Deep Learning for Power Quality Event Detection and Classification Based on Measured Grid Data
Nuno M. Rodrigues, Fernando M. Janeiro, Pedro M. Ramos
Abstract
Energy consumption has increased over the years and, due to the dependency on fossil energy, alternative and renewable energy sources have been integrated to address environmental concerns. However it is important to maintain the efficiency, reliability, and safety of the power grid amid the integration of different energy sources. IEEE and IEC standards regulate power quality and define thresholds for power quality events that traditionally have been detected through specialized algorithms. With machine learning it is possible to detect and classify those events using deep learning techniques that teach systems to learn by example, providing a more scalable approach to classification. Published studies in power quality with deep learning algorithms to detect disturbances rely only on simulated signals or imposed disturbances. In this paper, a deep learning neural network is trained and used to detect and classify power quality events from a database built with real electrical power grid signals measured with monitoring devices.