CONTROL OF POWER SUPPLIED TO AN ELECTRICAL POWER DISTRIBUTION NETWORK BY A HYBRIDIZING CONTROLLER

Abstract

To control the hybridization of an electrical production group comprising an electrical power production unit (1) and an energy storage unit (2), a hybridization controller (34) includes an inference controller (343) which receives as input a first signal (ESS_SOC) representing the state of charge of the storage unit, a second signal (PESS) representing the power stored or delivered by the storage unit and a third input signal representing the time derivative of the power (PHPP_FCR) produced by the production unit, and produces as output a signal (F_HPPn) setting the power produced by the production unit as its contribution to the primary frequency control of the electrical network (3).The hybridization controller (34) is configured to control the power (PESS) stored or released by the energy storage unit (2) from the difference (342) between the expected primary frequency control power (PFCR) from the generating group as its contribution to the primary frequency control of the electrical grid (3) and the power (PHPP_FCR) produced for this purpose by the electrical power generation unit (1). Figure 3.

Description

Title of the invention: CONTROL OF POWER SUPPLIED TO AN ELECTRICAL POWER DISTRIBUTION NETWORK BY A HYBRIDIZING CONTROLLER Technical field

[0001] The invention relates generally to the field of electrical power generation and the management of electrical power distribution networks. More particularly, it relates to the strategies implemented within power generation plants to maintain the frequency stability of electrical power distribution networks.

[0002] More specifically, the invention relates to the control of the power delivered by an electrical power generation group to an electrical power distribution network under an alternating voltage of determined nominal frequency, said electrical power generation group comprising a controllable electrical power generation unit having a relatively long response time to a variation in the setpoint of the power produced and a controllable energy storage unit having a relatively fast response time to a variation in the setpoint of the power stored or delivered, controlled by a hybridization controller configured to ensure the management of the electrical power generation group and its contribution to the primary adjustment of the frequency of the alternating voltage according to a deviation of said frequency from the nominal frequency. State of the art

[0003] Electrical energy is transported in electrical power distribution networks under an alternating voltage of a determined nominal frequency, namely 50 Hertz (Hz) in Europe or 60 Hz in other geographical areas. This frequency is commonly referred to as the "network frequency".

[0004] The management of electricity distribution networks consists of maintaining a constant balance between the electricity produced and injected into the network, on the one hand, and the electricity consumed at any given moment by all subscribers to the distribution service. Any imbalance between overall electricity production and consumption results in a deviation of the network frequency from the aforementioned nominal frequency. The network frequency changes constantly, with a deviation ranging from a few millihertz to a few hundred millihertz. For example, a network frequency deviation of 0.2 Hz corresponds to an imbalance of 3 gigawatts (GW) between the electricity injected into the network by all power generation units. supplying it and the electrical energy consumed by all the electrical consumer equipment connected to it by subscribers to use the distributed electrical energy.

[0005] To automatically ensure network stability, several mechanisms, generally standardized, exist that allow the balance between electricity consumption and production to be maintained at all times. These mechanisms make it possible to stabilize the electrical characteristics of the network at the local, national, or larger scale.

[0006] For example, in Europe there is a mechanism called "primary frequency control" or FCR (Frequency Containment Reserve), which allows the power individually produced by each of the electricity generating units supplying the grid to be modulated according to the deviation in the electricity grid frequency reflecting an imbalance between overall production and consumption. Many specific solutions are also implemented by electricity producers at the level of their power generation units, which aim to contribute to primary frequency control by responding more or less rapidly to variations in grid demand from consumers.Implementing these solutions at each production unit provides frequency stabilization reserves that automatically react to compensate for frequency deviations and stabilize the network frequency at a value as close as possible to its nominal frequency. The main technical requirement for such an implementation is an automatic and proportional response to network frequency deviations within the shortest possible time, a few seconds at most.

[0007] In the existing art disclosed by European patent application EP2139090A1, it is proposed to provide FCR service using a battery storage system. A methodology based on predicting grid frequency variations is disclosed to dynamically optimize the minimum and maximum state-of-charge thresholds of the batteries. Specifically, when a battery's state of charge falls below the minimum threshold, recharging from the grid is necessary, and conversely for the maximum threshold. The aim is to reduce costly recharging periods. The success of this method relies on predicting grid events and is based on a record of past events. The authors justify their invention by arguing that periods such as morning and early evening correspond to times when the grid frequency is generally below its nominal value and are therefore more easily predictable.One drawback of this solution is the need to establish robust predictions over a long time horizon. However, this document does not propose any technical means to achieve this. Furthermore, with... With the increasing share of renewable energy in the energy mix, these network fluctuation events become difficult to anticipate given the local meteorological aspects at solar or wind power plants.

[0008] Furthermore, electrical grids are evolving towards a hybridization of traditional power plants with energy storage systems (batteries, flywheels, etc.). This hybridization improves the overall performance of a generating unit. Indeed, compared to a power generation unit, which has a relatively long response time (considering the response times expected for primary frequency control) to a change in the setpoint for the power produced, an energy storage unit generally has a relatively fast response time (again, considering the response times expected for the needs of the FCR service) to a change in the setpoint for the stored or delivered power.Therefore, appropriate control of these two units by a hybrid controller allows for more efficient management of the power generation unit and its individual contribution to the primary frequency regulation of the grid. In particular, hybridizing a generation unit with a storage unit improves the overall performance of the resulting unit, while reducing the demands placed on this hybrid unit, which ultimately helps to reduce its maintenance requirements.

[0009] International patent application WO2012177633A2, corresponding to European patent EP2721710, discloses a power generation unit comprising a hybrid power plant with a battery energy storage system whose function is to assist the power plant in maintaining the nominal frequency of the electrical grid. To this end, it is described that a regulation signal representing the grid frequency deviation is filtered so that the high-frequency component is applied to the energy storage system, while the low-frequency component is applied to the power generation unit. However, only this generic principle is presented, and no technical description of solutions to be implemented in practice is provided.

[0010] The principle of filtering a control signal representative of the grid frequency deviation is also described in the scientific article by C. Jin, N. Lu, S. Lu, Y. Makarov and R.A. Dougal, "Coordinated contract algorithm for hybrid energy storage systems", 2011 IEEE Power and Energy Society General Meeting, Detroit, MI, USA, 2011, pp. 1-7. According to this document, a hydroelectric power plant is hybridized with a flywheel, and the principle of managing this unit, based on filtering a control signal representative of the grid frequency deviation, is described, specifying methods for managing the flywheel. and flywheel control, which are based on minimum and maximum operating limits. When the flywheel is discharged and its state of charge (SOC) reaches its minimum limit, it is recharged using hydraulic energy until it returns to its charging point.

[0011] The scientific article by Ritu Raj Shrivastwa, Ahmad Hably, Seddik Bacha, Hugo Mesnage, Renaud Guillaume, "An overview of Hybridization of Power sources for Ancillary Service", ICIT 2020 - IEEE International Conference on Industrial Technology, Feb 2020, Buenos Aires, Argentina, also discloses a teaching related to a filtering of a regulation signal representative of the frequency deviation of the network to control the operation of several energy sources.

[0012] Finally, patent applications FR3131473 and FR3131474, from the same applicant as the present application, also describe solutions for managing the hybridization of a power plant with an electrical energy storage unit that are based on filtering a control signal representative of the grid's frequency deviation. These solutions provide good results, but continued research and development efforts in this area make it possible to propose an alternative, which is not based on the principle of filtering a control signal representative of the grid's frequency deviation to control the power generation unit and the energy storage unit. Description of the invention

[0013] The invention thus relates firstly to a method for controlling power supplied to an electrical power distribution network (3) under an alternating voltage of determined nominal frequency, by an electrical power generation unit comprising: • an electrical energy production unit controllable by a controller regulating the energy produced, having a relatively long response time to a variation in the setpoint of the power produced; • a controllable energy storage unit, having a relatively short response time to a change in the setpoint of the stored or delivered power, and used for primary frequency control (FCR: "Frequency Containment Reserve") of the alternating voltage; and, • a hybridization controller configured to manage the power generation unit and its contribution to the primary frequency regulation of the AC voltage based on a signal representing a deviation of said frequency from the nominal frequency, by controlling the power stored or released by the energy storage unit based on the difference between the expected primary frequency regulation power of the unit production as a contribution to the primary frequency regulation of the electrical grid and the power produced for this purpose by the electrical power generation unit, said process comprising the following steps, implemented in the hybridization controller: • acquisition of a first input signal which is representative of the state of charge of the energy storage unit; • acquisition of a second input signal which is representative of the power stored or released by the energy storage unit; • acquisition of a third input signal which is representative of the time derivative of the power produced by the electrical power generation unit as its contribution to the primary frequency regulation of the electrical network; • production of a signal of the setpoint of the power produced by the electrical power production unit as its contribution to the primary frequency regulation of the electrical network, said setpoint signal being produced using at least one inference processor with at least three inputs, as a function of the first input signal, the second input signal, and the third input signal.

[0014] As a person skilled in the art will have understood, the invention does not fall within the scope of the known solutions, set out in the introduction above, which are based on filtering a regulation signal representative of the frequency deviation of the network and which consist of using the high frequency component of this signal to control the energy storage unit while its low frequency component is used to control the electrical power generation group.

[0015] The state of charge (SOC) of the energy storage unit is taken into account, making it possible to monitor whether it deviates too far from its optimal operating point, around 50%. If necessary, the electrical power generation unit is used to increase or decrease its power, respectively when the SOC is low and when it is high, and thus restore an optimum charge level to the energy storage unit.

