Abstracts Category : Other

Fragility assessment and reliability analysis of transmission lines subjected to climatic hazards

by Seyedeh Nasim Rezaei

The main objective of this study is focused on developing a probabilistic framework for reliability analysis and fragility assessment of line components and systems subjected to wind and ice storms. The principal step to achieve this objective is to address the issue of unknown implicit limit state functions. Two main approaches are proposed for this purpose: the component-based approach and the system-based approach. The component-based method focuses on the novel application of the concepts of Statistical Learning Theory (SLT) and Vapnik-Chervonenkis (VC) generalization bounds for model complexity control in the field of structural engineering and for reliability analysis purposes. SLT and VC-generalization bounds have been previously used for model selection in solving practical problems such as motion estimation and tracking as well as traffic flow time series prediction. Our goal is to select a surrogate demand model of the finite element solver for each tower element or line component from several possible models that has good generalization (prediction) properties. Using this approach, component and system fragility curves are developed and further integrated with climatic hazards to estimate component and system failure probabilities. The relative importanceof different line components is further quantified and components are ranked based on the improvement potential importance measure and the Bayesian reliability importance measure. The proposed rankings are respectively useful for the purpose of upgrading lines with limited resources and fault diagnosis after an extreme climatic event. The system-based approach focuses on the novel application of a probabilistic support vector machine (PSVM) classification technique called the Venn Predictor (VP) in system-based fragility assessment and reliability analysis of transmission lines systems. SVM is a state-of-the-art machine learning technique which is used to discover patterns in highly dimensional and complex data sets. In this application, PSVM is used to determine the limit state function of the line and to provide estimates of the probability that the line fails or survives for given climatic conditions. This method allows developing system fragility curves without requiring information about component fragilities and is particularly useful in fragility analysis of complex structural systems with hundreds of members such as power networks. Both approaches are applied to a nine span segment of a typical 120 kV single circuit transmission line located in North America.This study also shows how the proposed methodologies can be used by researchers and engineers to solve some of the current challenges. For example, component and system fragility curves are developed for different scenarios of unbalanced ice loading that can result from non-uniform ice formation and/or ice shedding. In this application, the effect of wind direction on non-uniform ice formation and the resulting fragility curves is taken into account. Another application of the proposedAdvisors/Committee Members: Luc E Chouinard (Supervisor).