01460248

2004 Intematlonal Conference on Power System Technology - POWERCON 2004 Slngapore, 21-24 November 2004

Studies for the Integration of a TCSC in a Transmission System

Lutz Kirschner, Gerhard H. Thumm

A b s t ~ c b Series Capacitors are widely used in long distance transmission system to improve stability of power transmission. Thyristor- controlled series capacitors provide additional benefits for the transmission system compared to fixed series capacitors. During a project to realize a series capacitor several studies are required, starting with planning studies, studies that investigate pmible interaction with the surrounding system, and finally the capacitor component design. Different parties are involved in these studies, as illformation is required from the transmission system on one side, but also from the series capacitor and its components. This paper describes the studies and and shows typical results.

Index Terns--Power t " i ss iOn, series compensation, thyristor-controlled series compensation (TCSC), stability, FACTS

I. INTRODUCTION HE thyristor-controlled series capacitor (TCSC) is one system out of the family of FACTS systems, which has proven its bcnefits for transmission systems in several

installations [I]. In addition to a fixed series capacitor (FSC) the TCSC is able to vary the effective capacitor impedance within a very short time, thus increasing stability not only in steady state operation and during first swing. The TCSC is also able to actively clamp power oscillations in the transmission system by variation of the impedance between the generators involved. I n longitudinal series compensated transmission systems a specific interaction between the series capacitor and thermal generators came up, the sub-synchronous resonance (SSR). The TCSC is able to mitigate SSR just by operation with con- stant impedance under phase angle control, active counter- action to damp SSR is possible as well. These additional bene- fits justify the higher costs for a TCSC in comparison to an FSC installation.

The first TCSC is in commercial operation since more than ten years in USA [2]. TCSC installations in Brazil have demonstrated their capability to stabilize a transmission system With a length of more than 1000 km, which could not be oper- ated safe and stable without series compensation 131. Another TCSC installation is in operation since one year in China, within a project to transfer electric power over a distance of more than 1000 km on parallel AC- and DC-lines [4]. Also in

T

other countries with long distance power transmissions TCSCs are in commercial operation or under construction.

11. SYSTEMSTUDIES

A series compensation project starts with a number of sys- tem studies that show the benefit of the series capacitor for the transmission system and describe the overall characteristic. In a second stage the interaction with the surrounding system has to be investigated. During a project the results of studies can be compared with tests involving the actual control system and a digital system representation [ 5 ] , and eventually with on-site measurements to demonstrate the effect of the TCSC for the transmission system.

A. Loudflow Studies LoadfIow situations in a transmission system vary over a

wide range due to various parameters. Seasonal effects Like rain and temperature on one side have a big influence on loads as well as on generation capacity. On the other side contractual conditions can end up in restrictions and basic conditions for power flow. Load requirements vary with daily and weekly CYC~ES. Load forecast is a big challenge, with a time span From minutes until planning studies reaching over several years, resulting in investment in systems, when bottlenecks or eco- nomic advantages have been identified. Under all the situa- tions which are to be investigated, voltages and cutrents in the system shall not exceed limits as given from components and systems like hnsmission lines, eansformers, generators, etc. Series compensation is a system which can help in critical situations, SO that restrictions in power "fer can be avoided.

3. Stability Study and Power Oscillation Study Series capacitors in general are able to reduce the risk of

power oscillations by reducing the impedance between generators or networks involved in the oscillation. This effect will be investigated in a stability study. Additionally a TCSC can damp power oscillations by varying its impedance. This effect will be demonstrated in a subsequent power oscillation damping study, where the operating range of the TCSC and its dynamic behavior is incorporated.

From these studies a typical diagram for a TCSC can be drawn, showing the TCSC operating range as base for TCSC design, sec Fig. 1.

Lutz Kirschner, MIEEE, and Gerhard H. Thumm are with SIEMENS AG, Erlmgen, G m n y , ITD H166.91050 Erlangen, PO Box 3220. Paul Gossen Su. 100. Gemmy (e-nmil lutz.kirschnerbsietiiens.com and [email protected]).

