Achieving High Reliability: Design Considerations for Microgrids29th November 2017 Samson Shih

Why Microgrid• Improve operational security & reliability (SS / transient)

• Adoption of renewables to reduce environmental impact and improve self sustainability, reduce dependencies of fuel supply

• Especially beneficial for remote and critical infrastructures (eg. hospitals, airports, military applications)

2

Microgrid

Image:DillonThompsonDesign

Microgrid Design Dilemma

0.940.960.981.001.021.041.06

0.940.960.981.001.021.041.06

Frequency(HighPenetration)

Frequency(Low/Med.Penetration)

Irradiance(Wm-2)

200400600800100012001400

11:00 11:05 11:10 11:15 11:20 11:25 11:30 11:35 11:40 11:45 11:50 11:55

vSustainabilityvs.systemstability

3

[Wm

-2]

[p.u.]

[p.u.]

Microgrid Design Dilemma

Irradiance

Low/Med.Penetration

HighPenetration

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

0

250

500

750

1000

1250

1500

1750

2000

0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00

PowerFactor

Irrad

iance(W

m-2)

>Sustainabilityvs.systemstability

vMoreeffectivepowerfromrenewablesvs.powerquality

4

Microgrid Design DilemmaRe

liabiltiyInde

x

InvestmentCost

>Sustainabilityvs.systemstability

>Moreeffectivepowerfromrenewablesvs.powerquality

vReliabilityvs.economy

5Graphextractedfrom:Mosteller R.,Budget-constrainedPowerSystemReliabilityOptimization

Improving Power Reliability

• Modification or New Design

• Power availability requirements

• Renewables intermittency and desired penetration

• Operation mode (islanded, grid-connected or both)

v BasicDesignConsiderations

6

Improving Power Reliability

• ‘Strength’/‘rigidity’ of microgrid– Electrical inertia

• Network protection requirements– Bidirectional power flow requires bidirectional protection consideration– Non-radial network further complicate the problem

• Load demand & type

• Placement / Sizing of ESS & required functionalities

>BasicDesignConsiderations

v FurtherDesignConsiderations

7

Improving Power Reliability

• Impact of geographic and temporal characteristics of the renewables on the scheduling & dynamic behavior

• Interconnection of multiple microgrids

>BasicDesignConsiderations

>FurtherDesignConsiderations

v OperationConsiderations

8

Case Study

HIGH ADOPTION OF RENEWABLES IN A TRADITIONAL DISTRIBUTIONLEVEL MICROGRID

Case Network

• Distributionlevel(6.6kV– 400V)

• Radialnetwork

• SinglePointofFailure

• Critical/Non-criticalLoads

• ModelledinPowerFactory

10

Design Requirements• High reliability and resiliency against internal and external failures

– Unplanned islanding– Faults on microgrid network– Potential communication failure

• Integrate as much PV as possible

• Ensure system remains stable

• Minimise fuel cost during islanded scenario

11

Network Modification

• PVInstallation

• LoopCableInstallation

• LVDGInstallation

• ESSInstallation

• AdjacentMGCableInstallation

• Distributedcontrolarchitecture

12

PV Penetration Study

TxfPG-A

TxfB-LVBTxfQ-LVQTxfH-LVHPVLimit

0.0

1.2

2.4

0%

20%

40%

0%

40%

80% Transformers

LineCables

TotalPV

13

[Loa

ding]

[Loa

ding]

[MW

p]

PV Penetration Study (Islanded)

TxfGen-B

TxfB-LVBTxfE-LVETxfH-LVHPVLimit

0.0

1.2

2.4

0%

5%

10%

0%

40%

80% Transformers

LineCables

TotalPV

14

[Loa

ding]

[Loa

ding]

[MW

p]

Component Failure RateComponent FailureRate,λ"#$%& (1/y) MTTR(h)

Cable(LV/HV) 0.02670/0.02017 7.60/5.13

Switchgear(LV/HV) 0.00949/0.01794 7.29/2.27

DGs 0.58269 25.74

Inverters 0.00482 26.00

Failure(f)Theterminationoftheabilityofacomponent/systemtoperformarequiredfunction

Failurerate(λ)Arithmeticaveragefailureperunitexposuretime

λ'()%& =+,-./012+321.45

orλ"#$%& =+,-./012

+321.45∗789:

MeanTimetoRepair(MTTR)Totaldowntimeforunscheduledmaintenance(excludinglogisticstime)foragivenperiod

𝑀𝑇𝑇𝑅 =𝑅>(?@ABC#𝑇D$BE)%#

IEEEStandard493-2007 15

Power Security Improvement

SAIFI SAIDI

Base 0.028813 0.181

Loop 0.021437 0.147

ESS 0.009490 0.069

SYSTEM AVERAGE INTERRUPTION FREQUENCY INDEX= ∑ +(A$EH(.(DJ)&A(C#%K@A#%%)LA#>�

�+(A$EH(.(DJ)&A(C#%&N#%O#>

SYSTEM AVERAGE INTERRUPTION DURATION INDEX= ∑ J)&A(C#%PB@)A#&(DK@A#%%)LAB(@�

�+(A$EH(.(DJ)&A(C#%&N#%O#>

IEEEStandard1366

16

Seamless Unplanned Islanding

ESSPower

PCCPower

Voltage

Frequency

0.0

37.5

75.0

0.9975

1.0000

1.0025

Islanding

0 5 10 15 20 25 30 35 40 45 50 55 60 s17

[p.u.]

