solutions 2018 for integrating photovoltaicpanels …...of the xx power systemscomputation...

Solutions for IntegratingPhotovoltaic Panels into Low-voltage Distribution NetworksFrédéric OlivierAdvisor: Prof. Damien Ernst

2018

2

LV distribution network

• Distribution transformer• Electrical lines• Nodes (buses)• Houses• Smart meters

2

3

PV panels and LV distribution networks

Voltage too high

Disconnection of PV

Loss of renewable energy production

Loss of earnings

Slowing down the energy transition

4

How to increase the hosting capacity?

• Reinforcing the network• Balancing the phases of the network• Identifying the phases of the smart meters• Balancing the production and consumption on the three phases

• Actively managing the network• Controlling the active and reactive power flows inside the network

Part 1Phase Identification of Smart Meters

6

The phase identification problem

?

?

𝑎𝑏𝑐𝑛

Network

𝑟&

𝑟'

𝑠' 𝑡'

𝑠&

𝑡&

𝑟*

Cluster 𝒞&

Cluster 𝒞'

Cluster 𝒞*

6PSCC 2018

𝑎𝑏𝑐𝑛

Smar

t met

erSm

art m

eter

𝑟𝑠𝑡

𝑟

𝑛

𝑛

𝑀-&𝑀.&𝑀/&

𝑀-*

Measurements

7

Why is the phase information important?What are the existing solutions?What is the algorithm we propose?What is its performance ?

8

On the importance of phase identification

If you are a DSO

If you are a researcher

8

Change connection

phase of the customer

Reduce the imbalance in the network

Increase the hosting capacity

Phase-to-neutralactive and reactive

power measurements

Three-phasepower flow

Comparison between

measured and simulated voltages

9

Manual

Existing solutions

Automatic

9

Phase identifiers

Smart meters with PLC

k-means clustering

Graph theory

10

Contributions

1. Novel algorithm 1. Using the underlying structure of the network2. Using the advantages of both graph theory and correlation 3. Identifying the measurements that should be linked together and cluster

them. 2. Performance compared to those of a constrained k-means

clustering3. Tested on real measurements from a distribution network in

Belgium, in a variety of settings.

10

11

• Distance between two voltage measurements:• Pearson’s correlation

𝑑 𝑀1,𝑀3 = 1 − 𝑃𝐶 𝑀1,𝑀3 , ∀𝑙, 𝑖 ∈ ℐ

• Distance between a voltage measurement and a cluster

Δ 𝒞@,𝑀3 = min1∈𝒞D

𝑑(𝑀1,𝑀3)

Distances

11

12

Reference algorithmConstrained k-means Clustering

12

13

Proposed algorithmConstrained Multi-tree Clustering

13

Olivier,F.,Sutera,A.,Geurts,P.,Fonteneau,R.,&Ernst,D.(2018).Phaseidentificationofsmartmeters byclustering voltagemeasurements.In ProceedingsoftheXXPowerSystems ComputationConference (PSCC2018).

14

Test system

• Belgium LV distribution network• 5 feeders, star configuration 400 V• 79 three-phase smart meters• 2 single phase smart meters• Average phase-to-neutral voltage

measurements every minute

14

15

Results for the test setsDiscussions on the selection of the root

0

0,2

0,4

0,6

0,8

1

CKM CMT

Transformer House

Performance measureThe ratio between the measurements correctly identified and the total number of measurements

17

Influence of the voltage-averaging period

CMT

CKM

18

Influence of ratio single-phase – three-phase smart meters

CKM CMT

19

Influence of the number of smart meters

CKM CMT

Part 2 Active network management

21

Active network management

DesignstrategiestocontroltheactiveandreactivepowerinjectedbythePVpanelsintothedistributiongridsoasto

ensurethatthevoltagesstaywithintheirlimits

Objective

22

Active network management

Centralized controller• Model required• Communication infrastructure

Distributed controllers• Model-free• Little communication• Simpler logic

23

What are the key principles of the control scheme?What is the structure of the algorithm?How can we adapt the control scheme to three phases?What is its performance compared to other algorithms?