[0016] The power setpoint of the energy storage unit is monitored, and the electrical power generation unit is used to increase or decrease its output when the power reaches its saturation point during discharge and when the power reaches its saturation point during load. Indeed, technical and economic constraints may lead to undersizing the power component of the energy storage unit relative to the power required for the primary frequency regulation (FCR) service of the electrical grid. Consequently, the power delivered or absorbed by the unit Energy storage can sometimes reach its maximum operating point and thus become saturated. In these situations, the electrical power generation unit is called upon to relieve the energy storage unit so that it returns to an acceptable power range. A compromise can therefore be reached between acknowledging the occurrence of such situations, in which it is necessary to adjust the power produced by the generation unit, and the maximum installed electrical power for energy storage or release, corresponding to the power capacity of the batteries.Studies have shown that with a storage capacity covering only 15% to 20% of the maximum power variation required for the FCR service (as contracted with the grid operator), 99% of frequency deviations occurring annually are covered, meaning the storage unit is capable of supplying the required power. Therefore, it is not economically viable to have batteries whose capacity would cover the entire expected power variation for meeting FCR service requirements simply to fulfill contractual obligations in this regard. This invention offers cost savings on battery purchases, reducing the cost of installed battery capacity by 80% to 85%.The invention makes it possible to manage marginal situations, which may be limited to only a few occurrences per year, in which the storage system proves to be undersized in power, because the power delivered by the storage system is monitored and the power of the production unit is adapted to supplement the storage system when necessary.

[0017] Taking into account the time derivative of the power produced by the electrical energy production unit makes it possible to determine the rate of change of the energy thus produced. This allows for anticipating the increase or decrease in power produced by the production unit based on the inertia of the installation, and thus responding as quickly as possible to grid frequency variations without destabilizing the production unit. Indeed, by taking this time derivative into account at the input of the hybridization controller, for example, the controller will not request an increase in power production while it is already increasing. This is to avoid accelerating production when the optimum production point is reached and a change in the production setpoint would cause this operating point to be exceeded.Indeed, the hybridization controller would then request a reduction in power output to approach the optimum operating point, leading to significant overshoots or even oscillations in the power produced by the hydroelectric unit. These oscillations must be avoided as they cause accelerated aging of electromechanical components due to repeated stress. All of these... The disadvantages are addressed by the implementation of the process, and their consequences are substantially reduced through this implementation.

[0018] The embodiments of the invention offer a simple and robust charging strategy, in particular because its principle is not to control the charging of a flywheel directly from the hydraulic energy (non-electrical energy) of a hydroelectric power plant, nor to charge batteries from the power available on the electrical power distribution network, but rather to use the electrical power generation unit for this purpose. The operation is therefore transparent, as seen from the electrical grid.

[0019] Hybridization according to the embodiments of the invention makes it possible to reduce the cumulative displacements of the actuators of the electrical power generation unit. In this case, this controller was used on a hydroelectric turbine system and demonstrated a reduction in its cumulative displacements of more than 80% compared to a non-hybridized installation. It also makes it possible to reduce the changes in direction of the corresponding mechanical systems by 96%. This significantly reduces the use and therefore the wear of the mechanical parts of the hydroelectric power generation unit. The embodiments of the invention also make it possible to improve by a factor of 10 the performance of the contribution of the power generation unit to the primary frequency control (FCR) of the electrical grid to which said power generation unit is connected.

[0020] Some preferred but not limiting aspects of the process are as follows.

[0021] In some implementations of the method, the inference processor can put in use: • three fuzzification modules using three groups of membership functions to associate fuzzy input values ​​with normalized numerical values ​​of the first input signal, the second input signal, and the third input signal, respectively, • an inference engine, implementing inference rules ([Fig.8]) linking triplets of fuzzy input values ​​from the first input signal, the second input signal, and the third input signal, respectively, to respective fuzzy output values, • a defuzzification module using membership functions to associate fuzzy output values ​​with respective numerical output values.

[0022] The signal of the setpoint of the power produced by the electrical production unit as part of its contribution to the primary frequency regulation of the electrical network can be obtained by time integration of numerical output values ​​of the inference processor.

[0023] The power produced by the electrical power production unit as its contribution to the primary frequency regulation of the electrical network can be determined on the basis of a simplified model of the behavior of said production unit as a function of the setpoint signal of the power produced by the electrical power production unit as its contribution to the primary frequency regulation of the electrical network which is applied to said production unit.

[0024] The third input signal, which is representative of the time derivative of the power produced by the electrical energy production unit, can be obtained by calculation from the simplified model.

[0025] The setpoint signal for the power produced by the electrical power production unit as part of its contribution to the primary frequency regulation of the electrical network may advantageously be of the same nature as the signal representing a deviation of the frequency of the electrical network from the nominal frequency of said network, so as to be able to be substituted for it as an input to a controller regulating the power produced by said production unit.

[0026] Either of the signal representing a deviation of the frequency of the electrical network from the nominal frequency and the signal of the power produced by the electrical power production unit as its contribution to the primary frequency regulation of the electrical network can be selectively received at the input of the power regulation controller produced by the electrical power production unit.

[0027] The invention also relates to a second object: a control device for power supplied to an electrical energy distribution network under an alternating voltage of a determined nominal frequency, by an electrical energy production unit comprising: • a controllable electrical power generation unit, having a relatively long response time to a variation in the setpoint of the power produced; • a controllable energy storage unit, having a relatively short response time to a change in the setpoint of the stored or delivered power, and used for primary frequency control (FCR: "Frequency Containment Reserve") of the alternating voltage; and, said device comprising: • a controller for regulating the energy produced by the electrical power generation unit; and, • a hybridization controller configured to manage the power generation unit and contribute to the primary frequency regulation of the AC voltage based on a signal representing a deviation of said frequency relative to the nominal frequency, by controlling the power stored or delivered by the energy storage unit based on the difference between the expected primary frequency regulation power of the generating group as its contribution to the primary frequency regulation of the electrical grid and the power produced for this purpose by the electrical power generation unit, said hybridization controller comprising at least one inference processor with at least three inputs and being configured to: • receive as input a first input signal which is representative of the state of charge of the energy storage unit; • receive as input a second input signal which is representative of the power stored or released by the energy storage unit; • receive as input a third input signal which is representative of the time derivative of the power produced by the electrical energy production unit as its contribution to the primary frequency regulation of the electrical network; • produce at output a setpoint signal of the power produced by the electrical power production unit as its contribution to the primary frequency regulation of the electrical network, said setpoint variation signal being produced using the inference processor having at least three inputs, as a function of the first input signal, the second input signal and the third input signal, respectively.

[0028] In some embodiments of the device, the inference processor may include: • three fuzzification modules using three groups of membership functions to associate fuzzy input values ​​with normalized numerical values ​​of the first input signal, the second input signal and the third input signal, respectively; • an inference engine, implementing inference rules linking triplets of fuzzy input values ​​from the first input signal, the second input signal, and the third input signal, respectively, to respective fuzzy output values; and, • a defuzzification module using membership functions to associate fuzzy output values ​​with respective numerical output values.

[0029] The hybridization controller may include an integrator that can be configured to produce the signal of the power setpoint produced by the electricity generating unit as its contribution to the primary frequency regulation of the network electrical, by time integration of numerical output values ​​from the inference processor.

[0030] The power produced by the electrical power production unit as its contribution to the primary frequency regulation of the electrical network can be determined on the basis of a simplified model of the behavior of said production unit as a function of the setpoint signal of the power produced by the electrical power production unit as its contribution to the primary frequency regulation of the electrical network which is applied to said production unit.

[0031] The third input signal, which is representative of the time derivative of the power produced by the electrical energy production unit, can be obtained by calculation from the simplified model.

[0032] The setpoint signal for the power produced by the electrical power production unit as part of its contribution to the primary frequency regulation of the electrical network can advantageously be of the same nature as the signal representing a deviation of the frequency of the electrical network from the nominal frequency of said network, so as to be able to be substituted for it as an input to a controller regulating the power produced by said production unit.

[0033] In embodiments the device may further include a selector configured to provide either the signal representing a deviation of the frequency of the electrical network from the nominal frequency and the setpoint signal of the power produced by the electrical power generation unit as its contribution to the primary frequency regulation of the electrical network, as input to the power regulation controller produced by the electrical power generation unit.

[0034] A third and final object of the invention is an electrical power generation unit adapted to supply said produced electrical power to an electrical power distribution network under an alternating voltage of a determined nominal frequency, said unit comprising: • a controllable electrical power production unit, having a relatively long response time to a variation in the setpoint of the power produced; • a controllable energy storage unit, having a relatively short response time to a change in the setpoint of the stored or delivered power (PESS), and used for primary frequency control (FCR: "Frequency Containment Reserve") of the alternating voltage; as well as, • a device for controlling the power supplied to the electrical energy distribution network according to the second object above. Brief description of the drawings

[0035] Other aspects, objectives, advantages, and features of the invention will become clearer upon reading the following detailed description of preferred embodiments thereof. This description is given by way of non-limiting example. It is made with reference to the accompanying drawings, in which: [Fig.l] is a schematic representation of a non-hybridized electricity production unit connected to an electrical power distribution network, according to existing art; [Fig.2] is a schematic representation of a hybridized power generation group, also known in the prior art; [Fig.3] is a schematic representation of a hybridized electrical production group, according to embodiments of the invention; [Fig.4] is a functional diagram of a hybridization controller according to embodiments of the invention; [Fig.5] is a graph illustrating the membership functions used in embodiments to convert representative scalar values ​​of the state of charge of the energy storage unit into fuzzy variables; [Fig.6] is a graph illustrating the membership functions used in embodiments, to convert into fuzzy variables scalar values ​​representing the power supplied (positive) or absorbed (negative) by the energy storage unit; [Fig.7] is a graph illustrating the membership functions used in embodiments, to convert into fuzzy variables scalar values ​​representing the derivative of the power produced by the electrical power production unit in relation to the frequency deviation of the network within the framework of primary frequency control (FCR); [Fig. 8] is a table illustrating inference rules implemented in a fuzzy processor according to implementations of the invention; and, [Fig.9] is a graph illustrating the membership functions used in embodiments, to convert into scalar values ​​fuzzy variables representing the variation (in frequency) of the electrical power setpoint to be produced by the electrical power production unit in relation to the frequency deviation of the network within the framework of the primary frequency control (FCR). Detailed description of implementation methods

[0036] In the figures and in the following description, the same numerical references designate identical or similar elements. Furthermore, the various elements are not drawn to scale, in order to prioritize the clarity of the figures. By Moreover, the different embodiments and variants presented are not mutually exclusive and can be combined with each other.

