Method for operating a wind farm system

Abstract

The invention relates to a method for operating a wind farm system (310), in particular, an offshore wind farm system (310), wherein the wind farm system (310) is electrically decoupled from an external grid, and wherein the wind farm system (310) comprises at least one wind turbine (312) connected with a central electrical bus (322, 422) at least in a production operating mode of the wind farm system (310), at least one generator (342) connected with the central electrical bus (322, 422), at least one load (326) connected with the central electrical bus (322, 422) at least in the production operating mode, and at least one central controller (100, 300, 400), the method comprising supplying, by the at least one generator (342), power to the central electrical bus (322, 422) thereby providing at least a voltage reference and a frequency reference for the at least one connected wind turbine (312), determining, by the central controller (100, 300, 400), the voltage and the frequency at the central electrical bus (322, 422), determining, by the central controller (100, 300, 400), at least one turbine setpoint data set and at least one load setpoint data set depending at least on the determined voltage and the determined frequency, controlling the at least one connected wind turbine (312) according to the determined turbine setpoint data set, and controlling the at least one load (326) according to the determined load setpoint data set.

Description

[0001]3. April 2024 Method for operating a wind farm system The invention relates to a method for operating, in particular, controlling a wind farm system, in particular an islanded wind farm system. The invention also relates to a central controller, a wind farm system, and a computer program. In the present time, electrical power generation systems are increasingly used for the provision of electrical energy respectively power, in which the generating of electrical energy is based on so-called renewable energy sources. Electric power generation systems generally have at least one power generation device, preferably a plurality of power generation devices. More particularly, wind energy systems and wind farms, respectively, comprising at least one wind turbine as an energy generation device are used more and more as electrical energy generation systems. In particular, a wind turbine is configured to convert the kinetic wind energy into electrical energy. Such wind farms are not only located at onshore sites, but increasingly also at offshore sites. There are many reasons for choosing an offshore site instead of an onshore site: for example, the available space onshore may be limited. In addition, it has been shown that the energy yield can be increased at offshore wind farms, for example. Offshore locations are usually characterized by relatively continuous wind conditions and high average wind speeds (compared to onshore sites), so that offshore wind farms are increasingly being built. In prior art, an offshore wind farm may comprise a plurality of (stationary) offshore structures (i.e., no ships or the like), such as a plurality of offshore wind turbines and at least one offshore substation (also called converter station) by which an offshore wind farm can be electrically connected, for example, to an onshore substation or a further offshore substation. In particular, an offshore wind farm has an internal grid (e.g., comprising power cables respectively energy cables in the form of submarine power cables) configured to electrically connect at least two offshore structures of a wind farm. As described, in prior art, offshore wind farms usually comprise an onshore substation, which may be connected to an external grid, such as public power grid respectively an electrical distribution network. As described, the known offshore wind farms are usually electrically coupled to an external grid in order to supply the power generated by the one or more wind turbines into said external grid. However, more and more wind farms are being built to (exclusively) supply a specific load with power. Preferably, the at least one (specific) load is a hydrogen production plant. However, remote industrial or commercial installations may also be possible loads. A hydrogen production plant comprises at least one electrolyzer. An electrolyzer is configured to produce respectively generate hydrogen. A wind farm coupled with at least one specific load may be denoted as a wind farm system. Such a wind farm system may be preferably operated permanently in a macrogrid-disconnected mode otherwise known as power island mode. In other words, the wind farm system, in particular, the offshore wind farm system is (permanently) electrically decoupled from any external grid. However, as the operation of such a system is associated with problems in practice, a connection to an external grid is often established in the prior art nevertheless. A main issue is that the installed wind turbines are usually grid-following wind turbines. In particular, wind turbines can be equipped with full power converters which are grid-following. This means said wind turbines are not capable of external grid-independent parallel operations. In particular, grid-following converters utilize phase-locked loops (PLLs). A PLL is a closed-loop negative-feedback control system that maintains the phases of voltage and current signals in a well-defined phase relation to achieve synchronism to a voltage source. In the absence of an external HB / HB 230974WO 3. April 2024 electric grid respectively macrogrid, however, the PLLs lose a voltage and current reference needed to synthesize its frequency. Document US 2022 / 0294229 A1 describes a method to operate a stand-alone internal electrical network of a wind farm using a stabilizing controller and an impedance between an output of the stabilizing controller and the stand-alone internal electrical network node. The scope of this document is limited to the application of a stabilization controller. However, controllers have their stability limits, thus, limited stability performance of the stand-alone internal electrical network, in particular, in the event of faults. In addition, US 2022 / 0294229 A1 relates only to black start or power support from a local wind farm grid during commissioning. However, a normal energy production operating mode is not described in said document. Furthermore, document EP 2503146 B1 discloses a method for controlling an operation of a wind farm when disconnected from an external grid due to fault such that the auxiliary loads of the wind turbines may be continually supplied with electrical power. The scope of this document covers only temporary operation of the wind farm if disconnected from the grid due to faults and when they do not need to operate in normal energy production operating mode. The temporal operation only seeks to enable continuous supply of needed power to the auxiliaries of the wind turbines. Therefore, it is an object of the present invention to provide a possibility to reduce the drawbacks of the prior art and, in particular, to enable a safe and reliable operation of a wind farm system (permanently) decoupled from an external grid, in particular, in the normal energy production operating mode. The object is solved according to a first aspect of the invention by a (computer implemented) method according to claim 1 for operating a wind farm system, in particular, an offshore wind farm system, wherein the wind farm system is electrically decoupled from an external grid, and wherein the wind farm system comprises at HB / HB 230974WO 3. April 2024 least one (grid following) wind turbine connected with a central electrical bus at least in a (normal energy) production operating mode of the wind farm system, at least one (rotating power and / or synchronous) generator (driven, in particular, by a gas turbine) connected with the central electrical bus, at least one (specific) load connected with the central electrical bus at least in the production operating mode, and at least one central controller. The method comprises: - supplying, by the at least one generator, power to the central electrical bus thereby providing at least a voltage reference and a frequency reference for the at least one connected wind turbine, - determining, by the central controller, the voltage and the frequency at the central electrical bus, - determining, by the central controller, at least one turbine setpoint data set and at least one load setpoint data set depending at least on the determined voltage and the determined frequency, - controlling the at least one connected wind turbine according to the determined turbine setpoint data set, and - controlling the at least one load according to the determined load setpoint data set. A further aspect of the invention is a central controller according to claim 18 for a wind farm system, in particular, an offshore wind farm system, wherein the wind farm system is electrically decoupled from an external grid, and wherein the wind farm system comprises at least one (grid following) wind turbine connected with a central electrical bus at least in a production operating mode of the wind farm system, at least one (rotating power and / or synchronous) generator (driven , in particular, by a gas turbine) connected with the central electrical bus, at least one (specific) load connected with the central electrical bus at least in a (normal energy) production operating mode. The central controller is configured to: - control the at least one generator such that the at least one generator supplies power to the central electrical bus thereby providing at least a voltage HB / HB 230974WO 3. April 2024 reference and a frequency reference for the at least one connected wind turbine, - determine the voltage and the frequency at the central electrical bus, - determine at least one turbine setpoint data set and at least one load setpoint data set depending at least on the determined voltage and the determined frequency, - control the at least one connected wind turbine according to the determined turbine setpoint data set, and - control the at least one load according to the determined load setpoint data set. In contrast to the prior art, according to the invention, by providing a method (and a central controller) for controlling the one or more wind turbines and the one or more (specific) loads of an islanded wind farm system, wherein the wind turbine(s) and the load(s) are connected with a common central electrical bus, and wherein a (permanently operated) generator provides a voltage reference and a frequency reference to said central electrical bus such that load setpoint data sets and turbine setpoint data sets can be determined based on the voltage and the frequency at the central electrical bus and can be used for controlling said at least one wind turbine and said at least one load, a possibility to reduce the drawbacks of the prior art and, in particular, to enable a safe and reliable operation of a wind farm system (permanently) decoupled from an external grid, in particular, in the normal energy production operating mode. More particularly, the present invention provides a method for synchronizing (grid- following) wind turbines on a central electrical bus which is decoupled from any external electric grid, wherein the wind turbines produce (normal) energy respectively power for a continuous supply of power to at least one connected electrical load to form a power islanded wind farm system. HB / HB 230974WO 3. April 2024 The (computer implemented) method serves, in particular, to synchronize grid following wind turbines on a central electrical bus which is decoupled from an external grid. A wind farm system decoupled from external grid is, in particular, operated in grid- disconnected mode respectively power island mode. Such a system can also be denoted as an islanded wind farm system. Preferably, the wind farm system is an offshore wind farm system. In other words, the structures of the wind farm system are installed at an offshore site. In particular, (floatable or non-floatable) wind turbine(s), (floatable or non-floatable) substation(s), (floatable or non-floatable) load(s) and the like can be provided. The wind farm system may comprise a wind farm and at least one specific load suppliable with power (only) by the wind farm. The wind farm system, in particular, the wind farm comprises at least one wind turbine. According to a preferred embodiment, the wind farm system, in particular, the wind farm comprises a plurality of (offshore) wind turbines. The at least one wind turbine is configured to convert the kinetic wind energy in electrical energy. The at least one wind turbine may be an onshore wind turbine or, preferably, an offshore wind turbine. Each wind turbine of the wind farm system is, in particular, a grid following wind turbine. In particular, each wind turbine is equipped with a partial power converter or a full power converter. The one or more wind turbines are connectable with a (common) central electrical bus. In the production operating mode, all wind turbines are connected with the central electrical bus. Preferably, the (exact one) central electrical bus may be arranged on a substation. HB / HB 230974WO 3. April 2024 According to a preferred embodiment, the wind farm system, in particular, the wind farm may comprise the substation having at least the central electrical bus. The central electrical bus is, in particular, configured as an (common) interconnecting point for at least the wind