TECHNICAL FIELD

The present invention relates to an installation for transmission of electric power by means of high-voltage direct current, which installation comprises two phase-angle controlled converters, each of which is connected to an alternating-voltage network and which are connected to each other by a dc connection, and control members adapted to control one of the converters to operate as a rectifier and to control another converter to operate as an inverter, and to control one converter, the current-controlling converter for control of the direct current flowing in the dc connection, and to control another converter, the voltage-controlling converter, for control of the direct voltage in the dc connection.

BACKGROUND ART

Installations of the above-mentioned kind are previously well known and described through, for example,

Erich Uhlmann: "Power Transmission by Direct Current", Springer-Verlag Berlin-Heidelberg-New York 1975,

Ake Ekstrδm: "High Power Electronics, HVDC and SVC, The Royal Institute of Technology, Stockholm, 1990,

J. Arrilaga: "High Voltage Direct Current Transmission", Peter Peregrinus Ltd., London, 1983.

In an installation of the above kind, each of the two conver- ters is normally arranged in its own converter station, which are located at a distance from each other and interconnected through a dc-carrying cable and/or an overhead line. In certain cases - in case of so-called back-to-back connections - the

converters may be arranged in the same converter station, and the dc connection then consists of dc busbars in the station.

Usually, the converters are 12-pulse converters and each one consists of two series-connected six-pulse bridges with associated transformers as well as control and protective equipment.

An installation of the kind referred to here may be monopolar, that is, with one single dc line and with return of the direct current through ground, or bipolar, that is, with two conver¬ ters in each station, and with two dc lines, one with positive voltage relative to ground and the other with negative voltage relative to ground.

In installations of the above-mentioned kind, one of the converters is operating as a rectifier and the other as an inverter. The converters are then controlled such that one of the converters is current-controlling and the other is voltage- controlling. Usually, the rectifier is current-controlling, that is, it operates at a somewhat reduced dc voltage and with a control angle α which is influenced by a current controller such that the direct current is controlled towards a current reference value. The inverter is then the voltage-controlling converter. Its control angle is maintained as such a value that the converter operates at a minimum value of the extinction angle γ. The inverter then determines the direct voltage in tr.- transmission, and this direct voltage may be controlled with the aid of the tap changer of the inverter transformer such that the direct voltage, for example at the rectifier, is maintained at a voltage reference value.

In these known installations, the control angle of the recti¬ fier has a nominal value of, for example, 15°, and the tap changer of the rectifier transformer is controlled such that the control angle of the rectifier is maintained within a predetermined interval. This is determined, for example, such that, in case of a change of the control angle from the nominal value to the lower or upper limit of the interval, the change

of the direct voltage caused by the change of the control angle is ± 1 tap-changer step. The limits of the control-angle inter¬ val is typically 12.5° and 17.5° at the above-mentioned nominal control-angle value 15°. A tap-changer step typically corre- sponds to a voltage change of 1.25% of the nominal operating voltage.

The reactive-power consumption of a converter increases rapidly with the control angle and is approximately proportional to sin α. In these known installations, therefore, the reactive-power consumption of the rectifier is relatively high.

There is, therefore, a desire to be able to operate with lower control angles of the rectifier, but in most cases, for various reasons, this has not previously been possible in practice. In certain types of converters, however, it has proved to be possible to operate with low control angles (the control angles are assumed to be related to the line voltage) . This has especially proved to be the case in so-called series-compensa- ted HVDC converters, that is, converters which are connected to their alternating-voltage networks through series capacitors. Such converters may, as rectifiers, operate with very low and even negative control angles. In this way, the reactive-power consumption may be kept low.

in series-compensated converters, however, the firing and extinction voltage jumps, that is, the sudden voltage changes across a converter valve upon firing and extinction thereof, are relatively high and increase with the control angle. In such converters, therefore, it is important to maintain the control angle low.

