It should be noted, too, that only one rectifier is put out of To Mr. Kimbark's question regarding the connection for theoperation during the 0.2 to 0.8 second of blocking time required frequency converter in Japan, I would like to answer that thefor removing the causes of the are back in the rectifier. The motive for choosing the connection rectifier-inverter, is thatother series-connected rectifiers of the station go on operating as a common bypass valve can be used for each rectifier-before, their current passing through the bypass valve of the inverter pair instead of one for each converter. Further, nobackfiring one. The inverter station is arranged so that it higher potential to earth need appear than the voltage of oneautomatically lowers its voltage sufficiently to keep current up. converter, thereby saving in isolation.

Mr. Hauspurg also refers to my statement that tappings on d-c To answer the question, "Is direct current better or worse thanlines will require new thinking regarding control and protection. alternating current as regards contamination of the line in-I regret to have made a premature statement in this respect. sulators?"-direct current is worse.After my paper was written, a more thorough study of the tapping Mr. Holm has calculated the angle of overlap of an inverterproblem was started, and now it appears that the present system station which operates with a margin of commutation of 10of control and protection can, in its main principles, be adapted degrees and consumes an amount of reactive power of 50% toto a multiterminal transmission system. It may even develop 60% of the active power. He asks if the angle of overlap of 17 tothat the d-c power breaker will not be a necessary component, as 21 degrees so computed can be sustained by the ASEA valves.one would first assume. The answer is: Yes. Especially in inverter operation, the

In this connection, I would like to discuss an oral question, steepness of the current drop at the end of the conducting periodwhich I think was never answered, and which has been re- would not be too high with this magnitude of overlap. The usepeated. However, I wish to modify the question a little and say a of static capacitors will tend to reduce the angle of overlap whichfew words about tappings generally, hoping to be able to return may not be desirable in some cases as there is no correspondingwith a new report in the not-too-distant future. When a short reduction in direct voltage regulation.circuit occurs in an a-c system, a certain time is necessary for To Mr. Holm's inquiry as to whether forced commutation wasclearing the fault, during which period power transmission will developed for the purpose of power-factor improvement, I referbe interrupted. It seems to be possible in a d-c transmission to to my article, "Forced Commutation in Mercury-Are Convert-locate the faulty section, block all the rectifiers or reset them into ors," in Technical Achievements of ASEA Research, 1946. Someinverter operation, disconnect the faulty section with high-speed further studies have been made since that date, but forcedisolators, and then deblock the rectifiers again-and do this commutation remains too complicated to be introduced inwithin the same order of time span. present-day HV d-c transmissions. It is my opinion, however,

It will still be true, of course, that converters in the tapping that therein lies a potential future possibility for economic im-points are an additional expense, but from our preliminary esti- provements.mates, the protection cost including the high-speed isolators Mr. Briggs raises the question of parallel and sectionalized linesseems rather insignificant in comparison with the safety and with cross-switching facilities to be used on a d-c transmission inservice provided. much the same way as on a-c transmission. Theoretically, and

I think, on the whole, the difference between alternating and also practically, it should be possible to realize such a system bydirect current in a multiterminal line is not a qualitative one in placing high-speed isolators in the cross-switching stations, some-this respect but rather a quantitative one. With a-c transmis- what similar to what has been suggested for multiterminal d-csion, as the voltage is increased, the cost to the terminal stations transmission. A paper by Lamm, Uhlmann, and Danforswill weigh heavier. Each time we superimpose a network of presented this suggestion to the American Power Conference inhigher voltage over the existing system, we also reckon with 1963, but the case is somewhat different with d-c than with a-cmore sparsely situated transformer stations. If there is motiva- transmission.tion for the use of direct current, this will only be one further First, an a-c line is more vulnerable than a d-c line. Thestep in the same direction, and there will then be still longer former is unusable if only 1 conductor of 3 is out of order, whiledistances between terminals. a d-c line can operate independently on each of its two polesLet us look at the development of the HV d-c method in three with the earth as a spare return.

