Systems and methods for power variability reduction FIELD OF THE INVENTION

The present invention relates to a system, in particular to a control system, for reducing power variability of an energy generation system. The present invention further relates to a method for reducing power variability of an energy generation system and to a respective computer program.

BACKGROUND With the increasing commissioning of renewable energy genera tion systems, for example wind farms or solar farms, there is a rising concern about the problems associated with the vari ability of the power output thereof. Considering wind farms, the wind strength is variable, wind may even not be available at all and strongly depends on meteorological conditions.

This results in that the generated power output may comprise highly dynamic power variations over time, which may destabi lize the power system. Grid codes are therefore in force which impose limits on the power variability, and which re- quire smoothing of power variations to improve the stability of power systems. In order to at least partly compensate pow er variations output by a renewable energy generation system, such a system can be complemented by an energy storage sys tem, which is used in order to deliver or receive a compen- sating portion of power in case of power deviations, for ex ample in case of power overshoots or undershoots. Forming re serves of active power, which have to be available to the en ergy generation system in the required time, implies high costs for the system. Since the compensation of occurring power variations can require high dynamics of the power flow associated with the energy generation system, and since oc curring power variations can last over a longer period of time (30 minutes or more), the power requirements for the en- ergy storage system are highly demanding. The required high performance may also result in faster aging of the energy storage system. Further, while compensating power variations, the charging state of the energy storage system may drift to a fully charged or discharged state. To prevent the system from reaching such a system state, the energy storage system is appropriately sized, more specifically oversized, in order to ensure that the energy storage system does not reach its technical limits. Accordingly, such energy storage systems are designed generously or are oversized in order to meet the outlined demanding requirements. The oversizing results in a high consumption of resources and an increased maintaining effort, both of which are associated with increased costs. Document JP 2004147445 A relates to controller that extracts the power variations from an effective power of a load de tected by an effective power detector, and extracts a fre quency oscillation component from a frequency detected by a frequency detector. The controller controls charging and dis- charging of a power storage unit based on a sum of a first command value for suppressing the power variations, and a second command value for suppressing the power oscillation.

Document US 2015/194820 A1 is related to a control apparatus and corresponding control method which use per-unit filtering in a plurality of power-sharing controllers to obtain a pow er-sharing command signal for respective ones among a plural ity of different energy storage units in a hybrid energy storage system.

SUMMARY

Accordingly, there is the need to mitigate at least some of the drawbacks mentioned above and to reduce power reserve re- quirements for an energy storage system associated with the compensation of the power variability of an energy generation system. This need is met by the features of the independent claims. The dependent claims describe embodiments of the invention.

According to an embodiment of the invention, a control system for providing a control signal to an energy storage system to reduce power variability of an energy generation system, e.g. a power plant, is provided. The energy storage system is cou pled to the power output of the energy generation system and configured to at least one of receive and provide electrical power in response to the control signal. The control system comprises a monitoring unit configured to monitor the power output of the energy generation system and to provide a re spective monitoring signal. The control system further com prises a control signal generation unit configured to gener- ate the control signal from the monitoring signal, wherein the control signal generation unit implements a filtering unit that has a band pass response. The filtering unit passes a predetermined frequency band of the monitoring signal and the control signal generation unit is configured to generate the control signal in said frequency band so as to compensate power variations of the power output of the energy generation system in said frequency band.

In an example, the control signal for compensating said vari- ations is derived only from said monitoring signal in said predetermined frequency band so that the power variations are only compensated in said frequency band.

An energy generation system equipped with such a control sys- tern focuses the compensation of power variations on a defined frequency band and avoids compensation of power variations beyond it. The power output to be fed into the power grid is smoothed at the relevant frequencies. Accordingly, a reduced amount of energy has to be transmitted by the energy storage system in order to compensate power variations occurring in the output of the energy generation system. Reserve power re quirements for the energy storage system can therefore be re duced. Limiting the smoothing action to the frequency band also results in a reduced burden on the energy storage sys tem. As a consequence, associated costs and consumption of resources can be decreased, for example by adapting the siz ing of the energy storage system in accordance with the re- duced reserve power requirements.

The control signal is in particular generated such that by operation of the energy storage system, the power variations in the combined output of energy generation system and energy storage system are reduced (compared to operation of the pow er generation system alone). In other words, the control sig nal is generated such that the charging/discharging of the energy storage system smooths the power output of the power generation system.

The control signal may be a reference power signal indicating the power to be received or to be provided by the energy storage system. Further, the predetermined frequency band may comprise fre quencies that contribute majorly to the power variations of the power output of the energy generation system. Thus, it should be clear that that the predetermined frequency band may be tailored to the energy generation system characteris- tics of the energy generation system and/or the operation characteristics of the energy generation system. Major con tribution means that the predetermined frequency band in cludes frequencies that, if smoothed out, mainly reduce the probability of occurrence of non-admissible power variations, for example down to a probability below 50%, below 10%, or even below 1% or 0.1%. The frequency band may, e.g., include a continuous range of frequencies including at least 25% preferably at least 40% of the power of the variations, e.g. of the power in the frequency spectrum of the power varia- tions. For example, high frequency components are mostly re lated to power flicker but rather do not induce non- admissible power variations, and these can further be compen sated by internal controls of the power generation system. Lower frequencies are rather related to longer term drift and can be compensated on a grid level. By avoiding such frequen cies using the frequency band, a more optimal design of the energy storage system can be employed, leading to a signifi- cant decrease in its size and ageing. Frequencies that major- ly contribute to the power variations may be identified based on an analysis of power variation data and preferably by us ing an optimization algorithm. Said non-admissible power variations may occur when power variations occur that exceed a predetermined frequency range, and/or when variations occur that exceed a predetermined am plitude. Hence, the predetermined frequency band may be predetermined such that a probability of an occurrence of a non-admissible variation in the power output of the energy generation system is below a (predetermined) threshold. Said non-admissible variation may exceeds a predetermined frequency range and/or predetermined amplitude

In particular, the threshold may for example be a probability below 50%, below 10%, or even below 1% or 0.1% as mentioned above. The threshold may for example be derived from said historical/empirical data of the power variation data, e.g. based on an analysis of a probability density function and a cumulative distribution function of the power variations as herein described. As said, the frequency band that reduces the probability of occurrence of non-admissible power varia- tions down to the (predetermined) threshold may for example be identified based on said data and by using an optimization algorithm.

