Page 454 - Decision Making Applications in Modern Power Systems

Multistage and decentralized operations of Chapter | 16 413

population using different control approaches. Note that various services can
be provided by EVs, including peak shaving, regulation, voltage control, and
reserves, and many studies have quantified the benefits of smart charging
from various stakeholder perspectives [25].InRef. [25] the authors estimate
that smart charging will reduce the daily electricity costs of a plug-in hybrid
EV by $0.23. They also identify daily profits for the individual driver when
the charging of the vehicles can be regulated. The economic benefits of
fleets that participate in specific markets have also been extensively studied.
In Ref. [26] the authors use historical market data and charging data col-
lected from an EV located in a residential household to investigate financial
savings and peak demand reduction. The authors conclude that the peak EV
demand can be reduced by up to 56%.
In order to further explore the potential of EVs in reducing the total
microgrid operation cost, this chapter will present a comprehensive modeling
and decision-making framework for EVs under multiple DR markets in
California, such as proxy demand response (PDR), ancillary service, and
demand-based bid program (DBP) markets. It will provide combined strate-
gies to maximize the revenues via dynamic market participations.
Specifically, we analyze cost-saving performance of fleet EVs, considering
the EV fleet properties and market rules, including minimum consecutive
commitment, regulation market threshold, and the option to set baseline pro-
file. A multistage operation model is established for cost-effective microgrid
energy management, that is, day-ahead planning in the first stage and adap-
tive operations in the later stage(s). The preliminary results indicate that
these approaches can considerably reduce the total monthly energy cost
while satisfying the energy requests from both public and fleet EV drivers.
Asynchronous and decentralized algorithms [23,27] have been proposed
in the final control stage to allocate the prescheduled energy to individual
EVs, while preserving driver privacy and the overall robustness of the
decision-making framework. Heterogeneous energy requests, charging sche-
dules are considered in the framework, including arriving time and estimated
leaving. In the decentralized strategies, each EV agent will make asynchro-
nous local decisions, while coordinating with one centralized scheduler with
minimum amount of information exchange. In cases of communication
blackout or critical deviations of system states, the proposed decentralized
method can adaptively converge to the optimal solutions considering the new
inputs and updated system states. We investigate the robustness of the dis-
tribute algorithms with asynchronous coordination based on the alternating
direction method of multipliers (ADMM) [28].
The overall system architecture is illustrated in Section 16.2.Deterministic
problem formulation is provided in Section 16.3, followed by the more
implementable day-by-day operations in Section 16.4. Asynchronous and
decentralized algorithms are introduced in Section 16.5.Finally, Section 16.6
concludes this chapter with future work.