EPE '14 ECCE Europe Tutorial #2 - Challenges on the Road to Future High-Voltage Multi-Terminal DC Networks

Challenges on the Road to Future High-Voltage Multi-Terminal DC Networks

Introduction

Dr. R. Teixeira Pinto, Delft University of [email protected] /rteixeirapinto @Hredric

Co-authors:Prof.Dr. P. Bauer, TU Delft ([email protected])Dr. J. Enslin, UNC Charlotte ([email protected])

mailto:[email protected]

https://www.linkedin.com/profile/view?id=43469154

http://twitter.com/Hredric

Introduction

Table of Contents

1 IntroductionToday’s AgendaThe NSTG ProjectAdditional Material

2 BackgroundMotivationProblem DefinitionMTdc Network ChallengesObjectives and Research QuestionsOutline and Approach

Dr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: EPE’14-ECCE EU August 25, 2014 2 / 18

Introduction Today’s Agenda

Agenda

• 9:30–9:50 – P. Bauer / R. T. Pinto: Introduction

• 9:50–10:20 – P. Bauer: A review of MTdc networks with Classic HVdc

• 10:20–10:50 – P. Bauer: The modern VSC-HVdc technology

• 10:50–11:10 – Coffee Brake

• 11:10–11:40 – R. T. Pinto: Challenges on the road to future MTdcnetworks

• 11:40–12:20 – R. T. Pinto: Optimal load-flow control of MTdcnetworks

• 12:20–13:00 – R. T. Pinto : MTdc network modelling andexperimental validation

• 13:00–14:00 – Lunch break



Agenda




• 10:50–11:10 – Coffee Brake




• 13:00–14:00 – Lunch break



Agenda




• 10:50–11:10 – Coffee Brake




• 13:00–14:00 – Lunch break



Agenda




• 10:50–11:10 – Coffee Brake




• 13:00–14:00 – Lunch break



Agenda




• 10:50–11:10 – Coffee Brake




• 13:00–14:00 – Lunch break



Agenda




• 10:50–11:10 – Coffee Brake




• 13:00–14:00 – Lunch break



Agenda




• 10:50–11:10 – Coffee Brake




• 13:00–14:00 – Lunch break



Agenda




• 10:50–11:10 – Coffee Brake




• 13:00–14:00 – Lunch break


Introduction The NSTG Project

NSTG Project Timeline

Main project goal: identify and studytechnical and economical aspects ofa transnational electricity network inthe North Sea for the connection ofoffshore wind power and to promoteenergy trade between countries.

1 2 3 4

WP1

WP3

WP2

WP4

WP5

WP6

WP7


• WP1 – Inventory of available technology, modularity and flexibility

• WP2 – Technical and economic evaluation of different solutions

• WP3 - NSTG operation and control

• WP4 – Real-time multi-terminal network testing

• WP5 – Optimisation of NSTG solutions

• WP6 - Grid integration: planning, congestion and stability

• WP7 - Costs, benefits, regulations and market aspects of the NSTG




Main project goal: identify and studytechnical and economical aspects ofa transnational electricity network inthe North Sea for the connection ofoffshore wind power and to promoteenergy trade between countries. 1 2 3 4

WP1

WP3

WP2

WP4

WP5

WP6

WP7













WP1

WP3

WP2

WP4

WP5

WP6

WP7













WP1

WP3

WP2

WP4

WP5

WP6

WP7










Introduction Additional Material

More information and materials can be found on:

NSTG website:http://www.nstg-project.nl/

Ph.D. ThesisMulti-Terminal DC Networks:System Integration, Dynamics andControl


Background

Table of Contents

1 IntroductionToday’s AgendaThe NSTG ProjectAdditional Material

2 BackgroundMotivationProblem DefinitionMTdc Network ChallengesObjectives and Research QuestionsOutline and Approach


Background Motivation

General Background and Motivation

1970 1980 1990 2000 2010 2020 2030

1

3

5

7

911131517

Year

En

erg

y [

Bil

lio

n t

oe]

, Po

pu

lati

on

[B

illi

on

]

Primary Energy useElectricity Consumption

Population

Worldwide Primary Energy Use.

1970 1980 1990 2000 2010 2020 2030

0.030.050.070.1

0.30.50.7

1

357

10

305070

100

300500700

1000

30005000

Year

Ele

ctri

city

Gen

erat

ion

[T

Wh

]

Total

Renewables

Onshore Wind

Offshore Wind

Electricity generation in theEuropean Union.


Background Problem Definition

Why use HVdc Transmission

19th century - AC was preferred mainly because: easier to achieve highervoltages (lower transmission losses) / Three-phase synchronous generatorswas easier, cheaper and more efficient than dynamos.Reasons for choosing HVdc over HVac today:

1 Greater power per conductor: an overhead HVdc line can take 1.5 to2.1 times more power than a HVac overhead line, and anunderground HVdc line can take 2.9 to 3.8 times more power than anunderground HVac equivalent.

2 Higher voltages possible: Since 2010, HVdc voltages of up to 1600kV (± 800 kV) were already possible (Xiangjiaba – Shanghai HVdcline [1]) x 1200 kV for HVac (achieved in Russia 1988-1996).

3 Simpler line construction



50 m± 500 kV

100 m2 x 500 kV

ROW Comparison between HVac andHVdc lines [2].Transmission of 2000 MW:

• ± 500 kV HVdc line, the ROWis circa 50 m.

• For a HVac line, due to stabilitylimits, the ROW is doubled(additional three-phase circuit isneeded to transmit the same2000 MW [2].

Therefore (in this case) an HVdc lineis usually 30% cheaper than for itsHVac equivalent [3].

Right-of-way for an AC Line designed tocarry 3,000 MW is more than 70% wider.




4 Transmission distance is not limited by stability

0 1000 2000 3000 40000

100

200

300

400

500

600

700

800

Active power [MW]

Vo

ltag

e at

rec

eiv

ing

no

de

[kV

]

cos(φ)=0.95 capcos(φ)=1cos(φ)=0.95 ind

Transmission Distance = 250 km

0 1000 2000 3000 4000Active power [MW]

0

100

200

300

400

500

600

700

800

Vo

ltag

e at

rec

eiv

ing

no

de

[kV

]cos(φ)=0.95 capcos(φ)=1cos(φ)=0.95 ind

Transmission Distance = 500 km




4 Transmission distance is not limited by stability

50 100 150 200 250 300 350 400 450 500 550 6000

0.5

1

1.5

2

2.5

3

3.5

4

Transmission Distance [km]

Max

imu

m T

ran

smit

tab

le P

ow

er /

SIL

[p

u]

69 kV line115 kV line230 kV line345 kV line500 kV line765 kV line

Maximum transmittable power




5 No skin, proximity effect or need for reactive power compensation.6 Each conductor = independent circuit (if ground can be used).7 Synchronous operation is not required.8 Does not contribute to short-circuit current of the ac system.9 Less problems with resonances.

-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

conductor radius [mm]

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Cu

rren

t d

istr

ibu

tio

n [

pu

]

Drake

Honshu

Hokkaido

50 Hz60 Hz

Shikoku

substations

ac transmission line (500 kV)ac transmission line (187 ~ 275 kV)

back-to-back converterdc-ac converterdc transmission line

Hokkaido-Honshu HVdc link(600 MW, ±250 kV)

Anan-Kihoku HVdc link(2800 MW, ±250 kV)

Shin ShinanoCSC (600 MW)

SakumaCSC (300 MW)

Higashi-ShimizuCSC (300 MW)

Minami-Fukumitsu CSC (300 MW)

HVdc in Japan



When use HVdc Transmission for OWFs

Depends on a technical-economic analysis.

0 20 40 60 80 100 120 140 160 180 200 220 2400

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

Transmission Distance [km]

Tra

nsm

itta

ble

Po

wer

[p

u]

132 kV AC - 630 mm2

132 kV AC - 1000 mm2

220 kV AC - 630 mm2

220 kV AC - 1000 mm2

400 kV AC - 630 mm2

400 kV AC - 1000 mm2

80 kV DC - 300 mm2

80 kV DC - 1200 mm2

150 kV DC - 300 mm2

150 kV DC - 1200 mm2

320 kV DC - 300 mm2

320 kV DC - 1200 mm2

Maximum transferrable power x transmission distance for ac and dc submarine cables.



HVdc x HVac Transmission - Costs

Not an exact science: break-even area not break-even point.

AC Term.

AC Line

AC Losses

DC Term.

DC Line

DC Losses

Transmission distance

Costs

Break-even area

Cost Comparison between HVac and HVdctransmission systems.

Average transmission line cost in theUSA (a 500 kV line):

• HVDC - 1.5 – 2.0 M$/mile

• HVAC - 3.0 – 3.5 M$/mile

Data used in exercise at TU Delft:

• AC Cable Cost: 0.9 Me/km

• AC Substation Cost: 3 Me each

• Q (SVC) Compensation Cost:0.04 Me/MVAr

• DC Cable Cost: 0.7 Me/km

• Converter Cost: 70 Me each



The need for HVdc transmission systems

Future vision of the NSTG bythe OMA [4].

UK1

UK2

BE1

NL1

NL2

DE1

DE2

DK1

DK2

HUB1

HUB2

HUB3

NL

UK

BE

DK

DE

HUB4

HUB5

Future vision by the NSTGproject.

• Best renewable resources sites: remotelylocated.

• Using HVdc transmission is more efficient(than HVac).

• Offshore wind potential: 28% of wind energyshare [5]

• 44 GW of OWF by 2020: 31.3% annual growth(on average).

• New OWFs: further away and higher capacity[6]

• 30% of all generation goes ac/dc beforeconsumption. Prediction: staggering 80%.

• dc networks (level): microgrids (house) andsmart grids (district), electronic powerdistribution systems (cities/country) [7] andsupergrids (country/continent) [8, 9].



The need for HVdc transmission systems

Future vision of the NSTG bythe OMA [4].

UK1

UK2

BE1

NL1

NL2

DE1

DE2

DK1

DK2

HUB1

HUB2

HUB3

NL

UK

BE

DK

DE

HUB4

HUB5

Future vision by the NSTGproject.

• Best renewable resources sites: remotelylocated.

• Using HVdc transmission is more efficient(than HVac).

• Offshore wind potential: 28% of wind energyshare [5]

• 44 GW of OWF by 2020: 31.3% annual growth(on average).

• New OWFs: further away and higher capacity[6]

• 30% of all generation goes ac/dc beforeconsumption. Prediction: staggering 80%.

• dc networks (level): microgrids (house) andsmart grids (district), electronic powerdistribution systems (cities/country) [7] andsupergrids (country/continent) [8, 9].


Background MTdc Network Challenges

There are 5 main technical challenges

Challenge 1

System Integration

Challenge 2

Power Flow Control

Challenge 3

Dynamic Behaviour

Challenge 4

Stability

Challenge 5

Fault Behaviour


Background Objectives & Research Questions

Research Questions

• Chapter 2: What is the best HVdc technology and configuration for aMTdc network in the North Sea?

• Chapter 3: How does the power flow in MTdc networks?• Chapter 4: How to model the combined dynamic behaviour of a

MTdc network?• Chapter 5: What are the shortcomings of the main methods for

controlling the direct voltage in MTdc networks?• Chapter 6: What is needed to control the power flow in MTdc

networks?• Chapter 7: What is needed for a MTdc network to be able to

withstand and recover from faults in the connected ac systems, or inthe dc grid itself, without halting its operation?

• Chapter 8: What are the main variables affecting the small-signalstability of MTdc networks?

• Chapter 9: To what extent is it possible to reproduce the behaviourof a high-voltage MTdc network through a low-voltage one?



Research Questions


• Chapter 3: How does the power flow in MTdc networks?

• Chapter 4: How to model the combined dynamic behaviour of aMTdc network?

• Chapter 5: What are the shortcomings of the main methods forcontrolling the direct voltage in MTdc networks?

• Chapter 6: What is needed to control the power flow in MTdcnetworks?

• Chapter 7: What is needed for a MTdc network to be able towithstand and recover from faults in the connected ac systems, or inthe dc grid itself, without halting its operation?





Research Questions



MTdc network?

• Chapter 5: What are the shortcomings of the main methods forcontrolling the direct voltage in MTdc networks?







Research Questions




controlling the direct voltage in MTdc networks?







Research Questions





networks?






Research Questions











Research Questions











Research Questions










Background Outline

Outline and Approach

Part I.

Intro

du

ction

&

Literatu

re Rev

iew

Part III.

Dy

nam

ic An

alysis

Part V

. E

xperim

ental

Wo

rk &

C

on

clusio

ns

Part IV

. S

tability

An

alysis

Part II.

Stead

y-S

tate A

naly

sis

4. Stability5. Fault Behavior

1. System Integration2. Power Flow Control3. Dynamic Behavior4. Stability

1. System Integration2. Power Flow Control

2. Power Flow Control3. Dynamic Behavior4. Stability

1. System Integration

3. Network Operation and

Power Flow

2. HVdc Transmission

Systems

4. Dynamic Modelling

5. Control of MTdc Networks

8. Small-Signal Analysis

7. Fault Analysis

9. Laboratory Setup of a LV-MTdc

System

6. The Distributed Voltage Control

Strategy


References

[1] J. Dorn, H. Gambach, and D. Retzmann, “HVDC transmission technology forsustainable power supply,” in 9th International Multi-Conference on Systems,Signals and Devices (SSD), 2012, pp. 1–6.

[2] M. Rashid, Power Electronics Handbook: Devices, Circuits and Applications, ser.Engineering. Elsevier Science, 2010, iSBN: 9780080467658.

[3] M. Bahrman, “HVDC transmission overview,” in IEEE/PES Transmission andDistribution Conference and Exposition, 2008, pp. 1–7.

[4] The Office for Metropolitan Architecture. ZEEKRACHT, NETHERLANDS, THENORTH SEA, 2008: A masterplan for a renewable energy infrastructure in theNorth Sea. Last Accessed: 07 February, 2013. [Online]. Available:http://oma.eu/projects/2008/zeekracht

[5] Energy Company of the Netherlands (ECN), “Renewable Energy Projections asPublished in the National Renewable Energy Action Plans of the European MemberStates Summary report,” ECN, Petten, Technical Report November 2011, 2011.[Online]. Available:http://www.ecn.nl/docs/library/report/2010/e10069 summary.pdf

[6] European Wind Energy Association, “Wind in power: 2011 european statistics,”EWEA, Brussels, Technical Report, 2012. [Online]. Available:


http://oma.eu/projects/2008/zeekracht

http://www.ecn.nl/docs/library/report/2010/e10069_summary.pdf

References

http://www.ewea.org/fileadmin/files/library/publications/statistics/Wind in power 2011 European statistics.pdf

[7] D. Boroyevich, I. Cvetkovic, D. Dong, R. Burgos, F. Wang, and F. Lee, “Futureelectronic power distribution systems a contemplative view,” in 12th InternationalConference on Optimization of Electrical and Electronic Equipment. IEEE, May2010, pp. 1369–1380.

[8] S. Taggart, G. James, Z. Dong, and C. Russell, “The Future of Renewables Linkedby a Transnational Asian Grid,” Proceedings of the IEEE, vol. 100, no. 2, pp.348–359, 2012.

[9] M. Aredes, A. F. Da Cunha De Aquino, C. Portela, and E. Watanabe, “Going thedistance - Power-Electronics-Based Solutions for Long-Range Bulk PowerTransmission,” IEEE Industrical Electronics Magazine, no. March, p. 13, Mar. 2011.


http://www.ewea.org/fileadmin/files/library/publications/statistics/Wind_in_power_2011_European_statistics.pdf

http://www.ewea.org/fileadmin/files/library/publications/statistics/Wind_in_power_2011_European_statistics.pdf

A Review of MTdc Networks withClassic HVdc

Part I






Early HVdc Systems

Table of Contents

1 Early HVdc SystemsThe Thury SystemThe World’s 1st MTdc Network

2 HVdc Classic SystemsPastPresentFuture

3 HVdc Transmission Systems ConfigurationsMonopolar & HomopolarBipolar Configuration

4 From Point-to-Point to MTdc NetworksSACOIHydro-Quebec – New England

5 MTdc Network TopologiesTechnologySeries x ParallelClassification

Dr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: Part I August 25, 2014 2 / 23

Early HVdc Systems The Thury System

M M M M

12 kV

100 A

G

v

G

v

G

v

100 kW

50 kW

M

v

M

v

M

G

v

300 kW

300 kW

Diagram of a 1.2 MW Thury system.

Direct voltages in the Thury systemsbetween 1889 and 1925.

• Despite HVac success, efforts for HVdcdevelopment continued.

• Rene Thury (Swiss): dc generators andmotors in series. Constant current andvariable voltage to meet power demand.

• Generators/motors had series switch toshort-circuit when not used. Forgenerator operation: speed up untilnominal line current and thenshort-circuit switch was opened.

• By the 1910s: at least 15 Thurysystems in Europe (France, Hungary,Italy, Russia, Spain and Switzerland)[1].

• Bipolar voltages of up to 150 kV wheresuccessfully achieved.



M M M M

12 kV

100 A

G

v

G

v

G

v

100 kW

50 kW

M

v

M

v

M

G

v

300 kW

300 kW










M M M M

12 kV

100 A

G

v

G

v

G

v

100 kW

50 kW

M

v

M

v

M

G

v

300 kW

300 kW










1889 1897 1906 1912 1925

6

14

22

58

100

150

Gorzente River - GenoaLa Chaux - de-Fonds

St. Maurice - Lausanne

Lyon - Moutiers

Wilesden - Irongridge

Lyon - Moutiers - La Bridoire

Lyon - Moutiers - La Bridoire - Bozel

Chambéry

Year

Dir

ect

Vo

ltag

e [k

V]










1889 1897 1906 1912 1925

6

14

22

58

100

150

Gorzente River - GenoaLa Chaux - de-Fonds

St. Maurice - Lausanne

Lyon - Moutiers

Wilesden - Irongridge

Lyon - Moutiers - La Bridoire

Lyon - Moutiers - La Bridoire - Bozel

Chambéry

Year

Dir

ect

Vo

ltag

e [k

V]









Early HVdc Systems The World’s 1st MTdc Network

Moutier-Lyon-Bridoire-Bozel Thury system

• Start: 4.3 MW HVdc system betweenLyon and Moutier (hydroelectric plant)

• 75 A and the voltage up to 57.6 kV (16generators at Moutier).

