AT521551A1 - Device for stabilizing and filtering interference in a multi-phase AC network - Google Patents

Abstract

The invention relates to a device for stabilizing and filtering out disturbances within a predetermined frequency range in an at least three-phase AC network (100), the device comprising: - phase connections (200a, 200b, 200c) for connection to the cable connections (111, ... ) of the AC voltage network (100), - an inverter (203) with an intermediate circuit capacitor (C7), - a filter (202) arranged between the inverter (203) and the phase connections (200a, 200b, 200c) for filtering out the inverter (203 ) outgoing high-frequency switching components, - measuring devices (Va, Vb, Vc) for measuring the voltage profile (uNa, uNb, uNc) at the phase connections, - measuring devices (Aa, Ab, Ac) for measuring the current profile (iLa, iLb, iLc) between the inverter (203) and the phase connections (200a, 200b, 200c), - a control device (204), which stores the individual measured values (among others, ub, uc, uDC ia, ib, ic) of the measurement device (Va, Vb, Vc, VDC, Aa, Ab, Ac) are supplied, wherein - the control device (204) has a compensation unit which is designed to derive from the determined voltage profiles (ia, ub, uc; ud, uq) filter out the frequency component caused by the nominal network voltage, possibly also further frequency components, preferably arithmetically, and create an interference voltage curve (usa, usb, usc; usd, usq), and - the control device (204) is designed to convert the inverter (203) to be controlled with switching signals (s1, s2, s3) in such a way that the current profile (ia, ib, ic) between the inverter (203) and the phase connections (200a, 200b, 200c) corresponds to the current profile, if appropriate calculated, occurs when the disturbance voltage curve (usa, usb, usc; usd, usq) is fed to a compensation unit (305, 405) with an impedance defined by a predetermined transmission characteristic, which is reduced, preferably the lowest, in particular at a predetermined network resonance frequency to be compensated ,

Description

The invention relates to a device for stabilizing and filtering out

Interference within a predetermined frequency range in an at least three-phase AC network (100), the device comprising:

- phase connections (200a, 200b, 200c) for connection to the cable connections (111, ...) of the AC network (100),

- an inverter (203) with an intermediate circuit capacitor (C7),

a filter (202) arranged between the inverter (203) and the phase connections (200a, 200b, 200c) for filtering out high-frequency switching components originating from the inverter (203),

Measuring devices (V _a , V _b , V _c ) for measuring the voltage _profile (u _Na , u _N b, u _Nc ) at the phase connections,

- Measuring devices (A _a , A _b , A _c ) for measuring the current profile (i _La , ii_b, ii_c) between the inverter (203) and the phase connections (200a, 200b, 200c),

- A control device (204) which stores the individual measured values (u _a , u _b , u _c , u _D ci _a , ib, ic) of the measuring devices (V _a , V _b , V _c , V _D c, A _a , A _b , A _c ) are supplied, wherein

- The control device (204) has a compensation unit, which is designed to determine from the determined voltage profiles (u _a , u _b , u _c ; u _d , u _q ) the frequency component caused by the nominal network voltage, possibly also further frequency components, preferably arithmetically, filter and create an interference voltage curve (u _sa , u _S b, u _sc ; u _S d, u _sq ), and

- The control device (204) is designed to carry the inverter (203)

To trigger switching signals (si, s ₂ , s ₃ ) in such a way that the current profile (i _a , i _b , i _c ) between the inverter (203) and the phase connections (200a, 200b, 200c) corresponds to the current profile, if appropriate calculated , which occurs when the interference voltage curve (u _sa , u _S b, u _sc ; u _S d, u _sq ) is sent to a compensation unit (305, 405) by a predetermined one

Transmission characteristic of fixed impedance is supplied, which is reduced, preferably the lowest, in particular at a predetermined network resonance frequency to be compensated.

The invention relates to a device for stabilizing and filtering

Malfunctions in a multi-phase AC network.

The problem in the prior art is that local resonance phenomena can occur in three-phase or multiphase networks due to parasitic capacitances, such as are present, for example, in long cables or in transformers. These resonance phenomena can lead to network failures or the destruction of electrical components in the network concerned. For example, due to long cable distances or an increased proportion of cables in the network, there may be an increased proportion of inductance and capacitance. This is particularly the case when individual components in the network, such as Open-space PV systems, wind farms or offshore parks are located far away from the actual main load connection or supply.

The object of the invention is to provide improved network stability for three-phase or multi-phase networks. For this purpose, the invention provides a device for stabilizing the network, which - attached to a node or to a cable of the network - brings about a local stabilization of this network.

For this purpose, the invention provides a device for stabilizing the network, in which the control device has a compensation unit which is designed to filter out the frequency component caused by the nominal network voltage, possibly also further frequency components, preferably arithmetically, from the determined voltage profiles and an interference voltage profile and the control device is designed to control the inverter with switching signals in such a way that the current profile between the inverter and the phase connections corresponds to the current profile, if necessary calculated, that occurs when the interference voltage profile is sent to a compensation unit with a predetermined transmission characteristic Impedance is supplied, which is reduced, preferably the lowest, in particular at a predetermined network resonance frequency to be compensated.

2/38 ¹

By arranging a large number of such stabilizing devices, the latter can

Effect improved or achieved for a large number of nodes in the network concerned.

