EP3766170A1

EP3766170A1 - Ac/dc vsc connected to three or more dc lines

Info

Publication number: EP3766170A1
Application number: EP18712839.2A
Authority: EP
Inventors: Kalle ILVES; Nicklas Johansson; Lennart Harnefors; Stefanie HEINIG; Staffan Norrga; Hans-Peter Nee
Original assignee: ABB Power Grids Switzerland AG
Current assignee: Hitachi Energy Ltd
Priority date: 2018-03-15
Filing date: 2018-03-15
Publication date: 2021-01-20
Also published as: WO2019174732A1

Abstract

A voltage source converter (12) converts between AC and DC is connected between at least three DC terminals (DC1, DC2, DC3) and comprises three parallel AC phase modules each comprising an corresponding AC terminal (ACA1, ACB1, ACC3), where each comprises a lower converter arm (1a) connected between a first DC terminal (DC1) and the AC terminal (ACA1), a first upper converter arm (ua0, ua1) connected between the AC terminal (ACA1) and a second DC terminal (DC2) and a second upper converter arm (ua2), the lower and first upper converter arms forming a phase leg and the second upper converter arm stretching out from a connection point (CP) on the phase leg to a third DC terminal (DC3), where the phase leg comprises submodules (CLAL, CLAC, CLAB1) for supplying a first DC voltage, while the second upper converter arm comprises circuitry (CLAB2) for, together with the submodules between the first connection point and the first DC terminal, providing a second DC voltage and the first DC terminal is adapted to be connected to ground.

Description

AC/DC VSC CONNECTED TO THREE OR MORE DC LINES

FIELD OF INVENTION

The present invention generally relates to voltage source converters. More particularly the present invention relates to a voltage source converter for conversion between alternating current, AC, and direct current, DC.

BACKGROUND

Multiterminal High Voltage Direct Current (HVDC), MTDC, systems employing Voltage Source Converters (VSCs) is an emerging technology. This means that a number of converter stations may be interconnected in a Direct Current (DC) grid, where each converter station comprises a converter converting between AC and DC and at least one of these converters is connected to at least two different DC lines.

In meshed or multiterminal DC-grids the DC load flow controllability is, however, reduced. It is in many cases not possible to control the DC current in the different lines independently.

WO 20 13/ 071962 discloses a converter that is connected to two different DC systems. However, it is uncertain if the converter is intended to be used in a DC system or not. There is also no discussion about improving load flow controllability in such a system.

The invention is provided for addressing this problem of improving the load flow control capability in a DC system. SUMMARY OF THE INVENTION

The present invention is directed towards obtaining a converter for converting between alternating current and direct current and which converter is capable of improving the load flow control capability in a DC system.

This object is according to a first aspect of the present invention achieved through a voltage source converter for conversion between alternating current, AC, and direct current, DC, and being connected between at least three direct current, DC, terminals and comprising three parallel AC phase modules each comprising an corresponding AC terminal, each AC phase module further comprising:

a lower converter arm connected between a first of the DC terminals and the AC terminal, a first upper converter arm connected between the AC terminal and a second of the DC terminals and a second upper converter arm, the lower and first upper converter arms forming a first phase leg connected between the first and second DC terminals and the second upper converter arm stretching out from a first connection point on the first phase leg located between the AC terminal and the second DC terminal to a third DC terminal, where the phase leg comprises

submodules for supplying a first DC voltage between the first and second DC terminal for interfacing with a first DC line and an AC voltage on the AC terminal for connection to an AC network, while the second upper converter arm comprises circuitry for, together with the submodules between the first connection point and the first DC terminal, providing a second DC voltage between the first and third DC terminals for interfacing with a second DC line, wherein the first DC terminal is adapted to be connected to ground.

