Measuring the steering performance of ships during full-scale trials and model tests 1

During the last few years several papers have been published about adaptive autopilots, where the parameters of the controller are automatically tuned. To show the ‘optimality’ of the design, in most cases a simple quadratic criterion is used to measure the savings which may be expected. Full scale measurements which clearly show fuel-savings, are difficult and time consuming to carry out and until now they are hardly reported. This paper gives some results of a series of full-scale trials.

Another solution is to measure the difference in performance of various steering algorithms with the aid of a scale model in a sea-keeping basin, where the conditions during the measurements can be well controlled. A series of model tests has been carried out at the Netherlands Ship Model Basin (NSMB) to compare the performance of the adaptive autopilot of Delft University of Technology with a conventional PID controller. A description is given of the set-up of these measurements and the results which are obtained are discussed.