The behaviour of the NO96 membrane containment system when subjected to breaking waves is presented in this paper. The data obtained from the full scale impact tests in the Sloshel project are used. Uni-directional breaking waves were generated and interacted with a real membrane containment system. Three types of hydro-structural interaction can be distinguished based on the relative position between load durations and natural periods. The aim of this paper is to identify which type of interaction occurred between the breaking wave and the NO96. It is demonstrated that it is not straightforward to define the interaction type by a comparison of observed load durations and predefined mode shapes. As an alternative, it was investigated whether the type of interaction can be deduced from its effect on the load and on the response with respect to a quasi-static interaction.
The use of numerical simulations to assess membrane containment systems subjected to sloshing is beyond the state-of-the-art. The only tool available today for such an assessment is model tests at small scale. However, the tests represent an experimental modeling of the reality. One of the difficulties is that the response of the containment system is not directly obtained, because the tank walls are rigid in the model tests. The structural response needs therefore to be calculated based on the measured loads, taking the mutual interaction between load and response into account. The difficulty here lies in the fact that a large database of measured pressures needs to be processed. Furthermore, it is not straightforward to include the hydro-structural interaction after the interaction took place. Dynamic amplification factors are used today in combination with a statistical treatment of the measured pressures as an engineering approach to include the dynamic effect of the load on the structural response.