Operating a rectifier at low control angles, however, entails a significant drawback. This is illustrated schematically in Figure 1 which shows the relationship between the apparent power S, the active power P (P=Scosα) and the reactive power Q (Q=Ssinα) of a rectifier at a certain current and at two different control angles αl and α2. The direct voltage Ud of the rectifier is proportional to the active power P and is

shown on the same axis as this. As mentioned above, the control angle will need to be varied within a range which is so large that the change of the control angle causes a certain given direct-voltage change ΔUd, the magnitude of which is determined by the magnitude of the tap-changer step. At the higher control angle α2 and apparent power S2, a control-angle change Δα2 is required to achieve the voltage change ΔUd. This control-angle change entails a change ΔQ2 in the reactive power. At the lower control angle αl and the apparent power Si, a significantly larger control-angle change Δαl is required to achieve the same voltage change ΔUd, which entails a significantly larger reactive-power change ΔQ1. At still lower control angles, at or in the vicinity of the control-angle values where the direct voltage has its maximum, the required control-angle changes and hence the reactive-power changes will become very large. An operation at low control angles thus entails large reactive- power variations, which may be unfavourable at certain network configurations. While it is true that these reactive-power variations, at least partially, may be counteracted by the introduction of thyristor-switched shunt capacitors, this entails a significant disadvantage from the point of view of cost. The control-angle variations, and hence the reactive- power variations, could also theoretically be reduced by a reduction of the magnitude of the steps of the tap changer. Such a reduction, however, would entail a corresponding increase of the switching frequency of the tap changer, which would entail a considerable, and in most cases unacceptable, increase of the wear of the tap changer.

What has been described above for converters operating as rectifiers is valid - mutatis mutandis - for converters opera ^¬ ting as inverters as well. For such a converter, operation at low extinction angles (referred to the line voltage) is made possible, for example through series compensation, and in that case the above-mentioned large reactive-power variations then arise upon changes of the control angle.

SUMMARY OF THE INVENTION

The object of the invention is to provide an installation of the kind described in the introduction, in which the above- mentioned reactive-power variations during operation at low control angles (or extinction angles) are significantly reduced. The object of the invention is thus, without any need for additional components (such as thyristor-switched shunt capacitors) , and only by a new method of control, to make possible operation at significantly lower control angles and with lower reactive-power variations than what has hitherto been possible, whereby, respectively, the reactive-power consumption of the converters and the magnitude of the firing and extinction voltage jumps may be significantly reduced compared with prior art control principles.

What characterizes an installation according to the invention will become clear from the appended claims.

When stepping the tap changer of, for example, the rectifier, or upon a change of the voltage in the alternating-voltage network of the rectifier, the rectifier will change its control angle to maintain the ordered direct current. In an installa¬ tion according to the invention, the direct voltage of the transmision is changed such that the change in the control angle of the rectifier and hence its reactive-power variation are limited.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will be explained in greater detail in the following with reference to the accompanying Figures 1-7. Figure 1 shows, as mentioned above, the power relationships during operation with two different control angles. Figure 2a shows an overall diagram of an installation according to the invention, where the control-angle variation of the rectifier is limited by supplying a control-angle increment to the control angle of the inverter. Figure 2b shows a corresponding overall diagram for ari alternative embodiment, where the

inverter has a voltage controller and where the control-angle variation of the rectifier is limited by giving an increment to the voltage reference of the controller. Figures 3-7 show in more detail examples of embodiments of different units of the installations according to Figure 2. Figure 3 shows the control-angle generator of the rectifier in the installations according to Figures 2a and 2b. Figure 4a shows the control unit for the reference-value generator of the rectifier in the installation according to Figure 2a. Figures 4b and 4c show two alternative embodiments of the reference-value generator in the installation according to Figure 2b. Figure 5a shows the control device for the tap changer of the rectifier in the installation according to Figure 2a, and Figure 5b show the same control device in the installation according to Figure 2b. Figure 6a shows the control-angle generator of the inverter in the installation according to Figure 2a, and Figure 6b the control-angle generator in the installation according to Figure 2b. Figure 7a and Figure 7b show the control unit for the tap changer of the inverter in the installations according to Figure 2a and Figure 2b, respectively.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

Figure 1 has been described above.