The historical/empirical data may for example be collected during different operation modes of the energy generation system, e.g. while being connected and/or disconnected to an (external) grid, e.g. a DC or AC grid. It should be clear that the control system may then be tai lored to operate in said operation modes and smooth the power output of the energy generation system, e.g. while being con nected and/or disconnected to an (external) grid, e.g. a DC or AC grid. It should further be clear that such frequency band that is specifically tailored to the power generation system is advantageous since it reduces the power variability more effectively and, accordingly, reduces the stress acting on the system.

The predetermined frequency band may include a frequency of 0.1 Hz and preferably the predetermined frequency band at least comprises frequencies between 10 mHz and 30 mHz. The frequency band may for example include frequencies between 1 mHz and 30 mHz, or between 1 mHz and 1 Hz. The predetermined frequency band may for example exclude frequencies below 10 ^-4 Hz, or below 5*10 ^-4 Hz. The frequency band may be further configured in order to smooth effects associated with band- widths related to frequency deviations or voltage flicker. In particular, the frequency band may correspond to a frequency range determined for secondary frequency regulation by an op erator of the power grid. The frequency band is preferably continuous. In an embodiment, the filtering unit comprises a band-stop filter. The band-stop filter can be implemented by a combina tion of a high-pass filter and a low-pass filter. As an exam ple, the high-pass filter and the low-pass filter may be con nected in parallel, wherein the output signals of both fil- ters are summed, the high-pass filter defining an upper boundary of the frequency band and the low-pass filter defin ing a lower boundary of the frequency band.

The monitoring signal may be filtered by the band-stop fil- ter, and the monitoring signal is subtracted from the fil tered monitoring signal to obtain the band pass response. By the subtraction, the signal components passed by the bandstop filter cancel out, so that only the inverted frequency compo- nents of the stopband remain due to the subtraction. Accord ingly, a bandpass response between the lower and upper bound aries is obtained, the signal being inverted compared to the monitoring signal.

Alternatively, the filtering unit may directly implement a band-pass filter having a pass band between the lower and up per boundaries, which may for example be obtained by a series connection of a high-pass and a low-pass filter. An inverter may furthermore be used to invert the monitoring signal be fore or after filtering, so that the resulting control signal results in a power flow from the energy storage system that negates the variations. For example, the part of the control signal that compensates the power variations may only comprise frequencies within said frequency band. Frequencies outside the frequency band may be attenuated (significantly) or not be present. The con trol signal may accordingly have a frequency content such that only power variations at frequencies within said fre quency band are compensated. The control signal may only con sist of this part that compensates the power variations, or may include an additional part that is not providing compen sation of the power variations, for example a part that con- trols the charging state of the energy storage system, as ex plained in more detail below.

In an example, the control signal is (only) in said frequency band when the control signal is received at the energy stor- age system.

In an example, the control signal generation unit may output the control signal and operate on the monitoring signal as a filter having a band-pass response. Hence, the control signal may be generated as the band-pass filtered (and gained) moni toring signal. In an example, the control signal generation unit may direct ly provide the control signal to the energy storage system.

In an example, the control signal may be the only signal that is provided to the energy storage system and/or that is re ceived at the energy storage system.

In an example, the energy storage system may be only con trolled by the signal that is output by the control signal generation system.

In an example, the energy storage system is coupled to the power output of the energy generation system at a coupling point and the coupling point provides a variation compensated power output by combining the power output of the energy gen eration system and power that is at least one of received and provided by the energy storage system, wherein the variation compensated power output corresponds to a band-stop response on the power output of the energy generation system, wherein the stop-band corresponds to a pass-band of the band pass re sponse.

In an example the control system is configured to generate the control signal only within, or alternatively only within and below said frequency band.

Thereby, it may be ensured that the control signal is re stricted to the predetermined frequency band and smooths/reduces the variability of the power output of the energy generation system only in the predetermined frequency band. This renders the smoothing operation/ variability re duction more efficient since the energy/power required there fore is reduced. Hence, the stress acting on the system is also reduced.

In an example, the control signal may be generated inde pendently/without being influenced by the system behavior or input/output behavior of the energy storage system. In other words, the control signal may be generated based on the monitoring signal in a feedforward way. A signal that comprises information about the energy storage system, e.g. the output power, and is fed back, e.g. by a feedback loop, may not necessarily be present and may not be employed in the control signal generation. Such structure allows generating unitary control signals based on the same respective input for different energy storage systems. That may facilitate the planning, manufacturing, and implementation of the control system. Further, such structure requires less computation ef fort and may accordingly allow reducing a sampling time ac cording to which a value of the control signal is derived. Thus, the reduction of the variability may be more effective.

In an example, the control signal may be generated only from the monitoring signal and additionally or alternatively based only on such signals that originate from/are derived from/are based on the monitoring signal. Preferably, either the con- trol signal or the part that compensates power variations may be generated only from a signal that passes the filtering unit.

This may reduce the hardware and implementation effort due to the fact that, e.g., only one signal is required to be moni tored.

In an example, the control system may be configured to gener ate the control signal only in the frequency band. In another embodiment the control system may be configured to generate the control signal in the frequency band (for compensating power variations) and below the frequency band (e.g. for con trolling a charging state).

Such control signal controls the energy storage system such that it is still capable to efficiently reduce the power var iability. Such control signal may however be additionally ca pable to transport information related to a further function- ality, e.g. information that may ensure a stable state of charge of the energy storage system.

In an embodiment, the control system further comprises a charging state controlling unit, wherein the charging state controlling unit generates a charging state control signal provided to the energy storage system to adjust a charging state of the energy storage system in accordance with a charging state reference signal.

Preferably, the control system is configured to adjust the control signal of the control signal generation unit by the charging state control signal, for example by adding the charging state control signal to the control signal. The charging state is in particular controlled such that the en ergy storage system does not drift into a fully charged or fully discharged state (in which it could not take up or sup ply electrical power, respectively), but is driven towards this reference.

The charging state reference signal can be a fixed, predeter mined reference, for example representing a constant prede fined value, or a variable reference, which may be derived during operation.

