Intelligent Energy Management System Simulator for PHEVs ... · • US fleet’s 176 million light vehicles = power capacity of 19.5TW= 24 x power capacity of the electric generation

NC STATE UNIVERSITY

Intelligent Energy Management System Simulator for PHEVs at a Municipal Parking

Deck in a Smart Grid Environment

Preetika Kulshrestha, Student Member, IEEE, Lei Wang, Student Member, IEEE,

Mo-Yuen Chow, Fellow, IEEE and Srdjan Lukic, Member, IEEE

North Carolina State University

Raleigh, NC

NC STATE UNIVERSITY

Intelligent Energy Management System Simulator for PHEVs at a Municipal Parking Deck in a Smart Grid Environment 2

Outline

� Introduction

� System architecture

� Component description

� System simulator

� Sample system simulation

� Future work

NC STATE UNIVERSITY


Emergence of the “Smart Grid”

Optimization of power delivery - Capacity to deliver efficiently,

reliably and intelligently

Features

� Decentralization of control

� Services customized to user’s needs

� Use of energy efficient systems

� Rapid reconfiguration

NC STATE UNIVERSITY


Introduction to PHEVs

� HEV with larger battery pack

� Can be charged from standard wall outlet

� 40 mile all-electric range (Chevy Volt)

Benefits

� Reduction in GHG emissions

� Reduction on oil dependence

Time of Day

MW

Lo

ad

Off peak charging

Load leveling during extreme load events

Source: Prometheus Institute Time of Day

MW

Lo

ad

Off peak charging

Load leveling during extreme load events

Source: Prometheus Institute

• Load leveling

• Lower cost

Potential Synergistic

Relation

A cluster of vehicles is

a controllable load for the grid

NC STATE UNIVERSITY


Opportunities and Challenges

Opportunities

• US fleet’s 176 million light vehicles = power capacity of 19.5TW= 24 x power

capacity of the electric generation system.

• PHEV penetration by 2050 – 62% of the US fleet (EPRI prediction)

Challenges

• Potential load of 1000 cars => 4 MW load

• Potential dangers => Voltage instability and blackouts

• Infrastructure

• Need of an underlying framework to enable PHEV integration

A Solution

Intelligent Energy Management at a Municipal Parking Deck

NC STATE UNIVERSITY


Related Work

� J. Tomic and W. Kempton, “Using fleets of electric-drive vehicles for grid support”, J. Power Sources, vol. 168, issue 2, 2007

� M. Duvall and E. Knipping, “Environmental assessment of Plug-in Hybrid Electric vehicles”, EPRI, July 2007

� Hutson, G. K. Venayagamoorthy, K. A. Corzine, “Intelligent Scheduling of Hybrid and Electric Vehicle Storage Capacity in a Parking Lot for Profit Maximization in Grid Power Transactions”, in proc. IEEE Energy2030, Atlanta, GA, 2008

� S. B. Pollack et al, patent title “User interface and user control in a power aggregation system for distributed electric resources”, IPC8 Class: AG01R2106FI, USPC Class: 702 62

� GridPoint: http://www.gridpoint.com/

NC STATE UNIVERSITY


System Architecture

NC STATE UNIVERSITY


Information Flow

• Power Allocated

• Rate of charge

• Other control messages

iEMSOptimization

Pricing

Available power

Utility

• Time of availability

• Type of charge

• Pricing preferences

• Current state of charge

• Power consumed

• …..

Data Acquisition

(DSP, FPGA etc.)

Communication Medium: Wi-Fi, Bluetooth, Satellite ZigBee etc.

User Profile Entry

Vehicle 1

Vehicle 2

Vehicle 3

.

.

.

.

Vehicle n

NC STATE UNIVERSITY


System - Component Description

States

� Idle: There is no activity, the load may be waiting for a control action or it may have completed charging the battery.

� Data Acquisition: The data is acquired from the battery and the user.

� Communication: Data is communicated to controller.

� Charging: Battery charging is in progress.

� Error: There is an error in the system and system operation is halted until error is resolved. •A - Idle

•B - Data Acquisition

•C - Communication

•D - Charging

•E - StopState transitions

NC STATE UNIVERSITY


States

• Communication: Inform the loads of power allocated/listen for signals

• Optimize: Calculation of power allocation when:

• There is a change in utility power.

• A load has plugged-in/out.

