IEEE Electrification - June 2019 - 37

TABLE 4. Monthly revenue from the
Power (kW)

Part-Peak Region

Shifted Load

100
50

150

12:00

24:00
(a)

12:00

24:00

Shaved Peak

100
50
00:00

12:00

24:00
(b)

12:00

24:00

Baseload (kW)
Optimized EV Load (kW)
Original Building Load (kW)

Figure 3. An example load profile showing the smart-charge EV load
shifting to minimize electric costs for two consecutive days. In June:
(a) minimization of the first cost objective and (b) minimization of the
second cost objective.

Power (kW)

regulation market.

150

00:00
Power (kW)

Peak Region

120
110
100
90
80
70
60
50
40
30
00:00

Year

Ancillary
Service
Month Revenue

Year

Ancillary
Service
Month Revenue

2015

US$90.38

2016

US$74.53

2015

US$66.42

2016

US$66.33

2015

US$73.88

2016

US$90.37

2015

US$95.73

2016

US$71.75

2015

US$78.09

2016

US$68.62

2015

US$67.75

2016

US$95.70

2015

US$76.57

2016

US$69.22

2015

US$68.51

2016

US$77.28

2015

US$78.16

2016

US$61.50

2015

US$80.48

2016

US$78.24

2015

US$63.24

2016

US$79.70

2015

US$98.11

2016

US$115.14

The monthly revenue results are simulated by applying
EV management strategies over two selected years, as
shown in Table 4. The highest monthly revenue is calculated as US$115 for December 2016, and the lowest revenue
obtained is US$61.50 for September 2016.

PDR Market Participation
12:00

24:00

12:00

24:00

Baseload (kW)
Optimized EV Load (kW)
Original Building Load (kW)
Figure 4. An example load profile for two days of ancillary service
regulation up and down market participation.

participation, because optimization tends to change
power consumption when a high regulation price signal is
anticipated. However, with the possibility that the
increased EV charging load causes new demand peaks
and thus high demand charges, optimization evaluates
the tradeoff globally on a monthly basis. As depicted in
Figure 4, the adjusted power consumption profiles due to
participation in the regulation market are constrained so
as not to exceed the monthly demand peaks set by the
TOU-based optimization.
In this study, the duration of each regulation commitment during optimization is 15 min, within which the
actual regulation signals are dispatched every 4 s. A
15-min interval is considered as the finest resolution for
EV control. In addition, both regulation up and regulation
down bids are allowed in the same time periods.

The same analysis is conducted for PDR market integration
by combining the TOU charges with the revenues from the
PDR markets as the objective function. Figure 5(a) shows the
day-ahead PDR price for two consecutive days. In California,
CAISO requires that each participant in the PDR market have
at least a 1-h commitment. Therefore, additional constraints
must be applied on the optimization problem to guarantee
that each market participation event spans more than four
time steps (15 min per step). As shown in Figure 5(b), the
green curve indicates the actual EV power consumption profile, while the red curve represents the virtual sell power of
the EV aggregator given price signals from the PDR market.
Note that the total energy consumption according to the
actual power consumption profile is equal to that calculated
from the baseline profile. In addition, in this problem, the
participation of EVs in the PDR market is modeled with a
binary variable, which reflects that the EV aggregator does
not have to stay in the market for the entire day and can
plan to participate in the market when PDR prices are economically desirable. Table 5 summarizes the calculated
monthly revenues from PDR markets.

DBP Participation
DBP market integration has many similarities with PDR
market participation. The main difference is that the DBP
IEEE Elec trific ation Magazine / J UNE 2 0 1 9

IEEE Electrification - June 2019

Table of Contents for the Digital Edition of IEEE Electrification - June 2019