Environmental Engineering Reference
In-Depth Information
interval, all plugged PHEVs are aggregated per node, and this allows the model
to determine maximum rate of charge, the current SOC as well as the maximum
SOC of the batteries. As a result, the output data gives the spatial and temporal load
flexibility of the PHEV fleet. Note that in the ABM simulation the car battery is not
charged and travels about 30 mi per day; hence, the driving range of the vehicle is not
an obstacle to realise all its daily activities powered by the vehicle's battery only.
The ABM has been implemented in Repast Simphony building on top of the
open-source Repast City model [241]. Existing micro-simulation traffic models and
other agent-based models of driver behaviour may be more advanced and provide
more realistic traffic simulation (
e.g.
including congestion or traffic light control),
but instead here it was decided to start with a relatively simple model that focuses
solely on the temporal and spatial distribution of demand for electricity based on the
activities of the drivers. The simulation can be run for one or multiple days, during
which the activities of the agents change (
e.g.
fewer people have to go to work on a
weekend and people might come home later on a Friday than on a Thursday night).
The ABM produces activity logs for the individual agents and statistics on total
distances driven and energy consumed (for each PHEV as well as for the entire pop-
ulation) in addition to the results on load flexibility as mentioned above. The outputs
are then fed into the TCOPF model for optimisation of PHEV charging.
7.2.2 PHEV optimal power flow formulation
In this section a simplified version of the TCOPF presented in Chapter 5 is used
in order to incorporate the data provided by the agent-based simulation to resolve
PHEV charging for ideal integration to the grid. Taking a daily load profile, the solver
analyses the state of the electrical network in its multiple nodes as well as the energy
required by the PHEV fleet. After this data is processed the TCOPF can calculate
when it is convenient to command EV charging according to the objective function
in place. The objective function can, for example, focus on minimising the costs or
the CO
2
emissions for the generation of the electricity required to charge the PHEVs.
Thus, the TCOPF can focus on minimising or maximising a non-linear objective
function over multiple period intervals which are restrained by a set of non-linear
constraints.
The multi-objective function example presented here calculates the optimal
charging of PHEVs by trying to consider the key drivers that look to influence how
this technology can be charged efficiently from day to day. These drivers are repre-
sented in monetary terms as follows [242]:
Day-ahead spot electricity market prices;
●
Carbon cost from charging with carbon-intensive electricity;
●
Network operating costs for energy delivery.
●
The objective function focuses on minimising both energy and emission costs
incurred from charging PHEVs. Total costs are calculated from price signals given
by spot and carbon markets plus grid delivery costs, as seen in (7.1). The solver is
holistic and unbiased, thus giving no preference to any particular party either EV
Search WWH ::
Custom Search