[0037] In the diagram of [Fig. 1], a conventional non-hybridized electrical power generation unit 1, for example a hydroelectric power plant, is symbolically represented by a hydraulic turbine 4. In the figure, the water flow through the turbine 4 is symbolically represented by two curved arrows symmetrical with respect to the central axis of the turbine (vertical axis in the example shown). The turbine 4 includes water inlet valves 5, the degree of opening of which is controllable by actuators (not shown), between a fully closed and a fully open position, to allow a variable quantity of water to pass through the turbine according to a setpoint Cl for adjusting the valves. The turbine 4 further includes a rotor 7 with blades 6, which is driven in rotation by the action of the water moving against said blades.As it rotates, the rotor 7 drives the rotation of an electric generator 8 comprising a stator armature, and thus the production of an electric current in this armature. The pitch of the blades 5 is controlled by actuators (not shown), according to an angle determined by a setpoint C2, to vary the turbine pitch, which determines the response of the blades to the water flow circulating in the turbine. This, in combination with the degree of opening of the valves 5, allows the rotational speed of the rotor to be adapted in order to obtain optimal operating conditions for the turbine 4.

[0038] The generator 8 of the production unit 1 generates electrical power, denoted PHPP in [Fig. 1], which is supplied to the electrical power transmission and / or distribution network 3, which is symbolically represented by a high-voltage pylon in [Fig. 1]. In what follows, and for short, the terms "electrical network" or "network" are sometimes used to refer to the network 3.

[0039] In network 3, electrical energy is transported in the form of an alternating electrical voltage, having a nominal frequency Fo. For example, Fo is equal to 50 Hertz (Hz) in Europe, or 60 Hz in the United States. The value of the electrical voltage (expressed in volts, V, or kilovolts, kV) depends on the nature of the lines in the network under consideration. In Europe, the extra-high voltage (EHV2) and high-voltage (HV2) lines of the transmission network carry the current at voltages generally of 400 kV or 225 kV, and at voltages of 63 or 90 kV, respectively. In the distribution network, current flows under voltages between 15 kV and 30 kV on medium voltage (MV) lines which supply small industries on a local scale, and under a voltage of 230 V or 400 V for low voltage (LV) lines which supply homes and tradespeople.A network of electrical substations distributed across the territory allows the voltage to be converted and adapted. to the different parts of the network, between the place of production and the place of consumption of electrical energy.

[0040] All high-power electrical power generation units, whether so-called "thermal" power plants (fueled by coal, oil, gas, etc.) or hydroelectric power plants like generation unit 1 considered in the example, must contractually participate jointly in the primary frequency regulation (FCR) mechanism of the network. Electricity generation unit 1 is therefore sized and controlled by a power regulation controller 15 (block labeled "HPP-CTRL" in [Fig. 1]), not only to produce and deliver electrical power Phpp to network 3 at an agreed base value, according to a Power_SP setpoint negotiated between the respective operators of generation unit 1 and network 3, but also to vary it more or less slightly from its base value according to the needs of the primary frequency regulation (FCR) of said network.An expected response time, which can be configured according to the power plants and operational data specific to each plant, is, for example, around 30 seconds for the national electricity grid in France.

[0041] To this end, and under the control of the regulating controller 15, which is driven by a signal F_Grid representing the frequency of the electrical grid 3, the non-hybridized production unit 1 of [Fig. 1] adjusts, upwards or downwards, the Phpp power that is produced and injected into the grid 3 according to the difference between F_Grid and the known nominal grid frequency Fo. In this sense, F_Grid also represents the frequency deviation of the grid 3. This occurs as soon as a deviation of the grid frequency from Fo appears, which inherently reflects an imbalance between energy produced and energy consumed at the scale of the electrical grid as a whole. Of course, the other production units comparable to unit 1, which are also connected to the electrical grid, are controlled to do the same simultaneously. Each thus contributes to the frequency regulation of the grid.

[0042] In practice, the control controller 15 of the power plant 1 shown in [Fig.1] is configured to continuously adapt the setpoints Cl and C2 of the opening of the valves 5 and the orientation of the blades 6, respectively, of the turbine 4 as a function of the power setpoint Power_SP and the signal F_Grid representing the frequency deviation of the network 3. The power control controller 15 of a conventional (non-hybridized) hydroelectric power plant as shown in [Fig.1] is designed and operates, in a classic manner, on the principle of a PID controller (for "Proportional-Integral-Differential"), by implementing a control of the setpoints Cl and C2 as a function of the power setpoint Power_SP and the signal F_Grid.

[0043] The setpoints C1 and C2 control the servomotors of turbine 4 (not shown), which regulate the opening of the valves 5 and the pitch of the turbine blades 6 in order to adjust the electrical power output of power plant 1. It should be noted that setpoint changes can occur repeatedly, depending on small variations in the grid frequency, which require a response from the power plant, each time with a relatively short response time. However, the actuation of these control devices causes wear and aging of the electromechanical parts involved, necessitating more frequent maintenance and / or replacement of parts, as the stresses on these devices due to the contribution of power plant 1 to grid frequency regulation become more frequent.

[0044] The concept of hybridizing an electrical power production unit with a locally installed energy storage unit has been proposed to overcome this drawback.

[0045] With reference to the diagram in [Fig.2], a known hybrid power generation group, as disclosed in documents FR3131473 and FR3131474 cited in the introduction, includes for example the power generation plant 1 of [Fig.1] hybridized with an electrical energy storage unit 2.

[0046] In one example, the electrical energy storage unit 2 consists of batteries capable of storing energy in electrochemical form, such as lithium-ion batteries. The storage unit 2 is connected to the electrical network 3. It is adapted to deliver or consume, according to a setpoint, a predetermined electrical power from said network, within the framework of the frequency regulation mechanism (FCR) for said network.

[0047] The hybrid power generation group of [Fig.2] always includes an energy regulation controller 15, identical to that of the power generation group of [Fig.1]. It also includes a hybridization controller 14 (block labeled "HYB-CTRL" in [Fig.2]), which is interposed between the electrical network 3 and the energy regulation controller 15. This hybridization controller 14 receives as input the F_Grid frequency deviation signal from the network 3, as well as an ESS_SOC signal representing the state of charge or SOC (from the English "State-of-Charge") of the energy storage unit 2. In response to these two input signals, it delivers as output a setpoint signal ESS_Power_SP, adapted to control the energy stored or released by the storage unit 2. It also delivers as output a F_HPP signal to control the electrical power generation unit 1, replacing the F_Grid frequency deviation signal and which is of the same nature as the latter.By the terms "of the same nature" we mean that the F_HPP control signal encodes the information adapted to the control of unit 1 in the same way as the F_Grid signal: same voltage levels, . same modulation scheme and / or coding, same communication protocol where applicable, etc. Thus, the F_HPP control signal can be substituted for the F_Grid signal in order to control the production unit 1 without requiring modification of the energy regulation controller 15.

[0048] As indicated in the introduction to this description, the F_HPP signal is derived from the F_Grid signal by "low-pass" filtering of the latter, in that the fast (high-frequency) components of the network frequency variation around its nominal frequency Fo are processed within the FCR mechanism by the charging or discharging of the storage unit 2, while only the slower (low-frequency) components are transmitted to the power regulation controller 15 in the F_HPP signal so that a response can be provided to them by the production unit 1. The power regulation controller 15 processes the F_HPP signal by modifying the setpoints Cl and C2 of the production unit 1, without any change to its usual operating mode, but on the basis of a substitute control signal, namely the F_HPP signal instead of the F_Grid signal, which represents only the slow deviations of the network frequency 3.In the diagram in [Fig. 2], this filtering of the high frequencies of the F_Grid signal to produce the low-frequency F_HPP control signal is represented by the shape of corresponding portions of these two signals as a function of time, over the same time interval denoted AT. As can be seen, the F_HPP signal now contains only the low-frequency variations of the F_Grid signal. Thus, the changes in the setpoints Cl and C2 controlling the servomotors of production unit 1, driven by the F_HPP signal, are less frequent than if they were driven directly by the F_GRID signal. The result is reduced stress on these actuators and therefore less wear on all the electromechanical components associated with the valves 5 and the turbine blades 6 of turbine 4, in the example of the hydroelectric power plant 1 shown in [Fig. 2].

[0049] Those skilled in the art will appreciate that the substitution signal F_HPP for controlling power plant 1 is suitable for controlling, under the command of the control unit 15 unchanged from that of [Fig. 1], the power supplied by power plant 1 alone. The power supplied by power plant 1, still denoted Phpp in [Fig. 2], is entirely injected into the network 3, at the connection point with the electrical grid 3. This power is therefore supplied to the electrical grid and / or to the storage system 2 (particularly when the latter is under load). Those skilled in the art will also note that the energy storage unit 2 is used to supply or consume the additional power to the power expected by the grid operator, denoted here as PGrid. This additional power, denoted PESs, is supplied to the electrical grid 3 in addition to the power supplied by the production unit 1. denoted PHPP when the latter is insufficient with regard to the power expected by the grid operator. Conversely, this PESs power is consumed by the storage unit 2 when the PHPp power supplied by the production unit 1 is greater than the PGrid power expected by the grid operator. Consequently, a person skilled in the art will note that the storage unit does not store energy from the electrical grid 3, but only stores energy directly from the production unit 1.

[0050] Now with reference to the diagram in [Fig.3], a hybrid power generation group 1-2 according to embodiments of the invention comprises an electrical power generation unit 1 and a power regulation controller 15 like those in Figures 1 and 2, and an electrical power storage unit 2 like that in [Fig.2].

[0051] More generally, unit 1 can be any controllable electrical power production unit, such as for example a hydroelectric power plant with one or more turbines such as turbine 4 shown, for example KAPLAN type turbines, but also a photovoltaic solar power plant including a photovoltaic panel farm, a wind farm including a battery of wind turbines, etc.

[0052] The electrical energy storage unit 2, for its part, may comprise any means of energy storage capable of reversibly storing electrical energy in mechanical, electrochemical, or other forms. This may include, for example, a flywheel, a set of electrochemical batteries such as a lithium-ion battery pack or supercapacitors (or "supercaps" or EDLCs, for "Electrochemical Double Layer Capacitor" because commercially available supercapacitors are manufactured using the electrochemical double-layer process), one or more fuel cells such as hydrogen fuel cells, etc., or a combination of such means. Batteries, supercapacitors, or flywheels are preferred to fuel cells because the latter are relatively less agile, i.e., they have a longer response time to a change in their energy storage setpoint.A person skilled in the art will appreciate that, in all cases, the energy storage unit 2 has a shorter response time to a change in its setpoints than the electrical power generation unit 1, which is the basis of the hybridization principle. This is indeed likely to satisfy the expectations of the electricity grid operator for the efficiency of the primary frequency regulation of said grid.

