EP3616290A1

EP3616290A1 - Method for detecting formation of a separate system

Info

Publication number: EP3616290A1
Application number: EP18719568.0A
Authority: EP
Inventors: Volker Diedrichs
Original assignee: Wobben Properties GmbH
Current assignee: Wobben Properties GmbH
Priority date: 2017-04-24
Filing date: 2018-04-24
Publication date: 2020-03-04
Also published as: CA3060181C; WO2018197468A1; RU2734165C1; JP2020518216A; CN110546844A; JP6914358B2; DE102017108637A1; BR112019022067A2; CN110546844B; US11081886B2; KR20190137918A; CA3060181A1; US20200191840A1

Abstract

The invention relates to a method for controlling a generating unit (300) feeding into an electrical supply system (330), wherein the generating unit (300) feeds into the electrical supply system (330) by means of one or more converters or inverters (302), and the method is provided for the purpose of detecting a system separation or formation of a separate system, and the method comprises the steps of controlling the feed by means of a feed controller (334) operating with at least one current controller, detecting at least one current control error, testing the detected current control error for a disparity from a predetermined reference range and identifying a system separation involving a separate system, disconnected from the electrical supply system, arising, to which the generating unit (300) is connected if a disparity from the predetermined reference range has been detected.

Description

Method for detecting islanding

The present invention relates to a method for controlling a generating unit feeding into an electrical supply network, in particular a method for controlling such a wind turbine. Moreover, the present invention relates to such a generating unit, in particular such a wind turbine. Moreover, the present invention relates to a wind farm with at least one such wind turbine.

Wind turbines are known and nowadays they are often combined in a wind farm, so that many wind turbines, for example 50 or 100, feed into an electrical supply network at a grid connection point. It is often even the case that remote wind farms are additionally connected to the electrical supply network over a comparatively long spur line.

Such wind farms and the corresponding supply networks are provided with various protective devices. Especially in the electrical supply network, which may be, for example, the European grid, disconnectors may be provided to electrically disconnect parts or portions of the electrical supply network. Also, a separation of a wind farm mentioned here comes into consideration. In this case, such a separation can take place at very different locations. With regard to the wind farm, a separation can take place in the area of the grid connection point and, if a described long spur line is present, this can also take place, for example, at the end or beginning of this spur line. It is also contemplated that a separation takes place in which several wind farms, or in addition also other decentralized feeders such as solar systems, are affected. Thus, for example, it is considered that a separation takes place after a section is separated from the rest of the electrical supply network and this separated section further contains interconnected generating units. This section may include, for example, multiple electrically interconnected wind farms and solar panels. However, it is also possible that only a single wind farm is separated from a part of the electrical supply network. In any case, the areas separated by the separation may be referred to as island meshes. The separation with the result of the separation is called islanding.

There are different types of island network formation. One type of island network formation is one in which the isolated island grid includes only generating units that do not have directly coupled synchronous generators, which thus have no conventional large-scale power plants. Such an island network or the particular island grid is referred to here as island grid type A. This type of island network or this island network is also characterized by the fact that no additional significant consumers are present or operated in this separated part. Of course, every wind turbine also includes consumers, such as for operating a process computer. In this described island network formation to a type A island grid but no consumers are present in the separated part, which are not part of the generating units. The examples described above also relate to such islanding of an island grid type A.

Other possible island networks are not the subject of the present invention, in which, for example, significant consumers are present in the separated part and can be referred to as island grid type B, or where additionally also at least one large power plant is present and incidentally, the separated part is so large and different participants implies that it could continue to operate independently and can be described as fragmentation.

If such islanding of the island grid type A, it makes sense for the wind turbines and other producers in this island grid to reduce their power, in particular to reduce to zero, because it can no longer be fed into the rest of the supply network and no consumers for Slimming the power are available. The problem with this is, however, that such a separation can not be readily recognized. In particular, disconnection by a mains circuit breaker or other circuit breaker may occur on a regular basis without any advance notice or any warning or other indication. It also happens regularly that especially a wind farm or even a wind turbine has the corresponding disconnector neither under control nor under surveillance. Even if a central wind farm controller would have such a circuit breaker under surveillance, then there would still be the problem of quickly passing on such knowledge to the wind turbines of the park. Also, installing a sensor to detect if a circuit breaker disconnects or not is likely to be problematic because different circuit breakers exist in different locations and it is not possible to predict which circuit breaker will perform a disconnection. In addition, such an effort would be very high.

The German Patent and Trademark Office has in the priority application for the present application, the following prior art research: DE 195 03 180 A1, DE 10 2008 017 715 A1, DE 10 2014 104 287 A1, DE 691 15 081 T2, US 2013/0076134 A1 , US 5,493,485 A and WO 2017/009608 A1. The present invention is therefore based on the object to address at least one of the above-mentioned problems. In particular, a solution is to be created to detect islanding as quickly as possible and as reliably as possible. At least should be proposed to previously known solutions an alternative solution.