0-7803-8610-8/041$20.00 0 2004 IEEE 1540

c0rtJ)nin 10s & I \ \ \

p g i line MBt ,-

10s

Fig. 1. TCSC operating range as function of line curreni

Generally a series capacitor is designed for overloads. In the relevant IEC and IEEE standards [6, 71 such overload cycles are proposed, however each system may require its spe- cific overload cycles depending on the system characteristic during contingency situations. Accordingly overloads in the range of minutes to hours are specified to cover stresses from overcurrents. A short time overload for 10 s for example is feasible to cover stresses from power swings, as they may occur immediately after fault clearing or during power oscilla- tions.

Typically a stability study investigates the most severe system faults based on the performance of line protection, also including scenarios with component failures, which are very rare to occur. In general it is required that the transmission system remains stable in such situations. The most severe dis- turbance of active power unbalance can be expected from three-phase faults with a duration of 100 ms typically. Some- times a longer fault duration is simulated raking into account for example the effect of a stuck breaker, non-successful auto- reclosure or protection failures.

5 T i p dYYl NWln TUCURUI, withmt W D 0

8 -5001' I I I 0 10 20 30 40 50 60 70 80

Tim (3

M p d ~ M W i n T U C U I I U I . v i t b Z P O D I l ~ z a d S . M m ~

500

w 0 10 20 30 40 50 60 70 80 Timr is)

Fig. 2. Exanplc of a power oscillation, damped by TCSCs

Subject of the power oscillation damping study is to find an optimum strategy to damp oscillations with the TCSC, as they can occur after defined fault scenarios. This includes the con- trol strategy, the selection of an input Signal, and the size of the TCSC.

As an example Figure 2 shows the transferred power through the transmission line in the North-South-Interconnec- tion in Brazil. After tripping of a generator a power oscillation arises. Without any action to damp the oscillation the system would become instable. With the TCSCs with the power oscillation damping (POD) control active the oscillation is damped within a few cycles.

Experience shows that in most cases the active power flowing through the series compensated line i s the most suit- able signal to detect the power oscillation. It can be caIculated from line current and voltage at the series capacitor location. Both are local signals. The transmission of signals for example from a generator normally is not safe enough to rely on during contingency situations.

m. bJlTRAClloNSTUDIES Interaction between a TCSC and the surrounding system occurs in several ways. The series capacitor forms together with the inductive line impedance a resonant circuit, which can interact with other resonant circuits or oscillating systems. At line opening the line breaker sees a voltage between its termi- nals that can be influenced by the series capacitor. The TCSC generally generates harmonics. These harmonics may not exceed given limits. The TCSC design and the selection of the steady state operating point have influence on the harmonics in the system. A study is very difficult, as the impedance area of the system on both sides of the TCSC must be taken into account including contingency situations. The harmonic im- pedance area also must be valid over a certain range of har- monic frequencies. The calculations have to be verified by harmonic measurements during the commissioning period.

A. SSR study The presence of series capacitors in a power system may

cause sub-synchronous resonance (SSR). The dangerous im- pact of this phenomenon has first been noticed at the damage of two generator shafts at the Mojave Power Station (Califor- nia, USA) in 1970 and 1971. Since then, considerable effort has been made to analyze the phenomenon and seek ways to prevent damages in the future.

In a power system, conditions for sub-synchronous reso- nance may arise from the presence of a series capacitor in an otherwise inductive electrical circuit. A system fault or sudden load change may then excite transient power oscillations at sub-synchronous frequencies. The frequencies depend on the degree of line compensation and typically range between 10 and 45 Hz. If the frequency corresponds to an undamped natu- ral torsional frequency of a generator shaft, the electrical power system interacts with the shaft and causes sub-synchro- nous resonance.

1541

Fig. 3. Multi-mass-sHng oscillator of a turbine-generator unit Ji moment of inenia k. torsion spring constant Dlk speed-prop~rtional frictional damping D,j" hysteresis damping, proporlional to the turning speed

differential @, torqueangle

0 , -

LO

The shaft structure has a number of inherent resonance fre- quencies, so-called natural frequencies or eigenvalues. Each eigenvalue is associated with an eigenvector that indicates the condition, i.e. speed and angle displacement, at each location along the shaft, necessary in order to excite the natural fre- quency. The local displacement is typically plotted in a mode shape diagram. For example, the shaft system has a natural frequency at a cc:rtain frequency, at which the low pressure section LP2 is in complete phase opposition to the other sec- tions. Once excited with this frequency it would swing against the rest of the shaft sections, This example demonstrates that, if the frequency flwuc of the sub-synchronous electrical torque is neac to one of the natural shaft frequencies fmj, and if its component in phase with the rotor sped deviation exceeds the inherent damping torque of the rotating system, sub-synchro- nous resonance ciccurs. The torque oscillations then rapidly increase and may cause severe damage to the shaft.

