Method for feeding electrical power into an electrical supply network

Abstract

A method for feeding electric power into an electrical supply grid by means of a local feed unit. The feed unit is connected to a grid link point connected to a transformer point directly or via a supply connection. The transformer point is connected to a grid section via a transformer. The method includes feeding electrical real power into the electrical supply grid at the grid link point, feeding electrical reactive power into the electrical supply grid at the grid link point, detecting a change to be made in the real power to be fed in, and changing the fed-in real power in accordance with the detected change to be made. The method includes limiting a change in the fed-in reactive power over time when changing the fed-in real power and/or immediately thereafter or temporarily activating voltage control on the basis of the change in the fed-in real power.

Claims

1. A method for feeding electric power into an electrical supply grid using a feed unit, comprising: feeding electrical real power into the electrical supply grid at a grid link point, wherein the feed unit is connected to the grid link point for feeding electric power, the grid link point is connected to a transformer point directly or via a supply connection, and wherein the transformer point is connected to a grid section via a transformer for transmitting the electric power from the transformer point to the grid section via the transformer; feeding electrical reactive power into the electrical supply grid at the grid link point; detecting a change to be made in the real power to be fed in; changing the fed-in real power in accordance with the detected change to be made; and at least one of: limiting a change in the fed-in reactive power over time in response to changing the fed-in real power to counteract a voltage increase at the transformer point or in the grid section; or temporarily activating voltage control based on the change in the fed-in real power, in order to: perform the voltage control at a reference point in response to the change in the fed-in real power, and dynamically correct a voltage at the reference point or grid link point, or bring the voltage at the reference point or grid link point along a trajectory, wherein a power factor control for the feeding-in of electrical reactive power is provisioned, and the power factor control causes the fed-in reactive power to be adjusted based on the fed-in real power such that a stipulated power factor is obtained, and wherein the changing the fed-in real power includes: initially changing the fed-in reactive power concurrently such that the power factor remains unchanged; setting a new value for the power factor based on a voltage that changes as a result of the change in the real power and reactive power or a changed voltage that is to be expected at the grid link point; and reducing the power factor to the set power factor again, wherein the reduction is made with a delay or using a time function.

2. The method as claimed in claim 1, wherein the change in the fed-in reactive power over time is limited such that: the fed-in reactive power is changed in accordance with a change function having a ramp with a slope of limited absolute value, or a gradient limit value is set for the fed-in reactive power, the gradient limit value being a maximum change in an absolute value of the fed-in reactive power over time such that the fed-in reactive power is changed with a temporal gradient having an absolute value that does not exceed the gradient limit value.

3. The method as claimed in claim 1, wherein: the transformer has a primary side to which the transformer point is connected and a secondary side to which the grid section is connected, the transformer is a variable ratio transformer and is configured to adjust a transformation ratio of the primary side to the secondary side to control a voltage level at the transformer point or on the grid section, and the change in the fed-in reactive power over time is limited such that a voltage change at the transformer point or on the primary side resulting from the change in the fed-in reactive power is sufficiently slow to allow the variable ratio transformer to correct a resultant voltage change.

4. The method as claimed in claim 1, wherein limiting a change over time for the fed-in real power is not provisioned.

5. The method as claimed in claim 1, wherein the temporarily activated voltage control is active only during the change in the fed-in real power or substantially immediately thereafter and reduces the voltage at the grid link point from a value that is changed by the change in the fed-in real power wholly or partially to a value that the voltage at the grid link point had immediately before the change in the fed-in real power.

6. The method as claimed in claim 5, wherein the temporarily activated voltage control is performed when the grid link point is connected to the transformer point directly.

7. The method as claimed in claim 1, wherein: power factor control for the feeding-in of electrical reactive power is provided, and the power factor control causes the fed-in reactive power to be adjusted based on the fed-in real power such that a stipulated power factor is obtained, the change in the fed-in real power includes changing the fed-in real power along a ramp or trajectory, the change in the fed-in real power along the ramp or trajectory includes using the voltage control to keep the voltage at the reference point constant at least to counteract a voltage change as a result of the change in the fed-in real power, wherein the voltage control adjusts the reactive power, and the power factor control is deactivated during the voltage control, the change in the fed-in real power is followed by a reactive power value that results from the voltage control being taken to a new reactive power value that would be obtained as a result of the deactivated power factor control, the reactive power value is taken to the new reactive power value via a ramp or trajectory, and the deactivated power factor control is activated in response to the reactive power value reaching the new reactive power value.