turbine(s), the generator and the at least one load of the system. According to a preferred embodiment, the central electrical bus is a MV (medium voltage) electrical bus. According to a preferred embodiment, the generator of the wind farm system is a synchronous generator or an induction generator (doubly-fed or squirrel cage). In particular, the generator is driven by a gas turbine. The gas turbine may be a hydrogen-based gas turbine. By way of example, the generator and, in particular, the gas turbine may be arranged on the substation or a separate structure of the wind farm system. In particular, the generator is configured to supply power to the electrical bus at all times. Thereby, a voltage reference and a frequency reference and, in particular, a phase reference may be provided (at all times respectively during all operating modes). Further, the wind farm system comprises at least one (specific) load. This means, in particular, that the wind farm with its one or more wind turbine essentially serves to solely supply said specific load with electrical power and, in particular, auxiliary loads of the wind farm system. According to a preferred embodiment, the (specific) load may be a hydrogen production plant. A hydrogen production plant comprises at least one electrolyzer. The at least one electrolyzer may be configured to produce respectively generate a fluid medium, in particular, in form of hydrogen and / or ammonia. The hydrogen production plant may comprise at least one hydrogen power supply for the at least one electrolyzer. Furthermore, the hydrogen production plant may comprise at least one (hydrogen) processing module respectively (hydrogen) HB / HB 230974WO 3. April 2024 finishing module. The at least one processing module may be configured to process (wet) hydrogen produced by the at least one electrolyzer. In the present case, wet hydrogen refers in particular to hydrogen saturated with water respectively a two-phase mixture of gaseous hydrogen and liquid water. In the present case, dry hydrogen refers in particular to hydrogen that is not saturated with water, i.e., whose dew temperature is (significantly) below the actual temperature at the pressure present. In other words, in particular, wet hydrogen gas can be produced by the at least one electrolyzer which, when cooled, becomes a gas-liquid mixture in the form of a water-hydrogen mixture. The gas respectively gaseous phase carries the liquid phase with it. The at least one hydrogen processing module can comprise at least one hydrogen drying module. The hydrogen drying module can be configured to dry the produced wet hydrogen. Preferably, the hydrogen drying module (e.g., after separation of the liquid phase) can be an adsorption drying module (in particular, a TSA (temperature swing adsorption) module) and / or a refrigeration drying module. The adsorption drying module may be in particular configured to dry the wet hydrogen by adsorption using a (suitable) adsorber, in particular silica beads. The at least one hydrogen drying module can be configured to cool the wet hydrogen to a temperature of at least less than 5°C. It shall be understood that in variants of the invention, other hydrogen drying modules can be used alternatively or additionally. Furthermore, the at least one hydrogen processing module can comprise a catalytic deoxo stage, in particular, for treating the wet hydrogen before drying as described above. Such a stage serves to remove oxygen (≤0.5%). In variants of the invention, such treatment can also take place at a later stage, for example, at a hydrogen destination structure to which the treated hydrogen can be transported, for example, via a hydrogen pipeline network. HB / HB 230974WO 3. April 2024 Preferably additionally, according to a further embodiment of the method according to the invention, the at least one (hydrogen) processing module may comprise at least one (hydrogen) compression module. The compression module may be configured to compress the processed fluid medium, in particular the dried hydrogen. The compression module may comprise at least one (hydrogen) compressor (in particular a (vertical) piston compressor). The compression module can be configured to compress the produced fluid medium, such as the dried hydrogen, to at least 50 barg (and for example at most 250 barg), preferably to essentially 70 barg. Alternatively or additionally, the at least one (hydrogen) processing module may comprise at least one transforming module configured to transform the produced hydrogen to ammonia (e.g., via the Haber-Bosch-method or the like). Furthermore, the at least one hydrogen production system may comprise at least one water treatment module. The at least one water treatment module can be configured to treat the water such that it can be used by the at least one electrolyzer of the hydrogen production plant to produce the fluid medium, such as hydrogen. In particular, the water treatment module can treat seawater for the electrolysis process to be performed. Preferably, the at least one water treatment module can be a seawater desalination module with membrane-based pressure filtration. Such a seawater desalination module can carry out reverse osmosis, ultrafiltration and / or electrodialysis (also known as electrodeionization (EDI)) to treat the water respectively to treat the seawater accordingly. Such treatment can be used to treat seawater, in particular in an energy-efficient manner, which can be used for water electrolysis. The treated water can also be referred to as ultrapure water or "demin water". Alternatively or additionally, the at least one hydrogen production plant may comprise at least one inert gas generation module configured to generate an inert gas. For example, the inert gas generation module may comprise an inert gas storage tank, HB / HB 230974WO 3. April 2024 in particular in the form of a pressurized inert gas storage tank configured to (temporarily) store the generated inert gas and in particular for conveying the inert gas through an inert gas pipeline network. Preferably, a nitrogen generation module can be provided as the inert gas generation module, wherein the nitrogen generation module can be configured to generate nitrogen from air, preferably by pressure swing adsorption. It shall be understood that other inert gas generation modules can also be provided alternatively or additionally in variants of the invention. Further, it shall be understood that the at least one electrolyzer plant may comprise less modules, further modules, or other modules. For instance, the electrolyzer plant may additionally comprise an intake and filtration module. It shall be understood that the load in form of the hydrogen production plant may comprise a plurality of sub- loads in form of the at least one electrolyzer, at least one hydrogen processing device etc. A portion of the produced hydrogen may be forwarded to the gas turbine and / or via a gas storage tank to the gas turbine. In particular, the gas turbine, and thus, the generator, can be operated by the produced hydrogen or ammonia. The wind farm system comprises at least one central controller. The central controller may be, in particular, the wind farm controller of the wind farm of the wind farm system respectively may be implemented in the wind farm controller. The central controller determines (e.g., by measuring) the voltage (reference) and the frequency (reference) and, in particular the phase reference, at the central electrical bus. Based on said determined electrical values, at least one turbine setpoint data set for the at least one wind turbine and at least one load setpoint data set for the at least one load is determined. A setpoint data set may comprise one or more electrical setpoints to be set. HB / HB 230974WO 3. April 2024 Determining at least one turbine setpoint data set and at least one load setpoint data set means, in particular, that a harmonizing is performed, via the central controller, of the active and reactive power levels of each of the plurality of the connected wind turbines with the at least one connected electrical load at the interconnecting point respectively the central electrical bus. Determining at least one turbine setpoint data set and at least one load setpoint data may depend on the power demand of the at least one load. The determining is performed such that a stable internal electrical network is provided. The at least one connected wind turbine is controlled respectively operated according to the determined turbine setpoint data set. The at least one connected load is controlled respectively operated according to the determined load setpoint data set. According to a further embodiment of the method according to the invention, controlling the at least one connected wind turbine according to the determined turbine setpoint data set may comprise transmitting, by the central controller, the turbine setpoint data set to a wind turbine controller of the at least one wind turbine. The wind turbine controller may control the wind turbine in accordance with the received turbine setpoint data set. The wind turbine may comprise a PLL. Alternatively or additionally, controlling the at least one connected load according to the determined load setpoint data set comprises transmitting, by the central controller, the load setpoint data set to a load controller of the at least one load. The load controller may control the load in accordance with the received load setpoint data set. In particular the load controller may control supplying of power to the load(s). As already described, according to an embodiment of the method according to the invention, the at least one load may be a specific load. The specific load comprises, in particular, a (previously described) hydrogen production plant. It shall be understood HB / HB 230974WO 3. April 2024 that the specific load may comprise other remote industrial or commercial installations, such as a server installation. According to a further embodiment of the method according to the invention, the at least one load may comprise at least one auxiliary load of the wind farm system. In particular, there may be a plurality of auxiliary loads besides the specific load. Exemplified and non-exhaustive examples of auxiliary loads are hydraulics, pumps, heaters etc. for lubrication, cooling, climate conditioning and so on. In the production operating mode of the wind turbine, at least more than 90 % of the power produced by the wind turbine(s) is fed to the specific load, in particular, the hydrogen production plant. Furthermore, according to a preferred embodiment of the method according to the invention, a respective turbine setpoint data set may be determined for each connected wind turbine. In other words, for each wind turbine actually connected with the central electrical bus, a specific (separate) turbine setpoint data set may be determined. Each connected wind turbine may be controlled according to the determined respective turbine setpoint data set. Alternatively or additionally, a respective load setpoint data set may be determined for each connected load. In other words, for each load actually connected with the central electrical bus, a specific (separate) load setpoint data set may be determined. Each connected load may be controlled according to the determined respective load setpoint data set. According to a further preferred embodiment of the method according to the invention, the at least one turbine setpoint data set may comprise a turbine active power setpoint and a (corresponding) turbine reactive power setpoint. In particular, for each connected wind turbine, a turbine active power setpoint and a (corresponding) turbine reactive power setpoint may be determined. HB / HB 230974WO 3. April 2024 Alternatively or additionally, the at least one load setpoint data set may comprise a load active power setpoint and a (corresponding) load reactive power setpoint. In particular, for each connected load respectively sub-load, a load active power setpoint and a (corresponding) load reactive power setpoint may be determined. In order to provide a particularly reliable controlling method, according to a preferred embodiment of the method according to the invention, the at least one turbine setpoint data set may be determined based on a linear quadratic control scheme (respectively linear quadratic regulation scheme). The at least one load setpoint data set may be determined based on the linear quadratic control scheme (implemented in the central controller). In particular, the respective active and corresponding reactive power of the one and more connected wind turbines and the at least one load may be controlled using a linear quadratic control (LQC) scheme respectively linear quadratic regulation scheme (LQR) implemented in the central controller. According to a particular preferred embodiment of the method according to the invention, the linear quadratic control scheme may be based on a discrete state space representing the internal electrical network of the wind farm system. It has been recognized that by modelling the internal electrical network in respectively as a discrete state space, the controlling can be improved. The internal electrical network can be divided in different electrical subsystems. In particular, the different subsystems of the internal electrical network of the wind farm system can be modelled in a discrete state space. In particular, it has been recognized that a model according to the Fourier’s theory used in prior art is not suitable for a sufficiently reliable controlling of a