Studying SSR interaction in a series compensated trans- mission system can be split in two steps. The first step is a frequency scanning study, in which the frequency-dependent impedance of the electrical system is calculated for various situations with varying loads and also for a lot of contingency cases. The impedance of interest is when looking from the investigated generator into the system, where all generators are represented by their subtransient impedance in the direct axis, xd".

Fig. 4 shows an example of a frequency dependent impe- dance calculation. In this example a series resonance condition occurs at 19 Hz.

Note that this series resonance is a resonance of only the elech-ical power network. If the machine shaft also had a natu- ral mode at 41 Hz (60 Hz minus 19 Wz) the resonance could lead to SSR and cause damage to the machine shaft. This is mostly determined by the damping characteristic of the oscil- lating mode. However since the overall damping depends on both the damping of the electrical network and the damping of the shaft system, details of the machine behavior at the reso- nance frequency can only be determined in a more detailed SSR analysis study. The purpose of the frequency scans in this

I I 1 1 0 1 1 1 I I I I I I l l I I I 1 1 1 1 1 1 I 1 I 1 1 l i 1 1 - --e - I - + - 1 - + k + + - - - - 4- - - + - + - t + - L - + I 1 I I I I I I I 1 I 1 l l l l I I I I l l l I l I I I I l l 1 1

- - - - -

- - - - - - l - - - ~ - , - ~ - l - ~ r r ~ - - - - ~ - - - ~ - ~ ~ r ~ - l -

The network impedance generally varies with system load- ing and with operating conditions other than normal operation, e.g. contingency situations. Therefore the frequency scans have to be carried out for each inspected generator at various system conditions, which are then presented in such a diagram.

The second step of an SSR study is the so called time domain analysis. A program suited for computing electrc- magnetic transients may be used to simulate the transient time response at a variety of system events. Time-domain simula- tion uses full three-phase electrical representation of network and generators, and permits detailed modeling of the multi- mass shaft systems. Detailed representation of nonlinear effects is also possible. SSR can be identified by observing the time response of the torques at a particular shaft system, If they persist or grow in time, then the system almost surely has an SSR problem. Due to the amount and detail of data, time domain simulation i s mostly used to verify the existence of an SSR problem after identification with one of the other methods.

E. TRVstudy The TRV (transient voltage recovery) study deals with

voltage stresses of breakers immediately after opening. Of special interest is this voltage at fault clearing in long distance transmission systems together with series compensated trans- mission lines. When a fault occurs on a series compensated tmnsmission line, the protection of the series capacitor is allowed to operate, as soon as components tend to become

ovcrstrcssed. As in most cases the bypass dcvice short-circuits thc series capacitor with a time delay of about I ms, this hap- pens clearly beforc the line breaker opens. Ncvcrthcless an influencc can be seen, as the pre-fault stcady stale current in a series compensatcd tmnsmission system will bc higher than without series capacitors.

The voltage across the breaker build-up after current inter- ruption is tcrmcd recovery voltagc. Due to the circuits induc- tive and capacitive Characteristics on both sides of the breaker, the recovery voltage generally has high frequency components superimposcd to its fundamental sinusoidal shape, which typi- cally decay within a few cyclcs. Hence, the attribute transicnt was added to qualify the lerm recovery vollagc. High frcquency componcnts ace particalarly responsible for exceeding the dielectric capabilitics in the contact gap.