• 1911: 1st upgrade. 2nd hydro plant atLa Bridoire (6 MW). The current wasdoubled to 150 A.

• Finally, a third hydroelectric plant,rated at 9 MW, was added in Bozel.

• Until 1937, when it was dismantled, theMoutier-Lyon-Bridoire-Bozel Thurysystem operated with four terminalsand can be assumed the world’s 1stmulti-terminal HVdc network.

Bozel9 MW

v v vv

vv

vv

vv

vv

v

v

150 A

La Bridoire6 MW

Moutier4.3 MW

Lyon19.3 MW

12 kV

29 kV

40 k

V

129 kV

Layout of the 1st HV MTdc network

Lyon19.3 MW

La Bridoire6 MW

Moutier4.3 MW

89 kV

129 kV

direct voltage

90 km 45 km 11 km

transmission distance

60 kV

Bozel 9 MW

Voltage and distances the terminals.









Bozel9 MW

v v vv

vv

vv

vv

vv

v

v

150 A

La Bridoire6 MW

Moutier4.3 MW

Lyon19.3 MW

12 kV

29 kV

40 k

V

129 kV


Lyon19.3 MW

La Bridoire6 MW

Moutier4.3 MW

89 kV

129 kV

direct voltage

90 km 45 km 11 km


60 kV

Bozel 9 MW










Bozel9 MW

v v vv

vv

vv

vv

vv

v

v

150 A

La Bridoire6 MW

Moutier4.3 MW

Lyon19.3 MW

12 kV

29 kV

40 k

V

129 kV


Lyon19.3 MW

La Bridoire6 MW

Moutier4.3 MW

89 kV

129 kV

direct voltage

90 km 45 km 11 km


60 kV

Bozel 9 MW










Bozel9 MW

v v vv

vv

vv

vv

vv

v

v

150 A

La Bridoire6 MW

Moutier4.3 MW

Lyon19.3 MW

12 kV

29 kV

40 k

V

129 kV


Lyon19.3 MW

La Bridoire6 MW

Moutier4.3 MW

89 kV

129 kV

direct voltage

90 km 45 km 11 km


60 kV

Bozel 9 MW










Bozel9 MW

v v vv

vv

vv

vv

vv

v

v

150 A

La Bridoire6 MW

Moutier4.3 MW

Lyon19.3 MW

12 kV

29 kV

40 k

V

129 kV


Lyon19.3 MW

La Bridoire6 MW

Moutier4.3 MW

89 kV

129 kV

direct voltage

90 km 45 km 11 km


60 kV

Bozel 9 MW



HVdc Classic Systems

Table of Contents







HVdc Classic Systems Past

Mercury-arc Valve HVdc Systems

Photography from the

Gotland 1 mercury-arc valve

hall during the 1950s [2].

• 1st system: 1930 (USA). From hydroelectricpower plant in Mechanicville to Schenectady(NY). 37 km / 12 kV / 5 MW. Interesting fact:40 Hz at plant and 60 Hz in NY [1].

• 1939: Backfire issue solved by Uno Lamm(Sweden).

• 1939 - 1951: 7 experimental HVdc transmissionsystems using mercury-arc valves were built inSwitzerland, Germany, Sweden and Russia.

• 1st commercial system: 1954, Gotland 1(Sweden - ASEA) [2]. Connected the Swedishmainland, at Vstervik, to Ygne in the island ofGotland. 98 km / 20 MW / 100 kV [3].

































HVdc Systems Comeback

• After war of the currents: 60years for HVdc transmissionsystems to “fightback”.

• 1970s: voltages > 400 kV andcapacities > 1000 MW.

• Pacific Intertie: 1440 MW,500 kV

• Nelson River: 1620 MW,450 kV [1, 3, 4, 5, 6].

1930 1935 1940 1945 1950 1955 1960 1965 1970 19750

100

200

300

400

500

600

Mechanicville–Schenectady

Zurich-Wettingen

Charlottenburg-Moabit

Lehrte-Misburg

Elbe-Project

Trollhattan-Merud

Moscow–Kashira

Gotland 1

Cross-Channel

Konti-Skan 1

Volgograd-Donbass

SACOI 1

Sakuma B2B

Inter-Island 1

Vancouver Island 1

Pacific DC Intertie

Nelson River Bipole 1

Kingsnorth

Commissioning Year

Dir

ect

Vo

ltag

e [k

V]

Biggest 1620 MW

Average 357 MW

Legend

Evolution of mercury-arc valves HVdc systems.








1930 1935 1940 1945 1950 1955 1960 1965 1970 19750

100

200

300

400

500

600


Zurich-Wettingen


Lehrte-Misburg

Elbe-Project

Trollhattan-Merud

Moscow–Kashira

Gotland 1

Cross-Channel

Konti-Skan 1

Volgograd-Donbass

SACOI 1

Sakuma B2B

Inter-Island 1

Vancouver Island 1

Pacific DC Intertie


Kingsnorth

Commissioning Year

Dir

ect

Vo

ltag

e [k

V]

Biggest 1620 MW

Average 357 MW

Legend









1930 1935 1940 1945 1950 1955 1960 1965 1970 19750

100

200

300

400

500

600


Zurich-Wettingen


Lehrte-Misburg

Elbe-Project

Trollhattan-Merud

Moscow–Kashira

Gotland 1

Cross-Channel

Konti-Skan 1

Volgograd-Donbass

SACOI 1

Sakuma B2B

Inter-Island 1

Vancouver Island 1

Pacific DC Intertie


Kingsnorth

Commissioning Year

Dir

ect

Vo

ltag

e [k

V]

Biggest 1620 MW

Average 357 MW

Legend



HVdc Classic Systems Present

HVdc Classic: Thyristor Technology

• Thyristor (or SCR): possible toachieve higher voltages.

• A modern 6-inch thyristor: upto 4 kA / block up to 8.5 kV [7].

• Thyristor valves improvements:larger powers through longerdistances.

• 1st commercial system: 1972Eel River link in Canada (GE).B2B / 320 MW / 160 kV [3, 4].

• Very mature technology(> 140 HVdc systemsworldwide) [4].

6-inch HVdc thyristor.









Blo

ckin

g V

olt

age

[kV

]

Si-

area

[m

m²]

0

1

2

3

4

5

6

7

8

9

1.65 kV

1970 1975 1980 1985 1990 1995 2000 2005 2010

1.5 “3000

6000

9000

12000

15000

18000

21000

24000

1000 MW converter à 14.000 thyristors

1000 MW converter à 400 thyristors

6 “

8.5kV 27000

0

Evolution of thyristor technology.









0 500 1000 1500 2000 2500 30000

200

400

600

800

1000

1200

1400

1600

1800

Dir

ect

Vo

ltag

e [k

V]

Transmission distance [km]Evolution of CSC-HVdc voltage.









0 500 1000 1500 2000 2500 30000

200

400

600

800

1000

1200

1400

1600

1800

Dir

ect

Vo

ltag

e [k

V]










0 500 1000 1500 2000 2500 30000

200

400

600

800

1000

1200

1400

1600

1800

Dir

ect

Vo

ltag

e [k

V]




The HVdc Classic Station (24-pulse converter)

10

AC System

AC FILTERS

AC FILTERS

DC

FIL

TE

RS

DC

FIL

TE

RS

DC

LIN

E

DC

LIN

E

Earth electrode

Met

alli

c R

etu

rn

1 1

2 2

4 45 5

6 6

Neu

tral

b

us

1 1

3 3

7

8

9

111) Converter bridges2) Converter transformers3) Smoothing reactors4) AC filters5) Reactive power supply6) DC filters7) Surge arresters8) Neutral bus surgecapacitor9) Fast dc switches10) Earth electrode11) DC line


HVdc Classic Systems Future

Undisputed for bulk transmission

Boxplot distribution of HVdc Classic projects worldwide.

30500

1000

2500

5000

6000

7200

Inst

alle

d C

apac

ity

[M

W]

35

500

700

1000

1600

Tra

nsm

issi

on

Vo

ltag

e [k

V]

12

178

700

1000

2090

2375

Tra

nsm

issi

on

Dis

tan

ce [

km

]

• Interquartiles (50%):• 180 - 1000 km• (± 250 kV) - (± 500 kV)• 500 - 2500 MW


HVdc Classic Systems Future

Undisputed for bulk transmission

Worldwide installed capacity.

1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 20150

20

40

60

80

100

120

140

160

180

200

Inst

alle

d C

apac

ity

[G

W]

• UHVdc ratings up to 1600 kV (± 800 kV) and 7.2 GW.(e.g. Jinping - Sunan link in China [8]).

• Only in China more than 270 GW over 40 project are predictedbetween 2010 and 2020 (for further info)


http://www.ptd.siemens.de/ETG_GMA201306_Integration_HVDC.pdf

HVdc Transmission Systems Configurations

Table of Contents







HVdc Transmission Systems Configurations Monopolar & Homopolar

Homopolar Configurations

Idc

Vdc

Idc

Vdc

2Idc

Ground return

Idc

Vdc

Idc

Vdc

2Idc

Metallic return

Monopolar Configurations

Idc

Vdc

Ground return

Idc

Vdc

Metallic return




Idc

Vdc

Idc

Vdc

2Idc

Ground return

Idc

Vdc

Idc

Vdc

2Idc

Metallic return


Idc

Vdc

Ground return

Idc

Vdc

Metallic return




Idc

Vdc

Idc

Vdc

2Idc

Ground return

Idc

Vdc

Idc

Vdc

2Idc

Metallic return


Idc

Vdc

Ground return

Idc

Vdc

Metallic returnDr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: Part I August 25, 2014 12 / 23

HVdc Transmission Systems Configurations Bipolar Configuration

Idc

Vdc

Idc

Vdc

Ground return

Idc

Vdc

Idc

Vdc

Metallic return

Bipolar configurations during faults

Idc

Vdc

HVdc cable fault

Idc

Vdc

HVdc converter fault



Idc

Vdc

Idc

Vdc

Ground return

Idc

Vdc

Idc

Vdc

Metallic return


Idc

Vdc

HVdc cable fault

Idc

Vdc




Idc

Vdc

Idc

Vdc

Ground return

Idc

Vdc

Idc

Vdc

Metallic return


Idc

Vdc

HVdc cable fault

Idc

Vdc




Idc

Vdc

Idc

Vdc

Ground return

Idc

Vdc

Idc

Vdc

Metallic return


Idc

Vdc

HVdc cable fault

Idc

Vdc



From Point-to-Point to MTdc Networks

Table of Contents







From Point-to-Point to MTdc Networks SACOI

MainlandItaly

San Dalmazio

Lucciana

Corsica

SardiniaCodrogianos

200 kV dc overhead line

200 kV dc submarine cable

Electrode line

Converter stations

Insolator stations

MediterraneanSea

Aerial view

• 1967: 6-pulse mercury-arc converters,monopole, ground return.

• Line capacity: 200 MW @ 200 kV (300 kmOHL, and 120 km submarine [9]).

• 1987: 1st MTdc HVdc network. Luccianaconverter in Corsica (thyristor / 50 MW) [10].

• 1992: Sardinian and Italian converter upgradedto thyristor technology (from 200 to 300 MW).

Italy

200 MW

200 kV

1000 A

Corsica

50 MW

200 kV

250 A

Sardinia

200 MW

200 kV

1000 A220 kV 220 kV

90 kV

Single line diagram [10]



MainlandItaly

San Dalmazio

Lucciana

Corsica

SardiniaCodrogianos



Electrode line

Converter stations

Insolator stations

MediterraneanSea

Aerial view





Italy

200 MW

200 kV

1000 A

Corsica

50 MW

200 kV

250 A

Sardinia

200 MW

200 kV

1000 A220 kV 220 kV

90 kV




MainlandItaly

San Dalmazio

Lucciana

Corsica

SardiniaCodrogianos



Electrode line

Converter stations

Insolator stations

MediterraneanSea

Aerial view





Italy

200 MW

200 kV

1000 A

Corsica

50 MW

200 kV

250 A

Sardinia

200 MW

200 kV

1000 A220 kV 220 kV

90 kV




MainlandItaly

San Dalmazio

Lucciana

Corsica

SardiniaCodrogianos



Electrode line

Converter stations

Insolator stations

MediterraneanSea

Aerial view





Italy

200 MW

200 kV

1000 A

Corsica

50 MW

200 kV

250 A

Sardinia

200 MW

200 kV

1000 A220 kV 220 kV

90 kV



From Point-to-Point to MTdc Networks Hydro-Quebec – New England

• Phase I (Oct 1986): 2 terminals (DesCantons - Comerford). 172km / ± 450kV / 690 MW [11, 12].

• Phase II (1990): +3 converters.Radission (2250 MW) and Sandy Pond(2000 MW) [12]. Goal: power from LaGrande hydro plant to Boston.

• 1992: Nicolet (2138 MW), Montreal.

• No dc breakers (dc-side contingenciesare dealt via control actions).

• Initial 2 converters weredecommissioned after reassessment ofthe additional benefits they would bringto the three-terminal MTdc (ABB) [12].

• Remaining 3 form a MTdc network(1480 km of OHL).

Aerial view [12]

Radisson

NicoletDes Cantons

Comerford

Sandy Pond

Canada

USA









Aerial view [12]

Radisson

NicoletDes Cantons

Comerford

Sandy Pond

Canada

USA









Aerial view [12]

Radisson

NicoletDes Cantons

Comerford

Sandy Pond

Canada

USA









Single line diagram

Chissibi

Quebec

Nemiscau

735 kV

735 kV

315 kV

1022 km

105 km

383 km

735 kV

735 kV

Sherbrooke

New England

345 kV

Legend

Islanding breaker

Generation

DC converter

Decommissioned

Transformer

6 16

12

La Grande-2La Grande-2A

La Grande-1

Radisson2 250 MW

Nicolet2 138 MW

Des Cantons690 MW

Comerford690 MW

Sandy Pond2 000 MW

± 450 kV









Single line diagram

Chissibi

Quebec

Nemiscau

735 kV

735 kV

315 kV

1022 km

105 km

383 km

735 kV

735 kV

Sherbrooke

New England

345 kV

Legend

Islanding breaker

Generation

DC converter

Decommissioned

Transformer

6 16

12


La Grande-1

Radisson2 250 MW

Nicolet2 138 MW

Des Cantons690 MW

Comerford690 MW

Sandy Pond2 000 MW

± 450 kV









Single line diagram

Chissibi

Quebec

Nemiscau

735 kV

735 kV

315 kV

1022 km

105 km

383 km

735 kV

735 kV

Sherbrooke

New England

345 kV

Legend

Islanding breaker

Generation

DC converter

Decommissioned

Transformer

6 16

12


La Grande-1

Radisson2 250 MW

Nicolet2 138 MW

Des Cantons690 MW

Comerford690 MW

Sandy Pond2 000 MW

± 450 kV



Why MTdc Networks with HVdc Classic didn’t thrive?

Nowadays, the MTdc network can operate in three different configurations:

• point-to-point link (e.g. between Radisson and Sandy Pond or Nicoletand Sandy Pond);

• multi-terminal HVdc network (all 3 terminals);• hybrid configuration (pole outage at Radisson or Sandy Pont): 3

converters connected on one pole and 2 on the other pole.

Switchgear arrangement [13]

DUNCAN

ELECTRODE

RADISSON NICOLET DES CANTONS COMERFORD SANDY POND

DES CANTONS

ELECTRODE

LINE 1

LINE 2

DISCONNECT

RAPID DISCONNECT

COMMUTATION BREAKERS

DC CABLE


MTdc Network Topologies

Table of Contents







MTdc Network Topologies Technology

HVdc Technology

• CSC-MTdc: all the converter stations use the line commutatedcurrent-source converter HVdc technology;

• VSC-MTdc: all the converter stations use the forced commutatedvoltage-source converter HVdc technology;

• Hybrid-MTdc: when both HVdc technologies – CSC and VSC – areused together.

Note: Multiple infeed of HVdc lines into ac networks does not form aMTdc network!

AC System 1 AC System 2

AC System 3



HVdc Technology






AC System 3



HVdc Technology






AC System 3


MTdc Network Topologies Series x Parallel

Multi-terminal dc network with monopolar HVdc stations

Idc

Vdc1 Vdc3

Vdc2

Vdc1+Vdc2+Vdc3+Vdc4 = 0

Vdc4

Series MTdc

Vdc

Idc1

Idc2

Idc1+Idc2+Idc3+Idc4 = 0

Idc4

Idc3

Parallel MTdc



Multi-terminal dc network with monopolar HVdc stations

Idc

Vdc1 Vdc3

Vdc2

Vdc1+Vdc2+Vdc3+Vdc4 = 0

Vdc4

Series MTdc

Vdc

Idc1

Idc2

Idc1+Idc2+Idc3+Idc4 = 0

Idc4

Idc3

Parallel MTdc



Parallel Connected MTdc networks

Radial

Meshed



Parallel Connected MTdc networks

Radial Meshed



Comparison between series and parallel MTdc networks

Characteristic Series MTdc Parallel MTdc

Power FlowReversal

In CSC-MTdc power flow reversal can easily beachieved by inverting the converter voltages. WithVSC-MTdc it would not be easy to invert theconverters voltage polarity, thus power flow reversalwould involve mechanical switches.

In CSC-MTdc the current direction cannot beinverted, hence, there is need for mechanicalswitches. In VSC-MTdc the current direction caneasily be inverted, hence power flow reversal can beachieved via control actions.

HVdc TerminalPower Rating

Depends on converter voltage rating (cheaper forsmaller powers).

Depends on converter current rating.

LossesHigher losses, which can be minimised by alwaysoperating with the minimum current possible.

Lower losses.

InsulationIs difficult in series connection as the voltages in theMTdc network vary.

All converters need to be insulated to the ratedvoltage.

DC FaultsA permanent fault in a transmission line would makethe whole MTdc network unavailable.

A permanent fault in a transmission line would onlymake the affected terminal unavailable (in meshedMTdc networks normal operation is still possible).

AC Faults Leads to overvoltages in the remaining terminals. Leads to overcurrents in the remaining terminals.

ProtectionIn series CSC-MTdc, dc faults can be handled viacontrol actions. VSC-MTdc will need dc breakers.