wind farm system decoupled from any external grids. In particular, power generation and demand may vary continuously in an islanded wind farm system according to the invention. The controlling of such a system is difficult due to this unpredictability. In order to reflect these difficulties and for enabling a reliable controlling, according to HB / HB 230974WO 3. April 2024 preferred embodiments of the invention, it is proposed to apply discrete-time modelling. According to a further embodiment of the method according to the invention, a first transfer function, G(s), may represent the at least one wind turbine and the generator of the wind farm system. In particular, the energy producing units may form a first subsystem of the internal electrical network of the wind farm system. The first subsystem may be formed by the at least one wind turbine (preferably, all connected wind turbines) and the generator. The first transfer function may represent the first subsystem. A second transfer function, D(s), may represent the at least one load. In particular, the load(s) may form a second subsystem of the internal electrical network of the wind farm system. The second subsystem may be formed by the at least one load (preferably all loads). The second transfer function may represent the second subsystem. In particular, at least said first subsystem connected with the central electrical bus and said second subsystem connected with the central electrical bus can be modelled in a discrete state space by respective transfer functions. The central controller may function as the grid manager utilizing the aforementioned modelled state space in the linear quadratic regulation respectively control scheme. According to a further preferred embodiment of the method according to the invention, the wind farm system may comprise at least one static reactive power equipment connectable to the central electrical bus. The static reactive power equipment may comprise a Static Synchronous Compensator (STATCOM) and / or a static VAR (volt-ampere reactive) compensator (SVC) and / or the like. Preferably, a third transfer function, W(s), may represent the at least one static reactive power equipment. In particular, the piece(s) of static reactive power equipment may form a third subsystem of the internal electrical network of the wind HB / HB 230974WO 3. April 2024 farm system. The third subsystem may be formed by the at least one static reactive power equipment (preferably all loads). The third transfer function may represent the third subsystem. In particular, at least said first subsystem connected with the central electrical bus and said second subsystem connected with the central electrical bus and said third subsystem connected with the central electrical bus can be modelled in a discrete state space by respective transfer functions. According to a further embodiment of the method according to the invention, the output of the discrete state space may represent a stable state of the internal electrical network (comprising the at least two subsystems, preferably the at least three subsystems) of the wind farm system. By determining the output, the controlling of the respective subsystems with respective setpoint data sets can be performed by the controller. A particular reliable controlling can be provided. Exemplified requirements of an internal electrical network of a wind farm system are described in subsequent Table 1. Parameters Grid voltage 66 kV (±10%) (or any medium or high voltage level) Grid frequency 50 or 60 Hz (±5%) Harmonic condition standard limits, no other specification applies Generated power of grid voltage ≥ total internal grid consumption at idling reference (idle mode / initial state) Generated power of grid voltage < total internal grid consumption at idle reference (post-initial) mode (induction generator (doubly-fed or squirrel cage) Short-circuit ratio see eq. (26) Short-time grid overload (10s) load control, power control of the power generating components HB / HB 230974WO 3. April 2024 Settling time of max active power set ≤ 1 min point Tolerance band of active power settling ± 5% time Settling time of max reactive power set ≤ 1 min point Power limitations due to over frequency load control, power control of wind turbines Power limitations due to under load control, power control of wind frequency turbines Reactive power (idle mode) 0.9 pf at the wind farm kW or MW capacity Reactive power (post-initial) 0.9 pf at the wind farm kW or MW capacity Low voltage ride through ≥ 0.7pu < 0.9pu (time: 0-0.2s); Communication Explicit between the subsystems and central controller respectively internal grid manager Table 1 It shall be understood that other requirements may be set for an internal electric network of the wind farm system and that the following explanations can also be apply for other requirements. As has already been described, power generation and demand may vary continuously in a wind farm system according to the present invention. Since it has been found that the control of the wind farm system is difficult by using a model based on Fourier’s theory, it is proposed to apply, according to a preferred embodiment, discrete-time modelling. The three phases of a discrete 3-phase power system can be expressed according to equation (1), wherein Vabcare the voltage amplitudes, φabctheir corresponding phase angles, ω is the angular speed, k is the iteration number and T is the sampling period. va(k) = Va sin(ωkT + φa)HB / HB 230974WO 3. April 2024vb(k) = Vb sin(ωkT − φb) (1)vc(k) = Vc sin(ωkT + φc)Equation (1) is, in particular, valid for unequal respectively unbalanced Vabcas well astheir corresponding φabc and frequency f. The frequency f can be obtained from ω =2πf. The discrete 3-phase power system in equation (1) can further be expressed as a unique sum of symmetrical components, as exemplified in equation (2). Here, subscript “+” represents the positive sequence, the negative components are denoted with “−”and the zero components with “0”. Preferably, for modelling the internal electrical network, a subsequent conversion of equation (2) can be performed into a corresponding complex expression of positive, negative and zero sequence respectively using the Clarke 2-dimensional αβ-reference frame. Such a conversion results in equations (3) - (5), whereine−jωkT = cos(wkT) + jsin(wkT), ∀ the phasor matrix of the sequence componentsequals and ∝ is a phasor rotation operator. v+ (k) = v+ + α(k) + jv β(k) = ∀+e+jωkT(3) v− (k) = v− α(k) + jv− β(k) = ∀−e−jωkT(4) v0 (k) = e−jωkT(5) With the Clarke transformation, the resulting complex expression is, in particular, the sum of complex voltages corresponding to the zero, positive and negative components as given in following equation (6). HB / HB 230974WO 3. April 2024 Given the expected dynamics of power system (in particular, the internal grid of the island wind farm), preferably, the standard discrete state-space form in equations (7) and (8) can be used to describe it with the matrices in equations (9) - (11). Hereby, xis the state, y the output, A is the dynamics matrix (A ∈ ℝ3×3), B is the input matrix(B ∈ ℝ3×2), C is the output matrix (C ∈ ℝ2×3), D the feedthrough matrix is a nullmatrix. x(k + 1) = A(k)x(k) + B(k)u(k) (7)y(k) = C(k)x(k) + D(k)u(k) (8)A11A12A13A = [A21A22A23] (9) A31A32A33B11B12B = [B21B22] (10) B31B32C = [C11C12C13C21C22C23] (11) The state of the wind farm system is described in subsequent equation (12) by the three parameters ẋ0, ẋ1& ẋ2. These parameters correspond to the discrete-time states of the zero, positive and negative sequence respectively in equation (6). By deriving ẋ0, ẋ1& ẋ2as functions of equation (6), the system states in equations (7) and (8) can be determined by iterative method using the matrices given in equations (9) to (11) to obtain the system output y. As described above, the system output may describe the stable power system network respectively may represent a stable state of the internal electrical network. HB / HB 230974WO 3. April 2024 The state feedback control may be implemented in the central controller of the wind farm system. Preferably, the wind farm system to be controlled may be a system of acentral electrical bus (preferably a medium voltage bus) having G(s), W(s), D(s),wherein, as described hereinbefore, G(s)is the transfer function of the first subsystem, in particular, the power generating subsystem which comprises the (synchronous) generators and the one or more wind turbines in the internal electrical network, W(s) is the transfer function of the optional third subsystem comprising the at least one static (reactive) power equipment, and ^^(^^) is the transfer function of the second subsystem comprising the at least one interfacing electrical load. It shall be understood that, since the wind farm system respectively the internal electrical network comprises different subsystems, the input and output control signals dimension are in particular multiple, thus leading to a multiple-input, multiple- output (MIMO) system. In this way, at any point in time, the stability of the wind farm system comprising the aforementioned subsystems can be ensured via the feedback of the same states in equation (12) in the control loop respectively by means of the central controller. x= [ẋ0 ẋ1 ẋ2] (12)As has been already described, the central controller may function as the grid manager. Preferably, the central controller may be configured to utilize the aforementioned modelled state space in a linear quadratic regulation respectively control LQR / LQC scheme. It has been recognized that the LQR / LQC scheme enables an incorporation of cross dependencies between frequency and power as well as reactive power and voltage, and generally, the exploitation of the system knowledge in the controller design. In addition, since model certainty can be assured, utilizing LQC facilitates, in particular, the application of first principles-based state-space model. The state HB / HB 230974WO 3. April 2024 feedback control implemented by the central controller may use a matrix Kito calculate the control signal from the system state x and achieve system stability accuracy by multiplying a command signal w with a pre-factor Kvas represented by equation (13), wherein Kvw accommodates the use of a large-signal state estimator. u= −Kix + Kvw for all operating points i ∈ {1, … , nop} (13)The LQC may be preferably described by equation (14), wherein the states x are penalized with the weighting matrix Q and the control effort with the weight of matrix R in equations (15) and (16) respectively. The selection of the matrices Q and R is, in particular, performed according to Bryson’s rule which, in particular, corresponds to the criterion in equation (17), 1 wherein Qii=maxacceptable value of xi2, whereby i ∈ {1,2, … , l} and where Rjj =maxacceptable value of uj2, whereby j ∈ {1,2, … , k} and ρ is a positive constant. According to a preferred embodiment of the method according to the invention, simulations according to Bryson’s rule may be applied (e.g., by the central controller) to determine the values of Q and R wherein the first, second and optional third subsystems achieved fast reaction to the needed setpoint changes p and q in the input HB / HB 230974WO 3. April 2024 signal u according to equation (18), wherein p and q are active and reactive power setpoints respectively (forming the turbine setpoint data sets respectively load setpoint data sets respectively static reactive power equipment setpoint data sets) for the participating subsystems in the MIMO scheme. u= [p q] (18)After tuning the matrices according to the desired output matrix, the Q and R matrices can be used to solve the Algebraic Riccati Equation (ARE) to compute, e.g., by the central controller respectively a processor of the central controller, the full state feedback matrix, which is essentially the LQR control. For an even faster response of the central controller, according to a further embodiment of the method according to the present invention, the Q matrix values can be changed (preferably by the central controller), whereas for minimizing stability error while achieving the desired setpoints without focusing on the time of response, the R matrix values is in particular adjusted. The values of the matrices to achieve the particularly preferred control goal – fast response for needed level of grid stability have been tuned for A, B, C, Q and R in equations (19)-(23) respectively, by the central controller. HB / HB 230974WO 3. April 20241 0 0Q = [ 0 10 0] (22) 00 10R = [1 00 1] (23) Thus, the chosen weighting matrix Q can penalise only the zero sequence ẋ0. The values may be just the starting point to a trial-and-error iterative design procedure aimed at obtaining desirable properties for the closed-loop system, as described hereinbefore. According to a further preferred embodiment of the method according to the invention, the method may further comprise: - determining the network impedance of the internal electric network of the wind farm system in real time based on an equivalent circuit model of the internal electric network of the wind farm system thereby determining corresponding operating active and reactive power injection into the internal electric network and / or active and reactive power withdrawal from the internal electric network. In other words, the network impedance which may be required for the particularly exact (respectively correct) (previously described) setpoint data sets during operation of the wind farm system may be determined by the central controller. The electrical properties of