| dc.description.abstract | In the context of increasing concerns related to climate change, there has been a substantial
change to both the transportation and residential sectors. While the transportation sector
has previously been limited to conventional gasoline reliant vehicles, there has been a recent
shift towards the use of plug-in electric vehicles (PEVs), powered by electricity. This shift
will introduce new and different kinds of electric loads into the next generation of the
electrical distribution systems (EDSs). The PEV charging load is seen as an additional load
by EDSs; as such, an accurate and reliable model of these new loads is highly required. PEV
loads are not modeled as an additional conventional load. They are stochastic in nature
and impacted by the randomness of driver' behavior. There have also been significant
deployment of using solar photovoltaic (PV) and energy storage system with residential
premises, which has changed the role of residential customers from consumers to prosumers
(i.e. power producers and consumers). Therefore, it is necessary for utility companies to
accurately and realistically assess and quantify EDS asset conditions, with a consideration
of different penetrations of PEV with residential prosumers, and develop future necessary
infrastructure and policies in order to accommodate these and further changes.
This thesis focuses on developing a more realistic and accurate stochastic models of
PEV loads and PV generation and investigate the addition of these elements into EDS
assessment and planning. First, a new framework for modeling the stochastic nature of both
PEV loads and PV generation while considering the effect of the spatial-temporal (SAT)
characteristics of the driver' behavior, as well as solar irradiation and temperature, is
proposed. A trip chain, based on the Markov Chain Monte Carlo process, is developed to
properly model PEV daily driving activities and the PV uncertainty. Charging facilities are
assumed available at home, work, and fast-charging stations, having charging levels of 3.7
kW, 6.6 kW, and 50 kW, respectively. The proposed framework is examined, considering
the National Household Travel Survey global data, as well as the city of Buffalo and New
York state datasets. The impact of varying the penetration levels of PEV and PV resources
is also investigated. The present work strengthens the proposed models in the literature by
integrating the SAT characteristics of PEV charging demand into PV stochastic models.
Second, a new framework is proposed for evaluating and enhancing voltage quality, distribution transformer (DT) overload and aging, while considering residential prosumer
ownership of PEVs. The proposed work develops a probabilistic power flow in order to investigate the impact of the stochastic nature of PEVs, PVs, and conventional load. In
this work, the residential premises are modeled for supply through a detailed secondary
distribution system which is integrated as a part of the EDS. This work enhances the
existing research through the inclusion of PEV SAT charging activities into the
assessment models of DT overload and aging, voltage imbalance, and voltage deviation.
The proposed framework provides a more realistic life expectancy for DTs compared with
a simplified model in the literature. The results indicate that the use of the proposed
SAT-based approach has reduced DT lifetime to 6.30 years from 7.92 years for the same
PEV penetration level.
Finally, a multi-year community battery energy storage system (CBESS) planning
framework is proposed to accommodate a high PEV penetration level. A part from
considering the conventional load (residential and commercial), the framework considers
PEV charging demand based SAT approach, as well as PV options. Based on a
back-propagation algorithm, a heuristic methodology is developed to determine the
CBESS sites and sizes infrastructure plan while minimizing the total capital and
operation costs. The proposed heuristic approach starts from the terminal year of the
planning horizon, and propagate backward to the initial year, to determine the required
CBESS sites a long with their corresponding sizes. The proposed CBESS planning
framework is examined on the modified IEEE 123 primary distribution feeder and the
effect of using the proposed PEV SAT-based approach versus simplified model in the
literature is investigated. The results indicate that using the realistic proposed
SAT-based approach increases the required total CBESS investment budget from M$6.3
to M$6.7.
The SAT model was coded using MATLAB software environment. For the proposed
assessment framework, the three-phase power flow (PF) algorithm was programmed using
OpenDSS and be integrated with MATLAB and Python software environments. The
mathematical optimization model in the proposed planning framework was coded using
MATLAB software. | en |
