IEEE Electrification - June 2019 - 36

0.4

Minimum: 0
Maximum: 1
Mean: 0.3972
Median: 0.3602

Probability

0.3

0.2

0.1

0.2

0.4
0.6
Flexibility

0.8

Figure 1. The probability distribution of the charging session
flexibility index.

TABLE 3. PG&E's E-19 demand charge and energy
charge rates.
Demand Charges

US$/kW

Time Period

Maximum peak
demand, summer

US$18.74

Noon-6:00 p.m.

Maximum part-peak US$5.23
demand, summer

8:30 a.m.-Noon
and 6:00-9:30 p.m.

Maximum demand, US$17.33
summer

Any time

Maximum part-peak US$0.13
demand, winter

8:30 a.m.-9:30 p.m.

Maximum demand, US$17.33
winter

Any time

Energy Charges

US$/kWh

Time Period

Peak, summer

US$0.14726 Noon-6:00 p.m.

Part-peak,
summer

US$0.10714 8:30 a.m.-Noon
and 6:00-9:30 p.m.

Off-peak,
summer

US$0.08057 Any time

Part-peak,
winter

US$0.10166 8:30 a.m.-9:30 p.m.

Off-peak, winter

US$0.08717 Any time

Electric Bill (US$)

10,000
8,000

TOU Market Participation
The first study includes load shifting and cost reduction through smart charging and scheduling optimization under TOU prices only. Table 3 summarizes
demand charge and energy charge rates according to
PG&E tariffs. As shown in Table 3, the energy charge
and demand charge rates in winter are lower than
those in summer. As a result, the test site's actual total
monthly costs for energy charges in winter were slightly lower than those for summer, indicated by the blue
bars in Figure 2. In addition, the total monthly demand
charges in winter were considerably lower than those
in summer, shown in Figure 2 by the orange bars.
Two separate optimization problems with different
cost objectives are defined as use cases. The first cost
objective includes only the total monthly energy charge;
the second includes the summation of total monthly
energy and demand charge. The load profile of two consecutiive days with the maximum monthly demand in
June is shown in Figure 3, with (a) and (b) corresponding
to minimization of the first and second cost objectives,
respectively. In Figure 3(a), a large portion of the EV load
has been shifted to time periods with lower energy
charge rates; in Figure 3(b), the monthly load peak
around 10 a.m. is shaved to minimize the monthly
demand charge.

6,000
4,000

Ancillary Service Market Integration

2,000

In this section, another capability is enabled that allows
EVs to participate in ancillary service markets in addition
to the TOU market. The day-ahead prices for ancillary service regulation up and down are collected from CAISO's
Open Access Same-Time Information System (OASIS) for
two years. As shown in Figure 4, the EV load profile
becomes spikier when supporting ancillary service market

0
Months
Energy Charge Cost

Demand Charge Cost

Figure 2. The actual total monthly cost of energy and demand charges
at the test site for two years starting from January, in chronological order.

The deferability of the EV charging load can be denoted
by the flexibility index, defined by the actual charging
time in each session divided by the total plug-in time. The
distribution of the flexibility indices for all collected charging sessions in this study is shown in Figure 1.
In this case study, the charging scheduling problem
for EVs is formulated as an optimization problem based
on mathematical models of each market. The mathematical model includes the physical and market constraints imposed on the participation of EVs in the
electricity market (e.g., state of charge and power and
energy capacity) as well as an objective function that
aims to maximize EV revenue and minimize the testing
facility electricity bill.
This section demonstrates the optimization results of
EVs participating in the demand-response programs and
ancillary service markets described in the previous section. The benefit analysis is conducted based on real data
from the testing facility as well as pricing information
from PG&E and CAISO.

I E E E E l e c t r i f i cati o n M agaz ine / J UN E 2019

IEEE Electrification - June 2019

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