the central electrical bus, preferably a medium voltage bus, of the internal grid network may not only be characterised by the voltage, its frequency and phase angle but also by its internal impedance respectively network impedance. It has been recognized that for a particularly reliable controlling of the wind farm system respectively its subsystems, the value of minimum impedance at the nominal network frequency may be crucial to synchronism and stability since it determines the maximum possible short-circuit. HB / HB 230974WO 3. April 2024 In particular, a model of the equivalent circuit diagram of the internal electricalnetwork can be determined (e.g., by the central controller), wherein IWF(rated) < φi isthe vector of the rated nominal current of the wind farm system. Likewise, the (synchronous) generator of the wind farm system may provide the vector of theinternal grid voltage, Vg < φv in the formed internal electrical network.In an production state of the wind farm system (subsequently described in more details), the loads connected at the central electrical bus may be preferably limited by the auxiliary consumptions of the wind farm system. It has been found that the approximate amount of the internal impedance of the network to ensure power transfer to the loads may result from the relationships in equations (24) - (26), wherein is the equivalent longitudinal impedance of the connecting inter-array cable lines respectively strings of the participating wind turbine, Zgis the equivalent impedance of the interfacing generator transformer and the synchronous generator and SCR is the short-circuit ratio. Zn = Zg + Zl (24)Zn = Rg + Rl + j(Xg + Xl) (25) There may be no equivalent longitudinal impedance of an upstream grid network since the present wind farm system forms an islanded network, thus, resulting in a high impedance via the, e.g., medium voltage and apparent power levels provided by the synchronous generator and the interfacing generator transformer. Nonetheless, this may ensure full power flow if the network impedance is lowered at minimum possible Zd, the impedance of the connected external loads via means of load regulation. The regulation respectively control of the at least one load may ensure a HB / HB 230974WO 3. April 2024 particularly optimal impedance for power transfer and contributes to the internal electrical network stability alongside the contributions of the power generating units of the wind farm system. In equation (26), the maximum terminal short-circuit (un-faulted) Idtoccurs at the point of minimum possible Znto ensure power transfer when the maximum loads are switched on. In particular, the sizing of the static reactive power equipment for voltage regulation may be optimisable. In addition, the terminal short-circuit at any point in time may be controlled via a combination of sequential synchronisation and load regulation to achieve internal electrical network stability. Using the previously described equations (24) - (26), the central controller can be configured to determine the impedance in real time and the corresponding operating active and reactive power injection or their withdrawal from the internal electrical network. In this way, the ability of the wind farm system by means of its (coordinating) central controller to regulate the voltage at the central electrical bus (e.g., a medium voltage bus) can be ensured. By way of example, the behaviour of the internal electrical network under the different states is given in Table 2. #State at PCC (point of common coupling, as^^ ^^^^^^ shown in Figs.6a and 6b) (p.u) (p.u) 1closed-circuit1 Idok~ Zl(idle mode, unsynchronised) HB / HB 230974WO 3. April 20242 min. short-circuit 1 < 1 Zn ∥ Zdmax(post-initial, synchronised) [see sequence “# ^1” = “sync WTG-1” in table 4] 3max. short-circuit 1 1 Zn ∥ Zdmin(post-initial, synchronised) [see sequence “# ^n_2” = “sync WTG-n” in table 4] 4short-circuit (faulty) ≪ 1 ≫ 1 ≪ (Zn ∥(post-initial, synchronised) Zdmin) Table 2 Under short-circuit fault conditions, Znmay drop below the minimum short-circuit impedance alongside the central electrical bus voltage as given by #4 in Table 2. Voltage regulation, e.g., by the central controller, in this state may be provided by the support of the static power equipment which is controlled in response to the prevailing central electrical bus voltage condition and needed compensation. The voltage limit for the operation of the wind farm system under low voltage conditions is given in Table 1, otherwise, the faulted subsystems triggering the voltage drop may be disconnected and isolated from the remaining system for stability and continued operation of the internal electrical network. According to a further embodiment of the method according to the invention, the wind farm system may be at least operable in the production operating mode and in an idling operating mode. The (normal energy) production operating mode of the wind farm system means, in particular, a state of the wind farm system in which all (available) wind turbines of the wind farm system are operating and supplying HB / HB 230974WO 3. April 2024 electrical energy respectively power to the specific load, e.g., the at hydrogen production plant, via the central electrical bus. The idling operating mode of the wind farm system means, in particular, a state of the wind farm system in which no wind turbine is operating and supplying electrical energy respectively power to the central electrical bus. In the idling operating mode solely the generator may supply power to the central electrical bus. In particular, in the idling operating mode of the wind farm system solely the at least one auxiliary load withdraws power from the central electrical bus (provided by the generator). According to a preferred embodiment of the method according to the invention, in the idling operating mode the generator may supply power of at least approximately 0.87 % of the total apparent power capacity of the installed wind turbines of the wind farm system. In particular, it has been recognized that during complete idling respectively in the idling operating mode (i.e., no power production from any of the wind turbines), the total auxiliary loads may need approximately 0.87 % of the total installed apparent power capacity of the wind farm of the wind farm system. The generator can be dimensioned such that this requirement can be met. According to a further embodiment of the method according to the invention, the wind farm system may be at least operable in the production operating mode and in an initializing operating mode. The initializing operating mode of the wind farm system means, in particular, a state of the wind farm system during which the one or more wind turbine are switched to the central electrical bus such that power is supplied by said wind turbines to the central electrical bus. In particular, after all available wind turbines are switched to the central electrical bus, the operation mode of the wind farm system can be switched from the initializing operating mode to the production operating mode. In the initializing operating mode, the one or more wind turbines may be switched to the central electrical bus according to a predefined regulation sequence scheme, in HB / HB 230974WO 3. April 2024 particular, under control of the central controller. In particular, the predefined regulation sequence scheme can be stored in a data storage of the central controller. The regulation sequence scheme can define the order and the way one or more of the wind turbines are switched to the central electrical bus. According to a further embodiment of the method according to the invention, according to the predefined regulation sequence scheme, in particular, after an idling operating mode, a first wind turbine of a plurality of wind turbines may be switched to the central electrical bus in a first synchronization step. According to the predefined regulation sequence scheme, the internal electric network may be stabilized in a first stabilization step following the first synchronization step. According to the predefined regulation sequence scheme, a further wind turbine of the plurality of wind turbines is switched to the central electrical bus in a further synchronization step. According to the predefined regulation sequence scheme, the internal electric network is stabilized in a further stabilization step following the further synchronization step. The further synchronization step and the further stabilization step can be repeated until all available wind turbines of the wind farm system are connected with the central electrical bus. According to a further preferred embodiment of the method according to the invention, in the production operating mode, the power generated by the generator connected with the central electrical bus may depend on the number of the wind turbines of the plurality of wind turbines connected with the central electrical bus. In particular, the more the fed-in power can be reduced the more wind turbines are connected to the central electrical bus (respectively the more power the connected wind turbine(s) feed in). Preferably, if all wind turbines of the plurality of wind turbines are connected with the central electrical bus, the power generated by the generator connected with the central electrical bus may be minimized to the least possible torque of the generator (driven by the gas turbine). HB / HB 230974WO 3. April 2024 In particular, during the idling operating mode of the wind farm system, the wind farm system respectively the internal electrical network may behave like a source and load network only. Some elements or components of the subsystems participating in the internal electrical network such as the inter-array cable networks, transformers, needed auxiliaries etc. may need magnetization and energization. This can be provided at the central electrical bus (e.g., a medium voltage (MV) bus) by the synchronous generator driven by a gas turbine (typically hydrogen- or ammonia- fired). Optimally and preferably, the generator can be supported by a static power equipment like STATCOM, Static VAR etc. As has been already described, during idling operating mode, the total auxiliary needs may be ~1 % of the total installed apparent power capacity of the wind farm. The active power and reactive power delivery to these components may correspond to the load current d0k, whereby the operating setpoints of the control signal inputs uwand ugokto the subsystem driving d0kmay be according to the idling load demands. Thus, the state of the positive, negative and zero sequence component [ẋ0ẋ1ẋ2] of the established power flow, comprising the bus voltage, in particular, their corresponding phase angles and frequency per equations (3) - (5) can be determined by the central controller for the initializing state. The initializing operating mode respectively stage may define the readiness of one or more (in particular all) of the active wind turbines in the wind farm system to be synchronized to the central electrical bus from a completely idling state. At this stage, the state of the power production system, network impedance for the operating regimes of the coupling loads, as determined by the central controller, can be communicated to the wind turbine controllers of the participating wind turbines by the central controller for their share of active and reactive power regulation. The energization sequence to start, synchronize and form the islanded internal electrical network by control of the apparent power demand is described, by way of a preferred embodiment, in the algorithm in Tables 3 and 4 (see Figure 8). HB / HB 230974WO 3. April 2024 In particular, it has been recognized that not more than approximately 1.63% of the total installed apparent power capacity of the wind farm system may be needed for the energization sequence, which is supplied by the synchronous generator. In particular, the generator can be dimensioned accordingly. The wind farm system after initiation, as already described, can be made ready for the formation of an internal electrical network respectively internal grid since the states are now known as well as the network impedance for a reliable setpoints command. In the absence of an external electric grid, the wind farm controller takes on the role of the grid manager in determining the setpoints of the electrical parameters of the grid- following wind turbines and connected loads while maintaining the integrity of the internal grid network, as described above. The active and corresponding reactive power are preferably controlled using a LQC scheme implemented in the central controller corresponding to the load impedance of the desired operating point. In the formation and maintenance of internal electrical network stability, clear and explicit communication can be performed, in particular, between each component, subsystem and the central controller. The explicit communication may seek to exchange information such as active states or faulty conditions of the subsystems in the network, prevailing or forecasted wind speeds and corresponding power of the wind turbines in the system, their actual operating points, operating points of the loads, etc. The wind turbines in the internal electrical network can be switched and synchronized sequentially to the central electrical bus as their respective wind turbine controllers may regulate their active and reactive power contributions according to the turbine setpoint data sets (e.g., comprising ug) provided to each participating windturbine ug1, ug2… ugn as determined