24

Key principles

• Distributed algorithm• No central entity • Each inverter applies locally the algorithm• No model required• First use reactive power• Curtail active power as a last resort• Limited communication in the form of a distress signal to pool the

resources

Olivier,F.,Aristidou, P.,Ernst,D.,&VanCutsem,T.(2016).Activemanagementoflow-voltagenetworksformitigating overvoltages duetophotovoltaic units. IEEETransactionsonSmartGrid, 7(2),926-936.

25

Assuming that the three phases are balanced

M M M M M M M M M M

N1 N2 N3 N4 N5 N6 N7 N8 N9 N10

Feeder bus

27

Mode𝑄1HI↓↑ (𝑉)

Mod

es o

f the

alg

orith

m

28

Mode 𝑄1HI↓↑ (𝑉)

Qf

Vtm

V2

V3

�Q(i)

max

⇣P (i)

set,t

⌘

Q(i)

max

⇣P (i)

set,t

⌘

V1

V4

Mode Q#"loc

(V )

29

Capability curve of PV inverters

Q

P

Inverter rated forreactive support at

full output(Curve A)

Inverter rated forreactive support atpartial output only

(Curve B)

30

Mode𝑄 ↑

Mode𝑄 →

Mode𝑄 ↓

Mode𝑄1HI↓↑ (𝑉)

Mode𝑃 →

Mode𝑃 ↓

Mode𝑃 ↑

31

Single feeder test system

M M M M M M M M M M

N4AB1

N4AB2

N4AB3

N4AB4

N4AB5

N4AB6

N4AB7

N4AB8

N4AB9

Node 4B

0.4 kV15 kV

Node 4ANode 4

MV LV

32

0

5

10

15

20

25

30

0 400 800 1200 1600 2000 2400 2800 3200

t (s)

(kW)

First try to restore active power

Second try to restore active power

PV N4AB1 PV N4AB2 PV N4AB3 PV N4AB4 PV N4AB5

PV N4AB6 PV N4AB7 PV N4AB8 PV N4AB9 PV NODE4B

-10

-8

-6

-4

-2

0

0 400 800 1200 1600 2000 2400 2800 3200

t (s)

(kVAr)


PV N4AB6 PV N4AB7 PV N4AB8 PV N4AB9 PV NODE4B

0

5

10

15

20

25

30

35

0 400 800 1200 1600 2000 2400 2800 3200

t (s)

(kW)


PV N4AB6 PV N4AB7 PV N4AB8 PV N4AB9 PV NODE4B 0.98

1

1.02

1.04

1.06

1.08

1.1

0 400 800 1200 1600 2000 2400 2800 3200

t (s)

(pu)