Figure 2a shows an example of an HVDC installation according to the invention. It is arranged for power transmission between two three-phase power networks NI and Nil. A converter SRI is connected through its converter transformer TRI to the network NI, and a converter SRII is connected via its converter trans¬ former TRII to the network Nil. The transformers are provided with tap changers, which are supplied with the control signals tcl and tell. The invention is particularly advantageous in series-compensated converters, since as mentioned above these may operate with very low control angles. The converters are therefore shown connected to their alternating-voltage networks through series capacitors CI and CII. These are preferably located as shown in the figure between the converter transfor¬ mers and the converters, but in an installation according to

the invention they may alternatively be located on the network side of the converter transformers.

The converters are connected to each other by a high-voltage line L which may consist of a cable and/or an overhead line and which has the resistance R - The direct current flowing through the line and the converters is designated Id. The converters have dc-measuring members IMI and IMII and for the sake of simplicity the same designation Id is used for the measured signals from the measurement members as for the actual direct current.

Unless otherwise stated, it is assumed in the examples described here that the most common operating case prevails, that is SRI operates as a rectifier and SRII operates as an inverter, and SRI is current-controlling and SRII is voltage- controlling, and only those parts of the control equipment of the installation which are relevant to this operating case are shown. In the usual way, however, it is assumed that the con- verters and their control equipment are mutually identical such that anyone of the converters can operate as a rectifier (the other converter then operating as an inverter) and anyone of the converters can be current-controlling (the other converter then being voltage-controlling) .

The control equipment of each converter comprises a control- angle generator, CFC _re ct and CFCinv _' a reference-value gene¬ rator, VARrect and VARCinv _* and a tap-changer control device, TCCrect and TCCinv- Further, there is a superordinate control unit SC which supplies a current order 10 to the converters. The various parts of the control equipment are schematically shown as analog functional units, but the parts may either be designed in analog or in digital technique.

Transmission of the necessary control, reference and measure¬ ment signals (e.g. Udref< 10) to or between the two converter stations is performed in a known manner with the aid of a telecommunication link (not shown in greater detail) .

In the installation according to Figure 2a, a limitation of the control-angle variation is obtained, in the manner which will be described in more detail below, by supplying a control-angle increment - negative or positive - to the inverter.

Figure 2b shows an alternative embodiment of an installation according to the invention, in which the inverter has a voltage controller and in which the control-angle variation of the rectifier is limited by supplying an increment - negative or positive - to the voltage reference of the controller. Apart from the composition of some of the control units of the installation and the signals exchanged therebetween, the installation corresponds to the one shown in Figure 2a.

Figure 3 shows the control-angle generator CFC _re ct of the converter SRI in the installations according to Figures 2a and 2b. A summator SUl is supplied with the current order 10 and the current measurement value Id and forms the current deviation Δl _r ect- This is supplied to a current controller CCA _r ect which, in turn, delivers a signal Otrect• The signal is supplied to the converter and controls the control angle thereof. The control angle otrect is limited upwards and downwards to the values ctmax rect and α _m i _n rect• The angle α _m ax rect is the maximally allowed control angle and corre- sponds to the smallest allowed commutating margin. It is defined in the same way as the maximum control angle otmax inv for the inverter, defined by the circuit UAL in Figure 6a (see description of Figure 6a below) . The limitation α _m in rect ensures adequate firing voltage for the valves and that the control angle is not reduced below the point where maximum direct voltage is obtained.

