METHOD FOR STABILIZING THE DC VOLTAGE IN A DC GRID, AND DC-TO-DC CONVERTER FOR CONNECTING A PV GENERATOR TO A DC GRID

Abstract

A method for stabilizing a DC voltage in a DC grid that includes a DC bus connected to a higher-order grid and to which an energy generating system and at least one load are connected. A variable electric grid output is exchanged between the DC bus and the higher-order grid in order to keep the DC voltage in the DC bus at a nominal voltage. The energy generating system includes a PV generator connected to the DC bus via a DC-to-DC converter and which exchanges an electric generator output with the DC bus. In a normal operating mode, the generator output is set to a normal operating output by the DC-to-DC converter on the basis of an MPP output of the PV generator. In a grid support mode, the generator output is set to a grid support output on the basis of the DC voltage in the DC bus in order to counteract a power imbalance between the electric power supplied in total to the DC bus and the power drawn in total from the DC bus.

Claims

1. A method for stabilizing a DC voltage in a DC grid, wherein the DC grid comprises a DC bus that has the DC voltage, wherein an energy generating system and at least one load are connected to the DC bus, wherein the DC bus is connected to a higher-order grid, wherein an electrical grid power is exchanged between the DC bus and the higher-order grid, wherein the electrical grid power is varied in order to maintain a DC voltage in the DC bus at a nominal voltage, wherein the energy generating system comprises a PV generator that is connected to the DC bus via a DC-to-DC converter and exchanges electrical generator power with the DC bus, in a normal operating mode, setting the electrical generator power to a normal operating power by the DC-to-DC converter as a function of an MPP power of the PV generator, and setting the normal operating power variably in a predetermined relation to the MPP power of the PV generator or fixedly at a value below the MPP power of the PV generator, monitoring the DC voltage in the DC bus using the energy generating system, in a grid support mode, setting the electrical generator power as a function of the DC voltage in the DC bus to a grid support power in order to counteract a power imbalance between a total electrical power supplied to the DC bus and a total power withdrawn from the DC bus, and setting the grid support power as a function of a deviation of the DC voltage from its nominal value and/or a rate of change of the DC voltage.

2. The method according to claim 1, wherein setting the electrical generator power in the grid support mode occurs when the DC voltage in the DC bus exceeds a predetermined limit value.

3. The method according to claim 1, wherein the electrical generator power is set by clocking power semiconductors of the DC-to-DC converter, wherein a PV voltage applied to the PV generator is set to a normal operating voltage in the normal operating mode by means of a first clock rate, and wherein the electrical generator power in the grid support mode is reduced by the DC-to-DC converter being operated by a second clock rate such that the PV voltage changes in a direction of the DC voltage.

4. The method according to claim 3, wherein the grid support mode is entered and the DC-to-DC converter is operated at the second clock rate when the DC voltage exceeds a limit value or a value deviating upwards from the limit value by a hysteresis, and wherein the normal operating mode is entered and the DC-to-DC converter is operated at the first clock rate when the DC voltage falls below the limit value or a value deviating downwards from the limit value by a hysteresis.

5. The method according to claim 3, wherein the second clock rate has the value zero or one such that the DC-to-DC converter is not clocked in the grid support mode.

6. The method according to claim 3, wherein the DC-to-DC converter is operated in the grid support mode such that the PV voltage is increased further with respect to the DC voltage when the PV voltage is matched to the DC voltage and/or the DC voltage continues to exceed a limit value.

7. The method according to claim 1, wherein the grid support power is less than the normal operating power by a PV control power, and wherein the grid support power has an inverted sign with respect to the normal operating power when the deviation of the DC voltage in the DC bus from its nominal value and/or the rate of change of the DC voltage in the DC bus requires a PV control power greater than the MPP power.

8. The method according to claim 1, wherein the rate of change of the DC voltage is determined from a derivative of a voltage measurement or from a current measurement.

9. The method according to claim 1, wherein an energy store is connected to the DC bus, wherein the energy store is configured to exchange a storage power with the DC bus, wherein, in the normal operating mode, the storage power is equal to zero or comprises a charging of the energy store, wherein, in the grid support mode, the storage power comprises discharging of the energy store in order to counteract a reduction in the DC voltage in the DC bus with respect to its nominal value.

10. The method according to claim 9, wherein, in grid support operation, increasing the storage power in order to counteract a power deficit in the DC bus, and reducing a PV power in order to counteract an excess of power in the DC bus.

11. The method according to claim 9, wherein a state of charge of the energy store in the normal operating mode is between 90% and 100% of a charging capacity of the energy store.

12. The method according to claim 1, wherein the DC bus is connected via a bidirectional power converter to the higher-order grid, wherein the bidirectional power converter comprises a DC-to-DC converter when the higher-order grid comprises a further DC grid, or comprises an inverter when the higher-order gird comprises an AC grid, and setting the electrical grid power exchanged via the bidirectional power converter between the DC bus and the higher-order grid as a function of electrical properties of the higher-order grid in order to stabilize the higher-order grid.

13. The method according to claim 12, wherein the DC bus is connected via the DC-to-DC converter to a supply grid and via the inverter to an energy supply grid and is configured to exchange an electrical DC grid power with the further DC grid and an AC grid power with the AC grid.

