IEEE Electrification Magazine - September 2013 - 51

could result in a loss of revenue estimated at US$80/kWh
(value of lost load), which covers the replacement cost of
damaged equipment, and personnel and administrative
cost of restoring and sustaining research and education at
IIT. Once the real-time price exceeds 6-8 cents/kWh (marginal cost of microgrid generation), the campus load is supplied by the local microgrid generation. The master controller uses a security-constrained unit commitment to calculate the day-ahead optimal operation of microgrid. The
optimal hourly solution includes the dispatch of the
microgrid generation and renewable energy resources,
exchanges with the utility grid, charge/discharge schedule
of the battery storage unit, and adjustments to set points of
building loads.
To perform the tertiary control, the master -controller
procures the day-ahead forecasts for building loads and
renewable energy resources. The forecasted price of electricity is procured by ComEd. The forecasted values are
calculated based on the historical data and forecasted
weather data using nonlinear regression methods. The
integration of renewable energy generation in microgrids
will reduce carbon footprints while decreasing the cost of
supplying the campus load. The drawback of integrating
renewable technologies is the variability of their generation portfolio. To overcome this challenge in the microgrid,
several approaches are used including the coordination of
dispatch with the utility grid and hourly demand
response. The economical operation of microgrid is implemented by two master controller functions, which are
discussed below.

Unit Commitment and Economical Dispatch

f/V
B
Secondary
Control
Rated

Pri
m
Co ary
ntro
l

No Change
in Dispatch

P/Q

Figure 17. The secondary control in DER units.

and 24.5 cents/kWh at hours 16 and 17, respectively. The
cost of supplying the campus energy on this day was
US$15,524.
The real-time optimization is based on real-time information, such as the price of electricity, campus load,
renewable energy generation, and the topology of the
campus microgrid including the state of Vista switches
and cables. The master controller will perform the campus
energy management by procuring the optimal 15-min
economical demand response and the dispatch and commitment of campus generation.

Economical Demand Response
The master controller will adjust shiftable building load
schedules to calculate optimal generation schedules.
Shiftable loads can often be served at delayed hours without jeopardizing the convenience of campus residents.
Moreover, the tertiary control will schedule the charging/-
discharging sequence of battery storage to optimize the
supply of campus load with respect to the utility price of
electricity. In island mode, the microgrid load is supplied
by dispatchable DER units, which respond according to
their droop characteristics using primary and secondary
control scheme. The tertiary control would also set the
optimal operating point of dispatchable DER units. The
nondispatchable DER units including solar PV and wind
turbine units will not respond to deviations in real and
reactive campus loads. In Figure 22, the master controller
would apply demand response through tertiary control

To ensure the economical operation of the microgrid, the
master controller performs unit commitment and economical dispatch in island and grid-connected modes to
procure the -optimal generation scheduling of DER units as
well as the utility grid dispatch. In grid-connected mode,
the microgrid load is compensated by adjusting the power
generation exchange with the utility grid. Here, the primary and secondary controls of DER units will not respond to
disturbances, as the microgrid voltage and frequency are
set by the utility grid. Figure 21 shows the day-ahead
hourly control signals provided by the master controller
for supplying the campus load on 17 July 2012. On this
day, the campus reached its annual
peak load of 11.263 MW at hour 15.
f
The master controller dispatched
V
∆V = mq . ∆Q
∆f = mp . ∆P
the microgrid generation once the
electricity price was higher than 6
(f1, P1)
(V1, Q1)
(Vrated, Qrated)
cents/kWh. In Figure 21, the battery
(frated, Prated)
∆V
∆f
mp
mq
storage was charged when the elec∆P
∆Q
tricity price was lowered to 2.8 and
2.7 cents/kWh at hours 4 and 5,
P
(a)
(b)
respectively. In addition, the battery
storage was discharged as the price of
the electricity was increased to 22.4 Figure 18. The (a) frequency and (b) voltage droop characteristics of a DER unit.

IEEE Elec trific ation Magazine / s ep t em be r 2 0 1 3

Table of Contents for the Digital Edition of IEEE Electrification Magazine - September 2013