The objcctive of TRV analysis therefore is to determine the fastest initial build-up of the voltagc after current intcrruption. While many textbook cxamples of TRV analysis rely on a sirnplificd RLC equivalent circuit of the systcm in order to determine thc highest f1,equcncy components in the initial voltagc rise, a high-volhgc scries-compcnsatcd transinission system shows a much higher dcgree of complexity than could possibly be represented by means of an RLC circuil. Thc following is a list of influcnces that affect TRV levels in a high-voltagc series-compensatcd transmission systcm:

Transmission lines: a high-voltage transmission system typically involves long transmission lines that cxtcnd over hundreds of kilomcters. Due to finitc wave lravcling times, distinct voltagc shapes such as TRV may travel forth and back a linc, producing a harmonic oscillation on its own (e.g. trapped line charge). Series compensaiion: as already tlescribcd, a series capaci- tor is equipped with protective deviccs such as nietal-oxidc varistor (MOV), a triggered spark-gap, and a bypass- switch. Since TRV is a process or fault clearing, thcsc protective dcvices have already responded to thc fault situation. Thcir rcsponse has a trcmcndous impact on the TRV Ievels: c.g. a bypassed serics capacitor docs not inter- act with thc harmonics in thc circuit. Fault sequence: thc instant at which the fault occurrcd, at which voltage and cnrrcnts, and thc location of the Fault in- flncnce some of the pawmcters, that arc contributing to the currents across thc linc breakers, and therefore havc an eflect on TRV. 3 y the samc atxument, it is important to consider the mlt lypc, i.e. single-phase-to-ground, three- phasc, phase-to-phasc with 01 without ground. Obviously, there arc multiple possibilities of fault tlevclopmcnt, e.g. a single-phase fault that dcvelops into a three-phase fauli. Another inflocnce i s how long a fault lasls bcforc the cir- cuit breakers start opening their contacts.

+ Fault clearing: note that it matters if the faulted line is interrupted in only one phasc or in all three phascs. The opcning of the line breakers at one substation may not start at cxactly the samc timc as at the othcr tcriniiinl. There inay bc x rcw milliseconds of a diffcrence duc to different pro- tection equipmcnt, comniunicatioti signals or diifcicnt types of brcakcrs.

Arching: most faults on a high voltage transmission systcm are not solid mctal coiineclions betwecn phascs but result i n arcs. An arc dissipates electric energy and, therefore, adds rcsistive damping to high-frequency oscillations. Correct rcprcsentation of thc arc bears an efrect on thc TRV levels that can bc experienced by thc circuit breakers. By the same argument, the arc produced within the breakcr chambers also contribute to the damping.

9 Powcr system: during the fault, the entire power systcm in the vicinity is affcctcd and iti a state of transicnt oscillation so that, when thc faulted line is isolatcd, thc oscillations continue until a steady post-fault state is gained. Since TRV are voltagcs across the line breakers, the systcm os- cillations also affect the level of TRV. As a consetpencc it is not only important to analyze the affcctcd transmission line with regards to TRV, but also to considcr thc entire power system in the vicinity in all its oscillatory dehil.

Thc number o l cases, which has to bc investigatcd to de- tcrmine worst-case TRV levels for the circuit breakers is immense, particularly sincc the task involves detailcd repre- sentations of stochaslic phcnomena such as the arcs at thc Fault location or in the breaker chamhers.

Evaluation and intcrprctdon of results can bc in a statisti- cal diagram, showing the distribulioii of TRV Icvcls. Besides this the fact, that the allowable lcvcl or TRV is exceeded or nol, is of course of interest. For thesc cases it is necessary to evaluate, under which condilions this TRV lcvcl occurs and how likely such a fanlt is.

A comparison with TRV stresses in situations without series comparison is diflicull, as the load flow situation i s diffcrcnt due to the capacitor. However, as a gciicral rulc it can be stated that TRV is higher in cases with a scrics capacitor. This is especially thc case at system faults far away from the breaker tinder invcstigation. In these cases thc linc together with thc scrics capacitor forms an oscillating circuit. Also when the scrics capacitor is equippcd with a fast bypass devicc acting clearly faster than the line brcaker under investigation, an increase in TRV can bc found. These fast bypass dcvices operate within some milliseconds after a severe fault, whercas the timc to open a line breaker is clearly longer.

C. Line protection study Thc linc protection system has to take into account lhat the

impcdaoce as seen from the substation into lhe linc has an additional component from the serics capacitor, which is ca- pacitive in contrary to the inductive line impcdancc. Especially at a TCSC this impedancc may vary, and no information re- garding the capacitor impcdancc will be sent to the linc pro- tection system. Line protection and series capacitor protection shall bc able to opcratc indcpendent from each olhcr. This mcans for the line protcction, that it must be desigricd for a series compcnsated line,

Gcnct-ally the communication between the two protcction systcms is limited. In rnosl casts thc series caipacitor protection systcm will rcccive a signal, whcn thc line opens, and accord- ingly thc bypass switch of Lhc scrics capacitor will be closcd. On the other sidc, a1 n scvcrc failure of the series capacitor c.g.