The object is according to a second aspect achieved through a direct current power transmission system comprising a voltage source converter according to the first aspect. The present invention has a number of advantages. It improves the DC load-flow controllability, which allows an optimization of the load flow distribution to be made, leading to higher transmission capability and lower losses. It can also be used to boost the DC- voltage, which may be useful for example for a long DC line with a significant voltage drop along the line. It also provides a greater flexibility in that it allows the possibility to interconnect DC grids with different voltages without the provision of additional DC/ DC converters.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will in the following be described with reference being made to the accompanying drawings, where fig. 1 schematically shows a first type of multiterminal DC grid comprising a number of interconnected voltage source converters,

fig. 2 schematically shows an expansion of the first multiterminal DC grid, fig. 3 schematically shows two converters interconnected by a

tapping/ boosting converter,

fig. 4 schematically shows a first version of a voltage source converter having a number of arms provided for a number of phases, where for one phase there is a lower arm, common valve arm and two branched arms, fig. 5 schematically shows a second version of voltage source converter having a lower arm, common valve arm and three branched valve arms for a phase,

fig. 6 schematically shows a model of the common valve arm and two branched arms of the converter shown in fig. 4,

fig. 7 schematically shows a plotting of a boosting factor as a function of active AC input power for a converter modeled according to the model, fig. 8 schematically shows the boosting factor as a function of the branched arm voltage factor for a converter modeled according to the model, fig. 9 schematically shows a variation of the voltage source converter having a lower arm, common valve arm and two branched arms for a phase, where one of the branched arms is a branched valve arm and the other is an AC component handling arm, and

fig. 10 schematically shows the arms in a fourth version of voltage source converter associated with one phase, without a common valve arm, but comprising a lower arm, two branched arms and a disconnector.

DETAILED DESCRIPTION OF THE INVENTION

In the following, a detailed description of preferred embodiments of the invention will be given .

Fig. 1 shows a first type of multiterminal direct current system 10. In the system 10 there is a first converter 12 converting between alternating current (AC) and direct current (DC), a second converter 14 converting between AC and DC and a third converter 16 converting between AC and DC, which converters are typically voltage source converters (VSCs) and with advantage also modular multilevel converters (MMCs) comprising chain links made up of submodules such as half-bridge and/ or full-bridge submodules.

The first converter 12 is connected to the second converter 14 via a first DC line 18. The first converter 12 is also connected to the third converter 16 via a second DC line 20. There is finally a third DC line 22 interconnecting the second and third converters 14 and 16.

Fig. 2 schematically shows the system of fig. 1, after expansion with two further converters. There is here a fourth converter 24 converting between AC and DC and a fifth converter 26 also converting between AC and DC, where the fourth converter 24 is connected to the second converter 14 via a fourth power line 28 , while the fifth converter 26 is connected to the third converter 16 via a fifth power line 30. The fourth and fifth converters 24 and 26 are finally interconnected via a sixth power line 32.

Fig. 3 schematically shows a second type of“point-to-point” system where the first, second and third converters are connected in series on the DC side. The second converter 14 is here connected to the first converter 12 via the first power line 18 , which is connected to the third converter 16 via the second power line 20. There is thus no DC connection between the second and third converters. The first converter 12 may either tap or inject power and may also perform voltage boosting.

Fig. 4 schematically shows a first type of converter that may be used in the different previously disclosed converter stations, for instance as the first converter 12. The converter is an MMC type converter with integrated DC/ DC conversion functionality.

As stated earlier, the converter 12 is a modular multilevel converter (MMC) that converts between Direct Current (DC) and Alternating Current (AC) and may with advantage be connected as a converter in the grid 10 or as a tapping/boosting converter in a point-to-point system. The converter has three DC terminals DC1, DC2 and D3 for enabling the provision of at least two different DC voltages and a number of AC terminals ACA1, ACB1,

ACC1 for provision of a number of phases of an AC voltage. The converter comprises three parallel AC phase modules, each

comprising a corresponding AC terminal for handling a corresponding AC phase. For the first phase there is thus a first phase module comprising a lower converter arm la comprising a lower arm chain link CLAL connected between the first DC terminal DC1 and the AC terminal ACA1 of the first phase. The AC terminal ACA1 is further connected to an upper common arm uaO comprising a common chain link CLAC, where the upper common arm uaO with upper common chain link CLAC is in turn connected to a second DC terminal DC2 via a first branching arm ual comprising a first branching chain link CLAB1. The upper common arm uaO is in this embodiment furthermore connected to a third DC terminal DC3 via a second branching arm ua2 comprising a second branching chain link CLAB2.

