Markov Chain-Based Modeling of Electric Vehicle Power Consumption

Nicole Ga Yan Ng

Abstract


The power consumed by recharging Electric Vehicles (EVs) can significantly impact the electric grid. EV charging can potentially cause a sharp increase in electricity demand at the end of every workday as EV owners begin charging their cars upon returning home. Having an accurate short-term forecast of EV power consumption can mitigate some of these impacts. One way of improving forecast accuracy is to develop a Markov Chain model of the underlying process. This research develops a Markov Chain model of hourly aggregate EV power consumption based on EV charging data from the Puget Sound region in Washington State during the year 2012.  A Markov Chain model is a discrete time stochastic process where the future evolution of the process is conditionally independent of the past given the present.  Aggregate EV charging power consumption exhibits strong diurnal trends—low in morning and increasing over the evening as drivers return home to recharge their vehicles. To represent these characteristics, the Markov Chain model was partitioned into three segments based upon the time of day. Within each segment, there are four distinct states, each corresponding to a range of power consumption. Transition probabilities between the states within each segment were computed. This model was used to simulate aggregate EV power consumption over a 24-hour period.


Keywords


Electric Vehicles; Load Profiles; Load Forecasting; Markov Simulation

Full Text: PDF

Refbacks

  • There are currently no refbacks.