A further improvement to stabilize the intermediate circuit voltage provides that a measuring device for measuring the intermediate circuit voltage is provided on the intermediate circuit capacitor of the inverter and / or that the control device, which is optionally connected downstream of the measuring device for measuring the intermediate circuit voltage, is designed to switch the inverter with switching signals to be controlled such that the intermediate circuit voltage corresponds to a predetermined constant value.

The invention also provides a method for stabilizing the network in order to achieve the object set forth above, in which the frequency component caused by the nominal network voltage, optionally also further frequency components, preferably arithmetically, is filtered out from the determined voltage profiles, and an interference voltage profile is created in this way, and Inverter is controlled with switching signals in such a way that the current profile between the inverter and the phase connections corresponds to the current profile, if necessary calculated, that occurs when the interference voltage profile is fed to a compensation unit with an impedance defined by a predetermined transmission characteristic, in particular at a predetermined one compensating network resonance frequency is reduced, preferably the lowest.

A further improvement to stabilize the intermediate circuit voltage provides that the intermediate circuit voltage is measured at the intermediate circuit capacitor of the inverter and / or that the inverter is controlled with switching signals in such a way that the intermediate circuit voltage corresponds to a predetermined constant value.

A preferred adjustment of the filter characteristic provides that for the case that currently no exist for compensating disturbances, for different frequencies measured the impedance of the network and / or the voltage response at a / 38 ² predetermined current excitation at a predetermined frequency and for the case that the determined impedance and / or voltage exceeds a predetermined impedance or voltage threshold value, the transmission characteristic of the compensation unit is adapted to this frequency, the impedance of the compensation unit being reduced, in particular the least, at this frequency.

In order to avoid harmful influences on the network, in particular further resonance effects, it can be provided that the current intensity used for the current excitation is gradually increased in the individual measurements until a voltage threshold value is exceeded at least at one frequency and, if necessary, subsequently for the current excitation is reduced in steps until the voltage threshold value is exceeded at only one frequency.

A preferred embodiment and several variants of the invention are illustrated in more detail with the aid of the following drawing figures.

1 schematically shows an electricity network with a plurality of stabilization devices according to the invention. 2 schematically shows an embodiment of a stabilization device according to the invention. 3 shows a variant of a transformer used in FIG. 2 in detail. FIG. 4 shows a variant of a filter circuit used in FIG. 2 in detail. FIG. 5 shows a typical filter characteristic of a filter shown in FIG. 4. FIG. 6 shows an alternative embodiment for the filter shown in FIG. 4 with additional common mode filtering. FIG. 7 shows a first variant of an inverter used in FIG. 2 in detail. FIG. 8 shows a second variant of an inverter used in FIG. 2 with a neutral conductor connection in detail. FIG. 9 shows a variant of a control circuit used in FIG. 2 in detail. 10 shows a further variant of a control circuit used in FIG. 2 in detail. 11 and 12 show different embodiments of phase-locked loops that can be used in the control circuits shown in FIGS. 9 and 10.

4/38

Fig. 13 shows a filter which can be used in a further embodiment of the invention with a four-phase network and which enables active control of the neutral conductor potential. 14 shows an inverter which can be used in a further embodiment of the invention with a four-phase network and which enables active control of the neutral conductor potential.

1 shows such an electricity network 100 with a central generator station 101, from which a number of multi-phase, in the present case, three-phase, outlets cable connections 111-129 originate. These cable connections 111-129 are connected to one another either via distributors 140-144 or transformer stations 151-153. To avoid instabilities, active stabilization devices 200A, 200B are arranged in the network 100 on individual cable connections 111-129. These are designed to feed current into the network 100 in a predetermined manner via the cable connection 111-129 or to withdraw it from the network 100 via the cable connection 111-129. The control of the current feed or withdrawal takes place on the basis of a voltage measurement on the relevant cable connection 111-129.

A preferred embodiment of a stabilization device 200; 200a, 200b is shown schematically in more detail in FIG. 2. The stabilization device 200 comprises a multi-phase transformer 201, which is shown in more detail in FIG. 3 and which is connected to a cable connection of the network 100. The transformer has the function of transforming the voltage prevailing on the network 100 into a voltage which is advantageous for the stabilization device 200. Typically, a stabilization device 200 or its associated inverter operates with a phase-phase voltage of 400 Vrms, while medium and high voltage networks typically have voltages of up to approximately 110 kV or more than 110 kV. The use of a transformer 201 is basically optional: if the other components of the stabilization device 200 are matched to the voltage prevailing on the network 100, the transformer 201 can also be omitted.

5/38

As can be seen in FIG. 3, a transformer 201 can be operated, for example, in a Y-Y circuit. The transformer 201 comprises an iron core with three legs 201 a, 201 b, 201 c, on each of which a coil 2011 a, 2011 b, 2011 c of the primary side and a coil 2012a, 2012b, 2012c of the secondary side are arranged. A connection of each of the coils 2011a, 2011b, 2011c on the primary side is connected to a common star point 2011S on the primary side. The other connections of the coils 2011a, 2011b, 2011c on the primary side are each connected to a phase 111a, 111b, 111c of the network 100.

In each case one connection of each of the coils 2012a, 2012b, 2012c on the secondary side is connected to a common star point 2012S on the secondary side. The remaining connections of the coils 2012a, 2012b, 2012c on the secondary side are each connected to one phase of the filter 202 (FIG. 4).

