A novel international initiative is underway to establish a purely computational database of synthetic buckling resistances on the basis of which the safety margin of metal wind turbine support towers can be calibrated within the Eurocode framework. This initiative has recently demonstrated that existing databases of laboratory-scale destructive buckling tests of cylindrical shell specimens are not an appropriate basis upon which to calibrate safety margins of full-scale civil engineering shells, and asserts that such a database based on laboratory testing alone is realistically unlikely ever to come into existence. A bolder approach based on high-quality nonlinear finite element buckling calculations of models generated with genuinely realistic imperfections obtained from the optical scanning of actual constructed towers is thought to be the only feasible path forward.
By way of preparation for this difficult task, this paper presents a critical appraisal of the extensive surface geometric imperfection surveys of lab-scale cylindrical metal shell specimens reported on by Arbocz and Abramovich, an important reference dataset for modern work of this nature, as well as the results of a series of nonlinear buckling calculations on finite element models constructed to include these imperfections. Differences between the finite element predictions and original test results are inevitably found and an explanation offered. The paper ends with a set of recommendations on the minimum scanning resolution for surface geometric imperfection surveys if they are to hope to yield sufficient information to create finite element models of realistically imperfect wind turbine support towers and other thin-walled metal shell structures.