Time-variant reliability of dynamic systems by importance sampling and probabilistic analysis of ice loads
Doctoral thesis
Permanent lenke
http://hdl.handle.net/11250/236389Utgivelsesdato
2006Metadata
Vis full innførselSamlinger
Sammendrag
The main objective of this thesis is to develop an efficient simulation technique to estimate the failure probability of time-dependent systems, whose state is expressed as a solution of Itô stochastic differential equations.
The work is divided in two topics due to requirements of the scholarship.
The first part addresses to the problem of assessing the reliability of dynamic systems, where the first-passage probability is chosen as a performance measure of structures subjected to the irregular, stochastic environmental loads. This problem has received considerable attention recently but it still remains a challenge for a wide class of systems.
The improved importance sampling method is developed which allows the solution of the first-passage problem and its applicability for single degree of freedom linear and nonlinear systems. Firstly, it is efficient for assessing small probabilities and it increases the convergence rate compared with the crude Monte Carlo method. Secondly, the procedure uses as much known analytical information about systems as possible.
The second part of the thesis features the probabilistic analysis of the ice loads on the Norstr¨omsgrund lighthouse situated in the Baltic Sea. The objective is to verify the spatial correlation model of the ice forces on the structure and estimate the design values from the appropriately chosen extreme value distribution.