Method and wind power installation for feeding electrical power into a power supply grid

Abstract

The invention relates to a method for feeding electrical power into a three-phase power supply network by means of an inverter device (105, 204), wherein the power supply network (208) has a three-phase network voltage, the network voltage having a first, a second and a third network voltage phase, the method comprising the steps of: if a fault-free operation has been identified for the power supply network (208), then feeding electrical power during normal operation, wherein during normal operation the positive-sequence voltage (u+) and optionally the negative-sequence voltage (u−) are detected from the network voltage, and a reactive current is preset at least according to the positive-sequence voltage (u+) and optionally according to the negative-sequence voltage (u−), and if the voltage change of the network voltage meets a predetermined fault criterion, in particular when the voltage change exceeds a presettable minimum change amount or minimum change gradient amount, then switching to a fault operation, wherein during the fault operation the reactive current is preset at least directly after the switching according to a space vector voltage.

Description

Technical Field

[0001] The invention relates to a method for feeding electrical power into a three-phase power supply network by means of an inverter device. In particular, the invention relates to such a method using a wind power plant. The invention also relates to a wind power plant having an inverter device for feeding electrical power into a three-phase power supply network. Background Art

[0002] In particular, wind energy systems and photovoltaic systems, but also other decentralized feeders, usually feed into the power supply network with the aid of one or more inverters. Usually, a plurality of inverters are used and these inverters can together form an inverter device or part thereof. Such decentralized feeders often also have to be able to feed reactive power into the power supply network. In this case, in order to feed in reactive power, it can be provided that, in terms of the method of symmetrical components, corresponding reactive currents are fed in in the positive and negative sequence, at least in the positive sequence. This can also be a provision of the power supply network or its operator.

[0003] For this purpose, it is necessary to determine the positive and negative sequence voltages in the power supply network. Such determination of the positive and negative sequence voltages can be a relatively slow process, at least in comparison with the short reaction times that are partly to be observed in the power supply network.

[0004] That is, if the grid voltage changes, in particular due to a fault, the determination of the positive and negative sequence voltages is sometimes not fast enough to detect this voltage change quickly enough, or is not detected accurately enough even after a short time. In particular, the positive and negative sequence voltages can usually only be detected for the duration of one period of the grid voltage. Thus, in the case of a 50 Hz grid, at least 20 ms are required to detect a change in the grid voltage.

[0005] Accordingly, the reactive current feed-in derived therefrom can only be increased relatively slowly. In particular, in the case of voltage-dependent reactive current feed-in, that is, in which the amplitude of the reactive current fed-in or the amplitude of the reactive power fed-in is dependent on the grid voltage, the reactive power feed-in cannot track voltage changes quickly enough. In other words, the increase in reactive current feed-in derived from the grid voltage occurs too slowly.

[0006] The German Patent and Trademark Office searched the following prior art in the priority application of the present application: DE 197 34 722 A1, DE 10 2015 112 155 A1 and CA 2 300 306 A1. Summary of the invention

[0007] The invention is therefore based on the object of solving at least one of the above-mentioned problems. In particular, a solution is to be proposed which implements the feeding of reactive current in both the positive and negative sequence as far as possible and nevertheless achieves the fastest possible reactive power tracking in the case of rapid voltage changes in the grid voltage. At least an alternative solution to the previously known solutions is to be proposed.

[0008] According to the invention, a method according to an embodiment is proposed. According to this, a method for feeding electrical power into a three-phase power supply network by means of an inverter device is proposed. In particular, the inverter device can include one or more inverters or can be an inverter. Preferably, the inverter device is part of a decentralized feeder, such as a wind energy plant, a wind farm or a photovoltaic facility.

[0009] The power supply network has a three-phase power supply voltage, which has a first, a second and a third power supply voltage phase. Therefore, the method for feeding is adapted to this three-phase power supply network. The three power supply voltage phases are considered jointly.

[0010] It is further proposed that if a fault-free operation has been identified for the power supply network, the method feeds in electric power in normal operation. In this normal operation, the positive sequence voltage and optionally the negative sequence voltage are detected from the power supply voltage, and the reactive current is preset at least according to the positive sequence voltage and optionally according to the negative sequence voltage. Therefore, it is proposed to decompose the three-phase power supply voltage according to the symmetrical component method. The result of this decomposition is basically a positive sequence component, a negative sequence component and a steady-state component. As often happens, the steady-state component is not important here, but the positive sequence component and the negative sequence component are important. It is sufficient to observe the positive sequence component here. Even if only the positive sequence component, i.e. the positive sequence voltage, is considered, a total observation of the three-phase power supply voltage is performed in any case.

[0011] Therefore, the reactive current fed in as a function of the positive sequence voltage is also set and fed in three phases. The reactive current is here dependent on the positive sequence voltage by taking into account its amplitude, in particular the effective value of the positive sequence voltage. In particular, the deviation of this amplitude from a reference voltage can be observed and the amplitude of the reactive current to be fed in can be determined as a function of this deviation.

