The model test was aimed at the validation of practical non-ergodicity ( correct math term is cyclic non-stationary quality). It means that the ergodic hypothesis cannot be used for any practical calculation of nonlinear irregular rolling. In other words, all the probabilistic characteristics have to be averaged over a representative ensemble of realizations and any result based on one realization is not correct. This was first started on the base of numerical simulations.
The model experiment was carried out in the towing tank of the National Research Institute of Fisheries Engineering in Japan. There were two series of tests with two models. The first series of tests was done with free-drifting model of a Japanese purse seiner. Such a test yields about 10 minutes of realization. Since the absence of ergodic qualities can be checked only on a significant amount of time, the second series of tests was conducted with a restrained model that was not able to drift. The second series produced 30 and 40 minutes realizations of model time that is close to a quasi-stationary range of full-scale waves. It was meant that the first series can be used for validation of the second one.
The paper is focused on model test technique.