Deterministic prediction of waves and wave induced vessel motions: Future telling by using nautical radar as a remote wave sensor

With many operations at sea carried out by ships or or other floating vessels, risks are involved because of the waves and resulting motions of the ships. Examples are the landing of helicopters on ships, transferring crew from a ship to a wind turbine, or working on the deck of an anchor handling tug. It is common practice to assess the workability of a certain operation by statistical criteria. The rationale behind such criteria is in general that the probability of some phenomenon (e.g. the vertical motion of a helicopter landing deck) exceeding a certain threshold value has to be less than a chosen acceptable level. Operability analysis usually qualify a given wave condition as workable or not workable based on such statistical criteria, assuming that no information is available about when critical wave induced events occur. In this thesis, the feasibility is investigated to obtain a short term, ’deterministic’, i.e. time-specific prediction of the critical response: making available a short term prediction of approaching waves and vessel response real time, on-board, would give crew the opportunity to anticipate and chose the optimal moment to perform a critical operation. The research is motivated by two possible advantages of such a deterministic prediction: 1. It further enhances safety in conditions that were considered as workable from a statistical point of view.
2. It possibly increases workability by pointing out windows of opportunity in conditions that were considered as unworkable from a statistical point of view.
The chosen approach to obtain the mentioned deterministic prediction of waves and induced motion response, is to use the ship’s navigation radar as a remote wave sensor. The spatial domain that can be covered by a navigation radar to observe the sea surface is of course limited: both its minimum and maximum range is limited. Besides it is obvious that the wave observation will only be available in the past, and by no means in the future. Therefor, the first chapter answers the theoretical question where in the spatio-temporal domain waves can be accurately predicted, given a perfect spatio temporal observation of the waves. An indicator is proposed that specifies predictability in space and time based on the spatio-temporal observation and based on a given wave condition. It is confirmed that the group velocity of the waves is governing concerning this question. In the remaining chapters basically 2 different approaches are proposed and investigated to solve a linear wave representation based on input from synthesized radar images of sea waves. Finally the methods are applied to real radar data acquired during a sea trial. Based on the solution of the linear wave field, ship motions were predicted using pre-computed linear motion tranfer functions. Correlation coefficients up to 0.86 were obtained for the heave motion predicted 60 sec in advance.