Developing a Sound Process Safety Culture: A Persistent and Hard Work

Currently, many incidents or near misses are investigated using Root Cause Analysis (RCA) or other techniques.  DNV GL applied the barrier-based bowtie diagram in a novel manner for a customer in the nuclear industry so that they could determine the weak links in their work practice procedures for reactor isolation during maintenance.  This paper discusses the application of the same semi-quantitative methodology to evaluate incidents in downstream/petrochemicals industry.  Additionally, the paper demonstrates the visual effectiveness of a bowtie to summarize the results of reviewing multiple incidents in order to identify common factors contributing to incidents and near misses and assess the priorities for strengthening or adding barriers.

The first step is the development of a generic bowtie diagram showing left-hand (prevention) side and the right hand (mitigation) side for loss of containment for similar incidents based on information that is publically available most notably through the CSB. Next, the generic bowtie diagram is annotated for a set of different incidents or near misses.  This is accomplished by a review of the sequence of events for each incident to determine where/when the barrier failure(s) occurred.  The failed barriers are then tallied and categorized to determine the patterns associated with these failures. Finally, a summary bowtie diagram is developed to show how many times each barrier was compromised across the studied events, and which barriers were effective in detecting the deviation. 

The findings from application of this methodology to the nuclear power facility accomplished the following:

• identified the common factors contributing to incidents or near misses reported over a period of time

• revealed which barriers were effective in detecting or preventing the top event

• identified priorities for strengthening or adding barriers

• provided guidance for development of an action plan to strengthen existing barriers and add additional barriers to reduce the likelihood or consequences of future events

Overall, it is anticipated that the use of this same approach in reviewing incidents in petrochemical and refining facilities will proactively give the users insight into trends that might have previously been unnoticed.  The use of the methodology will also demonstrate which barriers are effective and help the user identify and focus on strengthening the barriers that previously failed.  It can also help users identify systemic issues with culture, organization, or layout between different facilities or regions based on the trends.