[0012] In addition, it is proposed that if the voltage change of the power grid voltage meets the predetermined fault standard, then the fault operation is changed. In particular, if the voltage change exceeds a preset minimum change amount or a minimum change gradient amount. Therefore, the fault operation is clearly considered. The fault operation can be identified and thus also defined in the following manner: the voltage change of the power grid voltage meets the predetermined fault standard. That is to say, check whether the power grid voltage has a fault. In particular, if the voltage change exceeds a preset minimum change amount, it is concluded that there is a fault. Then, if the voltage change is greater than a predetermined limit value, there is a fault operation. Instead of the absolute change, the change speed, that is, the change gradient of the voltage can also be considered. For this purpose, it can be based on the minimum change gradient amount as a threshold. In particular, there is a fault operation when the power grid voltage drops suddenly. Accordingly, by definition, if the fault operation is not detected or not identified, it means that there is a normal operation.

[0013] It is now proposed that in fault operation, at least directly after the change from normal operation to fault operation, a reactive current is specified as a function of the space vector voltage.

[0014] Space vector voltages are familiar to those skilled in the art and are also calculated from the three grid voltage phases. A space vector voltage is a voltage vector that rotates over time in the complex plane and can also be referred to as a space vector or space vector voltage. However, such a space vector voltage cannot distinguish between symmetrical and asymmetrical components of the three grid voltage phases.

[0015] In particular, it is proposed that the space vector voltage defined by the following equation

[0016]

[0017] Therein, u1 , u2 and u3 are each intended as an instantaneous value, in particular as an instantaneous measured value of the first, second or third system voltage phase.

[0018] That is, all three grid voltage phases are included, and the value of the space vector voltage can be calculated immediately from the three instantaneous values ​​of the three grid voltage phases. That is, the space vector voltage occurs very quickly and can therefore provide a voltage value faster than the positive sequence voltage. Therefore, by transformation, for determining the reactive current, a transformation is made from a reference positive sequence voltage to a reference space vector voltage.

[0019] That is to say, it is proposed that at least directly after the change from normal operation to fault operation, a change is made to the reference space vector voltage. In particular, it has been recognized that the reactive power feed-in should react quickly to faults, which can also be referred to as errors. For this purpose, a change of the reference is proposed. After this rapid reaction, even if the fault or error continues to exist, it is possible to change back to the reference positive sequence voltage. In particular, it has been recognized that when an error occurs, the transition is critical and therefore a rapid reaction must be made. If the disturbance or error continues to exist, but does not cause a strong change, then the reference positive sequence voltage is sufficient or even more suitable.

[0020] According to one embodiment, it is proposed that the space vector voltage is also continuously determined in normal operation, and fault operation is detected based on the detected space vector voltage. Here, too, the knowledge is based on the fact that the space vector voltage can detect changes in the grid voltage more quickly or reflect them more quickly. Preferably, although the feed-in and also the reactive power feed-in are controlled based on the positive sequence voltage in normal operation, the space vector voltage can be detected together without having to influence the control. The space vector voltage that is thus continuously determined in normal operation can then be evaluated with regard to whether there is a transition to fault operation. Fault operation can thus be detected quickly.

[0021] According to one embodiment, it is proposed that if the positive sequence voltage has essentially already had a stable value and / or if the change in the positive sequence voltage is below a preset limit gradient in magnitude, then the preset reactive current according to the space vector voltage is converted back to the preset reactive current according to the positive sequence voltage. Therefore, the positive sequence voltage continues to be detected during the fault operation, which is also proposed as a feasible feature for the remaining embodiments. In particular, the positive sequence voltage continues to be detected regardless of whether normal operation or fault operation exists. If the positive sequence voltage changes slightly in magnitude, it can be assumed that normal operation or fault operation has passed. In order to evaluate whether the positive sequence voltage has only changed slightly, it is proposed to preset a limit gradient. This can also be called a limit gradient of the positive sequence voltage change. Preferably, the value of the limit gradient is 3% to 5% of the rated voltage for each cycle duration.

[0022] In this case, it is also considered that, depending on the system, the positive sequence voltage has less noise than the space vector voltage due to its type of detection. Therefore, the positive sequence voltage is very suitable for checking the gradient.

[0023] It has also been found that the change from normal operation to fault operation is time-critical and is preferably carried out in relation to the space vector voltage, which allows voltage changes to be detected more quickly than the positive sequence voltage. At the same time, it has been found that the change back to normal operation is less time-critical and that the positive sequence voltage can therefore be used well as a criterion.

[0024] Therefore, according to one embodiment, it is proposed to switch from normal operation to fault operation according to the space vector voltage, and to switch from fault operation back to normal operation according to the positive sequence voltage.

