Control of power injected by a renewable energy production module into an electrical micro-grid by frequency control of an electrical energy storage system

Abstract

The present disclosure relates to a method for controlling power injected by at least one renewable energy production module into an islanded electrical microgrid relative to a main electrical distribution network.Such a method comprises: a connection, in the micro-grid, of an electrical energy storage system comprising an electrical energy storage module and a DC / AC inverter, said electrical energy storage module being capable of delivering a direct voltage to the terminals of said DC / AC inverter, said DC / AC inverter generating an alternating voltage capable of supplying the micro-grid; a control of a frequency setpoint value of the alternating voltage generated by the inverter from control laws of the frequency setpoint value established from laws of evolution of the frequency setpoint value as a function of an absorbed power measured at the terminals of the electrical energy storage module on the one hand, and of a charge level of the electrical energy storage module on the other hand, the evolution laws having a hysteresis cycle. Abstract figure: Figure 9.

Description

Title of the invention: Control of power injected by a renewable energy production module into an electrical micro-grid by frequency control of an electrical energy storage system Technical field

[0001] The present disclosure relates to the field of electrical distribution networks, to which one or more variable renewable energy production modules, ENRv, are connected, such as photovoltaic energy production modules or wind-generated electrical energy production modules, for example. Prior art

[0002] It is known to temporarily connect one or more generator sets to an electrical distribution network, to compensate for a lack of electrical energy supply linked to an incident on the network or to maintenance work for example. The generator set then supplies electrical energy to a sub-part of the tree structure of the distribution network, which is called a micro-network, and which is temporarily disconnected from the main electrical distribution network: this is then called an islanded electrical micro-network.

[0003] Prior to implementing such a means of temporary electricity re-supply, the operator of the electricity distribution network disconnects all the variable renewable energy production modules from the micro-grid having an installed power greater than 5kW. Indeed, the existing generators cannot absorb the power of these variable renewable energy production modules, which must therefore be disconnected before intervention, then reconnected at the end of the intervention on the micro-grid.

[0004] This procedure is time-consuming and human resource-intensive. In addition, it prevents any production of electrical energy by variable renewable energy production modules, such as photovoltaic panels, for the duration of the intervention, which represents a financial loss. Finally, for the duration of the intervention, the energy supplying the electrical micro-grid is of fossil origin, and the generator set can generate pollution linked to the consumption of diesel-type fuel on the one hand, and noise pollution on the other hand.

[0005] To overcome these various drawbacks, an advantageous solution consists of replacing these generator sets with battery-based electrical energy storage systems, which can act as voltage sources for the islanded micro-grid.

[0006] However, it is necessary to be able to control by clipping the power injected by the variable renewable energy production modules into the islanded micro-grid, depending on its load for example, and to disconnect them from the grid if necessary, to avoid a blackout (in French, power failure).

[0007] Some variable renewable energy production modules, and in particular photovoltaic energy, conventionally comprise a photovoltaic panel generating continuous power on the DC side of a DC / AC inverter, which converts it into AC power, before its injection into the network. Some of these inverters are controlled by means of a power limitation function, called P(f). In a microgrid, it represents a means of indirect control of the power injected by these inverters. This control is carried out by the voltage source, in other words by the equipment which imposes the frequency of the microgrid, and makes it possible to impose a maximum admissible value of the power injected by the inverter into the microgrid.If the power available at a given time for the renewable energy production module is lower than this maximum power value imposed by the function P(f), the power injected by the inverter into the micro-grid is equal to the available power. Otherwise, it is limited to this maximum power value imposed by the function P(f). The control laws of the power injected into the network as a function of the frequency measured on the latter, for most known inverters, follow affine functions, as illustrated by [Fig.l].

[0008] In the context of the use of battery electrical energy storage systems in an islanded micro-grid, it has therefore been considered to indirectly control the inverters of these variable renewable energy production modules, via the frequency imposed on the micro-grid by these battery storage systems, acting as voltage sources of the micro-grid. Thus, some of these systems implement a “droop” frequency adjustment, or a dedicated frequency curve adjustment for the inverters of the photovoltaic modules. In particular, some manufacturers have implemented an option for adjusting the frequency of the battery storage systems, depending on the power measured at the terminals of the battery on the one hand, and the battery charge level (SoC) on the other hand.Patent document US2019190274A1, titled “An Energy Management System And Method For Grid-Connected And Islanded Micro-Energy Generation” presents such a solution.

[0009] However, this frequency adjustment function is then generally described by an affine type law of the frequency for the absorbed power values on the one hand, and for the battery charge level values on the other hand (i.e. the functions existing remain limited to a straight line shape). Since most inverters of variable renewable energy production modules also have affine functions as operating laws of their power with respect to the measured frequency, as indicated above, this frequency control solution can therefore prove unstable. In particular, one of the most significant risks of frequency control by such battery storage systems, in an islanded microgrid comprising one or more variable renewable energy production modules, is that of a microgrid blackout, leading to a disconnection of all customers, and therefore an increase in undistributed energy. Another possible side effect is the creation of power oscillations (or pumping) which can result in a cycle of decoupling and re-coupling which can be endless under certain conditions.These oscillatory instabilities can be of the type initiated over time (the least impactful), of the permanent type (the most "tiring" for the network equipment), or even divergent (which can create a generalized power outage following the exceeding of the maximum power absorbed by the battery). Summary

[0010] The present disclosure improves the situation.

[0011] A method is proposed for controlling power injected by at least one renewable energy production module into an islanded electrical micro-grid relative to a main electrical distribution network, which comprises: a. a connection, in the islanded electrical microgrid, of an electrical energy storage system comprising at least one electrical energy storage module and at least one DC / AC inverter, said electrical energy storage module being capable of delivering a direct voltage to the terminals of a DC / AC inverter, said DC / AC inverter generating an alternating voltage capable of supplying the islanded electrical microgrid; b. a control of a frequency setpoint value of the alternating voltage generated by the inverter from: i. a first control law of the frequency setpoint value established from a first law of evolution of the frequency setpoint value as a function of an absorbed power measured at the terminals of the electrical energy storage system, the first evolution law having a hysteresis cycle; ii. of a second law for controlling the frequency setpoint value established from a second law for changing the frequency setpoint value as a function of a charge level of the electrical energy storage module, the second law of changing having a hysteresis cycle.

