High Voltage DirectCurrent Transmission –Proven Technologyfor Power Exchange

2

Contents

Chapter Theme Page

Contents 3

1 Why High Voltage Direct Current? 4

2 Main Types of HVDC Schemes 6

3 Converter Theory 8

4 Principle Arrangement of an 11HVDC Transmission Project

5 Main Components 145.1 Thyristor Valves 155.2 Converter Transformer 185.3 Smoothing Reactor 21

5.4 Harmonic Filters 225.4.1 AC Harmonic Filter 235.4.2 DC Harmonic Filter 255.4.3 Active Harmonic Filter 26

5.5 Surge Arrester 28

5.6 DC Transmission Circuit5.6.1 DC Transmission Line 315.6.2 DC Cable 335.6.3 High Speed DC Switches 345.6.4 Earth Electrode 36

5.7 Control & Protection 38

6 System Studies, Digital Models, 45Design Specifications

7 Project Management 46

3

4

Why High Voltage Direct Current ?1

1.1 Highlights from the HighVoltage Direct Current (HVDC) History

The transmission and distribution ofelectrical energy started with directcurrent. In 1882, a 50-km-long 2-kV DCtransmission line was built betweenMiesbach and Munich in Germany.At that time, conversion betweenreasonable consumer voltages andhigher DC transmission voltages couldonly be realized by means of rotatingDC machines.

In an AC system, voltage conversion issimple. An AC transformer allows highpower levels and high insulation levelswithin one unit, and has low losses. It isa relatively simple device, which requireslittle maintenance. Further, a three-phasesynchronous generator is superior to aDC generator in every respect. For thesereasons, AC technology was introducedat a very early stage in the developmentof electrical power systems. It was soonaccepted as the only feasible technologyfor generation, transmission and distri-bution of electrical energy.

However, high-voltage AC transmissionlinks have disadvantages, which maycompel a change to DC technology:

• Inductive and capacitive elements ofoverhead lines and cables put limits to the transmission capacity and the transmission distance of AC trans-mission links.

• This limitation is of particular signi-ficance for cables. Depending on therequired transmission capacity, the system frequency and the loss eva-luation, the achievable transmission distance for an AC cable will be in therange of 40 to 100 km. It will mainly be limited by the charging current.

• Direct connection between two AC systems with different frequencies isnot possible.

• Direct connection between two AC systems with the same frequency ora new connection within a meshed grid may be impossible because of system instability, too high short-circuitlevels or undesirable power flow scenarios.

Engineers were therefore engaged overgenerations in the development of atechnology for DC transmissions as asupplement to the AC transmissions.

Line-Commutated Current SourcedConverters

The invention of mercury arc rectifiers inthe nineteen-thirties made the design ofline-commutated current sourcedconverters possible.

In 1941, the first contract for a commer-cial HVDC system was signed inGermany: 60 MW were to be suppliedto the city of Berlin via an undergroundcable of 115 km length. The systemwith ±200 kV and 150 A was ready forenergizing in 1945. It was never putinto operation.

Since then, several large HVDC systemshave been realized with mercury arcvalves.

The replacement of mercury arc valvesby thyristor valves was the next majordevelopment. The first thyristor valveswere put into operation in the latenineteen-seventies.

The outdoor valves for Cahora Bassawere designed with oil-immersedthyristors with parallel/series connectionof thyristors and an electromagnetic firingsystem.

Further development went via air-insulated air-cooled valves to the air-insulated water-cooled design, which isstill state of the art in HVDC valve design.

The development of thyristors with highercurrent and voltage ratings has eliminatedthe need for parallel connection andreduced the number of series-connectedthyristors per valve. The development oflight-triggered thyristors has furtherreduced the overall number ofcomponents and thus contributed toincreased reliability.

Innovations in almost every other areaof HVDC have been constantly addingto the reliability of this technology witheconomic benefits for users throughoutthe world.

Self-Commutated Voltage SourcedConverters

Voltage sourced converters requiresemiconductor devices with turn-offcapability. The development of InsulatedGate Bipolar Transistors (IGBT) with highvoltage ratings have accelerated thedevelopment of voltage sourcedconverters for HVDC applications in thelower power range.

The main characteristics of the voltagesourced converters are a compact design,four-quadrant operation capability andhigh losses.

Siemens is offering voltage sourcedconverters for HVDC applications withratings up to 250 MW under the tradename HVDC plus Power Link UniversalSystems.

This paper focuses upon HVDC trans-mission systems with high ratings, i.e.with line-commutated current sourcedconverters.

HVDC = high voltage direct currentDC = direct currentAC = alternating currentIGBT = insulated gate bipolar

transistor

1.2 Technical Merits of HVDC

The advantages of a DC link over an AClink are:

• A DC link allows power transmissionbetween AC networks with differentfrequencies or networks, which can not be synchronized, for other reasons.

• Inductive and capacitive parameters do not limit the transmission capacityor the maximum length of a DC overhead line or cable. The conductorcross section is fully utilized becausethere is no skin effect.

For a long cable connection, e.g. beyond40 km, HVDC will in most cases offerthe only technical solution because ofthe high charging current of an AC cable.This is of particular interest for trans-mission across open sea or into largecities where a DC cable may provide theonly possible solution.

• A digital control system provides accurate and fast control of the activepower flow.

• Fast modulation of DC transmission power can be used to damp power oscillations in an AC grid and thus improve the system stability.

1.3 Economic Considerations

For a given transmission task, feasibilitystudies are carried out before the finaldecision on implementation of an HVACor HVDC system can be taken. Fig.1-1shows a typical cost comparison curvebetween AC and DC transmissionconsidering:

• AC vs. DC station terminal costs• AC vs. DC line costs• AC vs. DC capitalised value of losses

The DC curve is not as steep as the ACcurve because of considerably lower linecosts per kilometre. For long AC linesthe cost of intermediate reactive powercompensation has to be taken intoaccount.

The break-even distance is in the rangeof 500 to 800 km depending on a numberof other factors, like country-specific costelements, interest rates for projectfinancing, loss evaluation, cost of rightof way etc.

1.4 Environmental Issues

An HVDC transmission system is basi-cally environment-friendly becauseimproved energy transmission possi-bilities contribute to a more efficientutilization of existing power plants.

The land coverage and the associatedright-of-way cost for an HVDC overheadtransmission line is not as high as thatof an AC line. This reduces the visualimpact and saves land compensation fornew projects. It is also possible to in-crease the power transmission capacityfor existing rights of way. A comparisonbetween a DC and an AC overhead lineis shown in Fig. 1-2.

There are, however, some environmentalissues which must be considered for theconverter stations. The most importantones are:

• Audible noise• Visual impact• Electromagnetic compatibility• Use of ground or sea return path

in monopolar operation

In general, it can be said that an HVDCsystem is highly compatible with anyenvironment and can be integrated intoit without the need to compromise onany environmentally important issues oftoday.

Fig. 1-2: Typical transmission line structuresfor approx. 1000 MW

5

Fig. 1-1:Total cost/distance

Costs Total AC Cost

TransmissionDistance

Break-EvenDistance

ACTerminals

AC LineDCTerminals

AC Losses

TotalDC Cost

DC Losses

DC Line

AC-tower

DC-tower

HVDC

Fig. 2-2: Back-to-back converter

AC

Sys

tem

1

AC

Sys

tem

2

Fig. 2-5: Monopole with metallic return path

HVDCCable/OHL

LVDC

AC

Sys

tem

1

AC

Sys

tem

2

2.1 DC Circuit

The main types of HVDC converters aredistinguished by their DC circuit arrange-ments. The following equivalent circuitis a simplified representation of theDC circuit of an HVDC pole.

The current, and thus the power flow, iscontrolled by means of the differencebetween the controlled voltages. Thecurrent direction is fixed and the powerdirection is controlled by means of thevoltage polarity. The converter is de-scribed in the next section.

2.3 Monopolar Long-DistanceTransmissions

For very long distances and in particularfor very long sea cable transmissions, areturn path with ground/sea electrodeswill be the most feasible solution.

Fig. 2-1: Equivalent DC circuit

Ud1 Ud2

Id

In many cases, existing infrastructure orenvironmental constraints prevent theuse of electrodes. In such cases, ametallic return path is used in spite ofincreased cost and losses.

± ±

Main Types of HVDC Schemes2

Fig. 2-4: Monopole with ground return path

HVDCCable/OHL

Electrodes

AC

Sys

tem

1

AC

Sys

tem

2

2.2 Back-to-Back Converters

The expression Back-to-back indicatesthat the rectifier and inverter are locatedin the same station.

Back-to-back converters are mainly usedfor power transmission between adjacentAC grids which can not be synchronized.They can also be used within a meshedgrid in order to achieve a defined powerflow.

Fig. 2-3: Back-to-back converterStation Vienna Southeast

6

HVDC = high voltage direct currentDC = direct currentAC = alternating currentUd = DC voltage 12-pulseId = DC currentOHL = overhead lineLVDC = low voltage direct current

2.4 Bipolar Long-DistanceTransmissions

A bipole is a combination of two polesin such a way that a common low voltagereturn path, if available, will only carry asmall unbalance current during normaloperation.

This configuration is used if the requiredtransmission capacity exceeds that of asingle pole. It is also used if requirementto higher energy availability or lower loadrejection power makes it necessary tosplit the capacity on two poles.

During maintenance or outages of onepole, it is still possible to transmit partof the power. More than 50% of thetransmission capacity can be utilized,limited by the actual overload capacityof the remaining pole.

The advantages of a bipolar solution overa solution with two monopoles arereduced cost due to one common or noreturn path and lower losses. The maindisadvantage is that unavailability of thereturn path with adjacent componentswill affect both poles.

2.4.1 Bipole with Ground ReturnPath

This is a commonly used configurationfor a bipolar transmission system. Thesolution provides a high degree offlexibility with respect to operation withreduced capacity during contingenciesor maintenance.

Electrodes

AC

Sys

tem

1

AC

Sys

tem

2

HVDCCable/OHL

HVDCCable/OHL

Fig. 2-6: in bipolar balanced operation (normal)

Upon a single-pole fault, the current ofthe sound pole will be taken over by theground return path and the faulty polewill be isolated.

AC

Sys

tem

1

AC

Sys

tem

2

Fig. 2-7: in monopolar ground returnoperation (converter pole or OHL outage)

Electrodes

HVDCCable/OHL

HVDCCable/OHL

Following a pole outage caused by theconverter, the current can be commutatedfrom the ground return path into ametallic return path provided by theHVDC conductor of the faulty pole.

Electrodes

HVDCCable/OHL

HVDCCable/OHL

2.4.2 Bipole with Dedicated MetallicReturn Path for Monopolar Operation

If there are restrictions even to temporaryuse of electrodes, or if the transmissiondistance is relatively short, a dedicatedLVDC metallic return conductor can beconsidered as an alternative to a groundreturn path with electrodes.

2.4.3 Bipole without DedicatedReturn Path for Monopolar Operation

A scheme without electrodes or adedicated metallic return path formonopolar operation will give the lowestinitial cost.

Monopolar operation is possible bymeans of bypass switches during aconverter pole outage, but not during anHVDC conductor outage.

A short bipolar outage will follow aconverter pole outage before the bypassoperation can be established.

