EP3888219A1

EP3888219A1 - Arrangement having store for electrical energy and regenerative energy generator, in particular a wind power plant, and method for operating same

Info

Publication number: EP3888219A1
Application number: EP19809813.9A
Authority: EP
Inventors: Matthias Seidel; Atanas Dimov
Original assignee: Siemens Gamesa Renewable Energy Service GmbH
Current assignee: Siemens Gamesa Renewable Energy Service GmbH
Priority date: 2018-11-28
Filing date: 2019-11-26
Publication date: 2021-10-06
Also published as: CN113383477A; WO2020109300A1; DE102018009341A1; US20220052530A1

Abstract

The invention relates to an arrangement comprising a regenerative energy generator (1) and a store (2) for electrical energy to be output to a grid (9), the regenerative energy generator (1) comprising an energy converter (13) for converting renewable energy sources into electrical energy and a power control unit (10) which has an input for a target power output and controls the power generation by the regenerative energy converter (13), with a charge controller (20) for the store (2) also being provided, which sets a minimum operating reserve to be provided and outputs a signal for a target state of charge (SoC) of the store (2). In order to achieve improved utilisation of the store, a correction unit (4) is connected to the charge controller (10) and continuously modifies the signal for the target state of charge (SoC). A sequence control unit (5) is also provided, which is actuated by the correction unit (4) and influences the power control unit (10) of the regenerative energy generator according to an output value of the correction unit (4).

Description

Arrangement with storage for electrical energy and regenerative energy producers, in particular wind turbines, and method for their operation

The invention relates to an arrangement for energy generation and storage. It comprises a regenerative energy generator, in particular a wind energy installation and / or a photovoltaic installation, and a storage device for electrical energy.

To ensure high security of supply, electrical networks must be able to quickly compensate for differences between power generation on the one hand and power consumption on the other. The electrical power required for this is called control power. Positive control power is required to compensate for a deficit in power generation that suddenly occurs, such as the failure of a power plant. In the case of negative control power, the opposite is true, there is too much power generation in relation to the power consumption, so that the generation must be reduced. In both cases, the balancing capacity bridges the excess or deficit quickly and for a limited period of time, namely until additional reserve generation can be started up or generation units can be taken off the grid. What is meant by a limited period of time varies depending on the country, typically periods in the range between 15 and 60 minutes.

In conventional electricity networks, electrical power is predominantly generated by means of conventional power plants, which generally have synchronous generators driven by turbines. An oversupply or a lack of electrical power (active power) is noticeable through a change in the network frequency. Such changes in the network frequency are due to the fact that if there is no active power in the network (generation is too low), the synchronous generators slow down due to their design, or if there is too much active power (generation is too large for consumption), the synchronous generators are relieved and thus accelerated . This braking or acceleration is counteracted by the large mass inertia of the rotating parts of the synchronous generators, including those coupled to the driving turbines. Classic synchronous generators therefore have a balancing effect on changes in active power. This effect only occurs in the time range of a few seconds, so that networks with predominantly conventional power plants also rely on control energy. However, conventional power plants with their large synchronous generators are increasingly being replaced by smaller, decentralized power plants in power grids, the power that can be fed in depends on the primary energy supply (for example, wind speed). Examples of this are plants for regenerative energy generation from renewable energy sources, in particular wind energy plants or wind farms or photovoltaic plants. Power plants for regenerative energy generation can only provide balancing energy depending on the primary energy supply, and must be operated below the possible output (regulated) if positive balancing power is to be provided.

On the other hand, an intrinsic, imperative requirement for the provision of balancing power is that it can be guaranteed. On the one hand, it must be guaranteed that it can be fed in for a certain period of time (typically a period of 30 minutes), and it must be guaranteed for a certain period of time (typically a period of one week in Germany is available for provisioning) requests).

While conventional power plants can usually meet this requirement of guaranteed maintenance, this usually does not apply to plants that generate renewable energy. In particular, the regenerative energy generation systems lack a sufficiently reliable forecast for the required period (the period over which the control power is guaranteed to be maintained), and furthermore these systems can often not (or only to a very limited extent) bring positive control power (if the wind is not blowing enough, no additional power can be fed).

