METHOD FOR ATTENUATING LOW-FREQUENCY OSCILLATIONS IN AN ELECTRICAL POWER SUPPLY GRID

Abstract

A method for attenuating low-frequency oscillations in an electrical power supply grid by means of a feed device which feeds into the electrical power supply grid, in particular a wind power installation, wherein the electrical power supply grid has a grid voltage and a grid frequency, comprising the following steps: picking up a grid signal having the low-frequency oscillations, splitting a total frequency range of the grid signal in which oscillations to be attenuated are to be expected into a plurality of partial frequency ranges, each having a lower and an upper range frequency, performing in each case one frequency analysis of the grid signal for each partial frequency range in order to identify in each case one or more oscillations having an oscillation frequency in the partial frequency range, if present, identifying a low-frequency oscillation to be attenuated as target oscillation depending on the frequency analyses of all of the partial frequency ranges, detecting the target oscillation at least according to frequency and amplitude and optionally according to phase, determining a setpoint attenuation signal depending on the target oscillation detected according to frequency and amplitude and possibly phase for attenuating the detected target oscillation, generating a setpoint feed signal depending on the setpoint attenuation signal and a basic setpoint signal, and generating and feeding in a feed signal depending on the setpoint feed signal (QE).

Claims

1. A method comprising: attenuating low-frequency oscillations in an electrical power supply grid by a feed device which feeds into the electrical power supply grid, wherein the electrical power supply grid has a grid voltage and a grid frequency, wherein the low-frequency oscillations are less than 5 hertz, the attenuating comprising: detecting a grid signal having the low-frequency oscillations, splitting a total frequency range of the grid signal in which oscillations to be attenuated are to be expected into a plurality of partial frequency ranges, each partial frequency range having a lower range frequency and an upper range frequency, performing a frequency analysis of the grid signal for each partial frequency range to identify in each case one or more oscillations having an oscillation frequency in the partial frequency range, if present, identifying a low-frequency oscillation to be attenuated as target oscillation depending on the frequency analyses of all of the plurality of partial frequency ranges, detecting the target oscillation at least according to frequency and amplitude, determining a setpoint attenuation signal depending on the target oscillation detected according to frequency and amplitude for attenuating the detected target oscillation, generating a setpoint feed signal depending on the setpoint attenuation signal and a basic setpoint signal, and generating and feeding in a feed signal depending on the setpoint feed signal.

2. The method as claimed in claim 1, wherein: the setpoint attenuation signal describes a reactive power to be fed in, and the basic setpoint signal is preset by power factor correction as setpoint signal for a reactive power to be fed in.

3. The method as claimed in claim 1, wherein: to generate the setpoint feed signal, a core controller is provided which outputs a controller output signal depending on the basic setpoint signal and the fed-in feed signal, the setpoint feed signal is determined depending on the controller output signal, and the setpoint attenuation signal is: injected onto the basic setpoint signal and taken into consideration by the core controller, and/or injected onto the controller output signal and influences the setpoint feed signal, and/or wherein an attenuation compensation signal, which is dependent on the setpoint attenuation signal, is injected on an input side of the core controller, and is injected onto the basic setpoint signal to at least partially compensate for an influence of the setpoint attenuation signal, via the fed-back feed signal, on the core controller.

4. The method as claimed in claim 3, wherein: injecting the attenuation compensation signal on the input side of the core controller, and injecting the fed-in feed signal onto the basic setpoint signal, the attenuation compensation signal is generated by filtering out a compensation signal component from the detected feed signal, the compensation signal component has an oscillation frequency of the identified target oscillation, and the attenuation compensation signal is formed depending on the compensation signal component, and to filter out the attenuation compensation signal from the detected feed signal, using a bandpass filter and setting the bandpass filter to the oscillation frequency of the identified target oscillation.