The filter circuit addressed in FIGS. 2 and 3 is shown in more detail below with reference to FIG. 4. If, as is the case in FIG. 2, a transformer 201 is present at the input of the stabilization circuit 200, the inductance at the output of the transformer can also be used for filter purposes.

In the present case, FIG. 4 shows a filter circuit 202 in the form of a three-phase push-pull output filter stage. At its input, this has an input inductance L _1a , L _1b , L _1c with a parasitic resistance R _L ia, Ri_ib, Rlic. When dimensioning the stabilization circuit, care must also be taken to ensure that the individual coils 2012a, 2012b, 2012c coupled to one another on the secondary side of the transformer 201 have a specific inductance L _1a , L _1b , L _1c and also with parasitic resistances R _L ia, Ri_ib, Rlic are affected, so that this part of the filter circuit 202 can possibly already be implemented by the transformer 201. In the present exemplary embodiment, the inductances l_i _a , Li _b , l_i _c each have a value of 10 pH (micro-Henry). The parasitic resistances R _L ia, Ri_ib, Rlic each have a value of 10 πίΩ (MilliOhm).

6/38

The inductors Li _a , Li _b , Li _c are via further inductors l_ _2a , l_ ₂ b, L _2c , which are realized in the present case by real coils with a parasitic resistance R _L2a , Ri_2b, R _L2c , at the output for the Inverter 203 of the filter circuit 202 connected. The inductors L _2a , L _2b , L _2c each have a value of 500 pH (micro-Henry), the parasitic resistances R _L2a , Ri_2b, Rl2c on these coils have a value of 10πΩ (milli-ohms).

Current measuring devices A _a , A _b , A _{c are also} provided at this output, which measure the respective current flowing into the inverter 203 and continuously provide current measured values i _La , ii_b, ii_c. Due to the interconnection of the current measuring devices A _a , A _b , A _c , in the case of a three-phase network without a neutral conductor, it is sufficient to carry out only two of the three current measurements, since the relationship i _La + kb + ii_c = 0 results from two current measured values the third occurring current can be calculated.

Furthermore, the filter circuit 202 comprises three of the phase lines in nodes P _a , P _b , P _c between the respective inductances I_i _a , L _2a ; Li _b , L _2b ; I_i _c , L _2c outgoing capacitors Ci _a , Ci _b , Ci _c , which are brought together at a common star point 202S. The capacitors Ci _a , Ci _b , Ci _c each have parasitic resistances R _C i _a , Reib, Rcic in the present case. The capacitors Ci _a , Ci _b , Ci _c typically have a capacitance of 1pF (micro farad), the parasitic resistances on the capacitors R _C i _a , Reib, Rcic have a value of 50mD (milli-ohms).

At each of the capacitors RCi _a , RCi _b , RCi _c between the nodes P _a , P _b , P _c and the star point 202S, a phase voltage u _Na , u _Nb is measured by means of a voltage measuring device V _a , V _b , V _c , u _Nc measured, which essentially corresponds to the respective phase voltage in the network 100. Voltage _measured values u _Na , u _N b, u _{Nc are continuously} determined. Due to the connection of the voltage measuring devices V _a , V _b , V _c , in the case of a three-phase network without neutral conductor, it is sufficient to carry out only two of the three voltage measurements, since due to the relationship u _Na + u _Nb + u _Nc = 0, two voltage measurement _values each result the third voltage can be calculated.

7/38

As an alternative to the measurement or connection shown, there is also the possibility that the voltage measurements and the capacitors Ci _a , Ci _b , Ci _{c of} the filter arrangement 202 are connected in a delta connection instead of in the star connection shown. With a delta connection, the individual sizes of the star connection can be calculated.

Through this specific interconnection of inductors L _1a , L _2a ; L _1b , L _2b ; L _1c , L _2c and capacities Ci, C ₂ , C3, a push-pull output filter stage is implemented in the filter circuit 202. This output filter stage achieves a filter characteristic with which the switching frequency components of the inverter connected to the filter circuit 202 resulting from the switching processes are damped and are not passed on to the network. An exemplary filter characteristic for the above-described and dimensioned filter circuit 202 is shown in more detail in FIG. 5. From this it can be seen that high-frequency components, which usually occur during the operation of an inverter 203, do not get into the network 100, while the resistance in the range of the nominal network voltage is very low.

In addition, it can also be provided that the filter circuit 202 additionally has a common mode filter stage (FIG. 6). Such a common mode filter stage is basically known from the prior art. In the simplest case, this can be implemented by a number of coils L _3a , L _3b , L _3c which are coupled to one another and which are coupled to one another in the same direction. In addition, capacitors C _2a , C _2b , C _2c can also be provided, via which the phase lines are grounded. Parasitic resistances which are present in the present filter circuit are not shown in FIG. 6 for reasons of space. The voltage measurement can alternatively also be carried out at the respective other connection of the coils coupled to one another, ie at the node P _a ', P _b ' P _c 'or at the input of the filter 202.