7

AC

Sys

tem

1

AC

Sys

tem

2

Fig. 2-8: in monopolar metallic returnoperation (converter pole outage)

Electrodes

HVDCCable/OHL

HVDCCable/OHL

AC

Sys

tem

1

AC

Sys

tem

2

Fig. 2-10: in bipolar balanced operation (normal)

HVDCCable/OHL

HVDCCable/OHL

AC

Sys

tem

1

AC

Sys

tem

2

Fig. 2-9: in bipolar balanced operation (normal)

HVDCCable/OHL

HVDCCable/OHL

LVDCCable/OHL

Fig. 3-1: Six-pulse converter bridge Fig. 3-2: DC voltage of bridge converteras a function of α

Fig. 3-3: Arrangement of the valve branchesin a 12-pulse bridge

1 Valve Branch2 Double Valve3 Valve Tower4 6-pulse Bridge

213

4

Converter Theory3

180°

150°

90 °

60°

α = 0°

α = 60°

α = 90°γ = 90°

α = 150°γ = 30°

α = 180°γ = 0°

1

5

3

i

i

i

2i

4i

6i

dI

1 53

L1i

L2i

i

1 3 5 1

6 2 4 6 2

0°

60°

120° Ud

264

DC current in each valve and phase

3.1 Bridge Circuit Function

Current flows through the valves whenthe voltage between the anode andcathode is positive. For the valve tocommutate the current, there must bea positive potential (voltage), and thethyristor must have firing pulses. In thereverse direction, i.e. when the potentialbetween anode and cathode is negative,a firing pulse has no effect. The flow ofcurrent in a valve ends when the voltagebetween anode and cathode becomesnegative. The instant when current beginsto flow through a valve, or to commutatefrom one valve to another, can be delayedby postponing the firing. This methodpermits the average value of the outgoingvoltage of the rectifier to be changed.The firing pulses are generated by syn-chronizing the network using an elec-tronic control device. These pulses can bedisplaced from their ”natural firing“ point,which is the point where the two phasevoltages intersect. The method of firing-pulse displacement is called phasecontrol.

The angle between the time at whichthe valve voltage becomes positive andthe firing time (start of commutation) isreferred to as the firing delay. Fig. 3-2shows that for a firing delay of 90°, theaverage voltage equals zero. i.e. thepositive and negative areas of the curve– voltage against time – cancel eachother out. No active power flows throughthe converter.

When the firing delay is greater than90°, the negative voltage/time areasdominate, and the polarity of the averagedirect voltage changes. Due to physicalreasons, the direction of the current doesnot change. (The thyristor valves conductcurrent only in one direction.) When thedirection of energy flow is reversed, thedelivery changes to the supply side. Therectifier becomes an inverter whichdelivers energy to the AC network.

The average value of the direct voltageas a function of the firing delay is givenby:Udiα = 1.35 * UL * cos α UL = secondary side line voltageα = firing angleγ = extinction angle

3.2 12-Pulse Group and ConverterTransformer

HVDC converters are usually built as 12- pulse circuits. This is a serial connectionof two fully controlled 6-pulse converterbridges and requires two 3-phase sys-tems which are spaced apart from eachother by 30 electrical degrees. The phasedifference effected to cancel out the6-pulse harmonics on the AC and DCside.

0°

ω t

ω t

ω t

ω t

ω t

0°

Ud

Udi

Id

8

L1

L1

L2

L3

L2 L3

L3

HVDC DC Circuit

UdN = PdN Rec/IdN

UdN => nominal DC voltage 12-pulse

IdN => nominal DC current

PdNRec => nominal DC active power at the rectifier

3.3 Reactive Power as a Functionof Load

The curve of reactive power demand ofan HVDC station with changing activepower P can be calculated from equation:

Q = P * tan [ arc cos ( cos α - dx)]

In Fig. 3-5, the reactive power demandof a converter is presented under threedifferent control methods.

If the terminal DC voltage Ud and thefiring angle α (or the extinction angle γof an inverter) are held constant, curve(1) will be obtained. If, however, Uv isheld constant (Udi = const regulation), alinear curve such as (2) is obtained. Thepower of a converter can also be changedwhen the (nominal) current is held con-stant by varying the DC voltage. Curve (3)shows the reactive power demand forthis control method. It is important tonote that the entire area between curves(1) and (3) is available for reactive powercontrol. Each point within this area canbe set by the selection of firing anglesα and ß (or γ).

Fig. 3-4:Current displacement with angle control

Fig. 3-5:Reactive power demand of an HVDC converter

α

ϕ

α

α

Secondary Voltageof the Transformer

Basic AC Current

= 60°

Udi

1.00.5

123

Ud = const; α = const ( = const)Ud = const; Uv = constId = const

1

2

3

P/PN

Q/PN

0.2

0.4

0.6

0.8

1.0

1.2

9

ω t

dx = relative inductive voltagedrop

Uv = valve voltageUd = DC voltage 12-pulseα = firing angleß = 180°-αγ = extinction angle

10

Reactive-Power Balance

UAC in p.u. (AC bus voltage)

– cap. reactive-power reactive-power reactive-power reactive-power+ ind converter AC filters reactors capacitors

QNetwork = + QConv – QFK*UAC2 + QL*UAC

2 – QC*UAC2

420 kV 50 Hz

Q = 103 Mvar

Q = 103 Mvar

Q = 103 Mvar

420 kV 50 Hz

Q = 103 Mvar

Q = 103 Mvar

Q = 103 Mvar

100 200 300 400 500 600 700

0.2 0.4 0.6 0.8 1.0 1.2

60

40

20

100

80

- 60

- 80

- 20

- 40

- 100

P (MW)

PPN

____

Q (Mvar)

Capacitor bank

High pass f ilter 2

High pass f ilter 1

Electronicreactive-powerregulation

Fig. 3-6: Reactive-power compensation andcontrol of an HVDC back-to-back link

10

Converter Theory3

3.4 Reactive Power Control

The possibility of electronic reactivepower control as demonstrated in thepreceding section is used only to a verylimited degree in HVDC technology. Thisis due to economic reasons. Both controlreactive power and commutation reactivepower are increased by the reduction ofthe DC voltage and the correspondingincrease of current. However, load lossesincrease with the square of the current.For this reason, application is limited tothe light loads where the necessary filtercircuits produce a considerable overcom-pensation for the reactive power requiredby the converter.

Fig. 3-6 depicts the reactive power controlof the Dürnrohr HVDC link. In this system,a compensation to ± 60 Mvar was spe-cified. Compliance with the Q limit isachieved by load-dependent switchingof a capacitor bank and one of the twohigh-pass filters. Electronic reactivepower is used only in the light load range.Normally, there is a difference betweenthe connect and disconnect points ofthe reactive power elements. This pro-vides a ”switching hysteresis” whichprevents too many switching operationsor even a ”pumping” .

No

rmal

lo

ad

Ove

r lo

ad

Red

uce

d m

inim

um

lo

ad

No

rmal

min

imu

m l

oad

-

-

round UCa

UndergroundCables

Converter Station

Existing 275-kVTransmissionSystem

Undersea Cables

AlternatingCurrent

DirectCurrent

UndergroundCables

Overhead Line

Existing 275-kVTransmissionSystem

Converter Station

AlternatingCurrent

Northern Ireland Scotland

Moyle Interconnector

BallycronanMore

Auchencrosh

Principle Arrangement of an HVDC Transmission Project 4

1111

Date of contract 09/99Delivery period 27 months

System DataTransmission capacity 2 x 250 MWSystem voltages 250 kV DC

275 kV ACRated current 1000 ATransmission distance 63.5 km

The Principle Arrangement of an HVDCTransmission Project is reflected on theMoyle Interconnector project. The HVDCstations between Northern Ireland andScotland are operating with the followinghighlights:• Direct light triggered thyristor valves

for the complete HVDC system, with 1872 thyristors in total, with 20% betterreliability and all valve components freefrom oil.

• Triple tuned AC filter in both stations.• Unmanned stations, fully automatic

remote operation and automatic load schedule operation.

• Hybrid optical ohmic shunt for DC current measuring unit.

• Low noise station design for:– AC filter capacitor and reactors– converter transformer– converter valve water cooling system– DC hall with smoothing reactor

• Station design for DC see/land cable with integrated return conductor and fibre optic cable for control and communication.

4. Principle Arrangement of an HVDC Transmission Project4

155146

CONTROL

BUILDIN

G

1623

1

14

4

14

VALV

E HALL

DC HALL

250 k

V DC

CONVERTE

R TRANSF

ORMER

AREA

SHUNT

CAPA

CITO

R BA

NK

810

9

12

AC-Filter

C-Shunt

AC-Filter

AC-Filter

AC-Filter

AC-Filter

AC Bus

ThyristorValves

HVDC Station Auchencrosh

ThyristorValves

Smoothing Reactor

Smoothing Reactor

Pole 1, 250 MW

Pole 2, 250 MW

250 DC Power Cable.63,5 km to HVDC StationBallycronan MoreNorthern Ireland

13

275 kV AC SWITCHYARD

ST A

C F

ILTE

R

TT A

C F

ILTE

R

98

9

275 kV OHL

10

15

11

14

13

8

912

10

6 7 5 6 5 6

5

10

1 Quadruple Thyristor Valve 2 Converter Transformer 3 Air Core Smoothing Reactor 4 Control Room and Control Cubicle 5 AC Filter Capacitor 6 AC Filter Reactor 7 AC Filter Resistor 8 Circuit Breaker 9 Disconnector

10 Current Transformer11 Voltage Transformer12 Combined Current-Voltage Transformer13 Capacitive Voltage Transformer14 Surge Arrester15 Earthing Switch16 AC PLC Filter

Fig. 5.1-2: General arrangement ofa 500 kV MVU (valve tower)

Main Components5

5.1.1 Introduction

The thyristor valves make the conversionfrom AC into DC and thus are the centralcomponent of any HVDC converterstation. The thyristor valves are of theindoor type and air-insulated. Siemenshas more than 30 years experience inthe development and manufacturing ofthyristor valves and has maintained thetechnical leadership by introducing newinnovative concepts such as the corrosion- free water cooling and the self-protectingdirect-light-triggered thyristor. This directlyreflects in the high reliability of thesevalves.

14

Components

Siemens is a leading supplier of HVDCsystems all over the world. Our compo-nents are exceeding the usual qualitystandards and are system-tailored to theneeds of the grid.

12-pulse group

Fig. 5.1-1: Principle circuit diagramof a 12-pulse group consisting of threequadruple valves

Valvebranch

Multiplevalve unit –quadruplevalve

valve

valve

valve

valve

valve

valve

valve

valve

valve

valve

valve

valve

Thyristor Valves 5.1

5.1.2 Valve Design

The modular concept of the Siemensthyristor valves permits differentmechanical setups to best suit eachapplication: single, double, quadruplevalves or complete six-pulse bridges –either free – standing or suspended fromthe building structure.

For seismic requirement reasons whichexist in some regions of the world,the standard Siemens valves for longdistance transmission are suspendedfrom the ceiling of the valve hall. Thesuspension insulators are designed tocarry the weight and additional loadsoriginating for example from an unbal-anced weight distribution due to insulatorfailure, an earthquake or during main-tenance. Connections between modules(piping of cooling circuit, fibre optic ducts,buswork, and suspension insulatorfixtures) are flexible in order to allowstress-free deflections of the modulesinside an MVU (multiple valve unit)structure. Figure 5.1-2 shows a typicalquadruple valve tower for a 500 kV DCsystem. Each valve is made up of threemodules. Four arresters, each related toone valve, are located on one side of thevalve tower. Ease of access for main-tenance purposes, if required, is anotherbenefit of the Siemens valve design.By varying the number of thyristors permodule and the number of modules pervalve, the same design can be used forall transmission voltages that may berequired.

15

Fig. 5.1-4: Valve module with direct-light-triggered thyristor

Fig. 5.1-5: Silicon wafer and housing of a direct-light-triggered thyristor, fibre optic cable for gating

Fig. 5.1-6: The optical gate pulse is transmitteddirectly to the thyristor wafer

Light Pipe

Cu

Cu

Si

Mo

5.1.4 LTT (Light-Triggered Thyristor)

It has long been known that thyristorscan be turned on by injecting photonsinto the gate instead of electrons. Theuse of this new technology reduces thenumber of components in the thyristorvalve up to 80%. This simplificationresults in increased reliability andavailability of the transmission system.With LTT technology, the gating lightpulse is transmitted via a fibre optic cablethrough the thyristor housing directly tothe thyristor wafer and thus no elaborateelectronic circuits and auxiliary powersupplies are needed at high potential.The required gate power is just 40 mW.The forward overvoltage protection isintegrated in the wafer. Further benefitsof the direct light triggering are theunlimited black start capability and theoperation during system undervoltageor system faults without any limitations.In case of electrically triggered thyristors(ETT), this is only possible if enough firingenergy is stored long enough on thethyristor electronics.