[0025] According to one embodiment, it is proposed that when the space vector voltage reaches a minimum value in the case of a voltage sag in one or more grid voltage phases, the reactive current preset according to the space vector voltage transitions back to the reactive current preset according to the positive sequence voltage. It is then assumed that a grid fault is perceptible as a voltage sag or a voltage sag. It has also been recognized that a reference to the space vector voltage is only required in a short transition range. In this case, the space vector voltage also sags rapidly and leads to a corresponding increase in the reactive current, which is calculated according to the space vector voltage. The space vector voltage, in particular when not all voltage phases sag, may have a variation curve that oscillates with the grid frequency, and it is therefore first important to quickly achieve rapid support by reactive power. The reactive power can then be set more accurately and the positive sequence voltage can also be used again for this purpose.

[0026] According to a variant, it is proposed that when the variation curve of the space vector voltage has reached the inflection point, the reactive current preset according to the space vector voltage is converted back to the reactive current preset according to the positive sequence voltage. If the voltage phase is no longer symmetrical after a fault, especially a voltage sag, for example because not all voltage phases have sagged, the space vector voltage will continue to oscillate. For calculating the reactive power, the average value of this oscillating space vector voltage can be important, and the space vector voltage reaches this average value approximately at its inflection point, that is, when the magnitude of its voltage drop reaches a maximum.

[0027] According to one variant, it is proposed that the reactive current presetting according to the space vector voltage is switched back to the reactive current presetting according to the positive sequence voltage when a predetermined transition time has elapsed after the detection of the faulty operation. In particular, it has been recognized that a transition time can be calculated in advance which can be in the range of one quarter of a grid period and thus allows a faster reaction. In other words, a switch back to a reference related to the positive sequence voltage can be made without evaluating measurements.

[0028] According to one embodiment, since the transformation to fault operation, the reactive current is preset according to the space vector voltage until the return standard has been identified. When the return standard is identified, the reactive current value calculated according to the space vector voltage is maintained as the space vector reactive current value. Since the return standard is identified, the positive-sequence reactive current value is continuously calculated according to the positive-sequence voltage, and the preset reactive current is calculated in the following manner: the maintained space vector reactive current value is transitioned to the positive-sequence reactive current value via a preset transition change curve.

[0029] At the beginning of fault operation, reactive current is therefore initially calculated only from the space vector voltage until a return criterion occurs or is detected. The return criterion can be, for example, that the space vector voltage has reached a minimum value in the event of a voltage drop in one or more grid voltage phases, or that the profile of the space vector voltage has reached a turning point, or that a predetermined transition time has elapsed.

[0030] If this criterion is reached, the last reactive current value, which has been calculated from the space vector voltage, is approximately frozen. From this last reactive current value, a transition is then made to the reactive current value which is calculated from the positive sequence voltage. This can also be referred to as a smooth transition. And it can proceed analogously, as will also be described below in conjunction with the transition from the positive sequence voltage to the space vector voltage or vice versa. When switching from the space vector reactive current value to the positive sequence reactive current value, the positive sequence reactive current value can also continue to change by continuous further calculations. In particular, the positive sequence reactive current value can still change in this case by the positive sequence voltage still changing.

[0031] According to one embodiment, it is proposed that the fault operation and / or the presetting of the reactive current according to the space vector voltage is carried out during a fault period, wherein the fault period is less than a grid cycle, in particular in the range of 5% to 50% of the grid cycle. The fault period can also correspond to a transition time. It has been recognized that after a grid cycle at the latest, the positive sequence voltage is again a good reference variable and should then be quickly transformed back to the positive sequence voltage. The fault period of the reference space vector voltage preferably does not include a smoothing period.

[0032] In addition or alternatively, it is proposed that in a smoothing transition period, a transition is made from presetting the reactive current according to the space vector voltage to presetting the reactive current according to the positive sequence voltage, which can also be referred to as smoothing. It is proposed that the smoothing transition period is shorter than a grid cycle, in particular in the range of 20% to 90% of a grid cycle. This is also based on the knowledge that the positive sequence voltage is determined within a grid cycle and therefore has a good value after a grid cycle has passed and should therefore be smoothly converted to a reference positive sequence voltage within a grid cycle.

[0033] According to one embodiment, it is proposed that during the transition to fault operation, a transition is made from prescribing a reactive current according to the positive sequence voltage to prescribing a reactive current according to the space vector voltage, as follows. The reactive current is prescribing according to a reference value, and the reference value is transitioned from the positive sequence voltage to the space vector voltage via a predefinable transition profile. This transition profile can also be referred to as a transition profile of the positive sequence voltage.

[0034] This is based on the knowledge that a rapid transition of the reactive current from one dependency to another dependency is important, but an abrupt transition can be serious and should be avoided. For this purpose, it is proposed that the reactive current is preset according to a reference value. In this context, the reference value is an artificial value, which transitions from the current value of the positive sequence voltage to the current value of the space vector voltage in particular within a transition period. The input variable for presetting the reactive current thus no longer changes abruptly, but can be transitioned particularly continuously and via a predeterminable transition curve.

[0035] A preferred possibility is that the predeterminable transition curve is linear. According to this, the value of the positive sequence voltage transitions linearly, i.e. along a straight line, to the value of the space vector voltage in the time view within a predetermined change period. Therefore, in the time view, the value of the positive sequence voltage is intuitively connected to the value of the space vector voltage via a straight line. Then, the reference value moves along this straight line in the change period in a time-dependent manner.

