IEEE Power & Energy Magazine - March/April 2018 - 108

in my view

Ralph Masiello

is IT ready for them?
future energy & distribution management systems

A HUgE gAME CHANgER FOR
utilities and possibly market operators
will be the need to gain visibility and potentially control of hundreds of thousands
of distributed energy resources (DERs).
DERs can, and will, include photovoltaic
(PV) generation, electric vehicles (EVs),
demand response (DR), combined heat
and power, storage, small-scale wind,
and other technologies. Some might
say that this is the province of a distributed energy management system
(DERMS), but a supervisory control
and data acquisition (SCADA) system
with applications by any other name is
an energy management system (EMS)/
distribution management system (DMS)/
DERMS or even a battery energy management system.
Widespread DERs create a number
of technical and process issues that will
stress today's EMS/DMS architectures.
First, the sheer point count will exceed
the design goals of most, if not all,
systems. This is not an unachievable
requirement as today's IT should have
no trouble with the kinds of limitations
that taxed the 16-b real-time SCADA
systems of 40-50 years ago. It may stress
some particular software designs, especially ones that carried over details
from those old systems. The basic SCADA
functions of get data, alarm data, and
store/display data are amenable to parallelization, so performance should also
be manageable, assuming the system
Digital Object Identifier 10.1109/MPE.2017.2779556
Date of publication: 21 February 2018

108

ieee power & energy magazine

software architecture in question will
allow parallelization, which is another
question altogether.
The scale and performance issues
begin to bite harder as you move up the
analytical food chain. Incorporating
all those DERs in a distribution network model for an advanced distribution management system could lead to
a million-node load flow with possible
optimization or management of those
hundreds of thousands of DERs. Again,
this can be in parallel if the network
is broken down at the transmissiondistribution boundary. This is actually
doable today; so again, this may not be
an issue. Integrating that model with
transmission analysis functions is another degree of difficulty, and this poses
mathematical/algorithmic challenges as
well as performance problems.
At the extreme, representing and
optimizing the DERs in a set of market
analytics is definitely not feasible today.
Today's mixed integer programming
(MIP) engines, depending upon the details of the system, are close to the edge
of what can be solved within a marketclearing window of an hour, and adding
one or two orders of magnitude to the
problem, could be the death knell. Aggregating the DERs helps, of course,
but already aggregators are lobbying
for the ability to aggregate resources
across multiple distribution circuits or
stations, which defeats the purpose from
the MIP engines' viewpoint.
The industry has always faced performance problems, however, and al-

ways managed to somehow keep the
problem requirements within the reach
of affordable architectures. It could be
possible that someone will crack the nut
of parallelizing the optimal power flow
and security constrained unit commitment problem; someone may figure out
how to use high-performance graphics
CPUs for these applications instead of
general-purpose chips. We have to have
faith that performance issues that appear impossible today will be solvable
within a decade, if not sooner.
A more challenging problem, in my
mind, is the need for a different paradigm for database management. At the
one extreme, new transmission construction is planned years in advance,
and the network models for new transmission are known months in advance.
At the transmission or independent
system operator level, databases can
be prepared, tested, and ready to go
when the new line is energized. Outages occur within the domain of existing switchgear and are handled by
topology processing. The number of
new changes is entirely manageable in
a manual process.
The distribution world is different as
changes happen daily and in sizable numbers. Utilities already put considerable effort into capturing planned and unplanned
work on a daily basis and keeping databases current within a day or so. This
requires coordination of DMSs, outage
management systems, asset management
(continued on p. 106)
march/april 2018

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - March/April 2018