8. The method as claimed in claim 1, wherein the fed-in real power is changed using a real power step change, and the reactive power is changed using a time function.

9. The method as claimed in claim 8, wherein the time function is a ramp function.

10. The method as claimed in claim 1, wherein the limiting of the change in the fed-in reactive power over time is dependent on the type or size of the supply connection.

11. The method as claimed in claim 1, wherein the feed unit is at least one wind power installation.

12. The method as claimed in claim 1, wherein the supply connection is configured for transmitting the electric power from the grid link point to the transformer point via the supply connection.

13. The method as claimed in claim 1, wherein the reference point is the grid link point.

14. The method as claimed in claim 1, wherein bringing the voltage at the reference point or the grid link point along the trajectory includes bringing the voltage at the reference point or the grid link point along a ramp.

15. A wind power installation for feeding electric power into an electrical supply grid, comprising: an inverter for feeding electrical real power into the electrical supply grid at a grid link point, wherein the wind power installation is connected to the grid link point for feeding in electric power, the grid link point is connected to a transformer point directly or via a supply connection for the purpose of transmitting the electric power from the grid link point to the transformer point via the supply connection, and the transformer point is connected to a grid section via the transformer for transmitting the electric power from the transformer point to the grid section via the transformer; an inverter controller for controlling a feed of electrical reactive power into the electrical supply grid at the grid link point; and an input interface for detecting a change to be made in the real power to be fed in, wherein the inverter controller is configured to change the fed-in real power in accordance with the detected change to be made, and at least one of: the inverter controller is configured to control or limit a change in the fed-in reactive power over time in response to changing the fed-in real power to counteract a voltage increase at at least one of: the transformer point or the grid section, or the inverter controller is configured to temporarily activate voltage control based on the change in the fed-in real power to perform voltage control at a reference point in response to the change in the fed-in real power, and dynamically correct the voltage at the reference point or grid link point or to bring the voltage at the reference point or grid link point along a trajectory, wherein controlling or limiting the change in the fed-in reactive power over time includes: initially changing the fed-in reactive power concurrently such that the power factor remains unchanged; setting a new value for the power factor based on a voltage that changes as a result of the change in the real power and reactive power or a changed voltage that is to be expected at the grid link point; and reducing the power factor to the set power factor again, wherein the reduction is made with a delay or using a time function.

16. A wind farm, comprising: a plurality of wind power installations, wherein at least one of the plurality of wind power installations is the wind power installation as claimed in claim 15.

17. A feed arrangement for feeding electric power into an electrical supply grid, comprising: a local feed unit; a grid link point; a transformer having a transformer point, wherein: the local feed unit is connected to the grid link point for feeding in the electric power, the grid link point is connected to the transformer point directly or via a supply connection for transmitting the electric power from the grid link point to the transformer point via the supply connection, and the transformer point is connected to a grid section via the transformer for transmitting the electric power from the transformer point to the grid section via the transformer an inverter for feeding electrical real power into the electrical supply grid at the grid link point; an inverter controller for controlling a feed of electrical reactive power into the electrical supply grid at the grid link point; an input interface for detecting a change to be made in the real power to be fed in, wherein: the inverter controller is configured to change the fed-in real power in accordance with the detected change to be made, and wherein at least one of: the inverter controller is configured to limit a change in the fed-in reactive power over time in response to changing the fed-in real power to counteract a voltage increase at at least one of: the transformer point or the grid section, or the inverter controller is configured to temporarily activate voltage control based on the change in the fed-in real power to perform the voltage control at a reference point in response to the change in the fed-in real power and dynamically correct the voltage at the reference point or grid link point or bring the voltage at the reference point or grid link point along a trajectory, wherein limiting the change in the fed-in reactive power over time includes: initially changing the fed-in reactive power concurrently such that the power factor remains unchanged; setting a new value for the power factor based on a voltage that changes as a result of the change in the real power and reactive power or a changed voltage that is to be expected at the grid link point; and reducing the power factor to the set power factor again, wherein the reduction is made with a delay or using a time function.