[0036] If the value of the positive sequence voltage and / or the value of the space vector voltage changes during the change time period, a linear transition change curve can be followed. That is, regarding the intuitive example of connection by means of a straight line, the two end points of the straight line connecting the space vector voltage and the positive sequence voltage can move together.

[0037] In addition or alternatively, it is proposed that a predeterminable transition curve is realized in the following manner: the reference value is composed of a positive sequence voltage with a first weight and a space vector voltage with a second weight, and the first weight decreases over time, while the second weight increases over time. In particular, it is proposed that the first weight decreases from 1 to 0 over time, while the second weight increases from 0 to 1 over time. In particular, the reference value can be an average value between the weighted positive sequence voltage and the weighted space vector voltage. If the first weight decreases from 1 to 0 here, while the second weight increases from 0 to 1 at the same time, a linear transition curve is generated. By using the first and second weights in this proposed manner, changes in the positive sequence voltage and / or the space vector voltage that occur during the transition time can also be taken into account in a simple manner. The first and second weights can also be referred to as first and second weights of the transition of the positive sequence voltage.

[0038] According to one embodiment, it is proposed that the reactive current is transferred from being preset according to the space vector voltage to being preset according to the positive sequence voltage. In this case, the reactive current can be preset according to a reference value, wherein the reference value is transferred from the space vector voltage to the positive sequence voltage via a preset transition curve. It is therefore proposed that the change in the correlation of the reactive current is returned from the space vector voltage to the positive sequence voltage via a preset transition curve. This transition curve can also be referred to as the transition curve of the space vector voltage. The variant of the implementation is carried out similarly to what is explained above in conjunction with the following embodiment, which describes how to transform from the correlation of the reactive current with the positive sequence voltage to the correlation of the reactive current with the space vector voltage.

[0039] In particular, even when the dependency of the reactive current on the space vector voltage is transformed into the dependency of the reactive current on the positive sequence voltage, it is proposed that the predeterminable transition curve is linear. In addition or alternatively, it is proposed that the predeterminable transition curve is realized in that the reference value is composed of the positive sequence voltage with a first weight and the space vector voltage with a second weight, and the first weight increases over time, in particular from 0 to 1, while the second weight decreases, in particular from 1 to 0. As a result, the dependency of the reactive current can also return from fault operation to normal operation without a sudden change. The first and second weights can also be referred to as first and second weights of the transition of the space vector voltage.

[0040] Instead of switching the reference value from the positive sequence voltage to the space vector voltage or vice versa as an indirect method, it is also possible as a direct method to calculate both reactive currents and then switch the predefined reactive current from one reactive current to the other.

[0041] Therefore, it is preferably proposed that during the transition to the fault operation, the reactive current preset according to the positive sequence voltage is transitioned to the reactive current preset according to the space vector voltage according to a preset transition curve, in particular the positive sequence reactive current is calculated as the reactive current according to the positive sequence voltage and the space vector reactive current is calculated as the reactive current according to the space vector voltage, and the preset reactive current is transitioned from the positive sequence reactive current to the space vector reactive current according to the preset transition curve, and in particular the preset transition curve is linear and / or the preset curve is realized in the following manner: the preset reactive current is composed of a positive sequence reactive current having a first weight and a space vector reactive current having a second weight in an additive manner, and the first weight decreases over time, in particular from 1 to 0, and the second weight increases, in particular from 0 to 1. The transition curve can also be referred to as a transition curve of the positive sequence reactive current. The first and second weights can also be referred to as first and second weights of the transition of the positive sequence reactive current.

[0042] In addition or alternatively, it is proposed that the reactive current preset according to the space vector voltage is transitioned back to the reactive current preset according to the positive sequence voltage in such a way that the preset reactive current is transitioned from the space vector reactive current back to the positive sequence reactive current according to a preset transition curve, and in particular the preset transition curve is linear, and / or the preset transition curve is realized in such a way that the preset reactive current is additively composed of a positive sequence reactive current with a first weight and a space vector reactive current with a second weight, and the first weight increases over time, in particular from 0 to 1, while the second weight decreases, in particular from 1 to 0. The transition curve can also be referred to as a transition curve of the space vector reactive current. The first and second weights can also be referred to as first and second weights of the conversion of the space vector reactive current.

[0043] The explanations about the direct method also apply analogously to the indirect method. In the direct method, the weights can be used and predefined like the weights in the indirect method. In both methods, the predefined transition curves can also be essentially the same, wherein of course the adaptation to the physical units must be taken into account.

[0044] According to one embodiment, when the measurement of the grid voltage is interrupted, the space vector voltage is estimated by a rotating voltage vector. To this end, it is proposed that the rotating voltage vector is calculated according to the value of the space vector voltage before the measurement of the grid voltage is interrupted, and the rotating voltage vector is further calculated according to the rated frequency of the grid voltage and / or according to the frequency of the last detected grid voltage. In particular, it is proposed that the magnitude and phase of the voltage vector adopt the value of the space vector voltage before the measurement of the grid voltage is interrupted, and continuously rotate at an angular frequency corresponding to the rated frequency or to the frequency of the last detected grid voltage. That is, the voltage vector rotates at an angular frequency ω and the angular frequency is calculated as ω=2*π*f N Or ω=2π*f, where f N φ denotes the rated frequency of the mains voltage and f denotes the frequency of the last detected mains voltage, which can also be referred to as mains frequency.