[0012] According to another aspect, there is provided a mobile electrical energy storage system capable of being connected to an electrical micro-grid islanded with respect to a main electrical distribution network, such a storage system comprising at least one electrical energy storage module and at least one DC / AC inverter, the electrical energy storage module being capable of delivering a direct voltage to the terminals of the DC / AC inverter, the DC / AC inverter generating an alternating voltage capable of supplying the islanded electrical micro-grid. Such a storage system comprises a processor configured to execute a control of a frequency setpoint value of the alternating voltage generated by the inverter from: i. a first control law of the frequency setpoint value established from a first law of evolution of the frequency setpoint value as a function of an absorbed power measured at the terminals of the electrical energy storage system, the first evolution law having a hysteresis cycle; i. of a second control law of the frequency setpoint value established from a second law of evolution of the frequency setpoint value as a function of a charge level of the electrical energy storage module, the second evolution law having a hysteresis cycle.

[0013] According to another aspect, there is provided a computer program comprising instructions for implementing all or part of a method as defined herein when this program is executed by a processor. According to another aspect, there is provided a non-transitory, computer-readable recording medium on which such a program is recorded.

[0014] Thus, compared to prior techniques, the stability of the frequency control of the power injected by the ENRv production modules into an islanded micro-grid is improved by implementing, in the controller of a battery-based electrical energy storage system, a logic for controlling the frequency as a function of the AC power measured at the terminals of the storage system and the battery charge level, which is based on evolution laws implementing hysteresis cycles. It is thus possible to increase the production of green energy in the islanded micro-grid, and to avoid systematic decoupling of all the ENRv producers, without risking a blackout.

[0015] The features set out in the following paragraphs may, optionally, be implemented independently of one another or in combination with one another.

[0016] The hysteresis cycle of the first law of evolution is of the parallelepiped type and understand : - a straight line of increase in the frequency setpoint value as a function of an increase in the absorbed power measured from a starting power value, Pstart. For example, Pstart=0.4 pu, where Pfin=1 pu is the maximum value of the absorbed power measured on the straight line of increase of the hysteresis cycle; - a straight line of decrease in the frequency setpoint value as a function of a decrease in the measured absorbed power; and said increase and decrease lines are parallel and offset from each other on the abscissa by a power hysteresis value AP hysteresis, with for example A Physteresis=0.1 Pu.

[0017] The first control law also comprises, for a determined blocking duration starting when the frequency setpoint increases and reaches a first blocking threshold frequency according to the first evolution law, a blocking of a minimum value of the frequency setpoint at a second blocking threshold frequency greater than the first blocking threshold frequency. The stability of the method is thus ensured by blocking the downward variation of the frequency for a given configurable period, for example one hour. This makes it possible to ensure the proper operation of the inverters of the ENRv modules for which the P(f) function is not activated.

[0018] The hysteresis cycle of the second law of evolution is of the parallelepiped type and comprises: - a straight line of increase in the frequency setpoint value as a function of an increase in the charge level of the storage module from a starting charge level value, SoCslarl, for example substantially equal to 60%, and up to an ending charge level value, SoCfm, for example substantially equal to 90%; - a straight line of decrease in the frequency setpoint value as a function of a decrease in the charge level of the storage module; and, said increase and decrease lines are parallel and offset relative to each other on the abscissa by a charge level hysteresis value A SoC, for example substantially equal to 10%.

[0019] The second control law also comprises, when the frequency setpoint value increases and reaches a first blocking threshold frequency according to the second evolution law, a blocking of a minimum value of the frequency setpoint value at a second blocking threshold frequency greater than the first blocking threshold frequency until the frequency setpoint value increases and reaches the first blocking threshold frequency or until the frequency setpoint value decreases. and reaches a third blocking threshold frequency lower than the first blocking threshold frequency, according to the second evolution law. This ensures the stability of the process and the proper functioning of the inverters of the ENRv modules for which the P(f) function is not activated.

[0020] The first, second and third blocking threshold frequencies are determined as a function of a decoupling frequency configured for an inverter of the renewable energy production module(s).

[0021] The control of the frequency setpoint value is controlled by the maximum function of a variation in time of the frequency setpoint value according to the first control law and of a variation in time of the frequency setpoint value according to the second control law. A frequency threshold is thus obtained which respects the constraints taken into account by the two control laws; in addition, the choice of the maximum function makes it possible to impose the most restrictive settings dictated by each of the two control laws.

[0022] The control of the frequency setpoint value also comprises an additional application, to the maximum function of a maximum frequency setpoint value, of a minimum frequency setpoint value, of a limit value of the slope of increase of the frequency setpoint value as a function of time, of a limit value of the slope of decrease of the frequency setpoint value as a function of time, and the limit value of the increase slope is greater than the limit value of the decrease slope, in absolute values. This guarantees a well-controlled operating range, while ensuring a rapid rise in the frequency to quickly stop the variations in the power injected by the ENRv modules and a slow descent of the frequency to avoid instabilities. Brief description of the drawings

[0023] Other characteristics, details and advantages will appear on reading the detailed description below, and on analyzing the attached drawings, in which: Fig.l

[0024] [Fig.l] presents the representative curve of the function P(f) as implemented, according to prior techniques, in certain inverters for connecting photovoltaic panels. Fig. 2

[0025] [Fig.2] illustrates an islanded micro-grid according to one embodiment. Fig. 3

[0026] [Fig.3] presents a curve representative of the first law of evolution of a frequency setpoint value as a function of an absorbed power measured at the terminals of a battery storage system according to one embodiment. Fig. 4

[0027] [Fig.4] illustrates a treatment by decoupling levels of the curve of [Fig.3] according to one embodiment. Fig. 5

[0028] [Fig.5] presents a curve representative of the second law of evolution of a frequency setpoint value as a function of a charge level of a battery storage system according to one embodiment. Fig. 6

[0029] [Fig.6] illustrates a treatment by decoupling levels of the curve of [Fig.6] according to one embodiment. Fig. 7

[0030] [Fig.7] illustrates the principle of a third stage of construction of a control law of the set value of the frequency according to one embodiment. Fig. 8

[0031] [Fig.8] illustrates the principle of a fourth stage of construction of a control law of the set value of the frequency according to one embodiment. Fig. 9

[0032] [Fig.9] presents in synoptic form the control logic implemented in a controller of a battery storage system according to one embodiment. Fig. 10

[0033] [Fig. 10] shows a block diagram of a battery storage system im implementing the control logic of [Fig.9] according to one embodiment. Description of the embodiments

[0034] The general principle of the technique which is the subject of the present disclosure is to indirectly control the inverters of the variable renewable energy production modules of an islanded micro-grid, by the frequency imposed by an electrical energy storage system forming a voltage source of the micro-grid, by reducing possible instabilities, and therefore the risks of blackout which could cause the power supply to customers to be cut off.