7 shows a preferred exemplary embodiment of an inverter 203. In the present exemplary embodiment, the inverter is made up of a total of three half bridges, each comprising two switches S _a i, S _a2 ; S _b i, S _b2 ; S _c i, S _C 2 realized, each via separate control signals s _a1 , s _a2 ; s _b1 , s _b2 ; s _c i, s _C 2/38 ^{7 are} switched. Each of the phase connections of the inverter 203 is connected to the center connection of one of the half bridges. The external connections of the half bridges are each connected to one of the connections of an intermediate circuit capacitor Cdc, which in the following exemplary embodiment has a capacitance of 500 pF (micro farads) and a parasitic series resistance of 20 πΩ (milli-ohms).

The individual switches S _a i, S _a 2i Sbi, Sb2i S _c i, S _C 2 can be activated in different ways; In the present exemplary embodiment, a pulse-width-modulated signal with a basic frequency of approx. 100 kHz is used, the two switches of each half-bridge being each acted on by a different switching signal, i.e. when one switch is closed, the other is opened, so that by selecting the switch position as Potential at the phase, the potential at one of the two connections of the intermediate circuit capacitor C _D c can be selected.

Implementation of the inverter 202 is by no means absolutely necessary in half bridges. Rather, alternative variants for fixing the voltage can also be carried out, such as multi-stage switch cells, for example NPC, T-type, flying capacitors, etc. In addition, multi-cell systems can also be used in order to be able to switch high voltages and thus achieve a higher control bandwidth ,

In order to be able to stabilize the intermediate circuit voltage u _D c, it is provided that a further voltage measuring device V _D c is provided on the intermediate circuit capacitor Cdc, which continuously determines the intermediate circuit voltage and provides the relevant measured values u _DC for the intermediate circuit voltage.

Finally, if the network 100 provides a neutral conductor, there is also the possibility of also connecting this to the inverter 202 in order to be able to compensate for uneven loads on the individual phases. Such an alternative embodiment of a four-phase inverter with a neutral conductor connection is shown in more detail in FIG. 8. In this case, the neutral conductor of the network is direct or via the transformer 201 which may be present

9/38 led to the neutral conductor input of the inverter 203. If necessary, measures to suppress interference effects on the neutral conductor can also be carried out on the filter circuit.

In contrast to the inverter shown in FIG. 7, the inverter 203 shown in FIG. 8 has a total of two intermediate circuit capacitors C _D ci, C _D c2 connected, the voltage measuring unit V _D c the intermediate circuit voltage at the series connection of the intermediate circuit capacitors Cdci, Cdc2 measures. The inverter 203 has a neutral conductor connection, which is connected to the center between the two intermediate circuit capacitors C _D ci, C _D c2.

In addition, the inverter 203 has a balancing unit 2031, which in each case ensures that the voltages at the two intermediate circuit capacitors C _D ci, C _D c2 are the same. In the simplest case, the voltages are determined passively using two resistors. The first resistor is at one end of the series circuit of the intermediate circuit capacitors Cdci, C _D c2 and at the center of the intermediate circuit capacitors C _D ci, C _D c2, the second resistor is at the other end of the series circuit of the intermediate circuit capacitors Cdci, Cdc2 and again at the center of the intermediate circuit capacitors Cdci, Cdc2 connected.

Alternatively, the balancing unit 2031 can, for example, use voltage measuring units (not shown) to measure the voltages at the two intermediate circuit capacitors Cdci, Cdc2 to regulate the potential, preferably by means of a half-bridge circuit.

The control and stabilization measures for the control of the inverter 203 described below in FIGS. 9 and 10 can also be used for the four-phase embodiment described in FIG. 8.

FIG. 9 shows a first preferred control circuit 204, the input of which is supplied with the individually determined voltage and current measured values i _La , kb, kc, u _Na , u _Nb , u _Nc , u _D c and which control signals s _1a , s _2a ; s _1b , s _2b ; s _1c , s _2c for the, 9

10/38

Activation of the switches Si _a , S2 _a ; Sw, S2b; Sw, 82c or the inverter 203 created.

The voltage _measured values u _Na , u _N b, u _Nc are _fed to a phase-locked loop 301 (FIGS. 11, 12), which has at its output the phase angle 0 _N , the frequency or angular velocity ω _Ν and a fundamental oscillation peak value | ύ _Ν | which indicates voltage _values u _Na , u _Nb , u _Nc . If, as in the present case, a mains frequency of approximately 50 Hz is used, the angular velocity ω _{Ν has} a value of 2 ir50 Hz.

The phase-locked loop 301 can be implemented in different ways. A very simple case will be described with reference to Figs. 11 and 12. The determination is made on the basis of the d / q transformation 3011, the three measured voltage values u _Na , u _Nb , u _Nc being supplied as input variables. The more detailed functionality of the d / q transformation and the d / q reverse transformation are explained in more detail, for example, in https://de.wikipedia.org/wiki/D/qTransformation. The phase position θ _Ν , which is also required for the d / q transformation, is specified in more detail on the basis of the control system shown in FIGS. 11 and 12. Depending on the desired phase position of the space vector in the d / q coordinate system, a downstream PI controller 3012 can regulate either the d components (FIG. 11) or the q component (FIG. 12) to zero. The component that is to be regulated to zero in the relevant coordinate system is fed to the PI controller 3012. An angular velocity ω _{Ν is} added to the output value of the PI controller 3012 in an adder 3013. One possible implementation provides that the angular velocity to be added is defined as a constant ω _Ί , which corresponds to the expected value of the angular frequency or is as close as possible to this. The result of the addition is fed to an integrator 3014. This integrator 3014 has the property that it is reset to zero after reaching a count of 2π, so that the value of the output of the integrator 3014 corresponds to the applied phase position 0 _N and this is always between 0 and 2π. Subsequently, the variant proposed in FIG. 12 continues to be used, in which the q component is regulated to zero.