[0053] Regardless of its technology, which may be multiple, the electrical energy storage unit 2 is coupled to the electrical network 3 and is adapted to deliver, according to a setpoint, a predetermined electrical power to said network 3. It can also to consume electrical energy from the electrical grid 3, within the framework of the implementation of the FCR mechanism for frequency regulation of said grid. Just like unit 2 of the hybrid power generation group in [Fig. 2], which is part of the prior art, the electrical energy storage unit 2 in [Fig. 3] can also be charged, in certain cases, by surplus electrical energy produced by the power generation unit 1 of the generation group to which it belongs. Its control, however, is no longer based here on the principle of filtering the F_Grid signal representing the frequency drift of the electrical grid 3, as was the case in the prior art illustrated by [Fig. 2]. This difference will become clearer from the description of embodiments of the invention that follows.It is the source of the improvement provided by these embodiments compared to a hybrid power generation group according to [Fig.2], namely an even greater limitation of the setpoint variations of the power generation unit 1 compared to this prior art. This improvement can be seen from the point of view of the reduction of stresses on the electromechanical components of the generation unit 1, which participate in the regulation of the power it produces, in particular for the needs of the primary frequency control (FCR) of the electrical network 3.

[0054] Furthermore, and like that of [Fig. 2], the hybrid power generation unit of [Fig. 3] includes a hybridization controller 34 (block still labeled "HYB-CTRL" in [Fig. 3]), which is interposed between the electrical network 3 and the power regulation controller 15, in the same way as the hybridization controller 14 of [Fig. 2]. Like the latter, the hybridization controller 34 according to embodiments of the invention is therefore operationally interposed between the electrical network 3 and the power regulation controller 15, to manage the operation of the hybridization of the generation unit 1-2, in particular with regard to the contribution of said unit to the primary frequency regulation of the network, which is the rationale for this hybridization.In other words, the hybridization controller 34 receives as input (among other things) the F_Grid signal representing the frequency deviation of the electrical grid 3, as well as the ESS_SOC signal representing the state of charge (SOC) of the energy storage unit 2. It also receives from the energy storage unit 2 a PESS signal which represents the power delivered to or drawn from the electrical grid 3 by said storage unit 2. It should be noted that the values ​​of this signal are algebraic values, that is to say, they can be positive (in the case of power delivered to the grid 3) or negative (when power is drawn from the grid 3). The hybridization controller 34 communicates with the other equipment of the generating group digitally or analogically.

[0055] Thus, the F_Grid signal representing the frequency deviation of the electrical network 3 results from a precise measurement (on the order of millihertz, mHz) of the frequency of the alternating voltage system (typically 50 Hz or 60 Hz, or any other standardized frequency value) on the electrical network 3 to which the production group 1-2 is connected.

[0056] The ESS_SOC signal, representing the state of charge (SOC) of the storage unit 2 and resulting from a measurement of the charge level of said unit 2, can, for example, be a numerical value expressing a percentage, in this case the relative level of the electrical charge state of unit 2 between the lowest level (unit 2 fully discharged) and the highest level (unit 2 fully charged). In addition to this ESS_SOC signal, the hybridization controller 34 can also, in embodiments not shown, receive other measurements from the energy storage unit 2, such as a power measurement, or information such as alarms, allowing monitoring of said unit 2, possibly with a display of said information intended for a monitoring operator.

[0057] The PESS signal also results from a measurement of the power actually delivered to or drawn from the electrical network 3 by the energy storage unit 2. This measurement can be transmitted to the hybridization controller 34 in the form of a numerical value. For this purpose, one can, for example, use an industrial communication protocol of the Modbus type, for example Modbus-RTU (for "Remote Terminal Unit"). The invention is not intended to be limited to this example; any other type of digital or analog coding can also be used according to the specific requirements of each application.

[0058] In response to these three input signals, namely F_Grid, ESS_SOC, and PESs, the hybridization controller 34 outputs a control signal F_HPPn for the power produced by the power generation unit 1 as part of its participation in the primary frequency control (FCR) mechanism of the electrical grid 3. This signal varies one or more setpoints of the power generation unit 1, such as the setpoints C1 and C2 of the turbine 4 in the example shown, so that said unit 1 adapts its nominal power output based on the power setpoint Power_SP, to contribute appropriately to the FCR mechanism. The setpoint signal F_HPPn (or control signal for the power produced by the power generation unit 1) can be used as a substitute for the frequency deviation signal F_Grid, since it is of the same nature.By the terms "of the same nature" we mean here again that the control signal F_HPPn (like the signal F_HPP of [Fig.2]) encodes the information adapted to the control of the electrical power production unit 1 in the same way as the signal F_Grid: same voltage levels, same modulation scheme and / or coding, same protocol of. communication where applicable, etc. Thus, the F_HPPn control signal can be substituted for the F_Grid signal to control the generating unit 1 without requiring modification of the power regulation controller 15. It follows that the F_HPPn control signal allows the hydroelectric generating unit 1 to be controlled via the same power regulation controller 15 as that of the power generation units in [Fig. 1] and [Fig. 2], without modification thereof. This is not mandatory, but is a significant advantage insofar as the invention can be deployed without substantial modification of the already installed components of the generating unit.

[0059] In practice, the control signal F_HPPn can be an analog signal, such as a sinusoidal voltage whose frequency varies according to the setpoint calculated by the hybridization controller 34. But it can also be in another form of analog or digital information, provided that it is (preferably, as indicated above) of the same nature as the signal F_Grid which it replaces.

[0060] In some embodiments, a selector 151 allows, for example when hybridization is temporarily undesired and / or when maintenance of the energy storage unit 2 needs to be carried out, the direct supply of the input frequency signal F_Grid, instead of the substitute control signal F_HPPn, to be restored to the corresponding input of the power generation controller Phpp 15. This allows the system to continue operating with the generating unit without the energy storage unit 2, as would be the case with the existing installation shown in [Fig. 1]. It should be noted that the functionality described above with respect to the selector 151 is advantageous but by no means mandatory.

[0061] However, those skilled in the art will appreciate, from the description of implementation methods that will be given with reference to [Fig. 4] and the following figures, that the F_HPPn control signal of the power regulation controller 15 differs, from the point of view of its informational content, from the F_HPP control signal generated by the hybridization controller 14 of [Fig. 2]. Indeed, and unlike the said F_HPP signal, the F_HPPn control signal of the power regulation controller 15 generated by the hybridization controller 34 according to embodiments of the invention, does not result from a low-pass filtering of the F_Grid signal representing the frequency deviation of the network 3. Its informational content is more subtle, and incorporates a trend-based notion into the management of the electrical power generation group 1.2.In the implementations of the hybridization controller 34 proposed by the invention, this trend notion is dealt with by a Fuzzy Logic approach.

[0062] Consequently, the hybridization controller 34 according to embodiments differs substantially from the hybridization controller 14 conforming to art. previous according to [Fig.2], as will now be described in detail. It allows for even better optimization of the stress on the electromechanical components of production unit 2, resulting in a further reduction of fatigue and mechanical aging of this equipment.

[0063] The hybridization controller 34 outputs and sends, for example digitally, its power setpoint PESs_SP to the energy storage unit 2. The PEss_SP signal can then conform, for example, to the aforementioned Modbus industrial communication protocol. Alternatively, the PESs_SP setpoint can be transmitted in analog form, for example by a current control in the range of 4-20 mA. Regardless of the practical implementation, the PEss_SP signal functions to control the energy storage unit 2, that is, to control the PESS power that it consumes and temporarily stores from the grid, or that it returns to the electrical grid 3, in both cases for the purposes of the primary frequency control (FCR) of said grid 3.

[0064] The two sources of electrical energy delivered to the electrical grid 3 by the generating unit, namely the generating unit 1 and the storage unit 2, each deliver an electrical power to said grid that is a function of the instructions they receive from the hybridization controller 34, namely the instruction F_HPPn and the instruction Pess_SP, respectively. The power PEss supplied or consumed in the grid 3 by the energy storage unit 2, based on the PEss_SP instruction calculated by the controller 34, is added to the power PHpp_fcr supplied by the generating unit 1 as its contribution to the primary frequency regulation (FCR) mechanism, in order to have a total power PFCr injected into the grid 3 that is related to the frequency deviation of said grid reflected by the signal F_Grid.As explained in the introduction, the electrical power generation units involved in primary frequency regulation are primarily sized and configured to produce a nominal power output. In the diagram in [Fig. 3], this is symbolized by a nominal power setpoint, Power_SP, shown in the lower left of the figure. This setpoint is supplied as input to the power regulation controller 15. This setpoint is modulated by said controller 15 according to the frequency deviation observed at the level of the electrical grid 3, and indicated by the signal F_Grid. According to the principle of hybridization of the electrical power generation group, however, it is not the F_Grid signal that directly controls the generation unit 2 from the point of view of the FCR mechanism, but the F_HPPn signal generated by the hybridization controller 34.

[0065] A person skilled in the art will appreciate that, on hydroelectric power plants, the nominal power setpoint Power_SP can be either an electrical power setpoint to produce either a setpoint for the water flow rate to be turbined. Indeed, on dams of the "in-river" type, equipped for example with Kaplan turbines, one objective may be the regulation of water levels upstream and / or downstream of the installation in order to ensure the navigability of the river, the availability of water for crop irrigation and / or for water sports on reservoirs, etc. To meet this objective, the hydraulic production units are, in this case, controlled according to a nominal setpoint of the type of water flow rate to be turbined ("Flow rate set point").

[0066] In all cases, the power produced by the electrical production unit, as seen from the electrical grid, is the algebraic sum of the powers produced by the production unit 1 and by the storage unit 2: [Math. 1] PGrid = Phpp + PESS (equation 1) with: • PGnd: power supplied to the electrical network 3 by the electrical energy production unit 1 and the storage unit 2; • PHpp: power supplied by the electrical power production unit 1 as illustrated by equation 2; • Pess: power supplied or absorbed by the energy storage unit 2.