steps: The first step, in which a number of installations are Second, a d-c line can be overloaded without endangering thenow in use or under construction, is intended to be only a point- stability. It may, therefore, be an alternative solution toto-point operation. The second step, which we are preparing to switching all converters to one pole or line when a fault occurs inenter, contemplates the possibility of arranging one or two the other pole or line.intermediate terminals on a d-c line where this is economically This is presently all I can say in a general way regarding Mr.motivated. The third step, which today we can only vaguely Briggs' interesting point. I think this subject could be bestvisualize, would be to use direct current in something which studied on actual cases comprised of possible long-distance trans-might be called a meshed network. missions.

Operation and Control of HV D-C TransmissionP. G. Engstr6m

Summary: The special requirements on control of inverters for system, it is therefore sufficient to study such a 3-pulse con-HV (high-voltage) d-c transmission have led to a special control verter as shown in Fig. 1. The maximum direct voltage Edsystem, which is described. Co-operation between regulators ateach end of the transmission is necessary in reducing the conse- from the converter in Fig. 1, is buiilt up by the top parts of thequences of disturbances in the converters. Also described are voltages eR, es, and eT in the three phases R, S, and T asthe switching and blocking of valve groups and how to accom- seen in Fig. 2. The instant the anodes pick up currents canplish this without using circuit breakers. The type of protective be delayed by an angle ae by means of grid control as in Fig. 3,system is discussed in a general way.______________________________

Paper 63-83, recommended by the AJEE Transmission and Dis-tribution Committee and approved by the AIEE Technical Opera-

In HV d-c schemes, two or more series-connected converters tions Committee for presentation at the IEEE Winter Generalare used. Each converter has, in principle, six single-anode Meeting, New York, N. Y., January 27-February 1, 1963. Manu-

'. ~~~~script submitted October 30, 1962; made available for printingmercury-arc valves in 2-way 6-pulse connection. However, December 4, 1962.such a converter can be regarded as consisting of two 3-pulse P. G. ENGSTROM is with Allmiinna Svenska Elektriska Aktiebolagetconverters. In a discussion of principles in the grid-control (ASEA), Ludvika, Sweden.

JANUARY 1964 Enqstreim-Operation and Control of HV D-C Transmission 71

AC- NETWORK Fig. 1. Diagram of 3-pulse Fig. 2. Voltage diagram of 3- elconverter pulse converter at et = 1 elec- eR es eT

trical degrees

(At

eR e, e e

L tdL L Ed ed ~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~eR e s eT

negative alue, Fig5.Intid latteeas,thFig. 3. Voltage diagram at ce25 electrical degrees

which gives a reduced direct voltage. By increasing thedelay angle a, the mean value of the direct voltage decreases,passes through zero as in Fig. 4, and it can even attain a ec = V\2E. sin wot(1negative value, Fig. 5. In this latter case, the power flow has The commutating current i, meets an impedance, which canchanged its direction; the power is now fed into the a-c be regarded as a pure inductance L per phase and there-system from the d-c line, the converter acting as an inverter. fore:

Control of Inverters ec=2LdicInverter operation involves special requirements on the dt (2)

control system such as limitation of the delay angle a, whichlimitation varies with the load and the voltages in the a-c Atd mit=W =t1, the commutation must start in order to obtain asystem. During commutation, two valves are conducting .desi marin. Winsrtd tcurrent simultaneously, let us say the valves in phases R and S aas in Fig. 1, and a short-circuit current i, is forced by the com- 't V/EC s cotdt=2Lfo c =d di (3)mutating voltage eC=eS-eR from phase S to phase R. If aninfinitely large smoothing reactor is introduced in the d-c which gives:line, this commutation is finished when ic has reached the / 7r\same value as the outgoing direct current Id, thanks to the uni- V\2EEC sit --2 = V2Ec cos Y - XCJd (4)directional action for current in the valve in phase R.The current Id is transferred in this manner from valve R where the commutation reactance Xc is

