Modern society is increasingly dependent on high quality electricity for its economy, security, cultural life, public health, safety and governance. As such, ensuring the resilience of electric power systems against deliberate attacks and natural disasters is critical to the continuous function, particularly of urban cities. This thesis proposes an efficient for assessing the resilience of electric power systems under hurricane hazards. The thesis also explores the use of wind turbines as distributed generation to provide back-up electricity during hurricane-induced outages. The study develops computationally efficient models for evaluating outages in electric grids, while demonstrating their applicability through modeling a large real system subject to natural hazards, structural and system responses, and restoration processes. It employs a Bayesian networks approach and uses influence networks constructed via N-1 contingence Direct Current (DC) flow analyses to make the framework computationally tractable, time-efficient and amenable for real-time updating of information via data fusion in the future. The framework computes hurricane-induced customer outages in distributed 1 km2 blocks across the entire system, and simulates system restoration according to resource mobilization practices and sequences identified from historic events. The study uses the Harris County electric grid in Texas under Hurricane Ike in 2008 to illustrate the framework’s application. The framework yields system responses that are in agreement with observed outages, with a mean error of 15.4% in outages aggregated at the ZIP code level. Performance comparison of the proposed framework with two previously existing models shows that the model has a better prediction accuracy and requires a significantly lower computation time than the existing models. The model takes minute and half as compared to more than an hour required by the previous models to run 50 simulations. Having observed widespread outages in the electric power system, with some lasting several days or weeks before power restoration, the study also looked at the reliability of wind turbines to support their integration in the form of distributed generation in power systems. The study introduced a closed-form methodology for computing the system failure probabilities of wind turbines considering different failure event definitions. The methodology is enhanced to incorporate consequences such as downtimes and repair costs of individual component failures, and to determine the turbine unavailability or cost risks. It yields vital reliability information that could be readily used for planning maintenance and forecasting wind power outputs necessary for widespread distributed wind generation. Furthermore, the study examines the use of tuned liquid column dampers (TLCDs) to increase the reliability of wind turbines. Comparison of results for wind turbines with and without the damper shows that a baseline TLCD of 1% mass ratio significantly reduces the structural vibrations (by as much as 47%), and considerably decreases the unavailability probability of a turbine (by up to 8%). Armed with the resilience assessment model for power systems and the reliability analysis tools for wind turbines, the study also develops a probabilistic model for quantifying the impact of distributed wind generation (DWG) on an electric grid during hurricane-induced outage periods. The model incorporates energy adequacy assessment principles while accounting for the uncertainty in electricity demands and in power output due to variability in the wind resource, unavailability of DWG units owing to turbine failures, as well as component failures in the main utility system. An application of the model to Harris County’s power system equipped with turbines of total rated capacity of 1.8 GW shows that the DWG can provide back-up power to up to 85% of the customers in a distribution area which directly connects a DWG unit, while reducing the overall outages in the entire county by 8.5%. Thus, DWG can help improve the resilience of electric grids, support the rapid recovery of hurricane-affect communities and reduce economic losses associated with widespread and prolonged outages. In summary, the study provides computationally efficient tools for exploring a wide range of what-if scenarios in large real energy systems. The models can be readily adapted to consider other emerging technologies such as storage systems, vehicle grids, smart grids and micro grids in electric grid resilience assessments. Thus, they can support resilience-based decisions for hurricane preparedness and mitigation, and restoration strategies that could ensure rapid recovery of the systems. They support efforts in ensuring a reliable and a sustainable supply of electricity during normal conditions or in the immediate aftermath of hurricane events. The outage assessment model, for instance, is directly implemented in the City of Houston’s Storm Risk Calculator, an online tool that informs resident users about the local risks they face from an in-coming hurricane, and the city’s emergency managers in hurricane disaster management. '