Using such a charging state controlling unit, it can be en sured that the charging state cannot drift to a destabilized fully charged or discharged state but is kept at a predeter mined level (i.e. close to the reference). Accordingly, bur- den on the energy storage system is decreased, and it can al so be ensured that smoothing of the power variations can be performed over prolonged periods of time, even if losses oc cur in the energy storage system. The charging state controlling unit may be configured to de rive the charging state reference signal from the monitoring signal under consideration of the transfer function of the filtering unit. The charging state controlling unit preferably includes a modelled filter that represents at least a model of the charging state of the energy storage system to derive the charging state reference signal.

In an example, the charging state reference signal is derived such that the charging state reference signal at least repre sents a modelled actual charging state of the energy storage system. In particular, the modelled filter may represent the charging state of the energy storage system in the ideal case, i.e. in the case in which a lossless system is assumed.

The modelled filter preferably includes at least a second or- der low-pass filter or the modelled filter preferably in cludes at least an integrator.

In an example, the model of the charging state of the energy storage system derives the charging state reference signal from the monitoring signal. For example, when the modelled filter includes at least the second order low-pass filter, the output signal of the modelled filter may be directly de rived from the monitored signal, and when the modelled filter includes at least the integrator, the output signal of the modelled filter may be derived from the control signal, the control signal being based on the monitoring signal.

A model-based charging state control ensures a stable state- of-charge while minimizing the effect of the charging state control on the in parallel performed smoothing operations that are applied to the power output of the energy generation system. Even though the state of charge is controlled and drifts therefrom are prevented, the system consumes still ap proximately the same energy as in the ideal case (lossless system).

In an embodiment, the charging state controlling unit further comprises a gain unit in order to tune the effect of the charging state controlling unit on the reduction of power variability. For example, a scaling factor may be applied to the charging state reference signal. Preferably the scaling factor comprises a range from zero to one.

Such a gain unit allows tuning or adjusting the effect of the charging state control on the performance of the variability reduction, i.e. how much the control signal that controls the power flow from the energy storage system for smoothing is 'disturbed' by the charging state reference signal. Using an appropriate scaling factor allows further optimization of the efficiency of both operations.

According to an example, the charging state reference signal comprises a minimum charging state value that ensures that the charging state of the energy storage system is kept to the minimum charging state value, wherein the minimum charg ing state value is preferably predetermined. The minimum charging state value may be the only signal com prised in the charging state reference signal or additionally comprised in the charging state reference signal together with other signals, e.g. as a sum of the comprised signals.

According to another embodiment of the invention, a system configured to reduce power variability of an energy genera tion system is provided. The system comprises an energy stor age system coupled to the power output of the energy genera- tion system. The energy storage system is configured to at least one of receive and provide electrical power in response to a control signal. Further, the system comprises a control system having any of the configurations described herein, which is coupled to the energy storage system to provide the control signal to the energy storage system to reduce power variability of the energy generation system. The system may further comprise the power generation system. The energy generation system may further comprise a renewable energy power source. The renewable energy power source pref erably is at least one of one or more wind turbines and one or more solar power systems.

According to another embodiment of the invention, a method for providing a control signal to an energy storage system to reduce power variability of an energy generation system is provided, wherein the energy storage system is coupled to the power output of the energy generation system and configured to at least one of receive and provide electrical power in response to the control signal. The method comprises monitor ing, by a monitoring unit, the power output of the energy generation system and providing a respective monitoring sig- nal and generating, by a control signal generation unit, the control signal from the monitoring signal, wherein the gener ating includes at least the use of a filtering unit that has a band pass response, the filtering unit passing a predeter mined frequency band of the monitoring signal, wherein the control signal is generated in said frequency band so as to compensate power variations of the power output of the energy generation system in said frequency band. By such method, ad vantages similar to the ones outlined further above may be achieved.

The control signal may in particular be generated as a band pass filtered and inverted version of the monitoring signal, wherein further signal processing can be applied, such as a real or complex gain or the like.

The method may further comprise generating, by a charging state controlling unit, a charging state control signal, ei ther based on a predefined charging state reference or in ac cordance with the response of a modelled filter, and provid- ing the charging state control signal to the energy storage system. In an embodiment, the method further comprises adjusting the control signal of the control signal generation unit by the charging state control signal, for example by adding the sig nals.

According to another embodiment of the invention, a computer program for providing a reference power signal for an energy storage system to reduce power variability in an energy gen eration system is provided. The computer program comprises control instructions which, when executed by a processing unit of a control system coupled to the energy storage sys tem, cause the processing unit to perform any of the methods described herein. According to another embodiment of the invention, a control system controlling the charging state of an energy storage system is provided. The energy storage system is configured to reduce power variability of an energy generation system and is coupled to the power output of the energy generation system and configured to at least one of receive and provide electrical power in response to a control signal generated by processing a monitoring signal with a filtering unit, wherein the monitoring signal is indicative of a power output of the energy generation system. The control system comprises a charging state controlling unit configured to generate a charging state control signal for adjusting the charging state of the energy storage system, wherein the charging state control signal is generated based on a monitored charg ing state of the energy storage device and a charging state reference signal, wherein the charging state controlling unit is configured to derive the charging state reference signal from the monitoring signal under consideration of the trans fer function of the filtering unit. As the transfer function is decisive for the generation of the control signal and thus for the flow of energy into and out of the energy storage de vice, considering this transfer function in the generation of the reference for the charging state may result in a control that minimizes the power reserve requirements and that main- tains the charging state at a level suitable for compensating the occurring power variations.

The charging state control signal may for example be formed by subtracting the monitored charging state of the energy storage device from the charging state reference signal.

Further, the control system controlling the charging state may be configured to add the charging state controlling sig- nal to the control signal to obtain an adjusted control sig nal, and to control the energy generation system by means of the adjusted control signal.

The charging state controlling unit may include a modelled filter, wherein the modelled filter represents at least a model of the charging state of the energy storage system, the modelled filter being determined based on the transfer func tion of the filtering unit. The modelled filter may be con figured as described above, it may in particular be imple- mented as an at least second order low pass filter.

The charging state controlling unit may further comprise a gain unit in order to tune or adjust the effect of the charg ing state controlling unit on the reduction of power varia- bility. Such gain unit may be implemented as described above and may provide the respective advantages.

