Uncertainty and imprecision in safety assessment of offshore structures.

Non-deterministic phenomena unavoidably exist in offshore engineering practice and the available information to engineers regarding these phenomena is frequently not certain or precise but rather characterized with uncertainty and/or imprecision. This complicates an appropriate mathematical modeling and the associated engineering analysis required to derive reliable evaluations of structural performance and safety. This thesis aims to explore a realistic modeling of the available information regarding marine corrosion and marine growth in the complex ocean environment, and develop efficient computational procedures for an adequate processing of the information through safety assessment of offshore structures. Non-probabilistic models and the concept of imprecise probabilities are utilized in the modeling of marine corrosion and marine growth effects due to scarce and imprecise information. These effects on offshore structural reliability, robustness and damage detection, have been captured in the computational results, such as the probability of failure, robustness measures and the detected damage, through the developed procedures with high efficiency and accuracy. More reliable engineering decisions concerning the safety conditions of offshore structures can be achieved based on the computational results. Furthermore, it shows that these mathematical concepts in modeling of uncertainty and imprecision possess certain advantages in the case of limited information and associated difficulties of a traditional probabilistic modeling.