18. The feed arrangement as claimed in claim 17, wherein the local feed unit is a wind power installation.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

(1) The invention is described in more detail by way of example below on the basis of embodiments with reference to the accompanying figures.

(2) FIG. 1 shows a perspective depiction of a wind power installation.

(3) FIG. 2 shows a schematic depiction of a wind farm.

(4) FIG. 3 shows a schematic depiction of a feed arrangement.

(5) FIG. 4 shows an equivalent circuit diagram for the feed arrangement in FIG. 3.

(6) FIG. 5 shows a graph of a change in feed according to the prior art.

(7) FIG. 6 shows a further graph of a change in feed.

(8) FIG. 7 shows a graph of a change in feed according to a proposed embodiment.

(9) FIG. 8 shows a graph of a change in feed according to a further embodiment.

(10) FIG. 9 shows a graph of a change in feed according to a further embodiment still.

DETAILED DESCRIPTION

(11) FIG. 1 shows a wind power installation 100 having a tower 102 and a nacelle 104. The nacelle 104 has a rotor 106 arranged on it, having three rotor blades 108 and a spinner 110. The rotor 106 is set in a rotary motion by the wind during operation and thereby drives a generator in the nacelle 104. To provide a feed, there is provision for an inverter 130 that can be controlled by means of an inverter controller 132. The inverter controller is moreover provided with an input interface 134 via which setpoint values, particularly a real power setpoint value Ps, can be input.

(12) FIG. 2 shows a wind farm 112 having, in exemplary fashion, three wind power installations 100, which may be identical or different. The three wind power installations 100 are therefore representative of basically any number of wind power installations on a wind farm 112. The wind power installations 100 provide their power, namely in particular the generated current, via an electrical farm grid 114. The respective currents or powers generated by each of the individual wind power installations 100 are added and there is usually provision for a transformer 116 that steps up the voltage on the farm, so as then to feed into the supply grid 120 at the feed point 118, which is also referred to generally as a PCC. There is also provision for a variable ratio transformer 124, which is connected to the feed point 118, which can also be referred to as a grid link point, via a supply connection 122. The variable ratio transformer 124 is also used to step up the voltage further to the voltage in the supply grid 120. The wind power installations 100 may be in a form as shown in FIG. 1, including inverter 130 and inverter controller 132 with input interface 134. There may alternatively be provision for a central controller for the farm 112.

(13) FIG. 2 is just a simplified depiction of a wind farm 112 that does not show a controller, for example, even though a controller is naturally present. The farm grid 114 may for example also be in a different form, for example by virtue of there also being a transformer at the output of each wind power installation 100, to mention just one other exemplary embodiment.

(14) The feed arrangement 300 in FIG. 3 has a wind farm 302 as local feed unit. This wind farm 302 is connected to a grid link point 304 and connected to a transformer point 310 by means of a connecting cable 306, which in this instance forms a supply connection, via a busbar 308. The busbar 308 can also be regarded as part of the transformer point 310 because the busbar 308 basically forms the connection hardware of the transformer point 310. The connecting cable 306 is therefore connected to the transformer point 310.

(15) The connecting cable 306 is designed for example for a voltage level of 20 kV in this instance, and this voltage can be transformed by means of a transformer 312 to a higher voltage, which in this instance is 110 kV, for example, and forms the voltage level of the grid section 314. The transformer 312 is accordingly connected to the grid section 314. The transformer in this instance has a primary side 316 and a secondary side 318.

(16) As an illustration, the connecting cable 306 has a parallel line 320 shown in parallel with it as an illustration, said parallel line likewise being able to connect the grid link point 304 to the busbar 308 or the transformer point 310 when a likewise illustrative isolator switch 322 is closed. Likewise as an illustration, various loads, namely for example industrial loads 324 and nonindustrial loads 326, are indicated in the region of the connecting cable 306 or the parallel line 320, said loads each being able to be connected there. Despite the same reference sign 324 or 326 these loads may nevertheless be different.

