Machine Learning Prediction of the Morison Equation Coefficients

Abstract Prediction of forces on cylindrical structures in marine environments such as jacket platforms and monopiles typically relies on the Morison equation. The Morison equation relies on two coefficients to accurately predict the magnitude of the inertial and drag forces experienced by the structure. These coefficients are sensitive to the diameter and the external shape of the cylinders. Both of these factors change in the presence of biocolonization through an increase in diameter and in macro-roughness. This article details how machine learning can be used to estimate these coefficients in a number of scenarios and demonstrates the applicability through experiments in a wave basin. Several machine learning methods are compared. The approach allows for improved accuracy of force estimates when assessing loads on marine structures, especially for scaled testing in wave basins, where progression of technological readiness is typically done. The benefit of this approach, as opposed to the current binary approach of rough or smooth structures, is that we can examine the evolution of force over time allowing for improved estimates of fatigue lifetimes or improved maintenance cycles.