The control system may furthermore include any of the fea tures and aspects described herein above and further below with respect to the generation of the control signal for con trolling the energy storage device, in particular the moni toring unit and the control signal generation unit.

According to another embodiment of the invention, a corre- sponding method of controlling the charging state of an ener gy storage system is provided. The method may include gener ating, by a charging state controlling unit, a charging state control signal for adjusting the charging state of the energy storage system, wherein the charging state control signal is generated based on a monitored charging state of the energy storage device and a charging state reference signal, and de riving the charging state reference signal from the monitor- ing signal under consideration of the transfer function of the filtering unit.

According to a further embodiment, a respective computer pro gram that comprises control instructions which implement the respective method of controlling the charging state is pro vided.

Each computer program may be implemented on a storage medium or data carrier, such as a volatile or non-volatile storage medium.

Any of the methods or method steps described herein may be performed by the control system in any of the configurations disclosed hereinabove or further below. Likewise, the control system may be configured to carry out any of the method steps described herein.

It is to be understood that the features mentioned above and those yet to be explained below can be used not only in the respective combinations indicated, but also in other combina tions or in isolation, without leaving the scope of the pre sent invention. In particular, the features of the different aspects and embodiments of the invention can be combined with each other unless noted to the contrary.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing and other features and advantages of the inven tion will become further apparent from the following detailed description read in conjunction with the accompanying draw ings. In the drawings, like reference numerals refer to like elements. Fig. 1 is a schematic drawing illustrating a control sys tem generating a control signal for reducing power variabil ity of an energy generation system according to an embodiment of the invention.

Fig. 2 is a schematic signal flow chart illustrating a generation of a control signal according to an embodiment of the invention. Fig. 3 is a schematic flow diagram describing a method for generating a control signal to reduce the power variability of an energy generation system according to an embodiment of the invention. Fig. 4 is a schematic diagram illustrating an amplitude spectrum of a monitoring signal indicating the power output of an energy generation system and filtered variants thereof using either a low-pass filter or a band-stop filter. Fig. 5 is a schematic diagram illustrating probability density functions and cumulative distribution functions of power variations occurring during different operation modes of an energy generation system. Fig. 6 is a schematic signal flow chart illustrating a principle control scheme for controlling a charging state of an energy storage system according to an embodiment of the invention. Fig. 7 is a schematic signal flow chart illustrating a principle control scheme for controlling a charging state of an energy storage system according to an embodiment of the invention. Fig. 8 is a schematic flow diagram describing a method for controlling a charging state of an energy storage system and for reducing power variability of an energy generation system according to an embodiment of the invention. Fig. 9 is a schematic diagram illustrating the charging state of an energy storage system while applying and not ap plying a charging state control according to embodiments of the invention.

DETAILED DESCRIPTION

In the following, embodiments of the invention will be de- scribed in detail with reference to the accompanying draw ings. It is to be understood that the following description of the embodiments is given only for the purpose of illustra tion and is not to be taken in a limiting sense. It should be noted that the drawings are to be regarded as being schematic representations only, and elements in the drawings are not necessarily to scale with each other. Rather, the representa tion of the various elements is chosen such that their func tion and general purpose become apparent to a person skilled in the art. As used herein, the singular forms "a," "an," and "the" are intended to include the plural forms as well, un less the context clearly indicates otherwise. The terms "com prising," "having," "including," and "containing" are to be construed as open-ended terms (i.e., meaning "including, but not limited to,") unless otherwise noted.

Fig. 1 shows a schematic signal flow chart illustrating a principle control scheme for reducing power variability of an energy generation system 101 according to an embodiment of the invention. The energy generation system 101 generates electrical power Ps and the output thereof is coupled to the input of a power grid 111 via a summation point 109. The en ergy generation system can be, but is not limited to, a sin gle power plant generating electrical power or general source of power. It is also possible that the energy generation sys- tern is composed by a plurality of power plants or general sources of power and these can be of different types. Exam ples are power plants of renewable energy, e.g. solar power plants and/or wind farms. The electrical power Ps may include inadmissible - e.g. according to grid codes - power varia tions. The power variations are compensated at the summation point 109 by combining the generated power Ps with a compen sating power flow Pc provided by an energy storage system 107 in accordance with a power reference signal generated by a control system 100. Thereby, variability reduced power Pvr is generated.

An output of the energy storage system 107 is connected to the summation point 109. The magnitude of the compensating power flow Pc is controlled by a control signal Pr input to the energy storage system 107. The control signal Pr may be a power reference signal actuating the energy storage system 107 in such a way that the output compensating power flow Pc corresponds to the input control signal Pr. The energy stor age system 107 is capable of receiving and providing electri cal power as needed in accordance with the control signal Pr. The energy storage system 107 may be, but is not limited to, an electro-chemical energy storage system, e.g. a battery or an accumulator system. An example for an energy storage sys tem 107 other than the electro-chemical system type can be a thermal energy storage system.

Control system 100 comprises a control signal generation unit 103, whose output is connected to the input of the energy storage system 107. The control signal generation unit 103 generates the control signal Pr on the basis of a monitoring signal Psm. The monitoring signal Psm is obtained by monitor ing the electrical power Ps provided by the energy generation system 101 using a monitoring unit 112. The output of the monitoring unit 112 provides the monitoring signal Psm to the control signal generation unit 103.

The monitoring unit 112 may include an interface for receiv- ing a signal indicating the power output of system 101, or it may be coupled to or include a single sensor or a sensoring system measuring the output power, the unit providing the re spective monitoring signal Psm. It is also possible that the monitoring signal Psm is obtained based on data via a respec tive interface of monitoring unit 112, which data comprises information about the power generated by the energy genera tion system 101 itself, for example if the information is provided by a controller of the energy generation system 101

(e.g. power plant controller). Further, it is possible that the signal Psm is based on data, which may be used to approx imate, estimate or reconstruct the power Ps generated by the energy generation system as monitoring signal Psm, for exam- pie by an appropriate model or observer system. It should be clear that the monitoring of the monitoring unit 112 is not limited to the given examples.

