Friday, July 16, 2010

Deepwater Drilling Design: Shell vs BP

Two Shell drilling experts, Joe Leimkuhler and John Hollowell, recently gave an exceptionally informative talk to the Aspen Ideas Festival.

The talk begins with some surprising statistics on the proportion of current US oil production generated from deepwater drilling and then proceeds to an overview of the technology and process of deepwater drilling. The heart of the talk, found around the 18 minute mark and again in the discussion period beginning at the 27 minute mark, describes the differences between how Shell designs and drills deepwater wells and how BP did. There is also some interesting discussion of drilling design requirements that the US government now requires but were not required/followed in the BP well.

These guys clearly have an interest in maintaining the viability of deepwater drilling and claim that good well design, which BP didn't follow, would have minimized the liklihood of an accident and, if one occurred, would have provided a larger number of potential ways to close down the well early on. They are, however, up front that the specific cause of the accident is not yet known and, hence, they don't go so far as to say such accidents are impossible with Shell rigs. Everything considered, they come across as both honest and forthright.

That said, the details they provide are a sad testimony to anyone familiar with the sociological literature on technological accidents (e.g., Perrow's Normal Accidents or Vaughan's Challenger Launch Decision). The talk illustrates both a failure to incorporate the principals of design advocated by Perrow (e.g., independent and redundant components) and the malignancy of poor institutional decision making chronicled by Vaughan. Sadly, it took a second shuttle accident before NASA took Vaughan's analysis seriously. Hopefully the oil industry has a better learning curve.



If the above video doesn't work, you can find the original here: Aspen Ideas Festival 6392010 Audio / Video Library

No comments:

Post a Comment