Direct light-triggered thyristors withintegrated overvoltage protection (LTT)is a proven technology meanwhile andthe Siemens standard. In 1997, it hasbeen implemented successfully for thefirst time (Celilo Converter Station of thePacific Intertie). It shows excellent per-formance and no thyristor failures ormalfunction of the gating system havebeen recorded. BPA has emphasized itsconfidence in this technology in 2001 byawarding Siemens the contract to replaceall mercury arc valves with direct-light-triggered thyristor valves. Furthermore,this valve technology is used for theMoyle Interconnector (2 x 250 MW),which went into service in 2001 and ison contract for the 3000-MW, ± 500-kVGuizhou-Guangdong system.

Monitoring of the thyristor performanceis achieved by a simple voltage dividercircuit made from standard off-the-shelfresistors and capacitors; monitoringsignals are transmitted to ground poten-tial through a dedicated set of fibre opticcables as for the ETT. However, all elec-tronic circuits needed for the evaluationof performance are now located atground potential in a protected environ-ment, further simplifying the system.The extent of monitoring is the same asfor the ETT.It can be expected that this technologywill become the industry standard inHVDC thyristor valves of the 21st century,paving the way towards maintenance-free thyristor valves.16

Thyristor Valves5.1

Thyristor BlockingVoltage (UDRM)

Thyristor BlockingVoltage

LTT

Thyristor Current(IdN)

1970 1980 1990 2000 2003

2

4

6

8kV

2

4

6

kA

Thyristor Current forLong-Distance Transmission

Fig. 5.1-3: Thyristor development

5.1.3 Thyristor Development

Thyristors are used as switches and thusthe valve becomes controllable. Thethyristors are made of highly pure mono-crystalline silicon. The high speed ofinnovation in power electronics technol-ogy is directly reflected in the develop-ment of the thyristor. The high perform-ance thyristors installed in HVDC plantstoday are characterized by silicon waferdiameters of up to 5’’ (125 mm), blocking

voltages up to 8 kV and current carryingcapacities up to 4 kA DC. Thus no parallelthyristors need to be installed in today’sHVDC systems for handling the DCcurrent. The required DC system voltagesare achieved by a series connection ofa sufficient number of thyristors.

Cu = CopperSi = SiliconMo = MolybdenumLTT = Light-triggered thyristorsETT = electrically triggered

thyristors

Fig. 5.1-7: Piping of module cooling circuit –parallel flow (top); series flow (bottom)a) thyristor; b) heat sink; c) connection piping;d) manifold

b a

d

c

ba c

5.1.5 Valve Cooling

Siemens has used the parallel watercooling principle for more than 25 years.No corrosion problems have ever beenencountered.

The thyristors are stacked in the modulewith a heat sink on either side. The waterconnection to the heat sinks can bedesigned in parallel or series as shownin figure 5.1-7. The parallel cooling circuitprovides all thyristors with the samecooling water temperature. This allows abetter utilization of the thyristor capability.Siemens makes use of this principle,which offers the additional advantagethat electrolytic currents through the heatsinks – the cause for electrolytic corrosion– can be avoided by placing gradingelectrodes at strategic locations in thewater circuit. Siemens water cooling alsodoes not require any de-oxygenizingequipment.

5.1.6 Flame Resistance

Much effort has been invested bySiemens to minimize the fire risk:

• All oil has been eliminated from the valve and its components. Snubber capacitors and grading capacitors useSF6 as a replacement for impregnatingoil.

• Only flame-retardant and self-extin-guishing plastic materials are used.

• A wide separation between the modular units ensures that any localoverheating will not affect neighbouringcomponents.

• Careful design of the electrical connections avoids loose contacts.

The past has shown that Siemens HVDCinstallations have never been exposedto a hazardous valve fire. The testsperformed on actual components andsamples in the actual configuration asused in the valve indicate that the im-proved design indeed is flame-retardantand the risk of a major fire followinga fault is extremely low or even non-existent.

Fig. 5.1-8: Converter valvesSylmar HVDC station, Los Angeles, USA

1517

Siemens supplies transformers whichmeet all requirements concerning power,voltage, mode of operation, low noiselevel, connection techniques, type ofcooling, transport and installation. Theyalso comply with special national designrequirements.

All over the world, power transformersfrom Nuremberg enjoy a great reputation.What the Nuremberg plant manufacturesreflects today’s state of the art andtestifies to the highest levels of qualityand reliability. Our quality managementsystem is certified to DIN 9001, theworld’s most stringent standard. Ouraccredited test laboratories likewise meetthe latest specifications.

Project: Tian GuangHVDC bipolar long-distance transmissionPN = 2 x 900 MWUd = ± 500 kVTransformers:SN = 354/177/177 MVA1-ph/3-w unitUAC = 220 kV

Fig. 5.2-1: Converter transformer for the Tian GuangHVDC project during type test

Converter Transformer5.2

Converter transformer for the Three GorgesHVDC project 284 MVA, 1-ph/2-w unit

1618

Transformer Rating

STrafo Rec (6-pulse) = √2 * IdN * Usec Rec

IdN nominal DC currentUsec Rec Transformer-voltage

valve side (Rectifier)

STrafo Inv (6-pulse) = √2 * IdN * Usec Inv

Usec Inv Transformer voltage valve side (inverter)

Single-phase – two-winding transformer 1

Single-phase – three-winding transformer 1.6

Three-phase – two-winding transformer 2.2

Three-phase – three-winding transformer 3.6

5.2.1 Functions of the HVDCConverter Transformer

The converter transformers transformthe voltage of the AC busbar to therequired entry voltage of the converter.

The 12-pulse converter requires two3-phase systems which are spaced apartfrom each other by 30 or 150 electricaldegrees. This is achieved by installing atransformer on each network side in thevector groups Yy0 and Yd5.

At the same time, they ensure thevoltage insulation necessary in order tomake it possible to connect converterbridges in series on the DC side, as isnecessary for HVDC technology. Thetransformer main insulation, therefore,is stressed by both the AC voltage andthe direct voltage potential betweenvalve-side winding and ground. Theconverter transformers are equipped withon-load tap-changers in order to providethe correct valve voltage.

5.2.2 Transformer Design Variations

There are several aspects which play arole in selecting the transformer design:

Transportation Weight and Dimensions

In systems of high power, weight canbe an important consideration, inparticular where transportation is difficult.The relative transportation weights ofthe 4 major design types areapproximately as follows:

The transport dimension and the weightof the converter transformer dependson the limitations for street, railway andshipping, especially in the case ofbridges, subways and tunnels.

5.2.3 HVDC Makes Special Demandson Transformers

HVDC transformers are subject tooperating conditions that set them apartfrom conventional system or powertransformers. These conditions include:

• Combined voltage stresses• High harmonics content of the

operating current• DC premagnetization of the core

The valve windings which are connectedto the rectifier and the converter circuitare subject to the combined load stressof DC and AC voltage. Added to thisstress are transient voltages from outsidecaused by lightning strikes or switchingoperations.

The high harmonics content of theoperating current results from the virtuallyquadratic current blocks of the powerconverter. The odd-numbered harmonicswith the ordinal numbers of 5, 7, 11, 13,17 … cause additional losses in thewindings and other structural parts.

1719

5.2.4 Main Components of theConverter Transformer

Core

HVDC transformers are normally single-phase transformers, whereby the valvewindings for the star and delta connectionare configured either for one core withat least two main limbs or separately fortwo cores with at least one main limb,depending on the rated power and thesystem voltage. Appropriately sized returnlimbs ensure good decoupling for acombined arrangement of windings.

The quality of the core sheets, thelamination of the sheets, and the nominalinduction must all conform to specialrequirements covering losses, noise level,over-excitation, etc. Special attentionmust be paid to the DC premagnetizationof the core due to small asymmetriesduring operation and stray DC currentsfrom the AC voltage network. The effectsof DC premagnetization must becompensated by appropriate design andmanufacturing efforts (e.g. additional corecooling ducts, avoidance of flux pinchingin the core sheet).

Windings

The large number of parameters con-cerning transport limitations, rated power,transformer ratio, short-circuit voltage,and guaranteed losses require significantflexibility in the design of windings.

In concentric winding arrangements, staror delta valve windings lying directly onthe core have proven optimal in manycases. The line winding, normally with atapped winding, is then mounted radialoutside this core configuration.

The valve windings with high insulationlevels and a large portion of currentharmonics make particular demands onthe design and the quality of the windingmanufacturing. Together with itspressboard barriers, each limb set, in-cluding a valve, an overvoltage and atapped winding, forms a compact unit,which is able to cope with the demandmade by voltage stress, loss dissipation,and short-circuit withstand capability.

Tank

The unconventional tank design in HVDCtransformers result from the followingrequirements:

• The valve-side bushing should extendinto the valve hall

• The cooling system is mounted on theopposite side to facilitate rapid trans-former exchange

For HVDC transformers with delta andstar valve winding in one tank, the valvebushing must be arranged so that theirends conform to the geometry of thethyristor valve towers. This frequentlyleads to very high connection heightsand the need to mount the oil expansiontank at a significant height.

In close cooperation with the equipmentdesign department, the engineeringspecialists at the Nuremberg TransformerPlant have always been able to find adesign suited to every customerrequirement.

Bushings

Compared to porcelain, compositebushings provide better protectionagainst dust and debris. A 15% higherDC voltage testing level compared to thewindings underscores the particularsafety aspect of these components.

Special Tests for HVDC Transformers

Special tests for verifying operatingfunctionality are required for HVDCtransformers. The applicable internationalstandards are subject to constant furtherdevelopment. Separate tests with DCvoltage, switching and lightning impulsevoltages cover the range of differentvoltage loads. The 2-MV DC voltagegenerator in the Nuremberg TransformerPlant is well-suited for all required DCvoltage and reverse poling tests. Themost important criterion is partialdischarge. A maximum of 10 dischargesover 2000 pC during the last 10 minutesof the test is permitted.

Converter Transformer5.2

20

Fig. 5.3-2: Air-insulated smoothingreactor – Tian Guang project• Inductance: 150 mH• Rated voltage: 500 kV DC• Rated current: 1800 A DC

Fig. 5.3-1: Oil-insulated smoothing reactor –Three Gorges project• Inductance: 270 mH• Rated voltage: 500 kV DC• Rated current: 3000 A DC

The wall bushing in composite designis the state-of-the-art technology whichprovides superior insulation perform-ance.

Smoothing Reactor 5.3

smoothing reactors is often selectedin the range of 100 to 300 mH for long-distance DC links and 30 to 80 mH forback-to-back stations.

5.3.3 Arrangement of theSmoothing Reactor

In an HVDC long-distance transmissionsystem, it seems quite logical that thesmoothing reactor will be connected inseries with the DC line of the station pole.This is the normal arrangement.However in back-to-back schemes, thesmoothing reactor can also be connectedto the low-voltage terminal.

5.3.4 Reactor Design Alternatives

There are basically two types of reactordesign:• Air-insulated dry-type reactors• Oil-insulated reactors in a tank

The reactor type should be selected takingthe following aspects into consideration:• Inductance• Costs• Maintenance and location of spare

units• Seismic requirements

An advantage of the dry-type reactor isthat maintaining spare units (to the extentnecessary) is not very expensive becausethey usually consist of several partialcoils. However for very large inductancesit is possible to have more than one unitand it could be a problem if much spaceis not available.

In high seismic regions, setting them onpost-insulators or on an insulating plat-form is a possible problem. Oil-insulatedsmoothing reactors are then the preferredsolution.

The oil-insulated reactor is economicalfor very high power (Id2 * Ldr).It is the best option for regions with highseismic requirements.

One bushing of the oil-insulatedsmoothing reactor penetrates usuallyinto the valve hall, while the otherbushing is normally in a vertical position.For the air-insulated dry-type smoothingreactor, a wall bushing is needed toconnect with the valves.

5.3.1 Functions of the SmoothingReactor

• Prevention of intermittent current• Limitation of the DC fault currents• Prevention of resonance in the

DC circuit• Reducing harmonic currents including

limitation of telephone interference

Prevention of intermittent current

The intermittent current due to thecurrent ripple can cause high over-volt-ages in the transformer and the smoothingreactor. The smoothing reactor is usedto prevent the current interruption atminimum load.

Limitation of the DC fault current

The smoothing reactor can reduce thefault current and its rate of rise forcommutation failures and DC line faults.