14. The method according to claim 12, wherein, in the normal operating mode, a grid power flows from the DC bus into the AC grid and/or the further DC grid, wherein the grid power in the normal operating mode comprises an excess power that corresponds to a difference between a generator power generated by the energy generating system and a power consumed by a consumer and/or a machine.

15. A DC-to-DC converter configured to connect a PV generator to a DC bus of a DC grid, wherein the DC-to-DC converter is configured to exchange electrical power between the PV generator and the DC bus, wherein the DC-to-DC converter further comprises a controller, wherein the controller is configured to: in a normal operating mode, set the electrical power to a normal operating power by the DC-to-DC converter as a function of an MPP power of the PV generator, and set the normal operating power variably in a predetermined relation to the MPP power of the PV generator or fixedly at a value below the MPP power of the PV generator, monitor a DC voltage in the DC bus using an energy generating system, in a grid support mode, set the electrical power as a function of the DC voltage in the DC bus to a grid support power in order to counteract a power imbalance between a total electrical power supplied to the DC bus and a total power withdrawn from the DC bus, and set the grid support power as a function of a deviation of the DC voltage from its nominal value and/or a rate of change of the DC voltage.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0067] The disclosure is further explained and described below with reference to exemplary embodiments illustrated in the figures.

[0068] FIG. 1 shows a DC grid in a first embodiment,

[0069] FIG. 2 shows a DC grid in a second embodiment,

[0070] FIG. 3 shows a characteristic curve from one embodiment of a method for stabilizing a DC voltage in a DC grid,

[0071] FIG. 4 shows a characteristic curve from a further embodiment of a method for stabilizing a DC voltage in a DC grid,

[0072] FIG. 5 shows a characteristic curve from a further embodiment of a method for stabilizing a DC voltage in a DC grid,

[0073] FIG. 6 shows a control from one embodiment of a method for stabilizing a DC voltage in a DC grid,

[0074] FIG. 7 shows a DC-to-DC converter for electrical connection between a PV generator and a DC grid,

[0075] FIG. 8a shows time profiles of a DC voltage in a DC grid and of powers of a disruption and of a chopper resistor in the DC grid, and

[0076] FIG. 8b shows time profiles of a DC voltage in a DC grid and of powers of a disruption and of a PV generator in the DC grid.

DETAILED DESCRIPTION

[0077] The disclosure relates to stabilization of a direct current grid (DC grid) for supplying electrical consumers (loads). Direct current consumers (DC loads) can be connected to such a DC grid via direct current converters (DC-to-DC converters), and/or alternating current consumers (AC loads) can be connected to such a DC grid via inverters (DC-to-AC converters). In addition, converters for connecting the DC grid to an alternating current grid (AC grid), electrical energy stores, and energy generating systems having energy sources and further DC-to-DC converters can be connected to the DC grid. The energy sources can comprise photovoltaic generators (PV generators) or wind turbines.

[0078] FIG. 1 shows a DC grid 1 with a DC bus 10 for energy transmission between different electrical units 11-18. Electrical power flows between these units 11-18 within the DC grid 1 via the DC bus 10. Such a DC grid 1 can, for example, be located in an industrial environment, e.g., a factory or a production plant, and have a considerable spatial extent—for example, comprise dozens or hundreds of square meters.

[0079] The DC bus 10 is connected to a higher-order AC grid 11 and/or to a further DC grid 12. The AC grid 11 can, for example, comprise a public energy supply grid 11a, while the further DC grid 12 comprises, for example, a higher-order or adjacent supply grid 12a. Electrical power can be exchanged between the DC bus 10 and the AC grid 11 or the DC grid 12 by arranging a bidirectional inverter (DC-to-AC converter) 11b between the DC bus 10 and the energy supply grid 11a or a bidirectional DC-to-DC converter 12b between the DC bus 10 and the supply grid 12a. Via the DC-to-AC converter 11b, an electrical AC grid power is taken from the energy supply grid 11a and fed into the DC bus 10 or vice versa. Via the DC-to-DC converter 12b, an electrical DC grid power is taken from the supply grid 12a and fed into the DC bus 10 or vice versa.

[0080] A load 13, which comprises a consumer 13a and a power supply 13b, is connected to the DC bus 10. The power supply 13b is configured to extract an electrical consumer power from the DC bus 10 and to make it available to the consumer 13a in a suitable form with respect to current and voltage. Depending upon the type of consumer 13a (DC or AC), the power supply 13b can in particular comprise a DC-to-DC converter and/or an inverter—optionally, an inverter having variable output frequency. The actual consumer power flowing from the DC bus 10 into the load 13 can be set both by the consumer 13a and the power supply 13b—for example, also as a function of a higher-order control unit—for example, a process controller (not shown).

[0081] In one embodiment, when the consumer 13a is part of a larger system, e.g., a motor or a cooling system in an industrial or commercial plant, the consumer power can be predefined by a higher-order control unit in such a way that safe and continuous operation of the system is sought. In this respect, the consumer power is generally assumed, in the context of the operation of the DC grid 10, as being given and can be modified only in exceptional cases for other reasons, in order to contribute, for example, to the stabilization of the DC grid. However, the load 13 can, due to its operation, have considerable feedback effects on the DC bus 10—particularly if the consumer 13a performs switching operations and/or has a rotating flywheel mass, and thus an electromechanical inertia that generates considerable currents by accelerating and decelerating the flywheel mass.