5. The method as claimed in claim 1, wherein: at least three overlapping partial frequency ranges are provided, the upper range frequency of a partial frequency range is in a region of 1.5 times to 10 times a value of the lower range frequency of the same partial frequency range, the respective frequency analysis for each partial frequency range uses different time segments of the detected grid signal and has different scanning rates, a time segment of the grid signal is assigned to each partial frequency range for the evaluation, wherein time segments of a plurality of partial frequency ranges overlap one another, and a scanning rate has been assigned to each partial frequency range for performing the frequency analysis, and/or a number of scans per time segment is used which is identical for different partial frequency ranges, the duration of the time segment of the partial frequency range corresponds at least to half inverse value of the lower range frequency, and/or corresponds at most to five times the inverse value of the lower range frequency, and/or the scanning rate of, in each case, one partial frequency range corresponds to at least twice the upper range frequency, and/or the scanning rate of, in each case, one partial frequency range corresponds up to one hundred times the upper range frequency.

6. The method as claimed in claim 1, wherein: when, in a first of the partial frequency ranges having a higher upper range frequency than a further one of the partial frequency ranges, an oscillation having an oscillation frequency has been identified, the oscillation frequency of the identified oscillation is considered as potential aliasing frequency, and for the frequency analysis of at least a second of the partial frequency ranges having a lower upper range frequency than in a case of the first partial frequency range, a filtered signal of the grid signal is used from which signal components having the aliasing frequency are filtered out.

7. The method as claimed in claim 1, wherein: in a first analysis step, the frequency analysis is performed for each partial frequency range to identify at least one oscillation having a first frequency, in a second analysis step, frequency, amplitude, and phase of the oscillation are identified by a signal investigation which is adapted to the identified oscillation, and the adapted signal investigation of the identified oscillation has a higher resolution than the frequency analysis and/or has been tuned in a targeted manner to the first frequency identified in the first analysis step.

8. The method as claimed in claim 1, wherein: to determine the setpoint attenuation signal, an input signal which corresponds to the picked-up grid signal or is derived therefrom is filtered using a bandpass filter, the bandpass filter is set depending on the identified target oscillation to allow, from the input signal, a signal component having the oscillation frequency of the identified target oscillation to pass to only allow the target oscillation from the input signal to pass as extracted grid oscillation, and converting the extracted grid oscillation into the setpoint attenuation signal.

9. The method as claimed in claim 8, wherein: to convert the extracted grid oscillation into the setpoint attenuation signal, at least one conversion element is used from the list comprising: a low-pass filter for filtering out a noise component, an amplifying element for amplifying the extracted grid oscillation, a high-pass filter for filtering out low-frequency signal components which occurs in the case of feeding wind power installations as a result of fluctuations in the wind, and at least one lead-lag filter for compensating for communications-related delay times.

10. The method as claimed in claim 1, wherein: the determination and/or injection of the setpoint attenuation signal is activated or deactivated depending on a property of the identified target oscillation depending on an amplitude of the identified target oscillation, to activate an activation threshold and to deactivate a deactivation threshold is provided in each case as comparison value for the amplitude of the target oscillation, and the activation threshold is greater than the deactivation threshold.

11. The method as claimed in claim 1, wherein: to inject the attenuation signal, an activation function is generated which ramps up the setpoint attenuation signal to be injected in a controlled manner at the beginning of the injection, and/or to end the injection of the attenuation signal, a deactivation function is used which ramps down the injected setpoint attenuation signal in a controlled manner in order to end the injection.

12. The method as claimed in claim 1, wherein: in the case of a plurality of identified oscillations, one is selected as target oscillation depending on an amplitude in such a way that, of a plurality of identified oscillations, that which has the highest amplitude is selected as target oscillation, the setpoint attenuation signal is generated depending on the selected target oscillation, and after the activation of the determination and/or injection of the attenuation signal, the target oscillation is maintained, and only after a deactivation of the determination or injection is one of the identified oscillations selected as target oscillation.

13. The method as claimed in claim 1, wherein: the total frequency range is preset, and the splitting the total frequency range into the plurality of partial frequency ranges occurs after presetting of the total frequency range.

14. The method as claimed in claim 1, wherein: the feeding in the feed signal comprises using a wind farm comprising a plurality of wind power installations, each of the plurality of wind power installations generate a part of the feed signal, and each of the plurality of wind power installations takes into consideration the same target oscillation for generating the feed signal.

15. The method as claimed in claim 1, wherein the detecting the target oscillation is further according to a phase, wherein the determining the setpoint attenuation signal further depends on the phase.