(17) For this feed arrangement 300 shown there is now provision for a method in which the wind farm as a representative local feed unit feeds in both real power and reactive power at the grid link point 304. The fed-in real power can then be changed in accordance with an applicable stipulation, that is to say a change to be made. By way of example it can be halved given an applicable stipulation. At the same time a change in the fed-in reactive power over time is limited. This is achieved by counteracting a voltage increase at the transformer 312, namely at the transformer point 310 and/or at the grid link point 304. This does not have to mean that a voltage increase is precluded here, but the absolute value thereof is at least reduced in comparison with a voltage increase that would be obtained without such a change in the fed-in reactive power over time.

(18) The effect that can be achieved thereby is that unfavorable or even incorrect parameterization of cos(φ) feedback control or open-loop control results in voltage step changes at the transformer and/or at the grid link point being limited. Such unfavorable parameterization can occur for example on the basis of a reactive power budget, for example if a reactive power feed or effect is incorrectly assessed, or reactive power cannot be delivered as provided for.

(19) In the case of correct, optimum or ideal parameterization, the cos(φ) at the grid link point would be adjusted such that the voltage would barely change in the event of a power change. If a wind farm is connected to a busbar directly, like the one on the busbar 310, a purely inductive character would ideally need to be assumed and cos(φ) would be parametrized to unity. Divergences from this ideal state regularly occur, however, which are allowed for here.

(20) FIG. 4 now shows an equivalent circuit diagram by way of example for the feed arrangement 300 in FIG. 3 or for a portion thereof. The equivalent circuit diagram in FIG. 4 fundamentally initially assumes that the isolator switch 322 is open as shown in FIG. 3. If it were closed, however, a quite similar equivalent circuit diagram would be obtained, but with different specific values, namely with different specific impedances or at least one different impedance.

(21) Accordingly, FIG. 4 initially again shows the wind farm 302, which provides a feed at the grid link point 304. The connection section for the busbar 308 and hence for the transformer point 310 has a connection impedance 406. This connection impedance 406 is referred to as impedance Z.sub.3. Very specific values are indicated here, to which the results of the graphs shown below also relate. These values cited and explained below are cited only by way of example, however, and may have either small or large divergences therefrom.

(22) At any rate the exemplary connection impedance 406 has a resistive component, that is to say real component, of 5.78Ω and a reactance value, that is to say reactive component, of 4.83Ω. As already explained, the exact values do not matter, but it should be stated that the connection impedance 406 has a real component and a reactive component of identical magnitude in this case, for example. A real current and a reactive current each of identical amplitude would therefore lead to identical voltage drops across this connection impedance 406, for example.

(23) The busbar 308 or the transformer point is followed by a transformer impedance 412 and a grid impedance 414 of a superimposed grid. The grid or the grid section 314, transformed to the primary side 316 of the transformer 312, can be regarded as a voltage source 415 having the voltage level 20 kV. Both the transformer impedance 412 and the grid impedance 414 are therefore impedances that are transformed to the low voltage side, that is to say the voltage side of the primary side 316 of the transformer 312. The transformer impedance 412 accordingly represents a transformed impedance of the transformer. The grid impedance 414 represents a transformed impedance of the grid section 314.

(24) Exemplary values are indicated in this case too, allowance having been made for them below, but they may also end up differently. At any rate the values of the transformer impedance 412 are 0.14Ω for the resistive component and 3Ω for the component of the reactance. The grid impedance 414 has a resistive component of 0.07Ω and a reactance component of 0.4Ω. It can be seen that particularly the transformer impedance 412 has a very much higher reactance value than the value of the resistance. In this case there is a factor of 20 between the two components, for example. It can therefore be seen that a change in the reactive current through this transformer impedance 412 results in a substantially larger voltage change than a change in the real current. Accordingly, an undesirable voltage change can occur particularly, but not only, at the transformer impedance 412 if real power and reactive power fed in by the wind farm 302 are changed in equal measure.

(25) In principle, the feeding-in of real and reactive power, which also includes the negative feeding-in of reactive power, may be selected such that—in illustrative terms—the reactive power counteracts a voltage increase as a result of the real power. This applies at any rate at the grid link point if a supply connection that dominates here, that is to say a dominating cable impedance of the supply connection, is present. If both values are now changed to the same extent, however, this can lead to a new situation at another point in which this compensation is no longer available. This has been recognized particularly as a problem for wind farms in which the grid link point is connected to a transformer, in particular a substation, directly or with a weakly dominant supply connection. This applies particularly if the grid link point is connected to the transformer directly, or is even located in the substation, and if the wind farms are under cos(φ) feedback control. This may be the case for example if applicable control of the reactive power budget requires this. It has been recognized that in such a case the real current does not have much of a voltage-increasing effect, but the reactive current has a voltage-lowering effect. If the reactive current is absent in such a case, the power increase results in a sudden increase in the voltage.