to valve S. It is obvious that this normal commutation is Xc = (5)possible only during the interval when ec is positive; that is,between wct= 0 and cot= Y. See Fig. 6. This means the delay In most practical cases, the reactance Xc is constant andangle a must not exceed 180 electrical degrees. In a practical corresponds to the transformner reactance with a compara-case, however, the maximum value must have a certain tively small addition due to the network reactance. Themargin. The commutation must be finished at a given time load (Id) and the commutating emf can vary within a widecorresponding to -y electrical degrees before e, becomes nega- range. The consecutive grid-control system therefore sensestive. these two parameters for each valve. It consists in principle

This margin of commutation y is normally kept at the lowest of three units, one giving a d-c output voltage proportional topossible value in order to reduce the consumption of reactive the direct current Id; one with a d-c output voltage propor-power, which is taken from the a-c network. On the other tional to commutation emf, e., and depending upon the sethand, the valve must have time enough for its deionization margin of commutation y; and the last unit gives an alterna-before the instant cot= 7r, after which a positive voltage from ting voltage proportional to e, but displaced 90 electrical de-anode to cathode is applied across the valve in phase R. In grees after e, as indicated in Fig. 7.addition, the attainable accuracy of control and other de- These three voltages are connected in series and the result-sirable safety margins must be considered, leading to a prac- ing voltage triggers a grid pulse-generator, which gives a firingtical value of 15 to 20 electrical degrees for the margin of pulse to the control grid in the main valve, when the triggeringcommutation y. voltage passes zero; see Fig. 8. The current-sensing im-A syst.em called the consecut.ive control system' has been pedance, which gives the voltage proportional to the direct

developed to maintain the margin of commutation at a con- current, is fed from a transductor, measuring the current in thestant value independent of load, voltage variations, and asym- d-c line, and consists in principle of a resistor. The delaymetric disturbances of the a-c system. Based on an invention angle computer, which gives the two voltages proportional toby Dr. U. Lamm, the general thoughts behind thiseconsecutive the commutating emf, is connected to this latter voltage.control system can be described in connection with a study of This emf is normally obtained from the primary voltage ofcommutation in inverter operation. the converter transformer, with compensation for the voltageThe commutating emf (electromotivre force), ec, is derived drop due to variations in the converter load.

from the voltage difference between phases R and S and can be As this equipment is connected only to one phase, it iswritten: necessary to use a phase-shifting impedance in order to

72 Engstr8m-Operation and Control of HV D-C Transmission JANUARY 1964

e eR es eT Fig. 4. Voltage diagram ata s current of the converter. This control is accomplished in90 electrical degrees such a way that the valves are made to fire at an earlier in-

stant than that determined by the consecutive grid control.Wt \AM< A direct voltage whose polarity cannot be changed, added

to the control voltages and taken from the output of a controlCC amplifier, produces this earlier firing. The control amplifier

is, in principle, an electronic d-c amplifier with built-in sta-bilizing feedbacks. It can be proved that the total amplifi-cation-including this amplifier, other control circuits, andthe converter itself-is constant within the whole operating

e eR es eT range of the delay angle. The system can therefore operateat maximum control accuracy in all operating points.

Fig. /.Voltage diagram in in- In a rectifier, with constant delay angle a, the direct volt-

Fig.5.* Voltage diagram in in- age decreases when the current increases. This internalvV\,Fe9rter operation, f150 elec- voltage drop is mainly due to the commutation reactances.CC _trical degrees An inverter reduces its counteracting direct voltage by about

the same amount when the load increases and the commuta-tion margin is kept constant. Difficulties may arise, there-fore, in maintaining stable operation, when the characteristics