For clearing dc faults parallel MTdc networks willneed dc breakers.


MTdc Network Topologies Classification

SeriesParallel

Radial Meshed

MonopolarHomopolar Bipolar

Classic VSC Hybrid

HVdc technology

MTdc network

topology

Return Path

GroundMetallic

HVdc configuration

Classification of MTdc transmission systems.


References

[1] E. W. Kimbark, Direct current transmission. Wiley-Interscience, 1971, vol. 1,iSBN: 9780471475804.

[2] ABB AB Grid Systems - HVDC, “HVDC Light - It’s time to connect,” ABB,Ludvika, Technical Report, December 2012. [Online]. Available:http://www05.abb.com/global/scot/scot221.nsf/veritydisplay/2742b98db321b5bfc1257b26003e7835/$file/Pow0038%20R7%20LR.pdf

[3] D. A. Woodford, “HVDC Transmission,” Manitoba HVDC Research Centre, pp.1–27, March 1998, Last Accessed on 03 February 2013. [Online]. Available:http://www.sari-energy.org/PageFiles/What We Do/activities/HVDC Training/Materials/BasisPrinciplesofHVDC.pdf

[4] Working Group on HVDC and FACTS Bibliography and Records, “HVDCPROJECTS LISTING,” IEEE Transmission and Distribution Committee: DC andFlexible AC Transmission Subcommittee, Winnipeg, Technical Report, 2006.[Online]. Available:http://www.ece.uidaho.edu/hvdcfacts/Projects/HVDCProjectsListingDec2006.pdf

[5] R. S. Thallam, High-Voltage Direct-Current Transmission, ser. The ElectricalEngineering Handbook. CRC Press, 2000, ch. 61.3, pp. 1402–1416, iSBN:9780849301858.


http://www05.abb.com/global/scot/scot221.nsf/veritydisplay/2742b98db321b5bfc1257b26003e7835/$file/Pow0038%20R7%20LR.pdf

http://www05.abb.com/global/scot/scot221.nsf/veritydisplay/2742b98db321b5bfc1257b26003e7835/$file/Pow0038%20R7%20LR.pdf

http://www.sari-energy.org/PageFiles/What_We_Do/activities/HVDC_Training/Materials/BasisPrinciplesofHVDC.pdf

http://www.sari-energy.org/PageFiles/What_We_Do/activities/HVDC_Training/Materials/BasisPrinciplesofHVDC.pdf

http://www.ece.uidaho.edu/hvdcfacts/Projects/HVDCProjectsListingDec2006.pdf

References

[6] M. P. Bahrman, DIRECT CURRENT POWER TRANSMISSION, ed., ser. StandardHandbook for Electrical Engineers. Mcgraw-hill, 2006, ch. 15, pp. 1–34, iSBN:9780071491495.

[7] V. Botan, J. Waldmeyer, M. Kunow, and K. Akurati, “Six Inch Thyristors forUHVDC Transmission,” in PCIM Europe: International Exhibition and Conferencefor Power Electronics, Intelligent Motion, Renewable Energy and EnergyManagement. Nuremberg: Mesago, 2010, pp. 1–4. [Online]. Available:http://www05.abb.com/global/scot/scot256.nsf/veritydisplay/c22b8e970d5455e3c1257af3004ef622/$file/6Inch%20Thyristor%20for%20UHVDC%20transmission.pdf

[8] Z. Kunpeng, W. Xiaoguang, and T. Guangfu, “Research and Development of ±800kV / 4750A UHVDC Valve,” in 2nd International Conference on IntelligentSystem Design and Engineering Application (ISDEA), 2012, pp. 1466–1469.

[9] J. Arrillaga, Y. Liu, and N. Watson, Flexible Power Transmission: The HVDCOptions. Wiley, 2007, iSBN: 9780470511855.

[10] V. Billon, J.-P. Taisne, V. Arcidiacono, and F. Mazzoldi, “The Corsican tapping:from design to commissioning tests of the third terminal of the Sardinia-Corsica-ItalyHVDC,” IEEE Transactions on Power Delivery, vol. 4, no. 1, pp. 794–799, 1989.


http://www05.abb.com/global/scot/scot256.nsf/veritydisplay/c22b8e970d5455e3c1257af3004ef622/$file/6Inch%20Thyristor%20for%20UHVDC%20transmission.pdf



References

[11] G. Morin, L. Bui, S. Casoria, and J. Reeve, “Modeling of the Hydro-Quebec-NewEngland HVDC system and digital controls with EMTP,” IEEE Transactions onPower Delivery, vol. 8, no. 2, pp. 559–566, 1993.

[12] ABB. The HVDC Transmission Quebec - New England: The first large scalemutiterminal HVDC transmission in the world. Last Accessed: 31 July, 2013.[Online]. Available: http://www.abb.com/industries/ap/db0003db004333/87f88a41a0be97afc125774b003e6109.aspx

[13] D. McNabbn, “Feedback on the Quebec - New England Multiterminal HVDC Line:20 years of Operation,” in Cigree B4 Open Session. Paris: Cigree, 2010. [Online].Available: http://b4.cigre.org/content/download/15754/604515/version/1/


http://www.abb.com/industries/ap/db0003db004333/87f88a41a0be97afc125774b003e6109.aspx

http://www.abb.com/industries/ap/db0003db004333/87f88a41a0be97afc125774b003e6109.aspx

http://b4.cigre.org/content/download/15754/604515/version/1/

The Modern VSC-HVdc TechnologyPart II






Introduction

Table of Contents

1 IntroductionBackgroundComparison with CSC-HVdc

2 Voltage-Source Converter TechnologyInitial DevelopmentsMulti-level VSC technologyVSC Commissioning

3 VSC-HVdc StructureAC Circuit BreakersTransformersFiltersPhase ReactorsDC Capacitors

Dr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: Part II August 25, 2014 2 / 22

Introduction Background

• Previous HVdc systems used thyristor (only turn off).

• VSC-HVdc uses self-commutating devices – e.g. GTO, IGBT (turn onand off).

• Therefore, VSC = more controllable sinusoidal voltages and currentswhereas CSC = much higher power ratings (and lower losses) [1].

• 1999: 1st commercial VSC (ABB). 50 MW Link between Gotlandisland and Sweden [1].

• Well stablished: 2010 at least 10 systems in operation [2].

• Most projects are point-to-point, but control flexibility means VSCtechnology eases implementation of MTdc networks (1st VSC-MTdcnetwork was built in China in 2013).

• Main converter topologies used: 2-level, 3-level and multi-level.
























































Introduction Comparison with CSC-HVdc

Characteristic CSC-HVdc VSC-HVdc

Converter Line-commutated current-source. Self-commutated voltage-source.Switch Thyristor: turn on capability only. IGBT: turn-on and turn-off capabilities.Age Old: First commercial project in 1954. New: First commercial project in 1999.Projects Worldwide 146 15Power Rating up to 8000 MW up to 1000 MWVoltage Rating up to ± 800 kV up to ± 320 kV

FiltersHarmonic orders are low (e.g. 11-th and 13-th),hence high filtering efforts are needed.

Filters are tuned to higher frequencies and are,therefore, smaller and cheaper.

Footprint Very high. Lower.

ControlAlways consume reactive power (two-quadrantoperation).

Independent control of active and reactive power(four-quadrant operation).

AC NetworkRequirements

Needs a reasonably strong ac system to operate(high minimum short-circuit ratio, e.g. SCR > 3)

Can operate with a weak ac network or be usedto feed islands and passive ac networks providingfrequency control. Black start capability.

AC FaultsPresents commutation failure during ac faults. Incase of repeated commutation failures theconverter is blocked.

Can maintain active power transfer even under acfaults, fault-ride through capable.

DC FaultsIs capable of extinguishing dc-side faults viacontrol actions.

Has no way of limiting dc-fault currents (becauseof the free-wheeling diodes), therefore dc breakersare needed.

Losses [% of RatedPower]

0.7% 1.5% (two-level) or 1.0% (multi-level)

CommunicationSpecial arrangements are needed to coordinatethe operation of converter stations.

Communication between the rectifier station andthe inverter station in theory is not necessary.The control of each converter station operates inan independent way.

Multi-terminalOperation

Difficult since there is need for coordinationbetween the converters (current ordersynchronisation) and power-flow reversal involvespolarity changes through mechanical switches.

Easier to accomplish since there is little need forcoordination between the interconnectedconverters and power-flow reversal does notinvolve mechanical switches.


Voltage-Source Converter Technology

Table of Contents





Voltage-Source Converter Technology Initial Developments

2-level VSC

The Graetz bridge is the most straightforward VSC configuration: eachphase can be connected either to the positive dc terminal, or the negativedc terminal.

+

-

Vdc / 2

Vdc / 2

n

va

1

4

1'

4'

vb

3

6

3'

6'

vc

5

2

5'

2'

or

Vdc

Idc

ia ib ic

Circuit of the 3-phase two-level voltage-source converter.



2-level VSC

VSC consists of 6 valves. Each valve contains a switching device (IGBT)and an anti-parallel diode.

+

-

Vdc / 2

Vdc / 2

n

va

1

4

1'

4'

vb

3

6

3'

6'

vc

5

2

5'

2'

or

Vdc

Idc

ia ib ic




2-level VSC

Series connection of IGBTs needed to handle the higher voltages. For ex-ample, for ± 150 kV, at least a 100 IGBTss for each valve (if 3.0 kV IGBTsare used).

+

-

Vdc / 2

Vdc / 2

n

va

1

4

1'

4'

vb

3

6

3'

6'

vc

5

2

5'

2'

or

Vdc

Idc

ia ib ic




2-level VSC

Several IGBTs in series: not an easy task. Standard wire-bond IGBT fail inopen circuit (bad for series connection).

+

-

Vdc / 2

Vdc / 2

n

va

1

4

1'

4'

vb

3

6

3'

6'

vc

5

2

5'

2'

or

Vdc

Idc

ia ib ic




2-level VSC

On early VSC-HVdc ABB used presspack IGBTs which fails in short-circuit(good for series connection).

+

-

Vdc / 2

Vdc / 2

n

va

1

4

1'

4'

vb

3

6

3'

6'

vc

5

2

5'

2'

or

Vdc

Idc

ia ib ic




2-level VSC

Presspack IGBT drawback: only a few suppliers worldwide.Wire-bond IGBT s many suppliers and bigger overall market. Siemens andAlstom VSCs use wire-bound (maybe because ABB patented series connec-tion of IGBTs?) [3].

+

-

Vdc / 2

Vdc / 2

n

va

1

4

1'

4'

vb

3

6

3'

6'

vc

5

2

5'

2'

or

Vdc

Idc

ia ib ic




3-level VSC

+

-

Vdc / 2

Vdc / 2

Na

1

4

1'

4'

1*

4*

1°

1°

diode clamping

+

-

Vdc / 2

Vdc / 2

Na

1

4

1'

4'

1*

4*

1°

1°

Vdc / 2

flying capacitor

Switching logic for the 3-level VSC.

TopologyVoltage Level

+Vdc/2 0 −Vdc/2

Diode Clamping 1 & 4 4 & 1* 1* &4*

Flying Capacitor 1 & 1* 4 & 1* or 1 & 4* 4 & 4*

• Lower dv/dt, less harmonics and lowerswitching losses.

• Most used topologies: NPC and CFC(switching logic differs).

• Total number of IGBTs ' 2-level VSCs.However, 3-level NPC require morediodes for clamping.

• The three-level concept can be extendedto a higher number of voltage levels [4].

• No NPC > 3-levels for HVdc: complexinsulation and cooling design of valves(option for B2B links - higher currentsand lower voltages).


Voltage-Source Converter Technology Multi-level VSC technology

• Proposed in 2003 by Prof. RainerMarquardt (University of Bundeswehr -Munich) [5, 6].

• Applications: HVdc, STATCOM, Railwayand other large drives (tens of MW).

• DC capacitors distributed in modules(converter is built up by cascadingmodules) [7].

• Each MMC module consists of two valves

3 different switching ways:

1 lower IGBT on and upper IGBT off: thecapacitor inserted into from A to B.

• Current from A to B (charging);• Current from B to A (discharging)

2 upper IGBT on and lower IGBT off: bypassed

3 Both IGBTs off: blocked

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

MMC (or M2C) topology.












AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB


B

A

MMC Submodule.












AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB


B

A

MMC Submodule.









• Current from A to B (charging);

• Current from B to A (discharging)



AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB


B

A

Charging.












AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB


B

A

Discharging.Dr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: Part II August 25, 2014 8 / 22











AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB


B

A

Bypassed.












AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB

AB


B

A

Blocked.



Continuing...

• The major challenge with the MMC is mainly a control problem: tomake sure that the capacitor voltages in all the submodules arestrictly controlled.

• The control system is responsible for maintaining the average sum ofinserted submodules in the upper and the lower arm at a constantlevel (voltage balancing).

• The alternating voltage output is obtained by varying the differencebetween the number of inserted submodules in the upper and thelower arms (switching frequencies as low as 150 Hz could beachieved).

• The MMC has low switching losses and its harmonic content is verysmall (very small filters) [8].

• Smaller voltage steps involving only a few semiconductor devices inthe converter arm. Lower stresses on the phase reactor.



HVC Plus

IGBT (wirebond)

Siemens HVDC PLUS components [8].

• For the submodules: off-the-shelfwire-bonded IGBT modules with plasticcases.

• The submodule: half-bridge with abypass thyristor and vacuum switch(wire-bond fail as an open circuit).

• MMC because ABB patented seriesconnection of IGBTs? [3].

• A 400-MW MMC ' 200submodules/converter arm (dependingon voltage level)

• Nov. 2010: 1st MMC for HVdc (TransBay Cable project: 88 km / ± 200 kV /400 MW [8, 9]).



HVC Plus

B

A

Submodule Diagram









HVC Plus

Submodule PictureSiemens HVDC PLUS components [8].








HVC Plus

Submodule PictureSiemens HVDC PLUS components [8].








HVC Plus

California

Livermore

PleasantonSan Leandro

Oakland

Richmond

MartinezConcord

Antioch

Pittsburg

Vallejo

San Rafael

Novato

SanFrancisco

Daly City

Potrero Hill

±200 kV (400 MW)88 km

The Transbay project [9].









HVDC Light

2

1

36

2

1

36

2

1

36

2

1

36

2

1

36

2

1

36

Doublesubmodule

CTL converter diagram

ABB HVDC Light components [8].

• ABB Light 4th generation: CascadedTwo-Level (CTL) converters (directvoltages up to ± 320 kV) [10, 11].

• CLT submodules: half-bridges with a smallnumber ( 8) series-connected IGBTs(press-pack) [12].

• CTL converter configuration ' MMCconfiguration (difference = IGBT switchesinside the submodule).

• July 2010: Awarded connection ofDolWin 1 OWF (Converter: 800 MW / ±320 kV / 75 km submarine cable / 90 kmunderground cable part). Will serve 2 400MW OWFs (2nd to be connected later).



HVDC Light

submodule IGBT (presspack)








HVDC Light

IGBT1

IGBT8

IGBT1

IGBT8

double submodule diagramABB HVDC Light components [8].







HVDC Light

double submodule picture.








HVDC Maxsine

A B

Vdc

SW1

SW2

SW3

SW4

Diagram of a FBS.

Alstom HVDC Maxsine components

[8].

• MMC topology, but might use full-bridgesubmodules (FBS) [13, 3].

• Off-shelf IGBTs: operated (rated) 2 kV(3.3 kV) / 1.2 kA (1.5 kA) for safety.

• FBS: Possible to invert dc-side voltage (forHybrid MTdc networks) and to fullyinterrupt the dc fault currents.

• Submodule: IGBTs, dc capacitor (oil free),gate drives and a fast mechanical bypassswitch.

• 8 submodules form a power module (circa160 kg / 150 x 65 x 30 cm) [3].

• FBS: double the switching devices (morecostly).



HVDC Maxsine

A B

Vdc

SW1

SW2

SW3

SW4

Diagram of a FBS.


[8].









HVDC Maxsine

FBS voltage output

Switches(SW)

Output voltage(vAB)

1 & 4 +Vdc

1 & 3 02 & 4 02 & 3 −Vdc


[8].









HVDC Maxsine

Alstom HVDC Maxsine submodule [3].


[8].









HVDC Maxsine



[8].









HVDC Maxsine



[8].








Voltage-Source Converter Technology VSC Commissioning

Planning VSC-HVdc Links (2-3 years)

Application and Feasibility Studies:

• Comparison of Line-Commutated Converter and VSC

• Economic Justification of a VSC Scheme

• Comparing Alternative Termination Points for the VSC Scheme

• Cable route and voltage options

• Comparing the Selected Scheme with Alternative Solutions

• Preparing an Outline Specification of the VSC Transmission Project

Specification Studies

• Specifying the Performance Requirements for the VSC Scheme

• AC System Data for the Design of the VSC Scheme

Modeling of VSC System

• Load Flow, Short-circuit, Harmonic Performance Modelling

• EM Transient & Stability System Impact Modelling

• Environmental Impact studies: Cable route and termination stations environmental impactanalysis, Audible EMF and EMC


Voltage-Source Converter Technology VSC Commissioning

Commissioning VSC-HVdc (1 year)

All equipment within the scope of supply shall be comprehensively tested and run fortrail period in order to demonstrate that it meets the specified requirements.

Factory Acceptance Tests Equipment QA and component testing

• Factory Acceptance Tests - Equipment QA and component testing

• Control Verification Tests with Real-Time Simulator: TSO and Owner witnessthese tests.

• Commissioning Tests: performed to confirm, without undue disturbance to thepower system, that the system meets the performance specifications.

The on-site inspection and test activities:

• Equipment tests;

• Subsystem tests;

• System tests;

• Trial operation;

• Acceptance tests.


VSC-HVdc Structure

Table of Contents





VSC-HVdc Structure

Typical layout of a VSC-HVdc Station

HVDC Plus station layout [14] Aerial view of the Shoreham VSC [15]

The smaller footprint and more flexible control make VSC-HVdc systemsthe most convenient choice for the connection of offshore wind farms [16,17, 18].


VSC-HVdc Structure


HVDC Plus station layout [14] Aerial view of the Shoreham VSC [15]

VSC-HVdc transmission system: 2 converter stations (nowadays with multi-level topology); an ac transformer; dc-side capacitors; and a dc cable oroverhead line.


