Robustness, repeatability and resilience of intercontinental distributed computing for the purposes of seismic assessment of bridges

Hybrid simulation is a promising tool that permits the integration of different laboratories in real time to undertake a set of joint experiments on individual components assembling a structural system. The components studied either analytically or experimentally can be physically distributed and thus potentially located anywhere. The same concept has also been applied for the numerical coordination of distributed sub-structures in the framework of multi-platform simulation. The advantage of the latter approach is that it permits the use of different numerical analysis packages, enabling the concurrent use of the most sophisticated constitutive laws, element types and features that each specialized software provides for each individual part of the system. Notwithstanding the major advancements made so far in hybrid experimentation, it is notable that the vast majority of such trans-laboratory campaigns has taken place within the U.S. The scope of this paper therefore, is to present the recent findings and technical challenges encountered in an attempt to run similar (hybrid and multi-platform) experiments both within Europe and between E.U. and the U.S. for the study of seismic soil-structure interaction effects in bridge structures. The paper also discusses the parameters that were found to affect the efficiency and repeatability of distributed analysis while it investigates the resilience of the procedure in cases of abrupt network failure. It also foresees the conditions under which the implementation of a geographically distributed hybrid simulation is indeed feasible, robust and repetitive between transatlantic partners. For this scope, the seismic response of a 110m long, four span, reinforced concrete bridge is preliminary assessed, through its numerical sub-structuring into five structural components (modules) that are analyzed in computers located in the cities of Thessaloniki (Greece), Patras (Greece), Naples (Italy), Urbana-Champaign (U.S.), and Toronto (Canada). The analyses results as well as the implications of the parameters involved in the feasibility of an intercontinental hybrid experiment are also critically discussed.