Spatial and Temporal Model of Electric Vehicle Charging Demand

Spatial and Temporal Model of Electric Vehicle Charging Demand

Presented by: Hao Liang

2012.5.31Broadband Communications Research (BBCR) Lab

Smart Grid Research Group

Reference: S. Bae and A. Kwasinski, “Spatial and temporal model of electric vehicle charging demand,” IEEE Transactions on Smart Grid (SI on Transportation Electrification and Vehicle-to-Grid Applications),vol. 3, no. 1, pp. 394-403, Mar. 2012.

Outline

2

• Introduction

• Highway Model Description

• Model Formulations

• Numerical Example and Discussions

• Conclusions

Broadband Communications Research (BBCR) LabSmart Grid Research Group

Introduction

3


4


Transportation ElectrificationPlug-in Electric Vehicle (PEV)

Plug-in Hybrid Electric Vehicle (PHEV)

BMW Electric Mini Cooper

Nissan Leaf

Tesla Model S

Toyota Prius

Chevrolet Volt

5


• Reduce gasoline consumption

• Decrease greenhouse gas emission

• Reduce energy bill of vehicle owner ?

• Increase profit of vehicle manufacturer ?

Benefits of Transportation Electrification

6


Challenges in Transportation Electrification

• Stress on the power system during the peak time (temporal changing nature)

• Limitations in distribution transformers, which are aggravated by the uneven PEV and PHEV penetration favoring high-income areas, e.g., downtown vs. rural (spatial changing nature)

7


Main Contributions of the Paper

• Present a mathematical model of rapid charging station’s electricity demand which may vary both spatially and temporally

• The arrival rate of discharged electric vehicles at a specific charging station is anticipated by the fluid traffic model (modified based on highway Poisson-arrival-location model (PALM))

• EV charging demand is predicted by the M/M/s queueing theory (Poisson vehicle arrival, exponential vehicle charging time, s identical charging pumps)

Highway Model Description

8


9


Traffic Model for Semi-infinite, One-way, Single-lane Freeway

• x – Distance along the highway from the spatial origin which is the beginning point of the highway

• v(x, t) – Velocity field of each vehicle • Charging stations are located on each exit or entrance • Vehicle arrives at each entrance with the Poisson distribution

10


Extensions of the Traffic Model

Multiple-lane Highway Model– Combine basic highway models in which vehicles have different velocities

Bidirectional Highway Model– Eastbound ve(x, t) ≥ 0 – Westbound vw(x, t) ≤ 0

Elaborate Highway Network Model– Superimpose groups of multiple-lane and bidirectional highways

Model Formulations

11


12


Assumptions

• The battery for an already charged vehicle entering the highway has a full state-of-charge (e.g., due to the night-time charging at home)

• Fully charged batteries can last for the entire range of the trip. Hence, the user of an EV that enters the highway fully charged may exit the highway not because the batteries are discharged but rather because he/she may require to rest

=> This study focuses on the discharged EV’s user who forgets to charge it at night, thus requiring visiting a charging station on a highway (consider the highway as a prison or jail)

13


Definitions of Variables

– The number of discharged EVs remaining in the interval (0, x] at time t

– The number of discharged EVs that have already passed through the position x before time t

– The number of discharged EVs that have entered the highway along the interval (0, x] before time t

– The number of discharged EVs that have exited the highway along the interval (0, x] before time t

14


Representations of Variables

Discharged EVs leaving the system (i.e., ) are divided into:

1) Permanently depart from the highway and recharge at their final destinations (e.g., arrive home)

2) Temporarily leave the highway in order to recharge their batteries at the highway exit charging station, and will return to the highway after recharging their batteries

15


Deterministic Fluid Dynamic Model Conservation Equation (Foundation)

Density of Discharged Vehicles, veh/km (Will show: this is the only unknown variable)

Traffic Flow of Discharged Vehicles, veh/min

Densities of Discharged Vehicles Entering or Leaving the Highway

16


Deterministic Fluid Dynamic Model (Cont’d)

In traffic theory, traffic flow can be defined as the multiplication of a traffic density by a vehicle’s velocity

Known (or Measurable)

17


– All discharged EVs actually arriving at the highway in the interval (0, x] before time t. Equivalent to

– All discharged EVs permanently departing from the highway in the interval (0, x] before time t

– All discharged EVs temporarily leaving the highway in order to recharge their batteries in the interval (0, x] before time t


18



These rate densities can be identified with the actual arrival rate (i.e., i(t)) and the permanent departure rate (i.e., βi(t)) of discharged EVs at the ith highway entrance/exit and at time t typically measured in the number of vehicles per minute

The condition which discharged vehicles can only arrive at and depart from the highway through entrances/exits:

Dirac Delta Function

Distance from the Spatial Origin to the ith highway entrance/exit

Known (or Measurable)

19


Deterministic Fluid Dynamic Model (Cont’d) Assume that discharged EVs will return to the highway immediately after finishing to recharge their batteries

Temporarily Departing Rate per Minute

Charging Completion Rate per Minute

Average Charging Power per Vehicle

Average Recharged SOC per Vehicle

1/60

20



By substituting the equations into the conservation equation, we have

An ordinary differential equation (ODE) which can be solved with numerical methods without many difficulties, given certain boundary condition

The arrival rate of discharged EVs at the ith highway charging station

21

EVs’ Charging Demand by the M/M/s Queueing Theory

• The queueing system is stable if and only if the occupation rate (ρ) of charging pumps is less than 1

• The minimum number of charging pumps


22


EVs’ Charging Demand by the M/M/s Queueing Theory (Cont’d)

• The expected number of busy charging pumps

• The power demand of charging station

23


Stochastic Model

• The purpose of the stochastic model presented here is to identify the expected value of the stochastic EV charging demand

Analogous to the deterministic model

Numerical Example and Discussions

24


25


Basic Highway Model for a Numerical Example (1/2)

• Charging power: 70 kW (level 3 charging station )• Battery capacity: 8.6 kWh• Average charge per vehicle at the highway charging station is 4 kWh

which is about 50% of the battery capacity (3.4 min to recharge)

26


Basic Highway Model for a Numerical Example (2/2)

• Velocity fields of vehicles on the highway is 1 km/min for all x ≥ 0 when t ≤ 40 or t > 55 min. During the time interval (40, 55] min corresponding to rush hour, given by

27


Simulated Mean Density of Discharged Vehicles

Non-Rush Hour Rush Hour

28


Simulated Mean Traffic Flow of Discharged Vehicles


29


Expected Number of Charging Pumps in Service and Expected Charging Demand


30


Application: Sizing the Energy Storage System

Off-Peak Time Peak Time

Conclusions

31


32


• Highway EV model is based on the fluid traffic model

• EV charging demand is calculated with the arrival rate of discharged EVs by the M/M/s queueing theory

• Application I: Sizing the energy storage system

• Application II: Distribution system planning

- Traditionally, the planning focuses on local demand

- With EVs, demand may move from another utility into the area of the utility under consideration => coordination among neighboring utilities is necessary

Thank you!

33


Spatial and Temporal Model of Electric Vehicle Charging Demand

Documents

Transcript of Spatial and Temporal Model of Electric Vehicle Charging Demand