[0067] Furthermore, the PHPP power supplied by the electrical power generation unit 1 is given by: [Math. 2] Phpp = Ppower_sp + Phpp_fcr (equation 2) with: • Ppower_sp: power produced relative to the power setpoint Power_SP; and, • Phpp_fcr: power variation of said power produced as a function of the frequency deviation with respect to the setpoint F_HPPn.

[0068] However, in the primary frequency control (FCR), the power corresponding to this control is only the portion of the PGrid power that is related to the contribution of the grid-connected generating units to the overall response to the frequency deviation of the electrical grid 3. For the generating unit considered here, the contribution to this regulating power is the algebraic sum of the PHPP_FCR fraction of the PHPP power produced by the generating unit 1 as its contribution to the FCR mechanism and which is related only to the frequency deviation of the grid (independently of the nominal power setpoint Power_SP of the generating unit in question), on the one hand, and the PESS power which is supplied or consumed by the energy storage unit 2, on the other hand, which is used only for the purposes of the primary frequency control (FCR): [Math. 3] Pfcr = Phpp_fcr + Pess (equation 3) with : • PFCr: primary frequency control power, i.e. the total power corresponding to the contribution of the production group intended to compensate for the frequency deviation of the network 3. Note that PFcR can be positive when it is necessary to raise the network frequency towards Fo (i.e., when the overall production is less than the total consumption on the network 3), but can also be negative when the network frequency is greater than Fo (i.e., when the overall production is greater than the total consumption on the network) and it is necessary to lower it by storing energy in the storage unit 2; • Phpp_fcr: power supplied by production unit 1 linked only to the primary frequency setting and with regard to the setpoint F_HPPn; • Pess: power supplied or consumed by the energy storage unit 2, which can be positive or negative (for the reasons indicated above concerning PFCr) and in relation to the PFSs_SP setpoint.

[0069] The primary frequency control power PFCR is contractually agreed upon between the electricity network operator 3 and the energy producer, who is the operator of the electricity generation unit. A coefficient (KFCR), the value of which may be linked at least in part to the nominal power setpoint Power_SP, but not necessarily and / or not solely to this setpoint, is fixed and contractually agreed upon between the two parties: [Math. 4] PFCr = KFCr x AF (equation 4) with: • Pfcr: primary frequency control power, expected from the production group by the electricity network operator 3; • KFCr: contractual coefficient agreed between the producer and the network operator 3; • AF: difference between the nominal frequency and the network frequency at time t.

[0070] The control of the two energy sources of the production group (i.e., the production unit 1 and the storage unit 2) is carried out under the control of the hybridization controller 34 in order to satisfy at each instant the operating equations (1), (2), (3) and (4) above.

[0071] With reference to the block diagram in [Fig.3], the main functional elements of the hybridization controller 34 are as follows.

[0072] The hybridization controller 34 includes a first block 341 (denoted "FCR" in the figure) which calculates, at time t, the target PFCr power that is expected by The electricity grid operator contributes to the primary frequency regulation (PFR) contribution from the generating unit in question. Block 341 receives as input the F_Grid signal, representing the grid frequency deviation and the KFCR coefficient contractually agreed upon with the producer. It outputs a setpoint corresponding to the primary frequency regulation (PFCr) power, i.e., the expected power output from the generating unit (see equation 4 above), independent of the nominal power setpoint (Power_SP).

[0073] The hybridization controller 34 includes a second block 342 representing a subtraction operator, to which the PFCR setpoint is supplied as input as delivered by the block 341 above. More specifically, the PFCr setpoint is delivered on the positive "+" input, or additive input, of the operator 342.

[0074] A third block 343 of the hybridization controller 34, labeled "ESS_Mngt" in [Fig. 3], is a management module for the energy storage unit 2. Those skilled in the art will understand that this module 343 generates the F_HPPn control signal for the power regulation controller 15, which in turn controls the production unit 1 via the setpoints C1 and C2 based on this signal and also on the nominal power setpoint Power_SP, as previously stated. The management module 343 receives as input the three signals ESS_SOC, PFSSet, and the time variation dPHPP_FCR / dt of the power PHpp_fcr, and outputs the F_HPPn control signal for the electrical power production unit 1. Embodiments of this management module will be described below.

[0075] Still referring to [Fig. 3], the hybridization controller 34 finally includes a fourth block 344, which represents a time-domain model of the hydroelectric generating unit 1. This model 344 allows the power PHpp_fcr produced by unit 1 to be calculated based on the setpoint F_HPPn sent by the control module 343. In the example, model 344 is a numerical model of the behavior of the electrical generating unit 1, for example, here a hydroelectric unit 1. This numerical model is, for example, a first-order model based on a 3D map, making it easy to obtain an estimate of the power component PHpp_fcr produced by unit 1 in relation to the setpoint signal F_HPPn. It should be recalled that the power PHpp produced by generating unit 1 has two components (see equation 2 given above): • a major component linked to the nominal power setpoint Power_SP, corresponding to the electrical power to be produced or the turbine flow rate, and which generally represents more than approximately 90% of the power actually produced by unit 1; and, • a minority component PHpp_fcr linked to the frequency deviation of the electrical network 3, which corresponds to the contribution of the production group considered to the primary frequency regulation mechanism, and which generally represents less than about 10% of the power actually produced by unit 1.

[0076] However, it is difficult to distinguish these two components in a measurement of the electrical power PHpp produced by the hydroelectric unit 1 that could be carried out in practice. In other words, it is difficult to extract the aforementioned minor component PHpp_fcr. Indeed, the use of available filtering techniques lacks robustness because any measurement of the actual PHpp power is highly disturbed and noisy. In hydraulic installations, the flow variations induced by the primary frequency control (FCR) mechanism generate shock and pressure waves which, through echoes, generate new power variations in addition to those sought by the primary frequency control. As a result, a power measurement is highly noisy and difficult to use.Determining an estimated value for the Phpp_fcr power, corresponding solely to the generating unit's contribution to the primary frequency control (FCR) through its F_HPPn setpoint, based on model 344, provides a solution to this problem. Those skilled in the art will appreciate that the problem mentioned above may not arise for some electrical power generation units, or may be less pronounced. Therefore, using model 344 is an advantageous option, but by no means essential.

[0077] In all cases, a measurement or, as here, an estimate of the value of the minority component PHpp_fcr of the power produced by generating unit 1 as part of the generating group's contribution to primary frequency control (FCR mechanism) is delivered to the negative input (subtractive input) of the subtraction operator 342. Thus, and based on equation 3 given above, which links the power PFcr to the power PHpp_fcr and to the power PEss, the operator 342 can deliver the PEss_SP setpoint for controlling the power PESS of the storage unit 2. As will be understood, this PEss_SP setpoint is positive if the power PECR for the generating group is greater than the power PHpp_fcr produced by unit 1 as part of the FCR mechanism, so that the latter must be supplemented by energy returned by the storage unit 2.Conversely, the PESs_SP setpoint is negative if the total PECR power for the production group is less than the PHpp_fcr power produced by unit 1 under the FCR mechanism, so that the latter must be reduced by energy storage in unit 2.

[0078] In other words, the power setpoint PESs_SP is addressed by the hybridization controller 34 to the storage unit 2, and it is the result at the output of the operator 342: it is the difference between the power PFCr expected from the generating group by the network operator for the primary frequency control (FCR) mechanism for the electrical network 3, on the one hand, and the estimated value of the power PHpp_Fcr produced by the generating unit (see equation 2 above) under said primary control, on the other hand.

[0079] According to the embodiments of the invention, and contrary to the case of the low-pass filtering hybridization of the F_Grid signal implemented in the known production group, illustrated in [Fig.2] previously described, the load management strategy of the energy storage unit 2 does not have the effect of controlling the charge or discharge of said unit 2 on the electrical network 3 without any consideration other than the power that should be supplied or, on the contrary, absorbed, respectively, in application of the primary frequency control (FCR) mechanism.

[0080] On the contrary, the invention proposes to use, opportunely, an established configuration of the production unit 1 in which it is located but which should be modified due to the evolution of the F_Grid signal representing the frequency deviation of the network, possibly maintaining this configuration, depending on the state of load of the storage unit 2. In particular, if the production unit 1 is configured via its setpoints Cl and C2 so as to produce more energy than becomes required by the FCR mechanism due to an increase in the frequency of the network 3 reflected by the F_Grid signal, the configuration which has become "oversized" from the point of view of the FCR mechanism alone, can be maintained, in order to load the storage unit further when its state of load (SOC) is less than 100%, and all the more (i.e., all the more time) if its state of load is less than 50%.In reality, the "oversized" configuration of production unit 1 can be maintained, more or less, as long as the state of charge (SOC) of storage unit 2 remains below an optimum of approximately 70%, i.e., as long as said unit 2 can absorb the excess electrical energy produced by unit 1. Other thresholds on the state of charge (SOC) can also be used.

[0081] In other words, the strategy implemented in module 343 allows, when the energy storage unit 2 is discharged and when the power generation unit 1 produces more power than demanded, for example when the grid frequency is below 50 Hz but is increasing, for the production setpoint PHpp_SP of power generation unit 1 to remain at its operating point instead of being lowered, and for this surplus power produced (considered as the "surplus" relative to the power PFCr_hpp expected by the grid 3) to be used Recharge storage unit 2 as long as it can accept more energy to store. Thus, if the energy demand for the primary grid setting were to reverse again before storage unit 2 is fully charged, the configuration of generating unit 1, which has been maintained, can become suitable again. In such a scenario, the setpoints Cl and C2 of turbine 4 in hydroelectric power plant 1 were not modified. This prevented damage to the actuators of valves 5 and rotor blades 6 of rotor 7, extending their operational life.

[0082] Of course, the same strategy according to implementations of the invention leads, conversely, if the production unit 1 is in a configuration in which it does not generate, or generates only a little PHpp_fcr power as part of its contribution to the FCR mechanism, for example because the network frequency is lower than the nominal frequency Fo, but the evolution of the signal F_Grid indicates or suggests that this configuration should be modified, to maintaining the configuration of the production unit (i.e., to keeping its Cl and C2 setpoints unchanged when it would be appropriate to raise them by strict application of the FCR mechanism), if it also turns out that the storage unit 2 is at a high state of charge SOC, or even as long as its state of charge is not considered too low, i.e. not too close to 0%.In this scenario as well, we avoid adjusting the Cl and C2 setpoints of the production unit, which may prove unnecessary in the event of a further trend reversal.

