Probabilistic Methodology for Design of Arctic Ships

Abstract

With the increased demand for hydrocarbon and mineral resources, as well as tourism, marine transportation in the arctic continues to increase. The region is harsh and fragile, making safety and environmental protection paramount. A key concern is how to estimate extreme design events that first and foremost satisfy safety and then economy. A rational approach to design of arctic ships based on probabilistic methods is detailed in this paper, including global impact forces and local panel design. Using a probabilistic approach, extreme design events can be identified by combining annual, seasonal and regional variability in environmental conditions with model uncertainty, and integrating these directly in the design methodology. Extreme design loads are estimated based on the annual number of interaction events, and the design strategy - target exceedence criteria established based on general public safety. The approach also provides a comprehensive basis for the selection of an appropriate ice class given certain operational requirements (e.g. an icebreaker for facility protection, or suitability of a cruise liner, having minimal ice class, to operate through a particular season). Otherwise, design for extremes is largely based on observational experience and judgment from one or more experts and such experience only reflects a relatively short span of natural occurrence. Neither is it appropriate to arbitrarily establish the most extreme condition imaginable. For extremal analysis, a parent distribution of global impact forces is first developed either through numerical ship-ice simulations or directly from measured full scale ship ram trial data. Using the parent distribution, and the expected annual number of ram events, a new design distribution representing the maximum of n annual force events is developed. Based on the global analysis, mean penetration and duration can be estimated which, when combined with number of interactions per year, provides a measure of exposure, a key input for local design analysis. A rational local pressure design model is presented that is derived based on measured ship ram data. Peak pressures through the full ram duration are considered and not just realizations at some arbitrary point of maximum force. A local scale effect is measured where pressures on smaller areas (i.e. <10m2) increase considerably above the global scale effect. A design pressure area curve based on a design strategy (e.g. 100 yr return period) is produced. A hypothetical design illustration is provided for a ship along a particular route including estimates of global forces and local design pressures. A linear trend in forces based on logarithmic trend in number of ramming events is observed. This illustrates that linking the design forces and pressures directly with expected number of interaction events is most reasonable and appropriate. Vertical impact forces estimated using the Polar Class rules are compared with estimates using the rational probabilistic approach and measured full scale MV Arctic data. The analysis illustrates how measured forces and expected exposure can be used for design and classification, as well as calibration. Preliminary results indicate that the higher PC1 and PC2 class forces seem rather conservative and a large gap exists between PC2 and PC3. Further calibration is needed. Introducing different design strategies (e.g. elastic-plastic and fully plastic) for corresponding load levels (e.g. 10−2extreme and 10−4abnormal) should be considered, allowing the designer to better understand the performance of his design.