[kW]

• TotalLoad(1.44MVA@0.92pf)– 0.50MVA@0.95pf– 0.20MVA@0.85pf– 0.25MVA@0.95pf– 0.50MVA@0.90pf

• DieselGenerator– 2x1MVA@0.80pf

• PVPenetrationTest– 0.3MWp

(18.75%Penetration)– 1.5MWp

(93.75%Penetration)

PV Penetration Study

18

*PenetrationisdefinedasratioofinstalledPVcapacitytoDGrating

Power Quality Analysis

0% 10% 20% 30% 40% 50% 60% 70%

-0.075 -0.05 -0.025 0 0.025 0.05 0.075

DGInter-secondRampRate

0%

20%

40%

60%

80%

100%

0.4 0.5 0.6 0.7 0.8 0.9 1

PowerFactor

HzMW/s

cosϕ

18.75%PV

56.25%PV 93.75%

PV 0%

5%

10%

15%

20%

25%

30%

49.75 49.85 49.95 50.05 50.15 50.25

Frequency

19

Increasing PV Penetration

• Network Stability:– Introduced network instability (frequency deviations)– Constant ramping of DGs– Deterioration of power factor at PCC and DG substation– Voltage rise during low loading

• Protection:– Change in flow of power in the network (due to PV and network topology

changes)– Disparity between available fault current in grid-connected and islanded

mode & changes in fault current flow

ProblemsFaced

20

• TotalLoad(1.44MVA@0.92pf)– 0.50MVA@0.95pf– 0.20MVA@0.85pf– 0.25MVA@0.95pf– 0.50MVA@0.90pf

• DieselGenerator– 2x1MVA@0.80pf

• ESS– 2MVA,1MWh

• PVPenetrationTest– 1.5MWp

(57.69%Penetration)– 2.7MWp

(103.84%Penetration)– 3.3MWp

(126.92%Penetration)

PV Penetration Study (ESS)

*PenetrationisdefinedasratioofinstalledPVcapacitytoDGrating

21

Power Quality Analysis with ESS

93.75%PV

57.69%PV

103.84%PV 126.92%

PV

0%

5%

10%

15%

20%

25%

30%

49.75 49.85 49.95 50.05 50.15 50.25

Frequency

0% 10% 20% 30% 40% 50% 60% 70%

-0.075 -0.05 -0.025 0 0.025 0.05 0.075

DGInter-secondRampRate

0%

20%

40%

60%

80%

100%

0.4 0.5 0.6 0.7 0.8 0.9 1

PowerFactor

HzMW/s

cosϕ22

Issues• Bi-directional&multiple

powerflow

• Changesinavailablefaultcurrents

Resultsin• Impartialdiscriminationof

faults

PotentialSolutions• Method1:

– Changetodifferentialprotection

– Directionalsensitivityforovercurrentrelays

– Overcurrentrelaysformsthebackupprotection

• Method2:– Makeuseof“loop”cableasa

backuptie-breaker

• Method3:– Adaptiveprotection

Protection Challenge

23

ESS Placement, Sizing & Scheduling

MULTIPLE MICROGRIDS

ESS – Multiple Microgrids

• Single ESS vs. Multiple ESS– Isolated MG with individual GESS– Connected MG with single GESS

v Placementlocation

MG1 MG2

MG3

ESS

ESS

ESSESS ESS

25

ESS – Multiple Microgrids

• Objective:– Minimise total fuel cost

• Using simple PV prediction based on ANN

• Considerations– Operations

• Load profile• PV prediction accuracy• DG fuel efficiency and ramp cost

• Simulation– 3 days with varying PV condition

>Placementlocation

v Sizing&Scheduling

• Levelised GESS cost• GESS efficiency• Power losses (transfer)

26

PV Prediction

• Consideration of time-series irradiance and weather data

• Training through backpropagation

>ArtificialNeuralNetworkbased

OutputsInputs

HiddenLayer(s)

backpropagation

27

PV Prediction

Predicted

Actual

PredictionError

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

01/11 02/11 03/11 04/11 05/11 06/11 07/11 08/11

Preidctio

nError/W

m-2

Irrad

iance(Predict&Actua

l)/Wm-2

v ArtificialNeuralNetworkbased

>PredictionResults

28

IndividualESS

• IndividualESSoneachMGinterconnectedtogetherreducestotalrequiredESScapacity&inverterrating

• Allowsforhigherreliability

AggregatedESS

• Loweroperatingcost

• AggregatenatureofloadandPVallowsforDGtooperateatmoreefficientpoint

ESS Study

MG1 MG2

MG3ESS

ESS

ESS

MG1 MG2

MG3ESSESS ESS

29

Summary• Determine RER capacity and variability

• Evaluate power reliability and quality requirements

• Provide network redundancy

• Provide power / energy redundancy

• Install corrective DERs

• Ensure network remains properly protected

30

End of Presentation