[0083] In summary, the load management strategy of the energy storage unit 2 implemented in module 343 consists of conditional timing, based on the current state of charge of said storage unit 2, of the adaptation of the operating setpoints of the electrical power production unit 1. As will be understood, this is achieved by module 343 via the generation of the signal F_HPPn, which induces the power setpoint PEss_SP of the storage unit 2 by means of equation 3, which is verified at the level of the subtraction operator 342. The technical effect obtained by this conditional timing strategy of the variations of the signal F_HPPn, which controls unit 1 within the framework of its hybridization with the storage unit 2, is that unit 1 is controlled with a virtuous optimization of the controlled variation of its operating point.This allows for the advantageous operation of energy storage unit 2 across its entire operating range, i.e., with a state of charge (SOC) that frequently and more broadly fluctuates within the range from approximately 0% to approximately 100%. The result of operating it more regularly and with greater charge and discharge amplitudes is a reduction in the degradation of its effective storage capacity. Furthermore, and simultaneously, the average aging of the mechanical components regulating energy production by generating unit 1 is further reduced. In effect, they are potentially called upon less often for modifications which, otherwise, would be called upon more frequently in time by the natural evolution of the frequency deviation of the network 3. Put another way, the ballast capacity of the energy storage unit 2 is now fully used, which is not the case with control by low-pass filtering of the F_Grid signal in implementations conforming to [Fig.2].

[0084] Those skilled in the art will appreciate that the notion of "trend" mentioned above in relation to the evolution of the grid frequency deviation can be deduced from observing the derivative, with respect to time t, of the power PHpp_fcr produced by the generating unit 1 for the needs of the primary frequency control mechanism of the electrical grid 3. This derivative dPHpp_FCR / dt is easier to obtain by calculation, particularly from model 344, which is a simplified model, than a trend in the evolution of the signal F_Grid representing the grid frequency deviation, which would have to be obtained by measurements in said signal. Therefore, in some embodiments, the control module 343 receives as input the derivative with respect to time dPHPP_FCR / dt of the power PHpp_FCRt as estimated by model 344 of the electrical power generating unit 1.This implementation is particularly advantageous because PHpp_fcr is already calculated from model 344. Calculating its derivative dPHPP_FCR / dt is therefore easy, and it is particularly beneficial to do so via model 344 because such a derivative would be much more difficult to obtain, sensitive, and imprecise if the power PHpp_FcR were obtained through physical measurement.

[0085] Furthermore, another lesson of the present invention is that the conditions and the extent to which the timing of the modifications of the signal F_HPPn supplied as input to the controller 15, which drives the production unit 1, is achieved do not need to result from highly sophisticated statistical analyses, which are, moreover, difficult to perform even with a deep learning mechanism. On the contrary, an approach to defining the modifications of the signal F_HPPn that are necessary to implement more dynamic management of the energy storage unit 2, as described above, relies instead on experimental designs based on understanding the tangible phenomena behind equations (1), (2), (3), and (4), not only at each instant but also, and especially, in their evolution observed over a certain observation time window.In the real world, we never have access to all the data we need to map every possible relationship between the different variables involved. But we can identify some key strategies for isolating and exploring the mechanisms between some of these variables. This is what is proposed with the implementation methods of the fuzzy logic-based management module 343, which will now be described.

[0086] With reference to the functional diagram in [Fig. 4], the control module 343 can be implemented as a fuzzy logic processor 41 (also called a "fuzzy inference processor" or simply an "inference processor"). This computer is adapted to deliver one output variable, namely the F_HPPn control signal for the production unit 1, to be supplied as input to the power regulation controller 15, and to receive three input variables, namely: • the ESS_SOC signal representing the state of charge (SOC) of the energy storage unit 2; • PESS information representing the electrical power delivered or absorbed by the energy storage unit in the electrical network 3; and, • the derivative with respect to time dPHpp_EcR / dt of the power PHpp_FCRtelle as estimated by model 344 shown in the diagram of [Fig.3].

[0087] In some embodiments, the inference processor 41 receives as input the triplet formed by the signals ESS_SOC, the power measurement PEss of the storage unit 2 and the time derivative dPHpp_EcR / dt of the power PHpp_fcr- The relevant numerical values ​​are normalized, for example: • between 0 and 1 for the ESS_SOC signal, the normalization resulting from the ratio between the energy stored in the storage unit and the maximum storage capacity of the storage unit; • between -5 and 5 for the PESS power delivered or absorbed by the energy storage unit 2, this normalization resulting from the ratio of the PESS power supplied or absorbed to the maximum power of the hybridized and contracted generation unit with the grid operator PEcR_max. In our example, these values ​​of -5 and +5 for the normalized variation range for the PESS power derive from the ratio between the maximum power of the storage unit 2 (for example, 650 kW) and the maximum power of the generation unit (for example, 3 megawatts, MW). It should be noted that the values ​​of PEss can also be negative when the PESS power is absorbed for storage (see equation (1) when PG rid must decrease at a constant value of Phpp); and, • between -1 and +1 for the time derivative dPHpp_EcR / dt of the power PHpp_fcr expressed in W / s, the normalization resulting from the ratio of the power variation measured in W / s to the maximum power variation. It should be noted that the derivative dPHpp_EcR / dt can take negative values ​​when the power PHpp_Fcr decreases.

[0088] A person skilled in the art will appreciate that the indicator, the PESS power delivered or absorbed by the energy storage unit 2, reflects the power level demanded from the batteries. Between -1 and 1, the system is within its nominal operating range; beyond this range, it is in a saturation range, and the production unit must be used because the storage unit is saturated. If the batteries were not undersized, in the sense that the entire power demanded by the grid operator for frequency regulation could be supplied by the batteries (i.e., in 100% of cases), then the ratio would be strictly between -1 and +1, and consequently, the production unit would never be required to provide power to the storage unit. Ultimately, the operating parameters of the production unit would only need to be adjusted to correct the state of charge of the energy storage unit.

[0089] The input values ​​thus normalized are expressed in the form of a scalar (a signed scalar, where appropriate) in a so-called "unit basis", that is to say, relative to the unit or "per unit" or "pu" ("per unit" in English). They are denoted ESS_SOC_pu, PEss_pu and dPHpp_EcR_pu in the following and in Figures 5, 6 and 7, respectively.

[0090] The steps of the process implemented in the fuzzy inference processor 41 are as follows: • a fuzzification step of the normalized input values, which is executed by the fuzzification modules 41la, 411b and 41le in order to obtain a symbolic representation (in the form of linguistic symbols) of the numerical input values, namely the normalized values ​​ESS_SOC_pu, PESs_pu and dPHpp_EcR_pu, respectively; • an inference rule evaluation step (or inference step) which is executed by block 412 of processor 41, and which consists of determining the fuzzy partition of the output value by combining the fuzzified inputs using fuzzy operators and a set of inference rules; and, finally, • a defuzzification step of the results of the inference step, which is executed by the defuzzification module 413 and which allows, from a result in symbolic representation, to obtain a numerical output value noted HPP_AF.

[0091] The graphs shown in [Fig. 5] are an example of membership functions that are implemented in the fuzzification module 41 of the fuzz processor 4L. These membership functions assign to each numerical value ESS_SOC_pu of the state of charge of the energy storage unit 2 (expressed in pu) that is received as input, a more easily understood symbolic description or representationmanipulable thanks to the fuzzy logic paradigm. More specifically, membership functions describe the degree of membership of the symbols in question as a function of the numerical value of the input. [Fig. 5] thus presents an example of membership functions that map the numerical value of the charge state of storage unit 2 to different symbols (also called "fuzzy symbols" or "fuzzy values") that allow this input to be described. In the non-limiting example as shown, there are five of these symbols, which can be defined through the membership functions shown from left to right in [Fig. 5].[5]: "VeryLow" when the received input numerical value ESS_SOC_pu is considered to mean that the SOC of unit 2 is very low; "Low" for an SOC considered low; "OK" for an SOC considered optimal; "High" when the SOC is considered high; and "VeryHigh" when it is considered very high. The membership functions are identified in [Fig. 5] by numbers 1 to 5 in a small circle, in that order from left to right in the figure. In the example shown in [Fig. 5], they have the classic shape of trapezoids. This shape is not mandatory, however. Any other shape, for example Gaussians ("bell-shaped" functions), may be preferred, even though trapezoids have the advantage of simplicity. In practice, these tables can be stored as one-dimensional value tables, with one table for each of the membership functions.Each table gives a value (between 0 and 1) of the degree of belonging (also called degree of truth) to the symbol in question, of the numerical value ESS_SOC_pu received as input.

[0092] Thus, for example, if the received normalized numeric value ESS_SOC_pu is equal to 0.18, the membership functions ® and ® associate with this normalized numeric value the fuzzy value "VeryLow" and the fuzzy value "Low" with a respective membership degree of 0.5 in both cases (meaning that the membership of said normalized numeric value to the symbol "VeryLow" and to the symbol "Low" is true in both cases to a degree of 50%). The membership degree of the numeric value ESS_SOC_pu = 0.18 to the other fuzzy symbols is zero. According to another example, the membership function ® associates the fuzzy value "OK" with the numeric values ​​ESS_SOC_pu = 0.4, ESS_SOC_pu = 0.5, and ESS_SOC_pu = 0.6 if they are received as input, and the four other membership functions associate their associate the four other fuzzy symbols / values ​​with each having a degree of truth that is zero.