First try to restore active powerSecond try to restore active power

N4AB1 N4AB2 N4AB3 N4AB4 N4AB5

N4AB6 N4AB7 N4AB8 N4AB9 NODE4B

33

When the three phases are unbalanced

• Four groups of inverters• One for each phase and one for three-phase inverters

• Three distress signals• One per phase

• Each group applies the control scheme independently

34

Unbalanced test system

PV0

L0

PV1

L1

PV2

L2

PV3

L3

PV4

L4

PV5

L5

PV6

L6

PV7

L7

PV8

L8

PV9

L9

PV10

L10

PV11

L11

PV12

L12

PV13

L13

PV14

L14

PV15

L15

PV16

L16

PV17

L17

PV18

L18

External network

35

PV connection scenarios

PV0

L0

PV1

L1

PV2

L2

PV3

L3

PV4

L4

PV5

L5

PV6

L6

PV7

L7

PV8

L8

PV9

L9

PV10

L10

PV11

L11

PV12

L12

PV13

L13

PV14

L14

PV15

L15

PV16

L16

PV17

L17

PV18

L18

External network

Thre

e-ph

ase

Thre

e-ph

ase

Scenarios 3P

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

Thre

e-ph

ase

36


PV0

L0

PV1

L1

PV2

L2

PV3

L3

PV4

L4

PV5

L5

PV6

L6

PV7

L7

PV8

L8

PV9

L9

PV10

L10

PV11

L11

PV12

L12

PV13

L13

PV14

L14

PV15

L15

PV16

L16

PV17

L17

PV18

L18

External network

Thre

e-ph

ase

Thre

e-ph

aseA B C A B C A B C A B C A B C A B

Scenarios ABC

37


PV0

L0

PV1

L1

PV2

L2

PV3

L3

PV4

L4

PV5

L5

PV6

L6

PV7

L7

PV8

L8

PV9

L9

PV10

L10

PV11

L11

PV12

L12

PV13

L13

PV14

L14

PV15

L15

PV16

L16

PV17

L17

PV18

L18

External network

Thre

e-ph

ase

Thre

e-ph

aseA A B C A A B C A A B C A A B C A

Scenarios AABC

38


PV0

L0

PV1

L1

PV2

L2

PV3

L3

PV4

L4

PV5

L5

PV6

L6

PV7

L7

PV8

L8

PV9

L9

PV10

L10

PV11

L11

PV12

L12

PV13

L13

PV14

L14

PV15

L15

PV16

L16

PV17

L17

PV18

L18

External network

Thre

e-ph

ase

Thre

e-ph

aseA A A B C A A A B C A A A B C A B

Scenarios AAABC

39


PV0

L0

PV1

L1

PV2

L2

PV3

L3

PV4

L4

PV5

L5

PV6

L6

PV7

L7

PV8

L8

PV9

L9

PV10

L10

PV11

L11

PV12

L12

PV13

L13

PV14

L14

PV15

L15

PV16

L16

PV17

L17

PV18

L18

External network

Thre

e-ph

ase

Thre

e-ph

aseA A A A A A A A A A A A A A A A A

Scenarios AAA

40

Simulations for a sunny day

06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00

0

0.2

0.4

0.6

0.8

Time

[kW

/kW

p]

41

Comparison between algorithms

On-off• The one already implemented in commercial

inverters

Distributed• The one we propose

Optimal power flow• The one centrally optimizing the power

produced by the PV

42

Energy curtailed

3P ABC AABC AAABC AAA0

100

200

300

400

500

50

95

155

205

450

41 31

113

185

447

14 1.2 38

98

324

Ene

rgy

curt

aile

d[k

Wh]

On-off Distributed OPF

43

Reactive energy usage


10

20

30

40

50

60

0 0 0 0 0

31

22

28

25

22

30

23

51 53

35

Rea

ctiv

een

ergy

[kVA

rh]


44

Losses


10

15

20

25

30

19.3

18.8

17.5

18

16.5

20.1 22

.3

17.9

14.8

10.8

21.8

24.7

23.6

20.7

16.2

Los

ses

[kW

h]


45

Batteries

46

Local energy communities

• Definition

Electricitydistributionsystemcontainingloadsanddistributedenergyresources(suchasdistributedgenerators,storagedevices,orcontrollableloads),thatcanbeoperatedina

controlled,coordinatedway

47

Local energy communities

• Drivers• With a shared infrastructure between the members• Without a shared infrastructure

• Network operation• Energy market

Communities extend theperimeter ofself-consumption fromoneprosumer toseveral topool

productionandflexibility means

47

48

Electric vehicles

49

Conclusion – From research to industry

Question1Identifying thenetworktopology andthephasesofthesmartmeters

Question2Computing thehostingcapacity networks

Question3Preventive actions

Question4Correctiveactions

Solutions for IntegratingPhotovoltaic Panels into Low-voltage Distribution NetworksFrédéric OlivierAdvisor: Prof. Damien Ernst

2018

solutions 2018 for integrating photovoltaicpanels …...of the xx power systemscomputation...

Documents