VSC-HVdc Structure


AC System Phase

Reactor

circuit breaker

AC Transformer

VSC converter

station

DC Capacitor

DC LINE

Neutral point grounding

AC FILTERS

Single-line diagram of a VSC-HVdc transmission station.

Other components include: ac circuit breakers, surge arresters, ac-harmonicand radio-interference filters, transformers tap-changers, phase reactors, dc-side harmonic filters, dc chopper and braking resistor; and grounding equip-ment [19, 1]. Some of these components are briefly discussed next.


VSC-HVdc Structure AC Circuit Breakers

High-voltage ac circuit breaker [20]

(800 kV, breaking currents up to

63 kA in 2 cycles [20]).

• Triple purpose: 1st it connects the acsystem to the converter station when,during system start-up, the dc-sidecapacitor is charged to the nominalvoltage.

• 2nd: disconnect the converters in caseof a contingency in the ac system.

• 3rd: VSC have no control means forclearing dc-side faults.

• SF6 GIS switchgear substituted traditional oil-based ac breakers.Breaking times: circa 2 to 3 cycles. [21].

• Interruption processes:

1 terminals are mechanically separated creating an electrical arc2 arc is finally quenched once there is a current zero crossing.


VSC-HVdc Structure Transformers

East West Interconnector Transformer.

• Main task: adapt the ac system voltagefor the converter (e.g. 0.8 SPWMmodulation index).

• Helps filtering the currents and limitingfault currents.

• Decouples zero sequence harmonics (ifsecondary side is not grounded).

• Basic design principle (physical dimensions) [21]: Acu,Aco ∝Sn

fJϕB• Main differences with CSC-HVdc transformers [22]:

1 lower insulation requirements: produced ac-side voltage has null dcoffset w.r.t. to ground;

2 lower harmonic content in the current (lower stresses and losses);3 No need for OLTC for control purposes (only to optimise operation and

reduce losses).4 Standard two-winding transformers can be used.


VSC-HVdc Structure Filters

LCL filter

Cf

Lg Lc

HVdc ac-side

capacitor bank

• Negative effects of harmonics [19]:

1 Extra losses (heating) in components;2 Overvoltages due to resonance;3 Inaccuracy or instability of control systems;4 Noise on voice-frequency telephone lines.

• Smaller and cheaper filters (PWM operation)

• Very difficult to optimise filter design (networkharmonic impedances are needed).

• Simpler LCL filter is usually used for low andmedium-power VSC applications [23].

• Theoretically a high-pass filter would be sufficient. Inpractice, 2 or 3 branches of tuned filters may benecessary [21]. A radio interference filter may beneeded to avoid telecommunication disturbances [19].

• The filter capacitors reactive power are usually selectedbetween 10 and 20% the VSC rated power.


VSC-HVdc Structure Filters

Size matters...

Filtering greatly influences the total footprint of a HVdc converter station.

Aerial view (to scale from GoogleEarth) of a HVdc Classic and a VSC-HVdc converter stations.

Furnas HVdc Classic Station from theItaipu transmission system, in Brazil.0.86 km2 (0.86 km x 1 km), 3150 MW,±600 kV [24]

VSC-HVdc station in Diele – part ofthe BorWin1 HVdc transmission systemin Germany. 0.17 km2 (0.33 km x0.50 km), 400 MW, ±150 kV [25].


VSC-HVdc Structure Phase Reactors

HVdc Classic [26]VSC-HVdc[27]

Air coil reactors for HVdc systems

(on LCC dc-side / on VSC ac-side).

Serves the following purposes: [1, 22]:

• reduce the converter ac-side currenthigh frequency content;

• help to avoid sudden change of polaritydue to valve switching;

• decouple active and reactive powercontrol;

• limit amplitude and rate-of-rise ofshort-circuit currents.

• Figures: an 800 kV reactor for UHVdc projects in China [26]; and 350 kVreactor for the Caprivi VSC system in Africa (300 MW, monopole) [27, 28].

• Typical impedance values are in the 0.10 - 0.15 pu range [29, 30].

• Air coil is the most common technology (iron core also possible).

• Main design aspects: rated current, impedance, insulation level, overalllosses, noise levels, thermal and dynamic stability [31].

• A basic scaling rule for air coils is given by [21]: Co,M ∝ In√L


VSC-HVdc Structure DC Capacitors

Dry: 750 µF, 1210 V [32]

Oil: 1600 µF, 2900 V [32]

HVDC Maxsine submodule

[13]

• Energy storage (act as a voltage source).

• Serve as a filter for high frequency currents [1].

• Larger capacitors, lower voltage ripples (costly).

• Basic dimensioning rule: Co,M,Vo ∝ 1

2CdcV

2dc

• Design: essential to consider the transient voltageconstraints.

• Dry metallised film capacitors (more environmentalfriendly, explosion safe, 2x the capacitance in 0.5xthe volume, self-healing, shorter production timeand simpler to install) [21].

• Up to 10x more capacitors are needed in MMCthan 2-level VSC [33].

• Ultimately, the dc capacitor can be characterisedby a time constant [34]: τ = CdcV

2dc

/2Sn


References

[1] J. Arrillaga, Y. Liu, and N. Watson, Flexible Power Transmission: The HVDCOptions. Wiley, 2007, iSBN: 9780470511855.

[2] E. Koldby and M. Hyttinen, “Challenges on the Road to an Offshore HVDC Grid,”in Nordic Wind Power Conference, 2009.

[3] Hans-Peter Nee and Lennart Angquist, “Perspectives on Power Electronics and GridSolutions for Offshore Wind farms,” Elforsk AB, Stockholm, Technical Report,November 2010. [Online]. Available:http://www.elforsk.se/Rapporter/?download=report&rid=10 96

[4] N. Flourentzou, V. Agelidis, and G. Demetriades, “VSC-Based HVDC PowerTransmission Systems: An Overview,” IEEE Transactions on Power Electronics,vol. 24, no. 3, pp. 592–602, Mar. 2009. [Online]. Available:http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=4773229

[5] A. Lesnicar and R. Marquardt, “An innovative modular multilevel converter topologysuitable for a wide power range,” in Power Tech Conference Proceedings, 2003 IEEEBologna, vol. 3, 2003, pp. 1–6.

[6] M. Glinka and R. Marquardt, “A new AC/AC multilevel converter family,” IEEETransactions on Industrial Electronics, vol. 52, no. 3, pp. 662–669, 2005.


http://www.elforsk.se/Rapporter/?download=report&rid=10_96_

http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=4773229

References

[7] S. Allebrod, R. Hamerski, and R. Marquardt, “New transformerless, scalableModular Multilevel Converters for HVDC-transmission,” in IEEE Power ElectronicsSpecialists Conference (PESC), 2008, pp. 174–179.

[8] J. Dorn, H. Gambach, and D. Retzmann, “HVDC transmission technology forsustainable power supply,” in 9th International Multi-Conference on Systems,Signals and Devices (SSD), 2012, pp. 1–6.

[9] R. Adapa, “High-Wire Act,” IEEE Power and Energy Magazine, pp. 18–29,Nov./Dec. 2012.

[10] L. Harnefors, A. Antonopoulos, S. Norrga, L. Angquist, and H.-P. Nee, “Dynamicanalysis of modular multilevel converters,” IEEE Transactions on IndustrialElectronics, vol. 60, no. 7, pp. 2526–2537, 2013.

[11] B. Jacobson, P. Karlsson, G. Asplund, L. Harnefors, and T. Jonsson, “VSC-HVDCtransmission with cascaded two-level converters,” in Cigree Session B4. Paris:Cigre, 2010, pp. 1–8. [Online]. Available:http://www05.abb.com/global/scot/scot221.nsf/veritydisplay/422dcbc564d7a3e1c125781c00507e47/$file/b4-110 2010%20-%20vsc-hvdc%20transmission%20with%20cascaded%20two-level%20converters.pdf

[12] J. Hafner and B. Jacobson, “Proactive hybrid HVDC breakers – A key innovationfor reliable HVDC grids,” Paper presented at the International Symposium on


http://www05.abb.com/global/scot/scot221.nsf/veritydisplay/422dcbc564d7a3e1c125781c00507e47/$file/b4-110_2010%20-%20vsc-hvdc%20transmission%20with%20cascaded%20two-level%20converters.pdf



References

Integrating supergrids and microgrids. Bologna: CIGRE, September 2011, paper264, pp. 1–8. [Online]. Available:http://search.abb.com/library/Download.aspx?DocumentID=9AKK105408A3383&LanguageCode=en&DocumentPartId=&Action=Launch

[13] Alstom Grid, “HVDC-VSC: transmission technology of the future,” France, THINKGRID #08: Magazine, 2011. [Online]. Available: http://www.alstom.com/Global/Grid/Resources/Documents/Smart%20Grid/Think-Grid-08-%20EN.pdf

[14] D. Retzmann. HVDC Station Layout, Equipment LCC & VSC and Integration ofRenewables using HVDC. Cigre Tutorial 2012 – Last Accessed: 07 September, 2013.[Online]. Available:http://www.ptd.siemens.de/Cigre AUS 2011 HVDC & GridAccess tutorial Re.pdf

[15] Railing, BD and Miller, JJ and Steckley, P and Moreau, G and Bard, P andRonstrom, L and Lindberg, J, “Cross Sound cable project–Second generation VSCtechnology for HVDC,” in Cigre conference, Session B4, France, Aug 2004. [Online].Available: http://www05.abb.com/global/scot/scot221.nsf/veritydisplay/c062bb7ae84536fec1256fda004c8cc9/$file/b4-102.pdf

[16] P. Bresesti, W. L. Kling, R. L. Hendriks, and R. Vailati, “HVDC Connection ofOffshore Wind Farms to the Transmission System,” IEEE Transactions on EnergyConversion, vol. 22, no. 1, pp. 37–43, mar 2007. [Online]. Available:http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=4105997


http://search.abb.com/library/Download.aspx?DocumentID=9AKK105408A3383&LanguageCode=en&DocumentPartId=&Action=Launch


http://www.alstom.com/Global/Grid/Resources/Documents/Smart%20Grid/Think-Grid-08-%20EN.pdf

http://www.alstom.com/Global/Grid/Resources/Documents/Smart%20Grid/Think-Grid-08-%20EN.pdf

http://www.ptd.siemens.de/Cigre_AUS_2011_HVDC_&_GridAccess_tutorial_Re.pdf

http://www05.abb.com/global/scot/scot221.nsf/veritydisplay/c062bb7ae84536fec1256fda004c8cc9/$file/b4-102.pdf

http://www05.abb.com/global/scot/scot221.nsf/veritydisplay/c062bb7ae84536fec1256fda004c8cc9/$file/b4-102.pdf

http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=4105997

References

[17] W. Kling, R. Hendriks, and J. H. den Boon, “Advanced transmission solutions foroffshore wind farms,” Power and Energy Society, Nov. 2008. [Online]. Available:http://ieeexplore.ieee.org/xpls/abs all.jsp?arnumber=4596257

[18] A. van der Meer, R. Hendriks, and W. Kling, “Combined stability andelectro-magnetic transients simulation of offshore wind power connected throughmulti-terminal VSC-HVDC,” in IEEE Power and Energy Society General Meeting.IEEE, 2010, pp. 1–7. [Online]. Available:http://ieeexplore.ieee.org/xpls/abs all.jsp?arnumber=5589619

[19] E. W. Kimbark, Direct current transmission. Wiley-Interscience, 1971, vol. 1,iSBN: 9780471475804.

[20] Siemens AG Energy Sector, “High-Voltage Circuit Breakers: from 72.5 kV up to800 kV,” , Erlangen, Germany, Technical Report, 2012. [Online]. Available:http://www.energy.siemens.com/hq/pool/hq/power-transmission/high-voltage-products/circuit-breaker/Portfolio en.pdf

[21] Staffan Norrga, “Power Electronics for High Voltage Direct Current (HVDC)Applications,” ABB Sweden, Conference Tutorial: 14th European Conference onPower Electronics and Applications (EPE 2011), Birmingham, 2011.


http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4596257

http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=5589619

http://www.energy.siemens.com/hq/pool/hq/power-transmission/high-voltage-products/circuit-breaker/Portfolio_en.pdf

http://www.energy.siemens.com/hq/pool/hq/power-transmission/high-voltage-products/circuit-breaker/Portfolio_en.pdf

References

[22] B. Gemmell, J. Dorn, D. Retzmann, and D. Soerangr, “Prospects of multilevel VSCtechnologies for power transmission,” in Proceedings of IEEE/PES Transmission andDistribution Conference and Exposition, Chicago, April 2008, pp. 1–16.

[23] M. Liserre, F. Blaabjerg, and A. Dell’Aquila, “Step-by-step design procedure for agrid-connected three-phase PWM voltage source converter,” International journal ofelectronics, vol. 91, no. 8, pp. 445–460, 2004.

[24] M. P. Bahrman, DIRECT CURRENT POWER TRANSMISSION, ed., ser.Standard Handbook for Electrical Engineers. Mcgraw-hill, 2006, ch. 15, pp. 1–34,iSBN: 9780071491495.

[25] J. Glasdam, J. Hjerrild, L. Kocewiak, and C. Bak, “Review on multi-level voltagesource converter based HVDC technologies for grid connection of large offshore windfarms,” in IEEE International Conference on Power System Technology(POWERCON), Auckland, November 2012, pp. 1–6.

[26] BPEG Reactors. BPEG’s smoothing reactors for ± 800kV HVDC Projects. LastAccessed: 07 September, 2013. [Online]. Available:http://www.bpeg-usa.com/News.html

[27] Coil Innovation Power Inductors. CI awarded contract for “Caprivi” HVDCSmoothing Reactors. Last Accessed: 07 September, 2013. [Online]. Available:http://www.coilinnovation.com/company/news-archive/?L=1


http://www.bpeg-usa.com/News.html

http://www.coilinnovation.com/company/news-archive/?L=1

References

[28] ABB. Caprivi Link Interconnector. Last Accessed: 07 September, 2013. [Online].Available: http://www.abb.nl/industries/ap/db0003db004333/86144ba5ad4bd540c12577490030e833.aspx

[29] B. R. Andersen. VSC Transmission. CIGRE B4 HVDC and Power Electronics.HVDC Colloquium, Oslo, April 2006 – Last Accessed: 07 September, 2013. [Online].Available: http://www.andersenpes.talktalk.net/VSC%20Transmission%20Oslo%20-%20final.pdf

[30] M. Liserre, F. Blaabjerg, and S. Hansen, “Design and control of anLCL-filter-based three-phase active rectifier,” IEEE Transactions onIndustryApplications, vol. 41, no. 5, pp. 1281–1291, 2005.

[31] Siemens Energy Sector, Transformers, 7th ed., ser. Power Engineering Guide.Siemens AG, 2006, ch. 5, pp. 232–261.

[32] EPCOS, “Power Capacitors,” , Munich, Germany, Product Brief, 2011. [Online].Available: http://www.epcos.com/inf/20/50/MKK PB.pdf

[33] TDK. Minimizing energy losses during longdistance transmission of electricalenergy. Last Accessed: 07 September, 2013. [Online]. Available:http://www.global.tdk.com/csr/csr highlights/csr04600.htm#anchor 01


http://www.abb.nl/industries/ap/db0003db004333/86144ba5ad4bd540c12577490030e833.aspx

http://www.abb.nl/industries/ap/db0003db004333/86144ba5ad4bd540c12577490030e833.aspx

http://www.andersenpes.talktalk.net/VSC%20Transmission%20Oslo%20-%20final.pdf

http://www.andersenpes.talktalk.net/VSC%20Transmission%20Oslo%20-%20final.pdf

http://www.epcos.com/inf/20/50/MKK_PB.pdf

http://www.global.tdk.com/csr/csr_highlights/csr04600.htm#anchor_01

References

[34] M. M. d. Oliveira, “Power Electronics for Mitigation of Voltage Sags and ImprovedControl of AC Power Systems,” PhD. Thesis, Royal Institute of Technology, 2000.[Online]. Available: http://kth.diva-portal.org/smash/record.jsf?pid=diva2:8765


http://kth.diva-portal.org/smash/record.jsf?pid=diva2:8765

Challenges on the road to future MTdcnetworks

Part III






System Integration

Table of Contents

1 System Integration

2 Power Flow Control

3 Dynamic Behaviour

4 Stability

5 Fault Behaviour

Dr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: Part III August 25, 2014 2 / 32

System Integration Description

Introduction

• AC infrastructure to generate, transmit and distribute electricity is theresult of evolution and improvements for > 100 years.

• Italy (1886): 1st line from the Cerchi (l’Aventino) thermoelectricpower plant to Rome. Ganz system: 300 HP / 2 kV / 30 km [1, 2].

• MTdc networks will have to organically grow, in a much shorter time,from simple phases → complex meshed topologies [3].

• The North Seas Countries Offshore Grid Initiative: 40-70 GW ofoffshore wind power in Northwest Europe up to 2030.

• It is extremely important to choose the most suitable systemarchitecture from the start [4].

• Next step: distinguish clearly the primary functions of system modulesand submodules, and how they interact.

• System integration: establishing the functionalities inside the systemand assuring optimal performance [5].


System Integration Performed Work

Performed Work on System Integration

• Analyse what has prevented the development of MTdc networks inthe past.

• Perform a review of HVdc transmission technology, andconfigurations, to identify the best option for the development of aMTdc network in the North Sea.

• Compare different topologies for construction of MTdc networks.

• Establish when the transmission system of an offshore wind farmshould be done in direct current.

• Study the role and impacts of system architecture on the developmentof MTdc networks for offshore wind applications;

• Form a LVdc network by connecting more than two power electronicconverter on their dc sides.



System Architecture

According to Baldwin [6]: “Modularity is the practice of building complexsystems or processes from smaller subsystems that can be designedindependently yet function together as a whole”.

NSTG Global Design Rules

HVDC Stations Design rules

DC Transmission Cables

Protection SystemsWind FarmsDesign rules

Turbine Technology

ControlSystem

MV System

Tower

InterfaceInterfaceSystem

Integration & Communication

Plattaform

Cable Technology

MechanicProtection

Circuit Breakers

SwitchesControl Strategy

Filters

Converter Technology

Possible design hierarchy for the NSTG system with four modules.



