Considering impacts of PEVs in planning optimal hybrid systems

Considering impacts of PEVs in planning optimal hybrid systems

Hamed V. HAGHIM. A. GOLKARS. M. HAKIMI

[email protected]

Frankfurt (Germany), 6-9 June 2011

General Outline

Hybrid Active System with PEVs - Modeling

The Stochastic-Heuristic Algorithm

Results

Conclusion

Haghi – Iran – RIF Session 4 – Paper 0664

Main Topics

2



Studied Problem: Optimal Sizing - both the generation side and the load side are distributed

H OH

HH

H+ - H

Wind Generation

Converter/Controller

PEVs

Base Load

Electrolyzer H2 Tank Fuel Cell Reactor/Reformer

3


High penetration of stochastic energy flows spatially distributed throughout Microgrid

Variable generation (Wind, PV, etc) Variable load demand (PEVs, etc)

Representation of PEV load variations daily load shape locational displacement


General Outline

4


Strong dependence structure of load, generation and storage behavior over a year

Time dependence

Wind power autoregressive behavior impacts in planning storage (Markov chain fails for example)

Multivariate dependence

Correlation between load and generation


General Outline

5


Planning for net load capture both spatial and temporal diversity of PEV

Stochastic simulation (Monte Carlo approach) variability of PEV load on a multivariate modeling

Particle swarm optimization (PSO) optimization subroutine


General Outline – Scenario-based Optimization

6


General Outline



Results

Conclusion


Main Topics

7


Multi-objective optimization problem - weighted sum method


Hybrid Active System - Optimization

tan ( ) max

_ ( ) _ ( ) ( )

Minimize

with respect to

subject to 0

0

/

nx

n

n

n

k i

wt comv i fc conv i load i conv

NPC

N

N

E E

P P P

_ ( ) fc conv i fc fcP N P 8


By inserting impacts of PEVs on net load of system through a multivariate modeling

PEVs can cause a reversal of power flow through the distribution system

distribution network rely on a coincidence factor of loads for sizing all of the system’s components


Hybrid Active System with PEVs

9


Probability of coincident operation of PEVs is much higher

PEV controlled charging Actual demands are quite modest compared to

normal electricity demands Additional benefits as some kind of DSM controlled charging, with 20% randomness


Hybrid Active System with PEVs

10



PEVs Impact – Scenario-based Representation

0.2 0.4 0.6 0.8 10

200

400

600

800

Peruint Net Load With 20% Controlled PEV

Pro

babi

lity

0.2 0.4 0.6 0.8 10

100

200

300

400

500

600

Peruint Net Load without any PEV

Pro

babi

lity

11



Modeled planning dataset

500 1000 1500

500

1000

1500

500 1000 1500

500

1000

1500

500 1000 1500

500

1000

1500

0.75

0.8

0.85

0.9

0.95

1

0.75

0.8

0.85

0.9

0.95

1

-0.5

0

0.5

500 1000 1500

500

1000

1500

500 1000 1500

500

1000

1500

500 1000 1500

500

1000

1500

0.75

0.8

0.85

0.9

0.95

1

0.75

0.8

0.85

0.9

0.95

1

-0.5

0

0.5

500 1000 1500

500

1000

1500

500 1000 1500

500

1000

1500

500 1000 1500

500

1000

1500

0.75

0.8

0.85

0.9

0.95

1

0.75

0.8

0.85

0.9

0.95

1

-0.5

0

0.5

Net load with no PEV

Net load with 20% partially controlled PEV demand based on DSM indexes

wind speed

12


General Outline



Results

Conclusion


Main Topics

13



Scenario-based Optimization

Base load

Modeling PEV Load

Net Load ModelModeling wind

generation

Correcting for errors

Select scenario

Optimization

Save size set

No

Correlation analysis

All scenarios covered?

Scenarios, all together, represent long-term behaviour of PEV load and wind

Optimal set, considering uncertain variables space, to be analysed

14


General Outline



Results

Conclusion


Main Topics

15



Results – Benefits of Adding Controlled PEV

-0.2 0 0.20

1000

2000

3000

4000

Load Difference Distribution

Pro

babi

lity

-4 -2 0 2 4

x 107

0

10

20

30

X: 2.196e+006Y: 9.347

Pro

babi

lity

Optimal Costs Difference

Distributions of differences when the results of scenarios without PEV are subtracted from the results of scenarios with 20% PEV penetration

16



Differences – Optimal Sizes with and without PEVs

-400 -200 0 200 4000

5

10

15

20

25

30

X: 5.355Y: 12.4

WT Size Difference

Pro

babi

lity

-1000 0 10000

10

20

30

40

50

X: 82.01Y: 14.66

FC Sizes Difference

Pro

babi

lity

-2 -1 0 1 2

x 104

0

5

10

15

20

X: 258.7Y: 8.706

Pro

babi

lity

EL Sizes Difference-4000 -2000 0 2000 40000

5

10

15

20

25

X: -18.72Y: 9.967

Tank Size Difference

Pro

babili

ty

WT FC

EL HT

17



Results – Optimal Sizes Correlation

18



0 2000 4000 6000 8000 10000 120000

2

4x 10

7

Opt

imal

Cos

t

0

200

400

WT

Siz

e

0

5000

10000

EL

Siz

e

0

5000

HT

Siz

e

0 2000 4000 6000 8000 10000 120000

500

1000

FC

Siz

e

Simulated size sets for all 12,000 samples

19


General Outline



Results

Conclusion


Main Topics

20


A PSO-embedded stochastic simulation

Realistic modeling of the wind power and load demand data

A set of optimal sizes are obtained as final outputs which is then analyzed to provide a measure for making the optimal decision


Conclusions

21


A worthwhile optimal selection would be the mean values of all scenarios at the cost of reducing the reliability, but to an acceptable level most of the time

Sensitivity analysis of optimal sets

Other relationships could also be implied to help decision-maker


Conclusions

22


Contact:Hamed VALIZADEH HAGHIPhDc, P.EngFaculty of Electrical and Computer EngineeringK. N. Toosi University of Technology, Tehran 16315-1355, Iran+98 (21) 2793 [email protected]

Thank You!

23

Considering impacts of PEVs in planning optimal hybrid systems

Documents

net load of system

impacts of pevs

testfrankfurt germany

pevnet load

etcvariable load demand

behaviour of pev load

multivariate modeling

net load capture