[0082] Alternatively or additionally, the load 13 can have a consumer 13a, the consumer power of which can be modified within certain limits without affecting the operation of the consumer 13a in principle. For example, the load 13 can comprise an electrolyzer that has a nominal power and can be operated in an operating range around, as well as below, the nominal power. Such an electrolyzer can be connected to the DC bus 10 via a power supply 13b, wherein the power supply 13b can change the consumer power of the electrolyzer due to an external control signal or also autonomously as a function of electrical measurement values.

[0083] In addition, in one embodiment a machine 14 can be connected to the DC bus, which machine has a generator 14a and a bidirectional DC-to-AC converter 14b. The machine 14 can exchange an electrical machine power with the DC bus 10, wherein the generator 14a is driven externally, depending upon the operating mode, e.g., by an internal combustion engine, such that the DC-to-AC converter 14b feeds electrical power into the DC bus, and/or itself develops driving force in that the DC-to-AC converter 14b draws electrical power from the DC bus 10 and feeds it into the generator 14a. The machine power actually exchanged with the machine 14 can in this embodiment be predefined via a control signal and/or can be set autonomously by the machine 14.

[0084] The DC grid 10 can have a reserve load 15 that is intended, in one embodiment, to contribute to the stabilization of the DC grid. The reserve load can have a load resistor 15a—for example, what is known as a chopper resistor—into which an electrical reserve power is fed by means of a DC-to-DC converter 15b. The reserve load 15 can, for example, absorb excess power from the DC bus 10, which excess power is then converted only to heat in the load resistor 15a and is therefore not available or is available only to a very limited extent for further use. The actual reserve power can thereby be predefined via a control signal and/or can be set autonomously by the reserve load 15—for example, as a function of electrical parameters of the DC bus 10—for example, the DC voltage U_DC in the DC bus 10.

[0085] Furthermore, an electrical energy store 16 can be connected to the DC bus 10. The energy store 16 can, for example, have a battery 16a that is connected to the DC bus 10 via a DC-to-DC converter 16b. An electrical storage power can be fed into the battery 16a or taken from the battery 16a via the DC-to-DC converter 16b. The actual storage power can thereby be predefined via a control signal and/or be set autonomously by the energy store 16.

[0086] In one embodiment, the DC grid can further comprise energy generating systems—for example, a wind energy installation (WEI) 17 and/or a photovoltaic system (PV system) 18. The WEI 17 can comprise a wind turbine 17a that generates an electric wind energy power and feeds it into the DC bus 10 via an AC-to-DC converter 17b. The PV system 18 can comprise a PV generator 18a that can generate an electrical PV power that is fed into the DC bus 10 via a DC-to-DC converter 18b. The actual wind energy power and the actual PV power are, in one embodiment, based upon the maximum possible power of the wind turbine 17a or of the PV generator 18a in order to exploit the energy originating from these regenerative sources as completely as possible. In a normal operation, the PV generator 18a can feed the maximum possible PV power into the DC grid via the controllable DC-to-DC converter 18b largely independently of the DC voltage in the DC bus 10. Depending upon the voltage ratio between the DC voltage U_DC in the DC bus 10 and the PV voltage, a correspondingly suitable DC-to-DC converter 18b can be used—for example, a step-up converter, a step-down converter, or a two-quadrant converter. In this case, the DC-to-DC converter 18b generally comprises an arrangement of power semiconductors, i.e., in particular, diodes and switches, as well as capacitors and/or inductors, wherein the power semiconductors can be operated in a clocked manner in order to generate a desired transmission ratio between the DC voltage in the DC bus 10 and the PV voltage U_PV at the PV generator 18a.

[0087] In one embodiment, a DC grid 1 according to FIG. 1 can be operated in such a way that the renewable energy from the WEI 17 and PV system 18 is used to cover the power requirement of the loads 13 and, optionally, the machines 14. Electrical power required beyond this can, optionally, be temporarily drawn from the energy store 16 and/or, in the long term, from the AC grid 11 and/or the DC grid 12. Conversely, in the event of an excess of power in the DC grid 1, electrical power can, optionally, be temporarily introduced into the reserve load 15 and/or into the energy store 16, or fed into the AC grid 11 and/or into the DC grid 12 in the long term. The economically optimal operation of such a DC grid 1 is generally obtained if the energy requirement within the DC grid 1 is completely covered by renewable energy sources, i.e., by the WEI 17 and/or the PV system 18, wherein the energy store 16 can optionally compensate for interim shortfalls. As a result, as far as possible, no current is drawn from the AC grid 11 or the DC grid 12, and instead as much as possible of the excess current generated in the DC grid 1 is fed into the AC grid 11 and/or the DC grid 12.

[0088] FIG. 2 shows a DC grid 1 that is largely identical to the DC grid 1 according to FIG. 1. In contrast to FIG. 1, the reserve load 15 in the DC grid 1 according to FIG. 2 is merely optional, i.e., it can be omitted. Any excess power in the DC bus 10 that was generated, for example, by the energy generating systems 17, 18, but does not flow off to the load 13 or the machine 14, can be fed into the AC grid 11 and/or the further DC grid 12, and can optionally be used to charge the energy store 16.