The common arm uaO and first branching arm ual in this case make up a first upper converter arm that is connected between the first AC terminal ACA1 and the second DC terminals DC2. Moreover, the lower and first upper converter arms la, uaO and ual together form a phase leg connected between the first and second DC terminals DC1 and DC2, where the chainlinks of this phase leg provides a first DC voltage between the first and second DC terminals DC1 and DC2 for interfacing with a DC line, for instance the first DC line. The second branching arm ua2 in turn forms a second upper converter arm that stretches out from a first connection point CP on the first phase leg to the third DC terminal DC3. Furthermore, the first connection point CP is located between the AC terminal ACA1 and the second DC terminal DC2, where the connection point CP in this embodiment is a connection point between the upper common arm uaO and the first branching arm ual, which is at a junction between the common chain link CLAC and the first branching chain link CLAB1. The common arm uaO thereby stretches between the AC terminal ACA1 and the connection point CP, while the first branching arm ual stretches between the connection point CP and the second DC terminal DC2. The chain links comprise submodules and therefore, it is also clear that the first

connection point CP is provided between two submodules of the first upper arm. The second upper converter arm comprises circuitry for, together with the chainlinks CLAL and CLAC between the first connection point and the first DC terminal DC1, provide a second DC voltage between the first and third DC terminals DC1 and DC3 for interfacing with another DC line, such as the second power line 20. In this embodiment the circuitry is made of the second branching chain link CLAB2 for forming the AC voltage and the second DC voltage. The converter realization in respect of the second and third phases is the same. The second converter module thus comprises a lower chain link CLBL connected between the first DC terminal DC1 and a second AC terminal ACB1 of the second phase. The second AC terminal ACB1 is also connected to an upper common chain link CLBC, where the upper common chain link CLBC is in turn connected to the second DC terminal DC2 via a first branching chain link CLBB1 and to the third DC terminal DC3 via a second branching chain link CLBB2. The third converter module comprises a lower chain link CLCL connected between the first DC terminal DC1 and a third AC terminal ACC1 of the third phase. The third AC terminal ACC1 is also connected to an upper common chain link CLCC, where the upper common chain link CLCC is in turn connected to the second DC terminal DC2 via a first branching chain link CLCB1 and to the third DC terminal DC3 via a second branching chain link CLCB2.

In the figure there is also shown a control unit 34 set to control the different chain links provided for the different phases, which control involves the forming of an AC voltage on a corresponding AC terminal for connection to an AC network. At times it may also be involved in handling various faults.

The chain links are used for forming a number of discrete voltage levels of an AC waveshape as well as contributing to the forming of a DC voltage. Therefore, each chain link may comprise a number of series-connected or cascaded submodules, where a submodule may be realized as a half-bridge submodule, a full-bridge submodule or as a hybrid between the two. As is known a half-bridge submodule comprises two switches connected in parallel with an energy storage element, for instance realized as a capacitor. One submodule terminal is then provided at the junction between the two switches and the other submodule terminal is provided at a junction between one of the switches and the energy storage element.

The half-bridge submodule is configured to either provide a zero voltage or a unipolar voltage corresponding to the voltage across the submodule capacitor. A full-bridge submodule comprises two strings of series connected switches connected in parallel with the energy storage element, where one submodule terminal is provided at the midpoint of one of the strings, while the other submodule terminal is provided at the midpoint of the other string. The full-bridge submodule has a a zero and bipolar voltage contribution capability corresponding to the voltage across the capacitor.

The converters shown so far are all asymmetrical monopole converters where the first DC terminals is connected to ground potential and the second, third and fourth DC terminals are connected to the same or different positive poles of a DC system such as the DC grid. Thereby the AC terminals have an AC waveshape with a DC offset, which DC offset may be removed using a transformer. It should be realized that as an alternative the converter may be a bipole converter, in which the previously described converter structure is complemented by a mirrored converter structure connected between the first DC terminal and a fifth, sixth and seventh DC terminal for connection to corresponding negative poles.

As another alternative the converter may be a symmetric monopole converter. In this case the lower arm would be connected to two or more branching arms in the same way as the common arm.