produce the 90-degree displacement of the a-c part of the out- of the rectifier at the sending end and that of the inverter atput voltage. The same phase-shifting device also provides the receiving end do not have a well-defined point of inter-an auxiliary phase, which, together with the incoming voltage, section. This is one reason for introducing an automatic-feeds a rectifier. This rectifier must have a large number of regulation device, regulating the transmission to a certainpulses in order to avoid smoothing of this d-c part of the out- preadjusted d-c value, independently of the direct voltage.put voltage. Under steady-state conditions, the described See Fig. 10. Another very important reason is the considera-system thus controls each valve with constant margin of com- ble advantages gained during disturbances of the converters.mutation, irrespective of load and voltage variations or of AIany such precautions are taken in the control system tounbalance in the a-c system. If suitable differentiations of reduce the consequences of disturbances and to restore normalthe commutating voltage and direct current are incorporated, operation as soon as possible thereafter. Thus, there is alsothe system can be given such characteristics that it will op- a current regulator acting on the grid control of the inverters.erate satisfactorily also for transients in both the a-c system This regulator is set to a somewhat lower value, normallyand the d-c link. 10% to 15%0 lower than that in the rectifier station; see Fig.

Fig. 9 shows oscillograms taken in the Gotland plant dur- 11. In normal operation, the rectifier end of the transmissioning a special test, where, on the inverter side, short circuits dominates and delivers the current for which it is set, whichwere applied suddenly reducing the commutating voltages means that the current regulator in the inverter end senses ato half the normal value. Not one single failure in commuta- too-high current and therefore, in an effort to bring the cur-tion was recorded during these very extensive tests. rent down to the set value, tries to increase its d-c counter

voltage. In doing so, however, it hits against the consecu-tive grid control, which it cannot override, remaining always

In rectifier operation and also for special conditions of dominant to ensure the required margin of commutation.inverter duty, it is desired to control2 the output voltage and The actual current in the transmission thus will be main-

tained at the set value in the current regulator of the rectifiers,if this end has a voltage high enough to deliver the current.

el During normal operation, the direct voltage of the trans-eR es eT mission is determined by the consecutively controlled inverter,

which results in the voltage of the transmission being mainlyproportional to the voltage of the a-c system at the receivingend.

Lov \j;2 <Regulation During Disturbances

If the direct voltage over the inverter varies without exceed-ing the voltage delivered by the rectifier, the transmitted cur-

Wio * 0 _rent will remain unchanged. Voltage variations of this kind_t- can, for example, depend on disturbances or variations

in the load in the receiving a-c system. In such cases, a d-cw.Id ~~~~~~~~~transmission exhibits an advantage; namely, that its current

o9 > ~~~~~~~inputdoes not change with the occurrence of a disturbance.COMMUTATIN6/X \ ~~~~~Thevoltage decrease of the d-c line may also be caused by a*2Ec sin wt 7/ Yi \ ~~~~disturbance in any inverter valve group. An arc back, or a62ECS;n(W^-T)D/ ,\ IXECOSS\ ~commutation failure repeating itself during a longer periodl / \ t \ _ ~Fig. 6. Voltage/Wt1 \ 6Dtdiagram, same/ ~~~~~~asFig. 5 except/ ~~~~commutation Fig. 7. Commutating volt l/ ~~~~~interval is age, control function, and Il x

> ~~~~~ ~ ~~~showngrid pulse for one valve GRiD PULSETOVALVE S

JANUARY 1964 Engstrten-Oper4tion and Control of HV D-C Transmissionz 73

DELAY ANGLE COMPUTER CONTROLPULSE MAIN VALVES Fig. 8. Block diagram of consecu-e_______________ e______ GENERATOR tive control system

CURRENT SENSING < tX

IdEc cos r VEcsin wtii) -

of time, causes the blocking of the disturbed valve group and the same amount of power will be maintained even duringthe connection of a by-pass valve into the circuit, that is, the such disturbances, which will reduce the direct voltage.direct voltage of the group will be equal to zero. Generally,two or more valve groups in series are employed in each sub- Regulation of Transmitted Power and Other Quantitiesstation and the resulting voltage decrease thus equals 50% or In all cases, when the transmission is grid controlled andless. See Fig. 12. regulated to a constant value of any quantity or according

