Retreating Arctic sea ice offers new opportunities to harvest natural resources and opens up new shipping routes. In order to secure the safety of maritime operations, knowledge of ice-induced loads is needed. However, the ice-breaking process is a complex one. Full-scale measurements offer an opportunity to study the ice-induced loads on a ship hull as all the complexities are embedded in the measurements. The measurements are commonly conducted by measuring strains and the loading is determined from the load-strain relation. Thus, full-scale measurements contain additional uncertainty and variation related to the measurement techniques. Although the measurement uncertainty is an important topic, it has not received a noteworthy attention in earlier publications. The aim of this dissertation is to study the uncertainty and variation in the measured loading on the ship hull related to the load length. The focus is placed on the ice load measurements based on the shear strain difference. In this study, an analytical grillage model is derived to determine the shear force and torsion on the ice-loaded and adjacent unloaded frames. The model is validated with a Finite Element Model and a calibration pull. The model is applied to define the structural parameters affecting the load transfer between frames. Furthermore, the theoretical model is used to obtain a robust estimate of the possible error and uncertainty related to the load length and the extent of the instrumentation. Full-scale experiments on board S.A. Agulhas II are carried out in the Baltic Sea with first-year sea ice and the load length distribution is retrieved from these measurements. In addition, the measurements show that the loading on a frame increases as a function of the external load length. Furthermore, it is shown experimentally and explained theoretically that the extension of the instrumentation does not affect the measurements when the loading is short. However, the uncertainty and variation in the measurements increases when the load length increases beyond the instrumented area as a result of the load transfer between the frames. The effect of the load length on the mean value, standard deviation, and probability distribution of load amplitude on a frame is explained. Furthermore, the study shows that the linear-like increase of the mean values of the ice-induced load on a frame as a function of the standard deviations is a mathematical phenomenon rather than a physical one. The study shows that the shape of the probability density function for the ice-induced load on a frame is exponential-like for short loads and lognormal-like for longer loads. In a case where a single probability distribution is to be fitted to both short and long loads, Weibull distribution gives the best fit to the measurement data.
|Translated title of the contribution||Epävarmuus ja hajonta laivan rungolta mitatuissa jääkuormissa|
|Publication status||Published - 2018|
|MoE publication type||G5 Doctoral dissertation (article)|
- ice load
- ice load statistics
- measurement uncertainty
- shear strain measurement