Fig. 5 shows a variation of the converter that is suitable when connection is made to three DC lines, such as the second converter 14 in fig. 2. The converter is the same as in fig. 4 except for the fact that there is a further branching arm for each phase connected to a further DC terminal. There is therefore in this case a fourth DC terminal DC4 for connection to the third DC line. In this case each upper common chain link CLAC, CLBC and CLCC is additionally connected to the fourth DC terminal DC4 via a corresponding third branching chain link CLAB3, CLBB3 and CLCB3. There are a number of problems that may exist when a converter is to have two or more DC connections.

As can be seen in fig. 1 and 2 a converter may be connected to a

multiterminal DC grid, in which case it may be connected to two or more DC lines. If these connections are made using the same DC terminal then there may be a problem of load flow controllability being reduced.

As an example, consider fig. 1. If the voltage at the first converter 12 is kept constant and the voltages in second and third converter 14 and 16 and are adjusted to control the DC current, there are three lines 18 , 20 and 22 but only two degrees of freedom. Hence it will not be possible to control the DC current in the three lines independently. The complexity of controlling the load flow increases even further if the DC grid 10 is later extended with the terminals 24 and 26 as shown in fig. 2. Improving the load flow controllability in a DC grid would allow for an optimization of the load flow distribution, leading to higher transmission capability and lower losses.

However, also in systems of the type shown in fig. 3, there may be problems. If a long-distance transmission line was built, for example between Europe and China, this may require both voltage boosters and tapping along the way, where the first converter 12 is a tapping and/ or boosting converter between two end point converters in the form of the second and third converters 14 and 16. Similar to the case in Fig. 1 and 2, such system would also benefit from independent control of the DC voltages, not only to improve DC load flow control, but more importantly to boost the DC-link voltage.

The above disclosed converters are provided for addressing the above described and associated problems.

Each phase leg of the above shown converter realizations comprises a lower arm l_a, a common upper arm uaO , and a number of split or branching arms ual,ua2,...,uaN as illustrated in Fig. 5. The voltage and current in the lower arm la in the first phase are given by: ^vi_a = ^vo + V_l cos(a>t) ( la) where Idn is the DC current in a branched arm u_an (where n £ [l; 2] for a two-terminal and n€ [ 1; 2; 3 ;....;N] for an N terminal MMC, which number are the number of terminals connected to the upper part of the converter), k is the ratio of the AC current distribution between the lower and upper arms la and ua, and Vo is the DC voltage in the lower arm l_a.

For stable operation the active power resulting from alternating and DC components in the lower arm must cancel out,

The lower arm quantities for a given operating point can then be described by equations ( 1) - (3) where the DC voltage Vo can be chosen freely. In a conventional MMC, however, the DC voltage Vo is typically half of the DC- link voltage.

The voltages in the upper common arm and branched arms can be defined as ^vuao = yV_{0 -} (1 - r)V₁ cos(fflf) (4a) v_uan = V_{An -} rV_{l C}os((at) (4b) where y is the DC voltage contribution in the upper common arm in relation to the lower arm, r is a distribution factor which must be chosen such that the arm energies remain stable, and „= v^ - d + y)v₀ (5) Similarly, the arm currents are given by

where x is a distribution factor for the alternating current between the split arms. Accordingly, the values of x_n must satisfy Similar to the lower arm, the active power resulting from the AC and DC components must also cancel out in the upper arm, that is,

Solving for r yields

r = - 1— (1 + y)k

( 11)

1 - k Finally, in order to maintain stable arm energies in the split arms,

Substituting ( 11) and solving for x_n yields

The distribution factor x_n of a branched arm thereby depends on the active power of the branched arm, the relationship y between the voltage contribution of the common arm and the voltage contribution Vo of the lower arm and the active power of the AC terminal. It can more

particularly be seen that the distribution factor x_n which may depend on the per phase active power being separate from zero, may be provided as the active power of the branched arm divided by the active power of the AC terminal and by a variable comprising the relationship y. The variable may more particularly be 1/ k - 1- y, where k is ratio of the AC current distribution between the lower and upper arms.