[0089] In one embodiment, the energy store 16 can be almost fully charged in a normal operating mode, i.e., have a state of charge between 90% and 100% of the charging capacity. In the case of a power deficit in the DC bus 10, storage power can be fed from the energy store 16 into the DC bus 10 in order to counteract an undervoltage in the DC bus 10. In the event of an excess of power, the PV power can be reduced—optionally, also to a negative value—by virtue of electrical power being fed back into the PV generator 18a. By means of this operational distribution between the energy store 16 and the PV system 18, the energy store 16 can, for example, be optimally designed and used, and the PV system 18 develops a surplus value that significantly exceeds the mere generation of electrical power.

[0090] Specific embodiments of the method according to the disclosure for operating the DC grid 1 are explained with reference to the following FIGS. 3-8.

[0091] FIG. 3 shows an embodiment of the method according to the disclosure in which the PV power P.sub.PV is changed along a P.sub.PV(U.sub.DC) characteristic curve as a function of the DC voltage U.sub.DC in the DC bus 10. The specific dependence of the PV power P.sub.PV upon the DC voltage U.sub.DC in the DC bus 10 can be predefined by a characteristic curve that, according to FIG. 3, runs in a characteristic curve range 42 that is defined by the upper characteristic curve limit 40 and the lower characteristic curve limit 41.

[0092] In a first region I, where U.sub.DC<U.sub.DC1, the DC voltage U.sub.DC in the DC bus 10 is in the region of the nominal voltage U.sub.Nom. In this region I, the PV system 18 operates in the normal operating mode and feeds the normal operating power P.sub.0 into the DC bus 10. The normal operating power P.sub.0 can correspond to the maximum possible MPP power of the PV generator 18a or be based upon the MPP power, i.e., for example, comprise 90% of the MPP power, and in this respect can be varied at a fluctuating MPP power; alternatively, the normal operating power P.sub.0 can have a fixed value, which value is less than the MPP power. In the region I, there is thus no or only a small degree of regulation of the PV power compared to the MPP power.

[0093] In a second region II, where U.sub.DC1<U.sub.DC<U.sub.DC2, the DC voltage U.sub.DC in the DC bus 10 is increased compared to the nominal voltage U.sub.Nom. In this region II, the PV system 18 switches into grid support operation, and curtailment of the PV power P.sub.PV is carried out according to a P.sub.PV(U.sub.DC) characteristic curve. For this purpose, the voltage U.sub.PV at the PV generator 18a can be changed in the direction of the no-load voltage of the PV generator 18a. If the DC voltage U.sub.DC in the DC bus 10 is higher than the no-load voltage of the PV generator 18a, the voltage U.sub.PV at the PV generator 18a approaches the DC voltage U.sub.DC in the DC bus 10. The PV power P.sub.PV can return to zero with increasing DC voltage U.sub.DC and become negative in the further course, i.e., electrical power can be fed back into the PV generator 18a. In this case, the characteristic curve can run along the upper characteristic curve limit 40. Alternatively or additionally, a stronger, i.e., disproportionate, curtailment can take place such that the P.sub.VP(U.sub.DC) characteristic curve has a nonlinear profile in the characteristic curve range 42. This is useful in particular if further measurement values are added and indicate instabilities—for example, if a high rate of change

[00004]$\frac{{\mathrm{dU}}_{DC}}{\mathrm{dt}}$

of the DC voltage U.sub.DC or a high DC current IDC occurs in the DC bus 10 (cf. FIG. 4 or FIG. 5).

[0094] In a region IIa, where U.sub.DC2<U.sub.DC<U.sub.DC3, increased curtailment of the PV generator 18a can take place—in particular, by the voltage P.sub.PV at the PV generator 18a being further changed in the direction of the no-load voltage of the PV generator 18a and, optionally, further, such that electrical power is fed back into the PV generator 18a. When the maximum possible feedback power −P.sub.Gen to the PV generator 18a is reached, the feedback power can be kept constant in the case of a further rise in the DC voltage U.sub.DC in the DC bus 10 by means of control, as long as the DC voltage U.sub.DC remains below the voltage threshold U.sub.DC3. In this case too, control is alternatively or additionally possible along a non-linear P.sub.PV(U.sub.DC) characteristic curve, such that the change in the PV power is amplified with increasing DC voltage U.sub.DC.

[0095] In the region III, where U.sub.DC≥U.sub.DC3, a protective shutdown of the PV system 18 can finally occur, as a result of which the DC-to-DC converter 18a is disconnected from the DC bus 10; in such an extreme situation, the entire DC grid 1 can, optionally, be switched off for safety reasons.

[0096] FIG. 4 shows a further embodiment of the method according to the disclosure, in which a DC instantaneous reserve power is provided by the PV system 18; this corresponds, analogously, to a virtual inertia or a virtual increase in the capacitance of the DC bus 10 in that the PV system 18 sets the PV power, exchanged with the DC bus 10, according to FIG. 4.

