Identification of ship coupled heave and pitch motions using neural networks

Investigation of the ship motion behavior in irregular sea states is an important step for ship seakeeping performance research. Ship motion identification from the full scale measurements is the only way to study the actual motion behavior and verify the motion predictions after ship constructions. A particular identification method for coupled heave and pitch motions was developed and validated in this research. The two-degree Random Decrement technique and the Neural Networks technique were combined in identification process.; This developed method was applied to several motion systems to test its effects. The random motion data were obtained from the ship model experiments and numerical simulations. The coupled heave and pitch Random Decrement signatures obtained from the random motion histories were used as the Neural Networks training data to identify the Random Decrement equations. The identification results were verified by comparing the predictions with the actual Random Decrement signatures, and with the free response signatures.; The application results suggested that the validation of the identified equations was mainly dependent on the nature of the Random Decrement signatures and the quality of the Neural Networks training. Only White Noise or broad-band spectrum excitations could yield the required agreement between identified Random Decrement equations and motion free response equations.