In order to avoid large alternating currents in the branched arms, the DC voltages Vo and yVo can be chosen such that It can thereby be seen that DC voltage contributions for the lower arm Vo and the common upper arm y*Vo may be chosen so that a maximum absolute value of an AC current distribution factor between the branched arms is kept below or equal to a maximum distribution factor value.

As an example consider an MMC connected to three DC- lines where the DC- link voltages in two of the lines differ by ± 10 % from the nominal voltage. This should cover most typical DC load flow cases.

It should be realized that the invention is in no way limited to voltage differences of ± 10 %. Higher differences such as connecting a 400 kV DC- grid to a 320 kV DC-grid is also possible. For different cases the voltage difference and range in which the voltage can be varied can have a different impact on the required dimensioning of the arms. In the considered example, the vdci, Vdc2, and vdc3 are 440 , 360 , and 400 kV, respectively, which may be the voltages between the second and first DC terminal, the voltage between the third and first DC terminal and the voltage between the fourth and first DC terminal. The direct current from each line flowing into the converter per phase may as an example be 0.8 kA, 0.7 kA, and -0.5 kA, respectively.

There may be two degrees of freedom in the converter design, the direct voltage V0 in the lower arm and the direct voltage yVO in the upper arm.

In this example, a parameter sweep was performed to find the

combination which requires the least number of submodules. In order to avoid large currents the considered solutions were limited to such cases where k G[0.45 0.55] and x_max = 1.1. The parameter sweep resulted in a DC voltage in the lower arm of 192 kV and y =0.58. Calculating the peak values of the arm voltages indicate that if 3 kV submodules are used, a minimum of 131 modules are needed in the lower arm, 73 in the upper arm, and 78 in each and one of the three branched arms. It is possible to analyze the operation of the converter shown in fig. 4 in another way. A model of the part of the converter for the first phase from the AC terminal and upwards that is used in this analysis is shown in fig. 6. In this analysis the following symbols are used: udc per-pole DC voltage, which corresponds to y*V0

u_v converter phase voltage

i_v converter input phase current

idi per-phase output direct current at terminal DC3

id2 per-phase output direct current at terminal DC2

P— u_viv per-phase input active power at the AC terminals

Pdi = 2udcidi per-phase output power at terminal DC3

Pd2 = 2udcid2 per-phase output power at terminal DC2

p 1,2 voltage boosting factors of terminals DC2 and DC3

q part of the phase voltage taken up by the branched arms

x part of the AC arm current routed through terminal DC3

Obviously, if the converter is approximated as lossless, Pd l + Pd2 = P. It should normally be sufficient to use small boosting, i.e., p 1,2 is in the range of a few percent. Moreover, it is desired that the number of submodules in the branched arms should be much smaller than that in the common-arm. This implies that

For zero mean power in each of the three chainlinks, the sum of the DC and AC active powers must be zero. Thus, the following relations must hold: qxP

P_lUjdl 0 branched arm 2 ( 16)

2

q( 1 - x)P

P2^UdJd 2 ^~ 0 branched arm 1 ( 17)

2 common arm ( 18 ) Suppose that a certain boosting p2 of the DC voltage at terminal DC2 is desired. Solving for id2, q, and p l yields The most important observation to make is that q is proportional to p2 and inversely proportional to P. Boosting therefore requires a nonzero P. For P = 0 , p l = p2 = 0 must be used, i.e., the cells in the split arms are bypassed.

Alternatively, p2 can be solved from the second relation in equation ( 19) as

This shows that, for a certain q, the maximum I p2l is obtained for x = 0. In other words, for a certain desired boosting level, the minimum voltage rating— although not the minimum current rating— of the branched arms is obtained for the special case of x = 0 => p l = 0. So, consider x = 0 henceforth, for simplicity at least. Then equation (20 ) simplifies to

It can be seen that a voltage difference p2 to a nominal DC voltage at the second DC terminal is set based on the power Pdl at the third DC terminal DC3 and on at least one scaling of the power P at the AC terminal, where equation 21 in fact sets out two such scalings. It can also be seen that the voltage difference p2 to the nominal DC voltage ud at the second DC terminal DC2 is set as a first scaled input power divided by a difference between a second scaled input power and the power Pdl at the third DC terminal DC3, where the first scaling factor comprises the fraction q of the AC voltage applied across the first branched arm ual times the part of the AC current ( l-x) running through the first branched arm ual and the second scaling factor is formed as one minus the first scaling factor.