[0035] As indicated above, variable renewable energy production modules, ENRv, such as photovoltaic or wind modules, may comprise one or more converters, or inverters, which transform the direct power generated by a photovoltaic panel, or the alternating or direct power of a wind turbine, into alternating power which can be injected into the electrical distribution network. Some of the ENRv inverters present on the national electrical distribution network in France have an option for adjusting P(f) the injected power as a function of the network frequency, as illustrated in [Fig.l].

[0036] The curve P(f) represents the maximum value of the power that can be injected by an ENRv inverter into the network. It is imposed by the inverter as a function of the measured frequency. Between this curve P(f) and the zero power line is an operating zone of the inverter, represented in hatched form in [Fig.l]: the power injected by the ENRv inverter takes a value that is in this hatched zone, between the curve P(f) and 0 (losses are ignored in this example). The power imposed by the function P(f) therefore only impacts the power injected by the ENRv inverter when the available power exceeds the value of P(f).

[0037] The curve P(f) of [Fig.l] has three important and configurable elements: - the value of the start frequency of the limitation of the maximum power injected by the inverter: fstart= fn+Y, where fn designates the nominal frequency of the electrical distribution network, i.e. 50 Hz in the example of [Fig.l] and throughout this document. For a measured frequency less than or equal to fstart= fn+Y, the inverter can inject all the installed power (Pn), and the injected power is equal to the available power (no clipping); - the value of the frequency of the end of limitation of the maximum power injected by the inverter: ffin=fn+X. Beyond this frequency value ffin=fn+X, the power is completely clipped (0% of the injected power, regardless of the available power); - the value of the ramp between the starting point of the coordinate curve (Pn; f stait) and the end point of the coordinate curve (0; ffin); for a measured frequency between fstart and ffin, the maximum power injected by the inverter is clipped by the law P(f) if the available power is greater than P(f). In practice, one of these 3 parameters may prove redundant, and only two of these three parameters, at the manufacturer's choice, are generally configurable on the human-machine interfaces (HMI) of most existing inverters.

[0038] The parameters usually used are fstart=50.2Hz and ffin=51.2Hz, i.e. a ramp of AP / Af= 1 [pu / Hz].

[0039] However, in France, only ENRv inverters installed after June 2020 implement active control of the injected power based on the power measured according to such a P(f) function. ENRv inverters installed in the electricity distribution network before June 2020 must be disconnected when islanding is detected. To avoid power injection into an islanded network, without voltage regulation, the standard imposes decoupling protection depending on the year of installation of the ENRv inverter. The different types of inverter currently present on the national electricity distribution network in France are presented in Table 1 below.

[0040] [Tables 1] Inverter type P(f) activated by default Decoupling frequency (in Hz) Installed power (in GW) Probability of presence on the national network (in %) 1 NO 50.2 2 43 2 NO 50.4 0.3 7 3 NO 50.6 1.7 37 4 YES 51.5 0.6 13

[0041] It should be noted that this diversity of inverter types present on the national electricity distribution network makes it complex to control the power injected by an ENRv module by frequency control imposed by equipment forming a voltage source, and this is all the more so since in the case of intervention on an islanded micro-network, it is difficult to know the exact distribution of the different types of inverters and their installed power. Furthermore, in other countries, the number of types of inverters present on the electricity distribution network may be different from four.

[0042] The following describes an example embodiment in which a battery-based electrical energy storage system is temporarily connected to an islanded microgrid due to scheduled maintenance or a power outage on the main electrical distribution network. It is assumed that the islanded microgrid comprises several variable renewable energy production modules, ENRv, for example several photovoltaic energy production modules, as illustrated in [Fig. 2]. The following describes an example applicable to the French electrical distribution network, comprising type 1 to 4 inverters as mentioned above.This example can of course be generalized to any other configuration of electrical distribution network, comprising more or less distinct types of inverters (for example N>1 distinct types of inverters), with or without a P(f) function activated by default, and decoupling frequency and installed power characteristics which may differ from those mentioned in Table 1 above.

[0043] An electrical energy storage system comprises a battery 10i providing a direct voltage to the terminals of a DC / AC inverter 102. The DC / AC inverter 102 is a power electronics device that generates an alternating voltage, which supplies the islanded micro-grid 11, for the duration of the maintenance intervention or the outage. More generally, the term battery 10i denotes the DC part of the electrical energy storage system, which comprises one or more battery cells, their individual control, as well as the battery management system or BMS (for English: "Battery Management System"). The electrical energy storage system is the assembly consisting of the battery 10i and the DC / AC inverter 102 which can produce an alternating voltage. The alternating voltage and current can be measured to obtain the power of the storage system (injected or withdrawn from the micro-grid) and the frequency (of its voltage). Several photovoltaic energy production modules PVj to PVn are also present on the micro-grid 11. Each of these PV modules; comprises a photovoltaic panel 12u which supplies direct power to a DC / AC inverter 122i, which converts it into alternating power before injection into the micro-grid 11. A set of loads Ci to Cm are connected to the micro-grid 11 and consume the electrical energy supplied by the storage system and the photovoltaic modules PVj.

[0044] Throughout this document, the “generator” convention is used for the power of the electrical energy storage system: a positive power P means that the storage system injects power into the micro-grid 11, and therefore that it discharges. It is also considered that the power absorbed by the storage system corresponds to the power generated by the photovoltaic energy production modules PVj to PVn, less the power of the loads Ci to Cm. Throughout this document, a positive measured power is therefore a power absorbed by the battery storage system.