[0097] The DC instantaneous reserve power is provided by the PV system 18 in that the PV power P.sub.PV is set as a function of the rate of change of the DC voltage

[00005]${\stackrel{.}{U}}_{\mathrm{DC}}=\frac{{\mathrm{dU}}_{\mathrm{DC}}}{\mathrm{dt}}.$

In this case, the PV power P.sub.PV, in a normal operating mode, i.e., when the rate of change

[00006]$\frac{{\mathrm{dU}}_{DC}}{\mathrm{dt}}$

is below a first limit value

[00007]$\frac{{\mathrm{dU}}_{\mathrm{DC},1}}{\mathrm{dt}},$

is set to the normal operating power P.sub.0 on the order of magnitude of the MPP power. If the rate of change

[00008]$\frac{{\mathrm{dU}}_{\mathrm{DC}}}{\mathrm{dt}}$

exceeds the first limit value

[00009]$\frac{{\mathrm{dU}}_{DC,1}}{\mathrm{dt}},$

the PV power is reduced with respect to the normal operating power P.sub.0. The PV system 18 thus reacts to a rapid change in the DC voltage U.sub.DC in the DC bus 10 with a proportional or disproportionate change in the PV power P.sub.PV. In this embodiment, the PV power or PV current change counteracts the deviation and the rate of change of the DC voltage U.sub.DC, and, optionally, comprises feedback into the PV generator 18a, i.e., a negative PV power −P.sub.PV, according to FIG. 4. If the rate of change

[00010]$\frac{{\mathrm{dU}}_{DC}}{\mathrm{dt}}$

exceeds the first limit value

[00011]$\frac{{\mathrm{dU}}_{DC,2}}{\mathrm{dt}},$

the maximum possible feedback power −P.sub.Gen is fed back into the PV generator 18a. By means of the method, the deviation of the DC voltage U.sub.DC from its nominal value U.sub.Nom is limited, reduced, and, ideally, completely avoided.

[0098] The specific dependence of the PV power P.sub.PV upon the rate of change of the DC voltage

[00012]${\stackrel{.}{U}}_{DC}=\frac{{\mathrm{dU}}_{DC}}{\mathrm{dt}}$

can be predefined by a characteristic curve. The characteristic curve runs in a characteristic curve range that is defined, analogously to FIG. 3, by an upper and a lower characteristic curve limitation and, within this characteristic curve range, optionally has a non-linear shape. In particular, a family of characteristic curves can be specified from which, during operation, a characteristic to be currently used is selected on the basis of further parameters—for example, on the basis of the current DC voltage U.sub.DC in the DC bus 10.

[0099] FIG. 5 shows a further embodiment of the method according to the disclosure in which a DC instantaneous reserve power is provided by the PV system 18. In contrast to the embodiment according to FIG. 4, the PV power P.sub.PV is in this case set as a function of an amplitude of a direct current IDC in the DC bus 10. The direct current IDC can be determined from a measurement of a partial current I.sub.DCX in a partial capacitance C.sub.DCX of the DC bus 10 with the total capacitance C.sub.DC.

[0100] The DC instantaneous reserve power according to FIG. 5 is provided by the PV system 18 in that the PV power P.sub.PV is set as a function of the partial current I.sub.DCX. The partial current I.sub.DCX can be measured at a partial capacitance C.sub.DCX—for example, at an output capacitance of the PV-based DC-to-DC converter 18b. The partial current I.sub.DCX is in this case in the same ratio to the direct current IDC in the DC bus 10 as the partial capacitance C.sub.DCX to the total capacitance C.sub.DC of the D bus 10. A corresponding scaling factor for determining the total current IDC from the partial current I.sub.DCX can be determined when the DC grid 1 is initially put into operation, or can be continuously determined during operation. In contrast to the method according to FIG. 4, only a current measurement is necessary here, and, in particular, a derivative of the DC voltage U.sub.DC can be dispensed with.

[0101] Specifically, the PV power P.sub.PV is set to the normal operating power P.sub.0 in a normal operating mode, i.e., when the partial current I.sub.DCX is below a first limit value I.sub.DCX,1. If the partial current I.sub.DCX exceeds the first limit value I.sub.DCX,1, the PV power is reduced with respect to the normal operating power P.sub.0. The PV power P.sub.PV follows a characteristic curve that, analogously to FIGS. 3 and 4, runs within a characteristic curve range that is defined by an upper and a lower characteristic curve limitation and can run non-linearly within the characteristic curve range.

[0102] As a result of the proportional or disproportional change in the PV power P.sub.PV as a function of the amplitude of the DC bus current IDC or of the partial current I.sub.DCX, the PV system 18 counteracts the change and the rate of change of the DC voltage U.sub.DC in the DC bus 10. The change in the PV power P.sub.PV can comprise power feedback into the PV generator 18a. If the partial current I.sub.DCX exceeds a second limit value I.sub.DCX,2, the maximum possible feedback power −P.sub.Gen can be fed back into the PV generator 18a. By means of the method, the deviation of the DC voltage U.sub.DC in the DC bus 10 from its nominal value U.sub.Nom is limited, reduced, and, ideally, completely avoided.

[0103] FIG. 6 shows an embodiment of a specific control 60 and the effect thereof on the DC voltage U.sub.DC in the DC bus 10 in the context of the method according to the disclosure. The embodiments explained with reference to FIGS. 3 through 5 can be combined by means of the control 60. The mechanisms of action of this embodiment of the method according to the disclosure and the resulting reaction of the DC voltage U.sub.DC to sudden power imbalances in the DC bus 10, i.e., in the case of a sudden, strongly unbalanced power balance ΔP=P.sub.IN−P.sub.OUT in this DC bus 10, are explained with reference to FIG. 6 and the following embodiments.