This is of primary importance if a longDC cable is used for the transmission.For an overhead line transmission, thecurrent stress in valves is lower than thestress which will occur during valve shortcircuit.

Prevention of resonance in theDC circuit

The smoothing reactor is selected to avoidresonance in the DC circuit at low orderharmonic frequencies like 100 or 150 Hz.This is important to avoid the ampli-fication effect for harmonics originallyfrom the AC system, like negative se-quence and transformer saturation.

Reducing harmonic currentsincluding limitation of telephoneinterference

Limitation of interference coming fromthe DC overhead line is an essentialfunction of the DC filter circuits. However,the smoothing reactor also plays animportant role to reduce harmoniccurrents acting as a series impedance.

5.3.2 Sizing of the smoothing Reactor

While the current and voltage rating ofthe smoothing reactor can be specifiedbased on the data of the DC circuit, theinductance is the determining factorin sizing the reactor.Taking all designaspects above into account, the size of

21

Fig. 5.4.1-2:AC filters and capacitor banksof Gezhouba/Shanghai

22

Harmonic Filters5.4

The filter arrangements on the AC sideof an HVDC converter station have twomain duties:• to absorb harmonic currents generated

by the HVDC converter and thus to reduce the impact of the harmonics on the connected AC systems, like ACvoltage distortion and telephone interference

• to supply reactive power for compen-sating the demand of the converter station

Each filter branch can have one to threetuning frequencies. Figure 5.4.1-1 showsdifferent harmonic filter types with theirimpedance frequency characteristics.

Fig. 5.4.1-1:Different harmonicfilter types

Harmonic Order

Filte

r Im

peda

nce

0 10 20 30 40 50

Harmonic Order

Filte

r Im

peda

nce 1000

800600400200

00 10 20 30 40 50

Harmonic Order

Filte

r Im

peda

nce 800

600

400

200

00 10 20 30 40 50

Single-tuned

1000800600400200

0

Double-tuned

Triple-tuned

5.4.1.1 Design Criteria for AC Filters

Reactive Power Requirements

The reactive power consumption of anHVDC converter depends on the activepower, the transformer reactance andthe control angle. It increases with in-creasing active power. A common re-quirement to a converter station is fullcompensation or overcompensation atrated load. In addition, a reactive band forthe load and voltage range and thepermitted voltage step during bankswitching must be determined. Thesefactors will determine the size andnumber of filter and shunt capacitorbanks.

Harmonic Performance Requirements

HVDC converter stations generatecharacteristic and non-characteristicharmonic currents. For a twelve-pulseconverter, the characteristic harmonicsare of the order n = (12 * k) ± 1 (k = 1,2,3...). These are the harmonic componentsthat are generated even during idealconditions, i.e. ideal smoothing of thedirect current, symmetrical AC voltages,transformer impedance and firing angles.The characteristic harmonic componentsare the ones with the highest currentlevel, but other components may alsobe of importance. The third harmonic,which is mainly caused by the negativesequence component of the AC system,will in many cases require filtering.

An equivalent circuit for determinationof harmonic performance is given infigure 5.4.1-3. The most commonly usedcriteria for harmonic performance are

Fig. 5.4.1-3: Equivalent circuit for calculationof harmonic voltages and currents in theAC system

related to the harmonic voltage on theconverter station busbar. The purpose ofthe filter circuit is to provide sufficientlylow impedances for the relevant harmo-nic components in order to reduce theharmonic voltages to an acceptable level.

The acceptance criteria for the harmonicdistortion depend on local conditions andregulations. A commonly used criterionfor all harmonic components up to the49th order is as follows:

Dn individual harmonic voltage distortion of order n in percent of the fundamental AC busbar voltage(typical limit 1%)

Drms total geometric sum of individual voltage distortion Dn (typical limit 2%)

The BTS Telephone Interference Factor(TIF) and the CCITT Telephone HarmonicForm Factor (THFF) are determined withweighted factors in order to evaluate thevoltage distortion level on the AC busbarwith respect to the expected interferencelevel in nearby analogue telephonesystems. The IT product is a criterion forharmonic current injected into AC over-head lines. The criteria based on tele-phone interference are in many casesirrelevant, because modern digital tele-phone systems are insensitive to har-monic interference.

Network Impedance

The distortion level on the AC busbardepends on the grid impedance as wellas the filter impedance. An open circuitmodel of the grid for all harmonics is noton the safe side. Parallel resonancebetween the filter impedance and thegrid impedance may create unacceptableamplification of harmonic componentsfor which the filters are not tuned. Forthis reason, an adequate impedancemodel of the grid for all relevant har-monics is required in order to optimizethe filter design.

HarmonicCurrentSource

FilterImpedance

NetworkImpedance

Ih ZF ZNUh

There are basically two methods toinclude the network impedance in thefilter calculations:

• to calculate impedance vectors for allrelevant harmonics and grid conditions

• to assume locus area for the imped-ance vectors

The modelling of a complete AC networkwith all its components is very complexand time-consuming. For this reason, thelocus method is very often used. It isbased on a limited number of measure-ments or calculations. Different locusareas for different harmonics or bandsare often determined to give a moreprecise base for the harmonic perform-ance calculation.

A typical locus area is shown infig. 5.4.1-4. It is assumed that the im-pedance vector will be somewhere insidethe perimeter of the coloured area.

The impedance vector of the filter istransformed into the Y plane for eachharmonic frequency.

With both the network and the filterimpedances plotted in the admittanceplane, the shortest vector between thefilter admittance point and the networkadmittance boundary gives the lowestpossible admittance value for the parallelcombination of the network and the filter.This value is used to determine thehighest possible harmonic voltage.

Fig. 5.4.1-4 Circle of network admittance andthe resonance conditions

Yf

Yres

Ymin = 1/Rmax Ymax = 1/Rmin

AC Harmonic Filter 5.4.1

Ymin = 1/Rmax Ymax = 1/Rmin

23

Ih = harmonic source currentZf = filter impedanceZN = network impedanceUh = harmonic voltage

The selective resonance method repre-sents a reasonable compromise. It takesinto consideration the fact that thehighest voltage distortion (highest har-monic voltage) occurs with a parallelresonance between filter and AC network.It is unrealistic however, to assume thatsuch a parallel resonance takes place atall frequencies. Normally it is sufficientto consider in the calculation of totaldistortion and TIF value only two maxi-mum individual distortions from theresonance calculation. The AC network isassumed to be open for the remainingharmonic currents.

The filter calculations must reflect de-tuning caused by AC network frequencydeviations and component parameterdeviations. Production tolerances, tempe-rature drift and failure of capacitor ele-ments are the main contributors toparameter deviations.

Requirements to Ratings

Steady-State Calculation

The voltage and current stresses of ACfilters consist of the fundamentalfrequency and harmonic components.Their magnitudes depend on the ACsystem voltage, harmonic currents,operating conditions and AC systemimpedances. The rating calculations arecarried out in the whole range ofoperation to determine the higheststeady-state current and voltage stressesfor each individual filter component.

Transient Calculation

The objective of the transient ratingcalculation is to determine the highesttransient stresses for each componentof the designed filter arrangement. Theresults of the transient calculation shouldcontain the voltage and current stressesfor each component, energy duty for filterresistors and arresters, and the insulationlevels for each filter component.

To calculate the highest stresses of bothlightning and switching surge type,different circuit configurations and faultcases should be studied:

• Single-Phase Ground FaultThe fault is applied on the converter AC bus next to the AC filter. It is assumed that the filter capacitor is charged to a voltage level corres-ponding to the switching impulse protective level of the AC bus arrester.

• Switching SurgeFor the calculation of switchingsurge stresses, a standard wave of 250/2500 µs with a crest value equalto the switching impulse protective level of the AC bus arrester is appliedat the AC converter bus.

• Filter EnergizationThe AC filter is assumed to be energized at the moment for the maximum AC bus peak voltage. This case is decisive for the inrush currentsof AC filters.

• Fault Recovery after Three-Phase Ground FaultVarious fault-clearing parameters should be investigated to determine the maximum energy stresses for ACfilter arresters and resistors. The worst-

case stresses are achieved if the HVDC converters are blocked after fault initiation, while the AC filters remain connected to the AC bus afterfault clearing and recovery of the AC system voltage. In this case, a tempo-rary overvoltage with high contents ofnon-characteristic harmonics will occurat the AC bus due to the effects of load rejection, transformer saturationand resonance between filter andAC network at low frequency.

Fig. 5.4.2-1 DC filter of Guangzhou/China

AC Harmonic Filter5.4.1

24

5.4.2.1 DC Filter Circuits

Harmonic voltages which occur on theDC side of a converter station cause ACcurrents which are superimposed on thedirect current in the transmission line.These alternating currents of higherfrequencies can create interference inneighbouring telephone systems despitelimitation by smoothing reactors.DC filter circuits, which are connectedin parallel to the station poles, are aneffective tool for combating these pro-blems. The configuration of the DC filtersvery strongly resembles the filters on theAC side of the HVDC station. There areseveral types of filter design. Single andmultiple-tuned filters with or without thehigh-pass feature are common. One orseveral types of DC filter can be utilizedin a converter station.

The equivalent disturbing current com-bines all harmonic currents with the aidof weighting factors to a single inter-ference current. With respect to tele-phone interference, it is the equivalentto the sum of all harmonic currents. Italso encompasses the factors whichdetermine the coupling between theHVDC and telephone lines:

• Operating mode of the HVDC system(bipolar or monopolar with metallic orground return)

• Specific ground resistance at point x

The intensity of interference currents isstrongly dependent on the operatingcondition of the HVDC. In monopolaroperation, telephone interference issignificantly stronger than in bipolaroperation.

25

DC Harmonic Filter 5.4.2

5.4.2.2 Design Criteria for DC FilterCircuits

The interference voltage induced on thetelephone line can be characterized bythe following equation:

Vin(x) = Interference voltage on thetelephone line at point x(in mV/km)

Hµ = Weighting factors which reflectthe frequency dependence ofthe coupling between tele-phone and HVDC lines

Cµ = “C message“ – weighting factors

Iµ(x) = Resulting harmonic currentof the ordinal number µ in the HVDC line at point x as thevector sum of the currents caused by the two HVDC stations

Ieq = Psophometric weighted equivalent disturbing current

Z = Mutual coupling impedance between the telephone andHVDC lines

where

Z * IeqVin(x) =

(Hµ * Cµ * Iµ(x))2∑√m

1Ieq =

Active filters can be a supplement topassive filters due to their superior per-formance. They can be installed on theDC side or on the AC side of the convert-er. The connection to the high-voltagesystem is achieved by means of a passivefilter, forming a so-called hybrid filter. Thisarrangement limits the voltage level andthe transient stresses on the active part,so that comparatively low equipmentratings can be used. Appropriate designallows the exploitation of the positivecharacteristics of both passive and activefilters. Additionally, the passive partcan be used as a conventional passivefilter if the active part is by-passed formaintenance purposes.

Fig. 5.4.3-1: Active DC filter on site(Tian Guang HVDC project)

The Siemens active filters use voltage-sourced IGBT converters with a highswitching frequency to produce an outputvoltage up to approximately 700 Vpeak,containing harmonics up to the 50th asrequired. A powerful high-speed con-trol and protection system processesthe currents and/or voltages measuredat the network by appropriate sensorsand produces the control pulses for theIGBT’s.

A transformer matches the voltage andcurrent levels at the converter outputand provides the required insulation level.The goal of the scheme is to inject har-monics in the network with the sameamplitude and the opposite phase of theharmonics at the measurement point inorder to cancel them.

The filter for AC application comprisesthree single-phase systems controlledby a common digital control system. Amajor difference is the measurement:instead of measuring the line current,the active filter at Tjele measures andeliminates harmonics at the 400 kV ACbusbars of the station. This has the ad-vantage that the harmonic control requi-res just one measurement point, compa-red to a current measuring scheme,which would require to measure thecurrent at several points and combiningthe measured signals. The other advan-tage is that the active filter works justlike a passive filter ideally should do, i.e.eliminating the voltage in the bus, thusrepresenting no change in philosophy.

The active filter is fully assembled in atransportable container and is tested atthe factory as a complete system beforeshipping. Fig. 5.4.3-5 shows the installedactive AC filter (in the container) at theTjele substation.