[0045] In one embodiment, a logic for controlling the setpoint value of the frequency f of the alternating voltage generated by the DC / AC inverter 102 and supplying the micro-grid 11 is implemented in the battery storage system, which logic must meet several objectives:

[0046] 1. Avoid a blackout, which may be caused, for example: a. excessive power absorbed by the storage system (greater than its installed power); b. too high a storage system charge level (SoC). If the battery is almost full (high SoC), the battery management system will limit the maximum current drawn to a total limitation. If the equivalent network current requires a current drawn higher than this limitation, the protections will open the main circuit breaker; c. the storage system's charge level (SoC) is too low, if it is too discharged. If the battery is almost empty (very low SoC) the battery management system will limit the maximum injected current up to a total limitation. If the equivalent network current requires an injected current greater than this limitation, the protections will open the main circuit breaker.

[0047] 2. Take into account the diversity of types of inverter settings 122.; which can exist on microgrid 11 (see Table 1).

[0048] 3. Promote the production of renewable energy as much as possible in order to be able to: a. Extend the theoretical islanding duration (i.e. prevent the battery from discharging too quickly); b. Produce so-called “green” energy to meet an environmental priority; c. To allow the owners of photovoltaic energy production modules PVi to PVn to benefit from the presence of the micro-grid 11, by allowing them to sell the electricity produced to the operator of the electricity distribution network. 4. To guarantee stable operation in the short and long term with regard to oscillatory variations between the set frequency and the active power of the inverters.

[0049] To meet these needs, the proposed technique is based on a logic for controlling the set value of the frequency f of the alternating voltage generated by the DC / AC inverter 102 and supplying the micro-network 11 which comprises two main components: - a first control law f(P), which controls the frequency f as a function of the power P measured at the terminals of the electrical storage system; - a second control law f(SoC), which controls the frequency f according to the charge level SoC (for the English “State of charge”) of the 10i battery.

[0050] Each of these two control laws (f(P), f(SoC)) operates according to a two-stage logic: - a first stage, implementing an evolution law presenting a hysteresis cycle, which ensures the correct operation of type 4 inverters, or more generally of all inverters in which the P(f) setting is activated; - a second stage, which operates a processing by decoupling levels, and which ensures the correct operation of type 1, 2 and 3 inverters, or more generally of all inverters in which the P(f) setting is not activated.

[0051] We now present, in relation to figures 3 to 5, the first control law f(P) implemented in the controller of the inverter 102 of the electrical energy storage system. This first control law is established from a first evolution law illustrated in [Fig.3].

[0052] The input data of this first evolution law is the power measured at the terminals of the electrical energy storage system, noted Pmes. This power Pmes is the AC power, injected into the network, or withdrawn from the network, by the electrical energy storage system. This first evolution law, corresponding to the first stage of the first control law f(P), makes it possible to ensure the correct operation of type 4 inverters, in which the P(f) setting is activated, and whose decoupling frequency is fixed at 51.5 Hz. It has a hysteresis cycle parai- lepipedic, which includes: - a "rise" line, i.e. an increase in frequency as a result of an increase in the measured power, which is already present in the "P(f)" settings of prior art batteries with this function. Unfortunately, this simple affine function is not sufficient to ensure stability. This line is delimited by two end points A with coordinates (Pstart, fstart) and B with coordinates (Pfin, ffin). The values of Pstart> Pfin, fstart and ffin can be configured by the human-machine interface of the electrical storage system. For example, for these four parameters, values optimized for all application cases can be chosen, or these values can be adjusted on a case-by-case basis, in order to obtain a fine optimization particularly suited to a given application case. As an example, here, Pfin is set at 100% of the installed power. We therefore denote Pfin=lp.u.; - a “downward” line, that is to say a decrease in frequency as a result of a decrease in the measured power. This line (CD) is parallel to the “upward” line (AB), to its left; . - an area located between the lines (AB) and (CD), which has a width APhysteresis ct allows to ensure stability against oscillations between the frequency and the power. In one embodiment APhysteresis = 0, ^PU- ; - two non-adjustment zones of f(P), in which the frequency can take a value between ffin and fmax, respectively between fnOminaie and fstart; - two out-of-bounds zones, such that no measured power value can result in a frequency greater than fmax or less than fnOminaie-

[0053] In one embodiment, making it possible to optimize this evolution law for the operation of type 4 inverters, we fix: - fstart=50.2Hz; - ffin=51.2Hz; -fmax=51.55Hz, where f - ftype^ + A f with =51 5Hz^ J max J decoupling J margin ■' decoupling At f = 0.05Hz

[0054] This margin A fmarge is necessary to ensure that the ENRv inverter takes into account a measured frequency higher than the threshold set for its decoupling frequency, and therefore that it opens its internal circuit breaker. The value of the margin A f maroe must cover the frequency generation error with respect to the frequency setpoint value imposed at the storage system inverter and the frequency measurement error at the ENRv inverter.

[0055] The hysteresis zone between the lines (AB) and (CD) implies two possible evolutions of the frequency, depending on the sign of the derivative of the power with respect to time.

[0056] Following a high inflection point (i.e., when the measured power Pmes stops increasing and starts decreasing): - for a variation of the measured power less than APhysteresis, the frequency f is kept constant; - for a variation of the measured power greater than APhvsteresis» 'the frequency f decreases along the line (CD).

[0057] Following a low inflection point (i.e., when the measured power Pmes stops decreasing and starts increasing): - for a variation of the measured power less than APhvsteresis> 'the frequency f is kept constant; - for a variation of the measured power greater than APhysteresis, the frequency f increases along the line (AB).

[0058] Furthermore, the upper limitation zone of this first evolution law, for which f=fmax, is reached: - when Pmes>Pfin and A Pmes 0 î - when Pmes > P- AP hysteresis and AP mes < 0.

[0059] The lower limitation zone of this first evolution law, for which f=f nominal, is reached: - when SMEs<Pstart et A Pmes > 0 ; - when Pmes < Pstart - AP^steresis and AP^ < 0.

