Power System Transient Stability Assessment Using Stacked Autoencoder and Voting Ensemble

Increased integration of renewable energy sources brings new challenges to the secure and stable power system operation. Operational emanating from reduced inertia, in particular, will have important repercussions on transient stability assessment (TSA). At same time, a rise “big data” system, development wide area monitoring systems, introduces paradigms for dealing with these challenges. Transient concerns are drawing attention various stakeholders as they can be leading causes major outages. The aim this paper is address TSA problem perspective data mining machine learning (ML). A novel 3.8 GB open dataset time-domain phasor measurements signals built dynamic simulations IEEE New England 39-bus test case system. processing pipeline developed features engineering statistical post-processing. complete ML model proposed analysis, denoising stacked autoencoder voting ensemble classifier. Ensemble consist pooling predictions support vector random forest. Results classifier application reported discussed. promising, since it able ingest huge amounts while retaining ability generalize real-time decisions.