Accordingly, different sizes are obtained from the phase-locked loop 301. At the output of the controller 3012 shows the fundamental vibration peak value | ύ _Ν | as well as various interference frequencies of the mains voltage. The input of the integrator 3014 gives the current mains frequency ω _Ν , the phase position 0 _N is obtained at the output of the integrator 3014.

The fundamental vibration peak value | ύ _Ν | is fed to a low pass filter 302; in the present case, it is, for example, a linear second-order low-pass filter which has a cut-off frequency at 50 Hz. Depending on the application or nominal network frequency, the low-pass filter can also have a higher or lower cut-off frequency.

Subsequently, the filtered fundamental wave peak value | ύ _Ν | available at the output of the low-pass filter 302 transmitted as a d value to a d / q reverse transformation unit 303. The value 0 is supplied to the d / q reverse transformation unit 303 as the q value. In addition, the phase position 0 _N is transmitted to the d / q reverse transformation unit 303 as a phase value.

At the output of the d / q back-transformation unit 303, the low-pass filtered, corrected fundamental vibration amplitude by means of d / q-alpha / beta back-transformation again determines three individual variables which only contain the basic vibration component and no higher-harmonic, higher-frequency interference voltages.

The inverse transformation basically requires two sizes, namely a d component and a q component. In the present case, the low-pass filtered, corrected fundamental oscillation amplitude is used for the d component. The value 0 is used as the q component. At the output of the d / qRücktransformationseinheit 303 to obtain the three outputs nd as a fundamental voltages u _a, f _U, Ub.fund, u _c, f _U nd voltage waveforms are referred to.

The filtering function could alternatively also take place via three bandpass filters applied to the individual phase voltage values, in which case filtering by means of PLL 301 and the subsequent lowpass filtering in d / q, 11

12/38

System could be eliminated. As a rule, however, one for multi-phase systems

PLL 301 is available for simpler control, it can already be used here advantageously in connection with the correction of the signal for the fundamental oscillation.

The three determined fundamental voltages u _a, nd _fU, u _b, f nd _U, u _c, f _U are nd in a subtractor 304 from the three measured voltages U _Na, U _Nb, U _Nc subtracted. The total attenuation of the fundamental oscillation achieved by subtracting the fundamental voltages from the measured voltages corresponds to the overall transfer function from PLL 301 and low-pass filter 302. At the output of subtraction unit 304 there are cleaned interference voltages u _sa , u _sb , u _sc , the compensation of which is based on a preferred control strategy in a compensation unit 305 is shown in more detail below.

In addition, further frequency components can be filtered out of the signal arriving at the compensation unit 305, if it is to be ensured that these frequency components in the network should not be filtered out as useful signals or should not be regarded as interference. In addition to frequency components in the area of the nominal network voltage, this also applies to other frequency components, in particular in the range of 100 Hz, which are reserved for ripple control signals, for example.

The setpoint current is preferably determined by the compensation unit, i.e. it is by no means necessary for the compensation unit to be realized by an analog circuit which is actually constructed. Rather, the compensation unit is usually implemented by a DSP or an FPGA.

These three interference voltages u _sa , u _sb , u _sc are fed to the compensation unit 305, which has a basically freely selectable network attenuation filter characteristic. The purpose of the compensation unit 305 is to specify a target current value i * _La , i * _Lb , i * _Lc for the current measurement values i determined, which ultimately leads to the interference voltages u _sa , u _sb , u _{sc being} reduced as _far as possible. The network attenuation filter characteristic that can be defined in the compensation unit establishes a connection between the

13/38

Interference voltages u _sa , u _S b, u _sc and the target current and can, if a linear

Filter characteristics is desired, can be formulated as frequency-dependent admittance. An RLC filter is selected as an example, the admittance of which is defined as follows:

Y (s) = sC / (s ² LC + sRC + 1).

For example, 400 Hz with a maximum attenuation of A _r = 1 is selected as the filter resonance frequency f _r . This gives the following values for L, R and C.

C = V2A _r / (2irf _r ) = 562pF

L = 1 / _((2πί Γ) ² · C) = 281 pH

R = V2 / (2TTf _r C) = 1Ω

Would one implement such a filter as a purely passive RLC element, i.e. connecting this filter directly to the mains would result in a fundamental current load of

Iq = 2 * K * f _N * C * V _N = 40A _rms , which would, however, not contribute to the damping of the more frequently used parts and would result in additional losses of R * lq, rms ^A 2 = 1600W.

The proposed control strategy enables the same filter characteristics and dynamics to be achieved as with the passive structure without, however, additionally burdening the network with reactive power. The resulting losses correspond to the resulting losses of the switch cell and not those of the RLC element since the switch cell only virtually adjusts the RLC characteristics.