by the central controller to match the needs of theconnected and concurrently controlled electrical loads ud1, ud2… udnand the respective load setpoint data sets. In this way, the internal electrical network can be regulated by both the injection of the apparent power of the wind turbines and the HB / HB 230974WO 3. April 2024 absorption of the injected apparent power by the interfacing electrical loads via their respective controllers. The sequential synchronisation managed by the central controller can be achieved in applying the predefined regulation sequence scheme, preferably, an algorithm described by the sequence steps in Table 4. The goal achievable by the predefined regulation sequence scheme is, in particular, a controlled synchronisation to ensure internal electrical network stability via sufficient provision of synchronisation power at each synchronisation sequence, i.e., at synchronisation of each wind turbine. This can be realised by providing (adequate) synchronising power or rigidity at the PCC forevery active wind turbine participating in the synchronisation by ensuring that d ≤0.5a. In wind turbines deploying partial converter with a doubly-fed induction generator, both the LSC and machine side converter (MSC) can coordinate the power injection at the prevailing wind speed. After synchronisation, in the event where all wind turbines in the wind farm system are producing energy in the internal electrical network, the synchronous generator can be retained in operation at the least possible torque to minimise fuel consumption while maintaining the voltage and frequency (V / f) reference for the grid following turbines, as previously described. If not all the wind turbines are synchronised or producing power, the synchronous generator may maintain an increased level of power production to deliver the power needs of the auxiliaries of the idling wind turbines and other subsystems. At any point in time, one or more wind turbines in the wind farm system can join or leave the existing synchronised internal electrical network after the step “# n_2” (see Table 4). Or the power produced by one or more of the wind turbines participating in the synchronised network may be reduced due to diminishing wind resources. Under any of these scenarios, the central controller may ensure that the corresponding equivalent amount of load can be controlled to maintain the internal HB / HB 230974WO 3. April 2024 electrical network integrity as power is injected or removed. In such a manner, (V / f) regulation of the network can be preserved. Variables a a unit of total WTG auxiliary idling consumption (kVA) total apparent power at idle mode state [energisation (consumption) of b strings, cable array, auxiliaries of connecting loads, transformers etc.] c total apparent power at initial state d max. operating power point for a unit of the external load (d ≤ 0.5a)during the synchronisation sequence n total number of wind turbine auxiliaries (corresponds to total no. of installed wind turbines) m (m = ∑l i=i mi ) number of external load units(whereby l is the number of synchronised wind turbine) e operating power point for the external load after all participating turbines are synchronised f no. of synchronised wind turbines Apparent Power (S) SSG operating point of the connected synchronous generator(s) SWTG-total total power of all synchronised wind turbines as made available by the line side converter Stotal total generated power available on the MV bus SL-eng total power for energisation (see b) SL-AUX-WTGs total power consumption of all the auxiliaries taking power from the MV bus SL-ext-load power consumption of the connected and controlled external load(s) taking power from the MV bus SL-total total power consumption as measured on the MV bus Table 3 A further aspect of the invention is a wind farm system, in particular, an offshore wind farm system. The wind farm system comprises a (common) central electrical bus connectable with at least one (specific) load (and auxiliary loads). The wind farm system comprises at least one wind turbine connectable to the central electrical bus. The wind farm system comprises at least one (rotating power and synchronous) HB / HB 230974WO 3. April 2024 generator (driven by a gas turbine) connected to the central electrical bus and providing at least a voltage reference and a frequency reference. The wind farm system comprises at least one previously described central controller (according to claim 18). According to a preferred embodiment of the wind farm system according to the invention, the wind farm may further comprise (as described hereinbefore): - the at least one load, and / or - at least one static reactive power equipment connectable to the central electrical bus. A further aspect of the invention is a computer program comprising instructions which, when the computer program is executed by at least one processor of a central controller, cause the processor to execute and / or control the previously described method, in particular, according to claim 1. It is noted that expressions such as "first", "second", etc. do not specify a series order, but only serve to distinguish between two elements (e.g., subsystems etc.). The features of the wind farm systems, central controllers, methods, and computer programs can be freely combined with one another. In particular, features of the description and / or the dependent claims, even when the features of the dependent claims are completely or partially avoided, may be independently inventive in isolation or freely combinable with one another. These and other aspects of the present patent invention become apparent from and will be elucidated with reference to the following figures. The features of the present application and of its exemplary embodiments, as presented above, are understood to be disclosed also in all possible combinations with each other. HB / HB 230974WO 3. April 2024 In the figures show: Fig.1 a schematic view of an embodiment of a central controller according to the present invention, Fig.2 a diagram of an embodiment of a method according to the invention, Fig.3 a schematic view of an embodiment of a wind farm system according to the present invention, Fig.4 a schematic view of a further embodiment of a central controller according to the present invention, Fig.5 a schematic view of an embodiment of a MIMO system according to the present invention, Fig.6a an equivalent circuit diagram of the internal electrical network in a first state, Fig.6b the equivalent circuit diagram of the internal electrical network in a second state, Fig.7 an example of a power generation unit of a wind turbine, and Fig.8 Table 4. Similar reference signs in different Figures indicate similar elements. Figure 1 shows a schematic view of an embodiment of a central controller 100 according to the present invention. The central controller 100 is, in particular, a central controller 100 of a wind farm system (not shown). Preferably, the functioning HB / HB 230974WO 3. April 2024 of the central controller 100 may be implemented in a wind farm controller of a wind farm of the wind farm system. The central controller 100 comprises at least one processor 102 and at least one data and / or program storage 104. In particular, the processor is configured to execute data / programs stored in the storage 104. Preferably, at least a computer program is stored in the storage 104 comprising instructions which, when the computer program is executed by the at least one processor 102 of the central controller 100, cause the processor 102 to execute and / or control the method according to the invention. The central controller 100 is described in more details with the aid of Figure 2. Figure 2 shows a diagram of an embodiment of a method according to the invention. The method can be executed by the central controller 100 for operating a wind farm system, in particular, an offshore wind farm system. The wind farm system is (permanently) electrically decoupled from an external grid. The wind farm system comprises at least one (grid following) wind turbine connected with a central electrical bus at least in a production operating mode of the wind farm system, at least one (rotating power and synchronous) generator (driven by a gas turbine) connected with the central electrical bus, at least one (specific) load connected with the central electrical bus at least in the production operating mode, and the central controller 100. In step 201, a supplying respectively feeding in, by the at least one generator of the wind farm system, of power to the central electrical bus is performed thereby providing at least a voltage reference and a frequency reference for the at least one connected wind turbine (at all times). In step 203, a determining, by the central controller 100, of the voltage and the frequency at the central electrical bus is performed. HB / HB 230974WO 3. April 2024 In step 205, a determining, by the central controller 100, of at least one turbine setpoint data set and at least one load setpoint data set is performed depending at least on the determined voltage and the determined frequency. In step 207, a controlling of the at least one connected wind turbine is performed according to the determined turbine setpoint data set. Preferably, the central controller may transmit the determined turbine setpoint data set to the respective wind turbine controller of the respective wind turbine to be controlled. The respective wind turbine controller may be configured to control the respective wind turbine in accordance with the received turbine setpoint data set. In step 209, a controlling the at least one load is performed according to the determined load setpoint data set. Preferably, the central controller may transmit the determined load setpoint data set to the respective load controller of the respective load to be controlled. The respective load controller may be configured to control the respective load in accordance with the received load setpoint data set. Preferably, each of the steps 201 to 209 is executed continuously and in particular in a parallel manner at least during the production operating mode of the wind farm system. Figure 3 shows a schematic view of an embodiment of a wind farm system 310 according to the present invention. The wind farm system 310 can be, in particular, an offshore wind farm system. The wind farm system comprises a (common) central electrical bus 322. As can be seen from Figure 3, at least one (specific) load 326 is connectable with the central electrical bus 322. Preferably, the wind farm system 310 comprises the at least one load 326. Further, at least one wind turbine 312.n of the wind farm system 310 is connectable to the central electrical bus 322. In the present embodiment, a plurality of wind turbines HB / HB 230974WO 3. April 2024 312.n (n may be, e.g., between 2 and 200) are connectable to the central electrical bus 322. The depicted wind farm system 310 comprises at least one (rotating power and synchronous) generator 342 connected to the central electrical bus 322 and providing at least a voltage reference and a frequency reference. Further, the wind farm system 310 comprises at least one (preferably exactly one) central controller 300, e.g., a controller as described in connection with Figure 1. The at least one wind turbine 312.n is configured to convert the kinetic wind energy into electrical energy respectively electrical power. The at least one wind turbine 312.n is, in particular, configured to supply the at least one specific load and at least one auxiliary load with electrical power. As described, the wind farm of the wind farm system 310 may comprise at least one string with a plurality of wind turbines 312.n, preferably a plurality of strings, wherein each string may comprise a plurality of wind turbines 312.n. The plurality of wind turbines 312.n, in particular, the one or more strings, may be electrically connectable to the central electrical bus 322, in particular, a (medium voltage) collector bus system. A wind turbine 312.n may comprise a wind turbine controller 314.n which may be connected with the central controller 300 via a (wired and / or wireless) data network 360. A wind turbine controller 314.n may be configured to control respectively set electrical parameters of the wind turbine 312.n, in particular, depending on a respective turbine setpoint data set received from the central controller 300. In particular, each wind turbine 312.n comprises a power generating unit 316 (only for a better overview, only one power generating unit 316 is shown). The power generating unit 316 (in particular, a Full Power Converter (FPC)), e.g., arranged in a nacelle of the wind turbine 312.n, may comprise a generator 352, switches, power HB / HB 230974WO 3. April 2024 converters 318, 320 (e.g., MSC 318, LSC 320), a transformer 354, a PGU (power generation unit) controller 328 and the like. The PGU controller 328 may be communicatively connected at least to the central controller 300 and (in particular the wind turbine controller 314.n) via a (wired and / or wireless) data network 360. The PGU controller 328 (e.g., together with the wind turbine controller 314.n) may be configured to control respectively set electrical parameters of the PGU 316, in particular, depending on a turbine setpoint data set received from the central controller 300. In particular, the wind farm system 310 may comprise at least one substation 324. The substation 324 may comprise the central electrical bus 322. The central electrical bus 322 is, in particular, configured