Equation (21) will now be evaluated for two cases, which correspond to two realistic applications of the converter.

Suppose that a converter equipped with branched arms is placed as a tapping station near the center of a long DC line, such as the converter 12 in fig. 3. The line from the sending-end converter 14 is connected to terminal DC3, so Pdl < 0.

The tapping converter could be located either at a generation center (P > 0) or at a load center (P < 0). The objective for the boosting functionality

(p2 > 0) is to compensate for the voltage drop across the long DC line, thereby obtaining a smaller id2, see the first relation in equation ( 19). Figure 7 shows p2 as a function of P for Pd l = -l p.u. (equal ratings of the sending-end and tapping converters) and Pdl = -2 p.u. (twice the rating of the sending-end converter as compared to that of the tapping converter).

In both cases, I ql = 10 %.

As can be seen in the figure, tapping at a load center (P < 0) gives a steeper rise of the boosting factor as I PI increases as compared to tapping at a generation center (P > 0 ). On the other hand, q < 0 must be used for P < 0 , so an increased rating of the common-arm cell string may be needed.

In both cases, however, boosting by a few percent can be obtained even for I PI in the range of 0.5 p.u. The second application is a converter station where the DC lines have the same power direction, i.e., the signs of Pdl and Pd2 are equal. This could be two (semi-) parallel DC lines, radiating from a converter situated at a generation center, whose loadings need to be actively controlled, such as the first converter 12 in fig. 1 or 2. Assuming that equal sharing is desired, i.e. the power P at the AC terminal is shared equally between the branched converter arms ual, ua2, Pd l = Pd2 = P/ 2, (7) yields

It can thereby be seen that the voltage difference p2 to the nominal DC voltage ud at the second DC terminal DC2 is set as a fraction q of the AC voltage applied across the first branched arm ual divided by a difference between one half and the fraction q. More generally speaking, if a fraction n of the total power P is to be provided by the second branching arm, then the boosting level p2 of the first branching arm may be set as q divided by ( 1 - q - n), where q is the part of the phase voltage taken up by the split arms.

Several percent boosting is here obtained even for modest I ql , which is shown in fig. 8 . The proposed converter is thus very effective as a power- flow controller. It was found that x = 0 => p l = 0 gives the minimum voltage rating of the split arms. This gives the option to replace the chain link in the second branched arm by a circuitry for removing AC

components from the output voltage, which circuitry may be in the form of a parallel LC circuit tuned to the fundamental frequency. This is shown in fig. 9, where the second branching arms of the three phases are realized as filters FA, FB and FC. As an alternative the circuitry could be realized as a pure inductance. This would result in an AC component in the second branched arm, but it would be 90° phase shifted to uv and thus would not produce active power in the first branched arm, where it would circulate. However, a very large inductance would be needed to obtain a low enough current. By using a tuned LC circuit, the AC component can be reduced to a minimum even with a relatively small inductance.

A number of observations may be made from the above-described analysis.

• The sum cell voltages in the branched arms are balanced.

• Tapping/ boosting (Pd l < 0) with inverter operation (P < 0 ) requires— as predicted by theory— a modest increase of the peaks of the common-arm insertion indices. Since the branched-arm insertion indices are within the range [0 , 1] , use of half-bridge cells in the branched arms of the second DC terminal is sufficient.

• Tapping/ boosting (Pd l < 0) with rectifier operation (P > 0 ) reduces the peaks of the common-arm insertion indices, but the branched-arm insertion indices are now periodically negative. This implies that at least some of the cells in the branched arms of the second DC terminal must use full bridges.

• Power-flow control (Pdl = P/ 2) only needs a modest amplitude of the branched-arm insertion indices.

• For power-flow control a boosting of p2 = 2% may be sufficient. For a converter operating at 800 kV to be boosted by 2%, the branch-arm chain link responsible for the boost can then be reduced to only six submodules rated at 3 kV.

Some further observations are:

• Active power transfer (P ¹ 0 ) is needed in order to obtain boosting. The higher I PI , the higher boosting can be done.

