Non-linear dynamic analysis with deterministic and random seas: the case of minimum platforms

Minimum facilities platforms have a very simple layout and are largely in shallow water environments. Since their natural period is four, five times smaller than the design wave period, the design is usually carried out via a quasi-static analysis amplified "a posteriori" by a dynamic amplification factor. In this paper, we investigate the limitations of this approach by comparing quasi-static and dynamic results of a nonlinear, time domain, finite element analysis. Three different configurations of minimum platforms are considered: one freestanding caisson and two braced monopods. We begin by investigating the response under deterministic seas, using the stream function formulation. We then extend the analysis to random seas, using the JONSWAP spectrum with parameters measured from the Northwest Shelf of Australia. The first important result is the existence of considerable dynamic amplification under both deterministic and random seas. Interestingly, braced configurations are dynamically more sensitive than the unbraced monopod, even if the latter exhibits the largest top displacements. This can be inferred in the deterministic case from the higher values of the dynamic amplification factor. Under random waves this is further confirmed by the fact that the dynamic response of braced monopods exhibits resonant phenomena, and in particular is very sensitive to ringing. Ringing is characterized by sudden, large responses lasting for relatively short periods of time. It is shown that, among the several formulations for the dynamic amplification factor (DAF) in random only the one based on most probable maximum values takes ringing into account. Since so far ringing has been described mainly qualitatively in the literature, we suggest an innovative, quantitative indicator of ringing based on a careful assessment of its phenomological properties. We are therefore in a position to quantitatively compare the ringing behavior of different structures. This analysis confirme d that braced monopods are particularly sensitive to ringing. In conclusion, we show that for design purposes the use of deterministic versus random seas as a simulation tool for real ocean is conservative, yielding higher values of the dynamic response for all configurations. However, particular resonant phenomena, such as ringing, are not detected by a deterministic simulation.