Methods for Ice-Model Updating Using a Mobile Sensor Network
MetadataShow full item record
The need for dynamic positioning to function in ice-infested waters is growing as the offshore oil and gas industry enters the Arctic. The ice introduces great challenges, some of which can be resolved through proper ice management. This requires good knowledge of the surrounding ice-environment.This thesis deals with the question of achieving a good state estimator for a sea-ice model. The dynamic thermodynamic sea-ice model of Hilber III (1979) is implemented, and it is shown through simulations that it reacts in a realistic manner to varying air temperature. The states of this model are estimated with an ensemble Kalman filter, and it is shown that different states can be estimated very well by ensemble Kalman filters based on different measurement configurations. This implemented nonlinear sea-ice model and state estimator is meant to serve as a platform where methods designed to select measurement configurations best suited for state estimation can be tested.A suggestion for a method which chooses measurement configurations on-line is presented. The idea is that this method allows for different measurement configurations to be applied at different time steps, all based on which one that provides the best estimate at the current time. Unfortunately there was no time to implement this method and test it on the previously mentioned test platform; it must be kept in mind that it is merely a theoretical suggestion which must be further tested.