System identification and generalisation of elastic mooring line forces on a multi-float wave energy converter platform in steep irregular waves

A computationally efficient model for the calculation of mooring line forces is essential for mooring system design, yet is a challenging problem for realistic irregular sea states which can be highly nonlinear and include wave breaking. Most existing numerical methods are computationally demanding and the limited work on data-driven methods only offer short term prediction, which cannot meet requirements of long-term prediction and unseen wave conditions for fatigue analysis. In this paper, computationally-efficient and long-term prediction models are proposed with data-driven system identification methods by using only limited physical experimental data for modelling mooring line forces of a multiple-float wave energy converter system. Further, it is the first time that mooring line force models trained with limited data are generalised to unseen wave conditions without requiring any measurements under these unseen conditions. This proposed method has been verified in physical experiments under eleven irregular wave conditions including steep and breaking waves. The peak mooring forces are shown to be dominated by low frequency surge motions, excited through second-order sub-harmonic wave components, making this a challenging problem to model. Despite this, the model is found to give comparable performance when using only linear surface elevation signals, convenient for future use as a design tool without the requirement for nonlinear surface elevation inputs. The promising results demonstrate that the proposed method potentially creates a transformative solution to the wave energy sector for mooring line force modelling.