(116i) Using Data Analytics to Generate PHA Checklists – Saving SME Time and Supporting SME Knowledge

The majority of those that have been exposed to Process Safety are aware of Hazard and Operability Studies (HAZOPs); a Process Hazard Analysis (PHA) methodology that is regularly used in industry because of its scalability and relative ease of execution when analyzing the risks of a process unit. Those who have been exposed to HAZOPs are also aware that these studies are qualitative, subjective and incredibly dependent on the people involved in the analysis session. Everyone who participates in a HAZOP does the best they can to do a complete and comprehensive analysis, however their contributions can be limited and biased based on their human knowledge and personal experience. This can lead to inconsistencies in identifying and capturing risks in comparable facilities/units, which can result in increased risks at site.

In order to strive towards higher quality, more comprehensive and more consistent HAZOPs, operation companies are starting to create HAZOP/PHA Checklists to help their teams better prepare for their sessions and be aware of potential high risk scenarios.

Unfortunately, HAZOP/PHA Checklists often take a large amount of time and effort for a Subject Matter Expert (SME) to create, and this amount of SME time is not always available in smaller operating companies.

So, what to do? There is obviously a benefit in having HAZOP/PHA Checklists, so how can the time of an SME required to generate one of these checklists be reduced while still creating a high quality reference tool for HAZOP teams? Also, if the insights of a few SMEs are so beneficial, what about the knowledge of ten or a hundred SMEs?

From these wants, a new method has been found to accomplish this task of generating HAZOP/PHA Checklists by leaning on the power of automatic data processing and data analysis of already completed PHAs. By combining the SME knowledge of tens to hundreds of HAZOPs of the same technology process together, a complete HAZOP/PHA Checklist can be generated with minimum manual SME input.

This paper shows a case study comparison between an SME generated HAZOP Checklist and a Big Data generated HAZOP Checklist, and highlights the high similarity/accuracy when compared, as well as the time savings of the a Big Data generated HAZOP Checklist methodology.