(26) The graphs below illustrate various options for changing the real and reactive powers and show possible effects resulting therefrom.

(27) FIG. 5 shows a graph showing a known type of the variation.

(28) The graph in FIG. 5 plots power and voltage characteristics over time. A stable feed situation is initially assumed in that case, until, at the time t.sub.R, a reduction of real power is stipulated and then also implemented. In this respect the graph shows a fed-in farm real power P.sub.WEA, which in that case is reduced for example at the reduction time t.sub.R from for example 10 MW to approximately 3 MW. The reduction fundamentally takes place immediately, which means that the graph in FIG. 5 approximately shows a step-like characteristic for the reduction of the real power P.sub.WEA. On the basis of a firmly stipulated phase angle, the fed-in reactive power Q.sub.WEA is likewise reduced at the reduction time t.sub.R, namely also fundamentally in step-like fashion. The graph then shows divergences between the real power P.sub.WEA and the reactive power Q.sub.WEA, which are sometimes also related to the representation, however, and are less relevant here. At any rate the fed-in reactive power Q.sub.WEA is also reduced quickly together with the real power.

(29) As a result or as related consequences, both the voltage V.sub.NVP at the grid link point 304 and the voltage V.sub.SS on the busbar 308, which in turn corresponds to the voltage at the transformer point 310, can change.

(30) It can be seen that the reduction of the real power P.sub.WEA and of the reactive power Q.sub.WEA leads to a reduction of the voltage U.sub.NVP at the grid link point. This is related particularly to the fact that the connection impedance 406 has a larger resistive component in comparison with the reactive power component. A decrease in the real power therefore leads to a voltage reduction, which is counteracted by the simultaneous reduction of the reactive power. Since the reactance component of the connection impedance 406 is smaller than the resistive component, however, the simultaneous reactive power reduction cannot counteract the lowering of the voltage on account of the real power reduction completely. In the event of a phase angle being adjusted for this grid link point in optimum fashion, the voltage change at the grid link point (NVP) would be minimal, however.

(31) The response in the case of the voltage V.sub.SS on the busbar is somewhat different, because particularly the transformer impedance 412 but also, albeit to a slightly lesser extent, the grid impedance or transformed grid impedance 414 each have a significantly larger reactance component in comparison with the resistive component. The voltage-increasing effect as a result of the reactive power reduction is greater in that case than the voltage-lowering effect as a result of the real power decrease. For this reason, the voltage on the busbar V.sub.SS rises undesirably at the reduction time t.sub.R.

(32) A particular problem in this instance is that this voltage rise accompanies the real power reduction and reactive power reduction so quickly, namely immediately, that a variable ratio transformer cannot compensate for this quickly enough. Particularly the transformer 312 may be in the form of a variable ratio transformer in this case in order to counteract a slow voltage rise at the transformer point 310.

(33) The first countermeasure proposed in the prior art is to reduce the real power P.sub.WEA slowly so that the reactive power Q.sub.WEA is also reduced slowly. This is shown in FIG. 6.

(34) The graph in FIG. 6, however, reveals that to reach a target value for the real power P.sub.WEA that is to be reduced, said real power needing to be reached at the reduction time t.sub.R, the reduction needs to start earlier. For this, the reduction of the real power is already started at the lead time t.sub.V. In fact, the effect achieved thereby is that the voltage at the grid link point V.sub.NVP and the voltage on the busbar also change correspondingly slowly. In this case there is provision for a period of 400 seconds from the lead time t.sub.V to the reduction time t.sub.R. The changes shown for the voltage, particularly the change shown for the voltage on the busbar V.sub.SS, is then sufficiently slow for a variable ratio transformer to be able to compensate for this voltage change. However, this is achieved with an excessively early power reduction, which, given the 400 seconds cited by way of example, starts 6.5 minutes too early and hence leads to losses. These losses can be illustrated by the indicated energy triangle 630.

