Sammendrag
This paper describes the need for improved methods for validating numerical models used in shiphandling simulators. Such models vary in complexity, from rather simplistic models used for initial shiphandling training at maritime training centers to high-quality models used in the study of advanced marine operations. High-quality simulation models are also used in investigations of maritime accidents such as collisions and groundings. The SIMMAN 2008 conference presented the results of benchmarking studies of simulation tools currently used by research institutes, universities and training centers around the world. Many of these tools employ models based on numerical calculations using methods based on potential or viscous fluid flow, experiments using scale ship models (free running or captive) or semi empirical expressions based on regression analysis of previous model tests. The organizers of SIMMAN 2008 made the hull characteristics of certain ship types available for a comparative study of simulation maneuvering models. The outcome of the benchmark study (using IMO standard maneuvers as case study maneuvers) showed that simulated results varied significantly. In the opinion of the authors, there is an urgent need for new validation studies. The first part of this paper discusses the concepts of simulation model fidelity, verification and validation and the present guidelines issued by ITTC for validation of maneuvering simulation models. The second part looks at the outcomes of the SIMMAN 2008 conference and describes MARINTEK's contribution to the benchmark study. The use of real-world measurements in model validation is briefly discussed. The need for registration of actual test conditions, as well as the types of tests that should be included in a test scheme, are presented. Finally, the authors discuss validation requirements with respect to the actual application of the selected simulation model as an engineering tool that can be transferred to training simulators used by maritime training centers. It is assumed that simplified simulation models may reduce the quality of simulator based training for ship officers. It is believed that increased quality of simulator model will improve the transfer of training from simulators to real life operations and remove some of the uncertainties related to investigation of maritime accidents.
Vis fullstendig beskrivelse