The control angle oc _r ect of the rectifier will be influenced - under otherwise constant conditions - by the voltage in the network NI and by operations of the tap changer of the recti ^¬ fier. An increase of the line voltage or a corresponding tap- changer operation tends to give an increase of the direct voltage of the rectifier and hence an increase of the direct current. The current controller of the rectifier will then

increase the control angle to attempt to maintain the direct current at the current reference. In a corresponding way, a reduction of the line voltage or a corresponding tap-changer operation causes a reduction of the control angle of the rectifier.

Figure 4a shows the reference-value generator VARCrect i ⁿ the installation according to Figure 2a. It is supplied with the control angle otrect of the rectifier and delivers a signal ot _m ax CCA via the reference-value generator VARCinv of the inverter to the control-angle generator CFCinv and the tap- changer control device TCCinv of the inverter. In the summators SU2 and SU3, otrect is compared with the upper and lower limit values α _ma χ nom rect and α _m i _n nom rect _* which define that interval within which it is desired to maintain the control angle under steady-state conditions.

The reference value otmax nom rect is that control angle which gives a direct voltage Ud which corresponds to a reduction of 0.5 tap-changer steps below nominal direct voltage at the control angle ctnom rect- I ⁿ corresponding way, the reference value ot _m in nom rect i that control angle which gives a direct voltage Ud which corresponds to an increase of 0.5 tap-changer steps above the nominal direct voltage at the control angle ctnom rect•

The output signals of the summators are supplied to integrator: IA1 and IA2, the output signals of which are supplied to a sum ator SU6 together with a value ctnom inv which denotes th« nominal control angle of the inverter (see the description o: Figure 6a below for definition of the control angle ctnom inv

The output signal of the integrator IAl is limited upwards tc value determined by the output signal from a summator SU4 and downwards to the value zero. The output signal of the integra¬ tor IA2 is limited upwards to zero and downwards to a value determined by the output signal from a summator SU5.

The summator SU4 is supplied with a quantity otmax inv- which denotes the upper limit to the dynamic operating range of the inverter, and with the quantity α _n om inv- The upper limit to the output signal of the integrator IA1 thus becomes equal to the difference between the upper limit Otmax inv of the control angle and the nominal control angle α _n om inv-

The summator SU5 is supplied with a quantity ct _m in nov inv- which denotes the lower limit to the steady-state operating range of the inverter, and with the quantity Otnom inv- The lower limit to the output signal of the integrator IA2 thus becomes equal to the difference between the nominal control angle ctnom inv and the lower limit o-max inv of the control angle.

The quantities α _n om inv. ct _m in inv and Ct ax inv are supplied to the reference-value generator of the rectifier from the control-angle generator CFCinv of the inverter.

Normally, the control angle of the rectifier will be within the nominal range determined by the values Otmax nom rect and ^α min nom rect- h ^e output signal from the summator SU2 is negative and that from the summator SU3 positive, the output signals from the integrators are zero, and the output signal otmax CCA of the reference-value generator has the nominal value otnom inv- If the control signal of the rectifier, for example because of a voltage change in the network NI, or because of a tap-changer operation, increases such that it arrives above the upper limit for the nominal range, the output signal of the integrator IAl will become positive and increasing. It provides an increment to the signal otmax CCA' which (as will be shown below with reference to Figure 6a) entails an increase of the control angle of the inverter and hence an increase of the direct voltage of the transmission and, in turn, a limitation of the control-angle change of the rectifier. In a corresponding way, if the control angle of the rectifier arrives below the lower limit of the nominal range, the output signal of the integrator IA2 will be negative and

increasing, which entails a negative increment to the control angle of the inverter, a reduction of the direct voltage of the transmission, and hence a limitation of the control-angle change of the rectifier.

In the manner described, the control angle of the rectifier will - except during dynamic states - be maintained within the predetermined nominal range.

Figure 4b shows how the reference-value generator VARCrect is designed in the installation shown in Figure 2b. The generator is supplied with the control angle ot _r ect of the rectifier and delivers a direct-voltage reference Udref via the reference- value generator VARCinv of the inverter to a voltage controller arranged in the control-angle generator CFCinv of the inverter. The generator is built up and functions, in principle, in the same way as the generator described above with reference to Figure 4a. However, the summator SU6 is here supplied with a value Udref nom which constitutes a reference value for the nominal direct voltage of the transmission at the rectifier end.

