research highlights
We exploited this low-rank property of PMU data matrix for multiple data management tasks. Take missing
data recovery as an example. Data losses can happen in an unpredictable way during the communication
between PMUs and the phasor data concentrator at the central operator. Losing measurements makes the
system unobservable and degrades the performance of the state estimator. We formulated the missing data
recovery problem as a low-rank matrix completion problem and developed computationally efficient data
recovery methods. Fig. 5 compares the recovery performance of multiple recovery methods. For the dataset
shown in Figs. 2~4, even when 30% of the measurements are lost at random locations, our methods can
accurately recovery the missing points. Using MATLAB running on a desktop with Intel i7-4770 @ 3.40GHz
and 12 GB DDR3 RAM, our developed online data recovery method took less than 1 millisecond to fill in the
missing points in each sampling instant. Hitachi America is interested in implementing our methods in their
prototype of new remedial action scheme (RAS) for Bonneville Power Administration (BPA).
Fun Fact Did you know that data loss and computer downtime can cost enterprises $1.7
trillion per year or the equivalent of nearly 50% of Germany’s GDP?
f
a
c
u
l
t
y
s
p
o
t
l
i
g
h
t
Professor Meng Wang is e the newest CURENT faculty member. She
says, “CURENT is a unique organization that brings people of diverse
backgrounds and expertise together. The intellectual communication
within CURRENT happens very naturally and smoothly and I really
enjoy the synergy of our research. It is exciting and fun to be a team
member here.”
Prof. Wang is also an assistant professor in the Department of Electrical, Computer and Systems Engineering at Rensselaer Polytechnic
Institute (RPI). She obtained her B.S. and M.S in Electrical Engineering
from Tsinghua University in 2005 and 2007, respectively, and earned
her PhD degree in Electrical and Computer Engineering from Cornell
University in August 2012. She was a postdoc research scholar at Duke
University before she joined RPI in Spring, 2013.
When asked how she came to the field of power systems in academia, Prof. Wang stated, “I like math and
physics and wanted to explore fundamental science. I also like to know how things work and want to do
something to contribute to everyday life. Electrical Engineering is a field that combines these two objectives perfectly. I can work on problems that have clear practical applications, while I still have the freedom
to explore fundamental theoretical developments beyond applications. For example, one project that our
group is currently working on is the data management and information extraction of large amounts of synchrophasor measurements in power systems. It is an important question in power system monitoring and
operation. The techniques we develop exploit low-dimensional models of signals in high-dimensional space.
These techniques are generic and thus can be applied to other fields like image processing, social network
analysis, etc. This type of exploration makes academia a natural choice for me. I very much enjoy the freedom of research. Another big bonus of academia life is that I get to work with many brilliant students, both
undergraduate and graduate students. I am happy to share my experiences and to learn with them.”
Prof. Wang’s current research focuses on high-dimensional data analysis and its application in power
system monitoring. Her boarder research interests include signal processing, optimization and networked
systems.
Welcome, Professor Wang!
newsletter
Spring
newsletter
Summer
2015
4