss function and some inverted probability density loss functions (IPLF) are investigated and compared. The results suggest that loss functions can be used as innovative means for determining the economic advantage of improving operational stability and system safety. From the loss function-based estimation, one can quantify the financial consequence associated with process variations and get the severity rating in an objective way. A numerical example is presented to demonstrate the industrial application of the proposed method by examining the response to a temperature surge for a reactor system, where the hazards associated with two alerts caused by low and high temperature are analyzed and integrated with response time for warning generation and prioritization.