[0093] The five graphs in [Fig. 6] similarly show an example of membership functions that are implemented in the fuzzification module 411b of the 4L fuzz processor. These membership functions, of which there are nine in this example, associate each normalized numerical value PESs_pu of the power of the energy storage unit 2 (expressed in pu) received as input, nine fuzzy symbols or values ​​which are for example, respectively, and from left to right on the [Fig.6] : "ChargeSatVeryHigh" when the PESS power is saturated and charges storage unit 2 very heavily; "ChargeSatHigh" when the PESS power is saturated and charges storage unit 2 heavily; "ChargeHigh" when the PESS power charges storage unit 2 heavily without saturation; "ChargeLow" when the PESS power adequately charges storage unit 2 adequately without saturation; "Zero" when the PESS power is negligible; "DischargeLow" when the PESS power adequately discharges storage unit 2 without saturation; "DischargeHigh" when the PESS power charges storage unit 2 heavily without saturation; "DischargeSatHigh" when the PESS power is saturated and discharges storage unit 2 heavily; and finally "DischargeSatVeryHigh" when the PESS power is saturated and discharges storage unit 2 very heavily.The number and form of the membership functions in [Fig.6] are given as a non-limiting example.

[0094] The graphs in [Fig. 7] show an example of five membership functions implemented in the fuzzification module 41 of the fuzz processor 41 to associate the normalized numerical values ​​dPHpp_EcR_pu, representing the variation in power PHpp_fcr of the hydroelectric unit 1 received as input, with the following five fuzzy symbols and an associated degree of membership: "DecreaseHigh" when the power PHpp_fcr decreases rapidly; "DecreaseLow" when the power PHpp_fcr decreases slowly; "Zero" when the power PHpp_fcr does not change; "IncreaseLow" when the power PHpp_fcr increases slowly; and "IncreaseHigh" when the power PHpp_fcr increases rapidly. The number and shape of the membership functions in [Fig. 7] are given by way of non-limiting example.

[0095] The inference rules implemented in block 412 are rules that express the influence of the fuzzy input variables of said block on the value of its fuzzy output variable. These logical rules allow the inputs to be combined and the output value of the fuzzy processor to be defined. They are expressed as an enumeration of N logical implication rules, where N is an integer greater than one (7V>1), which rules are of the following type: "Rule Rn: if "Condition ni" on input 1 and / or "Condition n2" on input 2 and / or "Condition n3" on input 3 then "Action n" on output" And this applies to n between 1 and N.

[0096] In the jargon of a person skilled in the art, for each of the N rules, the above conditions are called "antecedents" and the resulting action is called the "consequence" of the rule. The evaluation of these N different rules is done using the operator "OR" logic. Indeed, such an enumeration is understood, in accordance with common language practice, to mean: "If... and / or if... then... (Rule 1), or If... and / or if... then... (Rule 2), or (...) or If... and / or if... then... (Rule n), or (...) or If... and / or if... then... (Rule N)”.

[0097] Such rules define an inference engine, also called a fuzzy model, for example a model known as a Mamdani-type model, in which the antecedents and consequences of the rules are fuzzy propositions.

[0098] By way of example, the table in [Fig. 8] gives a non-exhaustive list of inference rules which, in an example, can be evaluated in the inference engine 412 of the fuzz processor 41 of [Fig. 4]. This list is partial because, in principle, one can have up to 225 rules combining the five possible fuzz values ​​of the first input ESS_SOC_pu, the nine possible fuzz values ​​of the second input PESS_pu and the five possible fuzz values ​​of the third input dPHpp_EcR_pu, which fuzz values ​​were presented above with reference to Figures 5, 6 and 7, respectively. In this table, the first three columns give the fuzz values ​​of the three inputs (i.e., the antecedents), respectively ESS_SOC_pu, PESS_pu and dPHpp_EcR_pu after fuzzification by blocks 411a, 411b and 411e, respectively.The fifth and final column shows the fuzzy output values ​​impacted by the evaluation of each rule presented here, when said rule is verified in the output of the 412 inference engine (i.e., the consequences). Those skilled in the art will appreciate that the fuzzy values ​​of the input variables ESS_SOC_pu, PESs_pu, and dPHPp_ECR_pu, shown in the first three columns, and the fuzzy values ​​corresponding to the output variable HPP_AF, shown in the fifth column, are actually fuzzy logic symbols of said variables, each weighted by a certain degree of truth (which is not apparent from reading the table itself). Finally, to make the example clear, the fourth column of the table contains, for each of these rules, a comment in natural language on the expected effect on the system of the combination of fuzzy inputs and fuzzy outputs.As the person in the trade will have understood, each rule consists of a "fuzzy" reasoning which matches each combination of fuzzy values ​​of the input variables (the antecedents) with an implication or conclusion about the fuzzy values ​​of the output variable (the consequence) with a degree of truth of the value. the vagueness of each output (the consequence) conferred upon it by the respective degrees of truth of the associated antecedents.

[0099] The graphs in [Fig. 9] show an example of membership functions that are implemented in the defuzzification module 413 of the fuzzy processor 41. These membership functions, of which there are eleven in this example, map the numerical value of the HPP_AF value at the output of the fuzzy processor 41 to the different symbols that allow this output to be described. It should be noted that the HPP_AF value (expressed in mHz / s) at the output of the fuzzy processor 41 corresponds to the expected variation of the primary frequency control (FCR) setpoint F_HPPn of the electrical power generation unit 1, which is a function of the values ​​of the input variables ESS_SOC_pu, PESS_pu and dPHPP_FCR_pu supplied to the input of said controller 41. The eleven symbols or fuzzy values ​​that are respectively associated with the eleven membership functions represented from left to right in [Fig. 9]9], for example: "DecreaseSuperHigh" when the setpoint variation decreases very, very quickly; "DecreaseVeryHigh" when the setpoint variation decreases very quickly; "DecreaseHigh" when the setpoint variation decreases quickly; "DecreaseMed" when the setpoint variation decreases moderately; "DecreaseLow" when the setpoint variation decreases slowly; "Zero" when the setpoint variation shows no change; "IncreaseLow" when the setpoint variation increases slowly; "IncreaseMed" when the setpoint variation increases moderately; "IncreaseHigh" when the setpoint variation increases rapidly; "IncreaseVeryHigh" when the setpoint variation increases very quickly; and finally "IncreaseSuperHigh" when the setpoint variation increases very, very quickly.These eleven membership functions also correspond to the eleven possible fuzzy output values ​​for the fifth and final column of [Fig. 8]. The weighting of the eleven fuzzy output variable values, along with their associated degree of truth, following the evaluation of the N RI to RN inference rules, is handled by the aggregation operation, implemented by the fuzzy inference engine 412 in [Fig. 4]. Finally, the transition from the fuzzy result of the fuzzy inference engine 412 to the numerical value of the setpoint variation is performed in the defuzzification step implemented by the defuzzification module 413 of the 4L fuzzy processor. This step transforms the result of the aggregation of the fuzzy implications from the evaluation of the N RI to RN rules in symbolic form into a single numerical output value, denoted HPP_AF in [Fig. 4].This numerical value can be obtained via the calculation of the center of gravity or any other method, applied to the output result of the aggregation operation. In other words, the numerical value delivered at the output of the defuzzification module 413 is . a quantitative value that allows us to represent the combination of fuzzy output values ​​shown in the fifth column of the table in [Fig.8].

[0100] From a practical standpoint, in the embodiment proposed here, a Mamdani-type implementation is used, for example, to implement the inference engine 412, the aggregation of the inference engine results, and then the defuzzification step 413. Thus, during the implementation of each inference rule of 412 that is considered true for a combination of inputs, a non-zero truth value is associated with the fuzzy output variable corresponding to two fuzzy input variables having non-null truth degrees for the rule in question. The truth degree of the output variable can, for example, correspond to the minimum between the two truth degrees of the two non-zero input variables.This non-zero truth value of the output variable can be applied to the graph of the membership function of the output variable's symbol, for example, to clip it to a maximum level corresponding to the truth value in question. Applying this principle to all the rules of the inference engine allows us to generate, for each possible fuzzy symbol of the output variable, a set of clipped membership functions. The membership functions thus clipped for all possible fuzzy symbols of the output variable can then be aggregated, for example, by generating a graph of the maximum values ​​of these clipped membership functions. Finally, this graph of maximum values ​​can be used in the defuzzification step 413.Applying, for example, a center of gravity principle to the surfaces defined by this curve allows us to identify the numerical value of the abscissa HPP_AF.

[0101] Finally, since the fuzzy processor 41 calculates the variation of the control signal, the numerical value HPP_AF obtained at the output of the fuzzy processor 41 is integrated over time (i.e., temporally over a sequence of successive output values) by the integrator module 44, to form the setpoint signal F_HPPn to be applied to the power generation unit 1 to control its contribution to the primary frequency regulation (FCR) of the electrical network 3 (see [Fig. 3]). This integration allows the setpoint signal F_HPPn to be stored over time and the setpoint variations from the control block 41 to be applied to it, ensuring the stability and accuracy of the power generation unit's control system.

[0102] In conclusion, it should be noted that the embodiments of the invention described above make it possible, at the level of the electrical production unit, to drastically reduce the wear of mechanical parts related to primary frequency adjustment. Indeed, its implementation on a hybrid production group based on a hydroelectric production unit has made it possible to reduce wear by a factor of 20 the changes in direction of the actuators of the valves 5 and the blades 6 of the rotor of the hydroelectric turbine 4, while allowing for greater flexibility and speed in the response provided as a contribution to the primary frequency control (FCR) of the electrical network 3. In addition, it was determined on this demonstrator that implementations of the invention improve the performance of the primary frequency control by a factor of 10.

[0103] Thus, numerical simulations and implementation carried out on a hydroelectric power plant make it possible to assess the gain concerning the displacement indicators of the turbine actuators, namely: • Gain on the distance traveled by the valves (m): -80%; • reduction in the number of valve direction changes: -96%; • Gain in the distance traveled by the blades (m): -77%; and, • reduction in the number of blade direction changes -90%.

[0104] Various possible variants and modifications of the particular embodiments just described will become apparent to the person skilled in the art.

[0105] In particular, it goes without saying that the invention is not limited by the number of fuzzy values ​​that each of the three input variables can take and the output value of the fuzzy logic calculator implemented in the management module 343 of the proposed hybrid controller 34. The number and form of the corresponding membership functions are not limited to the examples given above with reference to [Fig. 5], [Fig. 6], [Fig. 7], [Fig. 8], and [Fig. 9], respectively. Similarly, the number N and the definition of the inference rules taken into account in the inference engine 412 are not limited to the example given here purely for illustrative purposes with reference to the table in [Fig. 8]. All these parameters must be adapted to the problem being addressed, in the context of each application concerned by the implementation of the invention.Their choice results from prior knowledge of the characteristics and operation of the electrical power generation unit in question, and the constraints applicable to it, all of which falls within the expertise of the professional.

