Seminar 11: Monday 23 June 2014, Time: 1pm-2:15pm
Room 2175, Level 2, Haydon Allen Building, The Australian National University
Speaker: Dr Jan Hayes, Senior Research Fellow, School of Sociology, Research School of Social Sciences, The Australian National University
On 9th September 2010, a high-pressure natural gas transmission pipeline ruptured under the suburb of San Bruno, near San Francisco, California. The resultant fire burned for two days. Eight people died, and thirty-eight homes were destroyed because a weld in the buried pipeline failed. The weld had been poorly made in 1956 when the pipeline was first constructed.
Using document analysis (investigation reports, transcripts of evidence and other primary source materials), in this paper we focus on the experts in this organisation who were responsible for pipeline integrity management. Clarke’s work regarding fantasy planning in the face of uncertainty explains both how and why the integrity management system had taken on a symbolic, rather than functional, role in this organization. ‘Knowing’ about the system was grounded in elaborate algorithms and graphs purported to show risk was declining, and yet this analysis was only tenuously linked to the actual level of danger. In an environment of cost cutting, expensive inspection work was tailored simply to meet available budgets. Small failures were repaired and then dismissed from the collective memory so that valuable opportunities for learning in this uncertain environment were ignored.
This professional disposition on the part of the expert engineers was embedded in a broader organisational context. The paper further argues that senior management decisions about cost cutting were divorced from the real world impact as a result of the historical restructuring of the domestic gas industry overall and the low status of relevant industry expertise at senior management levels. We conclude that this ‘organizational accident’ provides important lessons for all organisations relying on technical expertise to prevent low frequency, but high consequence events.