In order to nevertheless involve the regenerative energy generation plants in the provision of control power, electrical storage devices are increasingly being assigned to them. However, the provision of the memories is complex and their utilization is sometimes not optimal. Typically, they are operated with a charge level of around 50% in order to have both serve upwards to receive power (negative control power) or power downwards to deliver additional power (positive control power). The storages are thus to be dimensioned twice as large as required for the control energy to be provided.

In order to minimize the effort required for this, it is known to reduce the memory by providing an energy converter (EP 3 148 036 A1). This energy converter is trained to act as an additional consumer, particularly in the case of negative control power, and thus to dissipate power so that it does not have to be recorded in memory. This means that the storage can be operated with a higher charge level (ideally 100%) and can therefore be dimensioned smaller. This advantage is offset by the disadvantage that an additional element for dissipating excess energy is required; furthermore, this dissipation of the electrical energy is a waste of energy which has to be avoided due to the lack of storage possibility.

The object of the invention is to avoid this disadvantage and to achieve better use of the memory. The solution according to the invention lies in the features of the independent claims. Advantageous further developments are the subject of the dependent claims.

In the case of an arrangement comprising a regenerative energy generator and a store for electrical energy for delivery to a network, the regenerative energy producer having an energy converter for converting renewable energy sources into electrical energy and a power control unit which has an input for a desired power delivery and the power generation is controlled by the regenerative energy converter, and a charging control for the memory is also provided, which sets a minimum control energy to be provided and outputs a signal for a desired charge level of the memory, according to the invention it is provided that a charge control Correction unit is connected, which continuously modifies the signal for the target charge level, and a sequence control unit is provided, which is controlled by the correction unit and acts on the power control unit of the regenerative energy generator depending on an output value the correction unit.

First, some terms used are explained. A regenerative generator is a system with an energy converter for generating electrical energy from renewable primary energy. In particular, this includes wind turbines, photovoltaic systems, or thermal solar power plants. It is typical of this that the primary energy cannot be directly controlled (whether and how strong the wind is blowing, the sun is shining, etc.). The memory is at least designed to store or withdraw electrical energy generated by the energy generator and to release it into the network or to draw it from it.

The invention is based on the idea of not leaving the state of charge of the memory fixed while maintaining control energy as in the prior art (for example to 50% or 100%), but to modify it dynamically by means of a correction unit, depending on the correction is acted on the regenerative energy generator and its target output is changed accordingly. This results in a dynamic distribution of the negative control energy to be provided between storage on the one hand and the regenerative energy generator on the other. If there is a lot of wind, for example, the state of charge of the storage can be raised by the correction unit, since with a lot of wind and thus high performance of the wind energy installation, the negative control power that may be required can be achieved by throttling the wind energy installation without having to resort to the storage . The memory therefore does not need to have a corresponding amount of free space, i.e. it can be charged significantly higher than 50%. The invention has recognized that this can be achieved by the combination of two measures, namely on the one hand the dynamic design of the target state of charge at the memory, combined with a change in the target output of the regenerative energy generator from the memory by means of the sequence control. In this way, the sequential control system turns the regenerative energy generator into an “assistant” to the storage system. This is exactly the opposite approach to what has been customary in the prior art.

The advantages achieved are considerable. On the one hand, not only is the provision of control power improved, but in addition the capacity of the memory can now also be used for the delayed provision of the generated electrical energy, that is to say for the so-called generation delay. This shift in time, also referred to as “production shifting”, is of considerable importance in network operation, since from the network's point of view a shift in the power supply is possible. This means that peak demand can be met, making it particularly suitable for situations when energy is scarce and therefore particularly important in the network. This enables the regenerative energy producers equipped in accordance with the invention to find new fields of use, in particular participation in the free trade of electricity on the so-called power exchanges. According to the invention, the energy store is therefore used twice, namely on the one hand for the provision of control energy (as before), but also on the other hand for “production shifting” to satisfy peak demand (electricity exchange). The memory is thus used much better, and without the need for expensive expansion of the storage capacity as such. The invention achieves this solely through a clever use and control of the memory itself and the regenerative energy generator associated therewith. It is always ensured that the required balancing energy is available immediately on request. A measure for the available primary power is preferably applied to the correction unit as an input parameter, in particular it can be a prediction (forecast) for the available primary power. A particular advantage is that this prediction can be shorter than the period over which the control power is guaranteed. The forecast period can therefore be shorter than the guarantee period (also referred to above as the "required period"). This is a particularly advantageous aspect of the invention.