According to the invention, a method according to claim 1 is proposed. Accordingly, at least one generating unit feeding into an electrical supply network is controlled. In particular, such a generating unit is a wind energy plant. This generating unit feeds by means of one or more inverters or inverters in the electrical supply network. In this case, a converter is a device which generates an alternating current or an alternating voltage of another frequency from an alternating current or an alternating voltage of one frequency. An inverter generates an alternating current or an alternating voltage with a desired frequency from a direct current or a direct voltage. An inverter can be part of an inverter. The decisive factor is that an alternating current or an alternating voltage with a predefinable frequency is generated and, to that extent, all subsequent explanations on a converter apply mutatis mutandis to an inverter and vice versa.

It is particularly important that the generating unit does not feed by a synchronous generator coupled directly to the supply network, but by means of the converter or inverter.

The method is thus provided for detecting a network separation or island network formation. In particular, the island network is the result of grid separation, because the network separation forms a stand-alone grid. The proposed method comprises at least the following steps. Initially, feeding is controlled by means of feed-in control. The feed-in control works with at least one current control. Thus, the current is detected and returned to control the feeding and thus adjusting the current. Thus, a current regulation is used in the control-technical sense, which contains at least one control loop.

It is then further proposed that at least one current control deviation of the control device is detected. Said current control thus includes that there is a current control deviation, namely in particular a deviation between a current setpoint and a detected current actual value. This current deviation is thus part of the current control, but is here additionally detected for detecting a network separation or island network formation or further processed and evaluated.

Furthermore, it is checked whether the detected current control deviation has a deviation from a predetermined reference range. Such a reference range can thus be predetermined or predetermined and it describes an area in which the current control deviations may lie. In particular, this area describes an area in which the current regulation deviations are when operated in a typical non-islanding behavior.

Current control deviations occur naturally and are an essential part of the current regulation. The current control tries, as is usual for a regulation, to correct the current control deviations.

However, it was recognized that the current control deviations can additionally give information about an operating point or working range of the feed-in control and thus of the generating unit, in particular of the wind energy plant. Here, the basic behavior of the generating unit and the feed-in control and in particular their current control is known. In particular, it is known in which area the current control deviation is usually located, that is, in which range it lies when there is no grid separation or stand-alone network formation. Accordingly, this known range can be specified or predetermined as a reference range.

For this purpose, it is now checked whether the detected current control deviations leave this reference range, ie whether the current control deviation deviates from a predetermined reference range, ie lies outside of such a reference range. It is then detected on a network separation, in which a disconnected from the electrical supply network island network, detected when a deviation from the predetermined reference range has been detected. In this case, of course, only an island network or network separation is detected if the corresponding generation unit is also connected to this island network, which is caused by the detected grid separation or island network formation.

Particularly preferably, the island grid is a network to which only the generating unit and one or more further generating units are connected and to which, in particular, no further generating units with directly coupled synchronous generators are connected. Thus, no conventional large power plants are connected to the island grid under consideration.

The invention also proposes a method for controlling a generating unit feeding into an electrical supply network, wherein the generating unit feeds into the electrical supply network by means of one or more inverters or inverters and wherein the method for detecting a grid disconnection is prepared in which one of the electrical supply network! A separate island grid is formed, to which the generating unit is connected, the method distinguishes between a first degree island fault and the presence of a second island island fault and is first checked for detection of island fault of the first degree and after detecting a island fault of the first degree, the existence a second degree island fault is checked. The method can thus detect a network separation or island network formation and this can be done, for example, as already explained in accordance with at least one embodiment described above. Here as well, a generating unit that uses one or more inverters or inverters for feeding is used. In particular, the use of a wind turbine as a generating unit is also proposed here. This method distinguishes between a first degree islanding error and a second degree islanding error. Accordingly, a second degree island fault is a more serious error. The island fault of the second degree is particularly serious to a greater extent than an island fault of the first degree in that it poses a threat to the generating unit, in particular its converter or inverter. The island fault second degree may also cause a threat to other parts of the isolated island network.

For this purpose, it is thus proposed to first check for a recognition of an islanding error of the first degree. In particular, it is also proposed to respond to such a detected island fault first degree control technology, which will be explained below according to further embodiments.

Now, if a first degree island fault has been detected, it is proposed as a further step to check the presence of a second degree islanding fault. In that regard, a two-stage test is proposed here, namely first on the island fault of the first degree. If such is not available, there is no need to be tested for a second degree island fault.

Preferably, a first degree off-grid error is performed by a method as described in accordance with at least one embodiment above.

Furthermore, according to an embodiment, it is proposed that upon detection of the first-degree off-grid error, a current value at which the current control deviation is detected is set to zero. It is here so targeted and especially immediately, so as quickly as possible, responded to this island fault first degree. At least the reaction also looks like that the current output of the relevant generating unit should be regulated to zero. The generating unit thus remains electrically connected, for example, remains electrically connected to a wind farm network, but it receives the value zero as the current setpoint. It is thus intended to regulate the relevant current, ie in particular the output current of the generating unit, to the value zero.

