Extrapolation of Test Data in Time, Size and Risk: A Challenge for Concrete Design Codes

It is no secret that the lifetimes of concrete structures have been inadequate, and that structural failures are far too common, especially in the case of large structures of novel designs. This problem has two aspects: 1) randomness of loads such as those from traffic, environment and random vibrations, and 2) the uncertainty in material properties, role of corrosive agents, and failure mechanics. Significant progress has been achieved in the first, but less in the second, which is a problem of extrapolation. The second is just as important, yet has been mostly neglected by structural safety specialists. Only the second aspect is discussed in this lecture, and only in qualitative, non-mathematical, terms. A broad vision of the extrapolation problem is offered. Lapses of quality control are a different problem, not discussed here.