Some of the MTdc network development phases foreseen by the NSTG project [7].

Phase 1 Phase 3

Phase 5 Phase 10


System Integration Conclusion

• MMC VSC-HVdc transmission systems are currently the best optionfor the development of MTdc networks.

• Best starting configuration: parallel-radial topology using a symmetricmonopolar configuration. This will allow the system to developmodularly, will keep initial capital costs down and will make it easierto face dc contingencies.

• Not all OWFs should be connected with dc by default. Decision:technical assessment + thorough economic analysis.

• MTdc networks need to embrace a modular architecture.

• HV dc-dc converters would allow a more modular development.

• Before construction starts: system designers have to establish theglobal design rules, the performance standards, and the interfacesbetween components (independent TSO: responsible for an offshoregrid code and a road map).

• Harder on HVdc networks than in low or medium-voltage (multipleinternational stakeholders, larger scale and capital costs). Therefore,dc microgrids & smartgrids more likely to appear before dc supergrids.


Power Flow Control

Table of Contents



3 Dynamic Behaviour

4 Stability

5 Fault Behaviour


Power Flow Control Description

Introduction

• Controlling exactly how power flows in a transmission system (ac ordc) is not an easy task.

• Principle of least action [8], Feynman [9]:“if currents are [...] obeying Ohm’s law, the currents distribute themselves [...] so that therate at which heat is generated is as little as possible.”

• One of most important factors for the development of MTdc grids.

• AC control: FACTS devices which vary a transmission linecharacteristic impedance or its voltage phase angle (limited) [10].

• DC Control: voltages do not possess a phase angle; and thetransmission line impedances are purely resistive (steady state).

• Only variables for power flow control: voltages/currents amplitudes.

• For improved performance, reliability and safety reasons: MTdcnetworks will require a distributed power flow control strategy (directvoltage control shared amongst several nodes).


Power Flow Control Performed Work

Performed Work on Power Flow Control

• Develop a load flow algorithm to analyse the power flow in MTdcnetworks with more than one slack node.

• Compare the most common control strategies for MTdc networks.

• Evaluate how to optimise the power flow in MTdc networks.

• Develop a novel direct voltage control strategy capable of reliably andsecurely operating a MTdc network during steady-state and faultscenarios.

• Test the novel control strategy on the developed LV-MTdc laboratorysetup.


Power Flow Control Performed Work

Desirable: fast dynamic behaviour; high flexibility, high expandability andlow communication requirements.

dcv

0 +1-1 InverterRectifier

mindcv

maxdcv

0dcv

dci

1.05

1.00

0.95

Low flexibility

dcv


mindcv

max2dcv

max1dcv

dci

1.05

0.97

1.02

Low Expandability

dcv


mindcv

maxdcv

*1dcv

min2dcv

dci

1.05

1.00

0.95

1.02

Low Flexibility

1.02

dcv

0InverterRectifier

dcp

liminvp

*2high

dcv

*acp

maxdcv

limrecp

0.98

1.05

*1dcv

1.00A

*2low

dcv

Low Dynamic PerformanceDr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: Part III August 25, 2014 11 / 32

Power Flow Control Conclusions

• A dc load flow algorithm is crucial for developing power flow control.• In HV MTdc networks power flow control should be accomplished

indirectly through the control of the system direct voltages (increasesreliability).

• Common methods for controlling HVdc transmission system were notdeveloped with expansion and flexibility as their prime objective.

• Nodes which are not connected to generating plants, should beworking inside the system as slack nodes.

• Power flow control can minimise the system losses, maximise its socialwelfare or any other TSO need (multi-objective optimisation)

• The distributed voltage control strategy was developed withexpansion and flexibility as it prime objective.

• The DVC strategy needs communication in control cycles in the orderof minutes (fast telecom links not needed).

• The communication is only performed between a node and thetransmission system operator, and not amongst nodes.

• Without communication, the development of microgrids, smartgridsor supergrids will be deprived of strength and efficiency.


Dynamic Behaviour

Table of Contents



3 Dynamic Behaviour

4 Stability

5 Fault Behaviour


Dynamic Behaviour Description

Introduction

• AC networks: main dynamic component is the synchronous machine(active/reactive power) [11]. Dynamic phenomena: 5 - 30 s [12].

• DC networks: main dynamic component is the power electronicconverter. Dynamic phenomena: < 100 ms (5 cycles).

• Converters have much faster response (control capabilities) but alsomuch lower inertia

• Machines: H = Jω2/2Sn = 1-10 s [12] vs. 20 ms for a 1 GW /±320 kV / 100 µF VSC-HVdc: τ = CdcV

2dc/2Sn (50 to 500x lower).

• Modelling of the power electronic converters is key for understandingthe dynamics of MTdc networks.

• Switching behaviour: dynamic equations describing the converter arediscontinuous and complex to solve. Solution: averaged models [13].

• Dynamic models are needed to assess MTdc network behaviour duringsound/fault conditions, to study interactions between VSCs and othercomponents (e.g. OWFs), but also for designing control strategies.


Dynamic Behaviour Description

Performed Work on Dynamic Behaviour

• Create non-linear dynamic models of the most important modulesinside MTdc networks: i.e. the offshore wind farms, the VSC-HVdcconverters, the dc cables and control systems.

• Establish how the models are connected to each other based on theirsignal flow.

• Analyse the control structure of a VSC-HVdc converter station.

• Test and validate the developed models on real power electronicconverters.


Dynamic Behaviour Conclusions

• Dynamic models: to determine the dynamic behaviour of thecomplete MTdc system.

• Tradeoff between computational time and level of detail (modelpurpose).

• Averaged non-linear models: sufficient for control strategies design.

• Modular models: identifying how the different system modules areconnected to each other is more important than the module level ofdetail.

• The most important model: the power electronic converters. Theydecouple the offshore wind farms and ac network dynamics from thedynamics of the dc grid.

• Experiments: VSC dynamics are fast (Active/reactive power anddirect voltage step response transients < 100 ms and accurate.)

• With proper power electronics control design dynamic behaviourissues will not hamper the development of MTdc networks.


Stability

Table of Contents



3 Dynamic Behaviour

4 Stability

5 Fault Behaviour


Stability Description

Introduction

• Stability: the ability to reach a new equilibrium state – or operatingpoint – after being subject to a disturbance.

• In ac networks: rotor-angle, voltage and frequency stability [14].

• Rotor-angle: small-disturbance and transient stability. Both intimatelyrelated to the dynamic behaviour of synchronous generators.

• Voltage: associated with system capability to supply reactive powerfor a given demand (usually happens in highly loaded areas whenreactive-power support is insufficient [11]).

• Frequency stability: i.e. the system capacity to maintain generationand load in equilibrium, and to regain equilibrium after a severedisturbance while losing a minimum amount of loads.

• DC networks: stability is observed only through the system voltages(different analysis).

• Must include dc grid passive components, but also the powerelectronic converters and their feedback controllers [15].


Stability Performed Work

Performed Work on Stability

• Development of linear small-signal models of the MTdc networkcomponents for stability analysis.

• Analysis of the system eigenvalues and their sensibility to parametricvariations.


Stability Conclusions

• MTdc network: stable if the sum of the active power entering thesystem is equal the sum of the power exiting the system.

• More slack nodes ⇒ more stable dc network. Nodes controlling theMTdc network direct voltage do so by making sure the power insidethe system is balanced.

• Small-signal model: dc networks are marginally stable systems.Feedback control of the converter direct voltage move poles awayfrom the unstable region of the RLP.

• Small-signal stability of MTdc networks is dictated by the stability ofits power electronic converters.

• VSC eigenvalues are resilient from parametric variations in the valuesof its hardware components.

• Variations the LCL filter and output dc capacitor deteriorates dynamicperformance deteriorates but the converters eigenvalues remain stable.

• Careful consideration should be made when designing and tuning theconverter controllers, specially the reactive power and the PLLcontrollers, which relate with voltage stability on ac-side.


Fault Behaviour

Table of Contents



3 Dynamic Behaviour

4 Stability

5 Fault Behaviour


Fault Behaviour Description

Introduction

• 2 main types of fault scenarios can take place in MTdc networks.

• 1st: faults can occur on the power electronic converters ac side(single or multi-phase)

• They will represent a loss of generation or load to the dc network.

• It is imperative that a contingency in one ac power system is notpropagated to another, isolated ac system, through the dc grid.

• 2nd: faults can happen on the power electronic converters dc side.

• Much more challenging to handle than ac faults [16].

• All ac systems will contribute to the dc fault current and, because ofthe dc cables low impedances, the voltages in the MTdc will besubstantially reduced, nearly stopping the power flow.

• The development of protection in dc networks – specially for HVdcnetworks – is a critical issue [17].


Fault Behaviour Performed Work

Performed Work on Fault Behaviour

• Use the develop models to study the behaviour of MTdc networksduring ac and dc contingencies.

• Investigate the influence of grid code requirements on MTdc networksduring ac contingencies.

• Analyse the impact of the HVdc transmission system configuration ondc contingencies management.



AC Faults

Cdc1

ZT

AC network 1

AC network 2

VSC3

VSC4

VSC1

VSC2

Pdc2

OWF2

ZT Pdc4

Pdc1

Cdc2

Cdc3

Cdc4

Pdc3

ZT

ZT

Rdc1

Rdc2

Rdc3

Rdc4

Vdc5

Vdc1 Vdc3

Vdc4Vdc2

OWF1

80 km 250 km

150 km50 km

4-node parallel VSC-MTdc network under study.



AC Faults

4.5 4.6 4.7 4.8 4.9 5 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 60.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

time [s]

Vac

[p

u]

|e2|

LR

LR

HR

HR

|vf2|

|vc2|

|vf2|

|vc2|

VSC voltage|vc2|High GCR

filter voltage|vf2|High GCR

VSC voltage|vc2|low GCR

filter voltage|vf2|High GCR

AC grid 2 voltage profile during the dip. The graphic shows the voltage module at theconverter terminals, (|vc2|), and at the ac filter capacitance, (|vf 2|), for the low and high

grid code requirements.



AC Faults

1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 70.8

0.85

0.9

0.95

1

1.05

1.1

1.15

1.2

1.25

1.3

time [s]

Vd

c [p

u]

VSC 1

VSC 2

WF 1

WF 2

Scenario i

1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 70.8

0.85

0.9

0.95

1

1.05

1.1

1.15

1.2

1.25

1.3

time [s]

Vd

c [p

u]

Scenario iiDirect voltage of all converters during the ac fault simulation.


Fault Behaviour Conclusions

• DC microgrids or supergrids will have to remain operational whenfaced with contingencies.

• With the correct provisions a MTdc dc network can be resilientagainst ac or dc contingencies.

• AC faults: the VSC terminals have to quickly restore the active powerbalance inside the MTdc network. Faults in nodes controlling thesystem direct voltage can be very demanding (a sudden active powersurplus can bring the direct voltage above 1.2 pu in < 10 cycles).

• AC grid code requirements influence the MTdc network behaviourduring the fault (negative impact on all the converters in the MTdcnetwork and, consequently, also on the other interconnected acnetworks).

• There are several solutions to deal with this issue: load shedding, fastpower reduction methods, installation of dc choppers at the windfarms, overrating the power electronic converters, or making the acgrid codes less stringent with their reactive power requirements.



DC Breakers

DC breakers (needed) isolate the fault and resume normal operationwithout interruptions.

Mechanical breaker

LC-resonance path

Surge arrestor

A

B

C

Resonant dc breaker circuit [18, 19].

PTC resistorLoad switch

UFS / Breaker

GT

O

C

B

A

Fast disconnector

Hybrid II dc breaker circuit [20].

Fast disconnector

Auxiliary dc breaker

Main dc breaker

Limiting reactor

Residual dc current

breaker

Hybrid I dc breaker circuit [17].

Surge arrestors

IGBT and diode main path

Solid-state dc breaker circuit [21].



DC Fault Currents

Current response to a pole-to-ground on MTdc line 2 for the ground and metallic returnconfigurations.

-4

-2

0

2

4

6

8

10

12

Cu

rren

t [p

u]

Fault currentDC link capacitor currentStation 1 contributionStation 2 contributionStation 3 contributionStation 4 contributionLine 1,3,4 contributionsLine 2 contribution

710690 695 700 705 715 720 725 730time [ms]

Symmetric monopole

-5

0

5

10

15

20

Cu

rren

t [p

u]

690 695 700 705 715 720 725 730time [ms]

710

Bipolar Metallic Return



DC Fault Limiting Reactors

Peak dc fault currents as a function of the dc limiting reactor size.

3.563.41

2.552.43

3.132.99

2.122.00

2.632.39

1.551.43

2.40

2.01

1.281.15

2.19

1.79

0.990.86

0

1

2

3

4

VSC1 VSC2 VSC3 VSC4

Ipea

k [

pu

]

1 mH10 mH50 mH100 mH200 mH

Symmetric monopole

6.47 6.41

5.47 5.335.53 5.38

4.54 4.394.38 4.12

3.26 3.24

4.153.71

3.23

3.853.56

3.22

0

1

2

3

4

5

6

7

VSC1 VSC2 VSC3 VSC4

Ground Return

8.58

4.45

3.56 3.43

6.64

3.87

2.9 2.82

5.15

3.32.97 2.93

4.64

3.192.57 2.46

3.12.46 2.39

0123456789

Ipea

k [

pu

]

VSC1 VSC2 VSC3 VSC4

Metallic Return

11.54

22.81

17.20

10.17

19.53

14.49

8.00

14.80 14.20

6.88

13.81 13.56

6.76

13.4912.51

0

5

10

15

20

25

Symmetric Ground Return Metallic Return

Total Fault Current



DC Fault Limiting Reactors

Overcurrent protection triggering time as a function of the dc limitingreactor size.

1.65

5.76.2

6.5

1.9

66.6 6.6

2.7

6.26.65 6.7

3.7

6.97.4 7.5

5.55

8.6 8.8 8.9

0123456789

VSC1 VSC2 VSC3 VSC4

Tim

e [m

s]

Metallic Return: Bipolar and Asymmetric Monopolar



DC Faults

• The situation regarding MTdc network fault changes when it comesto dc contingencies.

• There are not many options for dealing with dc fault situations.

• The development of dc circuit breakers is still incipient and the onlytwo breakers – the solid-state breaker and the ABB Hybrid breaker –can act fast enough to avoid interruption of the MTdc networkoperation.

• Nevertheless, they need to improve their ratings before they can beused in practice.

• Until the dc switch breakers are commercially available, the mostwell-suited solution to counteract dc contingencies is to build theMTdc network using a symmetric monopolar configuration inconjunction with direct current limiting reactors.


References

[1] Enciclopedia Italiana, “Energia Elettrica,” Trecanni, Last Accessed on 03 February2013. [Online]. Available:http://www.treccani.it/enciclopedia/energia-elettrica (Enciclopedia Italiana)/

[2] R. Nichols, “The first electric power transmission line in North America-OregonCity,” IEEE Industry Applications Magazine, vol. 9, no. 4, pp. 7 – 10, July-Aug.2003.

[3] North Seas Countries’ Offshore Grid Initiative, “Final report - grid configuration,”NSCOGI, Technical Report, November 2012. [Online]. Available:http://www.benelux.int/NSCOGI/NSCOGI WG1 OffshoreGridReport.pdf

[4] K. T. Ulrich and S. D. Eppinger, Product Design and Development, 4th ed.McGraw-Hill, 2008, ISBN: 978-0-07125947-7.

[5] C. Y. Baldwin and K. B. Clark, Design Rules. Vol. I: The power of Modularity.Cambridge, MA: MIT Press, 2000, ISBN: 0-262-02466-7.

[6] C. Baldwin and K. Clark, Design Rules: The power of modularity. Massachusetts:MIT Press, 2000, iSBN: 978-0262024662.

[7] J. Pierik. (2013) North Sea Transnational Grid - A better way to integrate largescale Offshore Wind Power. Last Accessed: 09 November, 2013. [Online]. Available:www.nstg-project.nl


http://www.treccani.it/enciclopedia/energia-elettrica_(Enciclopedia_Italiana)/

http://www.benelux.int/NSCOGI/NSCOGI_WG1_OffshoreGridReport.pdf

www.nstg-project.nl

References

[8] J. Hanc and E. F. Taylor, “From conservation of energy to the principle of leastaction: A story line,” American Journal of Physics, vol. 72, no. 4, pp. 514–521, April2004. [Online]. Available: http://www.eftaylor.com/pub/energy to action.html

[9] R. P. Feynman, The Feynman Lectures on Physics, Sixth printing ed.Massachusetts: Addison-Wesley Publishing, February 1977, vol. 2, ch. 19, p. 14,ISBN: 0-201-02010-6-H.

[10] N. G. Hingorani and L. Gyugyi, Understanding Facts: Concepts and Technology ofFlexible AC Transmission Systems, R. J. Herrick, Ed. New York: Wiley - IEEEPress, 2000, ISBN: 0.7803-3455-8.

[11] P. Kundur, Power System Stability and Control, ser. EPRI Power SystemEngineering Series. McGraw-Hill, 1994, ISBN 0-07-035958-X.

[12] D. P. Kothari and I. J. Nagrath, Modern Power System Analysis, 3rd ed.McGraw-Hill, 2003, ch. 12, pp. 433–455, ISBN: 0-201-02010-6-H. [Online].Available:http://highered.mcgraw-hill.com/sites/dl/free/0073404551/392048/chapter12.pdf

[13] P. Bauer, “Dynamic Analysis of Three-Phase AC Converters,” PhD. Thesis, DelftUniversity of Technology, Delft, January 1995, ISBN: 90-9007789. [Online].Available: http://repository.tudelft.nl


http://www.eftaylor.com/pub/energy_to_action.html

http://highered.mcgraw-hill.com/sites/dl/free/0073404551/392048/chapter12.pdf

http://repository.tudelft.nl

References

[14] P. Kundur et al, “Definition and Classification of Power System Stability,” IEEETransactions on Power Systems, vol. 19, no. 2, pp. 1387–1401, May 2004.

[15] L. Zhang, “Modeling and control of VSC-HVDC links connected to weak acsystems,” PhD Thesis, Royal Institute of Technology, Electrical Engineering Dept.,2010. [Online]. Available: http://www.ee.kth.se/php/modules/publications/reports/2010/TRITA-EE 2010 022.pdf

[16] R. Adapa, “High-Wire Act,” IEEE Power and Energy Magazine, pp. 18–29,Nov./Dec. 2012.