• Periodically, after sampling the instantaneous power consumed by loads.


iEMS flow chart

A

B

Plug- in/ Update

Request,

Tsample

•A - Communicate

•B- OptimizeOptimize on power

subject to constraints

Data from LoadsTime of Availability

User Preferences

Initial State of

Charge

Information

from UtilityTotal Power

Available

Pricing

Allocate Power

to loads

Data Acquisition

Sample

instantaneous power

and SOC of load

T=Tsample or

request received

for update or new

load has plugged

in?

Yes

Initialize

State transitions

NC STATE UNIVERSITY



Functions

� Periodically inform the iEMS about the power available and pricing information

NC STATE UNIVERSITY


iEMS - System Simulator

powergui

Discrete,

Ts = 5 s.

iEMS

Input from car I

Input from car II

Input from car III

Input from car IV

Input from car V

Power

Car I

Car II

Car III

Car IV

Car V

Utility

Power

Station V

Input Rate and Allocated Power

Notify Plug in

Vehicle Information

Completed

VE

UpdateV

Station IV


Notify Plug in

Vehicle Information

Completed

VD

UpdateIV

Station III


Notify Plug in

Vehicle Information

Completed

VC

UpdateIII

Station II


Notify Plug in

Vehicle Information

Completed

VB

UpdateII

Station I


Notify Plug in

Vehicle Information

Completed

VA

UpdateI

Simulation Time

time

Clock

Continuous Dynamics

Discrete

Hybrid System

NC STATE UNIVERSITY


Objective function:

wi(k): the priority assigned for to vehicle i at time step k

Priorities are assigned based on capacity required

and time remaining

Sample System Simulation

( )= +∑∑max ( ) ( )i ip

j i

J k w k SoC k j

12:00 am11:00 am10:00 am9:00 am

Tplug out = 11:36 am

SoCin = 40%

D


SoCin = 10%

B

Tplug out = 1:04 pm

SoCin = 50%

E


SoCin = 70%

A

Tplug out = 10:17 amSoCin = 0%

C

0 0.5 1 1.5 2 2.5 3

x 104

-2000

0

2000

4000

6000

8000

10000

12000Plot of power consumption of five cars

Time (s)

Pow

er

(W)

PowerA

PowerB

PowerC

PowerD

PowerE

Total Power

Available Power

8:00 am 1:00 pm

NC STATE UNIVERSITY


Monte Carlo Simulation

Simulation Parameters

State of Charge at plug-in: Uniform random number between 10% and 75%

Time of Availability: Uniform random number between 0.5 and 2 hours

Time of Plug-in: Uniform random number between 0 and 2 hours

Simulation Run Time: 4 hours

Battery Capacity: Uniformly distributed between 6 Ah and 15 Ah

Number of times the simulation was run for each algorithm: 100

-0.2 0 0.2 0.4 0.6 0.8 1 1.20

0.05

0.1

0.15

0.2

0.25

0.3

0.35

SOC at Plug out

Perc

enta

ge o

f V

ehic

les (

%)

State of Charge at Plug out - Optimal Allocation for SoC Maximization

40

8.0%

32

6.4%

2

0.4%

Number of vehicles leaving with SoC 35% or lower

67.6%

(339)

69.6%

(348)

81.8%

(409)

Percentage of vehicles leaving with SoC 55% or higher

Equal Priority Allocation

Dynamic Priority Allocation

Optimal Allocation for SoC Maximization

NC STATE UNIVERSITY


Conclusions

� This paper proposes an iEMS for managing power at a parking deck

� System components, functions and behavior are outlined

� A simulator (test-bed) is developed to simulate the real world scenario

� The simulator will contribute towards evaluation of varied scenarios and iEMS algorithms

� Optimization on a chosen objective is formulated and simulation results presented

NC STATE UNIVERSITY


Future Work

• Exploration of different objectives for optimization

• Extension of the problem to multi-objective optimization and incorporation

of additional constraints

• Network in the Loop iEMS – performance evaluation with communication delay, packet drop, and signal strength

• Decision on optimal sampling time

• Extension of the concept to distributed control

• Real world implementation and demonstration of the iEMS

NC STATE UNIVERSITY


Acknowledgement

This work was partially supported by the National Science

Foundation (NSF) under Award Number EEC-08212121

NC STATE UNIVERSITY


THANK YOU!

Documents

Intelligent Energy Management System Simulator for PHEVs ... · • US fleet’s 176 million light vehicles = power capacity of 19.5TW= 24 x power capacity of the electric generation