Simulation-based reliability assessment methods of structural systems.

摘要

Safety and functionality are important considerations in designing structural systems. Complex structural systems consist of several structural components that are connected in regular interrelationships for the purpose of achieving specific functional and performance requirements. The performance of a system depends on the performance of the components composing the system and the relationships among the components. Redundancy is used in structural systems to achieve desired performance and safety levels.; Reliability is a probabilistic measure of assurance of the performance of a system. Because uncertainties are inherent in structural systems, absolute safety of structural systems cannot be guaranteed. Uncertainties are due to incomplete information, and the reliability approaches account for such uncertainties. Structural systems may fail in different modes of failure. Each mode of failure depends on the sequence of failure combination of components in the system. Correlations usually exist among the components of a structural system. Correlations can affect the sequence of failure combinations and the reliability of the system.; In this dissertation, simulation-based methods were developed for assessing the reliability of structural systems. Fault tree analysis was selected as a method for assessing the reliability of systems. Simulation methods were used for assessing the reliability of structural components and their failure modes. The direct Monte Carlo method was used as the base simulation method. Variance reduction techniques were used to improve the simulation accuracy and efficiency. The antithetic variate technique, importance sampling technique, and conditional expectation technique were selected as variance reduction techniques. Discrete event simulation method was presented as a method for reliability assessment of structural systems.; The efficiency of the developed simulation methods was verified and assessed. In simulation methods developed in the dissertation, the probabilities of failure of the system, components, and their failure modes were estimated. The coefficients of variation of the estimated probabilities of failure were calculated. These methods showed accurate and efficient estimate of the probability of failure of structural systems especially for cases where variance reduction techniques were used. The existence of correlations among structural components and failure modes affected the redundancy, the estimated probability of failure, and the sequences of failure combination of components of the systems.