[0045] This is based in particular on the knowledge that, by using this voltage vector, even when the measurement of the grid voltage is interrupted, a voltage reference can be provided at least briefly in terms of magnitude and phase, which allows continued feeding in of power, in particular continued feeding in of reactive power. Interruptions in the measurement of the grid voltage can also be caused by the fact that the grid voltage actually collapses or at least drops briefly and for a short period of time, in particular within one second. In this case, it can be beneficial to feed reactive power into the power supply grid in spite of this in order to ensure grid support.

[0046] It is proposed that the current to be fed in and thus also the reactive current to be fed in are predefined and fed in according to the method of the tolerance band method. In the tolerance band method, the generated current is measured, the measurement or the measured value is fed back and the current is generated in dependence thereon. As a result, the desired current can be set and generated even in the event of fluctuations or even sudden drops in the grid voltage. Particularly preferably, if the space vector voltage is estimated by means of a rotating voltage vector, this is therefore proposed for the case where the measurement of the grid voltage is interrupted. However, the tolerance band method can also be used permanently for controlling the current to be fed in, that is to say also in normal operation.

[0047] In particular, it is proposed that a change is made from normal operation to fault operation in accordance with a voltage change of the grid voltage. In this case, in order to feed in reactive current in accordance with the positive sequence voltage, a change is made to feeding in reactive current in accordance with the space vector voltage. That is, in order to feed in reactive current, a change is made from a reference positive sequence voltage to a reference space vector voltage. This change can also be regarded as a first step, by means of which the reactive current can react quickly to voltage changes, in order to thereby achieve voltage stabilization in the power supply grid.

[0048] If the grid voltage still drops and / or the measurement of the grid voltage is interrupted, the space vector voltage can be estimated additionally from the rotating voltage vector. The reactive power feed-in is then referenced to the rotating voltage vector. Switching the reference to the rotating voltage vector can therefore be considered a second step. In this way, in particular, the first step and the second step together can advantageously enable rapid navigation through grid voltage errors.

[0049] According to the invention, a wind power plant for feeding electrical power into a three-phase power supply network is also proposed. For this purpose, the wind power plant has an inverter device. In addition, the wind power plant has a control device, and the wind power plant is configured to perform the feeding by means of a method according to at least one of the above-described embodiments. For this purpose, the corresponding method steps can be carried out in the control device in particular. Therefore, the control device is then configured to perform the feeding according to at least one of the above-described methods.

[0050] In particular, a measuring sensor is provided in the wind energy device or when the wind energy device is connected, and the measuring sensor detects the grid voltage in three phases. As a result, the positive sequence voltage and the negative sequence voltage as well as the space vector voltage can be determined. In particular, the positive sequence voltage or the space vector voltage can be used as an input variable for feeding in reactive power. The determination of the positive sequence voltage, the negative sequence voltage and the space vector voltage can be performed with the aid of a control device. The control device can also evaluate these voltage values ​​and perform a conversion between the positive sequence voltage and the space vector voltage as an input variable for starting reactive power. The control device can also control a transition for converting between the positive sequence voltage and the space vector voltage, or vice versa. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The invention will now be explained in more detail below on the basis of exemplary embodiments with reference to the accompanying drawings.

[0052] Figure 1 A perspective view of a wind power installation is shown.

[0053] Figure 2 The regulating structure is shown schematically.

[0054] Figure 3 Show Figure 2 elements of the regulatory structure.

[0055] Figure 4 and Figure 5 The diagrams each show a voltage profile and a reactive current profile calculated therefrom in each case in different ways. DETAILED DESCRIPTION

[0056] Figure 1 A wind power plant 100 is shown having a tower 102 and a nacelle 104. A rotor 106 having a spinner 110 and three rotor blades 108 is arranged on the nacelle 104. During operation, the rotor 106 is set into rotational motion by the wind and thereby drives a generator in the nacelle 104.

[0057] The wind energy device 100 has a generator 101 indicated in a pod 104. Electric power can be generated by means of the generator 101. A feed-in unit 105 is provided for feeding in electrical energy, which can be designed in particular as an inverter. As a result, a three-phase feed-in current and / or a three-phase feed-in voltage can be generated according to the amplitude, frequency and phase to be fed in at the grid connection point PCC. This can be done directly or also together with other wind energy devices in the wind farm. In order to control the wind energy device 100 and also the feed-in unit 105, an equipment control device 103 is provided. The equipment control device 103 can also obtain preset values ​​from an external park computer (Parkrechner), in particular a central park computer.