The output signals of the integrators are limited to values between zero and ±Δudmax- With the definition of the reference values α _m ax nom rect and ot in nom rect given above, the limi¬ ting value ΔUdmax should be set at a value which constitutes the change of the direct voltage which corresponds to 0.5 tap- changer steps.

In the normal case, that is, if the control angle of the rectifier lies within its nominal range, the output signals of the integrators are zero, the output signal Udref has the value Udref nom and the voltage of the transmission is controlled by the voltage controller of the inverter to this value. In the same way as described above, if the control angle of the recti¬ fier tends to go beyond the nominal range, one of the integra¬ tors will give a (positive or negative) increment to the value Udref nom- This entails such a change of the direct voltage of

the transmission that the control-angle change of the rectifier will be limited and the steady-state value of the control angle will be maintained within the nominal range.

Figure 4c shows an alternative embodiment of the reference- value generator VARCrect i ⁿ the installation according to Figure 2b. Like the circuit shown in Figure 4b, it is supplied with the control angle α _r ect of the rectifier and delivers a direct-voltage reference Udref- The summators SU7 and SU8 are supplied with the control angle α _r ect of the rectifier as well as with a constant signal Δα which corresponds to a small angle, and which may typically correspond to the angle 1°. In the summators the limit values α _m ax nom rect - Δα and αmin nom rect + Δα are formed, which in the level flip-flops NVl and NV2 are compared with αrect- If α _r ect becomes greater than the upper or lower than the lower of the limit values, a control signal is obtained and delivered via an OR circuit OGl to a switch SW, which then connects the output of the reference-value generator to the output signal from a calcu- lating circuit BC, which output signal is limited in a limiting circuit FB2. When αrect lies between the two stated limit values, the switch is in the position shown in the figure, in which case Udref = Udref nom-

The control signal αrect °f the rectifier is limited in a limiter FBI to the steady-state, nominal control-angle interval with the limits α _max nom rect and α _m i _n nom rect- The limited control angle is supplied to the calculating circuit BC. In addition, the calculating circuit is supplied with the maximum no-load voltage UdiO of the rectifier and with the direct current Id. From the known main circuit equation of the recti¬ fier, the circuit calculates Ud = f (UdiO- Id, αrect ⁾ that is, that value of the direct voltage which corresponds to the control angle (α _r ect) - supplied to the circuit, under the current operating conditions (UdiO- Id) . The output signal of the calculating circuit is designated Udref comp and is supplied to the limiting circuit FB2, where the signal is limited upwards to the value Udref nom+ΔUdmax and downwards to

the value Ud _r ef nom~ΔUd _m ax (for definition of these quantities, see the description of Figure 4b above) .

The function of the reference-value generator shown in Figure 4c corresponds in all essentials to the function described above with reference to Figure 4b. If the control angle α _r ect of the rectifier reaches any of the limits to the interval determined by the summators SU7 and SU8, the switch SW will become activated. The calculating circuit calculates that direct-voltage value which corresponds to the limited control angle obtained from the limiter FBI, and this value is forwar¬ ded by the switch as the voltage reference Ud _r ef- This means a change of the direct voltage Ud of the transmission to such a value that the control angle does not exceed the value α _m ax nom rect or fall below the value α _m in nom rect- As soon as the control angle α _r ect returns to the interval determined by the summators SU7 and SU8, the switch SW is reset, and the voltage reference Ud _r ef returns to the value Udref nom*

The reference-value generator shown in Figure 4c has one feed¬ back loop less than the generator according to Figure 4b but requires a higher calculating capacity for solution of the main circuit equation of the converter.