If the voltage delivered by the rectifier station does not to some other required program, the current settings in bothsuffice to drive the current through the consecutively con- stations have to be automatically adjusted together so thattrolled inverter side, the current regulator of the inverter station the above-described result can be achieved. A change of theregulates the transmitted power according to the somewhat setting made in order to increase or decrease the transmittedlower current value mentioned above; Fig. 13. Such a mode power is reflected in Fig. 11, as parallel displacements of theof operation may arise during disturbances in the a-c system vertical parts of the characteristics, thus maintaining the dif-of the sending side. This mode of operation is, however, ference between the set current values in the two stations atpossible also during disturbances within any rectifier valve a constant value. By rather simple means, some worth-group at the sending end, and also in this case the approximate while limitations can be introduced; for instance, a limitationdecrease in direct voltage is equal to 50%0 or less. of the transmitted current to rated current, which value can-

If the transmission is running at maximum load before the not be overridden by orders from the other control pro-instant of disturbance, the result will be a corresponding loss grams.of transmitted power. In case of transmission with true It is evident that with the flexibility and high speed offeredpower flow control and with the assumption that the current by grid control, the program of the automatic regulators canbefore the disturbance is sufficiently lower than maximum be varied in a great number of ways. It can be arranged tocurrent, the current setting will automatically be adjusted and meet almost any operational requirement, and a d-c trans-

_49 __

8^7 1 r Fi F(̂ ig. 9. Oscillogram, showing 2-{, _t- > ;X@"/<<l 9 a;^*><XA".s i5 v '~ !phase short circuit in 30-kv busI4 \<,/q 4,t yj . / J j v I bar on the Gotland converter sta-

BEGINNING TIME END cate:

1. Time calculation, 1/100 second between calibration peaks2. Direct current3. Incoming control voltage to grid pulse generator for valve connected between phase R and the plus pole of converter 14. The 6-pulse direct voltage of converter 25. Incoming control voltage to grid pulse generator for valve connected with phase R and the plus pole of converter 26. The 6-pulse direct voltage of converter 17. Phase-to-phase voltage R - S of converter 28. Phase-to-phase voltage R - S of converter 19. Grid timing voltage

74 Enqstreim-Operation and Control of HV D-C Transmission JANUARY 1964

RECT. Fig. 10. Characteristic Fig. 11. Characteristic Ed RECTIFIEREd INVERT, ~~~~~curves for rectifier and curves showing co-opera- --- INVERTER

J_ 1 inverter with current re- tion between rectifier andgulation on rectifier only inverter with current reg- OPERATING POINT

ulation in the inverter also

d SETTING MARGIN

REDUCED DC VOLTAGE Ed I REDUCED DC VOLTAGEEd OVER THE INVERTER OVER THE RECTIFIER

OPERATING POINT Fig. 12. Characteristc Fig. 13. CharacteristicOPERATING PONtcurves in case where in- curves in case where rec-verter d-c voltage drops, tifier d-c voltage drops,

d due to a fault duetoafault

mission system in this way offers a greater flexibility than an some cases also with impaired ability to meet certain dis-a-c link. As an example of this, it can be mentioned that d-c turbances, such as high-power short circuits in the receivingtransmission may be used to stabilize long a-c transmissions, system.connected to the same a-c system, by counteracting tendenciesof power oscillations in them. Switching of Power Flow DirectionWhere the operational requirements imply transmitting a In an emergency, the power flow direction may have to be

given amount of power, each terminal station is provided with changed very quickly. This can, in HV d-c transmission, bea handwheel for setting the power to the value desired. The done in a very easy way and without using circuit breakersprinciple of such a power flow control is shown in Fig. 14. or isolators in the main circuits. If the differenice in currentThe power order is transmitted to a computer circuit, which on setting of the two current regulators changes its polarity asthe basis of the prevailing direct voltage translates the power illustrated in Fig. 15, the normal working point is A and theorder to a current order. In the normal way, this current new one will be either B or C. B is reached if this change oforder is compared with the transmitted current, and the dif-ference is fed into the control amplifier.As was already mentioned, the resulting current order in

the rectifier station should exceed that in the inverter stationby a certain amount. When changing the power setting, it is 0necessary therefore to change the power order in each station