[17] J. Hafner and B. Jacobson, “Proactive hybrid HVDC breakers – A key innovationfor reliable HVDC grids,” Paper presented at the International Symposium onIntegrating supergrids and microgrids. Bologna: CIGRE, September 2011, paper264, pp. 1–8. [Online]. Available:http://search.abb.com/library/Download.aspx?DocumentID=9AKK105408A3383&LanguageCode=en&DocumentPartId=&Action=Launch

[18] B. Pauli, G. Mauthe, E. Ruoss, G. Ecklin, J. Porter, and J. Vithayathil,“Development of a high current HVDC circuit breaker with fast fault clearingcapability,” Power Delivery, IEEE Transactions on, vol. 3, no. 4, pp. 2072–2080,1988.


http://www.ee.kth.se/php/modules/publications/reports/2010/TRITA-EE_2010_022.pdf




References

[19] D. Andersson and A. Henriksson, “Passive and Active DC Breakers in the ThreeGorges-Changzhou HVDC Project,” in International Conference of Power Systems(ICPS), Wuhan, China, 13-15 September 2001, pp. 391–395.

[20] H. W. Steurer M., Frohlich K. and K. K., “A novel hybrid current-limiting circuitbreaker for medium voltage: principle and test results,” Power Delivery, IEEETransactions on, vol. 18, no. 2, pp. 460–467, 2003.

[21] C. Franck, “HVDC Circuit Breakers: A Review Identifying Future Research Needs,”IEEE Trans. on Power Delivery, vol. 26, no. 2, pp. 998–1007, 2011.


Optimal load-flow control ofMTdc networks

Part IV






Load flow in MTdc Networks

Table of Contents

1 Load flow in MTdc NetworksDC Load FlowExample of Load flow in a MTdc NetworkMTdc Network Security (N-1 Analysis)Conclusions

2 The Distributed Voltage Control StrategyControl MethodologyOptimal power flow for MTdc NetworksOptimisation Method: Steepest DescentCase Study: MTdc Losses MinimisationConclusions

Dr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: Part IV August 25, 2014 2 / 23

Load flow in MTdc Networks DC Load Flow

• Performing dynamic simulations for complex MTdc networks isdifficult and time consuming: Load flow analysis is a good initialapproach.

1 flow-based market simulations2 wind integration cost-benefit analysis3 design of high-level control strategies4 N-1 security assessments.

Let’s start from the classic ac load flow (Newton-Rhapson):x(k + 1) = x(k) + ∆x(k)

∆x(k) = −J(k)−1 · g(x(k))

where;x is the state variable vector;J is the Jacobian matrix and;g is the mismatch vector.



• Performing dynamic simulations for complex MTdc networks isdifficult and time consuming: Load flow analysis is a good initialapproach.

1 flow-based market simulations2 wind integration cost-benefit analysis3 design of high-level control strategies4 N-1 security assessments.

Let’s start from the classic ac load flow (Newton-Rhapson):x(k + 1) = x(k) + ∆x(k)

∆x(k) = −J(k)−1 · g(x(k))

where;x is the state variable vector;J is the Jacobian matrix and;g is the mismatch vector.



AC Load Flow – State variables are node phase angles and nodal voltages:

x = [δi , ..., δN−1,Vi , ...,VN−1]T

The mismatch vector (load flow equations):

g(x(k)) =[gP1 , ..., gPN−1

, gQ1 , ..., gQN−1

]T

gPi

= PGi− PLi −

N∑j=1

Vi Vj Yij cos(δi − δj − θij)

gQi= QGi

− QLi −N∑j=1

Vi Vj Yij sin(δi − δj − θij)

where;gPi is the active power mismatch at node i [W]; gQi is the reactive power mismatch atnode i [VAr]; PGi is the active power generation at node i [W]; PLi is the active powerconsumption (load) at node i [W]; QGi is the reactive power consumption at node i[VAr]; QLi is the reactive power consumption (load) at node i [VAr]; Vi is the systemvoltage at node i [V]; δi is phase angle at node i [rad]; Yij is the admittance matrix valueat position ij [S] and; θij is the admittance matrix angle at position ij [rad].



AC Load Flow – State variables are node phase angles and nodal voltages:

x = [δi , ..., δN−1,Vi , ...,VN−1]T

The mismatch vector (load flow equations):

g(x(k)) =[gP1 , ..., gPN−1

, gQ1 , ..., gQN−1

]T

gPi= PGi

− PLi −N∑j=1

Vi Vj Yij cos(δi − δj − θij)

gQi= QGi

− QLi −N∑j=1

Vi Vj Yij sin(δi − δj − θij)

where;gPi is the active power mismatch at node i [W]; gQi is the reactive power mismatch atnode i [VAr]; PGi is the active power generation at node i [W]; PLi is the active powerconsumption (load) at node i [W]; QGi is the reactive power consumption at node i[VAr]; QLi is the reactive power consumption (load) at node i [VAr]; Vi is the systemvoltage at node i [V]; δi is phase angle at node i [rad]; Yij is the admittance matrix valueat position ij [S] and; θij is the admittance matrix angle at position ij [rad].



DC Load Flow – State variables are simplified (no phase angles) [1]:

x = Vdc = [Vdci , ...,VdcN−1]T

where;Vdc is the MTdc network voltage vector [V] and; Vdci is the MTdcnetwork voltage at node i [V].

The mismatch equations are also simplified (no reactive power):

gp(Vdc(k)) =[gP1 , ..., gPN−1

]Twhere;gP is the active power mismatch vector [W].

The load flow equations become (in steady state cable are only resistive):

gPi= PGi − PLi −

∑j 6=i

Vdci Vdcj Yij − YiiVdci2







gp(Vdc(k)) =[gP1 , ..., gPN−1




∑j 6=i








gp(Vdc(k)) =[gP1 , ..., gPN−1




∑j 6=i




Next step: Calculate the dc Jacobian matrix:

J(Vdc(k)) =∂gP∂Vdc

=

∂gP1

∂Vdc1...∂gPN−1

∂Vdc1

. . .

. . .

· · ·

∂gP1

∂VdcN−1...∂gPN−1

∂VdcN−1

∂gPi∂Vdck

=

−YikVdci for k 6= i−∑j 6=i

Vdcj Yij − 2YiiVdci for k = i

The voltage difference between two load flow iterations, ∆Vdc(k):

∆Vdc(k) = −J(Vdc(k))−1 · gP(Vdc(k))

Finally, the direct voltages are updated:

Vdc(k + 1) = Vdc(k) + ∆Vdc(k)





=

∂gP1


∂Vdc1

. . .

. . .

· · ·

∂gP1


∂VdcN−1

∂gPi∂Vdck

=











=

∂gP1


∂Vdc1

. . .

. . .

· · ·

∂gP1


∂VdcN−1

∂gPi∂Vdck

=









The MTdc network admittance matrix, Y:Y = (IM)T · Yp · IM

Ypii =1

Rdci

Ypij = 0

where;IM is the incidence matrix;Yp is the primitive admittance matrix and;Rdci is the resistance of MTdc network line i [Ω].



Flowchart of the dc load flow algorithm [2]

Gather MTdc network data

Build incidence matrix IM andprimitive admittance matrix Yp

Calculate dc system bus admit-tance matrix Y according to (7)

Set PGiand PLi on all nodes

Initialise Vdc(k)k = 1;

Build Jacobian MatrixJ(Vdc(k)) according to (6)

Calculate Mismatch equationsgP(Vdc(k)) according to (5)

Calculate ∆Vdc(Vdc(k), J(k))according to (6)

Calculate Vdc(k + 1)according to (6);

k = k+1;k < 100 max(|gP|) < ε

STOPFailed to converge;

try a different initial Vdc(k).

STOPLoad flow converged;

Output Vdc(k).

no

yes

no

yes


Load flow in MTdc Networks Example of Load flow in a MTdc Network

• 5 countries: UK, Denmark (DN), Germany (DE), Netherlands (NL),and Belgium (BE).

• 19 dc nodes and 19 dc transmission lines.

UK1

UK2

BE1

NL1

NL2

DE1

DE2

DK1

DK2

HUB1

HUB2

HUB3

NL

UK

BE

DK

DE

HUB4

HUB5

NSTG layout for the load flow study with five EU countries.



The most important system parameters for the load flow analysis [2]

MTdc system parameters in the

load flow example.

Quantity Value

Power base (Sb) 1000 MVAVoltage base(Vb)

±300 kV

HVdc voltage(Vdc)

±300 kV

Cable resistance(Rdc)

0.023 Ω/km

The unit length resistancesof all the transmission linesis according to [3].

MTdc network lines in the load

flow example.

LineStart

LineEnd

Length[km]

Size[pu]

UK1 HUB1 100 3UK2 HUB1 40 2UK HUB1 120 5HUB1 HUB2 300 5BE1 HUB2 50 1BE HUB2 100 1HUB2 HUB3 120 5NL1 HUB3 100 2NL2 HUB3 40 1NL HUB3 70 3HUB3 HUB4 250 5DE1 HUB4 40 2DE2 HUB4 70 2DE HUB4 150 4HUB4 HUB5 120 5DK1 HUB5 40 1DK2 HUB5 50 1DK HUB5 150 2HUB1 HUB5 380 5

Offshore Wind Farms included

in the load flow example.

Wind Farm NodeSize[pu]

Doggers-bank

UK1 3

Hornsea UK2 2Thornton-bank

BE1 1

IJmuiden NL1 2Eemshaven NL2 1HochseeSud

DE1 2

HochseeNord

DE2 2

Horns Rev DN1 1Ringkobing DN2 1



UK1 UK2 BE1 NL1 NL2 DE1 DE2 DN1 DN2 HUB1 HUB2 HUB3 HUB4 HUB5 UK BE NL DE DN0.90

0.92

0.94

0.96

0.98

1

1.02

1.04

1.06

1.08

1.10

1.062

1.042 1.043

1.059

1.044

1.0541.059

1.047 1.048

1.0351.039 1.041

1.0471.043

0.976

1.0291.021

1.000

1.015

Dir

ect

Vo

ltag

e [k

V]

Direct voltages in all terminals

UK1 UK2 BE1 NL1 NL2 DE1 DE2 DN1 DN2 UK BE NL DE DN-4

-3

-2

-1

0

1

2

3

2.25

1.50

0.75

1.50

0.75

1.50 1.50

0.75 0.75

-3.75

-0.75

-2.25 -2.43

-1.50

Act

ive

Po

wer

[p

u]

Active power (onshore/offshore nodes)Results from the dc load flow algorithm for the NSTG studied

• OWFs: 75% of nominal power• Countries control power onshore to match OWF• Exception: Germany (slack node) controls its direct voltage at 1 pu• Highest voltage: UK1 OWF (1.062 pu or ± 318.6 kV)• Lowest voltage: UK (0.976 pu or ± 292.8 kV)• German node provides transmission losses (0.57 pu or 570 MW)• Losses are equal 5.06% (disregarding converter losses).


Load flow in MTdc Networks MTdc Network Security (N-1 Analysis)

• AC contingency onshore: the affectednode cannot exchange active power.

• The N-1 security analysis is performedfor 3 different scenarios

• During dc load flow: slacks control theirvoltages at 1 pu.

Scenarios for the N-1

Security Analysis.

Scenario Slack

(a) 1 Slack UK(b) 3 Slacks UK, NL & DE(c) 5 Slacks All countries

Normal UK BE NL DE DN0

2

4

6

8

10

5.06%

9.96%

5.51%

7.30%

8.28%

6.06%

Tra

nsm

issi

on

Lo

sses

[%

]

Scenario I


2

4

6

8

10

4.72%

8.02%

5.14%

7.31%6.53%

5.78%

Tra

nsm

issi

on

Lo

sses

[%

]

Scenario II


2

4

6

8

10

4.55%

6.89%

5.09%5.80% 5.74%5.59%

Tra

nsm

issi

on

Lo

sses

[%

]

Scenario IIIMTdc network losses in the different scenarios.



• AC contingency onshore: the affectednode cannot exchange active power.

• The N-1 security analysis is performedfor 3 different scenarios

• During dc load flow: slacks control theirvoltages at 1 pu.

Scenarios for the N-1

Security Analysis.

Scenario Slack

(a) 1 Slack UK(b) 3 Slacks UK, NL & DE(c) 5 Slacks All countries


2

4

6

8

10

5.06%

9.96%

5.51%

7.30%

8.28%

6.06%

Tra

nsm

issi

on

Lo

sses

[%

]

Scenario I


2

4

6

8

10

4.72%

8.02%

5.14%

7.31%6.53%

5.78%

Tra

nsm

issi

on

Lo

sses

[%

]

Scenario II


2

4

6

8

10

4.55%

6.89%

5.09%5.80% 5.74%5.59%

Tra

nsm

issi

on

Lo

sses

[%

]

Scenario IIIMTdc network losses in the different scenarios.



N-1 Analysis Results (Voltages)


0.925

0.95

0.975

1

1.025

1.05

1.075

1.1

1.125

1.15

1.175

1.2

1.225

1.25

Dir

ect

Vo

lta

ge

[pu

]

1 slack node: UK





0.925

0.95

0.975

1

1.025

1.05

1.075

1.1

1.125

1.15

1.175

1.2

1.225

1.25

Dir

ect

Vo

lta

ge

[pu

]

UK Fault

BE Fault

NL Fault

GE Fault

DN Fault

Normal Case

3 slack nodes: UK, NL and DE





0.925

0.95

0.975

1

1.025

1.05

1.075

1.1

1.125

1.15

1.175

1.2

1.225

1.25

Dir

ect

Vo

lta

ge

[pu

]

UK Fault

BE Fault

NL Fault

GE Fault

DN Fault

Normal Case

All countries are slack nodes



N-1 Analysis Results (Distribution)


1

1.05

1.1

1.15

1.2

1.25

Dir

ect

Vo

ltag

e [p

u]

1 slack node: UK





1

1.05

1.1

1.15

1.2

1.25

Dir

ect

Vo

ltag

e [p

u]

3 slack nodes: UK, NL and DE





1

1.05

1.1

1.15

1.2

1.25

Dir

ect

Vo

ltag

e [p

u]

All countries are slack nodes


Load flow in MTdc Networks Conclusions

A situation in which more nodes are controlling its direct voltage – thusworking as slack nodes – yielded better results when compared to asituation where only one node was given that task. The superiority wasfound regarding both N-1 contingencies scenarios and the overall losses inthe transmission system.


The Distributed Voltage Control Strategy

Table of Contents

1 Load flow in MTdc NetworksDC Load FlowExample of Load flow in a MTdc NetworkMTdc Network Security (N-1 Analysis)Conclusions

2 The Distributed Voltage Control StrategyControl MethodologyOptimal power flow for MTdc NetworksOptimisation Method: Steepest DescentCase Study: MTdc Losses MinimisationConclusions


The Distributed Voltage Control Strategy Control Methodology

Each voltagecontrolling VSC

receives a specificdirect voltage

set-point [4, 5, 1]

Distributed DC Load Flow

Algortihm

OPF Algorithm

No

Generation at OWFs

Voltage setpoints

to GS-VSCs

Optimum?

MTDC Secure?

Check for N-1 Security

No

Yes

Yes

DVC

SCADAsignal

SCADAsignal

Flowchart diagram for the DVC control strategy.

• Any predefined load flow scenario can be achieved

• No single converter is left alone with the responsibility of balancingthe power inside the transmission system

• The control of the dc system voltage is distributed between severalnodes inside the MTdc network

• As the DVC strategy relies only on a power flow solution, there is noneed for a fast communication link between the network terminals

• Info for dc load flow: a TSO can use a SCADA system as in acnetworks (e.g. 15 min. control cycle)


The Distributed Voltage Control Strategy Control Methodology

Each voltagecontrolling VSC

receives a specificdirect voltage

set-point [4, 5, 1]

Distributed DC Load Flow

Algortihm

OPF Algorithm

No

Generation at OWFs

Voltage setpoints

to GS-VSCs

Optimum?

MTDC Secure?

Check for N-1 Security

No

Yes

Yes

DVC

SCADAsignal

SCADAsignal

Flowchart diagram for the DVC control strategy.

• Any predefined load flow scenario can be achieved

• No single converter is left alone with the responsibility of balancingthe power inside the transmission system

• The control of the dc system voltage is distributed between severalnodes inside the MTdc network

• As the DVC strategy relies only on a power flow solution, there is noneed for a fast communication link between the network terminals

• Info for dc load flow: a TSO can use a SCADA system as in acnetworks (e.g. 15 min. control cycle)


The Distributed Voltage Control Strategy Optimal power flow for MTdc Networks

To solve the OPF problem:

• state variables: x

• specified variables: z =[U W

]T1 control variables, U,2 fixed variables, W)

HVac x HVdc bus types.

Bus Characteristic HVac HVdc

Voltage is unknown PQ bus P busVoltage is known PV bus V bus

P-type buses: HVdc stations controlling power output.V-type buses: HVdc stations controlling the dc system voltage(slack-buses).

• Control variables, U: voltage references of V-type buses• Fixed variables, W: power of the P-type buses

The OPF can then be defined as a minimisation problem:min f (x,U)

s.t. gP(x,U,W) = 0,

where;f (x,U) is the function to be optimised and;gP(x,U,W) are the load flow equations.


The Distributed Voltage Control Strategy Optimisation Method: Steepest Descent

Any optimisation method can be used to solve the OPF

Steepest Descent

1 Assign an initial value to control vector U0;

2 Solve the load flow eq.: ∇Lλ = gP (x,U,W) = 0;

3 With x and J = ∂gP/∂x, from step 2, solve ∇Lx w.r.t. λ :

λ = −[JT]−1∂f

∂x(1)

4 With the new value of λ, compute ∇LU;

5 If |∇LU| ≥ ε go to the next step. Otherwise, the optimisation isachieved;

6 Compute the new control vector Uk+1:

Uk+1 = Uk − β |∇LU| (2)

Restart, from step 2, with Uk+1.