The control signal generation unit 103 comprises a general filtering unit 106 smoothing the input monitoring signal Psm and providing a filtered monitoring signal at its output. For example, filtering unit 106 generates a filtered monitoring signal Pf of electrical power provided by the energy genera tion system 101 by applying a filter 104 or filter system to monitoring signal Psm. The filtering unit 106 has a band-pass response that passes a predetermined frequency band of the input signal Psm. Such band-pass response can be implemented in different ways. In some implementations, the output of the filter 104 and the input to the filter 104 (e.g. the signal Psm) are connected to summation point 105. At summation point 105, the monitoring signal Psm is subtracted from the fil tered monitoring signal Pf. Such implementation is particu larly suitable if filtering unit 104 implements a band-stop filter. In other implementations, summation point 105 may not be present, e.g. when implementing filter 104 as a band-pass filter. The output of the summation point 105 provides the output of the filtering unit 106, and in the example of fig ure 1 also the output of the control signal generation unit 103. The result is provided as control signal Pr to the ener- gy generation system 107. It should be clear that in other implementations, the control signal generation unit 103 may comprise further signal processing elements, such as real or complex gain elements or the like, and these may be applied prior to or after the filtering with filtering unit 106 to generate the final control signal Pr.

In general, the topology of control system 100 extracts in a first step an inverse signal of power variations occurring in the power Ps generated by the energy generation system based on frequency bands defined by filtering unit 106. The inverse signal of power variations is applied in a second step as control signal Pr to the energy storage system 107 to actuate it such that it applies at summation point 109 a power flow in accordance with control signal Pr to the power Ps generat ed by the energy generation system 101. Accordingly, occur ring power variations at frequencies predetermined by the filtering unit 106 are compensated. Resulting variability re- duced electrical power Pvr is then provided to the power grid 111.

The control system 100 may include a processing unit 113, which may comprise any type of processor, such as a micropro- cessor, an application specific integrated circuit, a digital signal processor or the like, and a memory 114. The memory 114 may include volatile and non-volatile memory, in particu lar RAM, ROM, FLASH-memory, hard disc drives and the like.

The memory 114 stores control instructions which, when exe- cuted by processing unit 113, cause the processing unit 113 to perform any of the methods described herein. Control sys tem 100 further includes input and output interfaces for re ceiving data (e.g. monitoring unit 112) and for transmitting control data (e.g. the control signal Pr) to the energy stor- age system. In some implementations, unit 103 and filtering unit 104 may be implemented in form of software, and the pro cessing of the monitoring signal may occur digitally. As such, it should be clear that the monitoring signal Psm and/or the control signal Pr may be a digital signal. In oth- er implementations, signal processing and in particular fil tering may occur in analog form, and filtering unit 104 may implement respective electric or electronic analog filters. Figure 2 is a schematic signal flow chart illustrating a gen eration of the control signal Pr according to an embodiment of the invention. Figure 2 shows an implementation of the control signal generation unit 103 according to figure 1. The control signal generation unit 103 is configured such that filtering of monitoring signal Psm acts on specific frequen cies that are of most interest for the electrical system op eration with regard to a reduction of power variability. For this purpose, the filtering unit 106 as shown in figure 1 im- plements the filter 104 as a band-stop filter 104a. The band- stop filter 104a can be implemented by a combination of a low-pass filter and a high-pass filter, which determine the lower and upper limits of the frequency bandwidth, respec tively. The band-stop filter can for example be implemented by a parallel structure of the low-pass and high-pass fil ters, wherein the response of both filters to a common input signal is added. As an alternative, the high-pass filter of the proposed structure can also be substituted by a second low-pass filter, wherein the response of the second low-pass filter is subtracted from the input signal of the second low- pass filter. In the exemplary implementation, a fist-order version of a low-pass filter and a high-pass filter connected in parallel are used. Accordingly, following the scheme pre sented in figure 2, the filtered monitoring signal Pf can be calculated in the s-plane based on the monitoring signal Psm as

1

Pf(s)= P _sm (s)( + (eq. 1)

¾s+1 T _H S+l)' wherein z _L is the time constant of the low-pass filter defin ing the lower bandwidth limit, t _H is the time constant of the high-pass filter defining the higher bandwidth limit and s is the complex frequency. This solution for the implementation of a band-stop filter can be understood as a general guide- line to designing a filter that affects specifically the de sired frequencies, which are related to reserve requirements and frequency deviation events. In the present example, a combination of a first order low- pass filter and a first-order high-pass filter are used. The use of a higher-order filter can also be considered if it is required to smooth the power variations in an appropriate way. Considering the topology of filtering unit 106, it is clear that the filtering unit 106 operates as a band-pass filter (of first or higher order) with a negative gain if a band-stop filter 104a is chosen for filter 104. Accordingly, the filtering unit 106 can alternatively be implemented by a respective band-pass filter or a filter system comprising a plurality of filters, wherein the input-output behavior of the filter system comprises an input-output behavior that corresponds to that of a band-pass filter. An example may be a parallel structure of low-pass filters or a serial struc- ture of a high-pass and a low-pass filter.

Further, taking into account the topology of figure 1 and as suming an ideal control system 100 and energy storage system 107, the input-output behavior of the system for reducing power variability, in which the generated power Ps is the in put signal and the variability reduced power Pvr is the out put signal, corresponds to a band-stop filter (power varia tions in the defined frequency band are smoothed). The band-stop filter topology as presented in figure 2 can be implemented using existing discretization methods for the im plementation of filters. In the following example, Euler- backward method is assumed. Thus, for a sampling time t _s , the response P _{S7rL i} {t) of the low-pass filter to the monitoring sig- nal Psm results in wherein t is time and k is an integer greater zero, and the start value may be initialized. The response i’ _sm .iiC °f the high-pass filter to the monitoring signal Psm results in wherein t is time and k is an integer greater zero, and the start value may be initialized. The response Pf of the re spective band-stop filter therefore is

Pf(t)= P _Sm,H (0 + Psm,L(t), (eq. 4). Accordingly, the control signal Pr can be calculated by

The control signal generation unit 103 hence generates a con trol signal Pr in a predetermined frequency band so as to compensate power variations of the power output of the energy generation system in said frequency band. Since the system actuation of the energy storage system 107 focuses on specif ic frequency bands and avoids others, energy consumption of the energy storage system 107 is reduced and therefore the generated compensating power flow Pc is applied in a more ef ficient way. Since the burden on the energy storage system 107 is reduced, also power requirements thereof are reduced and the system can be dimensioned more compact. Further, due to the reduction of stresses (reduced charging/discharging cycling; level of charging/discharging currents), the pre sented invention has also a positive effect on the aging of the energy storage system 107.

