Evaluation of inter-facility uncertainty for ship manoeuvring performance prediction

• Inter-facility tests evaluated with bivariant charts show systematic uncertainty.
• Previously unaccounted for Type-A uncertainty in PMM tests found to be very large.
• 3rd-order derivatives can be super-sensitive to small changes in 1st-order terms.
• Uncertainty for 3rd-order terms CMT/PMM does not overlap, questioning equivalence.
• Provides method to obtain estimate of confidence interval for manoeuvring criteria.

Ship manoeuvring, like many engineering problems, requires the prediction of motion performance through time-domain simulation based on force derivatives obtained by experiment. Procedures for obtaining the experimental uncertainty in the force measurements are typically given by the relevant national metrology institutions. However, the propagation of uncertainty from force measurements through to the predicted derivatives and onward into motion performance predictions, is less well defined. This paper presents a Monte-Carlo type approach for evaluating the propagation of uncertainty from force derivatives through to predicted performance parameters. Then, using a case study together with the published results from an inter-facility bias test, the paper identifies the like sources of uncertainty. The results show that, while some experimental uncertainty is evident, the likely cause of scatter (between facilities) is systematic in nature. That is to say, the current experimental procedures are ill contrived or insufficiently defined.