I keep being told infrastructures are complex technologies, even though they’re manifestly socio-technical and not just because the technologies have to be managed (i.e., this is more than “risk is socially constructed”).
We’re to believe regular operations are routine operations, but if routine means invariant, there is nothing invariant about normal infrastructure operations.
System reliability is probabilistic in the view of engineers, even though control room operators act deterministically, i.e., there’s a point at which system reliability cannot be traded off against other factors or else people would die.
I was assured that for reasons of tractability, the modeling of infrastructure operations has two stages, normal and failed. In actual practice, the temporary disruption of systemwide services–hardly ever modeled, let alone sufficiently–identifies highly relevant conditions for returning to normal operations or tipping into failure.
Engineers said the probability of infrastructure failure during post-disaster recovery of assets and operations was higher than the probability of failure during normal operations. Think: re-energizing line by line during a table-top Black Start exercise. Actually, nonmeasurable uncertainties–nothing like probabilities–are faced by operators post-disaster (the Black Start exercises for electric transmission infrastructure assume no asset destruction, as improbable as that is).
Consider the frequent “restore.” What’s it with respect to: interrupted services restored back to normal? Or services to be initially restored after major system failure? Or key equipment or facilities restored after a non-routine outage as part of normal maintenance and repair activities? Restore is one of the most ambiguous terms in infrastructure studies.
WHEN COMPLEX IS AS SIMPLE AS IT GETS 