[0060] In one embodiment, Pstart=0.4p.u., point D has an abscissa of 0.3p.u., and point C has an abscissa of 0.9p.u. The values chosen for Pstart and APhysteresis play an important role in the stability of the potential oscillations between the set frequency and the active power of the inverters. As Pfin=1p.u., the value of Pstart defines the variation range of the function f(P) for the battery storage system. The greater the variation range of the function f(P), the smoother the ramp Sf, and therefore the greater the AP frequency variation will be low for IkW of power variation: this gives greater stability. In fact, stability A / increases with the magnitude of the 5 AH / ) 5 P Pfin-Pstart range, which is 0.6p.u. in the example in [Fig.3].

[0061] In other embodiments, other values may be chosen for all of these parameters, and in particular for Pfin> Pstart and APhvsteresis, depending on the stability sought and the configuration of the islanded micro-grid to which the battery storage system is connected.

[0062] In one embodiment, the value of Pstart is calculated based on the installed power of the ENRv production modules and the battery storage system. In another embodiment, in which the battery storage system battery is mobile and can be used on all types of islanded micro-grids with variable configurations and installed power of ENRv modules, we set Pstart=0.4p.u. or less, up to Pstart=0p.u.

[0063] The APhysteresis parameter also has an impact on the stability of the control: it guarantees a constant frequency value for a change in the monotony of the variation of the measured power. Its value can be chosen in a range varying from 0.05p.u. to 0.5p.u. for example. Furthermore, the adjustment precision of the inverters, both of the ENRv production module and of the battery storage system, is increased by the magnitude of the frequency variation range. In the example of [Fig.3], we have ffin-fstart=1Hz, which is advantageous in that it is the value already imposed by the standard applied to type 4 inverters. Other values can also be considered, depending in particular on the characteristics of the inverters integrated in the ENRv production modules present on the islanded micro-grid.

[0064] The different numerical values proposed above make it possible to ensure satisfactory stability of the frequency control, while taking into account certain compromises. However, these values are configurable according to use. It should be noted that, although the most stable control is the one for which the frequency is constant and the power injected by the ENRv modules into the islanded micro-grid does not vary, this solution is not optimal because it involves complete decoupling of the ENRv modules from the micro-grid, therefore a financial loss and green energy.

[0065] The first evolution law described above in relation to [Fig.3] controls a frequency setpoint value which can be noted fhysteréSiS(PmeS), which feeds a second stage of construction of the first frequency control law as a function of the power measured at the terminals of the battery, presented below in relation to [Fig.4].

[0066] In Figure 4, a fictitious variation of the frequency fhysteréSiS(PmeS) is shown as a function of time. The processing by decoupling levels of the curve representing this variation ensures stability by a temporary blocking of the reduction of the frequency, if the latter exceeds a decoupling frequency threshold of a type 1, type 2, or type 3 inverter. In Figure 4, we have represented fseuii decoupling i=50.2Hz, which is the decoupling frequency of a type 1 inverter, fseuiidécouPiage 2=50.4Hz, which is the decoupling frequency of a type 2 inverter, and fseuiidécoupiage3 =50.6Hz, which is the decoupling frequency of a type 3 inverter (i.e. f . e { 50.2; 50.4; 50.6} where i is the type of inverter, ze 11 2; 3} )• We note that one could generalize from 3 to N decoupling frequency thresholds, where N>0, in a network which would include N types of inverters presenting N distinct decoupling frequencies for which the P(f) function is not activated.

[0067] When the frequency setpoint imposed by the first evolution law f (hysteresis) of [Fig.3] increases and reaches a first blocking threshold frequency equal to decoupling threshold 1 fmargin' -^"decoupling threshold 2 f margin OR decoupling threshold 3 " margin' the frequency setpoint value is prohibited from taking a value lower than a second blocking threshold frequency, which is higher than the first blocking threshold frequency by a value equal to 2 A f marse = 0.1Hz. Thus, the minimum frequency setpoint value is blocked, for a configurable duration of one hour for example, at fseuildécouplage l+0.05Hz, fseuildécouplage 2+0.05Hz OR fseuildécouplage 3 +0.05Hz. In another embodiment, the blocking duration can be set at thirty minutes, at 1h30, or even at two hours, depending on the configuration of the islanded micro-grid and the desired compromise between stability and use of the power of the ENRv production modules. In particular, this blocking duration can be optimized according to the ratio of the following powers: - the installed power of the battery storage system; - the installed power of the ENRv modules present on the islanded micro-grid; - the charging power (i.e. the installed charging power or the expected charging curve). For example, this duration can be reduced to optimize the production of renewable energy, when it is considered that there is no significant risk from the point of view of installed power, i.e. when the installed power of the battery storage system is much higher than the installed power of the ENRv modules.

[0068] With reference to Figure 4, it is therefore understood that the blocking of the frequency setpoint value is active and that a time counter is activated as soon as the frequency increases (f hysteresis(Pti) > fhysteresis(P^-i ) where and Pa designate the powers measured at two successive time instants ti -1 and h) and reaches the first blocking threshold frequency (y .. A f < f.,... (P, where j = 1, 2 or 4 The new frequency setpoint is immediately locked (downward) at the second locking threshold frequency, f . .+ A f . Thus, as soon as the frequency exceeds f .. ., . . - At f ... we have to jump the frequency to r J decoupling threshold j J margin' f . + Af to ensure that the inverters measure a value J decoupling threshold j J margin a lower than their decoupling threshold before blocking and that they measure a value higher than their decoupling threshold after blocking. Blocking is carried out at fdecoupling threshold jf margin'

[0069] This second stage of construction of the first control law makes it possible to ensure the correct operation of type 1, 2 or 3 inverters, or more generally of all the inverters in which the P(f) adjustment function is not activated. At the output of this second stage of construction of the control law, we obtain a frequency setpoint value corresponding to the first control law, which we denote by f(Pmes).

[0070] We now present, in relation to Figures 5 and 6, the second control law f(SoC) implemented in the controller of the inverter 102 of the electrical energy storage system. This second control law is established from a second law of evolution of the set value of the frequency as a function of a charge level of the electrical energy storage module, illustrated in [Fig.5]. This second evolution law is very similar, in its principle, to the first evolution law presented in relation to [Fig.3], and we therefore only describe below the main differences with the latter.