to electrically connect the one or more strings (e.g., between two and twelve) of (offshore) wind turbines 312.n. Each string may comprise a plurality of offshore wind turbines 312.n (e.g., between four and twelve). The substation 324 may comprise several components. In particular, the substation may comprise the (synchronous) generator 346. In other variants of the invention, the generator 346 may be arranged on another structure of the wind farm system 310. As can be seen from Figure 3, the generator 346 is driven by a gas turbine 348. The gas turbine 348 is, in particular, a hydrogen-fired or ammonia-fired turbine 348. Preferably, the hydrogen or ammonia can be supplied by the at least one specific load 326, which will be described hereinafter. A generator controller 350 may be provided. The generator controller 350 may be communicatively connected at least with the central controller 300 via the data network 360. The generator controller 350 may be configured to control respectively set electrical parameters of the generator 342, in particular, depending on a generator setpoint data set determined by and received from the central controller 300. HB / HB 230974WO 3. April 2024 The generator 342 may be suppliable with auxiliary power (e.g., LV AC) by means of a port 362. The generator 342 may be connectable to the central electrical bus 322 via a generator transformer 344. As can be seen from figure 3, preferably, the wind farm system 310 comprises at least one static reactive power equipment 346, 336, particularly preferably two pieces of reactive power equipment 346, 336. A reactive power equipment 346, 336 may comprise a Static Synchronous Compensator (STATCOM) and / or a static VAR (volt-ampere reactive) compensator (SVC) and / or the like. A first static reactive power equipment 346 may be connectable with the central electrical bus 322 via the generator transformer 344. A second static reactive power equipment 336 may be directly connectable with the central electrical bus 322. As can be further seen, each static reactive power equipment 346, 336 may comprise a static reactive power equipment controller 348, 338 communicatively connected at least with the central controller 300. A static reactive power equipment controller 348, 338 may be configured to control respectively set electrical parameters of the static reactive power equipment 346, 336, in particular, depending on a static reactive power equipment setpoint data set received from the central controller 300. It shall be understood that only the reactive power equipment 346 or only the reactive power equipment 336 may be provided. Preferably, the at least one static reactive power equipment 346, 336 can be arranged on the substation 324. Further, the wind farm system 310 may comprise a load power supply 356 preferably arranged on the substation 324. The load power supply 356 may be configured to control the power supply from the central electrical bus 322 to the at least one load 326. In particular, a load controller 366 may be provided. A load controller 366 may be configured to control respectively set electrical parameters of the load power supply 356, in particular, depending on a load setpoint data set received from the central controller 300 (via the data network 360). HB / HB 230974WO 3. April 2024 The load power supply 356 may comprise a transformer and a power converter. Further, power can be fed to the specific load 326 via power connection 334 (e.g., a MV connection). Further, power can be fed to the at least one (not shown) auxiliary load via at least one further power connection 364 (e.g., a LV connection). Exemplified and non-exhaustive examples of auxiliary loads are hydraulics, pumps, heaters etc. for lubrication, cooling, climate conditioning and so on. As already described, the specific load 326 may be preferably a hydrogen production plant. Merely for the sake of a better overview, the components of the hydrogen production plant are not depicted in Figure 3. A hydrogen production plant (also denoted as electrolyzer unit) may comprise at least one electrolyzer. The electrolyzer or a stack of electrolyzers may be configured to produce hydrogen at e.g., 30 bar. The at least one electrolyzer may be arranged outside the wind turbine, for instance, in a housing on a (not shown) main access platform. The electrolyzer plant may comprise at least one (hydrogen) processing module respectively (hydrogen) finishing module, e.g., arranged on a main access platform. The processing module may be configured to process the produced (wet) hydrogen. The at least one hydrogen processing module can comprise at least one (not shown) hydrogen drying module. The hydrogen drying module can be configured to dry the generated wet hydrogen. Preferably, the hydrogen drying module (after separation of the liquid phase) can be an adsorption drying module (in particular a TSA (temperature swing adsorption) module) and / or a refrigeration drying module. The adsorption drying module may be in particular configured to dry the wet hydrogen by adsorption using a (suitable) adsorber, in particular silica beads. The at least one hydrogen drying module can be configured to cool the wet hydrogen to a temperature of at least less than 5°C. Furthermore, the at least one hydrogen processing module HB / HB 230974WO 3. April 2024 can comprise a (not shown) catalytic deoxo stage, in particular, for treating the wet hydrogen before drying as described above. Such a stage serves to remove oxygen (≤0.5%). In variants of the invention, such treatment can also take place at a later stage, for example, at a hydrogen destination structure to which the treated hydrogen can be transported, for example, via a hydrogen pipeline network. Preferably additionally, the at least one hydrogen processing module may comprise at least one hydrogen compression module. The hydrogen compression module may be configured to compress the (processed) hydrogen, in particular, the dried hydrogen. The hydrogen compression module may comprise at least one hydrogen compressor (in particular a (vertical) piston compressor). The hydrogen compression module 430 can be configured to compress the dried hydrogen to at least 50 barg (and for example at most 250 barg), preferably to essentially 70 barg. The hydrogen compression module may comprise at least one hydrogen compression controller. The hydrogen compression controller may be communicatively connected at least to the hydrogen controller via the data network. The at least one hydrogen compression module may be arranged outside the wind turbine, for instance, in a housing on the main access platform. In further variants of the invention, the at least one hydrogen processing module may comprise at least one (not shown) transforming module configured to transform the produced hydrogen to ammonia (e.g., via the Haber-Bosch-method or the like). The at least one electrolyzer plant may comprise at least one water treatment module, e.g., arranged in a housing on the main access platform. The at least one water treatment module can be configured to treat the water such that it can be used by the at least one electrolyzer to produce hydrogen. In particular, the water treatment module can treat seawater for the electrolysis process. Preferably, the at least one water treatment module can be a seawater desalination module with membrane- based pressure filtration. Such a seawater desalination module can carry out reverse HB / HB 230974WO 3. April 2024 osmosis, ultrafiltration and / or electrodialysis (also known as electrodeionization (EDI)) to treat the water respectively to treat the seawater accordingly. Such treatment can be used to treat seawater, in particular in an energy-efficient manner, which can be used for water electrolysis. The treated water can also be referred to as ultrapure water or "demin water". The treated water can be fed to the electrolyzer module via at least one fluid pipe of a (not shown) internal fluid network. Alternatively or additionally, the at least one electrolyzer plant may comprise at least one (not shown) inert gas generation module configured to generate an inert gas. Only a portion of the produced hydrogen (or ammonia) may be fed to the gas turbine 340 (e.g., via a gas storage tank). The other portion may by fed into a gas pipeline network or the like. It is noted that the excitation power (of 367) may be 1% of rated rotating machine and that 365 denotes an input from bus (low voltage auxiliary powert). Further, ≥2 medium voltage units 50 / 60 Hz, 0.85 pf. Figure 4 shows a schematic view of a further embodiment of a central controller according to the present invention. In particular, Figure 4 shows the central controller 400 and the central electrical bus 422. In particular, Figure 4 shows the structure of the LQC scheme implemented in respectively by the central controller 400. This preferred control scheme enables, in particular, an incorporation of cross dependencies between frequency and power as well as reactive power and voltage, and generally, the exploitation of the system knowledge in the controller design of the central controller 400. In addition, since model certainty is assured, utilizing LQC facilitates the application of first principles-based state-space model of the wind farm system, e.g., according to Figure 3. In particular, as described above the state of the internal electrical grid of the HB / HB 230974WO 3. April 2024 wind farm system can be described by equation (12) by the three parameters ẋ0, ẋ1and ẋ2which correspond to the discrete-time states of the zero, positive and negative sequence respectively in equation (6). By deriving ẋ0, ẋ1& ẋ2as functions of equation (6), the system states in equations (7) and (8) can be determined by iterative method using the matrices given in equations (9) - (11) to obtain the system output y, which describes, in particular, the stable internal electrical network of the wind farm system. The state feedback control is implemented using the structure in Figure 4, wherein the wind farm system respectively the internal electric network of the wind farm systemto be controlled is a system of a central electrical bus 422 having G(s), W(s), D(s),where G(s), W(s) and D(s) are the transfer functions of the first subsystem (inparticular, the power generating subsystem which comprises the synchronous generator(s) and wind turbine(s) in the internal electrical network), the third subsystem (in particular, the at least one static (reactive) power equipment and the second subsystem (in particular, the interfacing electrical loads) respectively. Since the internal electric network of the wind farm system is formed, in particular, of different subsystems, the input and output control signals dimension are multiple, thus leading to a multiple-input, multiple-output (MIMO) system. This is detailed in Figure 5. In this way, at any point in time, the stability of the system containing the aforementioned subsystems is ensured via the feedback of the same states in equation (13) in the control loop seen in Figure 4. The state feedback control seen in Figure 4 uses, in particular, the matrix Kito calculate the control signal from the system state x and achieve system stability accuracy by multiplying the command signal w with the pre-factor Kvas represented by equation (13). The LQC is described by equation (14), wherein the states x are penalized with the weighting matrix Q and the control effort with the weight of matrix R in equations (15) and (16) respectively. HB / HB 230974WO 3. April 2024 The selection of the matrices Q and R is done according to Bryson’s rule which corresponds to the criterion in equation (17). Simulations according to Bryson’s rule can be applied by the central controller 400 to determine the values of Q and R where the subsystems achieved fast reaction to the needed setpoint changes p and q in the input signal u according to equation (18) and Figure 4, where p and q are active and reactive power setpoints of a respective load or turbine setpoint data set respectively for the participating subsystems in the MIMO scheme shown in Figure 5. After tuning the matrices according to the desired output matrix, the Q and R matrices can be used to solve the Algebraic Riccati Equation (ARE) to compute the full state feedback matrix, which is essentially the LQR control. For a faster response of the central controller 400, the Q matrix values can be changed, whereas for minimizing stability error while achieving the desired setpoints data sets without focusing on the time of response, the R matrix values can be preferably adjusted. The values of the matrices to achieve the preferred control goal – fast response for needed level of internal electrical network stability have been tuned for A, B, C, Q and R in equations (19) - (23) respectively. Thus, the chosen weighting matrix Q penalises only the zero sequence ẋ0. The values are just the starting point to a trial-and-error iterative design procedure aimed at obtaining desirable properties for the closed-loop system. Figure 6a and 6b show equivalent circuit diagrams of the internal electrical network. The electrical properties of the central electrical bus, in particular in the form of a medium voltage bus, of the internal electrical network are not only characterised by the voltage, its frequency and phase angle but also by their internal impedance respectively network impedance. The value of minimum impedance at the nominal network frequency is, in particular, crucial to synchronism and stability since it determines the maximum possible short-circuit. A