Ice loads on a bottomless cylinder under dynamic wave action: An experimental study

In this study, we report the results of a comprehensive experimental analysis of ice loads on a bottomless cylinder under dynamic wave action and explore advanced methods to identify, isolate, and quantify these loads. To achieve this, we investigate and discuss various system identification techniques, such as the Prony method, dynamic mode decomposition (DMD), and covariance-driven stochastic subspace identification (SSI-cov), highlighting their applicability to experimental data. We also employ advanced signal processing approaches, including Robust Principal Component Analysis (RPCA) and robust local regression, to separate ice loads from quasi-periodic wave loads. Our findings reveal that ice and wave loads exhibit a positive correlation with wave height. Further, when the wave height remains unchanged, the normalized collision duration increases with wave frequency and shows insignificant dependence on peak ice loads. Additionally, ice loads are found to be $3\sim 4$ times larger than wave loads. These results contribute to advancing our understanding of wave–ice–structure interactions and provide practical insights for designing resilient offshore structures in ice-infested waters.