1543

when the bypass switch does not close, the line has to be tripped.

IV. COMPONENTDESIGN S m u The component design study, of course, has to be done by

the manufacturer of the series compensation system, as he is responsible for the component design and series capacitor per- formance. On the other side the study is based on system requirements.

Figure 5 show!; the basic single line diagram of a TCSC. The design of the main components includes capacitor, varis- tor, the thyristor-valve and the reactor, and the bypass devices with the associated bypass damping circuit. Generally the essential data for disconnects, earthing switches, CTs etc. are already defined fiom customers side.

I bypass disconnect

) platform disconnect

el--___r____ AII- L___________-____ I L-- --- -L/J

bypass switch

Fig. 5. Principal single h e diagram of a TCSC

The first step of the component design covers steady state stresses. It takes into account continuous operation and opera- tion during overload situations for a limited time, see Fig. 1. In a TCSC installation steady state operation i s defined with a boost factor of 1.1 or 1.2 at nominal current. This defines nominal capacitor voltage. The TCSC also can operate con- tinuously at higher boost factors but with lower line current, so that the capacitor voltage does not exceed the rated value, see the dark area in Fig. 1. The continuous operating range to- gether with overloads defines the voltage rating for the varis- tor. The thyristor-valve and the reactor in series to the thyris- tor-valve must be designed to cover the associated stresses continuously. This also determines the rating of the cooling system for a water cooled thyristor-valve.

The second step of the component design study deals with [ransient stresses. They determine the design of all main com- ponents, for example the maximum capacitor voltage occur- ring at a severe system fault, the maximum voltage across the varistor, the associated current, and the energy the varistor accumulates during a system fault scenario. The protection strategy with the thyristor-valve as fast bypass device depends

on different conditions. Provided the steady state stress for the thyristor-valve allows the additional stress from bypass opera- tion, the thyristor-valve can be used as fast bypass device. Like in the thyristor-protected series capacitor (TPSC) the thyristor- valve can replace the triggered spark gap. When the protection strategy of the TPSC is acceptable, the energy rating of the varistor can be considerably reduced [8].

To find out maximum wansient stresses a number of tram sient stress simulations is necessary to find out the worst case system fault. The sequence during each simulation run is based on the strategy of line protection during and after system fault. The number of system faults for which the components shall be designed determines mainly the varistor rating. Auto-reclo- sure for example is usual in many cases, but this may be different for single-phase faults and multi-phase faults. At single-phase faults it is possible to open the line breakers only in the faulty phase and to reclose after a dead time of typically 1 s. As the most probable system fault is due to lightning strokes, it is most likely to clear the fault. This results in mini- mum disturbance for the remaining system, as the active power transfer continues during the dead time at a lower level. In other systems the faulty line opens in all three phases at single- phase faults. At multi-phase faults the line will be opened in three phases, and again it depends on the protection strategy, whether auto-reclosure is foreseen for multi-phase faults or not.

The series capacitor protection system is able to react in a similar way, i.e. it is possible that at single phase-faults only one phase will be bypassed. Before re-closing the line breaker the series capacitor returns into operation with the risk of a second stress, when the fault is still persistent (non-successful auto-reclosure). In other cases all three phases are bypassed arid they only return to operation after the success of redosure has been proven. All possible combinations between these two extreme cases are possible. In most cases the varistor shall be designed to withstand stresses from to subsequent external or internal faults in one, two, or three-phases. This covers worst case stresses from a fauIt with non-successfd auto-reclosure or from two subsequent faults. Accordingly the bypass damping circuit has to be designed, so that it is able to withstand stresses from two subsequent capacitor discharges from the maximum voltage.

Another result from this study are the parameters for ca- pacitor protection to distinguish between external and internal faults, as only at internal faults the protection system is a1- lowed to bypass the capacitor. Only when the varistor current or the accumulated energy exceeds the maximum value that has been found during the study at any external fault, the pro- tection system interprets a system fault as internal and initiates a bypass action.