IGBTconv.

400 kV Bus

1

2

3 4 5 6 7

8 9

Active AC Filter

CapacitivePotential

Transformer

CommonControl and AuxiliaryEquipment

From theOther Phases

ExistingPassive

FilterZF2

To/From MainComponentsof the OtherPhases (inthe Container)

Main Components(one Set for Each Phase)

Simadyn Dcubicle(control andprotection)

UndergroundCable

Active Harmonic Filter5.4.3

No.Component

1 IGBT converter2 Reactor for inductivity

adapting3 Thyristor switch for con-

verter overvoltage andovercurrent protection

4 Transformer5 Low-pass filter6 Vacuum switch7 ZnO arrester8 Isolators and grounding

switches9 LC branch for deviating

the 50-Hz currentcomponent

Main Components

OpticalInterface

LP

Fig. 5.4.3-2: Single-line diagramof the active AC filter.All phases have the same topology.

26

One harmonic controller is dedicated toeach harmonic selected for eliminationby the action of the active filter. In theseharmonic controllers, the particularharmonic is isolated and expressed bya complex signal in the frequency domain.

This is done through multiplication bysin (hωt) and cos (hωt), where h is theorder of the harmonic, ω the networkangular frequency and t the time. Thesetwo orthogonal signals are produced bya module synchronized by the funda-mental component of the filter current.The signal pair obtained after thementioned multiplication and filteringfeeds a complex controller with PIcharacteristic. The output of the controlleris then shifted back to the time domainby multiplication by cos (hωt) and sin (hωt).

The process is essentially linear, so thatall harmonic controllers can operatesimultaneously and the sum of all har-monic controller outputs gives the wave-form required by the active filter. Thissignal is then given to the IGBT controlmodule, which includes a pulse widthmodulator besides functions for pro-tection and supervision of the converter.

Fig. 5.4.3-5: Installation of the active AC filter, 400-kVsubstation Tjele (Denmark)

Fig. 5.4.3-4: Plots from measurement:left without, right with active filter control

Fig. 5.4.3-3: Principle block diagram of the harmonic control

IGBTcontrol

LPFilter

OpticalInput

Self-Tuningsystem

Σ

From OtherHarmonicControllersSynchronization

Harmonic Controller

cos (hωt)sin (hωt)

To the IGBTConverter

Voltage on the400-kV Busbar

PIController

To OtherHarmonicControllers

Harmonic voltages at the 400-kV bus(L1) without and…

…with active filter control (23th, 25th, 35th, 47th and 49th harmonics)

27

Siemens surge arresters are designedoptimally to the following requirements:

Excellent pollution performance forcoastal and desert regions or in areaswith extreme industrial air pollution.

High mechanical stability, e.g. for usein seismic zones.

Extremely reliable pressure reliefbehaviour for use in areas requiringspecial protection.

What is more, all Siemens surge arrestersare sized for decades and the materialused provides a contribution towards theprotection of the environment.

The main task of an arrester is to protectthe equipment from the effects of over-voltages. During normal operation, it shouldhave no negative effect on the powersystem. Moreover, the arrester must beable to withstand typical surges withoutincurring any damage. Non-linear resistorswith the following properties fulfil theserequirements:

• Low resistance during surges so that overvoltages are limited

• High resistance during normal operationin order to avoid negative effects on thepower system and

• Sufficient energy absorption capabilityfor stable operation

MO (Metal Oxide) arresters are used inmedium-, high-and extra-high-voltagepower systems.

Here, the very low protection level andthe high energy absorption capabilityprovided during switching surges areespecially important. For high voltagelevels, the simple construction of MOarresters is always an advantage.

Arresters with Polymer Housings

Fig. 5.5-2 shows two Siemens MOarresters with different types of housing.In addition to what has been usual up tonow – the porcelain housing – Siemensoffers also the latest generation of high-voltage surge arresters with polymerhousing.

Fig. 5.5-2: Measurement of residual voltageon porcelain-housed (foreground) andpolymer-housed (background) arresters

Fig. 5.5-1: Current/voltage characteristics of a non-linear MO arrester

Current through Arrester Ia [A]

150 °C

20 °C115 °C

2

1

010-4

Rated Voltage ÛR

Continuous OperatingVoltage ÛC

10-3 10-2 10-1 102 103 1041 10

28

Surge Arrester5.5

Arrester Voltage Referredto Continuous OperatingVoltage Û/ÛC

Fig. 5.5-3 shows the sectional view ofsuch an arrester. The housing consists ofa fibre-glass-reinforced plastic tube withinsulating sheds made of silicon rubber.The advantages of this design which hasthe same pressure relief device as anarrester with porcelain housing areabsolutely safe and reliable pressurerelief characteristic, high mechanicalstrength even after pressure relief andexcellent pollution-resistant properties.The very good mechanical features meanthat Siemens arresters with polymerhousing (type 3EQ/R) can serve as postinsulators as well. The pollution-resistantproperties are the result of the water-repellent effect (hydrophobicity of thesilicon rubber).

The polymer-housed high-voltage arresterdesign chosen by Siemens and the high-quality materials used by Siemensprovide a whole series of advantagesincluding long life and suitability foroutdoor use, high mechanical stabilityand ease of disposal.

For terminal voltage lower than thepermissible maximum operating voltage(MCOV), the arrester is capacitive andcarries only few milli-amps. Due to itsextreme non-linear characteristics, thearrester behaves at higher voltages aslow-ohmic resistor and is able to dis-charge high current surges. Throughparallel combination of two or morematched arrester columns, higher energyabsorption capability of the ZnO arrestercan be achieved.

Routine and type tests have beendetermined in accordance with theinternational standards:

IEC 60060 High-voltage test techniquesIEC 60071 Insulation coordinationIEC 60099 Surge arresters

Fig. 5.5-3: Cross-section of a polymer-housedarrester

Seal

Pressure Relief Diaphragm

Compressing Spring

Metal Oxide Resistors

Composite Polymer HousingFRP Tube/Silicon Sheds

Flange with Gas Diverter Nozzle

29

Arrester Type Location Main Task

AC bus arrester ’A’ The ZnO arrester will be installed close to Limit the overvoltages on the primary and secondarythe converter transformer line side bushing side of the converter transformer

AC filter bus The ZnO arrester will be installed at the Protect the AC filters busbar againstarrester ‘Aa’ busbar of the AC filter banks lightning surges

Valve-arrester ‘V’ 3-pulse commutation group The main events to be considered with respect to arrester discharge currents and energies are:a) Switching surges from the AC system through converter transformerb) Ground fault between valve and HV bushing of converter transformer during rectifier operation

Converter group 12-pulse converter group Protection against overvoltages from thearrester ‘C’ AC and DC side

DC bus arrester ‘D’ At the HV smoothing They will protect the smoothing reactor and thereactor and at the DC lines converter station (e.g. DC switchyard) against

overvoltages coming from the DC side

Neutral DC bus Neutral DC bus The neutral bus arresters protect the LV terminal of arrester ‘E’ the12-pulse group and the neutral bus equipment

AC filter arrester AC filter The operating voltage for the AC filter arresters consists‘Fac‘ of low fundamental frequency and harmonic voltages.

Overvoltages can occur transiently during faults

DC filter arrester DC filter The operating voltage for the DC filter arresters consists‘Fdc‘ of low DC component and harmonic voltages.

Overstresses may occur transiently during DC busfault to ground

4

A

6

7

V

9

E ECn

DC

D

10

12C3

8

Aa

Fac1

AC Filter Bank

ACFilter2

Fac2

AC Bus

A

Valve Hall Boundary

5

V

V

V

Neutral

DCFilter

D11

10

8Neutral

Fdc1

Fdc2

to DC line1 DC

Surge Arrester5.5

30

5.6.1 DC Transmission Line

DC transmission lines could be part ofoverall HVDC transmission contract eitherwithin a turnkey package or as separatelycontracted stand-alone item, later inte-grated into an HVDC link.

As an example of such a transmissionline design, an existing bipolar tower forthe 300-kV link between Thailand andMalaysia is shown in Fig.5.6.1-1 and abipolar AC transmission tower of TianGuang is shown in Fig. 5.6.1-2.

DC Transmission Circuit 5.6

31

5.6.1.1 Towers

Such DC transmission lines are mechani-cally designed as it is practice for normalAC transmission lines; the maindifferences are:• The conductor configuration• The electric field requirements• The insulation design

5.6.1.2 Insulation

The most critical aspect is the insulationdesign and therefore this topic isdescribed more detailed below:

For DC transmission lines, the correctinsulation design is the most essentialsubject for an undisturbed operationduring the lifetime of the DC plant.

Design Basics

• The general layout of insulation is based on the recommendations ofIEC 60815 which provides 4 pollutionclasses.

• This IEC is a standard for AC lines. It has to be observed that the creepagedistances recommended are based onthe phase-to-phase voltage (UL-L). When

transferring these creepage distancesrecommended by IEC 60815 to a DCline, it has to be observed that the DCvoltage is a peak voltage pole to ground value (UL-G). Therefore, these creepage distances have to be multi-plied by the factor √3.

• Insulators under DC voltage operationare subjected to more unfavourable conditions than under AC due to highercollection of surface contamination caused by constant unidirectional electric field. Therefore, a DC pollutionfactor as per recommendation ofCIGRE (CIGRE-Report WG04 ofCigre SC33, Mexico City 1989) hasto be applied.

The correction factors are valid for por-celain insulators only. When taking com-posite insulators into consideration,additional reduction factors based on theFGH report 291 “Oberflächenverhaltenvon Freiluftgeräten mit Kunststoffge-häusen“ must be applied.

Fig. 5.6.1-1:DC transmission line(bipolar tower 300-kV link)

Fig. 5.6.1-2:DC transmission tower(bipolar) Tian Guang(South China)

Long-Rod Porcelain Type

Positive Aspects:• Long-term experience/track record• Good mechanical strength• Puncture-proof• Good self-cleaning ability• Less intermediate metal parts• Due to caps on both insulator ends not

subjected to pin corrosion because of low track current density

• Moderate price

Negative Aspects:

• Heavy strings• String not very flexible• Under extreme vandalism failure of

string possible

Composite Long-Rod Type

Positive Aspects:• Small number of insulators in one string• Up to 400 kV per unit possible• Good mechanical strength, no chipping

of sheds possible• Very light – easy handling during

construction and maintenance, logisticaladvantages in areas with poor access

• Puncture-proof• Good self-cleaning behaviour –

hydrophobicity of surface which offersadvantages of less creepage distance up to pollution class II

• Very good RIV and corona behaviour• Good resistance against vandalism• Shorter insulator string length• Very competitive price

Negative Aspects:

• Relatively short track record in DC appli-cation (since 1985 first major applicationin the USA)

• Less tracking resistance against flash-over (can be improved by means of corona rings)

Cap and Pin Porcelain Long-Rod Composite Long-Rod

Insulator string length 5270 mm 5418 mm 4450 mm31 insulators 4 insulators 1 insulator

Creepage per unit 570 mm 4402 mm 17640 mm

Weight of string 332 kg 200 kg 28 kg

Breaking load 160 kN 160 kN 160 kN

32

Example/Comparison of Insulator Application for a 400 kV Transmission Line

Types of Insulators

There are 3 different types of insulatorsapplicable for DC transmission lines:• Cap and pin type• Long-rod porcelain type• Composite long-rod type

In detail:

Cap and Pin Type

Positive Aspects:• Long-term experience/track record• Good mechanical strength• Vandalism-proof• Flexibility within the insulator string

Negative Aspects:• Very heavy strings• Insulator not puncture-proof• Poor self-cleaning ability• Loss of strength/reliability due to

corrosion of pin in polluted areas causedby high track current density (this is extremely important for DC lines)

• Many intermediate metal parts• High RIV and corona level• For DC applications, special shed design

and porcelain material necessary• Very expensive

DC Transmission Line5.6.1

2) Oil-Filled CableIn comparison to mass-impregnatedcables, the conductor is insulated bypaper impregnated with a low-viscosityoil and incorporates a longitudinal ductto permit oil flow along the cable. Oil-filled cables are suitable for both AC andDC voltages with DC voltages up to600 kV DC and great sea depths. Dueto the required oil flow along the cable,the transmission line lengths are howeverlimited to <100 km and the risk of oilleakage into the environment is alwayssubject to discussions.