[0104] The total capacitance C.sub.DC of the DC bus 10 is virtually increased by a differential control component of the control 60 by C.sub.VIRT in that the PV system 18 sets the PV power P.sub.PV as a function of the rate of change

[00013]$\frac{{\mathrm{dU}}_{DC}}{\mathrm{dt}}.$

The differential control component limits the gradient of the DC voltage U.sub.DC in the DC bus 10 by providing a DC instantaneous reserve power (cf. FIG. 4 and the associated description, based on the derivation of the DC voltage U.sub.DC, or FIG. 5 as an embodiment with reserve power based on the determination of the direct current I.sub.DC). This behavior corresponds to a virtual increase in the total capacitance C.sub.DC of the DC bus 10 and, in the case of a power imbalance in the DC bus 10, leads to the DC voltage U.sub.DC drifting away more slowly, but not stopping, because the power balance in the DC bus 10 is not compensated for by the higher capacitance alone.

[0105] Compensation for the power balance in the DC bus 10 is sought by means of an additional proportional control component of the control 60. The PV system 18—optionally, supported by the energy store 16—adjusts its power as a function of the DC voltage U.sub.DC itself or of the deviation of the DC voltage U.sub.DC from its nominal value (cf. FIG. 3 and the associated description). The proportional control component has a voltage-maintaining function in that a power imbalance in the DC bus 10 is at least reduced, and drifting away of the DC voltage U.sub.DC is slowed down. The smaller a virtual parallel resistor 1/R.sub.VIRT used as a scaling factor in this proportional control component of the control 60 is designed, the greater the change is in the PV power P.sub.PV (represented here by the current I.sub.Rvirt) due to the proportional control component of the control 60 in the case of a voltage change that is otherwise the same, and the less far the DC voltage U.sub.DC drifts away from its nominal value U.sub.nom, and the faster the drifting away is stopped.

[0106] In addition to the PV system 18, further power converters can be connected to the DC bus 10, which power converters can also be configured to return the DC voltage U.sub.DC in the DC bus 10 to the nominal value. These further power converters—in particular, a bidirectional inverter (DC-to-AC converter) 11b between the DC bus 10 and the energy supply network 11a or a bidirectional DC-to-DC converter 12b between the DC bus 10 and the supply grid 12a (cf. FIGS. 1 and 2)—can, in the event of a continuous deviation of the DC voltage U.sub.DC from its nominal value, likewise suitably change the power exchanged with the DC bus 10 in order to compensate for the power imbalance; this corresponds to an additional integrating control component that can be implemented in the control 60 or as part of higher-order control of the DC grid 1. In this case, this integrating control component contributes to returning the DC voltage U_DC to the nominal value U.sub.Nom and can be implemented with reduced dynamics. The transient overvoltages are already effectively countered by the PV system 18 by means of the proportional and differential control components of the control 60 in that said control components limit the gradient of the DC voltage U.sub.DC or contribute to the voltage maintenance.

[0107] By adapting the power of the (further) power converters participating in the return of the DC voltage U.sub.DC to the nominal value U.sub.nom, the DC voltage U.sub.DC in the DC bus 10 is returned to a voltage band around its nominal value; in return, the change in the PV power P.sub.PV within the scope of the proportional or integral control component of the control 60 can be successively reduced such that the PV power P.sub.PV is returned to the normal operating power. As a result, in the case of any further (transient) power imbalance—in particular, starting from a DC voltage U.sub.DC near the nominal voltage U.sub.Nom—the full DC instantaneous reserve power of the PV system 18 is available. This also means that the effect of the described solutions is then particularly advantageously developed if either the deviation or the rate of change of the DC voltage U.sub.DC or the deviation of the direct current IDC in the DC bus 10 exceeds defined deadband values in each case; the dynamic behavior in the DC bus 10 is not influenced within the respective deadband.

[0108] FIG. 7 shows an embodiment of an energy generating system 18 having a bidirectional DC-to-DC converter 18b for connecting the PV generator 18a to the DC bus 10. The DC-to-DC converter 18b can be arranged between the PV generator 18a and the DC bus 10 and can manage the step-up and step-down voltage transmission types in different energy flow directions. Specifically, the DC-to-DC converter 18b can be configured as a two-quadrant converter. For a transfer of electrical power from the PV generator 18a to the DC bus 10, the DC-to-DC converter 18b comprises a step-up converter 71 having a first switch S.sub.L and a first diode DH, and, conversely, for feedback from the DC bus 10 into the PV generator 18a, a step-down converter 72 having a second switch S.sub.H and a second diode D.sub.L. The DC-to-DC converter 18b thus has, in particular, a half bridge having the power semiconductors S.sub.L, S.sub.H, D.sub.L, and DH and, by means of suitable clocking of the switches S.sub.L and S.sub.H, allows a demand-based power exchange between the PV generator 18a and DC bus 10. If both switches S.sub.L, S.sub.H are open and are not clocked, the PV voltage U.sub.PV approximates the DC voltage U.sub.DC in the DC bus 10. In addition, a feedback of electrical power from the DC bus 10 into the PV generator 18a can be made possible by bridging the first diode DH by means of closing the switch S.sub.H.