[0058] Figure 2 A control structure 202 is schematically shown, which controls an inverter 204 of a wind power installation 200. The wind power installation 200 can correspond to Figure 1 1. A wind power plant 100 is shown in FIG. 2. In particular, a wind power plant 200 is to be understood schematically and can include a control structure 202 or elements thereof. The control structure 202 can be included in a control device 206 or implemented therein. The inverter 204 can feed active power P and reactive power Q into a power supply grid 208.

[0059] According to one embodiment, the proposed method has the effect of measuring the three voltage phases u1, u2 and u3 of the power supply network 208. This can be done with the aid of a measurement recorder 210 in the power supply network 208 or as in Figure 2 As shown in FIG. 2 , this occurs at a connecting line 212 between inverter 204 and supply grid 208 .

[0060] In normal operation, the three phase voltages are decomposed or converted into positive sequence voltages u in decomposition block 214 according to the symmetrical component method. + and negative sequence voltage u - The positive sequence voltage u + is input into the transformation block 216. Optionally, the negative sequence voltage u - can also be input into the transform block 216, which is indicated by the corresponding dashed arrow.

[0061] In the conversion block 216, the positive sequence voltage u can be selected + As the voltage to be used for determining the reactive power. This is indicated by the corresponding switch position in the transformation block 216. The transformation block 216 then outputs the voltage to be used for controlling the reactive power as a reference value or reference voltage u ref .

[0062] Therefore, the transform block 216 Figure 2 The switch positions indicated in relate to normal operation, in which the reactive power control is dependent on the positive sequence voltage u + That is, the reference voltage u ref Corresponding to the positive sequence voltage u + .

[0063] For further implementation, the reference voltage u ref As an input variable, it is provided to the reactive power block 218. In the reactive power block 218, the desired reactive power is used as an input variable of the reactive power block 218, that is, the reference voltage u ref The relevant function is used to calculate and is used as the expected reactive power Q s Then, the expected reactive power Q s The reactive power expected value Q forms the input value of the inverter 204. s A reference variable is formed for controlling the inverter.

[0064] Still more variables are needed to control the inverter 204, but they are not shown here for simplicity. The inverter 204 can be supplied with energy via a DC voltage on the input side. This DC voltage can be obtained, for example, from a generator of the wind power plant 200, which generates rectified AC power.

[0065] In the case where normal operation must or should be left and fault operation should be used, the three phase voltages u1, u2, u3 are converted into space vector voltages in the conversion block 220. The space vector voltage The same is input to the transformation block 206. If necessary, ie in particular in fault operation, the transformation block 216 can transform to this space vector voltage via the indicated switches As an input variable for reactive power control. However, in this regard, the switches indicated in the transformation block 216 are only for illustration. In fact, it is proposed that, instead of the positive sequence voltage u + and space vector voltage Instead of hard switching between the two, the transformation is performed by a transition function. This will be explained in detail hereinafter by way of example, and possible previous or subsequent explanations on the transformation can be implemented in the transformation block 216.

[0066] One possible implementation of the transform block 216 is Figure 3 Here, the reference voltage u ref From the positive sequence voltage u + to space vector voltage The linear transformation of is implemented by weighting functions g1(t) and g2(t) as mathematical functions. Therefore, the first weighting function g1(t) decreases linearly from 1 to 0 in the transition period T, so that the positive sequence voltage u + The share of decreases from the maximum value to 0 in the transition period. At the same time, the weighting function g2(t) increases from 0 to 1 in the transition period T, so that the share of the space vector voltage increases from 0 to the maximum value in the transition period.

[0067] Accordingly, even if the reference voltage u ref The space vector voltage Convert back to positive sequence voltage u + , the same mathematical relationship can also be implemented, where the weight functions g1(t) and g2(t) may need to be swapped. However, it is also possible to use a shorter or longer transition period for the switch back, which is an example of a variant.

[0068] Figure 4 The feed-in of reactive current of the positive and negative sequence is shown. Since the determination of the positive and negative sequence voltage is not possible immediately but takes place over the duration of a period of the grid voltage, the increase in reactive current feed-in derived therefrom can only take place "slowly". This is Figure 4 In the description.

[0069] Figure 4Three separate diagrams are included, of which the upper diagram shows the profile of the three measured phase voltages u1, u2 and u3, ie the profile of the instantaneous values. At time t1, the two phase voltages u2 and u3 drop to low values.

[0070] The middle diagram shows the positive sequence voltage u + and space vector voltage It is calculated from the three phase voltages u1, u2 and u3 and normalized in the diagram to the rated voltage U N It can be seen that at time point t1, before the voltage sag, the positive sequence voltage u + and space vector voltage are approximately the same, at least indistinguishable in the diagram. Therefore, the three phase voltages u1, u2 and u3 are still approximately symmetrical to each other. After the voltage dip, the space vector voltage However, due to the principle, the space vector voltage Keep vibrating.

[0071] Positive sequence voltage u + The reaction to the voltage dip is slower and reaches the new value after a period T, ie at time t2.