Figure 5a shows the tap-changer control device TCC _r ect ⁱⁿ the installation shown in Figure 2a. The control device is supplied with the control angle α _r ect of the rectifier and with the signals α ax CCA- α _m ax inv and α _m i _n nom inv- In level flip- flops NV3 and ΝV4, αrect is compared with the limits αma nom rect and α in nom rect' respectively, for that interval within which it is desired to maintain the control angle under steady- state conditions. If αrect rises to the value α ax nom rect- a positive output signal is obtained from the circuit NV3 , and if αrect drops to the value α _m in nom rect- ^a positive output signal is obtained from the circuit NV4. In the circuits EC1 and EC2, the signal α _m ax CC _A is compared with the control- angle limits α _m ax inv and α in nom inv. which correspond to nominal direct voltage ±0.5 tap-changer steps.

I f both ^α max CCA = «max inv and αrect ≥ α _m ax nom rect an AND circuit AGl delivers a voltage reduction signal tcdl to the tap changer of the transformer TRI which is thus stepped down one step.

If both α _m ax CCA = α in nom inv ajci αrect ≤ αmin nom rect an AND circuit AG2 delivers a voltage increase signal tcul to the tap changer of the transformer TRI which is thus stepped up one step.

The two control signals to the tap changer are in Figures 2a and 5a commonly designated signal tcl.

The steps of the tap changer are adapted such that one step gives a voltage change of a suitable magnitude. A typical value of the voltage change per step is 1.25% of the nominal voltage.

The function of the tap-changer control device is thus such that stepping is ordered when the control angle goes beyond the predetermined interval on condition that the scope for varia¬ tion of the control angle of the inverter is fully utilized.

Figure 5b shows the tap-changer control device TCCrect ⁿ the installation shown in Figure 2b. The control device is supplied with the control angle αrect of the rectifier and the voltage reference Udref- m level flip-flops NV3 and NV4, α _r ect is compared with the limits α ax nom rect and α in nom rect- respectively, for that interval within which it is desired to maintain the control angle under steady-state conditions. If αrect rises to the value α _m ax nom rect< a positive output signal is obtained from the circuit NV3 , and if α _r ect drops to the value α _m in nom rect- a positive output signal is obtained from the circuit NV4. In the circuits EC1 and EC2, the voltage

reference Ud _r ef is compared with the values Ud _r ef nom ± ΔUd _m ax- which correspond to the nominal direct voltage ± 0.5 tap- changer steps.

If both

Ud _r ef = Udref nom + Δud _ma χ. and αrect - O-÷max nom rect an AND circuit AGl delivers a voltage reduction signal tcdl to the tap changer of the transformer TRI which is thus stepped down one step.

If both

Ud _re f = Udref nom - ΔUdmax< and αrect - αmin nom rect an AND circuit AG2 delivers a voltage increase signal tcul to the tap changer of the transformer TRI which is thus stepped up one step.

The two control signals to the tap changer are in Figures 2b and 5b commonly designated signal tcl.

The steps of the tap changer are adapted such that one step gives a voltage change of a suitable magnitude. A typical value of the voltage change per step is 1.25% of the nominal voltage.

The function of the tap-changer control device is thus such that stepping is ordered when the control angle goes beyond the predetermined interval on condition that the control scope of the voltage reference is fully utilized.

Figure 6a shows the embodiment of the control-angle generator ^CFC inv of the inverter in that embodiment according to Figure 2a where the control angle of the inverter is given an incre¬ ment for changing the direct voltage. The control-angle gene¬ rator has a current controller CCAinv a summator SU9 and a level flip-flop NV5.

The current order 10 is supplied to the summator SU9 , in which in a known manner a current margin Δl is subtracted from the current order. The actual current Id will therefore, as long as the rectifier is current-controlling, be greater than the resultant current order IO-ΔI, and the output signal αinv of the current controller is driven to its upper limit value. This is controlled by the signal αmax CCA received from the reference-value generator VARCrect of the rectifier. The output signal of the current controller is limited downwards to the value α _m in inv- which typically has a value just above 90° to prevent the inverter from dynamically changing to rectifier operation.

