Determination of ship roll damping coefficients by a differential evolution algorithm

Mauro, F and Nabergoj, R (2021) Determination of ship roll damping coefficients by a differential evolution algorithm. Journal of Physics: Conference Series, 2090 (1). 012135. ISSN 1742-6588 (https://doi.org/10.1088/1742-6596/2090/1/012135)

[thumbnail of Mauro-Nabergoj-JoP-2021-Determination-of-ship-roll-damping-coefficients-by-a-differential-evolution-algorithm]
Preview
 Text. Filename: Mauro_Nabergoj_JoP_2021_Determination_of_ship_roll_damping_coefficients_by_a_differential_evolution_algorithm.pdf
Final Published Version
License: Creative Commons Attribution 3.0 logo
Download (1MB)| Preview

Abstract

The execution of the so-called extinction tests represents the classical experimental method used to estimate the damping of an oscillatory system. For the specific case of ship roll motion, the roll decay tests are carried out at model-scale in a hydrodynamic basin. During these tests, the vessel is posed in an imbalance condition by an external moment and, after the release, the motion decays to the equilibrium condition. When the damping is far below the critical one, the transient decay is oscillatory. Here a new methodology is presented to determine the damping coefficients by fitting the roll decay curves directly, using a least-square fitting through a differential evolution algorithm of global optimisation. The results obtained with this methodology are compared with the predictions from standard methods. This kind of approach seems to be very promising when the motion model of the system under investigation is established with any level of non-linearities included. The usage of the fitting procedure on the approximate analytic solution of the differential equation of motion demonstrates the flexibility of the method. As a benchmark example, two experimentally measured roll extinction curves have been considered and suitably fitted. The newly predicted results, compared with the ones obtained from standard roll decay analysis, show that the algorithm is capable to perform a good regression on the experimental data.

CORE (COnnecting REpositories)