Process safety incident prediction using near miss data

Abstract

AbstractProcess safety is concerned with making sure that hazardous chemical industries are designed inherently safer and operated, and maintained by implementing safe operation and maintenance principles. It focuses on preventing events involving leaks, spills, fires, or explosions. Process industries are constantly using the lessons learned from prior mishaps and near misses to stop similar incidents from happening again. A robust near‐miss reporting system at HPCL‐Mittal Energy Limited (HMEL), which owns and operates the Guru Gobind Singh Refinery, enabled the development of an incident predictive model that uses big data to identify areas of safety concern and offer practical insights to help prevent future potential process safety incidents. Through this model, we have demonstrated that if a particular category of process safety‐related near miss, such as unsafe conditions or unsafe acts, are reported more frequently, it is a clear sign to the organization that there may be a potential weakness in the operational discipline, work processes, or safety layers of protection. The methodology also gives organizations the ability to create an action plan to address identified root causes and increase the integrity of safety systems. This paper examines HMEL's successes in preventing future big incidents and provides an illustration of the incident predictive model's operational methodology.