• From a perspective of voltage rating, it is beneficial to use boosting in only one of the terminals, here terminal DC2. (The direct voltage at the third DC terminal DC3 is still controllable by varying the common direct voltage udc.) A special case of the proposed topology is when the shared upper arm voltage-reference is set to zero, i.e. yVo or ua is set to zero. The resulting topology is then a converter with multiple branched arms but without a common arm.

In this case there is thus no common arm. Instead the two branched arms are directly connected to the AC terminal. The connection point CP from which the second branched arm stretches out, and if provided also the third branched arm, is thus provided at the AC terminal. This reduces the possibilities to optimize the dimensioning for certain operating conditions. However, it could provide reduced complexity and allow upgrading of existing MMC according to the proposed concept. It is thus possible to retrofit existing converters. If the DC voltage in the upper arms is at least 50 % of the full DC-link voltage and if full-bridge submodules are used, this configuration can also provide DC breaker functionality. This is schematically indicated in fig. 10 , which shows one phase of the converter 12 with two branched arms connected directly to the AC terminal ACA1. The first connection point CP is thus at the AC terminal ACA1. The first branched arm in this case comprises a first and a second first branched arm chainlink CLAB1A and CLAB1B, while the second branch arm comprises a first and second branched arm chain link CLAB2A and CLAB2B. The second and third DC terminals DC2 and DC3 are also interconnected by a mechanical disconnector MD.

The use of the converter may be the following:

There may, as an example be a line to ground fault in the first DC line 18 interconnecting the first and second converters 12 and 14 in the system in fig. 1. Accordingly, the faulty line 18 must be disconnected such that the second and third lines 20 and 22 can remain operational. According to this aspect of the invention, the converter is used to force the current through the mechanical disconnector MD to zero by diverting it through the submodules in the chain links CLAB1A, CLAB1B, CLAB2A and CLAB2B. When the current through it is zero, the mechanical disconnector MD may be opened Here it may be mentioned that it is also possible that the mechanical disconnector is a part of a DC circuit breaker that in this case is also connected between the two DC terminals DC2 and DC3.

Thereby it is also possible to eliminate the need for separate DC breakers.

The configuration in fig. 10 is also interesting without a mechanical disconnector MD. In this case a DC fault on one of the lines can be handled by the converter. The converter may also remain operational during and after the fault (with one DC voltage level reduced to zero).

The invention described so far has a number of advantages.

Increased DC load-flow controllability in meshed or multiterminal DC grids, which allows an optimization of the load flow distribution to be made, leading to higher transmission capability and lower losses.

Possibility to boost the DC-voltage in a tapping station of a very long DC- link. Possibility to interconnect DC grids with different voltages.

Possibility to eliminate separate DC breakers. The control unit may be realized in the form of discrete components.

However, it may also be implemented in the form of a processor with accompanying program memory comprising computer program code that performs the desired control functionality when being run on the processor. A computer program product carrying this code can be provided as a data carrier such as one or more CD ROM discs or one or more memory sticks carrying the computer program code, which performs the above-described control functionality when being loaded into a processor performing the role of control unit of the voltage source converter.

From the foregoing discussion it is evident that the present invention can be varied in a multitude of ways. It shall consequently be realized that the present invention is only to be limited by the following claims.

Claims

1. A voltage source converter ( 12; 14) for conversion between alternating current, AC, and direct current, DC, and being connected between at least three direct current, DC, terminals (DC1, DC2, DC3 ; DC4) and comprising three parallel AC phase modules each comprising an corresponding AC terminal (ACA1, ACB1, ACC1), each AC phase module further comprising:

a lower converter arm (la) connected between a first of the DC terminals (DC1) and the AC terminal (ACA1), a first upper converter arm (uaO , ual) connected between the AC terminal (ACA1) and a second of the DC terminals (DC2) and a second upper converter arm (ua2), said lower and first upper converter arms forming a first phase leg connected between the first and second DC terminals (DC1, DC2) and said second upper converter arm stretching out from a first connection point (CP) on the first phase leg located between the AC terminal (ACA1) and the second DC terminal (DC2) to a third DC terminal (DC3), where the phase leg comprises submodules (CLAL, CLAC, CLAB1; CLAB1A, CLAB1B) for supplying a first DC voltage (( l+p2) udc) between the first and second DC terminal for interfacing with a first DC line ( 18 ) and an AC voltage on the AC terminal (ACA1) for connection to an AC network, while the second upper converter arm comprises circuitry (CLAB2; FA; CLAB2A, CLAB2B) for, together with the submodules between the first connection point and the first DC terminal, providing a second DC voltage (( l+p l)ud) between the first and third DC terminals for interfacing with a second DC line, wherein the first