The sequence control expediently acts on the power control unit of the regenerative energy generator in such a way that it is in a master / slave relationship with the correction unit of the memory. Here, the correction unit is the master and accordingly the power control unit of the regenerative energy generator is the slave. It can be achieved that, for example, if there is sufficient wind, the storage can be fully charged, and the negative control power is then not provided via the storage by means of the sequential control, but rather by regulating the wind energy plant. If, for example, both sufficient wind and a period with a tendency to oversupply of energy in the network are predicted for a subsequent period (for example 48 hours), then the store can be fully charged according to the invention (and relatively less power is delivered to the network) ) to only discharge the storage at a later point in time when there are peaks in demand in the network (for a corresponding fee), and yet always have sufficient control power available.

The correction unit is advantageously designed for a plurality of input parameters. In particular, further parameters are predictive values for the available primary power (wind strength or sunshine intensity), predictive values for electricity requirements in the network, minimum values for control power to be kept available, storage capacity and power, target control power, and / or charge level of the storage. With such a plurality of input parameters for the correction unit, a better prediction or adaptation of the state of charge of the Storage (and depending on the sequence control also a corresponding power setting of the regenerative energy generator). Expediently, at least for some of the input parameters, statistical parameters are also additionally created, in particular those for a confidence interval. This enables a finer adjustment and higher reliability to be achieved, particularly in the area of predicting power production (wind forecast or sunshine forecast). The same applies to the prediction of the demand situation in the network.

The correction unit expediently has an optimization device which is designed to determine the signal for the desired state of charge using an optimization module on the basis of the input parameters. By means of optimization calculation known per se, the optimal target state of charge of the memory can be determined in relation to the input parameters and, via the sequence control, the power to be emitted by the regenerative energy generator. A gradient method, a neural network or an evolutionary algorithm are expediently implemented in the optimization module as the optimization method. Such optimization methods are known per se and therefore do not need to be explained in more detail here.

The input parameters are preferably time-dependent, that is to say they vary over time.

This applies in particular to the forecast for the available primary power, but also to the demand situation in the network. Furthermore, the correction unit is preferably designed such that it evaluates the input parameters in a staggered manner. This means that a dynamic process can be achieved when determining the target charge level and thus a more precise adaptation to the respective changed conditions.

According to a particularly preferred embodiment of the invention, weighting factors are also provided for at least one of the input parameters, in particular for the electricity requirement in the network and / or the power generated. The weighting factors can optionally be different depending on the sign (for positive or negative). With the weighting factors, a measure of the importance of the input parameters can be set, whereby the measure can also vary over time. It can be used to express the importance of a particular parameter at a particular point in time. This is particularly favorable with regard to the yield management of renewable energy generation, whereby the weighting factor can stand for a price to be achieved (for example at a power exchange). Weighting is a valuable tool for optimization. With regard to the target power output of the regenerative energy generator, it is advantageously provided that a reference signal for a target power output is applied externally. However, it can also be provided that it is generated internally, in particular by means of frequency statics. In the latter case, for example, the regenerative energy generator increases its output in the event of a falling frequency (below the normal value) and lowers its output in the opposite case when the frequency increases (above the normal value). In this way, self-regulation can be achieved which is basically similar to that of a synchronous generator in a conventional power plant.

The operation of the arrangement according to the invention is briefly outlined below using an example of a wind turbine as a regenerative energy generator. In the case when there is sufficient primary energy (because the wind is blowing strong enough), positive and negative control power can basically be provided both by the wind power plant itself and by the storage. In general, the choice of the distribution between the producer and the storage facility should on the one hand guarantee a minimum control capacity (positive and negative) at all times and on the other hand should be as efficient and economical as possible. On the other hand, if there is no or insufficient primary energy available (for example in the event of a lull), it must be ensured that the minimum control power offered can only be provided with the storage tank. This is done in a known manner by setting a defined state of charge, for example if a symmetrical primary control power is required (positive and negative equal), this state of charge is 50%. Since this control power must be available at all times, it is necessary that the charge level is set and achieved in this way. In other words, it is not enough to determine a lull and then regulate the storage, rather the lull has to be predicted in order to approach the storage condition with the decreasing wind.