Each filter block contained in the compensation unit 305 is designed separately from the other filter blocks and has a resulting target current specification, ie a separate target current i * _La , i * i_b, i * Lc is determined for each phase. This

14/38

Target currents i * _La, Lb i *, i * Lc be converted by a d / q transformation unit 3061 using the current phase angle _θ Ν in d / q-current command values i * _Ld * _Lq i. Likewise, the current measured values i _La , ii_ _b , ii_c are converted by a d / q transformation unit 3062 using the current phase position 0 _N into d / q current values i _L d, ii_q. Finally, the voltage _{measured values} u _Na , u _Nb , u _{Nc are converted} into d / q voltage _values u _Ld , u _Lq using the current phase position θ _Ν .

In addition to reducing high-frequency interference voltages in the network, the present regulation also has the task of keeping the intermediate circuit voltage u _D c as constant as possible. The measured intermediate circuit voltage u _D c is fed to an optional notch filter 307. This notch filter 307 is defined as follows for the simulated exemplary case:

Gnotch (s) = (s ² + (2π100) ² ) / (s ² + s-2-ττ100/20 + (2π100) ² )

The desired value of the desired intermediate circuit voltage u _D c and the value obtained at the output of the notch filter 307 are subtracted from one another in the difference forming unit 308. The difference value obtained in this way is then fed to a link 309, which in this example is dimensioned as follows:

Hpi (s) = KP + KI / s, with KP = 0.1 and Kl = 10 in particular

The output of PI element 309 is provided with an optional threshold value unit 310 which, when a predetermined threshold value is exceeded, provides an output value at its input which corresponds to the threshold value. Likewise, a further lower threshold value can be provided for this threshold value unit 310, with the relevant lower threshold value being kept available at the output in the event that a value falling below the lower threshold value is present at the output.

The value present at the output of the threshold value unit 310, the d component of the target current value i * _Ld and the negative value of the d component of the current value i _Ld are fed to an addition unit 311. That at the output of the addition unit

The result pending 311 is fed to a PI controller 312, which in this example is dimensioned as follows:

Hpi (s) = KP + KI / s, with KP = 50 and KI = 500000 in particular

A threshold value unit 312 can in turn optionally be provided at the output of the PI controller 312.

The q components i * _Lq , the target current value and i _{Lq of} the current value are subtracted from one another at the addition / subtraction unit 315, the difference value determined in this way is fed to a PI controller 316, which corresponds in structure to the PI controller 312. An optional threshold value unit 317 can be connected downstream of the PI controller 316.

Subsequently, two addition _units 319, 320 determine two values u _od , u _oq as follows:

The first addition unit 319 is supplied with the negative output of the PI controller 312 or the optional threshold value unit 313, the d value u _{Nd of} the measured voltage u and the positive q value i _{Lq of} the measured current, weighted by the factor K. The multiplication of the q value i _{Lq of} the measured current by the factor K is carried out in the multiplication unit 318. A voltage value u _{od is} obtained at the output of the first addition unit 319.

The second output unit 320 is supplied with the negative output of the PI controller 316 or the optional threshold value unit 317, the q value u _{Nq of} the measured voltage and the negative d value i _{Ld of} the measured current, weighted by the factor K. The multiplication of the d value i _{Ld of} the measured current by the factor K is carried out in the multiplication unit 314. A voltage _value u _{oq is} obtained at the output of the second addition unit 320.

For both multiplications carried out in the multiplication units 314, 318, the factor K corresponds to the value 2 mf _N Lf; For the present example, the factor K has the value 0.157.

, 15th

16/38

The results u _od and u _oq are transformed back into three individual voltage _quantities u * _a , u * _b , u * _c by means of d / q back-transformation using a back-transformation unit 321. These voltage quantities u * _a , u * _b , u * _{c are then divided} by the filtered intermediate circuit voltage which is present at the output of the notch filter 307 in order to obtain a normalized result with respect to the intermediate circuit voltage. Then, due to the concrete choice of half bridges in the inverter, the value 0.5 is added to the individual sizes, so that only positive values in the range between 0 and 1 are available for the modulation. These values are then used to control the switches as part of pulse width modulation.

A second preferred embodiment of the invention is shown in more detail in FIG. 10. This embodiment corresponds essentially to the embodiment shown in FIG. 9, only the differences being discussed in more detail here. The reference numbers used for the illustration are each increased by the value 100 or 1000 compared to the reference numbers in FIG. 9 and begin with the number 4 instead of the number 3.

While the phase-locked loop 301, 401 and the low-pass filter 302, 402 of the embodiments in FIGS. 6 and 6a are constructed essentially identically, the signal obtained from the low-pass filter 402 is not subjected to a d / q inverse transformation in the second embodiment , Rather, in the subtraction unit 404, the d / q-transformed voltage signal u _N d is increased by the value of the low-pass filtered fundamental wave peak value | LJn | reduced. Since - as in the first exemplary embodiment - it is specified that the q value of the signal generated by the phase-locked loop 401 should have the value 0, the q value of the voltage signal u _{Nq need} not be compensated. The thus compensated d value of the voltage signal u _Nd and the uncompensated q value of the voltage signal u _Nq are _fed to the compensation unit 405 as adjusted interference voltage _profiles u _S d, u _sq . In the present exemplary embodiment, the d / q-transformed signal is applied to the compensation unit 405, which is applied to the compensation unit 405 in the first exemplary embodiment. If the determined voltage signal u _N corresponds to an exact sine voltage and

17/38 in particular no high-frequency disturbances are contained in the voltage signal, a zero signal is present at the compensation unit 405.