Thus, at least as an immediate measure to this exceptional situation of network separation or island network formation was initially reacted quickly and appropriately. In this case, there is first an island network error of the first degree. However, it was recognized that despite this reaction to the island network formation an even greater exceptional case can be present, namely a second degree island fault. An island fault second degree is so far even rarer and even more critical than the island fault first degree. In particular, a second degree island fault can be distinguished by the fact that an undesired current occurs in the island grid, especially occurs at the output of the relevant generating unit, ie in particular occurs at the output of a wind turbine or its converter or inverter. Thus, such a current occurs here, although the setpoint value of the output current has been set to zero and insofar, if the regulation is functioning, an actual current of zero value must also be assumed. One cause of such a stream may be that further producers in the resulting island grid continue to feed electricity. It is now proposed that, when such a second degree island fault is detected after a first degree islanding fault, a circuit breaker is opened which interrupts a current underlying the current regulation deviation. This is particularly about the output current and if such a second degree island fault was detected, it is proposed to interrupt the output current by a corresponding circuit breaker.

In particular, it has been recognized here that initially in the presence of islanding, if it can still be assumed that the island grade error is of the first degree, the relevant generating unit initially remains connected and regulates only the current value zero. This can be ensured that the generating unit can also be fed back immediately, especially at the end of the occurred island network formation. This can be particularly important to support the network, which may have led to islanding due to a network problem. Only when this attempt fails and a second-degree islanding fault has been detected is an actual separation of the generating unit carried out. According to one embodiment, it is proposed that a test variable or test function be determined for checking the detected current control deviation to a deviation from a predetermined reference range from the current control deviation. For example, the amount of current deviation may be used as a test variable, to give a very simple example. However, it is also possible to consider the dynamics so that, for example, an increase in the control deviation per unit of time is used as the test variable. Additionally or alternatively, several values, especially in chronological order, can be recorded and used as a test function. It is also possible, a transformation of the control deviation or a Control deviation of a transformed current to be regarded as a test variable or test function.

Coordinated with this, a reference variable or reference function is formed and compared with the test variable or test function. In the simplest case, an upper limit for a control deviation can be determined in terms of amount. This upper limit is then the reference value and the amount of the detected control deviation the test value. If this amount exceeds the limit value, it is then assumed that there is a deviation from the predetermined reference range.

Likewise, however, it is also possible to use curves as a reference function here. For example, situations may occur in which there is a very high current deviation, very briefly, for example only for one sampling step, but which then drops back to a much lower value. Such a behavior may be a behavior that does not imply network disconnection or islanding and thus represents a behavior that does not represent a deviation of the current deviation from the predetermined reference range. In other words, this exemplified current waveform can be in the reference range with a short high current value.

In this way, further courses of current control deviations can form a course, which does not indicate a grid separation or island network formation. Thus, many different current waveforms can be in the reference range or thereby form the reference range in their entirety.

In order to check for a deviation from the reference range, it is thus also possible to consider that a detected current-control deviation and, for example, a time-normalized course form the test function. This test function can then be compared with several reference functions. If this test function exceeds a reference function depending on the consideration, this only means that the recorded test function does not match the examined reference function. If, however, another function is found under which this test function falls, then the reference function and thus the current control deviation lie in the predetermined reference range.

According to one embodiment, it is proposed that the predetermined reference range be determined as a function of an operating mode or operating state of the generation input. unit, in particular the feed-in control, is specified or changed. Thus, if the wind energy installation is in a normal operating mode in which there is a feed without any special features, then a corresponding reference function can be specified, which reproduces a correspondingly normal reference range. If the wind energy installation then changes to a support mode, for example by temporarily providing an instantaneous reserve power by briefly supplying more power to the electrical supply network than is possible due to the wind prevailing at the moment, or more than the rated power, a higher current control deviation, for example, can also occur to be expected. This is particularly due to the high dynamics of such a network support by providing an instantaneous reserve power. Accordingly, then the reference range or for the reference size or reference function can be adjusted. In other words, depending on the operating mode of the wind power plant, a variable-width control current deviation can lead to the evaluation that there is grid separation or stand-alone network formation. In this embodiment, depending on the operating mode, the relevant reference range would thus be selected or adapted or a corresponding reference variable or reference function selected or adapted. Another variant would be to establish corresponding reference ranges or reference variables or reference functions for different operating modes and then to test the test variable or test function for each of these reference ranges.

Preferably, a difference between a desired and an actual value or a desired and an actual current component of a current to be fed in is used as the current control deviation. Thus, the current to be injected and its control deviation are considered here in particular. Usually, a three-phase current is generated and, in addition, a solMstwertvergleich is made for each phase and thus considered for each phase, a current deviation. A phase may be a current component of the current to be injected. In this case, for example, only the current control deviation of a phase can be considered, or it can be considered for each phase, a current deviation. It is also possible to summarize the current deviation. It is also possible to use solMstwertvergleiche for transformed sizes. In particular, a transformation into a co-and a negative system comes into consideration, in which case preferably a SolMstwertvergleich for the Mitsystem component is proposed. It is also a transformation into dq components into consideration and it is proposed a SolMstwertvergleich for both components, namely in each case individually. Incidentally, these transformed quantities can also be understood as test parameters or test functions.

According to a further embodiment, it is proposed that a deviation of the current control deviation from the reference range is present when - the current deviation, test variable or test function exceeds a predetermined limit in magnitude, the current deviation, test variable or test function leaves a predetermined normal band or the current deviation, test variable or Test function changes with a temporal gradient, which exceeds the amount according to a predetermined limit gradient.