[0109] Depending upon the design of the PV generator 18a, the PV voltage U.sub.PV, which is oriented to the MPP voltage of the PV generator, to be set in normal operating mode can be less than or equal to the DC voltage U.sub.DC in the DC bus 10. Particularly if the maximum PV voltage U.sub.PV,max and the maximum DC voltage U.sub.DC,max are approximately the same (e.g., U.sub.PV,max=U.sub.DC,max=1,000 V), the PV generator 18a can be designed such that the system voltage is not exceeded, even in PV feedback operation. In this case, the PV voltage is always less than or equal to the maximum DC voltage U.sub.DC in the DC bus 10 and can be set by means of a step-up converter. If, on the other hand, the PV generator design is such that the PV voltage U.sub.PV can assume higher values than the DC voltage U.sub.DC in the DC bus 10, a multi-quadrant converter can be used for the feedback.

[0110] FIGS. 8a and 8b show, by way of example, time profiles of the DC voltage U.sub.DC and the power P in the DC bus 10 of a DC grid 1 for two different cases. In the respectively upper part of FIGS. 8a and 8b, voltage profiles 82, 84 of the DC voltage U.sub.DC in the DC bus 10 are shown over the time t, wherein the nominal voltage U.sub.0 is represented by the dashed line 80. In the respectively lower part of FIGS. 8a and 8b, power profiles 83, 85 of powers P in the DC bus 10 are shown over the time t, wherein the dotted line 81 represents the power balance in the DC bus 10.

[0111] In FIG. 8a, the DC voltage U.sub.DC at time t<t1 corresponds to the nominal voltage U.sub.0, wherein the sum of the powers fed into the DC bus 10 and the sum of the powers taken from the DC bus 10 are approximately equal; in other words, there is no disruption of the power balance, and the dotted line 81 is at zero.

[0112] At the time t1, a linearly increasing disruption of the power balance begins, which is caused, for example, by a consumer 13a having decreasing consumer power, a braking generator 14a, or a WEI 17 having increasing wind energy power, and the dotted line 81 rises. At the same time, the DC voltage U.sub.DC in the DC bus 10 also rises and, at the time t2, exceeds a threshold value U.sub.max or a value U.sub.max+ΔU deviating from the threshold value U.sub.max by a hysteresis ΔU. A load resistor 15a is then activated and draws an electrical reserve power, represented by the power profile 83, from the DC bus 10, as a result of which the DC voltage U.sub.DC decreases again. If the DC voltage U.sub.DC is again below the threshold value U.sub.max or below a value U.sub.max−ΔU deviating from the threshold value U.sub.max by a hysteresis ΔU, the load resistor is deactivated, and the DC voltage U.sub.DC rises again above the threshold value U.sub.max, and so forth.

[0113] The result is a clocked reserve power having the power profile 83 and the voltage profile 82. In the case of a further rise in the disruption of the power balance, represented by the rise in the dotted line 81 between the times t2 and t3, the duty cycle of the clocked reserve power also increases, i.e., the load resistor 15a is activated increasingly longer for each time unit and, consequently, increasingly heats up. At the time t3, the disruption of the power balance has reached a value that corresponds approximately to the instantaneous reserve power of the load resistor 15a such that the load resistor 15a remains activated for relatively long time periods in order to establish the power balance and to stabilize the DC voltage U.sub.DC. If the disruption of the power balance continues to rise, the reserve power would no longer be sufficient to stabilize the DC voltage U.sub.DC in the DC bus 10 and would further increase the DC voltage U.sub.DC. Between the times t3 and t4, however, the disruption of the power balance decreases to zero again, e.g., by a consumer 13a increasing its consumer power (again), or by further control reserves, connected to the DC bus 10, but reacting more slowly, drawing excess power from the DC bus 10—in particular, due to an integrating control component and/or at the request of a higher-order control of the DC grid. Accordingly, the duty cycle of the power profile 83 is reduced. At the time t4, the power balance is restored in the DC bus 10, the DC voltage U.sub.DC in the DC bus 10 corresponds to the nominal voltage U.sub.0, and the load resistor 15a is deactivated.

[0114] In FIG. 8b, the DC voltage U.sub.DC at a time t<t1 corresponds to the nominal voltage U.sub.0, for example, there is no disruption of the power balance, and the dotted line 81 is at zero. At the same time, a PV system 18 feeds an MPP power P.sub.MPP of a PV generator 18a into the DC bus 10 via a DC-to-DC converter 18b.

[0115] At the time t1, a linearly rising disruption of the power balance begins, i.e., both the dotted line 81 and the DC voltage U.sub.DC rise. At the time t2, the DC voltage U.sub.DC in the DC bus 10 exceeds the threshold value Umar or a value U.sub.max+ΔU deviating from the threshold value U.sub.max by a hysteresis ΔU. The DC-to-DC converter 18b then changes its operating mode such that the PV voltage moves away from the MPP in the direction of the DC voltage U.sub.DC. In particular, the DC-to-DC converter 18b can interrupt a clocking of its power semiconductors for this purpose such that the DC voltage U.sub.DC in the DC bus 10 is applied to the PV generator 18a via possible capacitances and inductances of the DC-to-DC converter 18b. This change in the PV voltage reduces the PV power fed into the DC bus, represented by the power profile 85. This counteracts the excess of power in the DC bus 10 such that the DC voltage U_DC decreases again. If the DC voltage U.sub.DC is again below the threshold value U_max or below a value U.sub.max−ΔU deviating from the threshold value U.sub.max by a hysteresis ΔU, the DC-to-DC converter 18b is operated again in such a way that the PV voltage is again moved in the MPP direction—in particular, by resuming the clocking of the power semiconductors of the DC-to-DC converter 18b—and the PV power rises. The DC voltage U.sub.DC then rises again above the threshold value U.sub.max or above U.sub.max+ΔU, and so forth.

