Page 263 - Practical Design Ships and Floating Structures

23 8

INTRODUCTION

Current structural design of surface ships is based on deterministic analysis methodologies and design
rules/requirements which are highly dependent upon test data and at-sea experience. Analysts and
designers are faced with numerous sources of uncertainty and product variability during ship design. In
traditional ship structure designs, uncertain parameters are treated as deterministic variables and
empirical factors-of-safety are used to account for uncertainties. Factors-of-safety are determined
based on past experience, but do not guarantee safety or satisfactory performance, nor provide any
information on how the different parameters of the ship structure influence safety. Thus, it is difficult
to design a ship system with a uniform distribution of safety among the different components using
factors-o f-safety .
Surface ships must survive a hostile environment. Ship vulnerability assessment is important in
identifying the levels of operational capability that must remain after a ship is damaged under extreme
dynamic loads. IJnder extreme environments, experirnenta! observations have shown that the loading
process is non-Gaussian and non-stationary and the ship structural response is nonlinear. Nonlinear
structural response is induced by the initiation and progression of multiple local damages, such as local
in-elastic/plastic deformation, stiffener tripping, panel buckling, or fracture. The complexity of fluid-
structure interaction phenomena renders the assumption on the loading process (Le. stationary and
Gaussian) invalid. The conventional approach, based on linear random vibration theories and peak
statistics, is inapplicable for the probabilistic vulnerability assessment of surface ships subjected to an
extreme environment. Therefore, it is imperative to develop a generalized, simulation-based,
probabilistic analysis tool which has no limit on the nature of input random processes (Gaussian, non-
Gaussian, stationary, or non-stationary) and system characteristics (linear or nonlinear).

The ship structural design process should maximize structural performance and minimize life-cycle
costs and weight, while ensuring an acceptable risk of failurc under operational, seaway and extreme
dynamic loads. Reliability-based design uses probabilistic methods to measure all uncertainties and
maximizes structural performance for an acceptable level of structural safety and reliability.
Reliability-based ship structural design will provide the best solution in the light of the available
knowledge, tools and consequent uncertainty (White et al., 1995). The two primary benefits of the
reliability-based approach are: 1) a formal and traceable measure of risk or safety in a new ship design
with the use of advanced materials and unconventional hull geometry; and 2) the ability to evaluate the
relative importance of various design options on the safety of ship structural components and provide a
consistent level of safety and eficiency throughout the ship (Mansour, 1990; Hess and Ayyub, 1997).
In order to perform the probabilistic vulnerability assessment of surface ships under extreme dynamic
loads, a stochastic finite element tool, SIMLAB has been developed by integrating the nonlinear finite
element code, DYNA3D, into a simulation based probabilistic analysis framework. SIMLAB can
provide probabilistic failure prediction of a structural system characterized by both random variables
and random processes (Lua, 2000). To demonstrate SIMLAB, an elastoplastic beam is subjected to a
random excitation to explore the effect of material nonlinearity on probability of failure and peak
statistics.
PULSTR is used to perform a reliability-based assessment of a surface ship structure at its preliminary
desigdanalysis stage. The uncertainties associated with global hull geometry, panels, hard comers,
stiffeners, and material properties are propagated into the hull ultimate strength prediction module. A
hybrid approach which combines both the MCS and FORM solution modules has been developed to
assess the reliability level of a Navy Combatant, Ship A, in both its single and double hull
configurations.