The compensation unit 405 in turn compensates for the voltage signals in such a way that a target current i _L * is determined, to which the current i _L in the area of the inverter must correspond in order to eliminate high-frequency interference in the voltage signal. In contrast to the first exemplary embodiment, in which the compensation unit 405 contains a total of three identical filter circuits, with which the individual phase voltages u _a , u _b , u _{c are} each filtered separately, only the d component and the q need in the present second embodiment -Proportion of filtering to be subjected. Especially in the case of filtering by using a computing method, the computing effort can be considerably reduced with this measure. The specific structure of the filter can be selected as in the first embodiment.

The voltage and current values (actual values), on the other hand, are fed to the d / q transformation units 4062, 4063; as in the first embodiment of the invention, the d / q-transformed voltage and current signals u _N d, u _Nq , iLd, iLq- Da are obtained in the present second embodiment, the target current i _L * or its d / q Shares i _L d *, ii_q * are determined immediately, no further d / q transformation is required, a d / q transformation unit corresponding to the d / q transformation units 3061 is therefore not provided in the second embodiment.

A further preferred possibility of using a four-phase system is described in more detail in FIGS. 13 and 14. FIG. 13 shows a further alternative embodiment of a filter 202 which has a neutral conductor input on the mains side, to which a coil Lw is connected, the dimensions of which correspond to the other coils Li _a , Li _b , Lie. The filter 202 also has a capacitor Ci _a . Ci _b , Ci _c corresponding neutral conductor capacitor Cw, via which the neutral conductor at node P _{o is} connected to star point 202S. The dimensioning of the neutral conductor capacitor Cw also corresponds to the dimensioning of the others

18/38

Capacitors Ci _a , Ci _b , Ci _c . Finally, the neutral conductor is connected to the neutral conductor connection of the inverter 203 via a further coil l_2o. Also the

Dimensioning of the further coil L ₂ o corresponds to the dimensioning of the corresponding coils L _2a , L _2b , L _2c for the other phases.

In the filter 202 shown in FIG. 13, the current i _L o via the neutral conductor and the voltage u _N0 between the neutral conductor or a node P _o and earth lying on the neutral conductor can also be measured. However, the provision of such voltage and current measuring devices A _o , V _o is optional and depends on the regulation of the potential on the neutral conductor.

The inverter 203 shown in FIG. 14 essentially corresponds to the inverter shown in FIG. 7, but has a further half bridge comprising two additional switches S _O i, S02, the center of which is connected to the neutral conductor input of the inverter 203.

These additional switches S01, Stades inverters 203 can be regulated in different ways, for example as a regulation condition, the sum of the measured voltages u _N1 , u _N2 , u _{N3 can be} a predetermined voltage curve, in particular the value zero or a voltage curve with a predetermined one Amplitude and frequency results. Other control conditions can also be specified with regard to the measured voltage and current measured values u _N0 , ii_o.

In all of the illustrated embodiments of the invention, in addition to improving the filter characteristic of the compensation unit 305, 405, a network resonance scan algorithm can be implemented which can be carried out cyclically, for example per minute, per hour, per day or per month, in order to use the specifically measurable Properties of the network 100 to detect current critical resonances. This information can advantageously be used in order to be able to use this information to adapt the virtual damping algorithm.

The network resonance detection is carried out as follows: First, a current setpoint or setpoint curve with a specific frequency and amplitude

Given 19/38, for example with an alternating current with a peak current I _pk = 1A and a frequency of 300Hz. The current is impressed into the network 100 by means of the inverter 203. The mains voltage response is then measured using the voltage measuring unit (V _a , V _b , V _c ). The component is then measured at the frequency that corresponds to the frequency of the impressed current, in this case at 300 Hz. This can be done, for example, using a PLL 301, 401 or a bandpass filtering. If the voltage amplitude at 300Hz exceeds a predetermined critical value, a critical resonance is detected.

If this is not the case, you can switch to the next frequency, for example 350 Hz, and the procedure described above can be repeated for this frequency. If changes in the network behavior are detected during these measurements, for example when newly detected resonance frequencies do not coincide with the resonance frequencies determined from the previous measurements, the control behavior specified in the compensation unit 305, 405 can be adapted by modifying the filter function.

In order to avoid an excessive load on the network, the current intensity used for the current excitation can be gradually increased in the individual measurements until a voltage threshold value, for example of 50 volts, is exceeded at least at one frequency. In addition, the current intensity used for the current excitation can subsequently be reduced in steps until the voltage threshold value is exceeded at only one frequency in order to find the resonance frequency of the network and to set the filter characteristic accordingly as described above. If current and voltage measurement have the necessary resolution, the network impedance can be determined in addition to the improved characterization of the network.

In addition, the system can also be connected via a data connection to other devices which have further information about the network 100. In this way, measurement results from other stabilization devices arranged in the same network 100 can also be used

20/38 to adjust the filter characteristic of the compensation unit 305, 405.

In addition, there is also the possibility that control or damping algorithms for the

Stabilization device can be specified.

If it can be determined on the basis of the voltage measurement that there are no high-frequency components at the respective node of the electricity network 100, the stabilization device can also be used to execute network-supporting standard algorithms, as are usually implemented in inverters.

If grid-stabilizing operation is not required at certain times, in addition to the described damping concept, a grid-supporting P / f or Q / U control strategy can also be overlaid if necessary, in order to enable grid-supporting operation as provided for inverter systems as standard or with classic grid control strategies , such as primary, secondary, or tertiary control.

