ACMD Seminar: Data-driven fractal modeling for blackout and malicious threat detection

Edmond Jonckheere
Professor, Ming Hsieh Department of Electrical Engineering, University of Southern California

Tuesday, January 24, 2017 15:00 - 16:00
Building 225, Room B111
Gaithersburg

Tuesday, January 24, 2017 13:00 - 14:00
Room 1-4058
Boulder

Slides

Abstract: This talk addresses power grid stability in a broad sense that involves the three main variables—voltage, frequency, and phase angles—with a view towards anticipating voltage collapse, blackout, even malware attacks. Instead of relying on "first principles," dubious at capturing the complex large-scale property of the loads and their interaction with the generators, we take a critical look at Phasor Measurement Units (PMU) data, from which it follows that the embedded data has long range dependence. This calls for a re-examination of the classical load models and the quest for models that could explain the strange behavior of the PMU data.

Preliminary investigations on the Texas ERCOT and the EPFL of Lausanne (Switzerland) have demonstrated scaling exponents alpha consistent with long range dependence. Next, data collected on the Indian grid before the 2012 voltage collapse shows shift in the frequency data, subtle at first sight, but much more pronounced as an increase of the Hurst exponent and the AR(1) coefficient before the actual collapse happened. This opens the road to anticipate and hopefully develop countermeasures to an imminent instability on the grid.

In view of this potential, it is imperative to understand why the PMU data is fractal.  Some early data taken on a Scandinavian microgrid revealed a strange dependency of the active and reactive power consumed by a load on the frequency to a noninteger exponent. The "mathware" of the presentation will be devoted to a rigorous re-interpretation of this frequency noninteger exponent as a fractional time derivative. But then the last question is how does the power grid generate fractional dynamics? It will be shown that a load aggregation effect—in mathematical terms, the strong connectivity of the graph of the hidden feedback structure of the grid—is a contributing factor.

Speaker Bio: Edmond Jonckheere received his "Electrical Engineer" degree from the University of Louvain, Belgium, in 1973. From 1973 to 1975, he was a Research Fellow of the European Space Agency and in 1975 he received the Dr.-Eng. degree in Aerospace Engineering from the Université Paul Sabatier, Toulouse, France. In 1978, he received the Ph.D. degree in Electrical Engineering from the University of Southern California, Los Angeles. In 1979, he was with the Philips Research Laboratory, Brussels, Belgium. In 1980, he returned to the University of Southern California, where he is currently a Professor of Electrical Engineering and Mathematics, a member of the Center for Applied Mathematical Sciences (CAMS) and a member of the Center for Quantum Information Science and Technology (CQUIST). Dr. Jonckheere has had short term visiting appointments with the Max-Planck-Institute, Gottingen, Germany, the Australian National University, Canberra, Australia, the University of Namur, Namur, Belgium, Cardiff University, Wales, UK, and Swansea University, Wales, UK. He has had consulting affiliations with Memorial Medical Center of Long Beach, Lockheed-Martin, the Aerospace Corporation, and Honeywell. Dr. Jonckheere is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE) for "Contribution to the spectral theory of linear-quadratic and H-Infinity problems." He was elected Life Fellow in 2016. His current research interests include conventional versus quantum control, adiabatic quantum computations, wireless networking, and the power grid.

Note: Visitors from outside NIST must contact Cathy Graham; (301) 975-3800; at least 24 hours in advance.

Part of the ACMD Seminar Series.