Abstract:
|
Electric vehicles are growing at a significant rate in the world and that makes it essential
for modern day electricity networks to be prepared for their integration. A common
approach of preparing the network for any kind of demand is to be able to predict or
estimate the same based on data and simulations using optimization techniques.
This work was aimed at the same in two distinct parts. In the first part, game theoretic
methods were tried to be applied to an existing multi agent probabilistic model
estimating net demand from electric vehicle. Owing to the complexity of the undertaking,
it was decided to only include a payoff based allocation of electric vehicle charging
scenarios to estimate electric vehicle demand which accounted for all scenarios rather
than all vehicles charging in a single scenario. In the second part, a smaller scenario of an
affluent household with two electric vehicles and typical mobility pattern was
formulated. Game theory solution concept of Nash Equilibrium was used to optimize the
charging of both electric vehicles over a week of usage.
The results from the first part, displayed an overall reduction in maximum loads while
there were certain shifts in loads observed as well. As an exercise without any inherent
optimization mechanism the overall results from this segment were inconclusive. The
results from the second part, demonstrated needs for charging the EVs shifting to offpeak
hours and charging of vehicles, a maximum of 1-2 times per week based on user
range anxiety, game theoretic competition and mobility needs. Further, savings from
charging at off-peak tariffs based on time of use electricity tariffs were also evaluated. |