Dynamic modeling and mitigation of cascading failure in power systems

Graduation date: 2015Recent blackout events consistently show that a variety of mechanisms are involved in cascading outages. These cascading mechanisms are irregularly modeled and validated within the existing literature and industry practices. Understanding the relative significance of these different mechanisms is important for choosing which one(s) needs to be modeled for specific applications. In this work, a cascading failure simulation model that captures fundamental dynamics of power networks and protection systems was developed in order to evaluate the usefulness of dynamic models for cascading outages. The results from a batch of N-2 contingency simulations revealed that the distributions of blackout sizes and event lengths from the proposed simulator correlate well with historical trends. In addition, the proposed model was compared against a quasi-steady state (QSS) model, and it was found that a wide set of dynamic cascading mechanisms are critical in the definition of later stages of the cascades. However, the early stages of cascades showed similar paths independently of the relative number of mechanisms implemented. \ud \ud This work also proposes a novel emergency control algorithm in the context of dynamic cascading outage simulations. It is a centralized, optimization-based control scheme that utilizes a dc power flow approximation to quickly provide load and generation adjustments when the system condition is compromised. Dynamic simulation results showed that cascading risk is considerably reduced with the assistance of successful emergency control actions. Lastly, several other significant elements in dynamic power system modeling were addressed in this study (for example, dynamic load models). They further demonstrate one of the advantages of using our proposed power system simulator: in-depth access to the model components within the simulator package, which is not feasible in commercial softwares