| dc.description.abstract | Buckling for thin metal shells are not an easy failure mode to design structures for. The large variety of
geometries, material properties, and uncertainties has med standards and best practice for designing
thin metal shells very conservative with their safety factor. The design is usually based on the theoretical
strength for buckling, and then, a knockdown factor, is multiplied with the strength in order to achieve
a strong enough structure. For many geometries, this knockdown factor reaches a value of 0.2, therefore
reducing the design strength to one fifth of the theoretical strength. Equinor is planning to create a
subsea shuttle tanker (SST) in order to transport liquids to and from subsea wells. The outer hull of
the SST will not be subjected to large external pressures dew to cargo tanks inside and a flooded inside.
However, the shell will be very large (100 meters long and 17 meters in diameter), and must be design
to withstand some pressure. This thesis will not look at the effect of stiffeners, as the main goal is to
establish a method to design a safe structure that is lighter and cheaper. Using existing studies done on
surface imperfections and physical tests (a study done on small nickel cylinders in the 1970’s known as
the A-shells), this thesis has put together both Final Element Analysis (FEA) and statistical analysis in
order to achieve a knockdown factor that is larger than what is standard and best practice today. Many
realisations of cylinders with different imperfections are simulated and then the result is fitted to a
probability distribution. The best distribution found was the distribution in the Estimation of Extreme
Values by the Average Conditional Exceedance Rate (ACER) Method. With this method, the knockdown
factor could be increased from 0.238 to 0.823 for the A7-shell and from 0.497 to 0.691 for the SST. | |