All of this can be achieved by the invention as described above. The behavior can be improved by selecting different optimization algorithms and, if necessary, by optional additional input parameters, such as one for an uncertainty of the weighting factors for the energy fed in (e.g. the market price forecast) or such a parameter that takes into account the storage efficiency or storage self-discharge (and thus one) rather long penalized).

The invention further relates to a corresponding method. For a more detailed explanation, reference is made to the above description. The invention is described in more detail below with reference to the accompanying drawing using an exemplary embodiment. Show it:

1 shows an overview of a wind power plant with a storage device according to an exemplary embodiment of the invention;

Figure 2 is an overview of a solar power plant with storage ge according to another embodiment of the invention.

3 shows a schematic block diagram for controlling the memory and the wind energy installation;

4a, b, c diagrams of the operating behavior of the wind power plant; and

Fig. 5 is an illustration of the gain.

A wind energy installation, designated in its entirety by reference number 1, forms, together with a storage unit 2, an arrangement according to an exemplary embodiment of the invention.

The wind power plant 1 is constructed conventionally per se. It has a tower 15, at the upper end of which a gondola 11 is arranged so as to be pivotable in the azimuth direction. A wind rotor 12, which drives a generator 13 via a rotor shaft (not shown), which rotates together with a converter 14 to generate electrical power, is rotatably arranged on one end face thereof. The electrical power generated in this way is routed via an internal line (not shown) to a system transformer 16 and via a connecting line 17 to a mostly network-internal collection network. The storage unit 2 is also connected to the connecting line 17. Further wind energy plants 1 of the wind farm can also be connected, which are constructed essentially in the same way and expediently connected to a common storage facility of the wind farm. be; but they can also each have their own memory 2. The electrical energy it generates is passed via a park transformer 18 to a transfer point 19, from where it is fed into a public network 9. The operation of the wind energy plant 1 is controlled by an operating control system 10. All of this is as far as conventional and therefore need not be explained further. It should be noted that a plurality of wind energy plants operated according to the invention do not necessarily have to be combined in one wind farm, they can also be arranged independently of one another.

The memory 2 is a basically conventional electrical energy storage, which in particular has a large number of batteries, but alternative storage technologies can also be provided, such as compressed air, liquid air, hydrogen or pumped storage. The operation of the memory 2 is controlled by a charge controller 20. This sets in particular a desired state of charge (SoC) that the memory 2 is to assume. The memory 2 is used, in particular, to compensate for fluctuations in the power output of the wind power installation 1 or to deliver or take up additional power on request, in short, to provide control power. All of this is also known per se and therefore need not be described further.

It should be noted that in the exemplary embodiment described here, the store 2 is in each case assigned to the wind energy installation 1 and is arranged externally to it. This is not mandatory. In particular, the store 2 can also be arranged within the wind energy installation 1 or once centrally in the collection network, as indicated in FIG. 1 by the broken line.

Another embodiment of the invention is shown in Figure 2. There are provided as a regenerative energy generator instead of the wind turbine 1 photovoltaic units 1 ', which are connected to each other and to the memory 2' via a connecting line 17 '. The photovoltaic unit T does not require rotating parts such as a wind rotor or rotating generator, it converts the radiation power applied by the sun 99 directly into electricity. Both of the wind power plant 1 and the photovoltaic unit T have in common that the respective primary source (wind or sunshine) cannot be controlled and can only be predicted to a limited extent. So there are considerable uncertainties regarding the power output. The invention is explained below using the example of a regenerative energy generator with the primary source “wind”. The same applies to other regenerative energy producers, such as special photovoltaic units 1 '.

In order to improve the performance and to make the memory 2 usable for other system services, such as delayed generation, in addition to providing control power, the invention provides an arrangement 3 comprising a correction unit 4 and a sequence control unit 5. At the correction unit 4, the control power PR to be maintained is applied as the main input variable. From this, it calculates a control signal for the state of charge SoC and applies this at its output to the charge controller 20 of the memory 2.