The Distributed Voltage Control Strategy Optimisation Method: Steepest Descent

Any optimisation method can be used to solve the OPF

Steepest Descent

1 Assign an initial value to control vector U0;

2 Solve the load flow eq.: ∇Lλ = gP (x,U,W) = 0;

3 With x and J = ∂gP/∂x, from step 2, solve ∇Lx w.r.t. λ :

λ = −[JT]−1∂f

∂x(1)

4 With the new value of λ, compute ∇LU;

5 If |∇LU| ≥ ε go to the next step. Otherwise, the optimisation isachieved;

6 Compute the new control vector Uk+1:

Uk+1 = Uk − β |∇LU| (2)

Restart, from step 2, with Uk+1.


The Distributed Voltage Control Strategy Case Study: MTdc Losses Minimisation

AC network 1VSC1 VSC2

N4 N5

OWF2OWF3

OWF1

AC network 2

N2

N6

N1

N7

N3

line 1

line 3line 2

line 4 line 6line 5

40 km

200 km 100 km 130 km

50 km 30 km

PWF3 = 0.5 puPWF2 = 0.3 pu

PWF1 = 0.4 pu

Radial 5-terminal VSC-MTdc network analysed in case study 1.



0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 70.96

0.98

1.00

1.02

1.04

1.06

1.08

1.10

1.12

1.14

time [s]

Vd

c [p

u]

VSC 1VSC 2WF 1WF 2WF 3

MTdc system voltages at the onshore and offshore wind farms VSCs



0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7-1.5-1.4-1.3-1.2-1.1-1.0-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.10.00.10.20.30.40.50.6

time [s]

Pd

c [p

u]

Active power at the onshore and offshore wind farms VSCs.


The Distributed Voltage Control Strategy Conclusions

• The DVC strategy is readily expandable to larger MTdc networks(more nodes controlling the network direct voltage will add moreflexibility to the method)

• Nodes not connected to generating plants work as slack nodes(controlling MTdc network voltage)

• The strategy is independent from the optimisation algorithm used tosolve the OPF problem

• OPF problem: minimise systems losses, or maximise the networksocial welfare – or even have multi-objective optimisation goals.

• The strategy is capable of reliably and safely controlling the powerflow in large MTdc networks with good dynamic performance.

• It needs telecommunication. As this communication can happen incontrol cycles in the order of minutes, fast telecommunication linksalthough desirable are, in principle, not needed.

• The distributed voltage control strategy was validated in a low-voltageMTdc network setup in the DC&S laboratory of TU Delft.


References

[1] R. Teixeira Pinto, P. Bauer, S. Rodrigues, E. Wiggelinkhuizen, J. Pierik, andB. Ferreira, “A Novel Distributed Direct-Voltage Control Strategy for GridIntegration of Offshore Wind Energy Systems Through MTDC Network,” IEEETransactions on Industrial Electronics, vol. 60, no. 6, pp. 2429 –2441, june 2013.

[2] R. Teixeira Pinto and P. Bauer, “The Role of Modularity Inside the North SeaTransnational Grid Project: Modular Concepts for the Construction and Operationof Large Offshore Grids,” Proceedings of the Renewable Energy World EuropeConference, Milan, Italy, June 2011, pp. 1–19.

[3] NationalGrid, “Round 3 Offshore Wind Farm Connection Study,” London, TechnicalReport, 2009. [Online]. Available:http://www.thecrownestate.co.uk/media/214799/round3 connection study.pdf

[4] R. Teixeira Pinto, S. F. Rodrigues, E. Wiggelinkhuizen, R. Scherrer, P. Bauer, andJ. Pierik, “Operation and Power Flow Control of Multi-Terminal DC Networks forGrid Integration of Offshore Wind Farms Using Genetic Algorithms,” Energies, vol. 6,no. 1, pp. 1–26, 2012. [Online]. Available: http://www.mdpi.com/1996-1073/6/1/1

[5] S. Rodrigues, R. Pinto, P. Bauer, and J. Pierik, “Optimal Power Flow Control ofVSC-Based Multiterminal DC Network for Offshore Wind Integration in the NorthSea,” IEEE Journal of Emerging and Selected Topics in Power Electronics, vol. 1,no. 4, pp. 260–268, 2013.


http://www.thecrownestate.co.uk/media/214799/round3_connection_study.pdf

http://www.mdpi.com/1996-1073/6/1/1

MTdc Network Modelling andExperimental Validation

Part V






Dynamic Modelling

Table of Contents

1 Dynamic ModellingModular Model of the NSTGVSC-HVdc ModelVSC-HVdc ControlMTdc Network Model

2 Small-Signal AnalysisLinear Model of MTdc NetworksLinear Model of VSC-HVdcControl Tuning & Eigenvalues

3 Experimental ValidationSTATCOMBack-to-backMTdc Network

Dr. R. Teixeira Pinto (TU Delft) MTdc Networks Tutorial: Part V August 25, 2014 2 / 38

Dynamic Modelling Modular Model of the NSTG

Wind Farm

1

Wind speed

PowerHVdc

Station1

Multi-TerminalDirect Current

Network(MTdc)

Power

DC Voltage

HVdcStation

J AC Current

AC GridN-J

HVdcStation

N

AC Voltage

Wind Farm

N

Wind speed

PowerHVdc

StationJ-1

Power

DC Voltage

AC Voltage

Power

DC Voltage

AC Grid1

Power

DC Voltage

AC Current

AC Voltage

AC Voltage

Modular representation of offshore MTdc networks.

• Dynamic models: assessment of system behaviour, interactionsbetween components, and control design.

• MTdc networks will be built modularly, hence, models should also bemodular.

• Modular dynamic models allow highly complex systems to be dividedinto smaller submodules.

• Approach: derive the equations driving the most important systemmodules, viz.: wind farms, HVdc stations, the multi-terminal dc grid,and the interconnected onshore ac systems.


Dynamic Modelling VSC-HVdc Model

Phase Reactor

Inner Current

Controller

Outer Controllers

Converter Model

Control References

ēdq īdq ῡdq*

īdq*

ῡdq

Pdc

MTdcNetwork:

HVdc cables&

VSC-HVdc capacitors

Vdc

Complete VSC-HVdc station

Converter Model

ῡdq* ῡdq*

Vdc

unitdelay ma_min

ma_max

ma_limma

1/2

ῡdq

Averaged converter model

• Switching behaviour: dynamic equations are discontinuous/complex.

• Solution: averaged dynamic models (enough details to understand thesystem dynamics and develop control strategies [1]).

• The closed-loop bandwidth of the VSC current controller is kept atleast 5 times lower than the converter switching frequency.

• The developed VSC-HVdc station model is modular [2].

• In average models: at least one time-step delay (controllercomputational time, blanking time, and algebraic loops) [3]


Dynamic Modelling VSC-HVdc Control

Direct control

VSCE Ln Lc

P,QCf

uf

ic

v

Uf

Uref

QQref

PPref

PLL

RPC/AVC

APC

Voltagereferencecontrol

vrefa

vrefb

vrefc

θv

ωt

∆V

• Straightforward: deals with the converteralternating currents and voltages directly in thethree-phase (abc) frame.

• Current references are calculated via load flowequations between two nodes [4, 5].

• Does not contain an inner-current loop (notpossible to protect the converter valves viacontrol.)

VSCE Ln Lc

P,QCf

ufic

v

vrefa

vrefb

vrefc

vrefα

vrefβ

vrefd

vrefq αβ

abcαβdq

Current-

reference

control

Current

controller

PPref

Q

Qref

Uref

∆Uref

∆V

AVCUf

irefd

irefq

kps

PSL

θv ωt

ωref t

kus

RPC

Power synchronisation control

• Main idea: have the grid-connected VSCdynamic response similar to that of asynchronous machine.

• Allows the VSC-HVdc to maintain steadyoperation for higher load angles. As a result,the VSC is then able to exchange more powerwith the ac network.

• Specially when the VSC is connected to a weak(low short-circuit ratio) ac network [6].



The VSC currents and voltages are transformed to the rotatingdirect-quadrature (dq) frame, which will be then synchronised with the acnetwork voltage through means of a phase-locked loop (PLL).

Inner CurrentController

AC System

Outer Controllers

Cdc

*

cdqi

*

dqv

Measurements

abceabc

i

dcV

TX

References

, ,| |ac ac dq

P Q e

dcV

Structure of VSC vector control strategy.



ICC

It is possible calculate the control bandwidth of a VSC inner currentcontrol, by applying the Laplace transformation the current at the VSCphase reactor:

sLT idq = edq − vdq − RT idq − jωLT idq

de1

T TR sL

TL

qe

1

T TR sL

TL

di

qi

-

+

+

--

( )C s

( )C s

-

-

*

di

*

qi

+

dv

qv

Block diagram representation of the VSC-HVdc

state-space equations with feedback control.

T

T

R

L

0.95 1 1.05 1.1 1.15 1.2 1.25

-3.22

-1.69

-1

0

1

2.73

time [s]

VS

C C

urr

ent

[A]

icd

icq

Root locus (a) and step response (b) of a VSC

without feedback current control.



q

p

e

K

1

1 T

p

Ls

K

T

p

L

K

di

qi

+

+

-

*

di

+

T

p

L

K

+

*

qi +

d

p

e

K

1

1 T

p

Ls

K

Equivalent diagram with ICC.

iq

i∗q=

1

1 +

(ωLT

Kp

)2≈ 1

iq

eq=

1

Kp

1 +

(ωLT

Kp

)2≈

1

Kp

iq

i∗d

=

ωLT

Kp

1 +

(ωLT

Kp

)2≈ωLT

Kp

-30

-20

-10

0

Am

pli

tud

e (d

B)

System: sysFrequency (rad/s): 2e+03Magnitude (dB): -3.01

102

103

104

105

-90

-60

-30

0

An

gle

(d

eg)

frequency [rad/s]

αc = 2000 rad/s

Bode plot

0 1 2 3 40

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

time [ms]

VS

C C

urr

ent

[pu

]

tr ≅ 1.1 ms

step response

• The system has a pole in −Kp/LT , which is equal its closed-loopbandwidth, αc = 2 krad/s.

• For a satisfactory dynamic performance: αc ≤ ωsw

5≤ ωs

10[7, 8]

• 1st order system: tr = ln(9)/αc

• In conclusion: Kp = αcLT is fixed (bandwidth/switching frequency).• Small error, in steady state, in the inner current controller response:

(ωLT/Kp)2 = (100pi · 0.160/320)2 < 2.5 %.



Active and Reactive Power Controllers

The inner current controller is responsible for generating the VSCreference voltages according to:

v∗d = ed − Kp(id − i∗d ) − Ki

∫(id − i∗d )dt + ωLT iq

v∗q = eq − Kp(iq − i∗q ) − Ki

∫(id − i∗d )dt − ωLT id

The outer controllers are the ones responsible for providing the referencessignals, (i∗d , i

∗q ), for the inner current controller.

pac*

pac

iq*Δpac

+ -

PP ip

KK

s

qac*

qac

id*Δqac

+ -

QQ ip

KK

s

Active/Reactive power outer controllers.

i∗d = (Q∗

ac − Qac ) ·(KQp +

KQi

s

)

i∗q = (P∗ac − Pac ) ·

(KPp +

KPi

s

)



Direct Voltage Controller

( )dc

I s

C

( )c

I s

( )L

I sdc

Z

Equivalent circuit of the VSC dc side (lossless).

ac dcP P

Wdc*

Wdc

iq*ΔWdc

+ -

+

wLT/Kp

-

+

1/Kp

id

eq

1

1 T

p

Ls

Keq

2

CPL

-

Pc1

s

Wc +

( )w

C s

Closed-loop diagram of the DVC.

0

0W

pK

1

8W

p CK C

1

8W

p CK C

C

2C

Root-locus diagram.

/ 2C/ 4C C

*

dc

dc open

W

W

-20 db/dec

-40 db/dec

*dc

dc closed

W

W

Theoretical Bode Plot

-40

-30

-20

-10

0

10

20

X: 0.2429Y: 0.0006857 X: 0.3211

Y: -3

frequency [pu]

Am

pli

tud

e |W

dc/W

dc*|

[d

B]

closed loopopen loop

0.01c 0.1c 0.3c 1c

Numerical Bode Plot



LW

p

P

K

0

dcW

dcW

0

LP

LP

LP

P regulator (droop control).

0

LP

LP

LP

0

dcW

/W W

p iK K

dcW

PI regulator (DVC strategy).

For ICC: αpuc =

kp

lT⇒ kp = αpu

c · lTLimiting the direct voltage controller bandwidthto the 0-dB frequency of the open-loop transferfunction:

αpudc ≤

αpuc

4

⇒ kwp ≤

1

8αpuc · cdc =

αpudc · cdc

2

where;αpudc is the bandwidth of the direct voltage

controller [pu].

Stability reasons: the integrator zero <open-loop transfer function 0-dB frequency

KWi

KWp

αc

4

ifkwi

kwp

≤αpuc

8=αpudc

2

⇒ kwi ≤

(αpudc

)2cdc

2

where;kwi is the DVC integral gain [pu].



• Reactive power controlller: difficult to perform a detailed analysis todetermine the bandwidth (good compromise: set it equal to theactive power controller bandwidth)

• The bandwidth of the active power controller, αp should be about thesame or less than the bandwidth of the direct voltage controller

Bandwidth and control gains of the different VSC-HVdc controllers.

ControllerBandwidth

[rad/s]Proportional gain Integral Gain

Inner-Current

αc αpuc lT αcLT αpu

c rT αcRT

DirectVoltage

αdc ≤ αc

4

kwp =

αpudc cdc

2

KWp = kwp · Ib

V 2dcb

kwi =kwp α

pudc

KWi = kwi · Ibωb

V 2dcb

ActivePower

αp ≤ αdc kpp ≤ kwp KPp = kpp · Ib

Sbkpi ≤ kwi KW

i = kwi · Ibωb

SbReactivePower

αq = αp kqp = kpp KQp = kqp · Ib

Sbkqi = kpi KQ

i = kqi · Ibωb

SbAC NetworkVoltage

αv = αq kvp = kqp KVp = kvp · Ib

Vbkvi = kqi KV

i = kvi · Ibωb

Vb


Dynamic Modelling MTdc Network Model

AC network 1

AC network j

jC

1C

_1dcV

_dc jV

...

WT-VSC

Offshore Wind farm 1

Offshore Wind farm i

iC

2C

_dc iVWT-VSC

GS-VSC

GS-VSC

_ 2dcV

cableR cableL

2

cableC

2

cableC

MTDCGrid

Generic representation of a MTdc grid

In the MTdc network model each dc cable is representedby a π-section circuit.

Applying Kirchhoff laws – j nodes andi lines – yields:

ILi =

Vdci

Ldci−

Vdcj

Ldci− ILi

Rdci

Ldci

sVdcj =1

Cdcj

(Idcj −

L∑i=1

IMij · ILi

)

where;ILi is the current flowing through line i[A]; Ldci is inductance of MTdc line i[H]; Cdcj is the sum of all capacitancesat node j [F] and; IMij is MTdcnetwork incidence matrix ij-thposition.

The state-space matrix representation of the MTdc system is given by:x = Ax + Bu

y = Cx + Du



The state variable vector, x is given by:

x =[

Vdc1 ... VdcN IL1 ... ILL]T

1×(N+L)

The input vector, u, is given by:

u =[

Idc1 ... IdcN]T

1×N

The state matrix, A, is composed out of 4matrices, given as:

A =

[A11

N×N A12

N×L

A21

L×N A22

L×L

](N+L)×(N+L)

A11 = 0

A12 = −

IM ·

1

Cdc1

0 · · · 0

01

Cdc2

. . ....

.

.

.. . .

. . . 0

0 · · · 01

CdcN

T

A21 =

1

Ldc1

0 · · · 0

01

Ldc2

. . ....

.

.

.. . .

. . . 0

0 · · · 01

LdcL

· IM

A22 =

−Rdc1

Ldc1

0 · · · 0

0 −Rdc2

Ldc2

. . ....

.

.

.. . .

. . . 0

0 · · · 0 −RdcL

LdcL

The input matrix, B, is constituted of only 2sub-matrices:

B =

[B

11N×N

B21L×N

](N+L)×N

B11 =

1

Cdc1

0 · · · 0

01

Cdc2

. . ....

.

.

.. . .

. . . 0

0 · · · 01

CdcN

B

21 = 0

The output matrix, C; and the feed-forwardmatrix, D; can be selected so as to obtain thedesired output vector, y.

C = I(N+L)×(N×L)

D = 0(N+L)×N



• VSC-HVdc systems have small distances: cables instead of OHL.• extruded polymeric insulation instead of oil-impregnated paper

insulation.• Lighter, more flexible, easier and quicker to install.

Cable 1 Cable 2 Cable 3 Cable 4Evolution of VSC-HVdc cable transmission capacity [9].

1 2000: Directlink – 354 km, ± 80 kV, 60 MW2 2001: Murraylink – 360 km, ± 150 kV, 220 MW3 2004: Estlink – 210 km, ± 150 kV, 350 MW4 2010: DolWin – 165 km, ± 320 kV, 800 MW


Small-Signal Analysis

Table of Contents





Small-Signal Analysis Linear Model of MTdc Networks

The dynamic model of MTdc networks presented is already linear(state-space matrix form)

C1 C2

LR

2-terminal dc network

C C

LR LR

C

LR

C

3-terminal MTdc networkExample of dc networks.

State matrix:

A2 =

0 0 −1/C10 0 1/C2

1/L −1/L −R/L

Eigenvalues:

λ1 = 0

λ2,3 = −R

2L±

√√√√ R2

4L2−

1

L

(1

C1

+1

C2

)

State matrix:

A3 =

0 0 0 0 −1/C 0 00 0 0 0 0 −1/C 00 0 0 0 0 0 −1/C0 0 0 0 1/C 1/C 1/C

1/L 0 0 −1/L −R/L 0 00 1/L 0 −1/L 0 −R/L 00 0 1/L −1/L 0 0 −R/L

Eigenvalues:

λ1 = 0

λ2,3 = −R

2L±

√R2

4L2−

1

LC

λ4,5 = −R

2L±

√R2

4L2−

4

LC


Small-Signal Analysis Linear Model of MTdc Networks

AC network 1VSC1 VSC2

N4 N5

OWF2OWF3

OWF1

AC network 2

N2

N6

N1

N7

N3

line 1

line 3line 2

line 4 line 6line 5

40 km

200 km 100 km 130 km

50 km 30 km

PWF3 = 0.5 puPWF2 = 0.3 pu

PWF1 = 0.4 pu

MTdc network withseven nodes.