[0072] Figure 4 The lower graph shows the positive sequence voltage u + The reactive current for feeding in reactive power determined in the above is called the desired reactive current I QS Therefore, the expected reactive current I QS The new value is also reached only after the period T, ie at the time t2. Figure 4 The same applies to the lower graph of Figure 5 For illustration purposes, assume that the expected reactive current reaches the maximum reactive current I QM , the diagram is normalized to the maximum reactive current.

[0073] In order to provide reactive current faster, a variant is proposed, which Figure 5 Illustration in Chinese. Figure 5 Three graphs are shown, and the upper and middle graphs are Figure 4 The upper and middle diagrams correspond, except for the deviation in time resolution, which is not significant here. Figure 4 Explanation Figure 5 The upper and middle graphs of .

[0074] However, the calculation of reactive current is Figure 4 and 5Therefore, the lower graph shows the calculated expected reactive current I QS , as in Figure 4 As in the lower diagram in , but where the desired reactive current I QS Calculated in different ways.

[0075] It is proposed that, based on the measured space vector voltage To perform the initial reactive current supply, i.e. initially from the detected fault criterion. Figure 5 The fault criterion in the example shown in is a voltage drop of the two phase voltages u2 and u3 to a small value. Therefore, this initial reactive current supply takes place starting from the time t1.

[0076] Therefore, starting from time t1, according to the measured space vector voltage Once the space vector voltage is identified, the initial reactive current supply, i.e. the desired reactive current, is determined. The first minimum in , freezes the expected value thus obtained. This is roughly at time t E This is the case. Then, a smooth transition from the desired value based on the space vector to the desired value based on the positive sequence can be achieved within a time period of 15 ms. The time period of 15 ms is slightly shorter than the period T, which is 20 ms in this case, because it is based on a 50 Hz power grid.

[0077] It has been recognized that in the event of an error, which can also be referred to as a fault event, it can be sensible to feed in reactive current more quickly, ie as far as possible within the first 10 ms.

[0078] Compared to what has been known so far in the case of reactive power feed-in based solely on the positive sequence voltage, the solution according to the invention makes it possible to provide reactive current more quickly for grid support, thereby also making it possible to achieve a higher grid stability.

[0079] It is therefore proposed to use the space vector voltage to detect errors and then switch to the space vector voltage for reference accordingly. This is faster than referencing the positive sequence. If the sequence is stable, it is possible to switch back to referencing the positive sequence.

[0080] It is a known problem to achieve a smooth transition from the reference positive sequence to the reference space vector voltage and back again. For this purpose, a smooth transition is proposed.

[0081] The smooth conversion from the reference space vector voltage to the reference positive sequence, i.e. the positive sequence voltage, can start at the minimum value of the space vector voltage. Other feasibility is also considered, such as only referencing the space vector voltage for a preset time. Such a preset time can be, for example, a quarter of the grid cycle. It is also possible to evaluate the space vector voltage and refer to the space vector voltage until it has a turning point, and then convert it to the reference positive sequence voltage.

Claims

1. A method for feeding electrical power into a three-phase power supply network by means of an inverter device (105, 204), wherein The power supply grid (208) has a three-phase grid voltage, wherein the three-phase grid voltage has a first grid voltage phase, a second grid voltage phase and a third grid voltage phase. The method comprises the following steps: If fault-free operation has been detected for the power supply grid (208), the electrical power is fed in during normal operation, wherein during normal operation: - Detecting the positive sequence voltage (u + ),and - According to the positive sequence voltage (u + ) preset reactive current, and If the voltage change of the grid voltage meets a predetermined fault criterion, a switch is made to a fault mode, wherein in the fault mode at least immediately after the switch - According to the space vector voltage The reactive current is preset, The space vector voltage Defined by the following equation: Among them, u1, u2 and u3 are instantaneous values, namely, instantaneous measured values ​​of the first grid voltage phase, the second grid voltage phase and the third grid voltage phase, respectively.

2. The method according to claim 1, characterized in that The space vector voltage is also continuously determined during normal operation. And according to the detected space vector voltage Switch to the fault operation.

3. The method according to claim 1 or 2, characterized in that: - If the positive sequence voltage (u + ) adopts a fixed value, and / or if the positive sequence voltage (u + ) is below a predeterminable limit gradient, and / or - If in the event of a voltage dip in one or more grid voltage phases the space vector voltage The minimum value has been reached, and / or - If the space vector voltage The change curve has reached a turning point, and / or - if a predetermined transition time has elapsed after the detection of the faulty operation, - then the space vector voltage The preset transition of the reactive current is back to: - According to the positive sequence voltage (u + ) preset the reactive current.

4. The method according to claim 1 or 2, characterized in that: - Since the transition to the fault operation, according to the space vector voltage The reactive current is preset until a return criterion is identified, - When the return criterion is identified, the space vector voltage The calculated reactive current value is kept as the space vector reactive current value. -Since the return standard is identified, according to the positive sequence voltage (u + ) continuously calculates the positive sequence reactive current value, and The predetermined reactive current is calculated by transitioning the stored space vector reactive current value to the positive sequence reactive current value via a predeterminable transition curve.