In a simple case, it is thus proposed to consider only the amount and to compare it with a predetermined limit value. Thus, a check of the absolute size of the current deviation is proposed. Instead of using an amount, a band may also be specified and a deviation exists when that band is exited up or down, or, in the case of a function, up and down. This also makes it possible to specify an upper and lower limit with different values.

When considering a temporal gradient, especially a dynamic behavior is considered. As a result, it may be possible to record particularly quickly if the current control deviation leaves the reference range. However, it is also possible to combine these test criteria. One possible combination is that deviation of the current control deviation from the reference range is assumed if one of several criteria detects a deviation. According to one embodiment, it is proposed that by means of a three-phase feed current is fed into the electrical supply network and this is composed of three phase currents. In this case, a current setpoint is specified for each phase current. For this purpose, it is proposed that the current control deviation takes into account a deviation of each phase current from its nominal value. Thus arise always three current control deviations. Especially the deviations of the instantaneous values are considered here.

For this purpose, it is particularly proposed that the current deviation which is used for testing whether there is a grid separation or stand-alone network formation is formed according to a vector metric from amounts of the deviations of each phase current from its current setpoint. Each phase current, current setpoint and also the respective deviation between them can each be described as a vector, possibly time-varying. Such a consideration may be referred to particularly as vector metrics. It is proposed here to consider deviations of each phase current from its current setpoint based thereon. In particular, the deviations can be described as vectors and the amounts are considered.

Preferably, the sum of such amounts is considered as a current deviation. For this purpose, it is proposed that a network separation or stand-alone network formation is detected if the current control deviation thus detected, that is to say in particular the sum of the amounts, exceeds a deviation limit value. As a result, all three phases of the feed-in current can be taken into account in a simple manner.

According to one embodiment, it is proposed that an amount of the current control deviation, test variable or test function is set in relation to a tolerance bandwidth, in particular to an average tolerance bandwidth. For this purpose, it is particularly proposed that a network separation is detected when the ratio of the amount of the deviation or the deviation sum to the tolerance bandwidth exceeds the deviation limit value.

It is particularly assumed that a network separation, if the amount of deviation is many times greater than the tolerance bandwidth. Then it is to be assumed that a network separation, because the underlying tolerance band method could not rudimentally correct the deviation and, in particular, this very high control deviation was exceeded by a multiple of the tolerance bandwidth. In the case of a tolerance band method, a deviation of the nominal current from the upper band limit can be assumed as a system deviation if this is exceeded or from the lower band limit if this is undershot. Alternatively, it is possible to base the evaluation of the current control deviation on a value curve within the tolerance band. According to an alternative embodiment, it is proposed that a three-phase feed-in current, in particular by predetermining current components by means of a vector control, be fed into the electrical supply network, with the three-phase feed current being controlled by means of a dq transformation into a d-component and a q- to control the feed-in. Component is decomposed. For this purpose, it is proposed to use a difference between a setpoint and an actual value of the d component and / or q component as the current control deviation.

In accordance with a further embodiment, it is proposed that the recognition of a network separation or island network formation by designing a deviation from the predetermined reference range is interpreted as detecting a first-degree islanding network error. For this purpose, it is then proposed that after detection of an islanding fault of the first degree, the generating unit continues to be operated. In particular, it continues to operate with a zero current setpoint.

For this purpose, it is preferably further proposed that subsequently the presence of a second degree islanding error be checked and then assuming the existence of a second islanding error second degree, if a current control deviation is still detected, although a current setpoint value is zero in the feed-in control. Thus, it has thus been recognized that a second degree islanding error then exists if the generating unit fails to actually comply with the current setpoint value of zero. Accordingly, there is a large exception error, namely a second degree island fault, in which the resulting island network of the generating unit basically imposes a current, be it positive or negative. Exactly this situation is preferably checked here.

According to a further embodiment, it is proposed that after detection of an islanding fault of the first degree, the generating unit remains connected to the electrical supply network or the island grid and after detection of a second island fault, the generating unit is disconnected from the electrical supply network or the island grid or parking network. In particular, a galvanic separation is proposed here. The separation can also be done by means of appropriate power semiconductors. It is thus not only advantageously tested in a second step for a second degree island fault, but it is also, should such a second degree island fault be recognized, a further security measure proposed. In particular, the method is implemented in the generating unit so that it carries out the proposed steps quickly in succession, and thus very quickly too, should that be necessary, perform this separation in case of second degree islanding error.

According to the invention, a generation unit, in particular a wind turbine, is also proposed. This includes at least one or more inverters or inverters for feeding electrical power into the electrical supply network. Whether a converter or inverter is used, depends on the specific design of the generating unit, especially the wind turbine. It is important that the generating unit is not designed in such a way that it feeds via a synchronous generator directly coupled to the grid, but via a converter or an inverter unit.

Furthermore, a feed-in control is provided, which is prepared to control the feeding by means of at least one current control. Thus, especially such a current regulation is implemented in the feed-in control. There are also corresponding measuring devices which carry out the corresponding current measurement for current regulation.