ISOLATING TRANSFORMER _simultaneously, or else to increase the power order to the e

rectifiers before the corresponding increment in the inverterCONTROL PULSE GENERATOR

order. To decrease the power, the change must occur inthe inverter station first. By rather simple means of re-mote control, this can be achieved automatically. A tele- GRIDBIAS

communication channel must be available during the setting eprocess. COMMUTATING VOLTAGE

. . . ~~~~DELAY ANGLE COMPUTER __-- .-t-In case the transmission is called upon to maintain con- J \stant frequency in the receiving or sending a-c system, theorder quantity for the current control is derived from anautomatic frequency discriminator. A more complete re-mote control channel, continuously available, must then be DIRECTCURRENTarranged between the two stations to ensure that a correct CURRENT SENSING i1margin between the currents required by the rectifier station .and the inverter station is maintained under all conditions.

In the Gotland transmission scheme,1 a radio link employstwo channels, which are mutually independent, whereby the e

control equipment changes automatically to the other CONTROL AMPLIFIER | llchannel, if the one being used is disturbed. If a telecom-| t| CURNODEmunication interruption occurs in spite of these precautions URHODg | IETLAGthe control system in both stations can be automatically re- RCMUEconnected so that the rectifier station maintains approxi-;lmately constant voltage on a somewhat lower level, and the [inverter station takes over current control as commanded, for -iPOWER ORDER Xinstance, by the frequency discriminator. During this ab- -normal state of control, the transmission thus can go on un- Fig. 14. Block diagram of complete control system for HV d-cdisturbed, but with somewhat inferior power factor and in transmission

JANUARY 1964 Engstrsm-Operation and Control of HV D-e- Tvransmission 75

Ed NORMAL OPERATING POINT Fig. 15 (left). Charac- 7__ _ ZA teristic curves, show-

C r--- ing reversal of power 6 3 4 5l flow direction 8 2

IIm

DC-LINE 2 AC-SYSTEMi dI < | 1 1° 1 01 t1 1t1 ~~~~~~~~~~~634 5 |||

X1: Fig. 16 (right). Main 8 2 7

circuit diagram of HV___c -'-'xs d-c station with two

converters in series 7

current setting is made in the receiving inverter station, and C Arc backs are detccted by means of a differential protectivewill be the new operating point if the change is made in the device, where the direct current of the valve group is com-rectifier station. pared with the alternating current on the valve side of the

converter transformer. If the alternating current is larger,Switching and Valve Blocking a fast-acting relay is energized, which effects the blocking ofAs shown by Fig. 16, no circuit breakers are used on the the whole valve group and the connection of the by-pass

valve side of the converter transformers. Neither are they valve into the circuit. The after effects of an are back gen-used with the d-c circuits. A rapid-action grid blocking sys- erally disappear in 0.5 to 1 second, and normal operation oftem is employed instead. On the other hand, isolators are the rectifier is therefore restored by automatic redeblockingused, which makes it possible to disconnect and ground a after 1 second, and during this sequence the by-pass switchvalve group, wvhen manual work has to be carried out on the is not used.group in question. Single commutation failures in the inverter station will be

In parallel with each valve group, a by-pass valve is con- cleared in a still simpler way. Experience in plants has provednected and it is blocked during normal operation; that is, its that inverterA resume normal commutation immediately aftergrid is fed with a negative voltage from the grid-bias device in occasional faults, thanks to the current limitation offered bythe valve. the high inductance of the d-c circuit, mainly in the smooth-