16. A wind power system, comprising: one or more wind power installations configured to: perform a method for attenuating low-frequency oscillations in an electrical power supply grid, wherein the electrical power supply grid has a grid voltage and a grid frequency, comprising: a measuring sensor for picking up a grid signal having the low-frequency oscillations, and a controller configured to; split a total frequency range of the grid signal in which oscillations to be attenuated are to be expected into a plurality of partial frequency ranges, each having a lower, an upper, and a middle range frequency, perform in each case one frequency analysis of the grid signal for each partial frequency range to identify in each case one or more oscillations having an oscillation frequency in the partial frequency range, if present, perform the identification of a low-frequency oscillation to be attenuated as target oscillation depending on the frequency analyses of all of the partial frequency ranges, perform a detection of the target oscillation at least according to frequency and amplitude and optionally according to phase, perform a determination of a setpoint attenuation signal depending on the target oscillation detected according to frequency and amplitude and possibly phase for attenuating the detected target oscillation, perform a generation of a setpoint feed signal depending on the setpoint attenuation signal and a basic setpoint signal, and perform a generation and feed a feed signal depending on the setpoint feed signal.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0121] The invention will now be explained in more detail below using embodiments by way of example with reference to the accompanying figures.

[0122] FIG. 1 shows a wind power installation in a perspective illustration.

[0123] FIG. 2 shows a wind farm in a schematic illustration.

[0124] FIG. 3 shows a graph for a possible frequency splitting.

[0125] FIG. 4 shows a closed-loop attenuation control structure for determining a setpoint attenuation signal.

[0126] FIG. 5 shows a structure comprising conversion components as part of the closed-loop control structure shown in FIG. 4.

[0127] FIGS. 6 to 8 show alternative total closed-loop control structures for injecting a setpoint attenuation signal.

DETAILED DESCRIPTION

[0128] FIG. 1 shows a schematic illustration of a wind power installation in accordance with the invention. The wind power installation 100 has a tower 102 and a nacelle 104 on the tower 102. An aerodynamic rotor 106 having three rotor blades 108 and a spinner 110 is provided on the nacelle 104. The aerodynamic rotor 106 is set in rotary motion by the wind during operation of the wind power installation and therefore also rotates an electrodynamic rotor of a generator, which is coupled directly or indirectly to the aerodynamic rotor 106. The electric generator is arranged in the nacelle 104 and generates electrical energy. The pitch angles of the rotor blades 108 can be changed by pitch motors at the rotor blade roots 109 of the respective rotor blades 108.

[0129] The wind power installation 100 in this case has an electric generator 101, which is shown within the nacelle 104. Electrical power can be generated by means of the generator 101. In order to feed in electrical power, a feed unit 105 is provided which can in particular be in the form of an inverter. Thus, a three-phase feed current and/or a three-phase feed voltage can be generated according to amplitude, frequency and phase for feeding in at a point of common coupling PCC. This can take place directly or else together with other wind power installations in a wind farm. In order to control the wind power installation 100 and also the feed unit 105, an installation controller 103 is provided. The installation controller 103 can also receive preset values from the outside, in particular from a central farm computer.

[0130] FIG. 2 shows a wind farm 112 having, by way of example, three wind power installations 100, which can be identical or different. The three wind power installations 100 are therefore representative of in principle any desired number of wind power installations of a wind farm 112. The wind power installations 100 provide their power, namely in particular the generated current, over an electrical farm grid 114. In this case, the respectively generated currents or powers of the individual wind power installations 100 are added up, and usually a transformer 116 is provided which steps up the voltage in the farm in order to then feed it into the power supply grid 120 at the feed point 118, which is also referred to generally as PCC. FIG. 2 is only a simplified illustration of a wind farm 112. It is possible for the farm grid 114 to have a different configuration by virtue of, for example, a transformer also being provided at the output of each wind power installation 100 in order to name but one other exemplary embodiment.

[0131] The wind farm 112 in addition has a central farm computer 122, which can also be referred to synonymously as central farm controller. This can be connected to the wind power installations 100 via data lines 124 or wirelessly in order to thereby exchange data with the wind power installations and in particular to receive measured values from the wind power installations 100 and to transmit control values to the wind power installations 100.

[0132] FIG. 3 shows a graph for illustrating possible frequency splits with a frequency ray 300 on which possible oscillation frequencies are plotted in Hz.