[0071] The input data of this second evolution law is the charge level of the 10i battery, noted SoC. It has a parallelepiped hysteresis cycle, which includes: - a "rise" line, i.e. an increase in frequency following an increase in the load level. This line is delimited by two end points A' with coordinates (SoCstart, fstart) and B' with coordinates (SoCfin, ffin). In one embodiment, SoCstart=60% and SoCfin=90%. These values can be configured on the HMI of the electrical energy storage system, and other values can be chosen, depending in particular on the configuration of the islanded micro-grid; - a "downward" line, i.e. a reduction in frequency as a result of a reduction in the load level. This line (C'D') is parallel to the "upward" line (A'B'), to its left; - an area located between the lines (A'B') and (C'D'), which has a width ASoChvsteresis ct allows to ensure stability against oscillations between the frequency and the charge level. In one embodiment ASoChysteiesis — 10%. Again, this value is configurable via the HMI of the electrical energy storage system, and can be adapted, depending on the application case; - two non-adjustment zones of f(SoC), in which the frequency can take a value between ffin and fmax, respectively between fnominaie and fstart; - two out-of-bounds zones, such that no load level value is for effect a frequency greater than fmax or less than or equal to fnominaie.

[0072] In one embodiment, making it possible to optimize this evolution law for the operation of type 4 inverters, we fix: - fstart=50.2Hz; - ffin=51.2Hz; - fmax=51.55Hz, where f = + A f with ftypeA = 5\5Hzcl J wax J decoupling J margin J decoupling ' Af„^ = 0.05 / fe.

[0073] It should be noted that it is important that these parameters are set to the same value for the first evolution law fhysteréSiS(PmeS) illustrated in [Fig.3] and for the second evolution law fhysteréSiS(Soc) illustrated in [Fig.5].

[0074] This second evolution law f(SoC), which has a hysteresis cycle, makes it possible to ensure the stability of the control of the power injected by the ENRv production modules, and in particular to overcome potential SoC jumps, resulting from reliability defects in the estimation of the battery charge level. Indeed, such defects are a known problem in battery storage systems, and depend on the battery technology on the one hand, and the technique for measuring or estimating their charge level on the other hand.

[0075] It should be noted that this second law of evolution only applies when the absorbed power measured at the terminals of the battery is negative (Pmes<0), i.e., in the event of charging of the battery lOp

[0076] The magnitude of the load level variation range, which is set to SoCfin-SoCstart =30% in the example of Figure 6, is important to ensure the stability of the proposed control. Indeed, the stability ......4 / .. increases with the magnitude of the SoCfin-SoC range This value may seem low, compared to that of the variation range of the measured absorbed power presented in Figure 3. However, the speed of variation of the load level is slower than the variation of the measured power, so that the function f(SoC) is more stable than the function f(Pmes).

[0077] Furthermore, this second law for controlling the frequency setpoint value as a function of the battery charge level must protect the microgrid from blackouts when the measured current (for an absorbed power) exceeds the dynamic limit imposed by the battery management system, BMS (for the English "Battery Management System"). A larger SoC variation range ensures greater stability, but the gain is not linear between the stability and the width of the variation range; in addition, it is necessary to limit the loss of renewable energy for low SoC charge levels. Poor management of shaving of the power injected into the microgrid by the ENRv production modules for low SoC charge levels can greatly reduce the islanding time: in fact, the battery can then discharge more quickly, reaching a SoC charge level = 0%, which results in a blackout of the microgrid 11.

[0078] The values of the parameters SoCfin, SoCstart and ASoChysteresis can be adapted according to the configuration of the islanded micro-grid, and the compromise sought between stability of the system and use of the power produced by the ENRv PV124 production modules. SoCstart can vary within a range substantially between between 0 and 60%; similarly SoCfin can vary in a range substantially between 60% and 100%. Finally, the value of ASoC^ysteresis Pcut can be chosen in a range substantially between 2% and 15% for example.

[0079] The second evolution law described above in relation to [Fig.5] controls a frequency setpoint value which can be noted fhysteresis(SoC), which feeds a second stage of construction of the frequency control law as a function of the battery charge level, presented below in relation to [Fig.6].

[0080] In [Fig.6], a fictitious variation of the frequency fhysteresis(SoC) as a function of time has been represented. The processing by decoupling levels of the curve representing this variation ensures stability by a temporary blocking of the minimum value of the frequency, if the latter increases and exceeds a decoupling frequency threshold of a type 1, type 2, or type 3 inverter. In [Fig.6], fSeuiidécoupiage i=50.2Hz, which is the decoupling frequency of a type 1 inverter, fSeuiidécoupiage2=50.4Hz, which is the decoupling frequency of a type 2 inverter, and f seuiidécouPiage3=50.6Hz, which is the decoupling frequency of a type 3 inverter, has been represented.

[0081] When the frequency setpoint value imposed by the second evolution law f hysteresis (SoC) of figure 5 increases and reaches a first blocking threshold frequency equal to decoupling 1 f margin' threshold decoupling 2 f margin OR f ., „ - At f , .we prohibit the frequency setpoint value from taking J decoupling threshold 3 J margin' & ir a value lower than a second blocking threshold frequency, which is higher than the first blocking threshold frequency by a value equal to 2 A fmar^e = 0.1Hz. Thus, the minimum frequency setpoint is blocked at fseuiidcoupi+0.05Hz, fseuiidcoupiage 2+0.05Hz or fseuiidcoupiage 3+0.05Hz. This blocking is lifted as soon as the frequency setpoint increases beyond the second blocking threshold frequency fseuiidcoupi+0.05Hz, fSeuiidcoupiage2+0.05Hz or fseuiidcoupiage3+0.05Hz, or as soon as it decreases below a third blocking threshold frequency lower than the first blocking threshold frequency, namely f .... . , - 2 A f ,f .... , ,-2Af or f , , 3-2Af J decoupling threshold 3 J margin

[0082] Such blocking of the frequency setpoint value is useful for inverters of type 1, 2 or 3 for which the P(f) function is not activated, and whose decoupling frequency is fixed at 50.2, 50.4 or 50.6 Hz.