5.6.2.3 Future Developments forHVDC Cables

Most of the research and developmentactivities for new cable types are donewith the insulation material. Theseinclude:

1) XLPE

To overcome the disadvantages of theabove mentioned cable types, extensiveR&D was conducted by the cablesuppliers. The result is the XLPE cable.XLPE means ‘cross-linked polyethylene‘and forms the insulation material. Theconductor is the segmented copperconductor insulated by extruded XLPElayers. The insulation material is suitablefor a conductor temperature of 90°C anda short-circuit temperature of 250°C.Although the main application for XLPEcables is the land installation and the off-shore industry, XLPE with extrudedinsulation material for HVDC systems oflower transmission capacities are underdevelopment.

2) Lapped Thin Film Insulation

As insulating material a lapped non-impregnated thin PP film is used insteadof the impregnated materials. The testsfor the cable itself are completed. Nowthe tests for the accessories such asjoints are under process.This type of cable can sustain up to 60%higher electrical stresses in operation,making it suitable for very long and deepsubmarine cables.

Another area of development are thecable arrangements. For monopolartransmission systems, either the returnpath was the ground (’ground return‘) ora second cable. The first solution alwaysprovokes environmental concerns where-as the second one has excessive impacton the costs for the overall transmissionscheme.Therefore, a new cable was developedwith an integrated return conductor. Thecable core is the traditional design for amass-impregnated cable and the returnconductor is wound outside the leadsheath. The conductor forms also part ofthe balanced armour, together with theflat steel wire layer on the outside of thereturn conductor insulation.This cable type was installed betweenScotland and Northern Ireland for250 kV and 250 MW. R&D is ongoing toincrease the voltage as well as the capa-city of the cable with integrated returnconductor.

33

DC Cable 5.6.2

5.6.2.1 General Application forDC Cables

An important application for HVDC aretransmission systems crossing the sea.Here, HVDC is the preferred technologyto overcome distances > 70 km and trans-mission capacities from several hundredto more than a thousand MW (for bipolarsystems). For the submarine transmissionpart, a special cable suitable for DCcurrent and voltage is required.

5.6.2.2 Different Cable Types

For HVDC submarine cables there aredifferent types available.

1) Mass-Impregnated CableThis cable type is used in most ofthe HVDC applications. It consistsof different layers as shown inFig. 5.6.2.2-1.

The conductor is built of stranding copperlayers of segments around a centralcircular rod. The conductor is covered byoil and resin-impregnated papers. Theinner layers are of carbon-loaded paperswhereas the outer layer consists ofcopper-woven fabrics.The fully impregnated cable is then lead-sheathed to keep the outside environ-ment away from the insulation. Thenext layer is the anti-corrosion protectionwhich consists of extruded polyethylene.Around the polyethylene layer galvanizedsteel tapes are applied to prevent thecable from permanent deformationduring cable loading. Over the steel tapesa polypropylene string is applied followedby galvanized steel wire armour.The technology is available for voltagesup to 500 kV and a transmission capacityof up to 800 MW in one cable with instal-lation depths of up to 1000 m under sealevel and nearly unlimited transmissionlengths. The capacity of mass-impreg-nated cables is limited by the conductortemperature which results in low overloadcapabilities.

1 Conductor of copper-shaped wires2 Insulation material3 Core screen4 Lead alloy sheath5 Polyethylene jacket6 Reinforcement of steel tapes7 Bedding8 Armour of steel flat wires

Fig. 5.6.2.2-1: Mass-impregnated cable

32

Duties

The HSNBS must commutate somedirect current into the ground electrodepath in case of faults to ground at thestation neutral.

The HSGS is needed to connect thestation neutral to the station ground gridif the ground electrode path becomesisolated.

If one pole of a bipolar system has to beblocked, monopolar operation of thesecond pole is achieved automatically,but with return current through ground(refer to Fig. 5.6.3-1). If the duration ofground return operation is restricted, analternate mode of monopolar operationis possible if the line of the blocked polecan be used for current return. This modeis called metallic return (refer to Fig.5.6.3-1). The MRTB is required for thetransfer from ground to metallic returnwithout interruption of power flow.

The GRTS is needed for the retransferfrom metallic return to bipolar operationvia ground return, also withoutinterruption of power flow.

Fig. 5.6.3-1: HVDC system configurations

MRTB at Tian Guang/China DC switchyard

34

High Speed DC Switches5.6.3

5.6.3.1 General

Like in AC substations, switching devicesare also needed in the DC yard of HVDCstations. One group of such devices canbe characterized as switches with directcurrent commutation capabilities,commonly called ”high-speed DCswitches” .

Siemens standard SF6 AC circuit-breakersof proven design are able to meet therequirements of high-speed DC switches.

Type

HSNBS(High-Speed Neutral Bus Switch)

HSGS(High-Speed Ground Switch)

MRTB(Metallic Return Transfer Breaker)

GRTS(Ground Return Transfer Switch)

5.6.3.2 Types and Duties of the High-Speed DC Switches

Electrodes

AC

Sys

tem

1

AC

Sys

tem

2

HVDCOverhead Line

HVDCOverhead Line

Bipolar

Monopolar ground return

AC

Sys

tem

1

AC

Sys

tem

2

Electrodes

HVDCOverhead Line

HVDCOverhead Line

monopolar metallic return

AC

Sys

tem

1

AC

Sys

tem

2

Electrodes

HVDCoverhead line

HVDCoverhead line

Monopolar metallic return

AC

Sys

tem

1

AC

Sys

tem

2

Electrodes

HVDCOverhead Line

HVDCOverhead Line

With reference to Fig 5.6.3-4, the principleof commutation is as follows: At t0,thecontacts of the breaker separate, therebyintroducing an arc into the circuit. Thecharacteristic of this arc sets up an oscilla-tory current (frequency determined byLp Cp) which is superimposed on thecurrent I1. As Rp is very small, the oscillationis not damped but increases. As soon asthe current I1 passes through zero (referto t1 in Fig. 5.6.3-4), the breaker current isinterrupted. I3 ,however, remains unchangednow charging the capacitor Cp until itreaches a voltage limited by the energyabsorber. This voltage acts as a countervoltage to reduce the current I3 and to in-crease the current I4 (refer to Fig. 5.6.3-4and Fig. 5.6.3-3). When the absorberlimiting voltage has been reached, thecurrent I3 flows into the absorber whichdissipates an amount of energy deter-mined by the counter voltage to bring I3to zero. When I3 has dropped to zero, I4equals I0 and the current commutationfrom ground to metallic return has beencompleted. It should be noted that thecurrent I0 of the system (refer to Fig. 5.6.3-2)did not change, i.e. the power transmissionwas never interrupted.There are also MRTB principles other thanthe explained one which are based oncomplex resonant circuits, externally exci-ted with additional auxiliary power sources.With respect to reliability and availability,the advantage of the above principle withpassive resonant circuit which is used bySiemens is quite evident. The nozzlesystem and specifically the flow of SF6

gas in the Siemens standard SF6 ACbreakers result in an arc characteristicwhich ensures reliable operation of thepassive resonant circuit. One unit of astandard three-phase AC breaker is used.Extensive series of laboratory tests haveshown the capabilities of Siemens SF6

breakers for this application. Furthermore,such switches are successfully in operationin various HVDC schemes.

The ground resistance Re is normally muchlower than the metallic resistance Rm.Therefore, during the transitional steady-state condition with both MRTB and GRTSclosed, most of the current is flowingthrough ground which determines thecommutation requirements for MRTB andGRTS. I3 may reach values of up to 90%of the total current I0 and I4 values of upto 25% of I0.The following considerations refer to MRTBonly. From the above it can be concludedthat the commutation duties for transferfrom ground to metallic return (MRTB) aremuch heavier than from metallic to groundreturn (GRTS).Fig. 5.6.3-3 shows the basic MRTB circuit.An energy absorber and the LpCp resonantcircuit (Rp represents the ohmic resistanceof that branch only) are connected inparallel to the main switch (MRTB) whichis a conventional SF6-type AC breaker.

5.6.3.3 Design Considerations

Details regarding the duties of ”HSNBS”and ”HSGS” are not discussed here butthe more severe requirements for “MRTB“and ”GRTS“ are explained.

Fig. 5.6.3-2 shows the disposition of MRTBand GRTS. Rm and Lm represent resistanceand inductance of the transmission line.Re and Le comprise resistances andinductances of the ground return path.

Fig. 5.6.3-2: Equivalent circuit relevant toMRTB and GRTS operation

LeRe

Rm Lm

MRTB

GRTS

I3

I4

I0

Fig. 5.6.3-3: Details of the MRTB circuit

Re Le

Energy Absorber

MRTB

I0

GRTSI3

Rm LmI4

Rp LpCp

Fig. 5.6.3-4: Principles of MRTB operation

tc

I1

Uabsorber

Uarc

t0 t1 t

I3

I3

Lp Cp

I1 Uarc

3335

As shown in figure 5.6.4-1, the electrodeconductor itself, which is generally madeof iron, is laid horizontally at a depth ofapproximately 2 m. It is embedded incoke which fills a trench having a crosssection of approximately 0.5 x 0.5 m2.

The advantage of this design becomesapparent in anodic operation. The passageof the current from the electrodeconductor into the coke bed is carriedprimarily by electrons, and is thus notassociated with loss of the material.

Several typical patterns of horizontalland electrodes are illustrated infigure 5.6.4-2

Fig. 5.6.4-1: Cross section through a horizontalland electrode

a) Line Electrodeb) Multi-Line Electrodec) Ring-Shaped Electroded) Ring-Shaped Electrode

with second ringe) Star-Shaped Electrodef) Forked Star Electrode

a )

c ) d ) e )e ) f)

b )

5.6.4.2.2 Vertical Land Electrode

If the ground strata near the surface havea high specific resistance, butunderneath, there is a conductive andsufficiently thick stratum at a depth ofseveral tens of meters, the vertical deepelectrode is one possible solution.

Figure 5.6.4-3 shows, as an example,one of the four deep electrodes at Apollo,the southern station of the Cahora BassaHVDC system.

Fig. 5.6.4-2: Plan view of a typical design ofhorizontal land electrodes

Fig. 5.6.4-3: Vertical electrode at Apollo, theSouthern Cahora Bassa HVDC station

Coarse GrainGraphite

ManholeFlexibleConnectionFeed Cable

ConcreteCover

Graphite Rod

CrushedStoneBorehole ø:0.57 m

ConductiveLayer40

m10 m

80 m

Earth Electrode5.6.4

5.6.4.1 Function of the EarthElectrode in the HVDC System

Earth electrodes are an essentialcomponent of the monopolar HVDCtransmission system, since they carrythe operating current on a continuousbasis. They contribute decisively to theprofitability of low-power HVDC systems,since the costs for a second conductor(with half the nominal voltage) aresignificantly higher, even for transmissionover short distances, than the costs forthe earth electrodes.

Earth electrodes are also found in allbipolar HVDC systems and in HVDCmulti-point systems. As in any high-voltage system, the power circuit of theHVDC system requires a reference pointfor the definition of the system voltageas the basis for the insulation coordinationand overvoltage protection. In a bipolarHVDC system, it would conceivably bepossible to connect the station neutralpoint to the ground mat of the HVDCstation to which the line-side star pointsof the converter transformers are alsoconnected. But since the direct currentsin the two poles of the HVDC are neverabsolutely equal, in spite of currentbalancing control, a differential currentflows continuously from the stationneutral point to ground. It is commonpractice to locate the grounding of thestation neutral point at some distance(10 to 50 kilometres) from the HVDCstation by means of special earthelectrodes.

5.6.4.2 Design of Earth Electrodes

Earth electrodes for HVDC systems maybe land, coastal or submarine electrodes.In monopolar HVDC systems, which existalmost exclusively in the form of sub-marine cable transmission systems, thereare fundamental differences betweenthe design of anode and cathodeelectrodes.

5.6.4.2.1 The Horizontal Land Electrode

If a sufficiently large area of flat land withrelatively homogeneous ground charac-teristics is available, the horizontal groundelectrode is the most economical formof a land electrode.