[0133] By way of example, a first frequency range 301 and a second frequency range 302 are illustrated. These first and second frequency ranges 301 and 302 characterize ranges in which the detection and attenuation of a low-frequency oscillation is of interest. Such frequency ranges can vary depending on the electrical power supply grid or grid section of an electrical power supply grid.

[0134] In order to detect at least one low-frequency oscillation, splitting into a plurality of in particular overlapping partial frequency ranges is proposed. FIG. 3 shows for this, by way of example, a first, second and third partial frequency range 311, 312 and 313.

[0135] The first two partial frequency ranges 311 and 312 cover the first frequency range 301, and all three partial frequency ranges 311 to 313 cover the second frequency range 302. However, it is also possible for all three partial frequency ranges 311 to 313 to be used for detecting a low-frequency oscillation for the first range 301. If, in this case, a low-frequency oscillation is detected outside of the first range 301, this can be discarded or does not need to be taken into consideration any further.

[0136] For the exemplary three partial frequency ranges 311 to 313 shown in FIG. 3, in each case different time windows are provided for the detection, and at the same time different resolutions are provided. The number of scans per window, and therefore per analysis, can be identical, however. For the first partial frequency range 311, a time window having a length of 25 seconds and a resolution of 0.5 second is provided. For the second partial frequency range 312, a time window having a length of 12.5 seconds and a resolution of 0.25 second is provided. For the third partial frequency range 313, a time window having a length of 5 seconds and a resolution of 0.1 second is provided.

[0137] Therefore, it is possible in particular for a relatively short time window to be provided for ranges having a relatively high frequency in order to thus also detect a low-frequency oscillation correspondingly quickly. Varying the resolution additionally makes it possible for there to be no need for an excessively large computation capacity, which would be the case if the resolution of a small time window for a high frequency range were to be maintained in the case of a long time window for a low frequency range. The proposed solution avoids this.

[0138] FIG. 4 shows a closed-loop attenuation control structure 400 for determining a setpoint attenuation signal for attenuating a low-frequency oscillation.

[0139] For this purpose, the closed-loop attenuation control structure 400 has a signal input 402 at which the grid voltage or a grid signal having the low-frequency oscillations is input. It is therefore possible for the grid voltage to be directly detected and input there, i.e., a picked-up grid signal. Preferably, the grid voltage is picked up by the detection of three conductor-to-ground voltages of the three-phase system. From this, a phasor is calculated, and of this phasor only the amplitude is taken into consideration. The rated voltage, i.e., a fixed value, can be subtracted from the amplitude of this phasor. Then there remains only a difference between the amplitude of the phasor and the fixed grid voltage value. This difference can be input at the signal input 402.

[0140] This difference is therefore an input signal or grid signal which still has the low-frequency oscillations since only a constant value was subtracted. This input signal is then passed via the signal process block 404, which can perform a first filtering in order to filter out in particular a measurement noise. The signal process block 404 can therefore have in particular low-pass responses.

[0141] Therefore, a prefiltered or preprocessed signal Si is output at the output of the signal process block 404 and input into one of the analysis blocks 411 to 413. Preferably, a grid fault, for example a voltage dip, can be identified in the signal process block 404, or in another component. If a grid fault is identified, a signal detected prior to the grid fault or a signal or value representative thereof, such as, for example, a mean value, can be used during the grid fault instead of the detected or measured signal. This relates in particular to a very short grid fault in the range of 100 to 500 ms. In this case, it has been recognized in particular that the proposed attenuation of oscillations is usually important after a grid fault. By virtue of the proposed bypassing of the voltage dip during the measurement, it is possible to achieve a situation whereby good detection of the low-frequency oscillation is present quickly after the fault.

[0142] In order to identify low-frequency oscillations in the prefiltered signal S.sub.1, therefore, a plurality of analysis blocks is provided, namely in this case a first, second and third analysis block 411, 412 and 413, respectively. Each of these three analysis blocks performs a frequency analysis in each case for one partial frequency range. For this purpose, each of the three analysis blocks 411 to 413 has a dedicated time window and a dedicated resolution. This could take place, for example, for three different partial frequency ranges 311 to 313 having the corresponding window lengths and resolutions, as has been explained in FIG. 3.

