Big Data in Smart Grids: Challenges and Opportunities

The IEEE PES Cyprus Chapter is organizing an IEEE PES Distinguished Lecture by Prof. Mladen Kezunovic, Texas A&M University, USA. The lecture will take place at the new campus of University of Cyprus, Room B104, Building ΧΩΔ02, on Friday 8 June at 11 am. The Distinguished Lecture is entitled: “Big Data in Smart Grids: Challenges and Opportunities”.

Abstract:
The issue of Big Data was introduced relatively recently (last 15 years) as the huge amounts of data became available through the space exploration, weather forecasting and medical biogenetic investigations. Social media and outlets such as Google, YouTube, and Facebook have also faced similar problems of handling huge data sets. The power systems are now experiencing huge amount of data obtained through field measurements as well. This talk focuses on the role of Big Data in managing and controlling future power systems, which will be characterized by “explosion” of data and unprecedented computational and communication capabilities to automatically extract the knowledge.
The focus is on different data sources that range from field measurements obtained through substation/feeder intelligent electronic devices such as Digital Protective Relays (DPRs), Digital Fault Recorders (DFRs), Phasor Measurement Units (PMUs), to other data sets obtained from specialized commercial and/or government/state databases: weather data of different types, lightning detection data, seismic data, fire detection data, electricity market data, etc. Due to the massive amount of such data (petabytes) available in real time and through historical records, processing and management of such data requires revisiting data analytics used to correlate data and extract features already developed in the Big Data industries such as banking, insurance and health care. This talk will point out the Big Data characteristics in the power industry where the temporal and spatial properties, as well as correlation to the power system and component models are necessary for an efficient data use.