Further input data are applied to the correction unit 4, in particular prediction data relating to the actual primary energy (in the example: wind) Vw and to the predicted demand for energy VB. Furthermore, data for at least power P _{min to be provided} (positive and negative), trust intervals o for the various parameters as well as possibly further parameters such as storage capacity and performance, target control power and state of charge of the memory are created. Weighting factors W ,, with which the energy or service provision and delivery can be weighted as a function of time are also created. For this purpose, a time module is also provided in the correction unit 4.

The correction unit 4 also has an optimization device in which an optimization method is implemented, for example a gradient method known per se.

The interaction of the correction unit 4 with the memory 2 and its charging control 20 is as follows:

If the supply of primary energy (wind) is low, it must nevertheless be ensured that the minimum control power offered can be made available with the help of storage 2. For this purpose, the memory 2 requires a certain state of charge, which it approaches in a defined manner. In Germany, for example, where primary control power has to be offered symmetrically, this state of charge (SoC) is 50%. This start-up must be done predictively, since the control power must be available at all times. The forecast values for the wind Vw are used in particular for this purpose. In this case, the store 2 is used solely to guarantee that the primary control power is provided as required. If, on the other hand, there is enough primary energy (i.e. the wind is blowing strong enough), then the (guaranteed) positive and negative control power can be maintained both by the storage unit 2 and by the wind energy installation 1. The correction unit 4, together with the sequence control 5, now distributes the power between the wind energy installation 1 and the storage 2 in such a way that the minimum control power is guaranteed at all times, but on the other hand the storage 2 can also be used as far as possible for other system services for the network 9 , such as a production shifting.

The sequence controller 5 receives reference signals for the target power. They are preferably applied via a frequency statics 51, which is designed to specify a lower target power at network frequencies above a nominal value (plus a standard tolerance) and one at network frequencies below the nominal value (again taking into account a standard tolerance) specify higher target output.

Furthermore, the sequence controller 5 exchanges signals with the charging controller 20 for the actual power output of the wind energy installation 1 and for the actual charging status of the memory 1.

This is explained using the example of a wind farm with several wind turbines and a total nominal output of 4 MW and a total storage capacity and capacity of 2 MWh or 2 MW. If there is a low supply of primary energy, in particular lulls, the store 2 must be kept at a charge level of 1 MWh, ie 50% of its storage capacity. This ensures that the guaranteed primary control power of 1 MW can be kept available for a sufficient period of time. - However, if the wind blows sufficiently strong, the store 2 can also be fully charged. In this case, the negative control power would be provided by reducing the wind energy installation 1. If, for example, sufficient wind as well as a period with low energy demand and low energy price is predicted for the next time (48 hours), the weighting is low and the storage can be charged at a time with low prices, taking into account the weighting factor, in order to then unload it at times with higher prices (production delay). The sequence control 4 has the effect that the wind energy plants 1 adapt their respective power generation and, for example, throttle the power generation when negative control power is requested and thus relieve the memory 2 of having to consume power accordingly. This does not only include the existing wind turbines 1 and storage 2 better exploited, but also an additional yield is realized. Here, according to the invention, it is nevertheless ensured that sufficient control power is always available.

The result is visualized in Figure 4 a), b) and c). In Figure 4a) is shown with a thick solid line, the actual power generation of the wind turbines 1 which actually occurs due to the wind conditions (the scaling for the power P arranged on the y-axis of Fig. 4a is decisive), and the charge level of the memory 2 (the scaling for the charge level Q arranged on the right on the y-axis of FIG. 4a is significant) is represented by the thin solid line. A period of one week is considered (equal to 168 hours, of which only the first 150 hours are shown). Important factors (here in the form of transfer prices on the electricity exchange) are shown in Figure 4b). Available power reserves are shown in FIG. 4c), namely the power reserve to be guaranteed with the horizontal dashed line and the power reserve actually present with a solid line, namely for positive power reserve (top) and negative power reserve (bottom).