(35) To avoid this, a slow reduction of the reactive power is proposed according to one embodiment, as illustrated by FIG. 7.

(36) According to this embodiment, which is illustrated in the graph in FIG. 7, it is therefore proposed that at the reduction time t.sub.R the real power P.sub.WEA fed in by the wind farm be reduced in step-like fashion again, whereas the fed-in reactive power Q.sub.WEA is initially kept constant. This leads to a larger inductive cos(φ) being obtained. The voltage at the grid link point V.sub.NVP therefore dips sharply at the grid link point 304.

(37) At any rate the voltage V.sub.SS on the busbar 308 dips only a little, however, since the configuration of the transformer impedance 412 and transformed grid impedance 414 means that said voltage has only little dependency on the real power and therefore undergoes only a very small dip as a result of the step-like reduction of the fed-in real power P.sub.WEA.

(38) The fed-in reactive power Q.sub.WEA is then lowered according to a ramp or linear function from the reduction time t.sub.R onward, however. Both the voltage V.sub.NVP at the grid link point 304 and the voltage V.sub.SS on the busbar 308 therefore increase gradually.

(39) The sharp dip in the voltage V.sub.NVP at the grid link point 304 is acceptable and can be increased again as a result of the gradual change in the reactive power Q.sub.WEA from the reduction time t.sub.R onward. The characteristic of said voltage is not a particular problem for the electrical supply grid, particularly the grid section 314. The increase in the voltage V.sub.SS on the busbar 308, however, is so slow that a variable ratio transformer can compensate for this change. The result of this is only a slight effect on the electrical supply grid; in particular, the voltage at the transformer point 310 can be kept substantially constant.

(40) In this case too, the gradual change in the reactive power fundamentally achieves a gradual change in the voltage V.sub.SS on the busbar that can be compensated for by a variable ratio transformer. However, the solution proposed here manages to nevertheless have the real power initially reduced in step-like fashion at the reduction time t.sub.R, with no loss of yield therefore. In this instance, it has particularly also been recognized that the voltage reactions at the grid link point 304, on the one hand, and a busbar 308, on the other hand, to real power changes and reactive power changes are different. Particularly if the grid link point is connected to the busbar directly and the wind farm uses cos(q) feedback control or is under such, advantages arise for the cited method. Specific allowance has been made for this and therefore in particular a solution that avoids an abrupt large voltage step change on the busbar and then avoids a corresponding voltage step change on the secondary side 316 of the transformer 312 has been proposed. As the grid voltage can be regarded as approximately constant, undesirable voltage differences between the primary and secondary sides of the transformer are avoided.

(41) FIG. 8 shows a variant that is similar to the variant in FIG. 7. In this case too the real power P.sub.WEA is reduced at the reduction time t.sub.R. The reactive power Q.sub.WEA is likewise reduced at the reduction time t.sub.R in this case, however. As a result, the voltage V.sub.SS on the busbar could likewise increase suddenly, as was the case according to the variant in FIG. 5. In this instance, however, additional reactive power control is now performed, particularly by an additional apparatus. This allows additional reactive power to be fed in or provided, which is referred to as Q.sub.Z in this case. This additional reactive power Q.sub.Z allows this voltage step change in the voltage V.sub.SS on the busbar to be counteracted.

(42) In particular, additional reactive power control that feeds in precisely enough reactive power, and in an appropriate manner, for the voltage V.sub.SS on the busbar not to rise as a result of the fall in the real power P.sub.WEA and in particular as a result of the parallel drop in the reactive power Q.sub.WEA is performed in this case. This additional reactive power control thus counteracts this. The additional reactive power control can provide for the additional reactive power Q.sub.Z to be reduced gradually again after the reduction time t.sub.R, particularly with a falling ramp, or to be reduced correspondingly linearly. This then leads to a slow increase in the voltage V.sub.SS on the busbar, which occurs so slowly, however, that a variable ratio transformer can counteract it. The voltage on the primary side 316 of the transformer 312 can therefore be maintained. This allows the additional reactive power control then to be returned to a value of 0, which means that no additional reactive power is then fed in.

(43) The additional reactive power in the reactive power control can fundamentally also be provided by the inverters of the wind power installations. The control value can be added to a reactive power value of a phase angle control loop. The setpoint value can be stipulated, e.g., by a central farm control system, which can also be referred to as an FCU, or a remote control terminal, which is also referred to as an RTU, whereas the phase angle is adjusted by a closed-loop or open-loop control system of a wind power installation.