In addition, a detection means is provided for detecting at least one current control deviation of the control device. The current control deviation is thus not only used for the current control of the feed-in control, but also used for further testing. The detection means can also be formed in that it receives the current deviation as a signal from the feed control. In particular, the detection means can also be provided as a software solution. Another evaluation of the flow control deviation in the feed-in control comes into consideration. In this case, the detection means would be a corresponding evaluation block in the software. Furthermore, a test means is provided for checking the detected current deviation to a deviation from a predetermined reference range. This test equipment can also be designed as software in a test block. It can also be implemented within the feed-in control.

The system controller is prepared to recognize a network separation when a deviation from the predetermined reference range has been detected. In this case, the grid separation is one in which a stand-alone grid separated from the electrical supply grid is created, namely that to which the generating unit is connected. It is particularly preferably provided that the generating unit, in particular a wind energy plant, is characterized in that it is prepared to carry out a method according to at least one of the embodiments described above.

According to the invention, a wind farm with several wind turbines is also proposed. At least one of the wind turbines, preferably all of these wind turbines, is or are each a generating unit or wind turbine according to an embodiment described above. It is particularly advantageous for such a wind farm that, in the event of grid disconnection, it can form the island grid or can form a significant part of such an isolated grid. The detection of such a network separation and the proposed action taken can thus lead to a protection of the wind turbines but thereby also to a protection of the wind farm as a whole. Therefore, it is advantageous to provide a wind farm with such wind turbines that can detect such separation or islanding.

The invention will be explained in more detail by means of embodiments with reference to the accompanying figures.

Figure 1 shows a wind turbine in a perspective view.

FIG. 2 shows a wind farm in a schematic representation.

Figure 3 shows schematically a part of a generating unit with elements for illustrating the behavior in the case of islanding according to an embodiment.

Figure 4 shows schematically a part of a generating unit and elements for illustrating the behavior in the case of islanding according to a second embodiment.

FIG. 1 shows a wind energy plant 100 with a tower 102 and a nacelle 104. A rotor 106 with three rotor blades 108 and a spinner 110 is arranged on the nacelle 104. The rotor 106 is set in rotation by the wind in rotation and thereby drives a generator in the nacelle 104 at. FIG. 2 shows a wind farm 1 12 with, by way of example, three wind turbines 100, which may be the same or different. The three wind turbines 100 are thus representative of virtually any number of wind turbines of a wind farm 1 12. The wind turbines 100 provide their power, namely in particular the 5 generated power via an electric parking network 1 14 ready. In this case, the respective generated currents or outputs of the individual wind turbines 100 are added up and usually a transformer 1 16 is provided, which transforms the voltage in the park, to then at the feed point 1 18, which is also commonly referred to as PCC, in the supply network 120th feed. Figure 2 is only a simplified illustration of a wind farm 1 12 which, for example, does not show control, although of course there is control. Also, for example, the parking network 1 14 be designed differently, in which, for example, a transformer at the output of each wind turbine 100 is present, to name just another embodiment.

FIG. 3 shows a part of a generation unit 300, namely, in particular, an inverter 302, with a system controller 304, which contains elements for measuring, evaluating and controlling the inverter 302.

The inverter 302 has a DC intermediate circuit 306, which receives power or energy from a generator part 308 of the generating unit 300. The generator part 308 is indicated here only schematically and can, for example, for a

20 generator of a wind turbine with downstream rectifier stand. The DC intermediate circuit 306 thus receives its power or energy from the generator part 308 and, based on this, the inverter 302 can generate a three-phase output current at the inverter output 310. This output current is output via the mains chokes or three-phase mains choke 312 and can there as well

25 are detected in the region of the three-phase line choke 312 with a current measuring means 314 as output current i (t). This output current i (t) is therefore representative of the entire three-phase current or representative of a measurement of a phase current of each of the phases.

For each of these phase currents i (t), a desired 30 / actual value comparison is respectively performed on the current comparator 316 between the detected actual current which was detected by the current measuring means 314 and a desired current. For a better explanation, the currents hi, and isi are plotted as actual values there as actual currents of the individual phases, which currents are subtracted from the respective desired current h _s , i 2s or i 3 _s . The desired currents h _s , 12s and i3s are specified in the transformation block 318 for each phase. This should also be illustrated with the indicated sine waves sin, which are shown in different phase positions.

In front of the current comparator 316, a multiplier arrangement 320 is arranged, which is intended to take into account the case of islanding and only then becomes relevant. As long as no island network formation has been detected and thus in particular there is also no island network error, the multipliers each receive the value 1 as error signal EF, so that the current nominal values which the transformation block 318 outputs reach the respective comparator 316 unchanged. The transformation block 318 receives current setpoint values in dq coordinates as input variables, namely the setpoint value ids and the setpoint value i _qs . The setpoint current component i _qs is thereby predefined essentially directly. The setpoint current component ids also takes into account a setpoint / actual comparison of the voltage comparator 322, which forms a sol d value difference between the voltage Vdc detected at the DC voltage intermediate circuit 306 and a predetermined voltage Vdcs.