[0116] The result is a modulated PV power having the power profile 85 and the voltage profile 84. In the event of a further rise in the disruption of the power balance, represented by the rise in the dotted line 81 between the times t2 and t3, the average value of the PV power decreases further, while the period of the modulation can largely remain constant. At the time t3, the disruption of the power balance has reached a value that is greater in magnitude than the MPP power of the PV generator 18a, such that the PV power has changed signs, and modulated power feedback into the PV generator 18a takes place. Depending upon the design of the PV voltages in relation to the DC voltage U.sub.DC in the DC bus 10, it may be necessary for this purpose to activate the DC-to-DC converter in a different operating mode in order to change the PV voltage across the DC voltage in the DC bus—in particular, to generate a PV voltage above the DC voltage U.sub.DC in the DC bus 10.

[0117] In principle, the PV generator 18a is able to absorb a feedback power on the order of magnitude of its nominal power. By means of a suitable design and control of the DC-to-DC converter 18b, the PV system 18 can thus provide a reserve power that always comprises at least the nominal power of the PV generator 18a and, in this case, more than twice the current MPP power of the PV generator 18a, this reserve power being available to at least limit a power imbalance due to a correspondingly large disruption and to stabilize the DC voltage U.sub.DC.

[0118] Between the times t3 and t4, the disruption of the power balance decreases to zero again, and the mean value of the PV power increases according to the power profile 85. At the time t4, the power balance is restored in the DC bus 10, the DC voltage U.sub.DC in the DC bus 10 corresponds to the nominal voltage U.sub.0, and the PV system 18 feeds the MPP power of the PV generator 18a into the DC bus 10.

[0119] The comparison of the dynamic aspects when using reserve loads 15 (FIG. 8a) or PV systems 18 (FIG. 8b) for stabilizing the DC voltage U.sub.DC in the DC bus 10 shows that a PV generator 18a that is connected via a suitable DC-to-DC converter 18b to the DC bus 10 can be operated similarly to a load resistor 15a, in order to contribute to the sufficiently rapid stabilization of the DC voltage in the DC bus 10. A load resistor 15a has a linear, time-invariant, unchangeable U/I characteristic curve. In contrast thereto, a PV generator has a non-linear, time-variant U/I characteristic curve whose equivalent internal resistance becomes increasingly smaller with increasing voltage. This non-linearity of the PV characteristic curve can advantageously be used to stabilize the DC voltage in the DC bus 10 and, in particular, to avoid overvoltages in the DC bus 10. By deactivating the DC-to-DC converter 18b of the PV system 18, the PV voltage rises above the MPP voltage, wherein the PV current decreases exponentially with increasing PV voltage. The PV power even gets negative at PV voltages above the PV no-load voltage. Due to the high slope of the PV characteristic curve at voltages above the MPP, small changes in the PV voltage already cause relatively large changes in the PV current.

[0120] In one embodiment, due to the dependence of the PV characteristic curve upon different external parameters—for example, upon irradiation and temperature—the specifically desired operating point is permanently monitored on the PV characteristic curve and adjusted by means of suitable control of the DC-to-DC converter.

[0121] In addition, in cases without irradiation, for example, in the dark, the PV generator 18a can absorb the simple PV nominal power as feedback power, wherein, as a result of lower PV voltages, a somewhat larger current flows compared to the load resistor of the same nominal power. When the PV generator 18a is fully irradiated, i.e., in a normal operation close to the PV nominal power, almost twice the PV nominal power is available as reserve power, in that the PV power is regulated from the MPP power to zero. Furthermore, a power on the order of magnitude of the PV nominal power may be fed back into the PV generator 18a. Thus, a PV system 18 provides at least as much reserve power as a reserve load 15 having a comparable rated power, wherein the available reserve power of the PV system 18, with increasing irradiation, increases to twice the PV nominal power.

[0122] When the DC voltage U.sub.DC is applied to the load resistor 15a or when the voltage is changed at the PV generator 18a, a respective corresponding operating point is generally established without delay, provided that parasitic effects are negligible. However, because the PV generator 18a is operated via a power-electronic DC-to-DC converter 18b that, for example, can comprise smoothing filters having inductive and/or capacitive filter energy stores, a shift of the operating point of the PV generator 18a does not take place at any speed, but with a time delay that depends upon the size of the filter energy store and the maximum permissible electrical quantities. This delay can be minimized by using advanced power semiconductors and/or by increasing the switching frequency and/or by reducing the filter energy store, if necessary. In one embodiment, the rate of change of current achievable at the filter energy store is greater than the rate of change of voltage in the DC bus 10. In addition, in one embodiment, suitable precautions are taken to prevent excessively high compensating currents, overvoltages, and/or excessive electrical powers in order to avoid damaging of components of the DC-to-DC converter 18b when the DC voltage U.sub.DC in the DC bus 10 is applied to the PV generator 18a. For this purpose, the determination and monitoring of the currents flowing and voltages occurring in the DC-to-DC converter 18b, which takes place anyway as part of the usual control of the DC-to-DC converter 18b, can be used. Alternatively or additionally, the power semiconductors of the DC-to-DC converter 18b responsible for the voltage adjustment can be bridged, for example, with an additional connected line.