-6 -5 -4 -3 -2 -1 0-200

-150

-100

-50

0

50

100

150

200

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

Eigenvalues from the MTdc

network shown.

• Determinant of the state matrix A is always null(eigenvalue located at the origin).

• MTdc network system is marginally stable (alsowhen KW

p of DVC was null).

• Stability of a MTdc network: feedback control ofthe system direct voltages is necessary (althoughnot sufficient).

• The eigenvalues shown, except for the one at theorigin, have all the same real part, equal to−Rdc/2Ldc


Small-Signal Analysis Linear Model of VSC-HVdc

The VSC model needs to belinearized.

In the analytical approach, a variable,x , is expanded in a small-signal part,x , and a steady-state part, X0. The

same applies to a non-linear function,f :

f (x1, x2) = f (x1, x2) + F (X01,X02)

eq

AC System

acQ

*acQ

dcV

*dcV

(*) reference=

Reactive Power Controller

DC Voltage Controller

Outer Controllers

C

*cdi

*cqi

*cdqv

abce

abc to dq

θ abc to dq

gabci

dq to abc

*abcv

dcVGL

PWM

CL

fC

abcu

gid

ged

giq

g-

PLL

Inner CurrentController

single-phase equivalent circuit

Grid Impedanceg

dqe Inner Current

Controller

Outer Controllers

*, acV

gdqi

References:

LP

*c

cdqi

Park Transformation

gdqu

PLLθ

Phase Reactor

ccdqi *c

dqv

cdqv

acPInverse Park

Transformation

gcdqi dcV

Qdc*

cdqu dc-side

Capacitor

Filter Capacitance

Converter Model

VSC small-signal

model.


Small-Signal Analysis Control Tuning & Eigenvalues

The VSC small-signal model, in the (dq) frame, has 13 state variables:

• the first 7: VSC hardware, i.e. they are the voltages and currents ofthe energy storing elements

• the later 6: VSC controllers.

The complete set of state variables is given by:

x =

igd igq︸︷︷︸grid−sidecurrent

ugd ugq︸︷︷︸LCLfiltervoltage

iccd iccq︸︷︷︸VSC−side

current

Wdc︸︷︷︸dc−sidevoltagesquare

εid εiq︸︷︷︸ICCerror

εdc︸︷︷︸DVCerror

εQ︸︷︷︸RPCerror

εpll︸︷︷︸PLLerror

θ︸︷︷︸PLLangle

The complete set of input variables is given by:

u =

egd egq︸︷︷︸grid−sidevoltage

PL︸︷︷︸dc−linepower

W ∗dc︸︷︷︸

dc−sidevoltage

reference

q∗g︸︷︷︸grid−side

reactive powerreference



To control the VSC using only PI controllers, 8 gains need to be set.It is possible to tune the controllers, by optimizing the location of thesmall-signal model Eigenvalues.

-100

-90

-80

-70

-60

-50

-40

-30

-3000-2500

-2000-1500

-1000-500

0

-2500

-2000

-1500

-1000

-500

0

largest complex part [rad/s]

slowest pole [rad/s]

curr

ent

con

tro

ller

ban

dw

idth

[ra

d/s

]

Each blue dot is a set

of the 8 VSC

control gains with a specific

eigenvalue signature.

Pareto front obtained from the MOGA optimal VSC control tuning.



VSC cabinet, real-time controller, and host PC.Laboratory setup for validating the VSC models

Rated parameters of the VSC used in the

laboratory experiments.Parameter Symbol Value Unit

VSC Rated Power Svsc 5000 VAVSC ac-side Voltage Vc 380 VGrid-side Inductance Lg 1.5 mHGrid-side Resistance Rg 0.2 ΩFilter Capacitance Cf 20 µFConverter Inductance Lc 1.5 mHConverter Resistance Rc 0.2 ΩDC-Side Voltage Vdc 700 VDC-Side Capacitor Cdc 500 µF

VSC

eigenvalues

-3000 -2500 -2000 -1500 -1000 -500 0-6000

-4000

-2000

0

2000

4000

6000

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

Optimised VSC Control gains

ControllerProportionalGain

Unit Integral Gain Unit

ICC Kp = 7.7080 [V /A] Ki = 6.3678 [kV /As]DVC KW

p = 0.3723 [mA/V 2] KWi = 48.6845 [mA/V 2s]

RPC KQp = 0 [1/V ] KQ

i = 1.3371 [1/Vs]

PLL Kpllp = 6.3919 [rad/Vs] Kpll

i = 3.3752 [krad/Vs2]



-138.05 -138.045 -138.04 -138.035 -138.03 -138.025 -138.02 -138.015-6000

-4000

-2000

0

2000

4000

6000

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

igd

igq

Grid current

-947.82 -947.8 -947.78 -947.76 -947.74 -947.72 -947.7 -947.68 -947.66 -947.64-5000

0

5000

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

ugd

ugq

Filter voltage

-3100 -3050 -3000 -2950 -2900 -2850-30

-20

-10

0

10

20

30

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

icd

icq

Converter current

-2600 -2400 -2200 -2000 -1800 -1600-800

-600

-400

-200

0

200

400

600

800

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

eid

eiq

ICC

-1200 -1000 -800 -600 -400 -200 0-1

-0.5

0

0.5

1

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

Wdc

edc

eQ

Wdc and outercontrollers

-500 -450 -400 -350 -300 -250 -200 -150-100

-50

0

50

100

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

epll

s

PLL

400x 10

-60

500 600

DC-side capacitor - Cdc

VSC eigenvalues with different dc output capacitor.



-300 -250 -200 -150 -100-1.5

-1

-0.5

0

0.5

1

1.5x 10

4

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

igd

igq

Grid current

-1300 -1200 -1100 -1000 -900 -800 -700 -600-1

-0.5

0

0.5

1x 10

4

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

ugd

ugq

Filter voltage

-3500 -3400 -3300 -3200 -3100 -3000 -2900 -2800 -2700 -2600-600

-400

-200

0

200

400

600

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

icd

icq

Converter current

-2500 -2400 -2300 -2200 -2100 -2000 -1900 -1800-300

-200

-100

0

100

200

300

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

eid

eiq

ICC

-1300 -1200 -1100 -1000 -900 -800 -700 -600 -500 -400-1

-0.5

0

0.5

1

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

Wdc

edc

eQ

Wdc and outercontrollers

-320 -300 -280 -260 -240 -220 -200-1

-0.5

0

0.5

1

Real axis [rad/s]

Imag

inar

y a

xis

[ra

d/s

]

epll

s

PLL

0.5 1 1.5 2 2.5x 10

-30

AC-side inductance - Lg

VSC eigenvalues with different ac grid inductance.


Experimental Validation

Table of Contents





Experimental Validation STATCOM

The dynamic models and controllers were tested in a 5 kVA VSC.

VSC as a static synchronous compensator (STATCOM)

+VdcGL

Lab

Grid

-Vdc

VSC

9.5 10 10.5 11 11.5 12 12.5-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

time [s]

Rea

ctiv

e P

ow

er [

pu

]

QACVSC

QACLSM

QACSSM

VSC reactive power.

VSC direct voltage.

9.8 9.9 10 10.1 10.2-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

time [s]

Rea

ctiv

e P

ow

er [

pu

]

VSC reactive power.

VSC direct voltage.


Experimental Validation STATCOM

The dynamic models and controllers were tested in a 5 kVA VSC.

VSC as a static synchronous compensator (STATCOM)

+VdcGL

Lab

Grid

-Vdc

VSC

14.5 15 15.5 16 16.5 17 17.50.80

0.85

0.90

0.95

1.00

1.05

1.10

1.15

1.20

time [s]

Dir

ect

Vo

ltag

e [p

u]

VDCSSM

VDCLSM

VDCVSC

VSC reactive power.

VSC direct voltage.

9.8 9.9 10 10.1 10.2-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

time [s]

Rea

ctiv

e P

ow

er [

pu

]

VSC reactive power.

15.8 15.9 16 16.1 16.20.85

0.90

0.95

1.00

1.05

1.10

1.15

1.20

time [s]

Dir

ect

Vo

ltag

e [p

u]

VSC direct voltage.


Experimental Validation Back-to-back

Two converters, VSC1 and VSC2, were connected via 4 mm2 LVdc cables(10 m long)

VSC1VSC2

AC Grid1AC Grid2

Experimental setup used to test the VSC operation in a back-to-back topology.



9 10 11 12 13 14 15 16 17 18 19 20 21 22 230.85

0.9

0.95

1

1.05

1.1

1.15

time [s]

Vo

ltag

e [p

u]

Direct voltage at VSC1

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23-1.5

-1.25

-1

-0.75

-0.5

-0.25

0

0.25

0.5

0.75

1

1.25

1.5

time [s]A

ctiv

e p

ow

er [

pu

]

Active power at VSC1

Results from the back-to-back configuration experiments.



16.9 17 17.1 17.2-0.25

0

0.25

0.5

0.75

1

1.25

time [s]

Act

ive

po

wer

[p

u]

Active power at ac grid 1

21.9 22 22.1 22.2-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.10.1

time [s]

Act

ive

Po

wer

[p

u]

Active power at ac grid 2

21.9 22 22.1 22.20.99

0.994

0.998

1.002

1.006

1.01

1.014

1.018

1.02

time [s]

Vo

ltag

e [p

u]

Direct voltage at VSC2

Zoom-in from the results during the back-to-back experiments.


Experimental Validation MTdc Network

The final setup:

1 Opal-RT real-time digital simulator2 the Triphase real-time controller3 the 3 converters – VSC1 to VSC34 The resistance boxes used to form the MTdc grid.

VSC1 VSC2 VSC3

Opal-RT: “offshore wind Farm”

3-terminal dc grid

Real-time controller



The final setup:

1 Opal-RT real-time digital simulator

2 the Triphase real-time controller

3 the 3 converters – VSC1 to VSC3

4 The resistance boxes used to form the MTdc grid.

+Vdc

VSC2

GL

Lab Grid

GL

Lab Grid

GL

Lab Grid

-Vdc

VSC1

Host PC

Real-Time Controller

VSC3



0 20 40 60 80 100 120 140 160 180 200 220 2400.3

0.35

0.4

0.45

0.5

0.55

time [s]

Po

wer

[p

u]

Wind power curve.Wind power signal from

OPAL-RT.

• Power curve: DutchOWF Egmond aanZee (OWEZ).

• 20 km offshore

• 27 km2 area

• 108 MW (36Vesta’s 3MW V90)

OPAL-RT Host PC

Real-Time Controller

Host PCOPAL-RT Real-time

Digital Simulator

EthernetSwitch

Analog Input

(EL3102)

EtherCat Coupler(EK1100)

Analog signal

Ethercatsignal

Flowchart of the offshore wind farm signal.



The MTdc network control isorganized in four layers:

1 DVC strategy & communication

2 The outer controllers

3 The current controller (ICC)

4 The modulator: PWM block

Target PCHost PC (TUD)

VSC1 VSC2

Communication structure

Voltage-Source Converter block



Control structure of the voltage-source converters.



1

[after(2,sec)]

1

1

1

2

1

1

STOP3

entry:

printf(“\n ERROR: \n”)

Printf(“VSC3 failed to startup \n”)

stop = 1;

STOP2

entry:



stop = 1;

STOP1

entry:



stop = 1;

VSC1_ON

entry:

switch[0] = 1;

switch[1] = 0;

switch[2] = 0;

VSC2_ON

entry:

switch[0] = 1;

switch[1] = 1;

switch[2] = 0;

VSC3_ON

entry:

switch[0] = 1;

switch[1] = 1;

switch[2] = 1;

[after(5,sec) & main_switch == 1] 1

[after(5,sec) & error[0] == 0 & Vdc[0] >= 630]

[error[0] == 1] 2

[error[1] == 1] 2


[error[2] == 1] 2

1

Delay


[after(0.01,sec) & main_switch == 1]

[g_error == 0 & main_switch == 1]

1

Optimization_Offline

entry:

[V_ini,Ybus,Pg,Pc,sn] = …

ml.System_Data(~)

Pg[0] = P_ref[0];

Pg[1] = P_ref[1];

V = ml.opt(V_ini,Ybus,Pg,Pc,sn);

Vdc_ref[0] = V[0];

Vdc_ref[1] = V[1];

Vdc_ref[2] = V[2];

g_error = error[0]+error[1]+error[2];

1

[g_error > 0 | main_switch == 0]

2

[g_error > 0]

[after(2,sec)]

VSC2_OFF

entry:

switch[1] = 0;

[after(2,sec)] STOP4

entry:

printf(“\n Succesful Shutoff. \n”)

stop = 1;

[after(2,sec)]

VSC1_OFF

entry:

switch[0] = 0;

Initialization

entry:

Vdc_ref[0] = 1; switch[0] = 1;



Initialize_Shutoff

entry:

Vdc_ref[0] = 1;

Vdc_ref[1] = 1;

Vdc_ref[2] = 1;

VSC3_OFF

entry:

switch[2] = 0;

Complete flowchart of the DVC strategy StateFlow block.



Results (1)

1 VSC1: slack node (controlling the MTdc network voltage);

2 VSC2: slack node (active power reference is established by the TSO)

3 VSC3: OWF (controls its active power reference)

Active power reference of VSC2 during the DVC validation experiments.

Time [s] VSC2∗ [pu] Time [s] VSC2∗ [pu]

0 0.00 140 -0.3025 -0.50 160 -0.6040 -0.25 180 -0.2060 0.00 200 0.2080 0.25 220 -0.40

100 0.10 240 0.00120 0.00 260 0.00



Results (2)

During the experiments, the DVC algorithm outputs the voltage referencesof VSC1 and VSC2 to minimise the losses in the MTdc network, andassure that the active power flow scheduling for VSC2 is fulfilled.

0.95

0.96

0.97

0.98

0.99

1

1.01

1.02

1.03

1.04

1.05

1.06

1.07

Dir

ect

Vo

ltag

e [p

u]

vdc1vdc2vdc3

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280time [s]

Direct Voltage

-0.8

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Act

ive

Po

wer

[p

u]

pac1pac2pac3

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280time [s]

Active Power



Results (3)

0 20 40 60 80 100120 140160 180 200220 240 260-0.6

-0.5

-0.4

-0.3

-0.2

0

0.10

0.2

0.25

0.30

Time [s]

Po

wer

[p

u]

pvsc2

pref

Reference x Measurements

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

Po

wer

dif

fere

nce

[p

u]

0 50 100 150 200 250Time [s]

Error

0 20 40 60 80 100120 140160 180 200220 240 260

0.02

0.04

0.06

0.08

0.1

0.12

0.14

To

tal

loss

es [

pu

]

Time [s]

MTdc Losses

Comparison between VSC2 averaged active power and its active power reference.



Conclusions

• Nowadays it is possible to successfully operate and control MTdcnetworks.

• The complete system can have very good dynamic response, withtransients lasting less than 100 ms.

• The DVC strategy can control the power flow inside MTdc networkswith more than 90% precision.


References

[1] P. Bauer, “Dynamic Analysis of Three-Phase AC Converters,” PhD. Thesis, DelftUniversity of Technology, Delft, January 1995, ISBN: 90-9007789. [Online].Available: http://repository.tudelft.nl

[2] R. Teixeira Pinto and P. Bauer, “Modular Dynamic Models of Large OffshoreMulti-Terminal DC (MTDC) Networks,” Proceedings of the European Wind EnergyAssociation Conference. Brussels: EWEA, March 2011, pp. 1–10. [Online].Available: http://proceedings.ewea.org/annual2011/proceedings/index2test.php?page=searchresult&day=10

[3] L. Harnefors, M. Bongiorno, and S. Lundberg, “Input-Admittance Calculation andShaping for Controlled Voltage-Source Converters,” IEEE Transactions On IndustrialElectronics, vol. 54, no. 6, pp. 3323–3334, December 2007.

[4] B. T. Ooi and X. Wang, “Voltage angle lock loop control of the boost type pwmconverter for hvdc application,” IEEE Transactions on Power Electronics, vol. 5,no. 2, pp. 229–235, 1990.

[5] R. Teixeira Pinto and P. La Seta, Dynamics and Control of VSC-based HVDCSystems: A Practical Approach to Modeling and Simulation. Lambert AcademicPublishing, 2012. [Online]. Available: http://www.amazon.de/Dynamics-Control-VSC-based-HVDC-Systems/dp/3845439742/


http://repository.tudelft.nl

http://proceedings.ewea.org/annual2011/proceedings/index2test.php?page=searchresult&day=10

http://proceedings.ewea.org/annual2011/proceedings/index2test.php?page=searchresult&day=10

http://www.amazon.de/Dynamics-Control-VSC-based-HVDC-Systems/dp/3845439742/

http://www.amazon.de/Dynamics-Control-VSC-based-HVDC-Systems/dp/3845439742/

References

[6] L. Zhang, “Modeling and control of VSC-HVDC links connected to weak acsystems,” PhD Thesis, Royal Institute of Technology, Electrical Engineering Dept.,2010. [Online]. Available: http://www.ee.kth.se/php/modules/publications/reports/2010/TRITA-EE 2010 022.pdf

[7] L. Harnefors and H.-P. Nee, “Model-Based Current Control of AC Machines Usingthe Internal Model Control Method,” IEEE Transactions on Industry Applications,vol. 34, no. 1, pp. 133–141, 1998.

[8] M. Kazmierkowski, R. Krishnan, and F. Blaabjerg, Control in Power Electronics.Academic Press, 2002, iSBN: 0-12-402772-5.

[9] P. Lundberg, M. Callavik, M. Bahrman, and P. Sandeberg, “Platforms for Change:High-Voltage DC Converters and Cable Technologies for Offshore RenewableIntegration and DC Grid Expansions,” IEEE Power and Energy Magazine, vol. 10,no. 6, pp. 30–38, 2012.




EPE '14 ECCE Europe Tutorial #2 - Challenges on the Road to Future High-Voltage Multi-Terminal DC Networks

Engineering

review of mtdc networks

delft university of

modern vschvdc technologydr

north sea

connection ofoshore

promoteenergy trade

classic hvdcdr

terminal network testingdr