Fill

CrushedStone

Conductor(Iron)Coke Bed(0.5 x 0.5m2)

1.5 – 2.5 m

36

Fig. 5.6.4.-4: Linear submarine electrode(anodic operation)

5.6.4.2.3 Cathodic Submarine Electrodes

The design and construction of thecathodic submarine electrodes of amonopolar HVDC system with sub-marine power transmission cable do notpresent any particular problems. Sincethere is no material corrosion, a coppercable laid on the bottom should theo-retically suffice. The length of the cablemust be designed so that the currentdensity on its surface causes an electricalfield of < 3 V/m in the surrounding water,which is also safe for swimmers anddivers.

5.6.4.2.4 Anodic Submarine Electrodes

Figure 5.6.4-4 shows an example of alinear submarine electrode for anodicoperation. The prefabricated electrodemodules are lowered to the ocean floorand then connected to the feed cable.When the submarine electrodes aredivided into sections which are con-nected to the HVDC station by meansof separate feed cables, the electrodecan be monitored from the land.

5.6.4.2.5 Anodic Coastal Electrode

The conventional design of a coastalelectrode is similar to that of a verticalland electrode. Graphite rods surroundedby a coke bed are installed in boreholeswhich are sunk along the coastline.

The advantage of the coastal electrodesis easy accessibility for inspection,maintenance and regeneration, ifnecessary.

A coastal electrode can also be con-figured in the form of a horizontal landelectrode if the ground has the necessaryconductivity or if the necessary con-ductivity can be achieved by irrigatingthe trench with salt water. In either case,it is assumed that even with a coastalelectrode, the current flow to theopposite electrode takes place almostexclusively through the water.

ElectrodeModule

FeedCable

Cable

Graphite RodCoke BedConcrete Cover

2 – 5 m0.5 – 1 m

0.5 – 1 m

37

38

Control & Protection5.7

5.7.1 General

The control system plays an importantrole in the successful implementation ofa high-voltage DC current transmission.Reliability through redundant and fault-tolerant design, flexibility through choicefrom optional control centres and highdynamic performance were the prere-quisites for the development of our controland protection system. Continuousfeedback during 30 years of operationalexperience and parallel use of similartechnology in related applications yieldedthe sophisticated technology we canoffer today.

Main objectives for the implementationof the HVDC control system are reliableenergy transmission which operateshighly efficient and flexible energy flowthat responds to sudden changes indemand thus contributing to networkstability.

Fig. 5.7.1-1:HVDC control hierarchy, one station(bipolar HVDC transmission scheme,redundancy is not shown in this figure)

Operator Control LevelWorkstations

Local Area Network

Control LevelDigitalControls

Field Bus

ProcessLevel

Optical Fibre

GPS

MasterClock

AC/DCI&M

RemoteControl Interface

Pole 1Control& VBE

Pole 2Control& VBE

InterstationTelecom

StationControl

DC YardMeasured

ValuesMeasured

Values DC YardTransformer 2Valves

Telecontr. Telecontr.

MeasuredValues AC Filter AC Feeder Transformer 1

The Control and Protection System isbased on Siemens standard products.This makes sure that many years ofexperience and thousands of applicationscontribute to today's control and pro-tection development. At the same timethe system performance has been in-creased substantially by the use of latesttechnology CPUs.

The control is divided into the followinghierarchical levels:• Operator control system• Control and protection systems level• Field level (I/Os, time tagging,

interlocking)

Internet

In the following, functions, tasks andcomponents are described to provide anoverview.

Fig. 5.7.1-2: Remote access connection

5.7.1.1 High Availability

The main design criteria for SiemensHVDC systems is to achieve maximumenergy availability. That as well appliesto the design of the control and pro-tection systems. A single fault of anypiece of equipment in the control andprotection systems may not lead to aloss of power. Therefore, the major controland protection components areconfigured as redundant systems.

5.7.1.2 Self-Testing Features

All control and protection systems areequipped with self-diagnostic featuresthat allow the operator to quickly identifyand replace the defective part to recoverredundancy as soon as possible.

5.7.1.3 Low Maintenance

With today’s digital systems there is norequirement for routine maintenance.However, should it be necessary toreplace single modules, the design issuch, that there is no operational impacton the HVDC system. This is achievedby designing all major components asredundant systems, where one channelcan be switched off without impact onthe other channel.

5.7.1.4 Best Support – Remote Access

As an optional feature, the control systemcan be accessed from the remote viathe Internet. This allows plant monitoring,detection of faults and from remotelocations. To ensure the data security, aVPN (Virtual Private Network) encryptedconnection is used. Furthermore, apassword protected access makes surethat only authorized personnel haveaccess.

With the use of a standard web browser,main diagnosis data can be monitored.Expert access to the control componentsis also possible. This remote accessfeature provides excellent support forthe commissioning and maintenancepersonell by our design engineers.

5.7.1.5 Modular Design

The control and protection systems usemultiprocessor hardware. This meansthat the computing capacity can bescaled according to the needs.

Therefore, the most economic solutioncan be found in the first place, whileadditional computing capacity can beadded at any time later, should it beneeded.

5.7.1.6 Communication Interfaces

The control and protection systems aswell as the operator control systemcommunicate via either Ethernet orProfibus protocols. For remote controlinterface, a number of major protocolsis available. Other protocols can beimplemented as an option.

39

Internet

Fig. 5.7.1-3: Operator workstation, typicalscreen layout for a bipolar HVDC systemoverview

Fig. 5.7.1-4: Sequence of events recording(SER), screen layout for display of SERinformation

40

Control & Protection5.7

5.7.2 Control Components

5.7.2.1 Operator Control System

The tasks of a modern operations andmonitoring system within the HVDCcontrol system include the following:• Status information: clear and

structured overview of the system forthe operator

• Operator guidance: prevent maloperation, explain conditions

• Monitoring of the entire installation and auxiliary equipment

• Graphic display providing structural overview of the entire system

• Support of operating personnel throughintegrated operator guidance

13 : 42 : 17

BP Pramp = 500 MW/minBP Pset = 1800 MWBP Pact = 1800 MW

Iramp = 500 A/minIset = 1800 AIact = 1800 APmax = 1300 MWPact = 900 MW

Iramp = 500 A/minIset = 1800 AIact = 1800 APmax = 1300 MWPact = 900 MW

Iramp = 500 A/minIset = 1800 AIact = 1800 APmax = 1300 MWPact = 900 MW

Iramp = 500 A/minIset = 1800 AIact = 1800 APmax = 1300 MWPact = 900 MW

Ud = 500 kV

Ud = 500 kV

Ud = 470 kV

Ud = 470 kV

γ = 18

γ = 18

α = 15

α = 15

DEBLOCKED

DEBLOCKED

P = const.

P = const.

Control LocationDC-SequencesDC - WS

Control LocationDC-Power/Current

DC - WS

Control Level

STATIONSystem Configuration

BIPOLAR

Runbacks

ENABLED

ENABLED

Runups

Power SwingStabilization

ENABLEDPower SwingDamping

ENABLED

DEBLOCKED

DEBLOCKEDDateTime

System OverviewBipole Control

03 - 16 - 94

I = 0 A I = 0 A

Energy TransferMode Pole 1

Energy TransferMode Pole 2

Imax = 1980 A Imax = 1980 A

Imax = 1980 A Imax = 1980 A

Station BStation A

F1 Control LevelF4 Power DirectionF7 Bipole BlockF10 Power Ramp Stop

F2 Energy Transfer ModeF5 Current Setting ValuesF8 Modulation Enable/DisableF11 Operator Notes

F3 BP Power Setting ValuesF6 Reduced VoltageF9 Emergency StopF12 MENU

• Troubleshooting: support operator with clear messages that allow quickresume of operation

• Display and sorting of events time tagged via global positioning system (GPS), automatic generation of process data reports

• Display and storage of messages concerning events, alarms and disturbances

DATE TIME NR. DESCRIPTIONDEVICEGROUP EVENT STATUS

EVENT SELECTION SORTING

DISPLAY SELECTION DISPLAY RANGE ACKNOWLEDGEONLINE ACTUAL DISPLAY

0 EVENTSTIMETRIP+WARN+STATEVENTS JUMPERED

FROM 1100 TO 1118 OF 1118 EVENTS

16:38:42:4220396 12 S16:38:42:423 12

SS

12 S1212 S12

SS

040016:38:42:7270949 27 =JS11

=XJ12=XJ12

16:38:42:8620946 27 =XS1116:38:42:9410949 27 =XS11

16:39:05:0820397 =XJ12039616:39:05:128

16:39:42:060 =XJ1216:39:42:061039716:39:42:2270806 21 W =XX1116:39:43:5590806 2116:39:52:4700981 29 =JS1116:39:52:5560978 29 =XS1116:39:52:5610301 09

S16:39:52:561 09 =V050300

16:39:52:9790781 20 =XJ1116:39:52:5960981 29 =js11

=v05

=xx11

=xj12

=xj12

s

s

w

s

s

s

s

olc: status standby

CA01Q0 BREAKER ON

OLC: STATUS BLOCKED

olc: status blockedTELE: H1 MESSAGE TRANS FAULTtele: h1 message trans fault

clc: converter blockedCLC: CONVERTER DEBLOCKED

OLC: Q-BAND UNDERFLOW

16:38:43:5790769 20 =XJ11

-

-

+

+

+

CA01Q0 BREAKER COMMAND ONON

cac01q0 breaker command off

CA03Q0 BREAKER COMMAND ONCA03Q0 BREAKER ON

ca03q0 breaker command offw

OLC: START SEQUENCE RUN

olc: start sequence stopOLC: START SEQUENCE RUN

-

+

-05-11-9305-11-9305-11-9305-11-9305-11-9305-11-9305-11-9305-11-9305-11-9305-11-93

05-11-9305-11-9305-11-9305-.11-9305-11-9305-11-9305-11-93

16:40:12EVENT RECORDING AND ALARM SYSTEM05-03-94

0396

S16:39:53:0210398 =XJ1212 OLC: STATUS DEBLOCKED +05-11-93=XJ12=XJ12

s120396olc: q-band underflows0781

16:39:53:06416:39:55:703 20 -

olc: start sequence stop05-11-9305-11-93

05-11-93

W OLC: C-BANK OFF BLOCKED +

NUMBER OF ALARMS: 1

HVDC Transmission System

F1 PAGE UPF4 EVENT SELECTIONF7 EVENT ARCHIVEF10 INITIATE GEN STATUS

F2 PAGE DOWNF5 SORT SELECTIONF8 JUMPER EVENTF11 RETURN TO PRESEN T CONDITIONS

F3 ACKNOWLEDGEF6 PRINTOUT SELECTIONF9 REMOVE EVENTF12MENU

Pole 1

Pole 2

Pole 1

Pole 2

HVDCTrend Display

Date: 18.08.1999 13:11:56

F1: DISPLAYCONFIGURATION

F4: SELECTF7:F10: DEFAULT

F2: PARAMETER SETCONFIGURATION

F5: ARCHIVEF8: FREEZEF11: ONLINE

F3: PRINT

F6: ARCHIVEF9: FREEZEF12: MENU

41

Fig. 5.7.1-5: Trend system,example for differenttrend displays

Pole 1

Pole 2

• Analysis of operating mode basedon user-defined and archive data(trend system)

Fig. 5.7.1-6: Principle of the hybrid opticalmeasuring scheme

5.7.2.2 Control and ProtectionSystems Level

The major tasks in this level are:• Measuring of actual values• Transmission of required amount of

power• Protection of equipment and personnel

Measuring of Actual Values

To determine the actual values, the hybridoptical DC measuring system is used.This system measures the voltage dropon a shunt in the high-voltage busbar,converts this voltage into a light telegramand transfers it to the controls via fibreoptics.

The scheme is designed completelyredundant, therefore loss of a signal doesnot lead to an impact on powertransmission. This measuring principlegreatly contributes to an increasedavailability of the control and protectionscheme.