Forecast data for the further course of the wind are shown in FIG. 4a) by a dotted line in the time range 100-150h (the actual course, which is real but is not yet known at this time, is shown by the dashed line). On the basis of the state of charge of the memory 2 represented by the thin line, it can be seen that the memory 2 is loaded differently by the correction device 4 according to the invention at times of sufficient wind, depending on the (predicted) weighting factors. If only a little wind blows, the state of charge is essentially kept at the conventional value of 50% (corresponding to 1 MWh), with a lot of wind higher charging states are approached.

The power reserve available at any time is shown in FIG. 4c). As can be clearly seen, the positive reserve is always 1 MW or greater (equal or above the dashed line above), and the same applies to the negative reserve, which is always on the other side of -1 MW. It can be seen that the guaranteed values (dashed horizontal lines) for 1 MW positive control power and 1 MW negative control power are always observed. The output used for the “production shifting” is shown by the hatched area. This results in advantages not only for the operational safety and power supply of the power network 9, but also from the point of view of earnings for the operator of the wind power plant. FIG. 5 shows the additional yield that can be achieved by the additionally provided system service within the scope of the generation delay (see hatched area). This is quite relevant, especially against the background that according to the invention no additional memory is required, but the existing memory is better used in the core.

Claims

1. Arrangement comprising a regenerative energy generator (1) and a memory (2) for electrical energy for delivery to a network (9), the regenerative energy generator (1) an energy converter (13) for converting renewable energy sources into electrical energy and a power control unit (10) which has an input for a desired power output and controls the power generation by the regenerative energy converter (13), and furthermore a charging control (20) for the memory (2) is provided which has a minimum to be provided Sets control energy and outputs a signal for a target charge level (SoC) of the store (2),

characterized in that

a correction unit (4) is connected to the charging control (10), which continuously modifies the signal for the target charge level (SoC), and a sequence control unit (5) is provided, which is controlled by the correction unit (4) and on which Power control unit (10) of the regenerative energy generator acts depending on an output value of the correction unit (4).

2. Arrangement according to claim 1, characterized in that the sequence control unit (5) acts on the power control unit (10) such that it is in a master / slave relationship with the correction unit (4) of the memory, the correction unit (4 ) is the master.

3. Arrangement according to claim 1 or 2, characterized in that a correction for available primary power, in particular prediction values for the available primary power, is applied as an input parameter to the correction unit (4).

4. Arrangement according to claim 3, characterized in that the correction unit (4) is designed for a plurality of input parameters, with further parameters predicting values for electricity requirements in the network, minimum values for power to be kept available, storage capacity and power, target control power, and / or Charge level of memory

5. Arrangement according to claim 4, characterized in that additional static parameters are created, in particular for a confidence interval (o).

6. Arrangement according to one of claims 3 to 5, characterized in that the correction unit (4) has an optimization device which is designed to determine the signal for the target state of charge (SoC) on the basis of the input parameters by means of an optimization module.

7. Arrangement according to claim 6, characterized in that a gradient method, a neural network or an evolutionary algorithm is implemented in the optimization device as the optimization method.

8. Arrangement according to one of claims 3 to 7, characterized in that the input parameters are staggered in time, and in particular the correction unit (4) evaluates the input parameters as a function of time.

9. Arrangement according to one of claims 3 to 8, characterized in that weighting factors (W,) are provided for at least one of the input parameters, in particular for forecast electricity demand in the network and / or he generated power, and preferably separately for positive and negative.

10. Arrangement according to one of the preceding claims, characterized in that a reference signal for a target power output is applied externally or generated internally, in particular by means of a frequency statics (51).

1 1 Method for operating an arrangement comprising a regenerative energy generator and a store for electrical energy for delivery to a network, wherein the regenerative energy generator has an energy converter for converting renewable energy sources into electrical energy and a power control unit, which has an input for a target power output and controls the power generation by the regenerative energy converter, and furthermore a charging control for the memory is provided, which a sets the minimum control energy to be reset and outputs a signal for a target charge level of the storage unit,

marked by

continuously modifying the signal for the target charge level by means of a correction unit connected to the charge control, and changing the power generated by the regenerative energy generator by means of a sequential control unit which is controlled by the correction unit and acts on the power control unit of the regenerative power generator as a function of one From the initial value of the correction unit.

12. The method according to claim 11, characterized by using an arrangement according to one of claims 2 to 10.