The output signal αinv of the current controller is supplied to the inverter SRII and determines the control angle thereof.

The output signal ICC from the level flip-flop NV5 is "0" as long as the rectifier is current-controlling and therefore αinv is at the upper limit value α ax CCA- If the inverter were temporarily to change into being current-controlling, its control angle αinv would generally fall below the upper limit value α _m ax CC _A and the signal ICC would become "1", which indicates that the inverter is current-controlling.

The quantities α _m ax inv. α _m in nom inv and α _n om inv are given by a circuit UAL.

The circuit UAL determines the upper limit value α _m ax inv or the control angle of the inverter according to two criteria. 1: the one hand, in a known manner, that control angle, at which * certain predetermined commutating margin is obtained, must nc ^• be exceeded. On the other hand, that control angle, at which the (negative) direct voltage of the inverter no longer incre-* ses with increasing control angle, should not be exceeded sinc-- otherwise the voltage control could have two possible operating points.

The lower limit value α in nom inv is that control angle which gives a reduction of the inverter direct voltage corresponding

to one tap-changer step relative to the direct voltage which is obtained at αi _nv = α _m ax inv-

The nominal operating point of the inverter is that control angle α _n om inv at which an inverter voltage is obtained which is 0.5 tap-changer steps below the voltage at αinv = αmax inv-

Figure 6b shows the control-angle generator CFCinv of the inverter in the case where this has a voltage controller and the direct voltage of the transmission is influenced by control of the voltage reference of the controller. The voltage con¬ troller VCAinv is supplied with the voltage reference Udref from the reference-value generator VARCrect of the rectifier and with a measured value Udrect o the direct voltage of the rectifier. It delivers the output signal αvc _A to the current controller CCAinv and to the control device TCCinv for the tap changer of the inverter transformer.

The output signal αinv of the current controller is driven - when the rectifier is current-controlling - to its upper limit value. This is controlled by the voltage controller VCAinv to the value αvc _A - The output signal of the voltage controller is limited upwards to the value α _m ax inv and downwards to the value α _m in nom inv- These values are obtained, as mentioned above with reference to Figure 6a, from a circuit UAL.

As mentioned above, the input signals to the voltage controller consist, on the one hand, of the voltage reference Udref and, on the other, of a measured value of the direct voltage Udrect of the transmission. Both the reference value and the measured value relate to the voltage at the rectifier side of the trans¬ mission and are transmitted from there via a telecommunications link. The measured value may alternatively be formed in the control equipment of the inverter by adding to the inverter direct voltage the product of the direct current and the known resistance of the line.

Figure 7a shows the control unit TCCinv for the tap changer of the inverter in the embodiment where the control angle of the inverter is given an increment for changing the direct voltage. The direct voltage Udrect of the transmission is supplied to the level flip-flops NV6 and NV7 together with the reference values Ud _m ax a and Ud _m in a - which correspond to the nominal direct voltage ± 1.0 tap-changer step. From NV6 a signal is obtained for stepping down the tap changer if Udrect > dmax a > and from NV7 a signal is obtained for stepping up the tap changer if

Udrect < Udmin a- These signals are supplied to the tap changer via AND circuits AG3 and AG4 and the step-down signal, in addition, via an OR circuit OG2. The resultant step-down signal is designated tcdil and the resultant step-up signal is designated tcull, and the two signals are denoted by the common designation tell.

If the inverter has temporarily taken over the current control, the signal ICC becomes "1", whereby the AND circuits AG3 and AG4 block the tap-changer operation. To ensure, in this opera¬ ting case, that the inverter does not operate with too low a control angle, an unconditional step-down signal is further given with the aid of the level flip-flop NV8 or the AND circuit OG2 if αi v < α _m in nom inv-

Figure 7b shows the control unit TCCinv for the tap changer of the inverter in the embodiment used in the case where the in¬ verter has a voltage controller and the direct voltage of the transmission is influenced by control of the voltage reference of the controller.