(44) Further variants proposed are that the control element used for generating additional reactive power is another wind power installation or another wind farm, a STATCOM or other compensating mechanisms.

(45) A disadvantage in this context may particularly be that such additional reactive power control is sophisticated.

(46) In the case of wind farms that are not connected to the busbar directly, it can be a conflict of interests if these voltage changes need to be limited both at the grid link point and on the busbar.

(47) In this case it may be critical if the wind farm uses stipulation of a cos(q) reactive power for feedback control. Control of the reactive power can also be referred to as control of the reactive power budget of the wind farm. A particularly critical case is when the reactive power control, particularly the stipulated power factor, that is to say cos(φ), does not or does not sufficiently accurately match the system in this case. Allowance can also be made for a reactive power budget in this instance, for example if a grid operator stipulates a reactive power budget.

(48) In FIG. 9 the voltage at the grid link point (NVP) is controlled temporarily. This indirectly results in a cos(φ) that matches the grid link point, which can also be referred to as a correct cos(φ).

(49) This can have significant effects on the voltage on the busbar in the case of the exemplary topology shown in FIG. 3.

(50) If the farm, unlike in the exemplary topology in FIG. 3, is connected to the busbar directly, the method proposed according to FIG. 9 leads to particular advantages.

(51) FIG. 9 shows a variant in which feedback control is used that controls the voltage V.sub.NVP at the grid link point 304. In this regard, it is assumed here too that the fed-in real power P.sub.WEA is reduced suddenly at the reduction time t.sub.R. However, there is then provision for feedback control that initially maintains the voltage V.sub.NVP at the grid link point. This is accomplished by lowering the fed-in reactive power Q.sub.WEA at the reduction time t.sub.R accordingly. The cos(φ) changes as a result. In order to get the cos(φ) to the earlier value again, the reactive power is then slowly raised, in particular linearly, to a setpoint value that corresponds to the initial cos(φ) stipulation, or which then produces the cos(φ), shortly after the reduction time t.sub.R.

(52) As a result of this measure, the voltage V.sub.NVP at the grid link point is initially maintained, but then lowered, likewise for example in ramp-like fashion, that is to say particularly linearly, to a later value, which the voltage V.sub.NVP at the grid link point then maintains, specifically as soon as the previous cos(φ) has been reached again. A new setpoint value for the cos(φ) could naturally also be stipulated in this case if desirable for other reasons.

(53) At the same time, however, the voltage V.sub.SS on the busbar undergoes a step change at the reduction time t.sub.R as a result of this measure. It too is then reduced again accordingly as a result of the gradual linear change in the fed-in reactive power Q.sub.WEA, however.

(54) According to one variant, it is proposed that both step changes, that is to say the step change in the voltage V.sub.SS on the busbar and the reactive power step change, be limited to a stipulated amount. In this regard, a small reactive power step change can be permitted in order to subsequently reduce the reactive power by means of a ramp. The effect that can be achieved thereby is that, instead of a large step change in the voltage V.sub.SS on the busbar, only a small step change occurs there, while a small voltage step change is additionally accepted at the grid link point. The voltage therefore changes in different directions at these two points, but the change is split over two small changes.

(55) It should be borne in mind, however, that the connection impedance 406 may also be in a totally different form. As an alternative to the topology illustrated in FIG. 3, the grid link point could also be connected to the busbar without or without significant impedance, that is to say for example without or only with a negligibly long connecting cable. In that case, the voltage V.sub.SS on the busbar would accordingly behave like the voltage V.sub.NVP at the grid link point, since the grid link point and the busbar would be electrically connected in identical, or at least almost identical, fashion. The voltage V.sub.SS on the busbar would then accordingly likewise not change suddenly, but rather only gradually, as also shown for the voltage V.sub.NVP at the grid link point in FIG. 9. A variable ratio transformer could accordingly counteract this gradual voltage drop.

(56) In this respect, however, FIG. 9 illustrates that such control for maintaining the voltage V.sub.NVP at the grid link point is not always advisable, but rather is dependent on the specific situation, namely also on the transmission response from the grid link point to the busbar.