model of the equivalent circuitdiagram of the network is described in Figure 6a, wherein IWF(rated) < φi is the vectorHB / HB 230974WO 3. April 2024 of the rated nominal current of the internal electrical network of the wind farm system. Likewise, the synchronous generator of the wind farm system provides the vector ofthe internal electrical network voltage, Vg < φv in the formed internal electricalnetwork. In an idling state respectively idling operating mode of the wind farm system, the internal electrical network can be described by Figure 6b, wherein the loads at the MV bus can be limited by the auxiliary consumptions of the wind farm system. The approximate amount of the internal impedance of the internal electrical network to ensure power transfer to the loads results, in particular, from the relationships in equations (24) - (26), wherein is the equivalent longitudinal impedance of the connecting inter-array cable lines respectively strings of the participating wind turbine, Zgis the equivalent impedance of the interfacing generator transformer and the synchronous generator and SCR is the short-circuit ratio. There is no equivalent longitudinal impedance of an upstream (external) grid network since the internal electrical network is an islanded network, thus, resulting in a high impedance via the medium voltage and apparent power levels provided by the synchronous generator and the interfacing generator transformer. Nonetheless, this can ensure full power flow if the network impedance is lowered at minimum possible Zd, the impedance of the connected external loads via means of load regulation. The regulation or control of the loads can ensure optimal impedance for power transfer and contributes to the internal electrical network stability alongside the contributions of the power generating units. As described, in equation (26), the maximum terminal short-circuit (un-faulted) Idtoccurs at the point of minimum possible Zn, in particular to ensure power transfer when the maximum loads are switched on. Therefore, the sizing of the static reactive power equipment to provide reactive power for voltage regulation may be HB / HB 230974WO 3. April 2024 optimisable. In addition, the terminal short-circuit at any point in time can be controlled, by the central controller, via a combination of sequential synchronisation and load regulation to achieve internal electrical network stability. By using, in particular equations (24) - (26), the central controller can determine the impedance in real time and the corresponding operating active and reactive power injection or their withdrawal from the internal electrical network. In this way, the ability of the wind farm system via its coordinating central controller to regulate the voltage at the central electrical bus can be ensured. The behaviour of the network under the different states shown in Figure 5 is, in particular, given in Table 2. Under short-circuit fault conditions, Zndrops below the minimum short-circuit impedance alongside the bus voltage as given by #4 in Table 4. Voltage regulation in this state can be provided by the support of the static power equipment which can be controlled, by the central controller, in response to the prevailing bus voltage condition and needed compensation. The voltage limit for the operation of the internal electrical network under low voltage conditions is given in Table 1, otherwise, the faulted subsystems triggering the voltage drop can be disconnected and isolated from the internal electrical network for stability and continued operation of the remaining internal electrical network respectively subsystem(s). Figure 7 shows an example of a power generation unit 716 of a wind turbine. In order to avoid repetitions, only the difference to the power generation unit of Figure 3 is explained below. For the other components, it is referred to the embodiment of Figure 3. In particular, Figure 7 shows an example of a power generation unit 716 of a wind turbine with a Partial Converter (PC). During the idling mode of the wind farm system (e.g., according to Figure 3), the internal electrical network behaves like a source and load network only. Some elements or components of the subsystem participating in the system such as the inter-array cable networks, transformers, needed auxiliaries etc. may need magnetization and energization. This may be provided for at the central electrical bus HB / HB 230974WO 3. April 2024 by the synchronous generator driven by a gas turbine and supported by static power equipment like STATCOM, Static VAR etc. During complete idling, i.e., no power production from any of the wind turbine(s), the total auxiliary needs may be ~0.87 % of the total installed apparent power capacity of the wind farm system. The active power and reactive power delivery to these components may correspond to the load current d0kas described in Figures 5 and 6, wherein the operating setpoints of the control signal inputs uwand ugok(i.e., the respective setpoint data sets) to the subsystem driving d0kare according to the idling load demands. Thus, the state of the positive, negative and zero sequence component[ẋ0ẋ1ẋ2]of the established power flow, consisting of the bus voltage, their corresponding phase angles and frequency per equations (3) - (5) can be determined by the central controller for the initialing state. The initialing operating mode respectively stage of the wind farm system may define the readiness of one or all of the available wind turbines in the wind farm system to be synchronized to the central electrical bus from a (completely) idling state. At this stage, the state of the power generating subsystem, network impedance for the operating regimes of the coupling loads, as determined by the central controller are communicated to the wind turbine controllers of the participating wind turbines for their share of active and reactive power regulation. The energization sequence to start, synchronize and form the islanded internal electrical network by control of the apparent power demand can be predefined by a predefined regulation sequence scheme preferably in form of the algorithm in Tables 3 and 4 (see Figure 8). Not more than ~1.63% of the total installed apparent power capacity may be needed for the energization sequence, which is supplied by the synchronous generator of the wind farm system. The internal electrical network after initiation as already described is made ready for the formation of an internal grid network with all available wind turbines since the states are now known as well as the network impedance for a reliable setpoints command respectively reliable determination of the respective setpoint data sets. In HB / HB 230974WO 3. April 2024 the absence of an external electric grid, the central controller takes on the role of the grid manager in determining the setpoints data sets with the electrical parameters of the grid-following wind turbines and connected loads while maintaining the integrity of the internal electrical network. The active and corresponding reactive power can be controlled using the described LQC scheme implemented in the central controller corresponding to the load impedance of the desired operating point. In the formation and maintenance of internal electrical network stability, communication can be performed between the controllers of the subsystems and the central controller, as indicated in Figure 3. The explicit communication may seek to exchange information such as active states or faulty conditions of a subsystem in the internal electrical network, prevailing or forecasted wind speeds and / or corresponding power of the wind turbines in the system, their actual operating points, operating points of the loads, etc. The wind turbines of the internal electrical network can be preferably switched and synchronized sequentially to the central electrical bus as their individual wind turbine controllers may regulate their active and reactive power contributions according to the received turbine setpoint data sets (comprising inputs ug), in particular providedto each participating wind turbine ug1, ug2… ugn as determined by the centralcontroller to match the needs of the connected and concurrently controlled electrical loads ud1, ud2… udn. In this way, the internal electrical network can be regulated by both the injection of the apparent power of the wind turbines and the absorption of the injected apparent power by the interfacing electrical loads, in particular, via their respective controllers. The sequential synchronisation managed by the central controller can be preferably achieved in an algorithm described by the sequence steps in table 4. The goal is in particular a controlled synchronisation to ensure grid stability via sufficient provision of synchronisation power at each synchronisation sequence, i.e., at synchronisation of each wind turbine. This can be realised by providing adequate synchronising power respectively rigidity at the PCC shown in Figure 6 for every active wind turbine HB / HB 230974WO 3. April 2024participating in the synchronisation by ensuring that d ≤ 0.5a. Only after synchronismof all participating wind turbines in the internal electrical network is achieved, is full injection of available active and reactive power by the line side converter (LSC) of wind turbines with full power converter implemented according to sequence step “# n_2”. In wind turbines deploying partial converter with a doubly-fed induction generator as shown in figure 7, both the LSC and machine side converter (MSC) coordinate the power injection at the prevailing wind speed. After synchronisation, in the event where all wind turbines in the system are producing energy in the internal electrical network, i.e., the wind farm system is in the production operating mode, the synchronous generator is preferably retained in operation at the least possible torque to minimise fuel consumption while maintaining the voltage and frequency (V / f) reference for the grid following turbines. If not all the wind turbines are synchronised respectively producing power, the synchronous generator may maintain an increased level of power production to deliver the power needs of the auxiliaries of the idling wind turbines and other subsystems. At any point in time, one or more wind turbines in the wind farm can join or leave the existing synchronised central electrical bus after the step “# n_2”. Or the power produced by one or more of the wind turbines participating in the synchronised network may be reduced due to diminishing wind resources. Under any of these scenarios, the central controller ensures that the corresponding equivalent amount of load is controlled to maintain the internal electrical network integrity as power is injected or removed. In such a manner, (V / f) regulation of the internal electrical network is preserved. Finally, some of the terms used hereinbefore are explained hereinafter: System – the term "System" is, in particular, used to describe the entire participating components within the borders of the wind farm or internal grid on the central electrical bus in Figure 3. There may be no integration of wind turbines or loads into HB / HB 230974WO 3. April 2024 the system other than the provided interfaces. The system is composed and operated by the interaction of subsystems. Subsystem – the term “Subsystem” is, in particular, used to describe entities within the wind farm system or internal electrical network that are composed of functions and interaction of components. They may provide functionality into the wind farm system and require a dedicated control. E.g., wind turbines, (rotating) power generators, external loads, etc. Components – the term “Components” is used, in particular, for entities within the subsystem to aid their functionality. They may or may not require a dedicated controller, e.g., power generating units (PGU), rotor blades of wind turbines, etc. External Grid – the small-area or wide-area macrogrid formed by an infinite bus or extensive network of power system network where statutory grid code applies. Internal electrical network – the grid formed without the aid of an external macrogrid; it is in particular a finite bus with controllable loads. Here, no statutory grid code may apply. Requirements for external grid support such as fault-ride through, current sequence injections, etc. may be needed but only to the extent at which grid stability is ensured. Some of the requirements of the internal electrical network are given in table 1 See also table 3 & 4. Central controller respectively Internal Grid Manager – a feature of the wind farm controller which computes power system parameters of the internal electrical network, keeps track of their values and communicates them in real time to the components and subsystems of the internal electrical network in order to maintain the requirements needed for the stability of the internal electrical network. It is also informed of, and keeps track of other relevant parameters by the participating subsystem, such as prevailing wind speed of each wind turbine, gas storage levels of the synchronous generator’s prime mover, etc. HB / HB 230974WO 3. April 2024 Auxiliaries respectively auxiliary loads – power consuming equipment which are the internal consumption of the wind farm, also known as balance of plant (BoP) equipment. Specific Load respectively External Load – power consuming equipment at the terminals of the internal electrical network, other than the internal load consumption (auxiliaries) of the islanding wind farm. HB / HB 230974WO 3. April 2024

