PhD Defence 14th June: "Estimation of Waves and Ship Responses Using Onboard Measurements"

Byggeri Skibe og off-shore-konstruktioner

Najmeh Montazeri from DTU Mechanical Engineering defends her PhD, "Estimation of Waves and Ship Responses Using Onboard Measurements", Tuesday, 14th June at 14:00. The defence takes place in Room S01, Building 101 at The Technical University of Denmark. Principal supervisor is Associate Professor Ulrik Dam Nielsen and co supervisor is Professor Jørgen Juncher Jensen. Examiners: Associate Professor Harry Bingham, DTU Mechanical Engineering, Professor Toshio Iseki, Tokyo University of Marine Science and Technology and Dr. Jesper Dietz, Maersk Maritime Technology.

This thesis focuses on estimation of waves and ship responses using shipboard measurements. This is useful for development of operational safety and performance efficiency in connection with the broader concept of onboard decision support systems. Estimation of sea state is studied using a set of measured ship responses, a parametric description of directional wave spectra (a generalised JONSWAP model) and the transfer functions of the ship responses. The difference between the spectral moments of the measured ship responses and the corresponding theoretically calculated moments formulates a cost function. A set of wave parameters, characterising the directional wave spectrum, is estimated through an optimisation problem using global search basin with proper constraints.

This approach applies a sequential partitioning procedure, which is able to classify swell and wind sea events using wind information. The model is tested on simulated data based on known unimodal and bimodal wave scenarios. The wave parameters in the output are then compared with the true wave parameters. In addition to the numerical experiments, two sets of full-scale measurements from container ships are analysed. Herein, the validation of the estimation method is assessed by comparing the results with the wave data from other tools, such as wave radar data and hindcast data.

The results show that the developed method is reasonably efficient. Automatic selection of a set of responses to be used for wave estimation is also studied using a sensitivity analysis of the wave parameters. This selection depends on the waves and the operational condition of the ship. Therefore, the method can be utilised based on initial knowledge about the waves and the operational condition in a specific location.

A dynamic trend model is proposed for tracking the evolution of the wave parameters during the voyage. This provides a prediction of the wave parameters, e.g. 20 minutes ahead of the measurements. Given the predicted parameters, a wave spectrum model and the transfer functions, forecasts of different wave-induced responses are made. The predicted variances of the responses are compared with actual measurements. The relatively good agreement in this comparison validates the model and the optimisation method. Finally, an uncertainty analysis of the presented approach is implemented to assess the reliability of the method.