The demand on the power grid is constantly increasing. Combining this with the growing amount of renewable energy and widespread integration of distributed generation creates many challenges. These challenges push the power grid closer to its limits and can result in system instability. This makes the system more vulnerable to large scale events caused by voltage collapse and cascading outages. Several major power outages in recent history have been a result of multiple contingencies occurring within a small timespan. This research focuses on determining contingencies such as transmission line and transformer outages using an arti cial neural network (ANN) algorithm with wide area monitoring systems (WAMS). Methods have been proposed in recent years for using phasor measurement unit (PMU) measurements to determine line outages in a power system with varying success. In this thesis voltage phasor measurements are used as inputs to an ANN which then determines if any transmission line or transformer is out of service. ANNs can detect patterns in data and classify this data with great accuracy, giving them many advantages over other methods of determining outages. Results found in this research show that ANNs can be used to accurately classify outages in a power system. In addition, in order for an ANN to classify all line outages accurately in a test system every bus in the power system does not need to be monitored, resulting in cost savings for PMU deployment. Results also indicate that the ANN based algorithm can withstand a single PMU loss and still detect line outages accurately without having redundant PMUs installed at every bus in the system.