Claims (6)

1. Device for stabilizing and filtering out interference within a predetermined frequency range in an at least three-phase AC network (100), the device comprising: - phase connections (200a, 200b, 200c) for connection to the cable connections (111, ...) of the AC network (100), - an inverter (203) with an intermediate circuit capacitor (C7), a filter (202) arranged between the inverter (203) and the phase connections (200a, 200b, 200c) for filtering out high-frequency switching components originating from the inverter (203), Measuring devices (V _a , V _b , V _c ) for measuring the voltage _profile (u _Na , u _N b, u _Nc ) at the phase connections, Measuring devices (A _a , A _b , A _c ) for measuring the current profile (i _La , ii_b, Ilc) between the inverter (203) and the phase connections (200a, 200b, 200c), - A control device (204) which stores the individual measured values (u _a , u _b , u _c , u _D c ia, ib, ic) of the measuring devices (V _a , V _b , V _c , V _D c, A _a , A _b , A _c ) are fed, characterized in that - The control device (204) has a compensation unit, which is designed to determine from the determined voltage profiles (u _a , u _b , u _c ; u _d , u _q ) the frequency component caused by the nominal network voltage, possibly also further frequency components, preferably arithmetically, filter and create an interference voltage curve (u _sa , u _S b, u _sc ; u _S d, u _sq ), and - The control device (204) is designed to carry the inverter (203) To trigger switching signals (s-ι, s ₂ , s ₃ ) in such a way that the current profile (i _a , i _b , i _c ) between the inverter (203) and the phase connections (200a, 200b, 200c) determines that, if necessary by calculation, Corresponds to the current profile that occurs when the interference voltage profile (u _sa , u _S b, u _sc ; u _S d, u _sq ) is sent to a compensation unit (305, 405) by a predetermined one Transmission characteristic of fixed impedance is supplied, which is reduced, preferably the lowest, in particular at a predetermined network resonance frequency to be compensated. 2. Device according to claim 1, characterized in that - A measuring device (V _D c) for measuring the intermediate circuit voltage (u _D c) ^{on the} intermediate circuit capacitor (C _D c) of the inverter (203) is provided, and / or - The control device (204), which is optionally connected downstream of the measuring device for measuring the intermediate circuit voltage, is designed to control the inverter (203) with switching signals (s _a , s _b , s _c ) in such a way that the intermediate circuit voltage (u _D c) corresponds to a predetermined constant value. 3. A method for stabilizing and filtering out interference within a predetermined frequency range in an at least three-phase AC network (100), the device comprising: - phase connections (200a, 200b, 200c) for connection to the cable connections (111, ...) of the AC network (100), - an inverter (203) with an intermediate circuit capacitor (C _D c), a filter (202) arranged between the inverter (203) and the phase connections (200a, 200b, 200c) for filtering out high-frequency switching components originating from the inverter (203), the voltage profile (u _a , u _b , u _c ) is measured at the phase connections, if appropriate after filtering has been carried out, - Measuring devices (A _a , A _b , A _c ) for measuring the current profile (i _a , i _b , i _c ) between the inverter (203) and the phase connections (200a, 200b, 200c), - A control device (204) which stores the individual measured values (u _a , u _b , u _c , u _D ci _a , ib, ic) of the measuring devices (V _a , V _b , V _c , V _D c, A _a , A _b , A _c ) are supplied, characterized in that from the determined voltage profiles (u _a , u _b , u _c ; u _d , u _q ) the frequency component caused by the nominal network voltage, and optionally also further frequency components, is preferably filtered out, and such an interference voltage curve (u _sa , u _S b, u _sc ; u _S d, u _sq ) is created, and - The inverter (203) is controlled with switching signals (s _a , s _b , s _c ) in such a way that the current profile (i _a , i _b , i _c ) between the inverter (203) and the phase connections (200a, 200b, 200c ) corresponds to that, optionally arithmetically determined, current profile that occurs when the interference voltage profile (u _sa , u _S b, u _sc ; u _S d, u _sq ) is fed to a compensation unit (305, 405) with an impedance defined by a predetermined transmission characteristic which is reduced, preferably the lowest, in particular at a predetermined network resonance frequency to be compensated for. 4. The method according to claim 3, characterized in that - The intermediate circuit voltage (u _D c) at the intermediate circuit capacitor (Cdc) of the inverter (203) is measured, and / or - The inverter (203) is controlled with switching signals (s _a , s _b , s _c ) such that the intermediate circuit voltage (u _D c) corresponds to a predetermined constant value. 5. The method according to claim 4 or 5, characterized in that in the event that there are currently no disturbances to be compensated, the impedance of the network and / or the voltage response at a given current excitation at a given frequency is measured for different frequencies and for If the determined impedance and / or voltage exceeds a predefined impedance or voltage threshold value, the transmission characteristic of the compensation unit (305, 405) is adapted to this frequency, the impedance of the compensation unit (305, 405) in particular being reduced at this frequency, in particular least, is. 6. The method according to claim 5, characterized in that in the individual measurements, the current strength used for the current excitation is increased step by step until at least at one frequency a voltage threshold value is exceeded, and optionally then the current strength used for the current excitation in steps as long is reduced until the voltage threshold is exceeded at only one frequency.