DC terminal is adapted to be connected to ground.

2. The voltage source converter ( 12; 14) according to claim 1, wherein the second upper converter arm (CLAB2; CLAB1A, CLAB1B) comprises submodules for forming the AC voltage and the second DC voltage.

3. The voltage source converter ( 12; 14) according to claim 1, wherein the second upper converter arm comprises circuitry for removing AC components from the third DC terminal (DC3).

4. The voltage source converter ( 12; 14) according to claim 2, wherein the first connection point (CP) is at the AC terminal (ACA1).

5. The voltage source converter ( 12; 14) according to claim 4, further comprising a disconnector (MD) connected between the second and third DC terminals.

6. The voltage source converter ( 12; 14) according to claim 5, further comprising a DC circuit breaker connected between the second and third DC terminals and comprising the disconnector.

7. The voltage source converter ( 12; 14) according to any of claims 1 - 3, wherein the first connection point (CP) is between two submodules of the first upper arm, thereby dividing the first upper converter arm into a common arm (uaO ) between the AC terminal (ACA1) and the connection point (CP) and a first branching arm (ual) between the connection point and the second DC terminal (DC2), thereby making the second upper converter arm into a second branching arm (ua2).

8. The voltage source converter ( 12; 14) according to claim 7, wherein a distribution (x_n) of the AC current between the first and second branching arms depends on the per phase active power being separate from zero.

9. The voltage source converter ( 12; 14) according to claim 8 , wherein the DC voltage contribution for the lower arm (VO) and the common arm (y*V0) are chosen so that a maximum absolute value of an AC current distribution factor (x_n) between the branched arms is kept below or equal to a maximum distribution factor value (x_max), where the distribution factor depends on the active power of the branched arm, the relationship (y) between the DC voltage contribution of the common arm (ucO) and the voltage contribution (Vo) the voltage lower arm (la) and active power of the AC terminal.

10. The voltage source converter ( 12; 14) according to claim 9, wherein the distribution factor is the active power of the branched arm divided by the active power of the AC terminal and by a variable

comprising said relationship (y).

11. The voltage source converter ( 12; 14) according to any of claims 7 - 10 , wherein a voltage difference (p2) to a nominal DC voltage (ud) at the second DC terminal (DC2) is set based on the power at the third DC terminal (DC3) and on at least one scaling of the power (P) at the AC terminal.

12. The voltage source converter ( 12; 14) according to claim 11, wherein the at least one scaling of the power (P) comprises a first and a second scaling of the power (P) and the voltage difference (p2) to the nominal DC voltage (ud) at the second DC terminal (DC2) is set as the first scaled input power divided by a difference between the second scaled input power and the power (Pd l) at the third DC terminal (DC3).

13. The voltage source converter ( 12; 14) according to claim 12, wherein the first scaled input power is scaled with a first scaling factor comprising the fraction (q) of the AC voltage applied across the first branched arm (ual) times the part of the AC current ( l-x) running through the first branched arm (ual).

14. The voltage source converter ( 12; 14) according to claim 13, wherein the second scaled input power is formed through a second scaling factor formed as one minus the first scaling factor.

15. The voltage source converter ( 12; 14) according to any of claims 7- 10 , wherein the power (P) at the AC terminal is shared equally between the branched converter arms (ual, ua2) and a voltage difference (p2) to a nominal DC voltage (ud) at the second DC terminal (DC2) is set as a fraction (q) of the AC voltage applied across the first branched arm (ual) divided by a difference between one half and said fraction .

16. A direct current power transmission system (10 ) comprising a voltage source converter according to any of the previous claims.