Blocking of a valve group is effected by stopping the con- ing reactors in combination with the current control of thetrol pulse generators for the main valves, hence, no pulses ar- rectifier station. The duration of such faults is thus only arive at their grids, which are biased with a negative block- fraction of a cycle.ing voltage from a grid-biasing device, one on each valve. At If the commutation failure persistently repeats itself; forthe same instant, a similar control pulse generator, designed example, if the cause is more or less permanent, a blockingfor the by-pass valve, is put into action, and delivers a posi- impulse is obtained from a delayed differential protectivetive grid voltage. The by-pass valve can thus take over the device, similar to that described for are-back blocking. In thiswhole direct current. An automatic relay device ensures that case, however, the direct current exceeds the alternating cur-the by-pass switch is closed and takes over the transmitted rent. This blocking may be made permanent or it may becurrent if the by-pass valve should remain in circuit during a followed by one or more automatic restoring attempts.longer period of time; for example, as a result of lasting The current control, being the primary element of thefaults or of manual blocking in order to take the valve group regulator system, makes it possible to have an effective cur-out of operation. rent limitation during transient faults. As protection againstWhen this group is to carry the current again at, for ex- a persistent short circuit on the d-c side-for instance, a line

ample, manual deblocking, the same action takes place, but fault-ordinary overcurrenit devices of course cannot be used.in the reverse direction. The by-pass valve is not, however, Instead, an undervoltage delayed relay in the rectifier stationable to strike and to take over the current before the voltage serves as protection, tripping the station in case the directover it has reached a certain value. This is achieved by the voltage drops to a low value over a certain period of time.opening of the by-pass switch, while it carries current, and Protection against faults on the d-c lines,3 such as flashoverthe voltage drop in the arc supplies the necessary ignition on a line insulator, may also be provided for by grid control.voltage for the by-pass valve, which then takes over the cur- A special advantage is gained by the simultaneous interven-rent very quickly. However, if this fails, in spite of the by- tion of the current control in both terminal stations. Thepass switch having made a certain fraction of its total move- current control system in the rectifier station reduces its volt-ment, the switch closes again automatically, thus avoiding age in order to keep its outgoing current to the set maximumpermanent arcing in the circuit, value. This current passes through the flashover arc, and

now the current control system in the inverter end senses aProtective Systems lower current than ordered. The inverter therefore adjusts itsA converter station contains many conventional units, voltage to obtain the ordered current and in the flashover

for example, transformers, auxiliary power systems, etc., arc there is, in the steady state, a current corresponding onlywhich are equipped with isolators, circuit breakers, and pro- to the difference in set current values in both stations.tective apparatus according to the same principles as in a-c The conditions are favorable therefore for a quick deioniza-systems. Here, only a few of the devices in question are dealt tion of the arc path and restoration of good insulation. Thewith: those which are wholly characteristic of HV d-c trans- d-c line can be de-energized by blocking the converters, but amission. better solution is to introduce the reduction of d-c line voltage

76 Enqstr6m-Operation and Control of HV D-C Transmission JANU.ARY 1964

in the control program for the rectifiers controlling them over which can be independent of the connected a-c system. Theto inverter operation. The arc will be extinguished and opera- control system has a very fast response, which is advantageoustion restored with least possible delay. in clearing external faults and reducing the consequences of

Finally, it may be mentioned that the grid control can also converter faults. Instead of d-c circuit breakers, a quick-be used for protecting the interior parts of the converter sta- acting grid-blocking system can be applied.tion. Surge diverters, for example, which are exposed to adirect-voltage component, may be helped by automatic grid Referencesblocking to extinguish after functioning. 1. THE GOTLAND D.C. LINK: THE GRID CONTROL AND REGULA-

TION EQUIPMENT, H. Forssell. Direct Current, London, England,vol. 2, no. 5, June 1955, pp. 109-14, and vol. 2, no. 7, Dec. 1955, pp.

Conclusions 16670.

