SYMPOSIUM SERIES NO 161 HAZARDS 26 © 2016 Texas A & M University 1 A Framework for Developing Leading Indicators for Offshore Drillwell Blowout Incidents Nafiz Tamim 1 , Delphine Laboureur 1 , Ray A. Mentzer 1 , M. Sam Mannan 1 and A. Rashid Hasan 2 1 Mary Kay O’Connor Process Safety Center, Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843-3122, USA 2 Harold Vance Department of Petroleum Engineering, Texas A&M University, College Station, TX 77843-3116, USA Offshore operations have always been very challenging due to technological and operational complexities in combination with harsh environmental conditions. Geological uncertainties, high pressure flammable fluids in the presence of ignition sources, complicated structural layouts, limited response time allowance, difficulty of control and communication are some of the critical factors that pose clear threats towards safe operations and may result in high consequence events i.e., blowouts. Developing well specified risk indicators is difficult due to such highly correlated factors and multifaceted operations. Leading indicators, which are able to identify critical events that could lead to high consequence events, have proven to be an effective tool that can help the operators in their decision making to react earlier to an event and to reduce the risk of an incident. Most of the research dedicated to leading and lagging indicators are applicable to the petrochemical industry, and there is not yet an agreement on a definition and classification of leading indicators related to drilling related blowouts. This paper discusses the approaches of different organizations and institutes on leading indicators characterization and development. The drilling industry is compared with the aviation industry to identify potential elements for developing a comprehensive leading indicators based risk model. A workable definition of leading indicators is proposed considering the intricacy of offshore operations. Leading risk indicators are broadly categorized into two classes which are further segmented into different groups. Proposed categorization is analyzed with a blowout case study and simple decision support algorithms are proposed for detecting gas kicks which are major precursor to blowouts. Keywords: Leading indicators; Risk metrics; Offshore blowouts; Risk assessment; Predictive tool Introduction Blowouts are considered to be the most notorious events in drilling operations and have caused hundreds of fatalities and injuries, millions of barrels of oil release to the environment and billions of dollars of property damage over the last few decades. As per US Bureau of Safety and Environmental Enforcement (BSEE) and US Bureau of Ocean Energy Management (BOEM) statistics, from 1980-2011 a total of 77 blowouts and 32 major well release events were reported from 31,574 drilled wells in US Gulf of Mexico [BOEM, 2014]. The picture is almost similar in other parts of the world. Many catastrophic events resulted from uncontrolled well releases while drilling or during other well-related activities. For instance, in recent times, the Deepwater Horizon blowout in the US Gulf of Mexico caused 11 fatalities [Marsh, 2013] and about 4.9 million barrels [U.S. Coast Guard, 2011] of oil spill in 2010 and the Montara blowout in Western Australia caused about 30,000 barrels [Koh Q., 2012] of crude oil spill in 2009. But incidents like Deepwater Horizon and Montara do not just happen due to a single failure and usually result from a complex combination of deficiencies that coincide – technical or operational failure, inadequate safeguards or safety management systems, and human factors. Focus on these factors can reveal any existing inconsistencies in the system that may initiate a blowout event. For general process industries, organizations have been using process safety metrics or risk indicators to evaluate and benchmark day to day safety performances. Historically, companies have been using lagging indicators i.e., the total recordable incident rate (TRIR), lost time incident rate (LTIR), number of fatalities or injuries in general to monitor and track organizational safety performances. But these are mostly personal safety measures and provide very little or no picture at all on overall process safety performances. So, industry started to consider leading indicators which are proactive or predictive measures and offer a closer look into operational and organizational safety culture. Having a workable set of process safety indicators came into discussion particularly after the Texas City refinery explosion in 2005 where the Baker panel report [Baker et al., 2007] recommended to establish leading and lagging process safety indicators to help prevent such incidents. Offshore drilling, being a very complex and high risk activity – can similarly be benefitted from implementation of well-specified leading indicators for early prediction of potential upsets. This work is dedicated to studying the scope and methodology of developing potential leading risk indicators for offshore drilling operations focusing on blowout incidents. Flow of uncontrolled well fluids into a wellbore and to the environment is called a blowout. As blowouts are low frequency-high consequence events, lagging indicators cannot offer a good measure because having a low past incident rate or low rate of gas kick events does not eliminate or help predict the chance of a future uncontrolled gas kick resulting in a blowout. Again, drilling is a multi-stakeholder process and organizational factors play a crucial role in risk management and acceptance which can only be taken into consideration with appropriate leading indicators. This paper includes a brief discussion of existing guidelines, recommended practices and relevant industrial works for developing process safety indicators. An arrow diagram relating lagging and leading indicators in drilling operations is proposed, followed by a detailed categorization of leading indicators. For effective safety performance monitoring and incident prevention, this work proposes incorporation of real-time indicators or process observables with