[0083] At the output of this second stage of construction of the second control law, we obtain a frequency setpoint value corresponding to the second control law, which we denote by f(SoC). It feeds, with the frequency setpoint value corresponding to the first control law, f(Pmes), a third stage of construction of the global control law, the operating principle of which is illustrated in figure 7. This third stage is a junction stage of the sub-functions f(SoC) and f(Pmes): it receives as input variables a function representing the evolution as a function of time of the sub-function f(Pmes), and a function representing the evolution as a function of time of the sub-function f(SoC). It delivers as output a function f (P, SoC) - max( / (p), f(SoC) ) which is the maximum function of the two input variables. Thus, at each instant, the most restrictive law is imposed: in fact, a higher frequency setpoint is calculated for a more restrictive case than a lower frequency setpoint, both for the variation of the frequency as a function of the measured power, f(Pmes), and for the variation of the frequency as a function of the battery charge level, f(SoC). The use of the maximum function to join the two sub-functions f(SoC) and f(Pmes) is possible because the frequency variation range is the same for these two sub-functions.

[0084] A fourth stage of construction of the global control law is supplied by the function f(P, SoC ) and delivers as output the set value of the frequency imposed on the 10i battery of the battery storage system, denoted fbatt. This value fbatt is calculated by applying limitations to the minimum and maximum values of the frequency, as well as its variation speed, as illustrated in Figure 8. Indeed, the limitation of the setpoint value of the calculated frequency f(P, SoC) is necessary to frame the operation of the battery storage system within strict limits that are globally valid for the adjustment of the P(f) function valid for the inverter side of the storage system.

[0085] In one embodiment, four limits are defined: - a minimum value of the frequency, fmin=50Hz, corresponding to the value of the nominal frequency in the electrical distribution network; - a maximum value of the frequency, fmax=51.55Hz, defined as follows: f = max ( 1 , ) + A f = 4 + A f J max decoupling JJ margin J decoupling J margin - a limit value for the frequency variation ramp on the rise _ 2 • This value is chosen because it corresponds to a usual value of variation of the frequency of an inverter or a generator (rotating group of type genset). It ensures a sufficiently rapid increase in frequency to quickly stop variations in the power injected by the modules. ENRv production, while avoiding overly rapid variations of the “jump” type frequency " ; - a limit value for the frequency variation ramp on the way down dt -0.000125 [Hz / s]^ueVon fixed at a low value to allow a slow reaction of the frequency evolution as a function of the measured power f(Pmes ), to avoid instabilities.

[0086] These values are given as an example of implementation. Other values can also be chosen, to adapt to the particular configuration of the islanded micro-grid, and in particular to the installed power of the ENRv production modules present on the network, and to the type of inverters they integrate.

[0087] In particular, the choice of the limit value of the frequency variation ramp during its decrease makes it possible to cover the entire power variation range in more than two hours. This choice therefore makes it possible to ensure a sufficiently slow variation of the frequency during its decrease to protect the islanded micro-grid against “yoyo” type instability, even for inverters of ENRv production modules with an installed power greater than the installed power of the storage system. It therefore makes it possible to size the battery storage system for the most critical case, since this system must be able to be used in any type of islanded micro-grid, whatever its configuration.However, for a battery storage system that would be dedicated to use in an islanded micro-grid in which the installed power of the ENRv modules is low, and below the installed power of the storage system, this limit value of the frequency decrease ramp can be reduced.

[0088] The different modules of the control logic implemented in a controller of a battery storage system according to one embodiment are summarized in synoptic form in [Fig.9].

[0089] The modules referenced 101 and 102, arranged in series, are dedicated to the first law of control of the frequency setpoint value as a function of the power measured at the terminals of the battery of the energy storage system. As illustrated in relation to [Fig.3], the module referenced 101 controls the first law of evolution of the frequency setpoint value as a function of the measured power, according to a hysteresis cycle; it delivers a sub-function fhysteréSiS(PmeS)- H is particularly dedicated to the control of the power injected by the ENRv inverters of type 4, having an activated P(f) function. The module referenced 102, arranged in series with the first module referenced 101, receives as input the sub-function fhysteréSiS(PmeS) delivered by the latter, to which it applies processing by decoupling levels, to control the control of the power injected by the ENRv inverters of type 1, 2 or 3 not having an activated P(f) function.It outputs a subfunction denoted f(P).

[0090] The modules referenced 201 and 202, also arranged in series, are dedicated to the second law of control of the set value of the frequency as a function of the charge level of the battery of the energy storage system. As illustrated in relation to [Fig.6], the module referenced 201 controls the second evolution law of the frequency setpoint value as a function of the load level, according to a hysteresis cycle; it delivers a sub-function fhysteresis(SoC). It is particularly dedicated to the control of the power injected by ENRv inverters of type 4, having an activated P(f) function. The module referenced 202, arranged in series with the first module referenced 201, receives as input the sub-function fhysteresis(SoC) delivered by the latter, to which it applies a processing by decoupling levels, to control the control of the power injected by ENRv inverters of type 1, 2 or 3 not having an activated P(f) function. It delivers as output a sub-function noted f(SoC).

[0091] The third stage referenced 300 of the control logic of [Fig.9] performs a junction of the sub-functions f(P) and f(SoC) by means of a maximum function of these two input variables, and delivers as output a function f(P, Soc) which feeds the fourth stage referenced 400. The latter applies limitations in value and in derivative to the frequency setpoint value f(P,SoC), which it transforms into a setpoint f batt which controls the frequency of the inverter of the battery electrical energy storage system.

[0092] The processing of modules 101, 102 on the one hand and 201, 202 on the other hand, are carried out in parallel. It will be noted that for a battery storage system intended to be used in an islanded micro-grid comprising only ENRv inverters of type 4 or later (i.e. only ENRv inverters having a power regulation function as a function of the activated frequency P(f)), the implementation of the processing modules 102, 202 by decoupling levels is optional.