In a-c power transmission, the power flow is controlled 2. OPERATION AND CONTROL OF HVDC POWER TRANSMISSION,by the alternating voltages in the producing and consuming G. Engstrdm. Electrical Journal, London, England, 163 (1959), pp.by th altrnatig votage in te prducig andconsming 1048-54.end. The d-c system is more flexible, and the transmitted 3. CLEARING OF EARTH FAULTS ON HVDC OVERHEAD LINESpower can be controlled in accordance with a desired program, E. Uhlmann. Direct Current, vol. 5, no. 2, Sept. 1960, pp. 45-47.

Discussion that this capability is not the only requirement for successfuloperation of a d-c link in an a-c transmission network. The

E. W. Kimbark (Bonneville Power Administration, Portland, additional requirement, and the most important one, is providingOreg.): One of the figures showed control characteristics for intelligence to direct the control of the d-c equipment. In otherrapid reversal of power. Under what circumstances would words, something is needed to tell the rectifier equipment whatrapid reversal be required, and how rapidly could it be accom- to do for the many and varied contingencies which arise on anplished? electric utility system. These are presently taken care of by the

inherent regulation of the a-c system itself. Other discussershave pointed out that HV d-c transmission is a very real and verycomplex operation. To design detecting equipment to replace

C. S. Schifreen (Philadelphia Electric Company, Philadelphia, the inherent intelligence of the a-c system and at the same timePa.): Mr. Engstr6m's paper is a welcome addition to the tech- provide its degree of accuracy and reliability will be a mammothnical record relating to the special features involved in the task, if even feasible.control of HV d-c transmission. First, I should point out that The author's experience with presently installed point-to-pointthere were several errors in Mr. Engstrom's preprint text under transmission is in no way comparable to meeting the requirementsthe caption "Control of Inverters" involving references to phases for d-c operation, either a parallel superimposed link or an internalR, S, and T in dealing with the commutation phenomenon. link in the a-c network. The conditions which can arise are soHowever, Mr. Engstrom, no doubt, will have corrected these numerous they could well defy the abilities of the finest engineersbefore submitting his text for further publication. to preprogram all that might occur.One could desire greater elucidation of the operation and con- In Mr. Engstrom's presentation, he points out that nearly

trol phenomena of the inverter than Mr. Engstrom included in anything can be done by changing settings, presumably byhis paper. These control functions are difficult to comprehend manual adjustment. This is not practical in the United States,from only generalized descriptions, and a clear understanding is since very few stations are manned, operating costs having led tonot readily achieved by simple armchair analysis. In my own general acceptance of automation. Also, even if attendants werecase, a satisfactory conception of the commutation technique in available, they would only add to the confusion, since com-the Graetz inverter circuit and its requirements of kvar supply munications of conditions from the many remote points affectedwas not attained until the opportunity was afforded me to discuss would not be readily available. Even if they were, it would bethe phenomena on a man-to-man basis with ASEA laboratory beyond the mental ability of any of us to interpret the conditionspersonnel at Ludvika, Sweden. rapidly enough to institute necessary corrective actions. This is

Simultaneously, however, my associates in the Philadelphia further complicated by the need to take simultaneous correctiveElectric Company gained the desired familiarity by investiga- action in many points.tions made possible by purchasing and testing relatively small- Even assuming that it became possible to prepare the necessaryscale models in our own laboratory. It is apparent that other programs and provide the intelligence and perfectly reliable com-utilities will sooner or later have to resort to similar laboratory munications systems which would be required, these would servesetups if they are to participate in developing and understanding as a definite penalty to the economic balance or break-even point.this device. For direct current, the proper controls would be very high, both

in capital and in operating costs. No one has yet ventured tostate these costs.

It would be well if the author would discuss some of theseH. C. Barnes (American Electric Power Service Corporation, problems with the operators of networks in the United StatesNew York, N. Y.): Mr. Engstrom's interesting paper clearly if they expect to apply direct current to such operations. Atestablishes the fact that the d-c rectifier equipment for d-c trans- present, we do not believe the application is feasible, eithermission is capable of control. However, it should be pointed out technically or economically.

JANUARY 1964 Engstrom-Operation and Control of HV D-C Transmission 77