[0143] The result of each of the three analysis blocks 411 to 413 can be an identified low-frequency oscillation. However, it is naturally also possible for in each case no low-frequency oscillation to be present, and therefore also that no such oscillation can be identified. It is also possible that in each case a plurality of low-frequency oscillations is identified in one or more of the analysis blocks.

[0144] In a case which is also used for explanatory purposes here, in each case one low-frequency oscillation is detected in each of the three analysis blocks 411 to 413. The respective analysis block 411 to 413 outputs frequency and amplitude for each detected low-frequency oscillation and passes these values on to a coordination unit 406.

[0145] In the coordination unit 406, these identified low-frequency oscillations are evaluated. First, an evaluation is performed to ascertain whether they are in the frequency range under consideration at all. If, therefore, for example with reference to FIG. 3, a low-frequency oscillation is in the region of 1.2 Hz but only the first frequency range 301 is relevant, which reaches only from 0.25 to 1 Hz, this low-frequency oscillation is discarded or not take into any further consideration.

[0146] Of the remaining low-frequency oscillations, one is identified as target oscillation and only this one is then taken into consideration any further. It is possible in particular for that low-frequency oscillation which has the greatest amplitude of all of the identified and relevant low-frequency oscillations to be considered as target oscillation. However, it is also possible for another or a further criterion to be considered. For example, a quotient of amplitude and oscillation frequency can be formed for each low-frequency oscillation, and the low-frequency oscillation with the largest quotient calculated hereby can be identified as the target oscillation, in order to name a further example.

[0147] It is naturally also possible for all three analysis blocks 411 to 413, which is also representative of a different number of analysis blocks, to identify in total only one single low-frequency oscillation which then therefore automatically forms the target oscillation.

[0148] Of the thus identified target oscillation, its oscillation frequency is output as oscillation frequency f.sub.PSOD of the low-frequency oscillation to be attenuated at the frequency output 408. This frequency f.sub.PSOD can be referred to simply as attenuation frequency f.sub.PSOD.

[0149] This attenuation frequency f.sub.PSOD is then input into a bandpass filter block 410. The bandpass filter block 410 is then set corresponding to the input attenuation frequency f.sub.PSOD, namely in particular in such a way that the attenuation frequency f.sub.PSOD forms the bandpass filter frequency.

[0150] The bandpass filter block 410, which has now been set to the attenuation frequency f.sub.PSOD, can therefore receive the prefiltered signal S.sub.1 as input signal. For this purpose, the coordination unit 406 can output an initialization trigger Ti, which namely then has the value 1. Therefore, the prefiltered signal S.sub.1 is multiplied by the initialization trigger Ti in the input multiplier 414. The prefiltered signal S.sub.1 is therefore as a result passed through to the bandpass filter block 410. In particular, provision is made for the initialization trigger Ti to only be able to assume the values 0 or 1. Prior to the initialization, the coordination unit 406 therefore only outputs the value 0 as initialization trigger Ti.

[0151] The prefiltered signal S.sub.1 is then passed through the bandpass filter block 410 and it naturally also has the low-frequency oscillation identified as dominant. The bandpass filter block 410 or the bandpass filter implemented therein is now precisely adapted to the dominant low-frequency oscillation to be attenuated, which results in that, of the prefiltered signal S.sub.1 which initially still has all of the frequency components, now only the target oscillation to be attenuated is passed through, and everything else is filtered out. The bandpass filter block 410 therefore outputs the oscillation signal Sd to be attenuated and in the process passes it on to the determination controller block 416.

[0152] The determination controller block 416 converts the oscillation signal Sd to be attenuated into a setpoint attenuation signal Q.sub.PSOD.

[0153] The determination controller block 416 can in this case change in particular amplitude and phase of the oscillation signal Sd to be attenuated and possibly perform further conversions such as filtering. Details in this regard are explained below with reference to FIG. 5, which in principle shows the inner structure of the determination controller block 416.

[0154] The setpoint attenuation signal Q.sub.PSOD is intended to be injected onto a further signal for the attenuation in order to generate thereby a setpoint feed signal which contains this setpoint attenuation signal. Details in this regard are described further below. First, provision is made, however, for the setpoint attenuation signal Q.sub.PSOD generated by the determination controller block 416 to be capable of being ramped up slowly in terms of its amplitude for activation purposes, in particular from zero. For this purpose, an output multiplier 418 is provided. The setpoint attenuation signal Q.sub.PSOD can be multiplied in the output multiplier 418 by a value which can increase from 0 to 1, for example as a ramp. Such a value can be output by the coordination unit 406 as output trigger T.sub.A. This output trigger T.sub.A can also be output as output trigger T.sub.A in order to then be used as activation indicator in subsequent closed-loop control structures.

