dc.description.abstract | In an ongoing project between DNV GL and a customer, the ultimate strength and
damage on a ship exposed to ice loading are to be assessed. The vessel has the DNV
GL class notation Ice(1A*), but is now considered for operation in more severe ice
conditions than it initially was designed for. The purpose of the thesis is to analyze
the response of a hull model subjected to different loads and to determine the residual
capacities after first yield. Ice properties and numerical models will also be reviewed.
Ice as a material is complex because its properties are strongly dependent on parameters
such as strain, strain rate, porosity, salinity, crystal structure and stress-state.
Moreover, the load distribution exerted by ice on impacting structures is highly irregular
due to brittle failure, including fracturing, crushing and spalling. These properties
make ice diffcult to model. The developed numerical models all have their flaws and there are currently no "verifed" material models for ice, despite decades
of research. The main issue seems to be their ability to reproduce the irregular load
distributions.
Class rules use a simplied approach where it is assumed that all energy will be
dissipated by the ice, implying crushing of ice. When exceeding design load level,
the hull will start to deform plastically and the residual capacity will be utilized.
Ice class vessels are actually expected to experience loads somewhat larger than the
design loads, and some deformation is accepted. The rules do, however, only have a
single design point at first yield and there are no further requirements on the residual
strength of the hull.
A finite element model of a part of the bow was considered. Rectangular, uniform
pressure patches of different sizes were applied at different locations of the model.
Linear response analyses were executed and the capacity with respect to first yield
was determined. The highest stress levels occurred in the stiffeners, and loads centered
the farthest away from the bulkheads were found to be the most critical. Linear
capacities ranged between 1.3 MPa and 3.7 MPa. It was also found that the rule
design loads are similar to the rst yield capacities calculated.
The residual capacities were determined by doing nonlinear finite element analyses.
A user defined ultimate strength criterion based on maximum allowable permanent
deformations was made. Permanent lateral deformation at stiffener midspan of 0.5
% or 1.0 % of the stiffener length was set as a criterion. The idea was to allow
deformations that can be regarded as acceptable, and the strain levels were found
to be well below material capacity using these criteria. The hull showed to exhibit
ultimate strengths in the range of 1.5 to 3.6 times the linear capacity, depending
on the load distribution. Two boundary conditions were considered, and they gave
equal results.
Even though the hull exhibits large residual capacities after first yield, it is questioned
whether this alone can be used as leverage for allowing the vessel to operate in more
severe ice conditions than it was designed for. Ice loads larger than design loads are
expected in the first place, so some of this residual strength is taken into account by
the class notations. Numerous damages on ice class vessels have been reported, so it is clear that the simplified method used in class notations has its weaknesses.
For future work it is recommended to do analyses with nonuniform pressure patches,
which is more realistic and more conservative. Doing long-term full scale measurements
is also recommended. A distribution function can be fit to the the measured
loads and from this a new design load can be determined using a probabilistic approach.
Using the capacities calculated in this thesis, it can be determined whether
the vessel can operate in harsher ice conditions than it was initially designed for or
not. | |