Figure 3 shows a schematic flow diagram describing a method 300 to be performed to reduce power variability of an energy generation system as for example presented in figure 1 and 2 according to an embodiment of the invention. The operational method comprises steps S302-S303. In some embodiments, the method may also comprise step S301 and/or step S304. Accord- ing to the optional first step S301, the frequency band con tributing majorly to the power variations of the power output of the energy generation system is predetermined. So, per forming step S301 configures the control signal generation unit in such a way that it generates a control signal operat- ing the energy storage system so as to compensate variations associated with the frequency band(s) specified. This allows an optimal tuning of the control system for a specific pur pose. For example, it can be avoided that the control system attempts to reduce variations at very low frequencies, which are usually related to energy shifting applications, or vari ations at very high frequencies, which are out of the scope of reserve requirements. Thus, the control system is opti- mized to smooth specifically the variations that are of most interest. Step S301 may be performed only once in order to initialize the filter system. For example, the frequency band contributing majorly to the power variations can be ideally predetermined by an optimization algorithm and the filter system can be tuned in accordance with the result of the op timization algorithm. Step S301 may also be performed repeat edly in order to adapt the reaction of the control signal generation unit on specific frequency bands dynamically. According to the second step S302, the power output of the energy generation system is monitored, e.g. by the monitoring unit 112. Further, the respective monitoring signal Psm is provided. The signal is received at the control signal gener ation unit and based on the monitoring signal, a control sig- nal is generated, wherein the generating includes at least the use of the filtering unit 106 that has a band pass re sponse. The filtering unit passes a predetermined frequency band of the monitoring signal, wherein the control signal is generated in said frequency band so as to compensate power variations of the power output of the energy generation sys tem in said frequency band. Performing step S303 may also comprise updating the response of the filtering unit. There by, the control signal may be generated based on the equa tions eq. 1-5. The control signal is received at the energy storage system and in accordance with the control signal, power is applied to the power output of the energy generation system (S304). Power variations of the power output of the energy generation system are thereby compensated. Figure 4 is a schematic diagram illustrating an amplitude spectrum of a power signal generated by an energy generation system as presented and of filtered variants thereof using either a low-pass filter or a band-stop filter. The diagram shows the absolute value of the amplitude of signals 401, 402 and 403 in the frequency space. The amplitude spectrum is ob tained by a fast Fourier transform (FFT) of the respective time signals of the power output by the energy generation systems and the filtered variants thereof. Considering fig ures 1 and 2, the diagram of figure 4 exemplarily shows sig nal 401 representing the electrical power Ps provided by en ergy generation system 101. Signals 402 and 403 represent the smoothed or variability reduced power Pvr using either a band-stop filter, BSF, (signal 402) or a low-pass filter,

LPF, (signal 403). The response using a LPF presents a low gain reducing variations for frequencies above 10 mHz. On the other hand, the response using a BSF only has an effect on a specific frequency band around 0.1Hz. More specifically, the BSF affects frequencies at least between 10 mHz and 30 mHz, preferably 1 mHz and 30 mHz, e.g. between 1 mHz and 1 Hz, but mostly around 0.1 Hz. This is one of the advantages of the solution of using a band-stop filter which can be tuned for the specific frequency range that should be smoothed.

Figure 5 is a schematic diagram illustrating the probability density functions (510, 530, 550), PDFs, and cumulative dis tribution functions (520, 540, 560), CDFs, of power varia tions occurring during different operation modes of an energy generation system as for example presented in figures 1 and

2. A CDF depending on a variable x and corresponding to a re spective PDF generally represents the probability p of an oc currence of an event according to a value less than or equal to the given value x. Considering figure 5, value x repre- sents an upper limit of power variation. Since reserves are not sized to compensate all possible variations, CDFs that correspond to respective PDFs of occurring power variations can be used to determine the size of the reserves required to compensate the system variability. Power reserves are consid- ered in different time frames according to the device's re quired actuation time. For energy generation systems, usual primary and secondary control level ranges of 1 minute and 30 minute variations are considered in figure 5. The decrease of reserve requirements is evaluated as shown in figure 5 using the cumulative distribution function (CDFs) of the variations at each time range. Diagrams 530 and 540 represent 1 minute variations and dia grams 550 and 560 represent 30 minute variations. Further, 1 second variations are shown that are represented by diagrams 510 and 520. Each diagram 510-560 contains values regarding variations occurring in generated power Ps (502, 504) and filtered variants thereof (signal Pvr) for both, a LPF (501, 506) and a BSF (503, 505). The CDFs at 99% probability are evaluated and obtained results are shown numerically in table 1. It can be seen that a filtering unit using a LPF reduces the CDF for the 1 second and 1 min variations but hardly af- fects 30 min variations. On the other hand, a filtering unit using a BSF affects predominantly the 1 min variations.

Table 1: Maximum power variations occurring with a probabil ity of 99%

Considering control signal Pr, the control system therefore focuses on the compensation of power variations in a prede termined frequency band and avoids compensation beyond it. In the following passage, the effect of the filtering on the stress of the energy storage system is analyzed. The used en ergy storage system stress parameters are summarized in table 2. It is noted that regarding Emax (maximum capacity), pre cise nominal capacity and state of charge (SOC) calculation is assumed and regarding Pmax (maximum power), symmetrical charge and discharge power requirements are assumed.

Table 2: Summary of energy storage system stress parameters

Table 1: Results of energy storage system stress evaluation

Table 3 underlines that an appropriate design of the filter ing unit correlates with reduced stress parameters of the en ergy storage system. This significantly reduces the burden on the system. For example, Eac and Pavr are both parameters which are responsible for the aging of the system, and both are reduced by about a third.