The transformation block 318 also takes into account a transformation angle γ, which is determined by a PLL control 324 from a measured output voltage v (t). The output voltage v (t) is detected by means of a voltage measuring means 326, for example in the region between the three-phase mains choke 312 and a power transformer 328. Incidentally, in the illustrated illustration of FIG. 3, the power transformer 328 is then connected to the indicated network 330. The grid 330 may be the electrical utility grid and the grid connection point 332 may be between the grid transformer 328 and the indicated grid 330.

For controlling the feeding, the current control deviations Δη, Δ, 2 and Δ, 3, ie the outputs of each comparator 316, are supplied to the control blocks 334. The drive blocks 334 respectively drive respective semiconductor switches in the inverter 302 to generate the output currents iii, 2I, and i3i from the DC voltage in the DC link 306. Incidentally, the drive blocks 334 together form feed-in control. In this case, the comparators 316 and, if appropriate, a setpoint value, in particular the transformation block 318, can be added to the feed-in control. The current control deviations Δπ, Δ, 2 and Δ, 3 are also input to the test block 336, which thus forms the test means. This test means or the test block 336 checks whether the current control deviation deviates from a predetermined reference range. The symbolic data feed 338 can to this extent also be regarded as detection means for detecting the current control deviations. The current control deviations are formed in the current comparators 316 to control the feed, but their forwarding to the test block 336 is another detection to that extent.

In the test block 336, it is thus checked whether these current control deviations or current control deviations Δπ, Δ, 2 and Δ13 deviate from a predetermined reference range. In the simplest case, a check of the absolute values of these three differential currents Δπ, Δ, 2 and Δ13 with a limit value is considered here. For this purpose, for example, an average of their amounts can be formed and compared with a corresponding limit value, or their amounts are added up and this sum is compared with an absolute limit value. If it turns out that there is a network separation and thus an island network, the error signal EF is output. This error signal EF can be given to another evaluation or control block 340. This evaluation and control block may include, for example, to inform the network operator or the plant operator or a park operator of the detected error. In addition, it is provided to give the error signal EF to the multiplier arrangement 320, there with the desired currents h _s o, i2so and. i3so to be multiplied. For this error signal can be designed so that it assumes the value zero in the event of an error. This would then the three desired currents h _s , 12s and. i3s have the value zero. However, this use of the multiplier arrangement 320 is to be understood symbolically in particular, and various other implementations come into consideration, such as the error signal already to be considered in the transformation block 318.

Also illustratively, an error propagation 342 is shown to indicate that the error signal may also directly affect the control values output from the drive blocks 334. In particular, it should be illustrated here that the fastest possible and immediate response or measure is proposed.

The result is thus that a current setpoint of zero or three times a current setpoint of zero is given for each phase. Is then used with the current measuring means 314 If, in essence, a current with the value zero is also detected, then the generating unit 300 can continue to remain connected, ie in particular also be connected to the network 330 via the network transformer 328 and the grid connection point 332. But it is detected that the output current i (t) or the three phase currents hi, and Ϊ3ί are not zero, especially if it is detected that they have a very high value, this may also be detected in the test block 336. In particular, it is implemented in the test block 336 that such monitoring is performed. Thus, such a current behavior is monitored in the test block 336, especially after occurrence of islanding. In that regard, the detection of an island mesh formation described so far with reference to FIG. 3 is the detection of a first-degree islanding fault.

If, after this first-degree islanding fault has been detected and the current setpoint or setpoint values have been set to zero, a current is still detected, in particular a high output current is detected, possibly also with a negative sign, this can be considered as second degree islanding error in the test block 336 be recorded. For this purpose, the special error signal EEF is output. This special error signal EEF can also be supplied to the evaluation and control block 340 in order to communicate this, for example, to a system, parking or network operator. In addition, it is proposed that upon detection of this island network error of the second degree of the intended disconnect switch 344 is controlled by the test block 336 immediately, namely so that it opens. The generating unit 300 and thus its inverter 302 is thus separated from the rest of the network. The network separation takes place as close as possible to the generating unit 300 and the inverter 302, namely directly behind the line choke 312. This may mean, for example, a separation of a wind park network 346, which is only hinted at here and also assumes that the generating unit 300th a wind turbine is.

It is thus possible in a simple manner to detect a network separation or island network formation in that the current control deviation, namely, in particular the differential currents Δπ, Δί2 and Δί3, are monitored. Particularly advantageous here is that an immediate evaluation of the voltage in particular wind park 346 is not required. This particularly takes into account the fact that the voltage detected in the wind farm 346 can be a very inaccurate criterion, in particular by means of the voltage measuring means 326. In principle, voltage peaks, voltage peaks or frequency changes from a normal state can occur. Sometimes you can Such deviations also assume large values, but this does not mean that a grid separation or islanding is present. By monitoring or evaluating the current control deviation, however, such network separation or stand-alone network formation can be reliably detected because the system behavior, namely the behavior of this current control deviation, is known in principle. If a current deviation occurs that leaves the predetermined reference range, this is a sure sign of a network separation or island network formation, in any case if the reference area contains all the behavior that can not be attributed to grid separation or stand-alone network formation. It was also recognized that the current control already takes into account the entire behavior of the network, including voltage and frequency changes as well as necessary support reactions by the generating unit.

