Catalogus

Around 2000 free running tests have been conducted in the Towing Tank for Manoeuvres in Confined Water (co-operation Flanders Hydraulics and Ghent University), Antwerp, Belgium, to investigate the effect of parameter variation, speed change and real physical ship-obstacle interaction on a model predictive controller for path following. Standard free running tests, accelerations and zigzags, are included for identifying the operating condition and the dynamical response of the ship model under consideration, a scaled, 4 m long bulk carrier. A model predictive controller is implemented to steer the ship along predefined routes, and a wide range of parameter variations allows to check the controller's limits and to propose future guidance for appropriate parameter settings. Batches of path following tests are then generated and carried out day and night, thanks to the unmanned carriage capabilities. All experiments were conducted in shallow water and at five speed classes. Two slim cubic blocks are installed in the tank at the later stage of the experimental program to compare the difference in control results between virtual and physical obstacles. The test results indicate some general parameter setting principles of the controller, show the speed influence in dynamic response and parameter setting, and reveal the non-negligible effects of hydrodynamical ship-obstacle interaction at close distance.