Claims

3. April 2024 Cl a i m s1. A method for operating a wind farm system (310), in particular, an offshore wind farm system (310), wherein the wind farm system (310) is electrically decoupled from an external grid, and wherein the wind farm system (310) comprises at least one wind turbine (312) connected with a central electrical bus (322, 422) at least in a production operating mode of the wind farm system (310), at least one generator (342) connected with the central electrical bus (322, 422), at least one load (326) connected with the central electrical bus (322, 422) at least in the production operating mode, and at least one central controller (100, 300, 400), the method comprising: - supplying, by the at least one generator (342), power to the central electrical bus (322, 422) thereby providing at least a voltage reference and a frequency reference for the at least one connected wind turbine (312), - determining, by the central controller (100, 300, 400), the voltage and the frequency at the central electrical bus (322, 422), - determining, by the central controller (100, 300, 400), at least one turbine setpoint data set and at least one load setpoint data set depending at least on the determined voltage and the determined frequency, - controlling the at least one connected wind turbine (312) according to the determined turbine setpoint data set, and - controlling the at least one load (326) according to the determined load setpoint data set.

2. The method according to claim 1, wherein - controlling the at least one connected wind turbine (312) according to the determined turbine setpoint data set comprises transmitting, by the central controller (100, 300, 400), the turbine setpoint data set to a wind turbine- 2 - controller (314) of the at least one wind turbine (312), and / or - controlling the at least one connected load (326) according to the determined load setpoint data set comprises transmitting, by the central controller (100, 300, 400), the load setpoint data set to a load controller (366) of the at least one load (326).

3. The method according to claim 1 or 2, wherein - the at least one load (326) comprises a hydrogen production plant.

4. The method according to any of the preceding claims, wherein - the at least one load (326) comprises an auxiliary load of the wind farm system (310).

5. The method according to any of the preceding claims, wherein - a respective turbine setpoint data set is determined for each connected wind turbine (312), - wherein each connected wind turbine (312) is controlled according to the determined respective turbine setpoint data set, and / or - a respective load setpoint data set is determined for each connected load (326), - wherein each connected load (326) is controlled according to the determined respective load setpoint data set.

6. The method according to any of the preceding claims, wherein - the at least one turbine setpoint data set comprises a turbine active power setpoint and a turbine reactive power setpoint, and / or - the at least one load setpoint data set comprises a load active power setpoint and a load reactive power setpoint. HB / HB 230974WO 3. April 2024- 3 - 7. The method according to any of the preceding claims, wherein - the at least one turbine setpoint data set is determined based on a linear quadratic control scheme, and - the at least one load setpoint data set is determined based on the linear quadratic control scheme.

8. The method according to claim 7, wherein - the linear quadratic control scheme is based on a discrete state space representing the internal electrical network of the wind farm system (310).

9. The method according to claim 8, wherein - a first transfer function (G(s)) represents the at least one wind turbine (312) and the generator (342), and - a second transfer function (D(s)) represents the at least one load.

10. The method according to claim 9, wherein - the wind farm system (310) comprises at least one static reactive power equipment (336, 346) connectable with the central electrical bus, and - a third transfer function (W(s)) represents the at least one static reactive power equipment (336, 346).

11. The method according to any of the preceding claims 8 to 10, wherein - the output of the discrete state space represents a stable state of the internal electrical network of the wind farm system (310).

12. The method according to any of the preceding claims, further comprising - determining the network impedance of the internal electric network of the wind farm system (310) in real time based on an equivalent circuit model of the internal electric network of the wind farm system (310) thereby determining corresponding operating active and reactive power injection into the internal HB / HB 230974WO 3. April 2024- 4 - electric network and / or active and reactive power withdrawal from the internal electric network.

13. The method according to any of the preceding claims, wherein - the wind farm system (310) is at least operable in the production operating mode and in an idling operating mode, - in the idling operating mode, solely the generator (342) supplies power to the central electrical bus (322, 422).

14. The method according to claim 13, wherein - in the idling operating mode, the generator (342) supplies power of at least approximately 0.87 % of the total apparent power capacity of the installed wind turbines (312) of the wind farm system (310).

15. The method according to any of the preceding claims, wherein - the wind farm system (310) is at least operable in the production operating mode and in an initializing operating mode, - in the initializing operating mode, the one or more wind turbines (312) are switched to the central electrical bus (322, 422) according to a predefined regulation sequence scheme.

16. The method according to claim 15, wherein according to the predefined regulation sequence scheme: - after an idling operating mode, a first wind turbine (312) of a plurality of wind turbines (312) is switched to the central electrical bus (322, 422) in a first synchronization step, - the internal electric network is stabilized in a first stabilization step following the first synchronization step, - a further wind turbine (312) of the plurality of wind turbines (312) is switched to the central electrical bus (322, 422) in a further synchronization step, and HB / HB 230974WO 3. April 2024- 5 - - the internal electric network is stabilized in a further stabilization step following the further synchronization step.

17. The method according to any of the preceding claims, wherein - in the production operating mode, the power generated by the generator (342) connected with the central electrical bus (322, 422) depends on the number of the wind turbines (312) of the plurality of wind turbines (312) connected with the central electrical bus (322, 422), - wherein if all wind turbines of the plurality of wind turbines, the power generated by the generator (342) connected with the central electrical bus (322, 422) is minimized to the least possible torque of the generator (342).

18. A central controller (100, 300, 400) for a wind farm system (310), in particular, an offshore wind farm system (310), wherein the wind farm system (310) is electrically decoupled from an external grid, and wherein the wind farm system (310) comprises at least one wind turbine (312) connected with a central electrical bus (322, 422) at least in a production operating mode of the wind farm system (310), at least one generator (342) connected with the central electrical bus (322, 422), at least one load (326) connected with the central electrical bus (322, 422) at least in the production operating mode, wherein the central controller (100, 300, 400) is configured to: - control the at least one generator (342) such that the at least one generator (342) supplies power to the central electrical bus (322, 422) thereby providing at least a voltage reference and a frequency reference for the at least one connected wind turbine (312), - determine the voltage and the frequency at the central electrical bus (322, 422), - determine at least one turbine setpoint data set and at least one load setpoint data set depending at least on the determined voltage and the determined frequency, - control the at least one connected wind turbine (312) according to the determined turbine setpoint data set, and HB / HB 230974WO 3. April 2024- 6 - - control the at least one load (326) according to the determined load setpoint data set.

19. Wind farm system (310), in particular, an offshore wind farm system (310), comprising - a central electrical bus (322, 422) connectable with at least one load (326), - at least one wind turbine (312) connectable to the central electrical bus (322, 422), - at least one generator (342) connected to the central electrical bus (322, 422) and providing at least a voltage reference and a frequency reference, and - at least one central controller (100, 300, 400) according to claim 18.

20. The wind farm system (310) according to claim 19, further comprising - the at least one load (326), and / or - at least one static reactive power equipment (336, 346) connectable to the central electrical bus (322, 422). HB / HB 230974WO 3. April 2024

Images