Also can be reacted by the proposed solution in a simple manner immediately on such a network formation and also by the proposed setting of the current setpoint to zero, the island network is considered and fed no further power, however, the generating unit in a standby state, from the out she could get back into grid support as soon as possible.

Similarly to the embodiment of FIG. 3, FIG. 4 shows a generation unit 400 or a part thereof, which has an inverter 402 and a system controller 404. Also provided is a DC link 406 which is powered by a generator section 408 and provides power to inverter 402 to provide inverter output 410 with output current i (t) and three phase currents in, respectively Also, a three-phase line choke 412 is present. In addition, a line filter 413 is indicated here.

The embodiment of FIG. 4 also shows a parking network 446, a network transformer 428, a network 430, and before that a network connection point 432.

For controlling the feeding, an arrangement of a plurality of drive blocks 434 is also provided. These drive blocks can at least partially form the feed-in control. Unlike in the embodiment of FIG. 3, however, a voltage control or a vector control is provided here. Especially a triangular modulation is proposed here. For this, the three-phase output current i (t) with the Current measuring means 414 detected, and transformed in a current transformation block 450 into a _q- component i _q and a d-component id. The transformation angle γ required therefor is also determined here by means of a PLL control 424 which uses a detected voltage v (t) of the voltage measurement means 426 as input variable. The components i _q and id transformed in the current transformation block 450 are then compared in a current comparator arrangement 416 with a desired current component ids or i _qs , respectively. The calculation is very similar to the embodiment of Figure 3 and in particular the current component ids is compared by means of a voltage comparator 422 from the voltages Vdc (t) and Vdcs. As a result, in comparison of the current comparators 416, the two differential current components Aid and Aiq arise. These two differential current components Aid and Ai _q thus form the control current deviation and this is supplied via the data feed 438 to the test block 436.

These two differential current components Aid and Aiq are also converted in a voltage specification arrangement 452 into the two nominal voltage components Vd and Vq, which in turn are converted in the transformation block 418 into three voltage profiles, namely one voltage profile per phase. These three predefined voltage profiles Vis, V2s and V3s are then ultimately converted in the drive blocks 434 into drive signals for the inverter 402. Thus, there is also an injection by means of a current control here. Namely, this current control performs a setpoint / actual value comparison between desired currents and actual currents. However, the variant of FIG. 4 uses the transformed current components id and i _q or ids and i _qs . The result is a current control deviation, namely the differential current components Aid and Aiq. These are required for the feed-in, but also recorded for the detection of grid disconnection or stand-alone grid formation. The data feed 438 can thus also be understood here as a detection means which detects this current control deviation from the feed-in control and supplies it to the test block 436.

The test block 436, which thus represents a test means, then first checks for a network separation or island network formation and thus for a first-island island fault. If such an islanding network error of the first degree is detected, the error signal EF is also output here. This error signal EF is for the sake of simplicity just as designated as the embodiment of Figure 3, but may differ in its values. Preferably, however, it does not differ and has the value zero or one. If it has the value one, this means that there is no error, ie no grid separation or stand-alone network formation has been detected. This then also means that this value one have no effect on the multiplier arrangement 420, so ids and i _qs does not change the depth there target current components because of the multiplication by the first But if a network error of the first degree is detected, this error signal EF can assume the value zero. This results in the setpoint current components being set to zero. Here too, as in the embodiment of FIG. 3, the symbolically represented error forwarding 442 indicates a direct and immediate effect on control signals of the control blocks 434 for driving the inverter 402.

This detection of a first-degree islanding network error thus also leads to the reaction that the nominal currents are set to the value zero. At the same time, the error signal EF can be supplied to the evaluation and control block 440 in order, for example, to continue to communicate this information and not only to use it for the regulation.

If such an islanding fault of the first degree has now been detected, the test block 436 continues the test and checks whether a second-degree off-grid error also occurs. Again, this is done based on the sensed current droop that is still provided to the test block 436 by the data feed 438. If an islanding network error of the second degree is detected, the disconnecting switch 444 is actuated, namely opened, and the inverter 402 is thus disconnected from the parking network 446. In addition, this particular error EEF is fed to the evaluation and control block 440.

Thus, if a first degree island fault is detected, the inverter 402 will continue to operate and not separate, but it will not feed any current. If, nevertheless, a current is detected, in particular one which has a high value and can not be explained by the control of the inverter 402, a second-degree stand-alone system fault is assumed and the disconnecting switch 444 is opened.

If the error has now been eliminated, in particular the island network formation is finished or can be terminated shortly, the evaluation and control block 440 can also be used to give a reset signal to the test block 436. As a result, if necessary, the circuit breaker 444 can be closed again and it can also accept the error signal EF again the value one, and thereby the target current can again Leave zero. It is also contemplated that only a first degree island fault was detected and the disconnect switch 444 was not opened. But even then the evaluation and control block 440 may give a reset signal to the test block 436 to reset at least the error signal EF again to a value which shows that there is no error, in particular to the value one.

Incidentally, this functionality is to give a reset signal from the evaluation and control block 440 to the test block 436 in the same manner described also for the embodiment of Figure 3, therefore there the evaluation and control block 340 is a reset signal can give to the test block 336.