[0155] The closed-loop attenuation control structure 400 therefore receives an input signal S.sub.0 which is representative in principle of the entire grid voltage signal, and in dependence thereon the closed-loop attenuation control structure 400 outputs the setpoint attenuation signal Q.sub.PSOD in addition to the output trigger T.sub.A, which can act as activation indicator.

[0156] FIG. 5 shows a structure comprising conversion elements, of which many can also be referred to as filters, which can be contained in the determination controller block 416. The oscillation signal Sd to be attenuated can be input at the converter input 502 of the converter structure 500. It is possible for in this case only the standardized differential voltage to be input, for example as a percentage value based on the rated voltage. The voltage difference is therefore, as has been explained above, the amplitude of the phasor, which is calculated from the three voltage values of the three phases, minus the value of the rated voltage. The rated voltage can correspond to 100%.

[0157] This input signal can first be given via an amplification element 504 in order to amplify it. A low-pass filter element 506 can follow, which can further reduce a still remaining noise component.

[0158] Further components which follow can each comprise a plurality of individual elements. FIG. 5 shows in this regard three examples each having three elements. However, it is also possible for in each case more or in particular fewer elements to be provided. In addition, reference is made to the fact that if the components are linear, they can also be arranged in a different order than that illustrated. Taking into consideration calculation which is not infinitely precise and signal rendering which is not infinitely precise, however, the proposed order can be advantageous.

[0159] Following on from the low-pass filter element 506 there is a high-pass filter component 510, which in this case is composed of three high-pass filter elements 511 to 513. The high-pass filter component 510 is provided for the purpose of filtering out low-frequency signal components which are still remaining from the oscillation signal to be attenuated. Such low-frequency signal components can occur in particular owing to feeding wind power installations due to fluctuations in the wind.

[0160] Both high-frequency noise components which are filtered out by the low-pass filter element 506 and low-frequency signal components which are filtered out by the high-pass filter component 510 have naturally also already been filtered out by the bandpass filter block 410 in FIG. 4. Since filters rarely function ideally, in particular are rarely exactly selective, however, such signal components can nevertheless also be present after the filtering by the bandpass filter block 410.

[0161] The low-pass filter element 506 and the high-pass filter component 510 are tuned to the attenuation frequency f.sub.PSOD. Therefore, this attenuation frequency f.sub.PSOD is input into the determination controller block 416 and can therefore be taken into consideration in this converter structure 500 in FIG. 5. The low-pass filter element 506, the high-pass filter component 510 and the lead-lag components or lead-lag elements yet to be explained below are likewise set depending on the attenuation frequency f.sub.PSOD.

[0162] The low-pass filter element 506 is set in such a way that its fundamental is above the attenuation frequency f.sub.PSOD. The high-pass filter component 510 or the high-pass filter elements 511 to 513 is/are set in such a way that its/their fundamental(s) is/are below the attenuation frequency f.sub.PSOD.

[0163] The high-pass filter component 510 is followed by a first lead-lad filter component 520, which is composed of three individual lead-lag filter elements 521 to 523. Owing to the use of a plurality of filter elements, in this case, therefore, the three lead-lag filter elements 521 to 523, a higher degree of selectivity can be achieved. This also applies to the high-pass filter component 510 and also the second lead-lag filter component yet to be described below.

[0164] By means of a lead-lag filter, signal shifts can be achieved which are frequency-dependent. Thus, in particular the oscillation signal to be attenuated, in particular after the described further filtering which further works out this oscillation signal to be attenuated, can be correspondingly shifted. As a result, communications-related delay times can be compensated for.

[0165] A lead-lag filter, which can also be referred to as a lead-lag element, can be realized by virtue of the fact that at least one filter element or filter component is in the form of a lead element or lead component and a further filter component or filter element is in the form of a lag filter or lag element.