[0106] Furthermore, the control of the electrical power generation unit 1 and the energy storage unit 2 may involve signals other than the three main signals ESS_SOC, PESs, and dPHDDFcR / dt considered in the preceding description. If necessary, a fourth or nth signal may be taken into account and integrated into the fuzzy logic reasoning. In other words, the fuzzy processor 41 of [Fig. 4] comprises at least two (because if the battery capacity were sufficient to allow the storage unit 1 to supply the FCR service on its own in 100% of cases, then the PESs input would no longer be of real interest), and preferably three inputs, and may have four or more. Similarly, other fuzzy processors comparable to fuzzy processor 41 can be envisioned to handle more input variables. However, despite the advantages of these embodiments, which have been outlined and are related to their ease of operational implementation, they are only non-essential.

Claims

1. Demands Method for controlling power supplied to an electrical power distribution network (3) under an alternating voltage of determined nominal frequency, by an electrical power generation unit comprising: - an electrical energy production unit (1) controllable by a controller (15) for regulating the energy produced, having a relatively long response time to a variation in the setpoint of the power produced (PHpp); - a controllable energy storage unit (2), having a relatively short response time to a change in the setpoint of the stored or delivered power (PESS), and used for the primary frequency control (FCR: "Frequency Containment Reserve") of the alternating voltage; and, - a hybridization controller (34) configured to ensure the management of the power generation unit and its contribution to the primary frequency regulation of the alternating voltage as a function of a signal (F_Grid) representing a deviation of said frequency from the nominal frequency, by controlling the power (PESs) stored or released by the energy storage unit (2) from the difference (342) between the primary frequency regulation power (PECr) expected from the power generation unit as its contribution to the primary frequency regulation of the electrical network (3) and the power (Phpp_fcr) produced for this purpose by the power generation unit (1), said process comprising the following steps, implemented in the hybridization controller (34): - acquisition of a first input signal (ESS_SOC) which is representative of the state of charge of the energy storage unit (2); - acquisition of a second input signal (PESs) which is representative of the power (PESs) stored or released by the energy storage unit (2);

2.

3. - acquisition of a third input signal (dPHPP_FCR / dt) which is representative of the time derivative of the power (Phpp_fcr) produced by the electrical power production unit (1) as part of its contribution to the primary frequency regulation of the electrical network (3); - production of a signal (F_HPPn) of the setpoint of the power produced by the electrical power production unit (1) as part of its contribution to the primary frequency control of the electrical network (3), said setpoint signal being produced using at least one inference processor (41) with at least three inputs, as a function of the first input signal (ESS_SOC), the second input signal (PESS), and the third input signal (dPHPP_FCR / dt). A method according to claim 1, wherein the inference processor (41) implements: - three fuzzification modules (41la, 41lb, 41le) using three membership function groups to associate fuzzy input values ​​with normalized numerical values ​​(ESS_SOC_pu, PESs_pu, dPHPP_FCR_pu) of the first input signal (ESS_SOC), the second input signal (PESS) and the third input signal (dPHPP_FCR / dt), respectively, - an inference engine (412), implementing inference rules linking triplets of fuzzy input values ​​from the first input signal (ESS_SOC), the second input signal (PESs) and the third input signal (dPHPPFcR / dt), respectively, to respective fuzzy output values, - a defuzzification module (413) using membership functions to associate the fuzzy output values ​​with respective numeric output values ​​(HPP_AF). Method according to claim 2, wherein the signal (F_HPPn) of the setpoint of the power produced by the electrical production unit (1) as its contribution to the primary frequency control of the electrical network (3), is obtained by integration (44) as a function of time of numerical output values ​​(HPP_AF) of the inference processor (41).

4. A method according to any one of claims 1 to 3, wherein the power (PHpp_fcr) produced by the electrical power generation unit (1) as its contribution to the primary frequency control of the electrical network (3) is determined on the basis of a simplified model (344) of the behavior of said generation unit (1) as a function of the signal (F_HPPn) of the setpoint of the power produced by the electrical power generation unit (1) as its contribution to the primary frequency control of the electrical network (3) which is applied to said generation unit (1).

5. Method according to claim 4, wherein the third input signal (dPHpp_FCR / dt) which is representative of the time derivative of the power (PHpp_fcr) produced by the electrical power generation unit (1), is obtained by calculation from the simplified model (344).

6. A method according to any one of claims 1 to 5, wherein the signal (F_HPPn) of the setpoint of the power produced by the electrical power production unit (1) as its contribution to the primary frequency regulation of the electrical network (3) is of the same nature as the signal (F_Grid) representing a deviation of the frequency of the electrical network (3) from the nominal frequency of said network (3), so as to be able to be substituted for it as an input of a controller (15) for regulating the power produced by said production unit (1).

7. A method according to any one of the preceding claims, wherein either of the signal (F_Grid) representing a deviation of the frequency of the electrical network (3) from the nominal frequency and the signal (F_HPPn) of the setpoint of the power produced by the electrical power generation unit (1) as its contribution to the primary frequency regulation of the electrical network (3) is selectively received at the input of the controller (15) regulating the power (PHpp) produced by the electrical power generation unit (1).

8. A control device for power supplied to an electrical power distribution network (3) under an alternating voltage of a determined nominal frequency, by an electrical power generation unit comprising: - a controllable electrical power production unit (1), having a relatively long response time to a variation in the setpoint of the power produced (PHpp); - a controllable energy storage unit (2), having a relatively short response time to a change in the setpoint of the stored or delivered power (PESS), and used for the primary frequency control (FCR: "Frequency Containment Reserve") of the alternating voltage; and, said device comprising: - a controller (15) for regulating the energy produced by the electrical power generation unit (1); and, - a hybridization controller (34) configured to manage the power generation unit and its contribution to the primary frequency regulation of the AC voltage based on a signal (F_Grid) representing a deviation of said frequency from the nominal frequency, by controlling the power (PESs) stored or delivered by the energy storage unit (2) from the difference (342) between the expected primary frequency regulation power (PECr) of the power generation unit as its contribution to the primary frequency regulation of the electrical grid (3) and the power (Phpp_fcr) produced for this purpose by the power generation unit (1), said hybridization controller (34) comprising at least one inference processor (41) with at least three inputs and being configured to: • receive as input a first input signal (ESS_SOC) which is representative of the state of charge of the energy storage unit (2); • receive as input a second input signal (PESS) which is representative of the power (PESs) stored or released by the energy storage unit (2); • receive as input a third input signal (dPHPp_ECR / dt) which is representative of the time derivative of the power (PHpp_fcr) produced by the electrical power production unit (1) as part of its contribution to the primary frequency regulation of the electrical network (3); • produce at output a setpoint signal (F_HPPn) of the power produced by the electrical power production unit (1) as its contribution to the primary frequency control of the electrical network (3), said setpoint variation signal being produced using the inference processor (41) having at least three inputs, as a function of the first input signal (ESS_SOC), the second input signal (PESs) and the third input signal (dPHpp_EcR / dt), respectively.

9. Device according to claim 8, wherein the inference processor (41) comprises: - three fuzzification modules (41 la,41 lb,41 le) using three membership function groups to associate fuzzy input values ​​with normalized numerical values ​​(ESS_SOC_pu, PEss_pu, dPHpp_EcR_pu) of the first input signal (ESS_SOC), the second input signal (PESs) and the third input signal (dPHpp_EcR / dt), respectively; - an inference engine (412), implementing inference rules linking triplets of fuzzy input values ​​from the first input signal (ESS_SOC), the second input signal (PESs), and the third input signal (dPHpp_EcR / dt), respectively, to respective fuzzy output values; and, - a defuzzification module (413) using membership functions to associate fuzzy output values ​​with respective numeric output values ​​(HPP_AF).

10. A device according to claim 9, wherein the hybridization controller (34) comprises an integrator (44) configured to produce the signal (F_HPPn) of the setpoint for the power produced by the electrical generating unit (1) as its contribution to the primary frequency regulation of the electrical grid (3), by time-dependent integration of numerical output values ​​(HPP_AF) from the inference processor (41).

11. Device according to any one of claims 8 to 10, wherein the power (PHpp_fcr) produced by the electrical power generation unit (1) as its contribution to the primary frequency control of the electrical network (3) is determined on the basis of a simplified model (344) of the behavior of said generation unit (1) as a function of the signal (F_HPPn) of the setpoint of the power produced by the electrical power generation unit (1) as its contribution to the primary frequency control of the electrical network (3) which is applied to said generation unit (1).

12. Device according to claim 11, wherein the third input signal (dPHpp_FCR / dt) which is representative of the time derivative of the power (PHpp_fcr) produced by the electrical power generation unit (1), is obtained by calculation from the simplified model (344).

13. Device according to any one of claims 8 to 12, wherein the signal (F_HPPn) of the setpoint of the power produced by the electrical power production unit (1) as its contribution to the primary frequency regulation of the electrical network (3) is of the same nature as the signal (F_Grid) representing a deviation of the frequency of the electrical network (3) from the nominal frequency of said network (3), so as to be able to be substituted for it as an input of a controller (15) for regulating the power produced by said production unit (1).

14. Device according to claim 13, further comprising a selector (151) configured to provide either of the signal (F_Grid) representing a deviation of the frequency of the electrical network (3) from the nominal frequency and the signal (F_HPPn) of the setpoint of the power produced by the electrical power generation unit (1) as its contribution to the primary frequency setting of the electrical network (3), into the input of the controller (15) regulating the power (PHpp) produced by the electrical power generation unit (1).

15. Electric power generation unit adapted to supply said produced electric power to an energy distribution network electrical (3) under an alternating voltage of determined nominal frequency, said group comprising: - a controllable electrical power production unit (1), having a relatively long response time to a variation in the setpoint of the power produced (PHpp); - a controllable energy storage unit (2), having a relatively short response time to a change in the setpoint of the stored or delivered power (PESS), and used for the primary frequency control (FCR: "Frequency Containment Reserve") of the alternating voltage; as well as, - a power control device supplied to the electrical power distribution network (3) according to any one of claims 8 to 14.

Images