Was there a sufficient discussion of uncertainty in the 2001 University of Minnesota report analyzing the bridge? The general conclusion of the report was that failure was not likely, and a page 11-14 analysis argued that the idealized models used for evaluating bridges are "inherently conservative." Given that the the report identified existing fatigue cracks, should there have been more of a discussion of possible unknown causes of failure? Was it too overconfident?
Initially, I don't immediately see a way in which there could have been a more nuanced discussion of uncertainty in the report. Historical evidence cited in the report seems to give a lot of credibility to assuming that the model calculations were conservative. The report indicated existing fatigue cracks but used a lot of empirical testing in their analysis of whether failure was imminent. If there are faults in the analysis or great inadequacies in the discussion of uncertainty, I don't yet see them.
Subscribe to:
Post Comments (Atom)
1 comment:
Thanks for pointing out this report Zach. I was wondering to what extent the concepts of Perrow detailed in his book 'Normal acidents' (1984) might prove useful in examining what led up to the castrophic collapse? The Minnesota report acknowledges that there wasn't much redundancy built into the bridge. In Perrow's terms we could say there was a high degree of 'coupling'. Perrow also advocates the examination of interactivenss of components of systems in terms of their 'complexity' and 'linearity'. The former to draw attention to the "possible incomprehensibility" and/or unexpected outcomes of interactions in a system. The latter to assess the degree to which interactions might be anticipated. Ultimately a judgement must be made as to what degree of understanding/ignorance is acceptable in any system that we chose to build. As the title of Perrow's book suggests, inevitably and unfortaunately we will sometimes get this call wrong.
Post a Comment