The even more obvious advantages ofsuch a scheme:• Reduced weight (100 kg)• Passive system (linear response)• Reduced EMC (due to fibre optics)• Integrated harmonic measurement

(Rogowsky coil) for use in active filtersor harmonic monitoring schemes.

redundant channelduplicated sensor head

Id

Shunt

Analog / Digital

Digital / Optical

Electrical Energy

Optical Energy

Sensor Head

Sensor Head Box at high voltage level Measuring at control systems

Signal fibre

Power fibre

Fibre optic cable

Power supply

Optical Energy

Electrical Energy

Optical

Digital

Digital control/protectionsystem

Control & Protection5.7

Protection of Equipment and Personnel

The DC protection system has the taskto protect equipment and personnelalso on a per pole basis.The protectionsystems can be divided into two areas.The HVDC-related protection functionsare referred to as DC protection. Theseinclude the classic DC protection functio-nality consisting of converter protection,DC busbar protection, DC filter protec-tion, electrode line protection and DCline protection as well as the AC filterprotection and converter transformerprotection.

The AC protection scheme consistsmainly of the AC busbar, the AC lineand the AC grid transformer protection.

The task of the protective equipment isto prevent damage of individual compo-nents caused by faults or overstresses.

Each protection zone is covered by atleast two independent protective units– the primary protective unit and thesecondary (or back-up) protective unit.

Fig. 5.7.1-7: Redundant pole control systemstructure (for one 12-pulse group)

Some control actions are initiated by theprotection scheme via signals to thecontrol systems. With this monitoringsystem, a false trip due to a hardwarefault of the protection hardware itself isalmost impossible.

The required functions of the variousprotective relays are executed reliablyfor all operating conditions. The selectedprotective systems ensure that all possi-ble faults are detected, selectively clearedand annunciated. The continuously activeself-monitoring system takes care thatdefects of the DC-protection Hardwarewill be detected.

Power Transmission

The pole control system is responsiblefor firing the thyristor valves such thatthe requested power is transmitted.The pole controls on each side of thetransmission link therefore have to fulfildifferent tasks. The pole control systemon the rectifier side mainly controls thecurrent such that the requested poweris achieved, whereas the pole controlsystem on the inverter side controls theDC voltage such that minimum lossesare achieved.

The pole control is configured asredundant scheme, so that a failure inone channel has no impact on powertransmission.

This channel will be repaired while theother channel stays in operation. Thenthe repaired channel will resume opera-tion. In bipolar schemes, a redundantpole control system is – of course –assigned to each pole. Failures in onepole will not have any impact on theremaining pole.

Fig. 5.7.1-8: Protection zones, one pole/onestation

Redundant Local Area Network

Red. Field Bus

PoleControl

ConverterControl

ValveBaseElectronic

PoleControl

ConverterControl

ValveBaseElectronic

Red.SystemSelectionControl

Fibre Optic

12-Pulse Group

ThyristorElectronic

Sys

tem

1

Sys

tem

2

42

Sys

tem

1

Sys

tem

2

1 AC-Busbar Protection 2 AC-Line Protection 3 AC-Filter Protection 4 Converter Transformer Protection 5 Converter Protection 6 DC-Busbar Protection 7 DC-Filter Protection 8 Electrode Line Protection 9 DC-Line Protection

1

2

3

45

67

8

9

43

5.7.4.2 Dynamic Performance Test

The offline simulation with EMTDC isalready an extremely accurate forecastof the real system behaviour. To verifythe findings and optimize the controllersettings, the control and protectionsystems are in addition tested duringthe dynamic performance test against areal-time simulator. During that phase,the customer may witness these per-formance tests of the final control andprotection software.

Fig. 5.7.1-9: Real-time simulator

All protective equipment in the HVDCconverter station is realized either withthe digital multi-microprocessor systemor with digital Siemens standard protec-tive relays. ”The DC Protection is of afully redundant design. Additionally bothprotection systems incorporate main-and back-up protection functions usingdifferent principles. The AC Protectionconsists of a main and back-up systemusing different principles. The differentProtection Systems are using differentmeasuring devices and power supplies.”

5.7.3 Control Aspects

5.7.3.1 Redundancy

All control and protection systems thatcontribute to the energy availability areconfigured in a redundant setup. Thisallows to cover any single fault in thecontrol and protection equipment withoutloss of power.

5.7.3.2 Operator Training

For Siemens HVDC application, an oper-ator training simulator is available as anoption. The simulator allows the operatorto work with the same hardware andsoftware than in the real process. Thissimulator consists of the original operatorworkstation plus a simulator PC, thenruns the HVDC process and feeds therelevant data to the workstations.

5.7.4 Testing and Quality Assurance

The design process has a number ofdefined review steps that allow to verifythe functionality and performance of thecontrols before they are delivered on site.

Already along with the tender, the useof accurate simulation tools allows toanswer specific performance issues thatare vital to the customer’s grid.

5.7.4.1 Offline Simulation EMTDC

Siemens uses a simulation tool thatincludes a complete control library,with all details of control and protectionfunctionality modelled in such detail thatforecast of real system behaviour is 100%reliable. Therefore it is possible to opti-mize the application to find the besteconomic solution while providing theoptimum performance.

Fig. 5.7.1-11: Example for a functionalperformance test setup

5.7.4.3 Functional Performance Test

In the functional performance test, thedelivered control and protection hard-ware is installed and tested against areal-time simulator. The purpose of theFPT is the proper signal exchangebetween the various control componentsas well as the verification of the specifiedcontrol sequences. This allows optimizedcommissioning time. Furthermore,customer personnel can participate inthis test for operator training and tobecome familiar with the control system.

5.7.4.4 On-Site Tests

On-site tests are basically divided intotest steps regarding the related station(station A, station B) and into the teststeps related to the whole HVDC system(see figure 5.7.1-10).

At the precommissioning stage, the basework for commissioning the controlsystem and protection system is re-quired. The main task is preparation andindividual testing of any single system.

44

Fig. 5.7.1-10: The main steps for the HVDCcontrol and protection versus the time startingfrom the contract award up to the on-sitetests

This is required to assure the systemsbeing free of transportation damage.The next station-related tests are thesubsystem tests. Subsystems consistof equipment items which are groupedtogether according to commonfunctions like AC filter banks or thyristorvalve systems. The main task is testingthe proper function of interconnectedsystems before switching on highvoltage. Following this, station testsunder high voltage but no energytransfer will take place. Finally, systemand acceptance tests with severaloperating points of energy transfer willbe used for final fine tuning andverification of system performance.

BidStage Contract Design Spec.

HardwareDesign

SoftwareDesign

EMTDCC&P Study

CubicleManufact.

Dynamic Perf.Test (TNA)

FunctionalPerf. Tests

On-SiteTests Operation

45

System Studies, Digital Models, Design Specifications 6

6.1 System Studies

During the planning stage of a HVDC project preliminarystudies are carried out in order to establish the basicdesign of the whole HVDC transmission project. Thisincludes the co-ordination of all relevant technical partsof the transmission system like HVDC converters, ACand DC overhead lines as well as the submarine cableif applicable.All specified requirements will be taken into accountand are the basis for the preliminary design of the HVDCtransmission link. In addition, special attention is paidto improve the stability of both connected AC systems.Several additional control functions like power modulation,frequency control, AC voltage limiter can be included inorder to provide excellent dynamic behaviour and toassist the AC systems if the studies show it necessary.Sub-synchronous oscillation will be avoided by specialcontrol functions if required. All the AC system conditionsand the environmental conditions as given in the relevantdocuments will be considered in the design calculations.The final design of the HVDC transmission systemincluding the operation characteristics will be definedduring the detailed system studies. All necessary studiesare carried out to confirm the appropriate performancerequirements and ratings of all the equipment.Due consideration is given to the interaction with theAC systems on both sides, the generation of reactivepower, system frequency variations, overvoltages, shortcircuit levels and system inertia during all systemconfigurations.

Typically the following studies are carried out:

a) Main Circuit Parametersb) Power Circuit Arrangementsc) Thermal Rating of Key Equipmentd) Reactive Power Managemente) Temporary Overvoltages and Ferro-resonance

Overvoltagesf) Overvoltage Protection and Insulation Coordinationg) Transient Current Requirementsh) AC Filter Performance and Ratingi) DC Filter Performance and Ratingj) AC Breaker and DC High-Speed Switch Requirementsk) Electromagnetic Interferencel) Reliability and Availability calculationsm) Loss Calculationn) Subsynchronous Resonanceo) Load Flow, Stability and Interaction between different

HVDC Systemsp) Audible Noise

6.2 Digital Models

Digital models of HVDC system can be developedaccording to the specified requirements. Typically adigital model of dc system is needed for a specific loadflow and stability simulation program, while anotherdigital model is required for simulation in a typicalelectromagnetic transients program such as EMTDC.The functionality and settings of HVDC control andprotection system shall be represented in a propermanner in such models, which allow suitable simulationof steady state and transient behavior of HVDC systemin the corresponding digital programs. Digital modelsconsistent with the actual dc control and protectionsystem are beneficial both for the operation of the HVDCscheme and for the network studies including DC link.Typically such models can be developed on request inthe detailed project design stage when all major designworks of control and protection functions are completed.

6.3 Control and Protection Design Specifications

Design Specifications are written for the control,protection and communication hardware and software.The control panels are then designed, manufacturedinspected and tested in accordance to the designspecification. The software for the control and protectionis also written in accordance to the design specification.It is tested using real time simulators in the dynamicperformance test and functional performance test.

Specifications for the topics below are typically written:

a) General Control and Protectionb) Interface Systemsc) Station Controld) Diagnosis Systemse) Pole Controlf) HVDC Protectiong) AC Protectionh) Metering and Measuringi) Operator Controlj) Communication

46

Project Management7

Project ManagementTechnicalProject Manager

CommercialProject Manager

ProcurementQualityAssurance

Basic Design Station Design

MainComponents

CivilEngineering

CivilConstruction

Installation

Control &Protection

EngineeringLogistic

DocumentationTraining

Commercial

Financing

TransientNetworkAnalyser

FunctionalPerformance

Test

Commissioning

CommunicationSystems

Activity Time

Award of ContractEngineering/System StudiesManufacturingTransportationCivil Works & BuildingsErection & PrecommissioningStation TestsSystem TestsCommercial Operation

7.1.3 Risk reduction

Any risks that could arise due to incorrectdeadlines, unclear technical concepts orexcessive costs shall be recognised earlyenough by a monitoring system so thatcounter measures can be taken. Thisincreases contract quality and createsthe basis for clear design criteria.

7.1.4 Progress Report

Periodically meetings with subcon-tractors, in house control working teamsand customer are recorded in progressreports which form an integral part ofthe quality insurance system.

7.1.5 Scheduling

The hierarchically structured bar-chartschedule is a high-level control tool inproject management. The clear structureof sequential processes and parallelactivities is crucial for execution of a24 to 36 month duration, according tothe project requirements.

Deadlines for project decisions –especially those of the critical path – caneasily be identified enabling the projectmanager to make up-to date pre-esti-mates and initiate suitable measures indue time.

Fig.7-1: Project organisation plan

Fig.7-2: Structured bar-chart timeschedule

7.1 Project Management in HVDCProjects

The success and functional completionof large projects depends on the struc-turing of the project team in accordanceto the related work and manpowercoordination. Periodically updates andadaptation of design guarantee theexecution of the project with constanthigh quality within the target time frame.Throughout all production, workingprocess and on-site activities, health,safety and environmental protection(HSE) measures s well as application ofcommonly agreed quality standards suchas DIN EN ISO 9001 are of primeimportance to Siemens.

7.1.1 Division Responsibilities

The overall project is divided andorganised according to design activitiesand technical component groups. Thesefeatures make it possible to define clearfunction packages which are toa great extent homogeneous withinthemselves and can be processed withminimised interfaces.

7.1.2 Transparency

A clear process structure plan (PSP)standardised for HVDC projects makesthe project contents and sequencestransparent in their commercial andtechnical aspects. Associations andinteractions are clarified according toprocedure of work.

47

47

Siemens AGPower Transmissionand DistributionHigh Voltage DivisionPostfach 32 2091050 ErlangenGermany

www.siemens.com/hvdc

Subject to change without prior notice

Order No. E50001-U131-A92-V2-7600Printed in GermanyDispo-Stelle 3000061D7085 TV/EV 02032.