Fatigue reliability of ship structures

摘要

Today we are sitting on a huge wealth of structural reliability theory but its application in ship design and construction is far behind. Researchers and practitioners face a daunting task of dove-tailing the theoretical achievements into the established processes in the industry. The research is aimed to create a computational framework to facilitate fatigue reliability of ship structures. Modeling, transformation and optimization, the three key elements underlying the success of computational mechanics are adopted as the basic methodology through the research. The whole work is presented in a way that is most suitable for software development. The foundation of the framework is constituted of reliability methods at component level. Looking at the second-moment reliability theory from a minimum distance point of view the author derives a generic set of formulations that incorporate all major first and second order reliability methods (FORM, SORM). Practical ways to treat correlation and non- Gaussian variables are discussed in detail. Monte Carlo simulation (MCS) also accounts for significant part of the research with emphasis on variance reduction techniques in a proposed Markov chain kernel method. Existing response surface methods (RSM) are reviewed and improved with much weight given to sampling techniques and determination of the quadratic form. Time-variant problem is touched upon and methods to convert it to nested reliability problems are discussed. In the upper layer of the framework common fatigue damage models are compared. Random process simulation and rain-flow counting are used to study effect of wide-banded non-Gaussian process. At the center of this layer is spectral fatigue analysis based on SN curve and first-principle stress and hydrodynamic analysis. Pseudo-excitation is introduced to get linear equivalent stress RAO in the non-linear ship-wave system. Finally response surface method is applied to this model to calculate probability of failure and design sensitivity in the case studies of a double hull oil tanker and a bulk carrier.