[0093] Figure 10 provides a block diagram of an electrical battery storage system according to one embodiment. As indicated above, such a system comprises one or more BATT 10i batteries for storing electrical energy, powering a DC / AC inverter 102 configured to supply an alternating voltage at the frequency fbatt to the electrical network to which it is connected. It also comprises one or more PROC processor(s) capable of controlling the execution and reading of the measurement of the charge level of the BATT battery, as well as the measurement of the power absorbed at the terminals of the storage system. It also comprises a memory MEM storing at least instructions of a computer program and accessible by the PROC processor(s) to implement the method described above in relation to Figures 2 to 9 when the PROC processor(s) of the storage system execute(s) the instructions of the program.In particular, the processor(s) PROC calculate(s) the setpoint value of the frequency fbatt at the output of the DC / AC inverter 102, as a function of the load level SoC and the measured power Pmes, from the first and second control laws described above. Such a system also comprises an interface INT, allowing an operator to configure the different variables of the first and second evolution laws, stored in the memory MEM,. such as the values of the parameters Pstart, Pfin, APhysteresis, SoCstart, SoCfin, ASoQysteresis, ffin, fstart the blocking times of the minimum frequency value for processing by decoupling levels, etc.; as described above in relation to figures 3 to 9. Industrial application

[0094] The subject of the present description may find applications in the upkeep, management and maintenance of electrical distribution networks to which variable renewable electrical energy production modules are connected, in particular for the purpose of optimizing the share of green energy in the network. List of cited documents Patent documents

[0095] For all useful purposes, the following patent document is cited: - patcitl: US2019190274A1 (publication number).

Claims

Claims

1. Method for controlling power injected by at least one renewable energy production module (PVb PV2, ...PVn) into an islanded electrical microgrid (11) relative to a main electrical distribution network, characterized in that it comprises: a. a connection, in said islanded electrical micro-grid, of an electrical energy storage system comprising at least one electrical energy storage module (BATT, 10i) and at least one DC / AC inverter (102), said electrical energy storage module (10i) being capable of delivering a direct voltage to the terminals of a DC / AC inverter (102), said DC / AC inverter generating an alternating voltage capable of supplying said islanded electrical micro-grid; b. a control of a frequency setpoint value of said alternating voltage generated by said inverter from: i. a first control law of said frequency setpoint value established from a first evolution law (101) of said frequency setpoint value as a function of an absorbed power measured at the terminals of said electrical energy storage system, said first evolution law having a hysteresis cycle; ii. a second control law of said frequency setpoint value established from a second evolution law (201) of said frequency setpoint value as a function of a charge level of said electrical energy storage module, said second evolution law having a hysteresis cycle, and in that said control of said frequency setpoint value is controlled (300) by the maximum function of a variation in time of said frequency setpoint value according to said first control law and of a variation in time of said frequency setpoint value according to said second control law.

2. Power control method according to claim 1, characterized in that said hysteresis cycle of said first evolution law is of parallelepipedal type and comprises: - a straight line of increase of said frequency setpoint value as a function of an increase of said absorbed power measured from a starting power value, Pstart and up to an end power value, Pfin; - a straight line of decrease of said frequency setpoint value as a function of a decrease of said measured absorbed power; and in that said increase and decrease straight lines are parallel and offset relative to each other on the abscissa by a power hysteresis value A Physteresis-

3. Power control method according to claim 1, characterized in that said first control law also comprises, during a determined blocking duration starting when said frequency setpoint value increases and reaches a first blocking threshold frequency according to said first evolution law, a blocking (102) of a minimum value of said frequency setpoint value at a second blocking threshold frequency greater than said first blocking threshold frequency.

4. Power control method according to any one of claims 1 to 3, characterized in that said hysteresis cycle of said second evolution law is of parallelepiped type and comprises: - a straight line of increase of said frequency setpoint value as a function of an increase of said charge level of said storage module from a starting charge level value, SoCstart and up to an end charge level value, SoCfin; - a straight line of decrease of said frequency setpoint value as a function of a decrease of said charge level of said storage module; and in that said increase and decrease straight lines are parallel and offset relative to each other on the abscissa by a charge level hysteresis value A SoC.

5. Power control method according to any one of claims 1 to 4 characterized in that said second control law also comprises, when said frequency setpoint value increases and reaches a first blocking threshold frequency according to said second evolution law, a blocking (202) of a minimum value of said value

6.

7.

8. frequency setpoint at a second blocking threshold frequency higher than said first blocking threshold frequency until said frequency setpoint value increases and reaches said first blocking threshold frequency or until said frequency setpoint value decreases and reaches a third blocking threshold frequency lower than said first blocking threshold frequency, according to said second evolution law. Power control method according to claims 3 and 5, characterized in that said first, second and third blocking threshold frequencies are determined as a function of a decoupling frequency set for an inverter of said at least one renewable energy production module. Power control method according to any one of claims 1 to 6, characterized in that said control of said frequency setpoint value also comprises an additional application (400), to said maximum function: - of a maximum frequency setpoint value; - a minimum frequency setpoint value; - a limit value for the slope of increase of said frequency setpoint value as a function of time; - a limit value of the slope of decrease of said frequency setpoint value as a function of time; and in that said increase slope limit value is greater than said decrease slope limit value, in absolute values. Mobile electrical energy storage system capable of being connected to an islanded electrical microgrid (11) with respect to a main electrical distribution network, said storage system comprising at least one electrical energy storage module (BATT, 10x) and at least one DC / AC inverter (102), said electrical energy storage module being capable of delivering a direct voltage to the terminals of said DC / AC inverter, said DC / AC inverter generating an alternating voltage capable of supplying said islanded electrical microgrid, characterized in that it comprises a processor (PROC) configured to execute a control of a frequency setpoint value (fbatt) of said alternating voltage generated by said inverter from: i. of a first control law of said setpoint value of frequency established from a first law of evolution of said frequency setpoint value as a function of an absorbed power measured at the terminals of said storage system, said first law of evolution having a hysteresis cycle; ii. of a second law of control of said frequency setpoint value established from a second law of evolution of said frequency setpoint value as a function of a charge level of said electrical energy storage module (BATT, 10i), said second law of evolution having a hysteresis cycle, and in that said control of said frequency setpoint value is controlled (300) by the maximum function of a variation in time of said frequency setpoint value according to said first control law and of a variation in time of said frequency setpoint value according to said second control law.

9. Computer program comprising instructions for implementing the method according to one of claims 1 to 7 when this program is executed by a processor of an electrical energy storage system according to claim 8.

10. Non-transitory recording medium readable by a computer on which is recorded a program for implementing the method according to one of claims 1 to 7 when this program is executed by a processor of an electrical energy storage system according to claim 8.

Images