Claims

claims

A method for controlling a generating unit (300) feeding into an electrical supply network (330), wherein the generating unit (300) feeds into the electrical supply network (330) by means of one or more inverters or inverters (302), and

the method is for detecting network disconnection or islanding, and the method comprises the steps

Controlling the feeding by means of a feed-in control (334), which works with at least one current control,

Detecting at least one current regulation deviation,

Checking the detected current deviation to a deviation from a predetermined reference range and

Detecting a grid separation, in which an isolated from the electrical supply network island grid is formed, to which the generating unit (300) is connected when a deviation has been detected by the predetermined reference range.

A method according to claim 1, characterized in that the island network designates a network to which only the generating unit (300) and one or more further generating units (100) are connected, in particular that no generating units are connected to the island grid with directly coupled synchronous generator.

A method for controlling a generating unit (300) feeding into an electric power grid, wherein the generating unit (300) feeds into the electric power grid (330) by means of one or more inverters or inverters (302), and wherein

the method is provided for detecting a grid disconnection, in which a stand-alone grid separated from the electrical supply network (330) is formed, to which the generating unit (300) is connected, the method comprising a first degree islanding fault and the presence of a second degree islanding fault differentiates and

is first checked for detecting a island fault of the first degree and

after detecting a first degree islanding fault, the presence of a second degree islanding fault is checked. A method according to claim 3, characterized in that

the detection of the island fault of the first degree according to claim 1 or 2 takes place, and

upon detection of the islanding fault of the first degree, a current setpoint at which the current control deviation is detected is set to zero and

then upon detection of a second degree islanding fault, a circuit breaker (344) is opened, which interrupts a current that is based on the current control deviation.

Method according to one of the preceding claims, characterized in that for checking the detected current control deviation to a deviation from a predetermined reference range

from the flow control deviation, a test variable or test function is determined and

is compared with at least one reference size or reference function.

Method according to one of the preceding claims, characterized in that the predetermined reference range depending on an operating mode or operating state of the generating unit (300), in particular the feed control (344), predetermined or changed.

Method according to one of the preceding claims, characterized in that a difference between a desired and actual current or a nominal and actual current component of a current to be fed is used as the current control deviation.

Method according to one of the preceding claims, characterized in that there is a deviation when

the current control deviation, test variable or test function exceeds a predefined limit in terms of amount,

the current control deviation, test variable or test function leaves a given normal band or

the current deviation, test variable or test function changes with a time gradient which exceeds a predetermined limit gradient in magnitude. Method according to one of the preceding claims, characterized in that

is fed by means of a three-phase feed current into the electrical supply network (330), which is composed of three phase currents, wherein for each phase current, a current setpoint is specified and the current control deviation takes into account a deviation of each phase current from its setpoint, in particular that

the current deviation according to a vector metric from amounts of the deviations of each phase current from its current setpoint, e.g. whose sum is formed and

A network separation or islanding is detected when the thus detected current deviation exceeds a deviation limit.

Method according to one of the preceding claims, characterized in that

an amount of the current control deviation, test variable or test function is set in relation to a tolerance bandwidth, in particular to an average tolerance bandwidth. 1. The method according to any one of claims 1 to 8, characterized in that

a three-phase supply current, in particular by presetting of power components by means of a vector control, is fed into the electrical supply network, wherein

for controlling the feeding of the three-phase feed-in current by means of a d-q transformation into a d-component and a q-component is decomposed and

- As current deviation, a difference between a desired and a

Actual value of the d component and / or the q component is used.

Method according to one of the preceding claims, characterized in that

the recognition of a grid separation or stand-alone network formation, in that a deviation from the predetermined reference area has been detected, is designed as detecting a stand-alone network error of the first degree, and the unit of detection is further operated after recognizing this island grid error of the first degree. A method according to claim 12, characterized in that

after detection of an islanding fault of the first degree

the generating unit continues to operate with a current setpoint of zero,

the presence of a second-degree island fault is checked, and then it is assumed that a second-degree islanding fault exists, if a current-control deviation is still detected, even though there is a zero current setpoint in the feed-in control.

Method according to one of the preceding claims, characterized in that

after detection of an islanding fault of the first degree, the generating unit remains connected to the electrical supply network or the island grid and

after detection of an islanding error of the second degree, the generating unit is disconnected from the electrical supply network or the island grid, in particular is electrically isolated.

Generating unit (300), in particular wind energy plant (100), comprising one or more inverters or inverters (302) for feeding electrical power into the electrical supply network (330),

a plant controller (304), which is prepared to detect a network separation or island network formation,

a feed control (334), which is prepared to control the feeding by means of at least one current control,

a detection means (338) for detecting at least one current control deviation,

a test means (336) for checking the detected current deviation to a deviation from a predetermined reference range, and wherein the system controller (304) is prepared to detect a grid disconnection which results in a stand-alone grid separated from the electrical grid (330) to which the generating unit (300) is connected when a deviation from the predetermined reference range has been detected. Generating unit (300), in particular wind energy plant (100), according to claim 15, characterized in that it is prepared to perform a method according to one of claims 1 to 14.

Wind farm (1 12) with a plurality of wind turbines (100), wherein at least one of the wind turbines (100) as a generating unit (300) or wind turbine (100) according to claim 15 or 16 is formed.