5. The method according to claim 1 or 2, characterized in that: - performing the fault operation during the fault period and / or according to the space vector voltage The reactive current is preset, wherein the fault time period is less than the grid cycle, and / or - During the smooth transition period, the space vector voltage The reactive current is preset to transition to the positive sequence voltage (u + ) presets the reactive current, wherein the smooth conversion time period is less than the grid cycle.

6. The method according to claim 1 or 2, characterized in that: When the transformation is run for the fault - will be based on the positive sequence voltage (u + ) The reactive current is preset to transition to: - According to the space vector voltage The reactive current is preset, The method is: - presetting the reactive current according to a reference value and changing the reference value from the positive sequence voltage (u + ) transitions to the space vector voltage and / or - will be based on the space vector voltage The reactive current transition is preset to return to: - According to the positive sequence voltage (u + ) preset the reactive current, The method is, The reactive current is preset according to a reference value, and the reference value is changed from the space vector voltage via a predeterminable transition curve. Transition to the positive sequence voltage (u + ).

7. The method according to claim 1 or 2, characterized in that: When changing to the fault runtime - According to the preset transition change curve, the positive sequence voltage (u + ) Preset the reactive current transition to - According to the space vector voltage The reactive current is preset, So that, - According to the positive sequence voltage (u + ) calculates the positive sequence reactive current as the reactive current, - According to the space vector voltage Calculate the space vector reactive current as reactive current, - transitioning the preset reactive current from the positive sequence reactive current to the space vector reactive current according to a preset transition curve, and / or - the predeterminable transition curve is realized in the following way: the predetermined reactive current is composed of the positive sequence reactive current with a first weight and the space vector reactive current with a second weight, and the first weight decreases over time while the second weight increases, and / or - will be based on the space vector voltage The reactive current transition is preset to return to: - According to the positive sequence voltage (u + ) preset the reactive current, The method is: - transitioning the predeterminable reactive current from the space vector reactive current back to the positive sequence reactive current according to a predeterminable transition curve, and / or The predefinable transition curve is achieved in that the predefinable reactive current is additively composed of the positive sequence reactive current with a first weighting and the space vector reactive current with a second weighting, and the first weighting increases over time while the second weighting decreases.

8. The method according to claim 1 or 2, characterized in that: - when the measurement of the grid voltage is interrupted - Estimation of the space vector voltage by rotating the voltage vector and - The rotating voltage vector - according to the space vector voltage before interrupting the measurement of the grid voltage , and according to - the expected frequency of the grid voltage and / or further calculation based on the last detected frequency of the grid voltage.

9. The method according to claim 1, characterized in that: If the voltage changes by more than a predefinable minimum change or a minimum change gradient, a switch is made to a fail-safe mode.

10. The method according to claim 5, characterized in that The fault time period is in the range of 5% to 50% of the grid cycle.

11. The method according to claim 5, characterized in that The smooth transition time period is in the range of 20% to 90% of the grid cycle.

12. The method according to claim 6, characterized in that - the predeterminable transition curve is linear, and / or The predeterminable transition curve is realized in the following way: the reference value is composed of the positive sequence voltage (u + ) and a space vector voltage with a second weight The first weight decreases over time, while the second weight increases.

13. The method according to claim 12, characterized in that The first weight decreases from 1 to 0.

14. The method according to claim 12, characterized in that The second weight increases from 0 to 1.

15. The method according to claim 7, characterized in that The predeterminable transition curve is linear.

16. The method according to claim 7, characterized in that The first weight decreases from 1 to 0.

17. The method according to claim 7, characterized in that The second weight decreases from 0 to 1.

18. The method according to claim 1, characterized in that In the normal operation - Detecting the negative sequence voltage (u - ),and - According to the negative sequence voltage (u - )Preset reactive current.

19. A wind power installation (100, 200) for feeding electrical power into a three-phase power supply network by means of an inverter device, wherein The power supply grid (208) has a three-phase grid voltage, the three-phase grid voltage having a first grid voltage phase, a second grid voltage phase and a third grid voltage phase, and The wind power installation has a control device (206), and the control device (206) is configured to control the feed-in according to the following steps: If fault-free operation is detected for the power supply grid (208), electrical power is fed in during normal operation, wherein during normal operation - Detecting the positive sequence voltage (u + ),and - According to the positive sequence voltage (u + ) preset reactive current, and If the voltage change of the grid voltage meets a predetermined fault criterion, a switch is made to a fault mode, wherein in the fault mode at least immediately after the switch - According to the space vector voltage The reactive current is preset, The wind power installation (100, 200) is configured to carry out the method according to any one of claims 1 to 17.

20. The wind energy installation (100, 200) according to claim 19, wherein: If the voltage changes by more than a predefinable minimum change or a minimum change gradient, a switch is made to a fail-safe mode.

21. The wind energy installation (100, 200) according to claim 19, wherein in normal operation: - Detecting the negative sequence voltage (u - ),and - According to the negative sequence voltage (u - )Preset reactive current.

Images