Insulation coordination is a part of the component design study, as it is focused On the insulation coordination of all components of the series compensation. The insulation levels between phases and ground are given from the surrounding system or substation. Based on the maximum voltage across the components, which has been found during the transient stress calculation, the insulation of all components, the creepage distances of housings, bushings and post insulators, and the distances between the components can be determined-

1544

V. CONCLUSION The TCSC is a powerful and flexible system that provides

benefits especially for long distance power transmission systems. The project of such an installation requires a number of studies that have to be carried out from the system owners side and from the manufacturer. This requires information from both parties, as they both together have the knowledge to allow a good presentation of transmission system and series capacitor during the studies, and guarantees study results that are as realistic as possible.

VI. REFERENCES [I] Kirschner, Thumm. Design and Experience with Thyristor-Controlled

and Thyristor-Protected Series Capacitors, presented at the IEEE PowerCon 2002, Kunming, China Chrisll, Luetzelberger, and Sad& System Studies and Basic Design for an Advanced Series Compensation Scheme, presented at the Inter- national Conference on Advances in Power System Control, Operation and Management APSCOM. Hong Kong, 1991 Gama, h n i , Gribel, Fraga, Eiras, Ping, Ricardo, Cavalcanti,, and Tenorio, Brazilian Noh-South Interconnection - Designing of Thy- ristw Controlled Series Compensation (TCSC) to Damp Inter-Area Oscillation Mode. presented at Sepope 1998, Brazil Kirschner, Design h p t s of the Chinese 500 kV Thyristor-Controlled Series Compensation Scheme TCSC Tian Guang, presented at the 2 International Conference on Electric Utility Deregulation, ResmcNring and Power Technologies, IEEE DRPT, HongKong 2004

[5 ] Retunano, Claus, Kuhn, Kumar, Lei, Baran, Forsyth. Maguire. Advanced Fully Digital TCSC Real-Xme Simulation, presented at CEPSl2001 Series Capacitors for Power Systems, Part 1: General, Inremcioaa] Standard, IEC 143-1 IEEE Standard for Series Capacitors in Powa Systems, IEEE 824-1995 Kirschner, Bohn, Sadek, l%yistor-Procected Series Capacitors. Part i: Design Aspects, presented at the Ci@ XERLAC Conference 2003, Argentina

121

[31

[4]

163

[7] [ R I

VII. BIOGRAPHIES

Lutz Kiwchner. Senior Project Engineer, MIEEE, received his Diploma in Electrical Engineering 1992 from the University of Aachen, Gemany. He joined SIEMENS conipany in h e HVDC D e m e n t as a system design engineer. He was involved with technical and commercial design of HVDC converter stations. He got his project experiences from the North-American Texas-Welsh Converter, the Chinese TianGuang Converter and the Asian Thailand-Malaysian Converter projects. Since 1995 he is alw, responsible for Fixed and Thyristor Conccolfed Series Capacitor design and was busy in the brazilian FurnadEletronorte series capacitor project as well as the noah- american LexingtonNalley and TCSC Tian Guang project carrying out the basic design system studies. He is workiog on the design of Thyristor Protected Series Capacitor system (TPSC) and FACTS devices. In the FSC Sao Joao do Piaui project he carried out the Final Basic Design Studies com- prising the transient fault calculation and main component ratings. His special fields are time domain digital simulations and system studies. Since 1998 he is a member of IEEE

Dr. Gcrhard H. m, Senior Project Engineer, received his Diploma in Electrical Engineering in 1977 from the T~chaical University of Stuttgart, Germany, and the h.-Jng. degree in 1991. From 1977 to 1982 he worked at the High Voltage Laboratory of the University in Stuttgart. He joined SEMENS AG. the system planning department in 1982, and changed to a group for Reactive Power Compensation in 1985. He was involved with tech- nical design of Static Var Compensators (SVC), and was responsible for the design of seven1 SVCs in Great Bricain, Australia and USA. Since 1995 he is also responsible for the design Fixed and Thyristor Controlled Series Capaci- tors, and was engaged in the Brazilian FurnasiEletronorte series capacitor project hteriigacao I. several fixed series capacitors in South Africa, Chine, and India, and in the system studies for the Tian Guang TCSC project.

1545