[0166] Furthermore, a second lead-lag filter component 530 can be provided which is likewise composed of three lead-lag filter elements 531 to 533. It is also possible for the first lead-lag filter component 520 to function substantially as the lead component in the frequency range in question, whereas the second lead-lag filter component 530 functions substantially as the lag component in the frequency range in question.

[0167] Finally, a limitation element 540 is also provided with which possibly a limitation of the generated signal can be performed. As a result, it is possible to avoid a situation whereby, as a result of the processing using the various filter components, an undesirably high signal amplitude occurs. Such an undesirably high signal amplitude can also occur when the oscillation signal to be attenuated has a correspondingly high amplitude. Although generating a corresponding setpoint attenuation signal which likewise has a high amplitude would on its merits be appropriate, it can have problems in terms of its implementation, with the result that this limitation is provided.

[0168] Therefore, a setpoint attenuation signal ΔQ.sub.PSOD is output by the converter structure 500 which can likewise be output as a percentage value. By way of reference, in this case a maximum reactive power value of the feed device can be used which can correspond in terms of magnitude to the rated power of the feed device. The designation with the Greek letter Δ indicates that this setpoint attenuation signal is intended to be injected onto an existing signal. It can correspond to the setpoint attenuation signal Q.sub.PSOD, in particular after multiplication by the activation trigger T.sub.A at the output multiplier 418.

[0169] In particular, provision is made here for a corresponding reactive power signal to be determined as setpoint attenuation signal and for the components or elements in the determination controller block 416 or the converter structure 500 to be adapted thereto.

[0170] FIGS. 6 to 8 show alternative structures for generating a setpoint feed signal depending on a basic setpoint signal and depending on a setpoint attenuation signal. All three structures in FIGS. 6 to 8 build upon a structure having a core controller 602, 702 and 802, respectively. Such a core controller is provided for converting a basic setpoint signal which still does not have a setpoint attenuation signal.

[0171] In this case, the basic setpoint signal is a basic reactive power setpoint signal. This basic reactive power setpoint signal Q.sub.S is intended, at least as long as there is no consideration of a setpoint attenuation signal, to be fed in as feed signal, and for this purpose, the core controller generates a controller output signal Q.sub.A. Without taking into consideration the setpoint attenuation signal, this can correspond to a setpoint feed signal Q.sub.E. For this purpose, the core controller receives, as input signal, a control error e (can also synonymously be referred to as system deviation) as a result of a setpoint value/actual value comparison between the basic reactive power setpoint signal Q.sub.S and an actually fed-in reactive power signal Qi. Deviations between the basic reactive power setpoint signal Q.sub.S and the actually fed-in reactive power signal Qi, i.e., the actual value, can result in particular owing to the response of the feed unit, i.e., in particular an inverter arrangement, and possibly further electrical components such as inductors, transformers and transmission lines.

[0172] For the additional feeding-in of the setpoint attenuation signal or of an attenuation signal corresponding to the setpoint attenuation signal, FIG. 6 provides a total structure 600 in which the setpoint attenuation signal Q.sub.PSOD is injected onto the basic reactive power setpoint signal Q.sub.S at a first summation element 604. A modified basic reactive power setpoint signal Q.sub.S′ results, and, in addition, the control error e is formed at the second summation element 606. The control error e is then input into the core controller 602 as previously.

[0173] The setpoint attenuation signal is generated by the attenuation controller 601, which can substantially correspond to the closed-loop attenuation control structure 400 in FIG. 4. Correspondingly, the attenuation controller 601 also receives an input signal S.sub.0, which can correspond to the input signal S.sub.0 in FIG. 4.

[0174] FIG. 6 therefore shows a simple variant in which there is the problem that the previous core controller 602 needs to also take into consideration this additional setpoint attenuation signal. Often, such a core controller 602 is provided for the conversion of a substantially constant reactive power setpoint signal, i.e., basic reactive power setpoint signal Q.sub.S, and therefore is not necessarily provided for the correction of an oscillating setpoint signal. However, if this core controller 602 is quick enough, i.e., is designed to be quick enough, and/or the oscillation to be attenuated is slow enough or a correspondingly slow oscillation to be attenuated is expected, the structure in FIG. 6 can be used.