Since the above-described control of the direct voltage with the aid of the voltage reference Ud _r ef will vary the direct voltage within the interval between Ud _m in b and Udmax b- which correspond to a nominal voltage ± 0.5 tap-changer steps, the equipment must be able to distinguish between, on the one hand, these voltage variations caused by the rectifier and, on the other hand, such voltage variations as are caused by voltage

changes in the alternating-voltage network of the inverter. This is achieved with the aid of comparison circuits EC3 and EC4, a level flip-flop NV8 and the gates AG3, AG4 and 0G2 mentioned above.

The circuits EC3 and EC4, respectively, deliver a "1" signal if the quantity αvcA is at either of the limits α _m ax inv or α in nom inv- The control angle of the inverter - if the inverter is current-controlling - is equal to the quantity αvc _A - This means that the inverter voltage controller VCAinv alone will counter¬ act deviations in the direct voltage as long as the signal αv _CA and hence the control angle lies between the limit values mentioned above. Only if the control possibilities of the voltage controller are exhausted, which is indicated by a "1" signal from either of the circuits EC3 and EC4, will an opera¬ ting signal to the tap changer be delivered.

In the foregoing a number of embodiments of the invention are described, but, of course, the invention is not limited to these embodiments.

The result of a control method according to the invention is that in case of variations in the UdiO of the rectifier or the inverter of up to ±1 tap-changer step, the control angle of the rectifier will be maintained between the limits α _m ax nom rect and α _m in nom rect- Further, the control angle of the inverter will be maintained between the limits α ax inv and min nom inv- In both cases, this gives a maximum control-angle variation corresponding to ±0.5 tap-changer steps. This limi- tation of the control-angle variations is achieved by con¬ trolling the direct voltage of the transmission in the way . described above. The maximum control-angle variation is thus approximately halved compared with prior art installations and control methods. This entails a corresponding reduction of the variations in the reactive-power consumption of both the recti¬ fier and the inverter. The changes of the reactive-power con¬ sumption will be distributed equally between the rectifier and the inverter.

Larger changes of the UdiO °f the rectifier than what corre¬ sponds to ±1 tap-changer step will cause the control angle of the rectifier to go outside the predetermined range and hence cause a movement of the tap-changer and therefore a restoration of the control angle and the direct voltage to the nominal values.

To avoid possible instabilities in the tap-changer control, it may be advantageous to synchronize the tap-changer stepping between the two stations (converters) of the transmission such that, when ordering stepping in one station, a stepping in the other is delayed until the stepping has been carried out in the former station.

The above-described approximate halving of the maximum ampli¬ tude of the line-voltage variations has been achieved with an unchanged magnitude of the steps of the tap-changers and hence without any increase at all of the operating frequency and wear of the tap changers.

The above description relates to an installation arranged for power transmission between two alternating-voltage networks. However, the invention may be applied also to so-called multistation transmissions, that is, installations where three or more alternating-voltage networks are connected to each other with the aid of converters and a common direct-voltage connection.

Alternatively, the rectifier may be voltage-controlling and the inverter current-controlling. Also in this case, according to the invention, it is the current-controlling converter - in this case the inverter - whose control-angle variations are reduced by causing the voltage-controlling converter to change the direct voltage of the transmission in such a way that the above-mentioned control-angle variations are counteracted.

In the embodiment described above, the limit values in the control equipment are chosen such that the control-angle changes and hence the reactive-power changes are distributed

equally between the current-controlling and the voltage- controlling converter. This distribution may be suitable if the two power networks are approximately of even strength. If one of the networks is strong and the other weak, the former with- stands higher reactive-power variations for a certain given level of the voltage variations. It may then be suitable to control the transmission such that the strong network is allotted a larger part and the weak network a smaller part of the control-angle variations, which may be achieved by a suitable choice of limit values in the control equipment.