Figure 6 is a schematic drawing illustrating a control system 100 for controlling a charging state of an energy storage system according to an embodiment. The charging state control can be operated in parallel with the solution of reducing power variability in an energy storage system described above with respect to figures 1-5 and extends the functionality thereof. The above explanations thus apply equally to the control system 100 of figure 6. In smoothing applications, the charging state of the energy storage system 107 varies according to the power required for power variability compensation. Thus, it cannot be ensured that the energy storage system will not be continuously charged or discharged over time. A charging state control is provided that ensures that the charging state is always kept between admissible limits, even during a continuous operation of the smoothing system. In order to optimize the use of en ergy of the energy storage system and minimize the effect on the power smoothing, the implementation of the charging state control is adapted according to the selected smoothing meth od, in particular to the transfer function of filtering unit 106. The solution of the present example adapts the charging state control to the disclosed band-stop filter solution. The charging state control is therefore capable of further reduc ing the burden on the energy storage system and preventing a drift to a fully charged or discharged system state which would be highly demanding for an energy storage system, e.g. batteries. For this purpose, the control system 100 presented in figure 1 and, especially, figure 2 is extended by a charg ing state controlling unit 601.

The charging state controlling unit 601 generates a charging state control signal Pres which may represent a power refer- ence signal and is output to the input of a summation point 612. Another input to the summation point 612 is the control signal Pr generated by control signal generation unit 103, as shown in figure 1. At summation point 612, the charging state control signal Pres is added to the control signal Pr in or- der to adjust the control signal Pr and to generate adjusted control signal Pr*. Instead of control signal Pr, the adjust ed control signal Pr* is then output to the energy storage system as indicated in figure 1. Accordingly, the control signal Pr* may be a power reference signal actuating the en- ergy storage system in such a way that the output compensat ing power flow Pc corresponds to the input control signal Pr*. The charging state control signal Pres of the illustrated em bodiment is generated based on a model. More specifically, it is based on a dynamic model that models the minimum energy variation required by the energy storage system 107 to reduce the variations of the electrical power Ps provided by the en ergy generation system 101. The energy variation corresponds to integration over time of the power Pc output by the energy storage system or the respective control signal Pr thereof. Such a calculation is performed in real time. Specifically, the energy variation AE _V in the energy generation system, if a band-stop filter smoothing solution is applied, can be es timated depending on the compensating electrical power Pc provided by the energy storage system and the reference sig nal thereof, respectively. Reference signal Pr depends on the structure of control signal generation unit 103 and filtering unit 106. Considering the structure of the control signal generation unit and the given transfer functions, AE _V results in:

Accordingly, equation 6 represents the model of the minimum energy variation depending on the electrical power Ps output by the energy generation system and the monitored signal Psm thereof, respectively. The input-output behavior of the mod elled filter 602 corresponds to that of a 2 ^nd order low-pass filter with a gain of K = (t _H — ti) . As an alternative to directing the monitoring signal Psm via the modelled filter 602 to summation point 606, which signal route implements the model of equation 6, it is possible to instantly integrate control signal Pr using integrator and gain element 602a. This is indicated in figure 6 by the dashed arrow directing signal Pr via an integration block to summation point 606. Since available computation results of the power variability reduction control are used by the al- ternative solution, required computing power for computing the output of the model may be reduced. However, the solution outlined above (using the modelled filter 602) has the ad vantage that it is faster if the integrator has to be imple- mented in discrete time.

In each case, using the model of equation 6, the reference signal Ecsr of the charging state can then be calculated by E _Csr E _csmin + ocAEp, (sq. 7), wherein a is a proportional factor and a = 1 is used for standard operation. However, factor a can be set to values deviating from 1 in order to tune the effect of the charging state control on the power variability reduction. The energy necessary for the regulation of the charging state of the en ergy storage system is obtained on the same power path as the power flow for reduction of power variability, so that the charging state control system influences the power flow and thus the reduction of power variability (which is evident from the adding of the control signals at 612). Thus, a value for a below 1 reduces the charging state control reference signal and reduces the energy to be exchanged for charging state regulation and therefore reduces the influence on the power variability reduction. Factor a can therefore be used to determine a compromise between the performance of the charging state control and of the power variability reduc tion. According to equations 6 and 7, figure 6 shows a second order low-pass filter 603, wherein the input thereof receives moni toring signal Psm of the electrical power provided by the en ergy generation system. The output of the second order low- pass filter is provided to the gain unit 604, which comprises a product of tuning factor a and gain K. The output of the gain unit 604 is connected to the input of summation point 606. Another input of summation point 606 is the minimum charging state signal Ecsmin. Ecsmin represents a minimum charging state value ensuring that the charging state of the energy storage system cannot drop below that value for a longer period of time. The output of summation point 606 pro vides the charging state reference signal Ecsr and is con- nected to the input of summation point 610. At summation 610 the input Ecsr is subtracted from another input signal Ecsm that represents the charging state monitoring signal. The charging state monitoring signal Ecsm is provided by a re spective charging state monitoring unit 608. Charging state monitoring signal Ecsm represents the actual value of the charging state of the energy storage system 107 and may be measured, e.g. by a sensor or sensor system or may be ob tained data-driven. The output of summation point 610 is con nected to the input of the charging state controlling filter 611. The charging state controlling filter 611 filters the control error provided by summation point 610. For this pur pose the charging state controlling filter 611 is designed such that the charging state of the energy storage system follows the charging state reference signal. The charging state controlling filter 611 may for example comprise a sca lar gain or may be partial integrative in order to compensate possible steady-state errors of the charging state. The im plementation of the charging state controlling filter 611 may also be model-based itself, can be of linear or non-linear type, and its design is not limited to the given examples.

The output of the charging state controlling filter 611 pro vides a charging state control signal Pres and is connected to the input of summation point 612. Signal Pres may be a power reference signal representing the required power flow to be applied in order to regulate the charging state of the energy storage system in accordance with the reference signal thereof. Hence, the charging state control signal Pres is added to the control signal Pr at summation point 612, which is another input to summation point 612. Accordingly, the output of summation point 612 provides an adjusted reference control signal Pr* to the energy storage system. Pr* can be given as: Herein, k _cs is an example for a gain of charging state con trolling filter 611. However, the charging state controlling filter 611 is not limited to a constant gain. It may also change over time, e.g. by implementing an adaptive control. Further, it may combine a plurality of filters, e.g. such filters comprising integrative or derivative parts. Pr* is a power reference signal that comprises a power reference sig- nal portion for charging state control and a power reference signal portion for power variability reduction.

