Moving grid operations from minutes to seconds
August 14, 2003 -- about 20 hours before the blackout.
In the last century, the electric power grid has grown from a
system that served one square mile in New York into a highlycomplex
interconnected system that serves all of North America.
Initially, individual local systems would connect to each
other to share resources and increase reliability. Today, utilities
rely upon that same interconnectedness to save money by
shipping large amounts of power over long distances, pushing
the system closer to its limits more and more often.
At Pacific Northwest National Laboratory, researchers are
exploring how changes in the way the nation's electrical grid is
operated can improve its reliability and effectiveness—lowering
costs and lessening
impacts. The focus
is on developing
real-time tools and
Rob Pratt, who leads
Initiative, there is a
need for grid operators to see beyond the individual control areas
within their purview. If operators understand what is happening
on neighboring systems, they could more quickly recognize and
address the potential impacts to the system they operate.
This need for "situational awareness" was a key finding
following the East Coast blackout of 2003, where events
leading up to the blackout could have been detected and
corrected before they spun out of control. Once started, the
blackout spread from Ohio to New York in only nine seconds.
PNNL's EIO Initiative focuses on improving situational
awareness by collecting more and richer data from a broader
area and developing tools that allow timely analyses and
therefore quick and appropriate action.
August 15, 2003 -- about 7 hours after the blackout.
For example, data collected by a conventional SCADA
(Supervisory Control and Data Acquisition) system currently
are refreshed only every four seconds. It can be another four
minutes before the data are analyzed and results are available
to operators. "Operators don't know for four minutes or more
if the news is good or bad," Pratt said. However, a new and
more precise type of data called phasor data that are timestamped
and synchronized could be shared with operators
across the system 30 times per second.But as we all know, having data is not enough," Pratt said.
"We will eventually have thousands of these data points
available, but we need new technology, including advanced
analysis and visualization tools, to know where and when to
look—and what we should be looking for."
Researchers are applying data-intensive computing to the
process used to reconcile data coming in from the grid and the
underlying fundamental physics of its operation. Called "state
estimation," it takes a few minutes to develop a comprehensive
picture of the grid that can be used also for contingency analysis
to determine likely outcomes and potential next steps.
the architecture and
the algorithms and
them to the problem
at hand, we've
already sped up this
process by an order
of magnitude," Pratt
said. Right now,
capabilities of DOE's
Office of Science for these breakthroughs, but Pratt said it won't
be long before this kind of processing will become available in
mainstream systems for utilities.
The EIO Initiative also is focused on research to help
understand the dynamics of the electric grid and build those
properties into improved real-time models. Today's operational
models assume that the grid is in a steady state, ignoring
oscillatory behavior that can potentially bring it down. Current
studies of the grid's dynamic stability are so slow that they're
typically done offline and can't be incorporated into control
processes. Without precise models, the grid is often operated with
more margin than needed, which results in real economic impact.
The Initiative centerpiece is the Electricity Infrastructure
Operations Center, a fully capable grid control center that
serves as a platform for grid research and development, testing
and training (see related story, next page).
As demand for electricity continues to grow, the existing
transmission and distribution system will continue to get pushed
to its limits, threatening its reliability. "We can't afford to goldplate
the grid to make it reliable," Pratt said. "Instead, we're
focused on how to safely live close to the edge—knowing with
confidence where that edge really is—while keeping the lights on."