Design of ships and offshore structures: a probabilistic approach for multi-year ice and iceberg impact loads for decision-making with uncertainty

Ralph, Freeman E. (2016) Design of ships and offshore structures: a probabilistic approach for multi-year ice and iceberg impact loads for decision-making with uncertainty. Doctoral (PhD) thesis, Memorial University of Newfoundland.

[img] [English] PDF - Accepted Version
Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

Download (52Mb)


Ice is a complex material that exhibits different failure properties depending on the loading rate, temperature and salinity. Under fast loading rates such as a ship ramming a multi-year (MY) ice, it fails as a brittle fracturing material. Fracture and spalling processes nonsimultaneously reshape the contact zone resulting in concentrated forces on localized contact areas. These localized High Pressure Zones (HPZs) are highly variable in time and space. The relationship between local and global processes is that the sum of n HPZs forces transferred into the structure at any point in time is the total global force transmitted to the structure. As with other fracturing materials, an inherent scale effect exists. Global pressures result from the sum of n HPZ forces averaged over the nominal contact area (e.g. the imprint of a ship’s bow into the ice without correction for spalling effects). The maximum global force will generally occur at the end of a ram at the maximum nominal contact area. Due to the random occurrence of natural flaws in the ice, pressures will vary as fractures occur, continually changing the contact face. A global scale effect exists such that pressures on larger contact areas, including zones of low and zero pressure, average out to be smaller. Unlike global pressures, maximum local pressures may occur on any panel and at any point through the ram duration. Modeling exposure is important as design pressures will increase for increasing number of interaction events as well as increased penetration or duration. The scale effect for local pressures within the nominal contact area is more demanding than for global pressures such that pressures on smaller areas are considerably higher. While this is expected, given confinement can suppress damage and limit fracturing events, a force limit exists where microstructural damage occurs, softening the ice and causing HPZs to fail. Local pressures on varying panel areas were studied based on spatial HPZ density and HPZ force. Building on earlier HPZ analysis using Louis S. St. Laurent data, in this thesis HPZ density and forces were derived from analysis of four Polar Sea data sets. The occurrence and intensity of HPZs on panel areas were simulated using a Poisson Process and an exponential distribution for HPZ force. The influence of modeling HPZ cutoff force on HPZ density, HPZ force distribution as well as local pressure parameters were studied and appropriate combinations recommended. Building on the Polar Sea HPZ analysis, a new model was developed for this thesis that considers HPZ occurrence in time through a ramming event, modeling HPZ rate. This was further enhanced by correlating HPZ rate with ship speed. Such a model allows the designer to determine baseline ‘parent’ local pressure design parameters based on vessel size and expected operational speed. The faster a ship operates through an ice regime, the greater the HPZ rate. Larger and faster ships will penetrate further, having longer interaction durations and hence a greater number of HPZs forming (unless, for example, the ship passes through a ridge). For design, we are interested in the maximum local pressure on a single panel area through the ram duration. Rates too will vary along the vessel being greater on the bow and least from mid-body to stern. For fixed structures designed for iceberg impacts, rate and duration based on iceberg size and drift can be used to model exposure in time. For floaters, modeling HPZ formation in time provides a means to estimate dynamic global forces and mooring loads illustrating benefit of compliance effects. Modeling of HPZ occurrence over a panel area is also very attractive for structure response analysis. The random placement of n HPZs over a structural panel gives a better basis to model stress localization, which is very important for limit states design. A preliminary review of the IACS Polar Class rules was carried out in this thesis. Global impact forces are estimated using a kinetic energy collision model. Consideration for modeling ice crushing strength assumes a pressure-area relationship that is proportional to A⁻⁰·¹ which is not consistent with experimental results demonstrating a scale effect proportional to A⁻⁰·⁴. The resultant design formulation models excessive semi-local pressures increasing with increasing semi-local contact area. While the intent is to model increasing pressures locally with increasing vessel displacement and subsequent penetration and contact area, justification for this trend suggests that there is no reason for traditional pressure area scale effects to exist and that with confinement, fracturing processes will be limited. But fracturing processes exist at all scales. The occurrence and behavior of HPZs either lead to very large stress localization that enhances fracture events or they undergo microstructure damage that softens the ice at the structure interface. While the design trend in the Polar Class rules may be okay, the background ice mechanics can be improved. An alternative collision model is developed in this thesis with an ice strength model based on data and an exposure algorithm to model pressures increasing locally with larger displacement vessels. In the mid 1990s as part of the Arctic Shipping Pollution Prevention Regulations (ASPPR) proposal reviews, a probabilistic time-step ship ram model was developed to estimate impact forces. Consistent with the ASPPR work, exposure based on annual number of collisions was mapped to each Polar Class (e.g. PC1, PC2, PC3 can expect on the order of 10000, 1000, 100 rams per year respectively). Using the MV Arctic as a test case and exercising extremal analysis, impact forces were estimated for each Polar Class. Characteristic 10⁻² global forces were compared with Polar Class rule estimates. Probabilistic local pressures were also compared with rule based estimates. Assuming impacts with MY ice, preliminary results show that plating design pressures may be reasonable, with recommendation for adjustment to the Polar Class 1 coefficients to reduce conservatism, and possible increases for lower classes. Analysis should be extended to other vessels and operating conditions. A probabilistic methodology for design of ships based on the principles of safety and consequences is important and necessary both for design and safety validation. Such approaches can consider the class of the vessel on the basis of expected number of annual interactions with extreme ice features. An example illustration of a design based on an arctic shipping route, ice conditions, design strategy, risk mitigation via detection and avoidance and resultant local pressures on the hull for structural design.

Item Type: Thesis (Doctoral (PhD))
Item ID: 12411
Additional Information: Includes bibliographical references (pages 328-331).
Keywords: Ice Structure Interaction, Arctic Shipping, Offshore Structures, Probabilistic Methods, Uncertainty, Extremal analysis, Ice Mechanics, Scale effects, Iceberg, Multi-year ice, Global Pressure, Local Pressure, High Pressure Zones, Exposure, Icebreaker, Polar Class Rules, Probability, Ship design
Department(s): Engineering and Applied Science, Faculty of
Date: October 2016
Date Type: Submission
Library of Congress Subject Heading: Ice--Impact testing--Mathematical models; Naval architecture; Ships--Design and construction; Offshore structures--Design and construction

Actions (login required)

View Item View Item


Downloads per month over the past year

View more statistics