Power System Disturbance Classification With Online Event-Driven Neuromorphic Computing

Accurate online classification of disturbance events in a transmission network is an important part wide-area monitoring. Although many conventional machine learning techniques are very successful classifying events, they rely on extracting information from PMU data at control centers and processing them through CPU/GPUs, which highly inefficient terms energy consumption. To solve this challenge without compromising accuracy, article presents novel methodology based event-driven neuromorphic computing architecture for power system disturbances. A Spiking Neural Network (SNN)-based framework proposed, exploits sparsity disturbances promotes local operation unsupervised inference incoming data. Spatio-temporal signals first extracted encoded into spike trains achieved with SNN-based supervised framework. In addition, benefits deep spiking networks complex multi-class event identification problem presented by leveraging increasing dynamic neural sparse depth. Moreover, QR decomposition-based selection technique proposed to identify participating the low rank subspace multiple events. Performance method validated collected 16-machine, 5-area New England-New York system.