Figure 7 is a schematic drawing illustrating a further embod iment of a control system 100 configured to control a charg- ing state of an energy storage system. The control system shown in figure 7 corresponds to that of figure 6, so the above explanations apply correspondingly. However, instead of basing the charging state reference signal Ecsr on the output of modelled filter 602, the charging state reference signal Ecsr is set to a constant value and is provided by constant charging state reference signal source 701. Again, it should be clear that the elements of charging state controlling unit 601 can be implemented digitally, as digital circuits or software running on processing unit 113, or as analog cir- cuits.

Figure 8 shows a schematic flow diagram of a method of con trolling a charging state of an energy storage system and of reducing power variability of an energy generation system ac- cording to an embodiment, which may be performed by the con trol system presented in figures 6 or 7. The method 800 com prises steps S302, S303, S802 and S803. In some embodiments, the method may also comprise at least one of steps S301, step S801 and S804. Steps S301-S303 are performed in correspond- ence to the ones of method 300 as shown in figure 3. However, the control signal generated in step S303 is adjusted in step S803 before it is provided to the energy storage system. The adjusted control signal is then received at the energy stor- age system, which applies power in accordance with the ad justed control signal, the power flow being added to the pow er output of the energy generation system (S804). For this purpose, the charging state control signal is generated (S802) based on the monitoring signal that represents the power output of the energy generation system. The generating thereby comprises that the monitoring signal is filtered by a modelled filter, wherein the modelled filter represents at least a model of the charging state of the energy storage system. This filtering may be based on the equations 6-8.

In an embodiment, the charging state control signal is gener ated (S802) additionally based on a second monitoring signal representing the charging state of the energy storage system. Before step S802, the monitoring signals of the charging state of the energy storage system (step S801) and of the power output of the energy generation system (step S302) are obtained and used subsequently to generate the charging state control signal (step S802). The method may also comprise a step of tuning the effect of the charging state controlling unit on the reduction of power variability (not shown). This may be performed by tuning the gain unit 604.

Besides the reduction of the power variability of the energy generation system, the method ensures that the energy storage system is prevented from drifting to a fully charged or dis charged state due to losses occurring while receiving or providing energy. It is therefore no longer necessary to oversize the energy storage system in order to prevent reach- ing these states of the energy storage system.

Figure 9 shows a schematic diagram illustrating the charging state of an energy storage system while applying and not ap plying a charging state control according to the above- described embodiments. Signal 901 represents the ideal state of charge over time without taking losses into account. Sig nal 902 represents the state of charge over time without charging state control and taking losses into account. Signal 903 represents the state of charge over time with charging state control and taking losses into account, and using a non-modelled constant charging state reference signal accord ing to an embodiment. Thus, signal 903 represents the effect of using the charging state control scheme as shown in figure 7. As an example, the charging state reference signal for the system generating signal 903 is set to a state of charge of 50%. Signal 904 represents the state of charge over time with charging state control and taking losses into account, and using a modelled reference signal according to an embodiment. Thus, signal 904 represents the effect of using the charging state control scheme as shown in figure 6.

It can be seen that signal 902 drifts from an initial state of charge of approximately 20% to a fully discharged state of the energy storage system due to non-compensated losses. Sig nal 903 shows a state of charge being increased until it reaches the reference state of charge. Since the state of charge is kept to the constant charging state reference sig- nal, it is ensured that the energy storage system cannot de stabilize and reach a fully charged or discharged state. Sig nal 904 shows a variation of the state of charge that depends on the power provided by the energy generation system. The behavior is close to the ideal case despite a consideration of the losses. Thereby, the configurable offset defines a minimum charging state. In this case, the minimum charging state is set to 0%. A different value may be used, for in stance due to manufacturers recommendations. Figure 9 also illustrates that the charging state control of curve 904 al- lows a further reduction of the power reserves of the energy storage system.

The effect of the described charging state controls according to figures 6 and 7 on the energy storage system is analyzed with respect to the parameters shown in table 4. The solution using a model-based charging state reference signal is capa ble of further reducing the required maximum energy compared to the solution using a constant charging state reference signal.

Table 2: Evaluation of required energy of model-based (mb) and non-model-based (nmb) charging state control

The maximum power and accumulated energy are also further re duced for the disclosed solution of a model-based charging state control compared to the non-model-based charging state control. The reduced stresses thereof positively affect the energy storage system over the scope of months or years of operation. These reduced stress factors contribute signifi cantly to slowing down the aging process of the energy stor age system. Further, the effect of the state of charge regu- lation on the power variability reduction can be minimized. Using the CDF for 1 second variations, 1 minute variations and 30 minutes variations as outlined above, the results of the evaluation for the application of a model-based and non model-based charging state control are shown in table 5. The results of the model-based solution are closer to the ideal line than those of the non-model-based solution. Hence, the state of charge control influences the reduction of the power variability (smoothing operation) less when the model-based charging state control is used than when the non-model-based charging control is used. Table 5 shows that the values of the CDFs of the ideal (lossless) system, i.e. table line 'Ps(BSF, ideal)', are close to the values of the CDFs of the system using the model-based charging control, i.e. table line ' Ps(BSF, mb)'. Thus, it can be concluded that the re- duction of the power variability (smoothing operation) can be performed in the case of the model-based charging control al most with the same efficiency as in the ideal case and still ensures that the state of charge cannot drift.

Table 5

Considering table 5, it can further be seen that the proposed solution of using a model-based charging state control en sures a stable SOC operation while minimizing the effect on smoothing operations and uses approximately the energy re quired in the ideal case. Accordingly, the burden on the en ergy storage system is reduced significantly, which results in reduced power requirements and therefore, a more compact system design is realizable.

The control system 100 may be provided as a separate stand alone control system, or it may for example be integrated in- to a power plant control system (e.g. wind farm controller) that operates the energy generation system 100 (e.g. a wind farm). The control system 100 or parts thereof, e.g. the charging state controlling unit 601, may also be implemented in the energy storage system 107. As such, the control system 100 can be a central control system or a distributed control system that is spread across the power plant, and the parts of which communicate by the respective signals.

While specific embodiments are disclosed herein, various changes and modifications can be made without departing from the scope of the invention. The present embodiments are to be considered in all respects as illustrative and non- restrictive, and all changes coming within the meaning and equivalency range of the appended claims are intended to be embraced therein.