[0175] FIG. 7 shows, as an alternative, a total structure 700 in which the setpoint attenuation signal Q.sub.PSOD is injected onto the controller output signal Q.sub.A downstream of the core controller 702 at a second summation element 706. This results directly in the setpoint feed signal Q.sub.E. In this case, too, an attenuation controller 701 is provided, which can correspond substantially to the closed-loop attenuation control structure 400 in FIG. 4 and can also receive the same input signal S.sub.0.

[0176] Likewise, a first summation element 704 is provided in the total structure 700 in FIG. 7, said first summation element forming a control error e, namely from the difference between the basic reactive power setpoint signal Q.sub.S and the fed-in reactive power actual value Qi. However, this results in the effect that an injected attenuation signal appears again in the control error e and is fed in in superimposed fashion on the basis of the setpoint attenuation signal. By injecting the setpoint attenuation signal Q.sub.PSOD at the second summation element 706, the fed-in reactive power deviates from the basic reactive power setpoint value Q.sub.S by this setpoint attenuation signal Q.sub.PSOD owing to the modified setpoint feed signal Q.sub.E. The fed-in reactive power actual value Qi is thus correspondingly changed. This can result in the core controller 702 attempting to correct this varying control error thus resulting. If, however, the core controller 702 is sufficiently slow in comparison with the oscillation signal Sd to be attenuated and therefore in comparison with the setpoint attenuation signal Q.sub.PSOD, this simple total structure 700 can be suitable or at least sufficient.

[0177] An improvement can be achieved, however, via the total structure 800 in FIG. 8. The total structure 800 in principle provides for the setpoint attenuation signal Q.sub.PSOD to be injected onto the controller output signal Q.sub.A at a second summation element 806 in order thus to obtain the setpoint feed signal Q.sub.E. In addition, provision is made for the fed-in reactive power actual value Qi to be adjusted by the attenuation signal fed in in accordance with the setpoint attenuation signal. As a result, it is possible to achieve a situation whereby the core controller 802 neither needs to convert the setpoint attenuation signal as well nor is influenced thereby in the control error e.

[0178] For this purpose, provision is made in principle for the fed-in reactive power actual signal Qi to be passed via a bandpass filter 810. The bandpass filter 810 can correspond to the bandpass filter block 410 in FIG. 4 and also receives the attenuation frequency f.sub.PSOD as input value, namely from the attenuation controller 801, which can substantially correspond to the closed-loop attenuation control structure 400. The bandpass filter 810 filters a compensation signal component out of the detected feed signal, i.e., out of the reactive power actual signal Qi. The compensation signal component in the ideal case corresponds at least in terms of its frequency to an attenuation signal Qd corresponding to the setpoint attenuation signal Q.sub.PSOD. However, it is influenced by the response of the feed device, in particular a converter arrangement, and possibly further components such as inductors, transformers and transmission lines. The attenuation signal Qd therefore actually corresponds imprecisely to the setpoint attenuation signal Q.sub.PSOD, with the result that the setpoint attenuation signal Q.sub.PSOD can also not simply be used for the adjustment.

[0179] In any case, the compensation signal component thus generated, as attenuation compensation signal, is subtracted from the reactive power actual signal Qi in a third summation element 808.

[0180] In idealizing fashion, therefore, a reactive power signal adjusted by the attenuation signal Qd remains and is subtracted from the basic reactive power setpoint signal Q.sub.S in the first summation element 804 in order thus to form the control error e, which has therefore been correspondingly adjusted by the attenuation signal Qd.

[0181] In accordance with the closed-loop attenuation control structure 400, the generation or output of the setpoint attenuation signal Q.sub.PSOD is dependent on the initialization trigger Ti and the activation trigger T.sub.A. This initialization or activation is therefore accordingly also implemented in the total structure 800 by an input multiplier 814 and an output multiplier 818. In other words, the adjustment of the reactive power actual signal Qi at the third summation element 808 is inactive even when no setpoint attenuation signal Q.sub.PSOD at all is generated or output.

[0182] The various embodiments described above can be combined to provide further embodiments. These and other changes can be made to the embodiments in light of the above-detailed description. In general, in the following claims, the terms used should not be construed to limit the claims to the specific embodiments disclosed in the specification and the claims, but should be construed to include all possible embodiments along with the full scope of equivalents to which such claims are entitled. Accordingly, the claims are not limited by the disclosure.